I have proven myself wrong

I keep working on a proof-of-concept paper for the idea I baptized ‘Energy Ponds’. You can consult two previous updates, namely ‘We keep going until we observe’ and ‘Ça semble expérimenter toujours’ to keep track of the intellectual drift I am taking. This time, I am focusing on the end of the technological pipeline, namely on the battery-powered charging station for electric cars. First, I want to make myself an idea of the market for charging.

I take the case of France. In December 2020, they had a total of 119 737 electric vehicles officially registered (matriculated), which made + 135% as compared to December 2019[1]. That number pertains only to 100% electrical ones, with plug-in hybrids left aside for the moment. When plug-in hybrids enter the game, France had, in December 2020, 470 295 vehicles that need or might need the services of charging stations. According to the same source, there were 28 928 charging stations in France at the time, which makes 13 EVs per charging station. That coefficient is presented for 4 other European countries: Norway (23 EVs per charging station), UK (12), Germany (9), and Netherlands (4).

I look up into other sources. According to Reuters[2], there was 250 000 charging stations in Europe by September 2020, as compared to 34 000 in 2014. That means an average increase by 36 000 a year. I find a different estimation with Statista[3]: 2010 – 3 201; 2011 – 7 018; 2012 – 17 498; 2013 – 28 824; 2014 – 40 910; 2015 – 67 064; 2016 – 98 669; 2017 – 136 059; 2018 – 153 841; 2019 – 211 438; 2020 – 285 796.

On the other hand, the European Alternative Fuels Observatory supplies their own data at https://www.eafo.eu/electric-vehicle-charging-infrastructure, as regards European Union.

Number of EVs per charging station (source: European Alternative Fuels Observatory):

EVs per charging station

The same EAFO site gives their own estimation as regards the number of charging stations in Europe:

Number of charging stations in Europe (source: European Alternative Fuels Observatory):

High-power recharging points (more than 22 kW) in EUNormal charging stations in EUTotal charging stations
201225710 25010 507
201375117 09317 844
20141 47424 91726 391
20153 39644 78648 182
20165 19070 01275 202
20178 72397 287106 010
201811 138107 446118 584
201915 136148 880164 016
202024 987199 250224 237

Two conclusions jump to the eye. Firstly, there is just a very approximate count of charging stations. Numbers differ substantially from source to source. I can just guess that one of the reasons for that discrepancy is the distinction between officially issued permits to build charging points, on the one hand, and the actually active charging points, on the other hand. In Europe, building charging points for electric vehicles has become sort of a virtue, which governments at all levels like signaling. I guess there is some boasting and chest-puffing in the numbers those individual countries report.  

Secondly, high-power stations, charging with direct current, with a power of at least 22 kWh,  gain in importance. In 2012, that category made 2,45% of the total charging network in Europe, and in 2020 that share climbed to 11,14%. This is an important piece of information as regards the proof-of-concept which I am building up for my idea of Energy Ponds. The charging station I placed at the end of the pipeline in the concept of Energy Ponds, and which is supposed to earn a living for all the technologies and installations upstream of it, is supposed to be powered from a power storage facility. That means direct current, and most likely, high power.   

On the whole, the www.eafo.eu site seems somehow more credible that Statista, with all the due respect for the latter, and thus I am reporting some data they present on the fleet of EVs in Europe. Here it comes, in a few consecutive tables below:

Passenger EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
20084 1554 155
20094 8414 841
20105 7855 785
201113 39516313 558
201225 8913 71229 603
201345 66232 47478 136
201475 47956 745132 224
2015119 618125 770245 388
2016165 137189 153354 290
2017245 347254 473499 820
2018376 398349 616726 014
2019615 878479 7061 095 584
20201 125 485967 7212 093 206

Light Commercial EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
20117 6697 669
20129 5279 527
201313 66913 669
201410 04910 049
201528 61028 610
201640 926140 927
201752 026152 027
201876 286176 287
201997 36311797 480
2020120 7111 054121 765

Bus EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
20178884451 333
20181 6084862 094
20193 6365254 161
20205 3115505 861

Truck EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
2020983291 012

Structure of EV fleet in Europe as regards the types of vehicles (source: European Alternative Fuels Observatory):

Passenger EVLight commercial EVBus EVTruck EV

Summing it up a bit. The market of Electric Vehicles in Europe seems being durably dominated by passenger cars. There is some fleet in other categories of vehicles, and there is even some increase, but, for the moment, in all looks more like an experiment. Well, maybe electric buses turn up sort of more systemically.

The proportion between the fleet of electric vehicles and the infrastructure of charging stations still seems to be in the phase of adjustment in the latter to the abundance of the former. Generally, the number of charging stations seems to be growing slower than the fleet of EVs. Thus, for my own concept, I assume that the coefficient of 9 EVs per charging station, on average, will stand still or will slightly increase. For the moment, I take 9. I assume that my charging stations will have like 9 habitual customers, plus a fringe of incidental ones.

From there, I think in the following terms. The number of times the average customer charges their car depends on the distance they cover. Apparently, there is like a 100 km  50 kWh equivalence. I did not find detailed statistics as regards distances covered by electric vehicles as such, however I came by some Eurostat data on distances covered by all passenger vehicles taken together: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Passenger_mobility_statistics#Distance_covered . There is a lot of discrepancy between the 11 European countries studied for that metric, but the average is 12,49 km per day. My average 9 customers would do, in total, an average of 410,27 of 50 kWh charging purchases per year. I checked the prices of fast charging with direct current: 2,3 PLN per 1 kWh in Poland[4],  €0,22 per 1 kWh in France[5], $0,13 per 1 kWh in US[6], 0,25 pence per 1 kWh in UK[7]. Once converted to US$, it gives $0,59 in Poland, $0,26 in France, $0,35 in UK, and, of course, $0,13 in US. Even at the highest price, namely that in Poland, those 410,27 charging stops give barely more than $12 000 a year.

If I want to have a station able to charge 2 EVs at the same time, fast charging, and counting 350 kW per charging pile (McKinsey 2018[8]), I need 700 kW it total. Investment in batteries is like $600 ÷ $800 per 1 kW (Cole & Frazier 2019[9]; Cole, Frazier, Augustine 2021[10]), thus 700 * ($600 ÷ $800) = $420 000 ÷ $560 000. There is no way that investment pays back with $12 000 a year in revenue, and I haven’t even started talking about paying off on investment in all the remaining infrastructure of Energy Ponds: ram pumps, elevated tanks, semi-artificial wetlands, and hydroelectric turbines.

Now, I revert my thinking. Investment in the range of $420 000 ÷ $560 000, in the charging station and its batteries, gives a middle-of-the-interval value of $490 000. I found a paper by Zhang et al. (2018[11]) who claim that a charging station has chances to pay off, as a business, when it sells some 5 000 000 kWh a year. When I put it back-to-back with the [50 kWh / 100 km] coefficient, it gives 10 000 000 km. Divided by the average annual distance covered by European drivers, thus by 4 558,55 km, it gives 2 193,68 customers per year, or some 6 charging stops per day. That seems hardly feasible with 9 customers. I assume that one customer would charge their electric vehicle no more than twice a week, and 6 chargings a day make 6*7 = 42 chargings, and therefore 21 customers.

I need to stop and think. Essentially, I have proven myself wrong. I had been assuming that putting a charging station for electric vehicles at the end of the internal value chain in the overall infrastructure of Energy Ponds will solve the problem of making money on selling electricity. Turns out it makes even more problems. I need time to wrap my mind around it.

[1] http://www.avere-france.org/Uploads/Documents/161011498173a9d7b7d55aef7bdda9008a7e50cb38-barometre-des-immatriculations-decembre-2020(9).pdf

[2] https://www.reuters.com/article/us-eu-autos-electric-charging-idUSKBN2C023C

[3] https://www.statista.com/statistics/955443/number-of-electric-vehicle-charging-stations-in-europe/

[4] https://elo.city/news/ile-kosztuje-ladowanie-samochodu-elektrycznego

[5] https://particulier.edf.fr/fr/accueil/guide-energie/electricite/cout-recharge-voiture-electrique.html

[6] https://afdc.energy.gov/fuels/electricity_charging_home.html

[7] https://pod-point.com/guides/driver/cost-of-charging-electric-car

[8] McKinsey Center for Future Mobility, How Battery Storage Can Help Charge the Electric-Vehicle Market?, February 2018,

[9] Cole, Wesley, and A. Will Frazier. 2019. Cost Projections for Utility-Scale Battery Storage.

Golden, CO: National Renewable Energy Laboratory. NREL/TP-6A20-73222. https://www.nrel.gov/docs/fy19osti/73222.pdf

[10] Cole, Wesley, A. Will Frazier, and Chad Augustine. 2021. Cost Projections for UtilityScale Battery Storage: 2021 Update. Golden, CO: National Renewable Energy

Laboratory. NREL/TP-6A20-79236. https://www.nrel.gov/docs/fy21osti/79236.pdf.

[11] Zhang, J., Liu, C., Yuan, R., Li, T., Li, K., Li, B., … & Jiang, Z. (2019). Design scheme for fast charging station for electric vehicles with distributed photovoltaic power generation. Global Energy Interconnection, 2(2), 150-159. https://doi.org/10.1016/j.gloei.2019.07.003

Ça semble expérimenter toujours

Je continue avec l’idée que j’avais baptisée « Projet Aqueduc ». Je suis en train de préparer un article sur ce sujet, du type « démonstration de faisabilité ». Je le prépare en anglais et je me suis dit que c’est une bonne idée de reformuler en français ce que j’ai écrit jusqu’à maintenant, l’histoire de changer l’angle intellectuel, me dégourdir un peu et prendre de la distance.

Une démonstration de faisabilité suit une logique similaire à tout autre article scientifique, sauf qu’au lieu d’explorer et vérifier une hypothèse théorique du type « les choses marchent de façon ABCD, sous conditions RTYU », j’explore et vérifie l’hypothèse qu’un concept pratique, comme celui du « Projet Aqueduc », a des fondements scientifiques suffisamment solides pour que ça vaille la peine de travailler dessus et de le tester en vie réelle. Les fondements scientifiques viennent en deux couches, en quelque sorte. La couche de base consiste à passer en revue la littérature du sujet pour voir si quelqu’un a déjà décrit des solutions similaires et là, le truc c’est explorer des différentes perspectives de similarité. Similaire ne veut pas dire identique, n’est-ce pas ? Cette revue de littérature doit apporter une structure logique – un modèle – applicable à la recherche empirique, avec des variables et des paramètres constants. C’est alors que vient la couche supérieure de démonstration de faisabilité, qui consiste à conduire de la recherche empirique proprement dite avec ce modèle.    

Moi, pour le moment, j’en suis à la couche de base. Je passe donc en revue la littérature pertinente aux solutions hydrologiques et hydroélectriques, tout en formant, progressivement, un modèle numérique du « Projet Aqueduc ». Dans cette mise à jour, je commence par une brève récapitulation du concept et j’enchaîne avec ce que j’ai réussi à trouver dans la littérature. Le concept de base du « Projet Aqueduc » consiste donc à placer dans le cours d’une rivière des pompes qui travaillent selon le principe du bélier hydraulique et qui donc utilisent l’énergie cinétique de l’eau pour pomper une partie de cette eau en dehors du lit de la rivière, vers des structures marécageuses qui ont pour fonction de retenir l’eau dans l’écosystème local. Le bélier hydraulique à la capacité de pomper à la verticale aussi bien qu’à l’horizontale et donc avant d’être retenue dans les marécages, l’eau passe par une structure similaire à un aqueduc élevé (d’où le nom du concept en français), avec des réservoirs d’égalisation de flux, et ensuite elle descend vers les marécages à travers des turbines hydroélectriques. Ces dernières produisent de l’énergie qui est ensuite emmagasinée dans une installation de stockage et de là, elle est vendue pour assurer la survie financière à la structure entière. On peut ajouter des installations éoliennes et/ou photovoltaïques pour optimiser la production de l’énergie sur le terrain occupé par la structure entière.  Vous pouvez trouver une description plus élaborée du concept dans ma mise à jour intitulée « Le Catch 22 dans ce jardin d’Eden ». La faisabilité dont je veux faire une démonstration c’est la capacité de cette structure à se financer entièrement sur la base des ventes d’électricité, comme un business régulier, donc de se développer et durer sans subventions publiques. La solution pratique que je prends en compte très sérieusement en termes de créneau de vente d’électricité est une station de chargement des véhicules électriques.   

L’approche de base que j’utilise dans la démonstration de faisabilité – donc mon modèle de base – consiste à représenter le concept en question comme une chaîne des technologies :

>> TCES – stockage d’énergie

>> TCCS – station de chargement des véhicules électriques

>> TCRP – pompage en bélier hydraulique

>> TCEW – réservoirs élevés d’égalisation

>> TCCW – acheminement et siphonage d’eau

>> TCWS – l’équipement artificiel des structures marécageuses

>> TCHE – les turbines hydroélectriques

>> TCSW – installations éoliennes et photovoltaïques     

Mon intuition de départ, que j’ai l’intention de vérifier dans ma recherche à travers la littérature, est que certaines de ces technologies sont plutôt prévisibles et bien calibrées, pendant qu’il y en a d’autres qui sont plus floues et sujettes au changement, donc moins prévisibles. Les technologies prévisibles sont une sorte d’ancrage pour the concept entier et celles plus floues sont l’objet d’expérimentation.

Je commence la revue de littérature par le contexte environnemental, donc avec l’hydrologie. Les variations au niveau de la nappe phréatiques, qui est un terme scientifique pour les eaux souterraines, semblent être le facteur numéro 1 des anomalies au niveau de rétention d’eau dans les réservoirs artificiels (Neves, Nunes, & Monteiro 2020[1]). D’autre part, même sans modélisation hydrologique détaillée, il y a des preuves empiriques substantielles que la taille des réservoirs naturels et artificiels dans les plaines fluviales, ainsi que la densité de placement de ces réservoirs et ma manière de les exploiter ont une influence majeure sur l’accès pratique à l’eau dans les écosystèmes locaux. Il semble que la taille et la densité des espaces boisés intervient comme un facteur d’égalisation dans l’influence environnementale des réservoirs (Chisola, Van der Laan, & Bristow 2020[2]). Par comparaison aux autres types de technologie, l’hydrologie semble être un peu en arrière en termes de rythme d’innovation et il semble aussi que des méthodes de gestion d’innovation appliquées ailleurs avec succès peuvent marcher pour l’hydrologie, par exemple des réseaux d’innovation ou des incubateurs des technologies (Wehn & Montalvo 2018[3]; Mvulirwenande & Wehn 2020[4]). L’hydrologie rurale et agriculturale semble être plus innovatrice que l’hydrologie urbaine, par ailleurs (Wong, Rogers & Brown 2020[5]).

Ce que je trouve assez surprenant est le manque apparent de consensus scientifique à propos de la quantité d’eau dont les sociétés humaines ont besoin. Toute évaluation à ce sujet commence avec « beaucoup et certainement trop » et à partir de là, le beaucoup et le trop deviennent plutôt flous. J’ai trouvé un seul calcul, pour le moment, chez Hogeboom (2020[6]), qui maintient que la personne moyenne dans les pays développés consomme 3800 litres d’eau par jour au total, mais c’est une estimation très holistique qui inclue la consommation indirecte à travers les biens et les services ainsi que le transport. Ce qui est consommé directement via le robinet et la chasse d’eau dans les toilettes, ça reste un mystère pour la science, apparemment, à moins que la science ne considère ce sujet comment trop terre-à-terre pour s’en occuper sérieusement.     

Il y a un créneau de recherche intéressant, que certains de ses représentants appellent « la socio-hydrologie », qui étudie les comportements collectifs vis-à-vis de l’eau et des systèmes hydrologiques et qui est basée sur l’observation empirique que lesdits comportements collectifs s’adaptent, d’une façon profonde et pernicieuse à la fois, aux conditions hydrologiques que la société en question vit avec (Kumar et al. 2020[7]). Il semble que nous nous adaptons collectivement à la consommation accrue de l’eau par une productivité croissante dans l’exploitation de nos ressources hydrologiques et le revenu moyen par tête d’habitant semble être positivement corrélé avec cette productivité (Bagstad et al. 2020[8]). Il paraît donc que l’accumulation et superposition de nombreuses technologies, caractéristique aux pays développés, contribue à utiliser l’eau de façon de plus en plus productive. Dans ce contexte, il y a une recherche intéressante conduite par Mohamed et al. (2020[9]) qui avance la thèse qu’un environnement aride est non seulement un état hydrologique mais aussi une façon de gérer les ressources hydrologiques, sur ma base des données qui sont toujours incomplètes par rapport à une situation qui change rapidement.

Il y a une question qui vient plus ou moins naturellement : dans la foulée de l’adaptation socio-hydrologique quelqu’un a-t-il présenté un concept similaire à ce que moi je présente comme « Projet Aqueduc » ? Eh bien, je n’ai rien trouvé d’identique, néanmoins il y a des idées intéressement proches. Dans l’hydrologie descriptive il y a ce concept de pseudo-réservoir, qui veut dire une structure comme les marécages ou des nappes phréatiques peu profondes qui ne retiennent pas l’eau de façons statique, comme un lac artificiel, mais qui ralentissent la circulation de l’eau dans le bassin fluvial d’une rivière suffisamment pour modifier les conditions hydrologiques dans l’écosystème (Harvey et al. 2009[10]; Phiri et al. 2021[11]). D’autre part, il y a une équipe des chercheurs australiens qui ont inventé une structure qu’ils appellent par l’acronyme STORES et dont le nom complet est « short-term off-river energy storage » (Lu et al. 2021[12]; Stocks et al. 2021[13]). STORES est une structure semi-artificielle d’accumulation par pompage, où on bâtit un réservoir artificiel au sommet d’un monticule naturel placé à une certaine distance de la rivière la plus proche et ce réservoir reçoit l’eau pompée artificiellement de la rivière. Ces chercheurs australiens avancent et donnent des preuves scientifiques pour appuyer la thèse qu’avec un peu d’astuce on peut faire fonctionner ce réservoir naturel en boucle fermée avec la rivière qui l’alimente et donc de créer un système de rétention d’eau. STORES semble être relativement le plus près de mon concept de « Projet Aqueduc » et ce qui est épatant est que moi, j’avais inventé mon idée pour l’environnement des plaines alluviales de l’Europe tandis que STORES avait été mis au point pour l’environnement aride et quasi-désertique d’Australie. Enfin, il y a l’idée des soi-disant « jardins de pluie » qui sont une technologie de rétention d’eau de pluie dans l’environnement urbain, dans des structures horticulturales, souvent placées sur les toits d’immeubles (Bortolini & Zanin 2019[14], par exemple).

Je peux conclure provisoirement que tout ce qui touche à l’hydrologie strictement dite dans le cadre du « Projet Aqueduc » est sujet aux changements plutôt imprévisible. Ce que j’ai pu déduire de la littérature ressemble à un potage bouillant sous couvercle. Il y a du potentiel pour changement technologique, il y a de la pression environnementale et sociale, mais il n’y pas encore de mécanismes institutionnels récurrents pour connecter l’un à l’autre. Les technologies TCEW (réservoirs élevés d’égalisation), TCCW (acheminement et siphonage d’eau), et TCWS (l’équipement artificiel des structures marécageuses) démontrant donc un avenir flou, je passe à la technologie TCRP de pompage en bélier hydraulique. J’ai trouvé deux articles chinois, qui se suivent chronologiquement et qui semblent par ailleurs avoir été écrits par la même équipe de chercheurs : Guo et al. (2018[15]), and Li et al. (2021[16]). Ils montrent la technologie du bélier hydraulique sous un angle intéressant. D’une part, les Chinois semblent avoir donné du vrai élan à l’innovation dans ce domaine spécifique, tout au moins beaucoup plus d’élan que j’ai pu observer en Europe. D’autre part, les estimations de la hauteur effective à laquelle l’eau peut être pompée avec les béliers hydrauliques dernier cri sont respectivement de 50 mètres dans l’article de 2018 et 30 mètres dans celui de 2021. Vu que les deux articles semblent être le fruit du même projet, il y a eu comme une fascination suivie par une correction vers le bas. Quoi qu’il en soit, même l’estimation plus conservative de 30 mètres c’est nettement mieux que les 20 mètres que j’assumais jusqu’à maintenant.

Cette élévation relative possible à atteindre avec la technologie du bélier hydraulique est importante pour la technologie suivante de ma chaîne, donc celle des petites turbines hydroélectriques, la TCHE. L’élévation relative de l’eau et le flux par seconde sont les deux paramètres clés qui déterminent la puissance électrique produite (Cai, Ye & Gholinia 2020[17]) et il se trouve que dans le « Projet Aqueduc », avec l’élévation et le flux largement contrôlés à travers la technologie du bélier hydraulique, les turbines deviennent un peu moins dépendantes sur les conditions naturelles.

J’ai trouvé une revue merveilleusement encyclopédique des paramètres pertinents aux petites turbines hydroélectriques chez Hatata, El-Saadawi, & Saad (2019[18]). La puissance électrique se calcule donc comme : Puissance = densité de l’eau (1000 kg/m3) * constante d’accélération gravitationnelle (9,8 m/s2) * élévation nette (mètres) * Q (flux par seconde m3/s).

L’investissement initial en de telles installations se calcule par unité de puissance, donc sur la base de 1 kilowatt et se divise en 6 catégories : la construction de la prise d’eau, la centrale électrique strictement dite, les turbines, le générateur, l’équipement auxiliaire, le transformateur et enfin le poste extérieur. Je me dis par ailleurs que – vu la structure du « Projet Aqueduc » – l’investissement en la construction de prise d’eau est en quelque sorte équivalent au système des béliers hydrauliques et réservoirs élevés. En tout cas :

>> la construction de la prise d’eau, par 1 kW de puissance  ($) 186,216 * Puissance-0,2368 * Élévation -0,597

>> la centrale électrique strictement dite, par 1 kW de puissance  ($) 1389,16 * Puissance-0,2351 * Élévation-0,0585

>> les turbines, par 1 kW de puissance  ($)

@ la turbine Kaplan: 39398 * Puissance-0,58338 * Élévation-0,113901

@ la turbine Frances: 30462 * Puissance-0,560135 * Élévation-0,127243

@ la turbine à impulsions radiales: 10486,65 * Puissance-0,3644725 * Élévation-0,281735

@ la turbine Pelton: 2 * la turbine à impulsions radiales

>> le générateur, par 1 kW de puissance  ($) 1179,86 * Puissance-0,1855 * Élévation-0,2083

>> l’équipement auxiliaire, par 1 kW de puissance  ($) 612,87 * Puissance-0,1892 * Élévation-0,2118

>> le transformateur et le poste extérieur, par 1 kW de puissance 

($) 281 * Puissance0,1803 * Élévation-0,2075

Une fois la puissance électrique calculée avec le paramètre d’élévation relative assurée par les béliers hydrauliques, je peux calculer l’investissement initial en hydro-génération comme la somme des positions mentionnées ci-dessus. Hatata, El-Saadawi, & Saad (2019 op. cit.) recommandent aussi de multiplier une telle somme par le facteur de 1,13 (c’est donc un facteur du type « on ne sait jamais ») et d’assumer que les frais courants d’exploitation annuelle vont se situer entre 1% et 6% de l’investissement initial.

Syahputra & Soesanti (2021[19]) étudient le cas de la rivière Progo, dotée d’un flux tout à fait modeste de 6,696 mètres cubes par seconde et située dans Kulon Progo Regency (une region spéciale au sein de Yogyakarta, Indonesia). Le système des petites turbines hydroélectriques y fournit l’électricité aux 962 ménages locaux, et crée un surplus de 4 263 951 kWh par an d’énergie à revendre aux consommateurs externes. Dans un autre article, Sterl et al. (2020[20]) étudient le cas de Suriname et avancent une thèse intéressante, notamment que le développement d’installations basées sur les énergies renouvelables crée un phénomène d’appétit d’énergie qui croît à mesure de manger et qu’un tel développement en une source d’énergie – le vent, par exemple – stimule l’investissement en installations basées sur d’autres sources, donc l’hydraulique et le photovoltaïque.  

Ces études relativement récentes corroborent celles d’il y a quelques années, comme celle de Vilanova & Balestieri (2014[21]) ou bien celle de Vieira et al. (2015[22]), avec une conclusion générale que les petites turbines hydroélectriques ont atteint un degré de sophistication technologique suffisante pour dégager une quantité d’énergie économiquement profitable. Par ailleurs, il semble qu’il y a beaucoup à gagner dans ce domaine à travers l’optimisation de la distribution de puissance entre les turbines différentes. De retour aux publications les plus récentes, j’ai trouvé des études de faisabilité tout à fait robustes pour les petites turbines hydroélectriques, qui indiquent que – pourvu qu’on soit prêt à accepter un retour d’environ 10 à 11 ans sur l’investissement initial – le petit hydro peut être exploité profitablement même avec une élévation relative en dessous de 20 mètres (Arthur et al. 2020[23] ; Ali et al. 2021[24]).

C’est ainsi que j’arrive donc à la portion finale dans la chaîne technologique du « Projet Aqueduc », donc au stockage d’énergie (TCES) ainsi que TCCS ou la station de chargement des véhicules électriques. La puissance à installer dans une station de chargement semble se situer entre 700 et 1000 kilowatts (Zhang et al. 2018[25]; McKinsey 2018[26]). En dessous de 700 kilowatt la station peut devenir si difficile à accéder pour le consommateur moyen, due aux files d’attente, qu’elle peut perdre la confiance des clients locaux. En revanche, tout ce qui va au-dessus de 1000 kilowatts est vraiment utile seulement aux heures de pointe dans des environnements urbains denses. Il y a des études de concept pour les stations de chargement où l’unité de stockage d’énergie est alimentée à partir des sources renouvelables (Al Wahedi & Bicer 2020[27]). Zhang et al. (2019[28]) présentent un concept d’entreprise tout fait pour une station de chargement située dans le milieu urbain. Apparemment, le seuil de profitabilité se situe aux environs de 5 100 000 kilowatt heures vendues par an.  

En termes de technologie de stockage strictement dite, les batteries Li-ion semblent être la solution de base pour maintenant, quoi qu’une combinaison avec les piles à combustible ou bien avec l’hydrogène semble prometteuse (Al Wahedi & Bicer 2020 op. cit. ; Sharma, Panvar & Tripati 2020[29]). En général, pour le moment, les batteries Li-Ion montrent le rythme d’innovation relativement le plus soutenu (Tomaszewska et al. 2019[30] ; de Simone & Piegari 2019[31]; Koohi-Fayegh & Rosen 2020[32]). Un article récent par Elmeligy et al. (2021[33]) présente un concept intéressant d’unité mobile de stockage qui pourrait se déplacer entre plusieurs stations de chargement. Quant à l’investissement initial requis pour une station de chargement, ça semble expérimenter toujours mais la marge de manœuvre se rétrécit pour tomber quelque part entre $600 ÷ $800 par 1 kW de puissance (Cole & Frazier 2019[34]; Cole, Frazier, Augustine 2021[35]).

[1] Neves, M. C., Nunes, L. M., & Monteiro, J. P. (2020). Evaluation of GRACE data for water resource management in Iberia: a case study of groundwater storage monitoring in the Algarve region. Journal of Hydrology: Regional Studies, 32, 100734. https://doi.org/10.1016/j.ejrh.2020.100734

[2] Chisola, M. N., Van der Laan, M., & Bristow, K. L. (2020). A landscape hydrology approach to inform sustainable water resource management under a changing environment. A case study for the Kaleya River Catchment, Zambia. Journal of Hydrology: Regional Studies, 32, 100762. https://doi.org/10.1016/j.ejrh.2020.100762

[3] Wehn, U., & Montalvo, C. (2018). Exploring the dynamics of water innovation: Foundations for water innovation studies. Journal of Cleaner Production, 171, S1-S19. https://doi.org/10.1016/j.jclepro.2017.10.118

[4] Mvulirwenande, S., & Wehn, U. (2020). Fostering water innovation in Africa through virtual incubation: Insights from the Dutch VIA Water programme. Environmental Science & Policy, 114, 119-127. https://doi.org/10.1016/j.envsci.2020.07.025

[5] Wong, T. H., Rogers, B. C., & Brown, R. R. (2020). Transforming cities through water-sensitive principles and practices. One Earth, 3(4), 436-447. https://doi.org/10.1016/j.oneear.2020.09.012

[6] Hogeboom, R. J. (2020). The Water Footprint Concept and Water’s Grand Environmental Challenges. One earth, 2(3), 218-222. https://doi.org/10.1016/j.oneear.2020.02.010

[7] Kumar, P., Avtar, R., Dasgupta, R., Johnson, B. A., Mukherjee, A., Ahsan, M. N., … & Mishra, B. K. (2020). Socio-hydrology: A key approach for adaptation to water scarcity and achieving human well-being in large riverine islands. Progress in Disaster Science, 8, 100134. https://doi.org/10.1016/j.pdisas.2020.100134

[8] Bagstad, K. J., Ancona, Z. H., Hass, J., Glynn, P. D., Wentland, S., Vardon, M., & Fay, J. (2020). Integrating physical and economic data into experimental water accounts for the United States: Lessons and opportunities. Ecosystem Services, 45, 101182. https://doi.org/10.1016/j.ecoser.2020.101182

[9] Mohamed, M. M., El-Shorbagy, W., Kizhisseri, M. I., Chowdhury, R., & McDonald, A. (2020). Evaluation of policy scenarios for water resources planning and management in an arid region. Journal of Hydrology: Regional Studies, 32, 100758. https://doi.org/10.1016/j.ejrh.2020.100758

[10] Harvey, J.W., Schaffranek, R.W., Noe, G.B., Larsen, L.G., Nowacki, D.J., O’Connor, B.L., 2009. Hydroecological factors governing surface water flow on a low-gradient floodplain. Water Resour. Res. 45, W03421, https://doi.org/10.1029/2008WR007129.

[11] Phiri, W. K., Vanzo, D., Banda, K., Nyirenda, E., & Nyambe, I. A. (2021). A pseudo-reservoir concept in SWAT model for the simulation of an alluvial floodplain in a complex tropical river system. Journal of Hydrology: Regional Studies, 33, 100770. https://doi.org/10.1016/j.ejrh.2020.100770.

[12] Lu, B., Blakers, A., Stocks, M., & Do, T. N. (2021). Low-cost, low-emission 100% renewable electricity in Southeast Asia supported by pumped hydro storage. Energy, 121387. https://doi.org/10.1016/j.energy.2021.121387

[13] Stocks, M., Stocks, R., Lu, B., Cheng, C., & Blakers, A. (2021). Global atlas of closed-loop pumped hydro energy storage. Joule, 5(1), 270-284. https://doi.org/10.1016/j.joule.2020.11.015

[14] Bortolini, L., & Zanin, G. (2019). Reprint of: Hydrological behaviour of rain gardens and plant suitability: A study in the Veneto plain (north-eastern Italy) conditions. Urban forestry & urban greening, 37, 74-86. https://doi.org/10.1016/j.ufug.2018.07.003

[15] Guo, X., Li, J., Yang, K., Fu, H., Wang, T., Guo, Y., … & Huang, W. (2018). Optimal design and performance analysis of hydraulic ram pump system. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Puissance and Energy, 232(7), 841-855. https://doi.org/10.1177%2F0957650918756761

[16] Li, J., Yang, K., Guo, X., Huang, W., Wang, T., Guo, Y., & Fu, H. (2021). Structural design and parameter optimization on a waste valve for hydraulic ram pumps. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Puissance and Energy, 235(4), 747–765. https://doi.org/10.1177/0957650920967489

[17] Cai, X., Ye, F., & Gholinia, F. (2020). Application of artificial neural network and Soil and Water Assessment Tools in evaluating Puissance generation of small hydroPuissance stations. Energy Reports, 6, 2106-2118. https://doi.org/10.1016/j.egyr.2020.08.010.

[18] Hatata, A. Y., El-Saadawi, M. M., & Saad, S. (2019). A feasibility study of small hydro Puissance for selected locations in Egypt. Energy Strategy Reviews, 24, 300-313. https://doi.org/10.1016/j.esr.2019.04.013

[19] Syahputra, R., & Soesanti, I. (2021). Renewable energy systems based on micro-hydro and solar photovoltaic for rural areas: A case study in Yogyakarta, Indonesia. Energy Reports, 7, 472-490. https://doi.org/10.1016/j.egyr.2021.01.015

[20] Sterl, S., Donk, P., Willems, P., & Thiery, W. (2020). Turbines of the Caribbean: Decarbonising Suriname’s electricity mix through hydro-supported integration of wind Puissance. Renewable and Sustainable Energy Reviews, 134, 110352. https://doi.org/10.1016/j.rser.2020.110352

[21] Vilanova, M. R. N., & Balestieri, J. A. P. (2014). HydroPuissance recovery in water supply systems: Models and case study. Energy conversion and management, 84, 414-426. https://doi.org/10.1016/j.enconman.2014.04.057

[22] Vieira, D. A. G., Guedes, L. S. M., Lisboa, A. C., & Saldanha, R. R. (2015). Formulations for hydroelectric energy production with optimality conditions. Energy Conversion and Management, 89, 781-788. https://doi.org/10.1016/j.enconman.2014.10.048

[23] Arthur, E., Anyemedu, F. O. K., Gyamfi, C., Asantewaa-Tannor, P., Adjei, K. A., Anornu, G. K., & Odai, S. N. (2020). Potential for small hydroPuissance development in the Lower Pra River Basin, Ghana. Journal of Hydrology: Regional Studies, 32, 100757. https://doi.org/10.1016/j.ejrh.2020.100757

[24] Ali, M., Wazir, R., Imran, K., Ullah, K., Janjua, A. K., Ulasyar, A., … & Guerrero, J. M. (2021). Techno-economic assessment and sustainability impact of hybrid energy systems in Gilgit-Baltistan, Pakistan. Energy Reports, 7, 2546-2562. https://doi.org/10.1016/j.egyr.2021.04.036

[25] Zhang, Y., He, Y., Wang, X., Wang, Y., Fang, C., Xue, H., & Fang, C. (2018). Modeling of fast charging station equipped with energy storage. Global Energy Interconnection, 1(2), 145-152. DOI:10.14171/j.2096-5117.gei.2018.02.006

[26] McKinsey Center for Future Mobility, How Battery Storage Can Help Charge the Electric-Vehicle Market?, February 2018,

[27] Al Wahedi, A., & Bicer, Y. (2020). Development of an off-grid electrical vehicle charging station hybridized with renewables including battery cooling system and multiple energy storage units. Energy Reports, 6, 2006-2021. https://doi.org/10.1016/j.egyr.2020.07.022

[28] Zhang, J., Liu, C., Yuan, R., Li, T., Li, K., Li, B., … & Jiang, Z. (2019). Design scheme for fast charging station for electric vehicles with distributed photovoltaic power generation. Global Energy Interconnection, 2(2), 150-159. https://doi.org/10.1016/j.gloei.2019.07.003

[29] Sharma, S., Panwar, A. K., & Tripathi, M. M. (2020). Storage technologies for electric vehicles. Journal of traffic and transportation engineering (english edition), 7(3), 340-361. https://doi.org/10.1016/j.jtte.2020.04.004

[30] Tomaszewska, A., Chu, Z., Feng, X., O’Kane, S., Liu, X., Chen, J., … & Wu, B. (2019). Lithium-ion battery fast charging: A review. ETransportation, 1, 100011. https://doi.org/10.1016/j.etran.2019.100011

[31] De Simone, D., & Piegari, L. (2019). Integration of stationary batteries for fast charge EV charging stations. Energies, 12(24), 4638. https://doi.org/10.3390/en12244638

[32] Koohi-Fayegh, S., & Rosen, M. A. (2020). A review of energy storage types, applications and recent developments. Journal of Energy Storage, 27, 101047. https://doi.org/10.1016/j.est.2019.101047

[33] Elmeligy, M. M., Shaaban, M. F., Azab, A., Azzouz, M. A., & Mokhtar, M. (2021). A Mobile Energy Storage Unit Serving Multiple EV Charging Stations. Energies, 14(10), 2969. https://doi.org/10.3390/en14102969

[34] Cole, Wesley, and A. Will Frazier. 2019. Cost Projections for Utility-Scale Battery Storage.

Golden, CO: National Renewable Energy Laboratory. NREL/TP-6A20-73222. https://www.nrel.gov/docs/fy19osti/73222.pdf

[35] Cole, Wesley, A. Will Frazier, and Chad Augustine. 2021. Cost Projections for UtilityScale Battery Storage: 2021 Update. Golden, CO: National Renewable Energy

Laboratory. NREL/TP-6A20-79236. https://www.nrel.gov/docs/fy21osti/79236.pdf.

We keep going until we observe

I keep working on a proof-of-concept paper for my idea of ‘Energy Ponds’. In my last two updates, namely in ‘Seasonal lakes’, and in ‘Le Catch 22 dans ce jardin d’Eden’, I sort of refreshed my ideas and set the canvas for painting. Now, I start sketching. What exact concept do I want to prove, and what kind of evidence can possibly confirm (or discard) that concept? The idea I am working on has a few different layers. The most general vision is that of purposefully storing water in spongy structures akin to swamps or wetlands. These can bear various degree of artificial construction, and can stretch from natural wetlands, through semi-artificial ones, all the way to urban technologies such as rain gardens and sponge cities. The most general proof corresponding to that vision is a review of publicly available research – peer-reviewed papers, preprints, databases etc. – on that general topic.

Against that general landscape, I sketch two more specific concepts: the idea of using ram pumps as a technology of forced water retention, and the possibility of locating those wetland structures in the broadly spoken Northern Europe, thus my home region. Correspondingly, I need to provide two streams of scientific proof: a review of literature on the technology of ram pumping, on the one hand, and on the actual natural conditions, as well as land management policies in Europe, on the other hand.  I need to consider the environmental impact of creating new wetland-like structures in Northern Europe, as well as the socio-economic impact, and legal feasibility of conducting such projects.

Next, I sort of build upwards. I hypothesise a complex technology, where ram-pumped water from the river goes into a sort of light elevated tanks, and from there, using the principle of Roman siphon, cascades down into wetlands, and through a series of small hydro-electric turbines. Turbines generate electricity, which is being stored and then sold outside.

At that point, I have a technology of water retention coupled with a technology of energy generation and storage. I further advance a second hypothesis that such a complex technology will be economically sustainable based on the corresponding sales of electricity. In other words, I want to figure out a configuration of that technology, which will be suitable for communities which either don’t care at all, or simply cannot afford to care about the positive environmental impact of the solution proposed.

Proof of concept for those two hypotheses is going to be complex. First, I need to pass in review the available technologies for energy storage, energy generation, as well as for the construction of elevated tanks and Roman siphons. I need to take into account various technological mixes, including the incorporation of wind turbines and photovoltaic installation into the whole thing, in order to optimize the output of energy. I will try to look for documented examples of small hydro-generation coupled with wind and solar. Then, I have to rack the literature as regards mathematical models for the optimization of such power systems and put them against my own idea of reverse engineering back from the storage technology. I take the technology of energy storage which seems the most suitable for the local market of energy, and for the hypothetical charging from hydro-wind-solar mixed generation. I build a control scenario where that storage facility just buys energy at wholesale prices from the power grid and then resells it. Next, I configure the hydro-wind-solar generation so as to make it economically competitive against the supply of energy from the power grid.

Now, I sketch. I keep in mind the levels of conceptualization outlined above, and I quickly move through published science along that logical path, quickly picking a few articles for each topic. I am going to put those nonchalantly collected pieces of science back-to-back and see how and whether at all it all makes sense together. I start with Bortolini & Zanin (2019[1]), who study the impact of rain gardens on water management in cities of the Veneto region in Italy. Rain gardens are vegetal structures, set up in the urban environment, with the specific purpose to retain rainwater.  Bortolini & Zanin (2019 op. cit.) use a simplified water balance, where the rain garden absorbs and retains a volume ‘I’ of water (‘I’ stands for infiltration), which is the difference between precipitations on the one hand, and the sum total of overflowing runoff from the rain garden plus evapotranspiration of water, on the other hand. Soil and plants in the rain garden have a given top capacity to retain water. Green plants typically hold 80 – 95% of their mass in water, whilst trees hold about 50%. Soil is considered wet when it contains about 25% of water. The rain garden absorbs water from precipitations at a rate determined by hydraulic conductivity, which means the relative ease of a fluid (usually water) to move through pore spaces or fractures, and which depends on the intrinsic permeability of the material, the degree of saturation, and on the density and viscosity of the fluid.

As I look at it, I can see that the actual capacity of water retention in a rain garden can hardly be determined a priori, unless we have really a lot of empirical data from the given location. For a new location of a new rain garden, it is safe to assume that we need an experimental phase when we empirically assess the retentive capacity of the rain garden with different configurations of soil and vegetation used. That leads me to generalizing that any porous structure we use for retaining rainwater, would it be something like wetlands, or something like a rain garden in urban environment, has a natural constraint of hydraulic conductivity, and that constraint determines the percentage of precipitations, and the metric volume thereof, which the given structure can retain.

Bortolini & Zanin (2019 op. cit.) bring forth empirical results which suggest that properly designed rain gardens located on rooftops in a city can absorb from 87% to 93% of the total input of water they receive. Cool. I move on and towards the issue of water management in Europe, with a working paper by Fribourg-Blanc, B. (2018[2]), and the most important takeaway from that paper is that we have something called European Platform for Natural Water Retention Measures AKA http://nwrm.eu , and that thing have both good properties and bad properties. The good thing about http://nwrm.eu is that it contains loads of data and publications about projects in Natural Water Retention in Europe. The bad thing is that http://nwrm.eu is not a secure website. Another paper, by Tóth et al. (2017[3]) tells me that another analytical tool exists, namely the European Soil Hydraulic Database (EU‐ SoilHydroGrids ver1.0).

So far, so good. I already know there is data and science for evaluating, with acceptable precision, the optimal structure and the capacity for water retention in porous structures such as rain gardens or wetlands, in the European context. I move to the technology of ram pumps. I grab two papers: Guo et al. (2018[4]), and Li et al. (2021[5]). They show me two important things. Firstly, China seems to be burning the rubber in the field of ram pumping technology. Secondly, the greatest uncertainty as for that technology seems to be the actual height those ram pumps can elevate water at, or, when coupled with hydropower, the hydraulic head which ram pumps can create. Guo et al. (2018 op. cit.) claim that 50 meters of elevation is the maximum which is both feasible and efficient. Li et al. (2021 op. cit.) are sort of vertically more conservative and claim that the whole thing should be kept below 30 meters of elevation. Both are better than 20 meters, which is what I thought was the best one can expect. Greater elevation of water means greater hydraulic head, and more hydropower to be generated. It pays off to review literature.

Lots of uncertainty as for the actual capacity and efficiency of ram pumping means quick technological change in that domain. This is economically interesting. It means that investing in projects which involve ram pumping means investment in quickly changing a technology. That means both high hopes for an even better technology in immediate future, and high needs for cash in the balance sheet of the entities involved.

I move to the end-of-the-pipeline technology in my concept, namely to energy storage. I study a paper by Koohi-Fayegh & Rosen (2020[6]), which suggests two things. Firstly, for a standalone installation in renewable energy, whatever combination of small hydropower, photovoltaic and small wind turbines we think of, lithium-ion batteries are always a good idea for power storage, Secondly, when we work with hydrogeneration, thus when we have any hydraulic head to make electricity with, pumped storage comes sort of natural. That leads me to an idea which looks even crazier than what I have imagined so far: what if we create an elevated garden with strong capacity for water retention. Ram pumps take water from the river and pump it up onto elevated platforms with rain gardens on it. Those platforms can be optimized as for their absorption of sunlight and thus as regards their interaction with whatever is underneath them.  

I move to small hydro, and I find two papers, namely Couto & Olden (2018[7]), and Lange et al. (2018[8]), which are both interestingly critical as regards small hydropower installations. Lange et al. (2018 op. cit.) claim that the overall environmental impact of small hydro should be closely monitored. Couto & Olden (2018 op. cit.) go further and claim there is a ‘craze’ about small hydro, and that craze has already lead to overinvestment in the corresponding installations, which can be damaging both environmentally and economically (overinvestment means financial collapse of many projects). Those critical views in mind, I turn to another paper, by Zhou et al. (2019[9]), who approach the issue as a case for optimization, within a broader framework called ‘Water-Food-Energy’ Nexus, WFE for closer friends. This paper, just as a few others it cites (Ming et al. 2018[10]; Uen et al. 2018[11]), advocates for using artificial intelligence in order to optimize for WFE.

Zhou et al. (2019 op.cit.) set three hydrological scenarios for empirical research and simulation. The baseline scenario corresponds to an average hydrological year, with average water levels and average precipitations. Next to it are: a dry year and a wet year. The authors assume that the cost of installation in small hydropower is $600 per kW on average.  They simulate the use of two technologies for hydro-electric turbines: Pelton and Vortex. Pelton turbines are optimized paddled wheels, essentially, whilst the Vortex technology consists in creating, precisely, a vortex of water, and that vortex moves a rotor placed in the middle of it.

Zhou et al. (2019 op.cit.) create a multi-objective function to optimize, with the following desired outcomes:

>> Objective 1: maximize the reliability of water supply by minimizing the probability of real water shortage occurring.

>> Objective 2: maximize water storage given the capacity of the reservoir. Note: reservoir is understood hydrologically, as any structure, natural or artificial, able to retain water.

>> Objective 3: maximize the average annual output of small hydro-electric turbines

Those objectives are being achieved under the corresponding sets of constraints. For water supply those constraints all turn around water balance, whilst for energy output it is more about the engineering properties of the technologies taken into account. The three objectives are hierarchized. First, Zhou et al. (2019 op.cit.) perform an optimization regarding Objectives 1 and 2, thus in order to find the optimal hydrological characteristics to meet, and then, on the basis of these, they optimize the technology to put in place, as regards power output.

The general tool for optimization used by Zhou et al. (2019 op.cit.) is a genetic algorithm called NSGA-II, AKA Non-dominated Sorting Genetic Algorithm. Apparently, NSGA-II has a long and successful history of good track in engineering, including water management and energy (see e.g. Chang et al. 2016[12]; Jain & Sachdeva 2017[13];  Assaf & Shabani 2018[14]). I want to stop for a while here and have a good look at this specific algorithm. The logic of NSGA-II starts with creating an initial population of cases/situations/configurations etc. Each case is a combination of observations as regards the objectives to meet, and the actual values observed in constraining variables, e.g. precipitations for water balance or hydraulic head for the output of hydropower. In the conventional lingo of this algorithm, those cases are called chromosomes. Yes, I know, a hydro-electric turbine placed in the context of water management hardly looks like a chromosome, but it is a genetic algorithm, and it just sounds fancy to use that biologically marked vocabulary.

As for me, I like staying close to real life, and therefore I call those cases solutions rather than chromosomes. Anyway, the underlying math is the same. Once I have that initial population of real-life solutions, I calculate two parameters for each of them: their rank as regards the objectives to maximize, and their so-called ‘crowded distance’. Ranking is done with the procedure of fast non-dominated sorting. It is a comparison in pairs, where the solution A dominates another solution B, if and only if there is no objective of A worse than that objective of B and there is at least one objective of A better than that objective of B. The solution which scores the most wins in such peer-to-peer comparisons is at the top of the ranking, the one with the second score of wins is the second etc. Crowding distance is essentially the same as what I call coefficient of coherence in my own research: Euclidean distance (or other mathematical distance) is calculated for each pair of solutions. As a result, each solution is associated with k Euclidean distances to the k remaining solutions, which can be reduced to an average distance, i.e. the crowded distance.

In the next step, an off-spring population is produced from that original population of solutions. It is created by taking relatively the fittest solutions from the initial population, recombining their characteristics in a 50/50 proportion, and adding them some capacity for endogenous mutation. Two out of these three genetic functions are de facto controlled. We choose relatively the fittest by establishing some kind of threshold for fitness, as regards the objectives pursued. It can be a required minimum, a quantile (e.g. the third quartile), or an average. In the first case, we arbitrarily impose a scale of fitness on our population, whilst in the latter two the hierarchy of fitness is generated endogenously from the population of solutions observed. Fitness can have shades and grades, by weighing the score in non-dominated sorting, thus the number of wins over other solutions, on the one hand, and the crowded distance on the other hand. In other words, we can go for solutions which have a lot of similar ones in the population (i.e. which have a low average crowded distance), or, conversely, we can privilege lone wolves, with a high average Euclidean distance from anything else on the plate.  

The capacity for endogenous mutation means that we can allow variance in all or in just the selected variables which make each solution. The number of degrees of freedom we allow in each variable dictates the number of mutations that can be created. Once again, discreet power is given to the analyst: we can choose the genetic traits which can mutate and we can determine their freedom to mutate. In an engineering problem, technological and environmental constraints should normally put a cap on the capacity for mutation. Still, we can think about an algorithm which definitely kicks the lid off the barrel of reality, and which generates mutations in the wildest registers of variables considered. It is a way to simulate a process when the presence of strong outliers has a strong impact on the whole population.

The same discreet cap on the freedom to evolve is to be found when we repeat the process. The offspring generation of solutions goes essentially through the same process as the initial one, to produce further offspring: ranking by non-dominated sorting and crowded distance, selection of the fittest, recombination, and endogenous mutation. At the starting point of this process, we can be two alternative versions of the Mother Nature. We can be a mean Mother Nature, and we shave off from the offspring population all those baby-solutions which do not meet the initial constraints, e.g. zero supply of water in this specific case. On the other hand, we can be even meaner a Mother Nature and allow those strange, dysfunctional mutants to keep going and see what happens to the whole species after a few rounds of genetic reproduction.

With each generation, we compute an average crowded distance between all the solutions created, i.e. we check how diverse is the species in this generation. As long as diversity grows or remains constant, we assume that the divergence between the solutions generated grows or stays the same. Similarly, we can compute an even more general crowded distance between each pair of generations, and therefore to assess how far has the current generation gone from the parent one. We keep going until we observe that the intra-generational crowded distance and the inter-generational one start narrowing down asymptotically to zero. In other words, we consider resuming evolution when solutions in the game become highly similar to each other and when genetic change stops bringing significant functional change.

Cool. When I want to optimize my concept of Energy Ponds, I need to add the objective of constrained return on investment, based on the sales of electricity. In comparison to Zhou et al. (2019 op.cit.), I need to add a third level of selection. I start with selecting environmentally the solutions which make sense in terms of water management. In the next step, I produce a range of solutions which assure the greatest output of power, in a possible mix with solar and wind. Then I take those and filter them through the NSGA-II procedure as regards their capacity to sustain themselves financially. Mind you, I can shake it off a bit by fusing together those levels of selection. I can simulate extreme cases, when, for example, good economic sustainability becomes an environmental problem. Still, it would be rather theoretical. In Europe, non-compliance with environmental requirements makes a project a non-starter per se: you just can get the necessary permits if your hydropower project messes with hydrological constraints legally imposed on the given location.     

Cool. It all starts making sense. There is apparently a lot of stir in the technology of making semi-artificial structures for retaining water, such as rain gardens and wetlands. That means a lot of experimentation, and that experimentation can be guided and optimized by testing the fitness of alternative solutions for meeting objectives of water management, power output and economic sustainability. I have some starting data, to produce the initial generation of solutions, and then try to optimize them with an algorithm such as NSGA-II.

[1] Bortolini, L., & Zanin, G. (2019). Reprint of: Hydrological behaviour of rain gardens and plant suitability: A study in the Veneto plain (north-eastern Italy) conditions. Urban forestry & urban greening, 37, 74-86. https://doi.org/10.1016/j.ufug.2018.07.003

[2] Fribourg-Blanc, B. (2018, April). Natural Water Retention Measures (NWRM), a tool to manage hydrological issues in Europe?. In EGU General Assembly Conference Abstracts (p. 19043). https://ui.adsabs.harvard.edu/abs/2018EGUGA..2019043F/abstract

[3] Tóth, B., Weynants, M., Pásztor, L., & Hengl, T. (2017). 3D soil hydraulic database of Europe at 250 m resolution. Hydrological Processes, 31(14), 2662-2666. https://onlinelibrary.wiley.com/doi/pdf/10.1002/hyp.11203

[4] Guo, X., Li, J., Yang, K., Fu, H., Wang, T., Guo, Y., … & Huang, W. (2018). Optimal design and performance analysis of hydraulic ram pump system. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 232(7), 841-855. https://doi.org/10.1177%2F0957650918756761

[5] Li, J., Yang, K., Guo, X., Huang, W., Wang, T., Guo, Y., & Fu, H. (2021). Structural design and parameter optimization on a waste valve for hydraulic ram pumps. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 235(4), 747–765. https://doi.org/10.1177/0957650920967489

[6] Koohi-Fayegh, S., & Rosen, M. A. (2020). A review of energy storage types, applications and recent developments. Journal of Energy Storage, 27, 101047. https://doi.org/10.1016/j.est.2019.101047

[7] Couto, T. B., & Olden, J. D. (2018). Global proliferation of small hydropower plants–science and policy. Frontiers in Ecology and the Environment, 16(2), 91-100. https://doi.org/10.1002/fee.1746

[8] Lange, K., Meier, P., Trautwein, C., Schmid, M., Robinson, C. T., Weber, C., & Brodersen, J. (2018). Basin‐scale effects of small hydropower on biodiversity dynamics. Frontiers in Ecology and the Environment, 16(7), 397-404.  https://doi.org/10.1002/fee.1823

[9] Zhou, Y., Chang, L. C., Uen, T. S., Guo, S., Xu, C. Y., & Chang, F. J. (2019). Prospect for small-hydropower installation settled upon optimal water allocation: An action to stimulate synergies of water-food-energy nexus. Applied Energy, 238, 668-682. https://doi.org/10.1016/j.apenergy.2019.01.069

[10] Ming, B., Liu, P., Cheng, L., Zhou, Y., & Wang, X. (2018). Optimal daily generation scheduling of large hydro–photovoltaic hybrid power plants. Energy Conversion and Management, 171, 528-540. https://doi.org/10.1016/j.enconman.2018.06.001

[11] Uen, T. S., Chang, F. J., Zhou, Y., & Tsai, W. P. (2018). Exploring synergistic benefits of Water-Food-Energy Nexus through multi-objective reservoir optimization schemes. Science of the Total Environment, 633, 341-351. https://doi.org/10.1016/j.scitotenv.2018.03.172

[12] Chang, F. J., Wang, Y. C., & Tsai, W. P. (2016). Modelling intelligent water resources allocation for multi-users. Water resources management, 30(4), 1395-1413. https://doi.org/10.1007/s11269-016-1229-6

[13] Jain, V., & Sachdeva, G. (2017). Energy, exergy, economic (3E) analyses and multi-objective optimization of vapor absorption heat transformer using NSGA-II technique. Energy Conversion and Management, 148, 1096-1113. https://doi.org/10.1016/j.enconman.2017.06.055

[14] Assaf, J., & Shabani, B. (2018). Multi-objective sizing optimisation of a solar-thermal system integrated with a solar-hydrogen combined heat and power system, using genetic algorithm. Energy Conversion and Management, 164, 518-532. https://doi.org/10.1016/j.enconman.2018.03.026

Seasonal lakes

Once again, been a while since I last blogged. What do you want, I am having a busy summer. Putting order in my own chaos, and, over the top of that, putting order in other people’s chaos, this is all quite demanding in terms of time and energy. What? Without trying to put order in chaos, that chaos might take less time and energy? Well, yes, but order look tidier than chaos.

I am returning to the technological concept which I labelled ‘Energy Ponds’ (or ‘projet Aqueduc’ in French >> see: Le Catch 22 dans ce jardin d’Eden). You can find a description of that concept onder the hyperlinked titles provided. I am focusing on refining my repertoire of skills in scientific validation of technological concepts. I am passing in review some recent literature, and I am trying to find good narrative practices in that domain.

The general background of ‘Energy Ponds’ consists in natural phenomena observable in Europe as the climate change progresses, namely: a) long-term shift in the structure of precipitations, from snow to rain b) increasing occurrence of floods and droughts c) spontaneous reemergence of wetlands. All these phenomena have one common denominator: increasingly volatile flow per second in rivers. The essential idea of Energy Ponds is to ‘financialize’ that volatile flow, so to say, i.e. to capture its local surpluses, store them for later, and use the very mechanism of storage itself as a source of economic value.

When water flows downstream, in a river, its retention can be approached as the opportunity for the same water to loop many times over the same specific portion of the collecting basin (of the river). Once such a loop is created, we can extend the average time that a liter of water spends in the whereabouts. Ram pumps, connected to storage structures akin to swamps, can give such an opportunity. A ram pump uses the kinetic energy of flowing water in order to pump some of that flow up and away from its mainstream. Ram pumps allow forcing a process, which we now as otherwise natural. Rivers, especially in geological plains, where they flow relatively slowly, tend to build, with time, multiple ramifications. Those branchings can be directly observable at the surface, as meanders, floodplains or seasonal lakes, but much of them is underground, as pockets of groundwater. In this respect, it is useful to keep in mind that mechanically, rivers are the drainpipes of rainwater from their respective basins. Another basic hydrological fact, useful to remember in the context of the Energy Ponds concept, is that strictly speaking retention of rainwater – i.e. a complete halt in its circulation through the collecting basin of the river – is rarely possible, and just as rarely it is a sensible idea to implement. Retention means rather a slowdown to the flow of rainwater through the collecting basin into the river.

One of the ways that water can be slowed down consists in making it loop many times over the same section of the river. Let’s imagine a simple looping sequence: water from the river is being ram-pumped up and away into retentive structures akin to swamps, i.e. moderately deep spongy structures underground, with high capacity for retention, covered with a superficial layer of shallow-rooted vegetation. With time, as the swamp fills with water, the surplus is evacuated back into the river, by a system of canals. Water stored in the swamp will be ultimately evacuated, too, minus evaporation, it will just happen much more slowly, by the intermediary of groundwaters. In order to illustrate the concept mathematically, let’ s suppose that we have water in the river flowing at the pace of, e.g. 45 m3 per second. We make it loop once via ram pumps and retentive swamps, and, if as a result of that looping, the speed of the flow is sliced by 3. On the long run we slow down the way that the river works as the local drainpipe: we slow it from 43 m3 per second down to [43/3 = 14,33…] m3 per second.  As water from the river flows slower overall, it can yield more environmental services: each cubic meter of water has more time to ‘work’ in the ecosystem.  

When I think of it, any human social structure, such as settlements, industries, infrastructures etc., needs to stay in balance with natural environment. That balance is to be understood broadly, as the capacity to stay, for a satisfactorily long time, within a ‘safety zone’, where the ecosystem simply doesn’t kill us. That view has little to do with the moral concepts of environment-friendliness or sustainability. As a matter of fact, most known human social structures sooner or later fall out of balance with the ecosystem, and this is how civilizations collapse. Thus, here comes the first important assumption: any human social structure is, at some level, an environmental project. The incumbent social structures, possible to consider as relatively stable, are environmental projects which have simply hold in place long enough to grow social institutions, and those institutions allow further seeking of environmental balance.

I am starting my review of literature with an article by Phiri et al. (2021[1]), where the authors present a model for assessing the way that alluvial floodplains behave. I chose this one because my concept of Energy Ponds is supposed to work precisely in alluvial floodplains, i.e. in places where we have: a) a big river b) a lot of volatility in the amount of water in that river, and, as a consequence, we have (c) an alternation of floods and droughts. Normal stuff where I come from, i.e. in Northern Europe. Phiri et al. use the general model, acronymically called SWAT, which comes from ‘Soil and Water Assessment Tool’ (see also: Arnold et al. 1998[2]; Neitsch et al. 2005[3]), and with that general tool, they study the concept of pseudo-reservoirs in alluvial plains. In short, a pseudo-reservoir is a hydrological structure which works like a reservoir but does not necessarily look like one. In that sense, wetlands in floodplains can work as reservoirs of water, even if from the hydrological point of view they are rather extensions of the main river channel (Harvey et al. 2009[4]).

Analytically, the SWAT model defines the way a reservoir works with the following equation: V = Vstored + Vflowin − Vflowout + Vpcp − Vevap − Vseep . People can rightly argue that it is a good thing to know what symbols mean in an equation, and therefore V stands for the volume of water in reservoir at the end of the day, Vstored corresponds to the amount of water stored at the beginning of the day, Vflowin means the quantity of water entering reservoir during the day, Vflowout is the metric outflow of water during the day, Vpcp is volume of precipitation falling on the water body during the day, Vevap is volume of water removed from the water body by evaporation during the day, Vseep is volume of water lost from the water body by seepage.

This is a good thing to know, as well, once we have a nice equation, what the hell are we supposed to do with it in real life. Well, the SWAT model has even its fan page (http://www.swatusers.com ), and, as Phiri et al. phrase it out, it seems that the best practical use is to control the so-called ‘target release’, i.e. the quantity of water released at a given point in space and time, designated as Vtarg. The target release is mostly used as a control metric for preventing or alleviating floods, and with that purpose in mind, two decision rules are formulated. During the non-flood season, no reservation for flood is needed, and target storage is set at emergency spillway volume. In other words, in the absence of imminent flood, we can keep the reservoir full. On the other hand, when the flood season is on, flood control reservation is a function of soil water content. This is set to maximum and 50 % of maximum for wet and dry grounds, respectively. In the context of the V = Vstored + Vflowin − Vflowout + Vpcp − Vevap − Vseep equation, Vtarg is a specific value (or interval of values) in the Vflowout component.

As I am wrapping my mind around those conditions, I am thinking about the opposite application, i.e. about preventing and alleviating droughts. Drought is recognizable by exceptionally low values in the amount of water stored at the end of the given period, thus in the basic V, in the presence of low precipitation, thus low Vpcp, and high evaporation, which corresponds to high Vevap. More generally, both floods and droughts occur when – or rather after – in a given Vflowin − Vflowout balance, precipitation and evaporation take one turn or another.

I feel like moving those exogenous meteorological factors on one side of the equation, which goes like  – Vpcp + Vevap =  – V + Vstored + Vflowin − Vflowout − Vseep and doesn’t make much sense, as there are not really many cases of negative precipitation. I need to switch signs, and then it is more presentable, as Vpcp – VevapV – Vstored – Vflowin + Vflowout + Vseep . Weeell, almost makes sense. I guess that Vflowin is sort of exogenous, too. The inflow of water into the basin of the river comes from a melting glacier, from another river, from an upstream section of the same river etc. I reframe: Vpcp – Vevap + Vflowin V – Vstored + Vflowout + Vseep  . Now, it makes sense. Precipitations plus the inflow of water through the main channel of the river, minus evaporation, all that stuff creates a residual quantity of water. That residual quantity seeps into the groundwaters (Vseep), flows out (Vflowout), and stays in the reservoir-like structure at the end of the day (V – Vstored).

I am having a look at how Phiri et al. (2021 op. cit.) phrase out their model of pseudo-reservoir. The output value they peg the whole thing on is Vpsrc, or the quantity of water retained in the pseudo-reservoir at the end of the day. The Vpsrc is modelled for two alternative situations: no flood (V ≤ Vtarg), or flood (V > Vtarg). I interpret drought as particularly uncomfortable a case of the absence of flood.

Whatever. If V ≤ Vtarg , then Vpsrc = Vstored + Vflowin − Vbaseflowout + Vpcp − Vevap − Vseep  , where, besides the already known variables, Vbaseflowoutstands for volume of water leaving PSRC during the day as base flow. When, on the other hand, we have flood, Vpsrc = Vstored + Vflowin − Vbaseflowout − Voverflowout + Vpcp − Vevap − Vseep .

Phiri et al. (2021 op. cit.) argue that once we incorporate the phenomenon of pseudo-reservoirs in the evaluation of possible water discharge from alluvial floodplains, the above-presented equations perform better than the standard SWAT model, or V = Vstored + Vflowin − Vflowout + Vpcp − Vevap − Vseep

My principal takeaway from the research by Phiri et al. (2021 op. cit.) is that wetlands matter significantly for the hydrological balance of areas with characteristics of floodplains. My concept of ‘Energy Ponds’ assumes, among other things, storing water in swamp-like structures, including urban and semi-urban ones, such as rain gardens (Sharma & Malaviya 2021[5] ; Li, Liu & Li 2020[6] ; Venvik & Boogaard 2020[7],) or sponge cities (Ma, Jiang & Swallow 2020[8] ; Sun, Cheshmehzangi & Wang 2020[9]).  

Now, I have a few papers which allow me to have sort of a bird’s eye view of the SWAT model as regards the actual predictability of flow and retention in fluvial basins. It turns out that identifying optimal sites for hydropower installations is a very complex task, prone to a lot of error, and only the introduction of digital data such as GIS allows acceptable precision. The problem is to estimate accurately both the flow and the head of the waterway in question at an exact location (Liu et al., 2017[10]; Gollou and Ghadimi 2017[11]; Aghajani & Ghadimi 2018[12]; Yu & Ghadimi 2019[13]; Cai, Ye & Gholinia 2020[14]). My concept of ‘Energy Ponds’ includes hydrogeneration, but makes one of those variables constant, by introducing something like Roman siphons, with a constant head, apparently possible to peg at 20 metres. The hydro-power generation seems to be pseudo-concave function (i.e. it hits quite a broad, concave peak of performance) if the hydraulic head (height differential) is constant, and the associated productivity function is strongly increasing. Analytically, it can be expressed as a polynomial, i.e. as a combination of independent factors with various powers (various impact) assigned to them (Cordova et al. 2014[15]; Vieira et al. 2015[16]). In other words, by introducing, in my technological concept, that constant head (height) makes the whole thing more prone to optimization.

Now, I take on a paper which shows how to present a proof of concept properly: Pradhan, A., Marence, M., & Franca, M. J. (2021). The adoption of Seawater Pump Storage Hydropower Systems increases the share of renewable energy production in Small Island Developing States. Renewable Energy, https://doi.org/10.1016/j.renene.2021.05.151 . This paper is quite close to my concept of ‘Energy Ponds’, as it includes the technology of pumped storage, which I think about morphing and changing into something slightly different. Such as presented by Pradhan, Marence & Franca (2021, op. cit.), the proof of concept is structured in two parts: the general concept is presented, and then a specific location is studied  – the island of Curaçao, in this case – as representative for a whole category. The substance of proof is articulated around the following points:

>> the basic diagnosis as for the needs of the local community in terms of energy sources, with the basic question whether Seawater Pumped Storage Hydropower System is locally suitable as technology. In this specific case, the main criterium was the possible reduction of dependency on fossils. Assumptions as for the electric power required have been made, specifically for the local community.  

>> a GIS tool has been tested for choosing the optimal location. GIS stands for Geographic Information System (https://en.wikipedia.org/wiki/Geographic_information_system ). In this specific thread the proof of concept consisted in checking whether the available GIS data, and the software available for processing it are sufficient for selecting an optimal location in Curaçao.

At the bottom line, the proof of concept sums up to checking, whether the available GIS technology allows calibrating a site for installing the required electrical power in a Seawater Pumped Storage Hydropower System.

That paper by Pradhan, Marence & Franca (2021, op. cit.) presents a few other interesting traits for me. Firstly, the author’s prove that combining hydropower with windmills and solar modules is a viable solution, and this is exactly what I thought, only I wasn’t sure. Secondly, the authors consider a very practical issue: corrosion, and the materials recommended in order to bypass that problem. Their choice is fiberglass. Secondly, they introduce an important parameter, namely the L/H aka ‘Length to Head’ ratio. This is the proportion between the length of water conductors and the hydraulic head (i.e. the relative denivelation) in the actual installation. Pradhan, Marence & Franca recommend distinguishing two types of installations: those with L/H < 15, on the one hand, and those with 15 ≤ L/H ≤ 25. However accurate is that assessment of theirs, it is a paremeter to consider. In my concept of ‘Energy Ponds’, I assume an artificially created hydraulic head of 20 metres, and thus the conductors leading from elevated tanks to the collecting wetland-type structure should be classified in two types, namely [(L/H < 15) (L < 15*20) (L < 300 metres)], on the one hand, and [(15 ≤ L/H ≤ 25) (300 metres ≤ L ≤ 500 metres)], on the other hand.  

Still, there is bad news for me. According to a report by Botterud, Levin & Koritarov (2014[17]), which Pradhan, Marence & Franca quote as an authoritative source, hydraulic head for pumped storage should be at least 100 metres in order to make the whole thing profitable. My working assumption with ‘Energy Ponds’ is 20 metres, and, obviously, I have to work through it.

I think I have the outline of a structure for writing a decent proof-of-concept article for my ‘Energy Ponds’ concept. I think I should start with something I have already done once, two years ago, namely with compiling data as regards places in Europe, located in fluvial plains, with relatively the large volatility in water level and flow. These places will need water retention.

Out of that list, I select locations eligible for creating wetland-type structures for retaining water, either in the form of swamps, or as porous architectural structures. Once that second list prepared, I assess the local need for electrical power. From there, I reverse engineer. With a given power of X megawatts, I reverse to the storage capacity needed for delivering that power efficiently and cost-effectively. I nail down the storage capacity as such, and I pass in review the available technologies of power storage.

Next, I choose the best storage technology for that specific place, and I estimate the investment outlays necessary for installing it. I calculate the hydropower required in hydroelectric turbines, as well as in adjacent windmills and photovoltaic. I check whether the local river can supply the amount of water that fits the bill. I pass in review literature as regards optimal combinations of those three sources of energy. I calculate the investment outlays needed to install all that stuff, and I add the investment required in ram pumping, elevated tanks, and water conductors.  

Then, I do a first approximation of cash flow: cash from sales of electricity, in that local installation, minus the possible maintenance costs. After I calculate that gross margin of cash,  I compare it to the investment capital I had calculated before, and I try to estimate provisionally the time of return on investment. Once this done, I add maintenance costs to my sauce. I think that the best way of estimating these is to assume a given lifecycle of complete depreciation in the technology installed, and to count maintenance costs as the corresponding annual amortization.         

[1] Phiri, W. K., Vanzo, D., Banda, K., Nyirenda, E., & Nyambe, I. A. (2021). A pseudo-reservoir concept in SWAT model for the simulation of an alluvial floodplain in a complex tropical river system. Journal of Hydrology: Regional Studies, 33, 100770. https://doi.org/10.1016/j.ejrh.2020.100770.

[2] Arnold, J.G., Srinivasan, R., Muttiah, R.S., Williams, J.R., 1998. Large area hydrological modelling and assessment: Part I. Model development. J. Am. Water Resour. Assoc. 34, 73–89.

[3] Neitsch, S.L., Arnold, J.G., Kiniry, J.R., Williams, J.R., 2005. “Soil and Water Assessment Tool Theoretical Documentation.” Version 2005. Blackland Research Center, Texas.

[4] Harvey, J.W., Schaffranek, R.W., Noe, G.B., Larsen, L.G., Nowacki, D.J., O’Connor, B.L., 2009. Hydroecological factors governing surface water flow on a low-gradient floodplain. Water Resour. Res. 45, W03421, https://doi.org/10.1029/2008WR007129.

[5] Sharma, R., & Malaviya, P. (2021). Management of stormwater pollution using green infrastructure: The role of rain gardens. Wiley Interdisciplinary Reviews: Water, 8(2), e1507. https://doi.org/10.1002/wat2.1507

[6] Li, J., Liu, F., & Li, Y. (2020). Simulation and design optimization of rain gardens via DRAINMOD and response surface methodology. Journal of Hydrology, 585, 124788. https://doi.org/10.1016/j.jhydrol.2020.124788

[7] Venvik, G., & Boogaard, F. C. (2020). Infiltration capacity of rain gardens using full-scale test method: effect of infiltration system on groundwater levels in Bergen, Norway. Land, 9(12), 520. https://doi.org/10.3390/land9120520

[8] Ma, Y., Jiang, Y., & Swallow, S. (2020). China’s sponge city development for urban water resilience and sustainability: A policy discussion. Science of the Total Environment, 729, 139078. https://doi.org/10.1016/j.scitotenv.2020.139078

[9] Sun, J., Cheshmehzangi, A., & Wang, S. (2020). Green infrastructure practice and a sustainability key performance indicators framework for neighbourhood-level construction of sponge city programme. Journal of Environmental Protection, 11(2), 82-109. https://doi.org/10.4236/jep.2020.112007

[10] Liu, Yan, Wang, Wei, Ghadimi, Noradin, 2017. Electricity load forecasting by an improved forecast engine for building level consumers. Energy 139, 18–30. https://doi.org/10.1016/j.energy.2017.07.150

[11] Gollou, Abbas Rahimi, Ghadimi, Noradin, 2017. A new feature selection and hybrid forecast engine for day-ahead price forecasting of electricity markets. J. Intell. Fuzzy Systems 32 (6), 4031–4045.

[12] Aghajani, Gholamreza, Ghadimi, Noradin, 2018. Multi-objective energy manage- ment in a micro-grid. Energy Rep. 4, 218–225.

[13] Yu, Dongmin, Ghadimi, Noradin, 2019. Reliability constraint stochastic UC by considering the correlation of random variables with Copula theory. IET Renew. Power Gener. 13 (14), 2587–2593.

[14] Cai, X., Ye, F., & Gholinia, F. (2020). Application of artificial neural network and Soil and Water Assessment Tools in evaluating power generation of small hydropower stations. Energy Reports, 6, 2106-2118. https://doi.org/10.1016/j.egyr.2020.08.010.

[15] Cordova M, Finardi E, Ribas F, de Matos V, Scuzziato M. Performance evaluation and energy production optimization in the real-time operation of hydropower plants. Electr Pow Syst Res 2014;116:201–7.   http://dx.doi.org/ 10.1016/j.epsr.2014.06.012  

[16] Vieira, D. A. G., Guedes, L. S. M., Lisboa, A. C., & Saldanha, R. R. (2015). Formulations for hydroelectric energy production with optimality conditions. Energy Conversion and Management, 89, 781-788.

[17] Botterud, A., Levin, T., & Koritarov, V. (2014). Pumped storage hydropower: benefits for grid reliability and integration of variable renewable energy (No. ANL/DIS-14/10). Argonne National Lab.(ANL), Argonne, IL (United States). https://publications.anl.gov/anlpubs/2014/12/106380.pdf

Le Catch 22 dans ce jardin d’Eden

Ça fait un sacré bout de temps depuis ma dernière mise à jour en français sur ce blog, « Discover Social Sciences ». Je n’avais pas écrit en français depuis printemps 2020. Pourquoi je recommence maintenant ? Probablement parce que j’ai besoin d’arranger les idées dans ma tête. Il se passe beaucoup de choses, cette année, et j’avais découvert, déjà en 2017, qu’écrire en français m’aide à mettre de l’ordre dans le flot de mes pensées.

Je me concentre sur un sujet que j’avais déjà développé dans le passé et que je vais présenter à une conférence, ce vendredi. Il s’agit du concept que j’avais nommé « Étangs énergétiques » auparavant et que je présente maintenant comme « Projet aqueduc ». Je commence avec une description générale du concept et ensuite je vais passer en revue un peu de littérature récente sur le sujet.

Oui, bon, le sujet. Le voilà. Il s’agit d’un concept technologique qui combine la rétention contrôlée de l’eau dans les écosystèmes placés le long des fleuves et des rivières avec de la génération d’électricité avec les turbines hydrauliques, le tout sur la base des structures marécageuses. Du point de vue purement hydrologique, une rivière est une gouttière qui collecte l’eau de pluie qui tombe sur la surface de son bassin. Le lit de la rivière est une vallée inclinée qui connecte les points le moins élevés du terrain en question et de de fait l’eau de pluie converge des tous les points du bassin fluvial vers l’embouchure de la rivière.

La civilisation humaine sédentaire est largement basée sur le fait que les bassins fluviaux ont la capacité de retenir l’eau de pluie pour un certain temps avant qu’elle s’évapore ou coule dans la rivière. Ça se retient à la surface – en forme des lacs, étangs ou marécages – et ça se retient sous terre, en forme des couches et des poches aquifères diverses. La rétention souterraine dans les poches aquifères rocheuses est naturellement permanente. L’eau retenue dans une couche aquifère reste là jusqu’au moment où nous la puisons. En revanche, la rétention superficielle ainsi que celle dans les couches aquifères souterraines est essentiellement temporaire. L’eau y est ralentie dans sa circulation, aussi bien dans son mouvement physique vers les points les plus bas du bassin local (la rivière du coin) que dans son évaporation vers l’atmosphère. L’existence même des fleuves et des rivières est aussi une manifestation de circulation ralentie. Le lit de la rivière n’arrive pas à évacuer en temps réel toute l’eau qui s’y agglomère et c’est ainsi que les rivières ont de la profondeur : cette profondeur est la mesure de rétention temporaire de l’eau de pluie.

Ces mécanismes fondamentaux fonctionnent différemment en fonction des conditions géologiques. Maintenant, je me concentre sur les conditions que je connais dans mon environnement à moi, donc sur les écosystèmes des plaines et des vallées de l’Europe du Nord, soit grosso modo au nord des Alpes. Ces écosystèmes sont pour la plupart des moraines post-glaciales de fond, donc c’est de la terre littéralement labourée, sculptée et dénivelée par les glaciers. Il n’y a pas vraiment beaucoup de poches aquifères profondes dans la roche de base, en revanche nous avons beaucoup de couches aquifères relativement proches de la surface. Par conséquent, il n’y a pas beaucoup d’accumulation durable de l’eau, à la différence de l’Europe du Sud et de l’Afrique du Nord, où les poches aquifères rocheuses peuvent retenir des quantités importantes d’eau pendant des décennies, voir des siècles. La circulation de l’eau dans ces écosystèmes des plaines est relativement lente – beaucoup plus lente que dans la montagne – ce qui favorise la présence des rivières larges et pas vraiment très profondes ainsi que la formation des marécages.

Dans ces plaines post-glaciales de l’Europe du Nord, l’eau coule lentement, s’accumule peu et s’évapore vite. La forme idéale des précipitations dans ces conditions géologiques c’est de la neige abondante en hiver – qui fond lentement, goutte par goute, au printemps – ainsi que des pluies lentes en longues. La moraine post-glaciale absorbe bien de l’eau qui arrive lentement, mais n’est pas vraiment faite pour absorber des pluies torrentielles. Avec le changement climatique, les précipitations ont changé. Il y a beaucoup moins de neige en hiver en beaucoup plus des pluies violentes. Si nous voulons avoir du contrôle de notre système hydrologique, il nous faut des technologies de rétention d’eau pour compenser des variations temporaires.

Bon, ça c’est le contexte de mon idée et voilà l’idée elle-même. Elle consiste à créer des structures marécageuses semi-artificielles dans la proximité des rivières et les remplir avec de l’eau pompée desdites rivières. La technologie de pompage est celle du bélier hydraulique : une pompe qui utilise l’énergie cinétique de l’eau courante. Le principe général est un truc ancien. D’après ce que j’ai lu à ce sujet, le principe de base, sous la forme de la roue à aubes , fût déjà en usage dans la Rome ancienne, était très utilisé dans les villes Européennes jusqu’à la fin du 18ème siècle. La technologie du bélier hydraulique – une pompe qui utilise ladite énergie cinétique de l’eau dans un mécanisme similaire au muscle cardiaque – fût victime des aléas de l’histoire. Inventée en 1792 par Joseph de Montgolfier (oui, l’un des fameux frères-ballon), cette technologie n’avait jamais eu l’occasion de montrer tous ses avantages. en 1792 (le même qui, quelques années plus tôt, fit voler, avec son frère Étienne, le premier ballon à air chaud). Au 19ème siècle, avec la création des systèmes hydrauliques modernes avec l’eau courante dans les robinets, les technologies de pompage devaient offrir assez de puissance pour assurer une pression suffisante au niveau des robinets et c’est ainsi que les pompes électriques avaient pris la relève. Néanmoins, lorsqu’il s’agit de pomper lentement de l’eau courante des rivières vers les marécages artificiels, le bélier hydraulique est suffisant.

« Suffisant pour faire quoi exactement ? », peut-on demander. Voilà donc le reste de mon idée. Un ou plusieurs béliers hydrauliques sont plongés dans une rivière. Ils pompent l’eau de la rivière vers des structures marécageuses semi-artificielles. Ces marécages servent à retenir l’eau de pluie (qui coule déjà dans le cours de la rivière). L’eau de la rivière que je pompe vers les marécages c’est l’eau de pluie qui avait gravité, en amont, vers le lit de la rivière. Une fois dans les marécages, cette eau va de toute façon finir par graviter vers le lit de la rivière à quelque distance en amont. Pompage et rétention dans les marécages servent à ralentir la circulation de l’eau dans l’écosystème local. Circulation ralentie veut dire que plus d’eau va s’accumuler dans cet écosystème, comme une réserve flottante. Il y aura plus d’eau dans les couches aquifères souterraines, donc plus d’eau dans les puits locaux et – à la longue – plus d’eau dans la rivière elle-même, puisque l’eau dans la rivière c’est l’eau qui y avait coulé depuis et à travers les réservoirs locaux.

Jusqu’à ce point-là, l’idée se présente donc de façon suivante : rivière => bélier hydraulique => marécages => rivière. Je passe plus loin. Le pompage consiste à utiliser l’énergie cinétique de l’eau courante. L’énergie, ça se conserve par transformation. L’énergie cinétique de l’eau courante se transforme en énergie cinétique de la pompe, qui à son tour se transforme en énergie cinétique du flux vers les marécages.

La surface des marécages est placée au-dessus du lit de la rivière, à moins qu’ils ne soient un polder, auquel cas il n’y a pas besoin de pompage. Une fois l’eau est déversée dans les marécages, ceux-là absorbent donc, dans leur masse, l’énergie cinétique du flux qui se transforme en énergie potentielle de dénivellation. Et si nous amplifions ce phénomène ? Si nous utilisions l’énergie cinétique captée par le bélier hydraulique de façon à minimiser la dispersion dans la masse des marécages et de créer un maximum d’énergie potentielle ? L’énergie potentielle et proportionnelle à l’élévation relative. Plus haut je pompe l’eau de la rivière, plus d’énergie potentielle je récupère à partir de l’énergie cinétique du flux pompé. La solution la plus évidente serait une installation de pompage-turbinage, donc le réservoir de rétention devrait être placé sérieusement plus haut que la rivière. Quoi qu’apparemment la plus évidente et porteuse des principes de base intéressants, cette solution a ses défauts en ce qui concerne sa flexibilité et son coût.

Le principe de base à retenir c’est l’idée d’utiliser l’énergie potentielle de l’eau pompée à une certaine élévation comme un de facto réservoir d’énergie électrique. Il suffit de placer des turbines hydro-électriques en aval de l’eau stockée en élévation. En revanche, les installations de pompage-turbinage sont très coûteuses et très exigeantes en termes d’espace. Le réservoir supérieur dans les installations de pompage-turbinage est censé être soit un lac semi-artificiel soit un réservoir complètement artificiel en de tour, certainement pas un marécage. Il est donc temps que j’explique pourquoi je suis tant attaché à cette forme hydrologique précise. Les marécages sont relativement peu chers à créer et à maintenir, tout en étant relativement faciles à placer près de et de combiner avec les habitations humaines. Par « relativement » je veux dire en comparaison au pompage-turbinage.

Le marécage est un endroit symboliquement négatif dans notre culture. Le mal est tapi dans les marécages. Les marécages sont malsains. Ma théorie tout à fait privée à ce sujet est que dans le passé les colonies humaines, fréquemment celles qui ont finalement donné naissance à des villes, étaient localisées près des marécages. Probablement c’était parce que le niveau d’eau souterraine dans des tels endroits est favorablement haut. Il est facile d’y creuser des puits, d’épandre des fossés d’irrigation, petit gibier y abonde. Seulement voilà, lorsque les homo sapiens abondent, ils se différencient inévitablement en hominides rustiques d’une part et les citadins d’autre part. Ce partage est un mécanisme de base de la civilisation humaine. La campagne produit de la nourriture, la ville produit des nouveaux rôles sociaux, à travers interaction intense dans un espace densément peuplé. L’un des aspects fondamentaux de la ville est qu’elle sert de laboratoire expérimental permanent pour nos technologies, à travers la construction et la reconstruction d’immeubles. Oui, l’architecture, en compagnie du textile, du bâtiment naval et de la guerre, ont toujours été les activités humaines par excellence orientées sur l’innovation technologique.

La ville veut donc dire le bâtiment et le bâtiment a besoin de terre vraiment ferme. Les marécages deviennent ennemis. Il faut les assécher et les séparer durablement de la circulation hydrologique naturelle qui les eût formés pendant des millénaires. Les humains et les marécages ce fût donc un mariage naturel au début, suivie par une crise conjugale due à la nécessité d’apprendre comment faire de la technologie nouvelle et maintenant la technologie vraiment nouvelle rend possible une médiation conjugale dans ce couple. Il y a tout un courant de recherche et innovation architecturale, concentré autour des concepts tels que « les jardins de pluie » (Sharma & Malaviya 2021[1] ; Li, Liu & Li 2020[2] ; Venvik & Boogaard 2020[3]) ou « les villes éponges » (Ma, Jiang & Swallow2020[4] ; Sun, Cheshmehzangi & Wang 2020[5]). Nous sommes en train de développer des technologies qui rendent la cohabitation entre villes et marécages non seulement possible mais bénéfique pour l’environnement et pour les citadins en même temps.

Question : comment utiliser le principe de base de pompage-turbinage, donc le stockage d’énergie potentielle de l’eau placée en élévation, sans construire des structures de pompage-turbinage et en présence des structures marécageuses à la limite de la ville et de la campagne ? Réponse : à travers la construction des tours relativement petites et légères, avec des petits réservoirs d’égalisation au sommet de chaque tour. Un bélier hydraulique bien construit rend possible d’élever l’eau par 20 mètres environ. On peut imaginer donc un réseau des béliers hydrauliques installés dans le cours d’une rivière et connectés à des petites tours de 20 mètres chacune, où chaque tour est équipée d’un tuyau de descente vers les marécages et le tuyau est équipé des petites turbines hydro-électriques.

L’idée complète se présente donc comme suit : rivière => bélier hydraulique => l’eau monte => tours légères de 20 mètres avec des petits réservoirs d’égalisation au sommet => l’eau descend => petites turbines hydro-électriques => marécages => l’eau s’accumule => circulation hydrologique naturelle à travers le sol => rivière.

Bon, où est le Catch 22 dans ce jardin d’Eden ? Dans l’aspect économique. Les béliers hydrauliques de bonne qualité, tels qu’ils sont produits aujourd’hui, sont chers. Il y a très peu de fournisseurs solides de cette technologie. La plupart des béliers hydrauliques en utilisation sont des machins artisanaux à faible puissance et petit débit. L’infrastructure des tours de siphonage avec les turbines hydro-électriques de bonne qualité, ça coûte aussi. Si on veut être sérieux côté électricité, faut équiper tout ce bazar avec des magasins d’énergie. Toute l’infrastructure aurait besoin des frais de maintenance que je ne sais même pas comment calculer. Selon mes calculs, la vente d’électricité produite dans ce circuit hydrologique pourrait assurer un retour sur l’investissement pas plus court que 8 – 9 ans et encore, c’est calculé avec des prix d’électricité vraiment élevés.

Faut que j’y pense (plus).    

[1] Sharma, R., & Malaviya, P. (2021). Management of stormwater pollution using green infrastructure: The role of rain gardens. Wiley Interdisciplinary Reviews: Water, 8(2), e1507. https://doi.org/10.1002/wat2.1507

[2] Li, J., Liu, F., & Li, Y. (2020). Simulation and design optimization of rain gardens via DRAINMOD and response surface methodology. Journal of Hydrology, 585, 124788. https://doi.org/10.1016/j.jhydrol.2020.124788

[3] Venvik, G., & Boogaard, F. C. (2020). Infiltration capacity of rain gardens using full-scale test method: effect of infiltration system on groundwater levels in Bergen, Norway. Land, 9(12), 520. https://doi.org/10.3390/land9120520

[4] Ma, Y., Jiang, Y., & Swallow, S. (2020). China’s sponge city development for urban water resilience and sustainability: A policy discussion. Science of the Total Environment, 729, 139078. https://doi.org/10.1016/j.scitotenv.2020.139078

[5] Sun, J., Cheshmehzangi, A., & Wang, S. (2020). Green infrastructure practice and a sustainability key performance indicators framework for neighbourhood-level construction of sponge city programme. Journal of Environmental Protection, 11(2), 82-109. https://doi.org/10.4236/jep.2020.112007

Big Black Swans

Oops! Another big break in blogging. Sometimes life happens so fast that thoughts in my head run faster than I can possibly write about them. This is one of those sometimeses. Topics for research and writing abound, projects abound, everything is changing at a pace which proves challenging to a gentleman in his 50ies, such as I am. Yes, I am a gentleman: even when I want to murder someone, I go out of my way to stay polite.

I need to do that thing I periodically do, on this blog. I need to use published writing as a method of putting order in the chaos. I start with sketching the contours of chaos and its main components, and then I sequence and compartmentalize.

My chaos is made of the following parts:

>> My research on collective intelligence

>> My research on energy systems, with focus on investment in energy storage

>> My research on the civilisational role of cities, and on the concept of the entire human civilisation, such as we know it today, being a combination of two productive mechanisms: production of food in the countryside, and production of new social roles in the cities.

>> Joint research which I run with a colleague of mine, on the reproduction of human capital

>> The project I once named Energy Ponds, and which I recently renamed ‘Project Aqueduct’, for the purposes of promoting it.

>> The project which I have just started, together with three other researchers, on the role of Territorial Defence Forces during the COVID-19 pandemic

>> An extremely interesting project, which both I and a bunch of psychiatrists from my university have provisionally failed to kickstart, on the analysis of natural language in diagnosing and treating psychoses

>> A concept which recently came to my mind, as I was working on a crowdfunding project: a game as method of behavioural research about complex decisional patterns.

Nice piece of chaos, isn’t it? How do I put order in my chaos? Well, I ask myself, and, of course, I do my best to answer honestly the following questions: What do I want? How will I know I have what I want? How will other people know I have what I want? Why should anyone bother? What is the point? What do I fear? How will I know my fears come true? How will other people know my fears come true? How do I want to sequence my steps? What skills do I need to learn?

I know I tend to be deceitful with myself. As a matter of fact, most of us tend to. We like confirming our ideas rather than challenging them. I think I can partly overcome that subjectivity of mine by interweaving my answers to those questions with references to published scientific research. Another way of staying close to real life with my thinking consists in trying to understand what specific external events have pushed me to engage in the different paths, which, as I walk down all of them at the same time, make my personal chaos.

In 2018, I started using artificial neural networks, just like that, mostly for fun, and in a very simple form. As I observed those things at work, I developed a deep fascination with intelligent structures, and just as deep (i.e. f**king hard to phrase out intelligibly) an intuition that neural networks can be used as simulators of collectively intelligent social structures.

Both of my parents died in 2019, exactly at the same age of 78, having spent the last 20 years of their respective individual lives in complete separation from each other to the point of not having exchanged a spoken or written word over those last 20 years. That changed my perspective as regards subjectivity. I became acutely aware how subjective I am in my judgement, and how subjective other people are, most likely. Pandemic started in early 2020, and, almost at the same moment, I started to invest in the stock market, after a few years of break. I had been learning at an accelerated pace. I had been adapting to the reality of high epidemic risk – something I almost forgot since I had a devastating scarlatina at the age of 9 – and I had been adapting to a subjectively new form of economic decisions (i.e. those in the stock market). That had been the hell of a ride.

Right now, another piece of experience comes into the game. Until recently, in my house, the attic was mine. The remaining two floors were my wife’s dominion, but the attic was mine. It was painted in joyful, eye-poking colours. There as a lot of yellow and orange. It was mine. Yet, my wife had an eye for that space. Wives do, quite frequently. A fearsome ally came to support her: an interior decorator. Change has just happened. Now, the attic is all dark brown and cream. To me, it looks like the inside of a coffin. Yes, I know what the inside of a coffin looks like: I saw it just before my father’s funeral. That attic has become an alien space for me. I still have hard times wrapping my mind around how shaken I am with that change. I realize how attached am I to the space around me. If I am so strongly bound to colours and shapes in my vicinity, other people probably feel the same, and that triggers another intuition: we, humans, are either simple dwellers in the surrounding space, or we are architects thereof, and these are two radically different frames of mind.

I am territorial as f**k. I have just clarified it inside my head. Now, it is time to go back to science. In a first step, I am trying to connect those experiences of mine to my hypothesis of collective intelligence. Step by step, I am phrasing it out. We are intelligent about each other. We are intelligent about the physical space around us. We are intelligent about us being subjective, and thus we have invented that thing called language, which allows us to produce a baseline for intersubjective description of the surrounding world.

I am conscious of my subjectivity, and of my strong emotions (that attic, f**k!). Therefore, I want to see my own ideas from other people’s point of view. Some review of literature is what I need. I start with Peeters, M. M., van Diggelen, J., Van Den Bosch, K., Bronkhorst, A., Neerincx, M. A., Schraagen, J. M., & Raaijmakers, S. (2021). Hybrid collective intelligence in a human–AI society. AI & SOCIETY, 36(1), 217-238. https://doi.org/10.1007/s00146-020-01005-y . I start with it because I could lay my hands on an open-access version of the full paper, and, as I read it, many bells ring in my head. Among all those bells ringing, the main one refers to the experience I had with the otherwise simplistic neural networks, namely the perceptrons possible to structure in an Excel spreadsheet. Back in 2018, I was observing the way that truly a dumb neural network was working, one made of just 6 equations, looped together, and had that realization: ‘Blast! Those six equations together are actually intelligent’. This is the core of that paper by Peeters et al.. The whole story of collective intelligence had become a thing when Artificial Intelligence started spreading throughout society, thus it is generally the same thing in scientific literature as it is with me individually. Conscious, inquisitive interaction with artificial intelligence seems to awaken an entirely new worldview, where we, humans, can see at work an intelligent fruit of our own intelligence.

I am trying to make one more step, from bewilderment to premises and hypotheses. In Peeters et al., three big intellectual streams are named: (1) the technology-centric perspective (2) the human-centric one, and finally (3) that focused on the concept of collective intelligence-centric perspective. The third one sounds familiar, and so I dig into it. The general idea here is that humans can put their individual intelligences into a kind of interaction which is smarter than those individual ones. This hypothesis is a little counterintuitive – if we consider electoral campaigns or Instagram – but it becomes much more plausible when we think about networks of inventors and scientists. Peeters et al. present an interesting extension to that, namely collectively intelligent agglomerations of people and technology. This is exactly what I do when I do empirical research and use a neural network as simulator, with quantitative data in it. I am one human interacting with one simple piece of AI, and interesting things come out of it. 

That paper by Peeters et al. cites a book: Sloman, S., Sloman, S. A., & Fernbach, P. (2018). The knowledge illusion: Why we never think alone (Penguin). Before I pass to my first impressions about that book, another sidekick. In 1993, one of the authors of that book, Aaron Sloman, wrote an introduction to another book, this one being a collection of proceedings (conference papers in plain human lingo) from a conference, grouped under the common title: Prospects for Artificial Intelligence (Hogg & Humphreys 1993[1]). In that introduction, Aaron Sloman claims that using Artificial Intelligence as simulator of General Intelligence requires a specific approach, which he calls ‘design-based’, where we investigate the capabilities and the constraints within which intelligence, understood as a general phenomenon, has to function. Based on those constraints, requirements can be defined, and, consequently, the way that intelligence is enabled to meet them, through its own architecture and mechanisms.  

We jump 25 years, from 1993, and this is what Sloman, Sloman & Fernbach wrote in the introduction to “The knowledge illusion…”: “This story illustrates a fundamental paradox of humankind. The human mind is both genius and pathetic, brilliant and idiotic. People are capable of the most remarkable feats, achievements that defy the gods. We went from discovering the atomic nucleus in 1911 to megaton nuclear weapons in just over forty years. We have mastered fire, created democratic institutions, stood on the moon, and developed genetically modified tomatoes. And yet we are equally capable of the most remarkable demonstrations of hubris and foolhardiness. Each of us is error-prone, sometimes irrational, and often ignorant. It is incredible that humans are capable of building thermonuclear bombs. It is equally incredible that humans do in fact build thermonuclear bombs (and blow them up even when they don’t fully understand how they work). It is incredible that we have developed governance systems and economies that provide the comforts of modern life even though most of us have only a vague sense of how those systems work. And yet human society works amazingly well, at least when we’re not irradiating native populations. How is it that people can simultaneously bowl us over with their ingenuity and disappoint us with their ignorance? How have we mastered so much despite how limited our understanding often is?” (Sloman, Steven; Fernbach, Philip. The Knowledge Illusion (p. 3). Penguin Publishing Group. Kindle Edition)

Those readings have given me a thread, and I am interweaving that thread with my own thoughts. Now, I return to another reading, namely to “The Black Swan” by Nassim Nicolas Taleb, where, on pages xxi – xxii of the introduction, the author writes: “What we call here a Black Swan (and capitalize it) is an event with the following three attributes. First, it is an outlier, as it lies outside the realm of regular expectations, because nothing in the past can convincingly point to its possibility. Second, it carries an extreme impact (unlike the bird). Third, in spite of its outlier status, human nature makes us concoct explanations for its occurrence after the fact, making it explainable and predictable. I stop and summarize the triplet: rarity, extreme impact, and retrospective (though not prospective) predictability.fn3 A small number of Black Swans explain almost everything in our world, from the success of ideas and religions, to the dynamics of historical events, to elements of our own personal lives. Ever since we left the Pleistocene, some ten millennia ago, the effect of these Black Swans has been increasing. It started accelerating during the industrial revolution, as the world started getting more complicated, while ordinary events, the ones we study and discuss and try to predict from reading the newspapers, have become increasingly inconsequential”.

I combine the idea by Sloman, Sloman & Fernbach, that we, humans, can be really smart collectively by interaction between very limited and subjective individual intelligences, with the concept of Black Swan by Nassim Nicolas Taleb. I reinterpret the Black Swan. This is an event we did not expect to happen, and yet it happened, and it has blown a hole in our self-sufficient certainty that we understand what the f**k is going on. When we do things, we expect certain outcomes. When we don’t get the outcomes we expected, this is a functional error. This is local instance of chaos. We take that local chaos as learning material, and we try again, and again, and again, in the expectation of bringing order into the general chaos of existence. Our collectively intelligent learning  

Nassim Nicolas Taleb claims that our culture tends to mask the occurrence of Black Swan-type outliers as just the manifestation of otherwise recurrent, predictable patterns. We collectively internalize Black Swans. This is what a neural network does. It takes an obvious failure – the local error of optimization – and utilizes it as valuable information in the next experimental round. The network internalizes a series of f**k-ups, because it never hits the output variable exactly home, there is always some discrepancy, at least a tiny one. The fact of being wrong about reality becomes normal. Every neural network I have worked with does the same thing: it starts with some substantial magnitude of error, and then it tends to reduce that error, at least temporarily, i.e. for at least a few more experimental rounds.   

This is what a simple neural network – one of those I use in my calculations – does with quantitative variables. It processes data so as to create error, i.e. so as to purposefully create outliers located out of the expected. Those neural networks purposefully create Black Swans, those abnormalities which allow us to learn. Now, what is so collective about neural networks? Why do I intuitively associate my experiments with neural networks to collective intelligence rather than to the individual one? Well, I work with socio-economic quantitative variables. The lowest level of aggregation I use is the probability of occurrence in a specific event, but this is really low aggregation for me. Most of my data is like Gross Domestic Product, average hours worked per person per year, average prices of electricity etc. This is essentially collective data, in the sense that no individual intelligence can possibly be observed as for its density of population or for its inflation. There needs to be a society in place for those metrics to exist at all.

When I work with that type of data, I assume that many people observe and gauge it, then report and aggregate their observations etc. Many people put a lot of work into making those quantitative variables both available and reliable. I guess it is important, then. When some kind of data is that important collectively, it is very likely to reflect some important aspect of collective reality. When I run that data through a neural network, the latter yields a simulation of collective action and its (always) provisional outcomes.

My neural network (I mean the one on my computer, not the one in my head) takes like 0.05 of local Gross Domestic Product, then 0.4 of average consumption of energy per capita, maybe 0.09 of inflation in consumer prices, plus some other stuff in random proportions, and sums up all those small portions of whatever is important as collectively measured socio-economic outcome. Usually, that sum is designated as ‘h’, or vector of aggregate input. Then, my network takes that h and puts it into a neural activation which, in most cases, is the hyperbolic tangent hyperbolic tangent AKA (e2h – 1) / (e2h +1). When we learn by trial and error, a hypothetical number e2h measures the force which the neural network reacts with to a complex stimulation from a set of variables xi. The ‘e2’ part of that hypothetical reaction is constant and equals e2 = 7,389056099, whilst h is a variable parameter, specific to the given phenomenal occurrence. The parameter h is roughly proportional to the number of variables in the source empirical set. The more complex the reality I process with the neural network, i.e. the more variables I split my reality into, the greater is the value of h. In other words, the more complexity, the more is the neuron, based on the expression e2h, driven away from its constant root e2. Complexity in variables induces greater swings in the hyperbolic tangent, i.e. greater magnitudes of error, and, consequently, longer strides in the process of learning.         

Logically, the more complex social reality I represent with quantitative variables, the bigger Black Swans the neural network produces as it tries to optimize one single variable chosen as the desired output of the neurally activated input.   

[1] Hogg, D., & Humphreys, G. W. (1993). Prospects for Artificial Intelligence: Proceedings of AISB’93, 29 March-2 April 1993, Birmingham, UK (Vol. 17). IOS Press.

When it plays out, it looks nasty

I feel like using my hypothesis of collectively intelligent social structures in other fields than just energy and urbanisation, which I have been largely doing so far. This time, I want to make a case for individual freedom as both a factor and a manifestation of collective intelligence. There is a population of humans. Each human has m possible states of being. As soon as two humans interact, one m states of being in the first human interacts with the other m states of being in the other human. It is like an existential geometrical square: those two humans together have m*m = m2 collective states of being. Generally, n humans, with m possible states of being in each of them, can produce mn different states of being together. When n gets substantial, like 38 million people in my home country, Poland, you can hardly expect all of us 38 million Poles having the repertoire of freedom in our behavioural patterns. Some of us will have 3m actually happening states of being, some other will soar into 6m alternative ways of being in the world, whilst still some other others will modestly stick to 0,3m. In that large population, the standard m ways of existing will be an expected state, thus an arithmetical average or an expected interval around it.   

Collectively intelligent structures learn by experimenting with many alternative states of themselves. Up to a point, the more such alternative states, the more and better we can learn. There is probably a point where ‘the more’ becomes ‘too much to process’, and then, we face a fork on the road: either we simply ignore some alternative versions of ourselves and we truly learn just from those which we can cover inside our cognitive span, or we try to experiment with everything we can possibly be, and chaos develops. I understand freedom, at the collective level, as the flexibility in shifting between those different states of being. Organized, collective freedom is the ability to explore the sweet spot of transition between order and chaos, and the ability to experiment with as many alternative versions of ourselves as we possibly can. Those collectively defined alternative realities always follow the basic logic of mn. At the end of the day, there are as many versions of us being together as there are us, for one, namely the ‘n’ exponent, and as many as there are possible states of being in the average individual among n, and this is the ‘m’ base.

Degrees of freedom in the average member of society are the foundation of collectively intelligent learning. I guess this is a mathematical argument for individual freedom in legal and political systems. As I think about my whole hypothesis of collectively intelligent social structures, I inevitably ask the question which any social scientist needs to ask: what is the practical usefulness of all that stuff? Social sciences are applied sciences, at the end of the day. However abstract I go in my intellectual peregrinations, my findings and methods need to serve in real life, for designing policies, business strategies, business plans etc. The empirical method I have developed around that whole thing of collective intelligence opens on two practical applications. Firstly, it allows non-arbitrary testing of various empirical observables as actual social outcomes. In policies and business strategies, and, by the way, in the whole realm of social sciences, there is that curse of arbitrary orientations. ‘People strive to maximize profit’. ‘No, they want to optimize dynamic equilibriums in their social games’. ‘Well, maybe, but we can and should educate people towards social justice and environmentally rational behaviour’ etc. etc. All that chatter abounds in literature which deems itself ‘scientific’, and yet it is 100% metaphysics, with no scientific grounds at all. I think my method allows working around that metaphysical part and testing human populations for the actual outcomes they collectively, objectively pursue. Here comes an interesting question: are our goals collective or individual? The more I think about it, the more I am convinced they are collective. When I ask myself about my own goals, at least those which I phrase out explicitly in my mind, they are all sort of categorical rather than idiosyncratically my own. I pursue the types of goals which many other people pursue in their existence. I just hop on those specific wagons, with my own backpack.  

Secondly, my method allows exploring the issue of Black Swans, i.e. outlier events, which suddenly become key drivers of social change. The method I have developed allows simulating something like a social chain reaction. An unexpected triggering event happens, and it is unexpected because from our point of view it is random. That triggers a collection of events which we could otherwise fathom, but they have been in the refrigerator of history so far. Now, they are triggered into existence, and, at the same time, the overall cohesion of the social structure weakens, at least temporarily. New things start happening, and old things happen sort of more loosely and chaotically than they used to. I have discovered that depending on the exact orientation assigned a priori to the social structure I study, those social chain reactions can we essentially predictable, completely unpredictable, or, in still another case, we can calm them down exaggeratedly quickly, without really learning from them.

All in all, the method of using a simple neural network as social simulator, which I developed in connection with my hypothesis of collectively intelligent social structures, allows what I perceive as very empiricist a study of social change, much freer of metaphysics than many other methods. Of course, a bit of metaphysics is unavoidable. What we use to call ‘quantitative variables’ in social sciences are always the mathematics of something we think that happens, and we think in terms of our language and culture.

Ooops, pardon my manners, I have gone into philosophy again. Philosophy is nice, but when I stay in this realm longer than what is strictly necessary for feeling like an intellectual, I start feeling as too much of an intellectual and my apish side calls for more ground under my feet. I use this blog for providing a current account of my intellectual journey, and of the actual projects which I am working on. I hope that the paragraphs above are (provisionally) sufficient as regards the intellectual journey, and I can pass to debriefing on my projects.

One of the projects I start working on is a platform for debt-based crowdfunding. This is some sort of comeback to the interest I had in financial schemes for the implementation of small installations in renewable energies. For the less initiated readers, I am quickly going through the basics. You probably know that if your cousin asks you to invest in his or her business, you can do it, on the basis of a private contract of partnership, and, in most countries, you don’t even go to jail afterwards. This is the market of private equity. You can also lend money to your cousin, you can agree as for the exact terms of the loan, and this is financing through private debt. The opposite of private is public, and therefore we have public capital markets on the opposite end of the spectrum. Stock markets are the most visible ones, and sort of next to them are the markets of publicly traded debt, where you can buy and sell bonds of all kinds: corporate, municipal, and sovereign. 

Between the strictly private and the regulated public, a transitional zone, of many shades and colours, is to be found.  Crowdfunding, sometimes called ‘societal funding’ or ‘communitarian funding’ dwells in this zone, precisely. The basic difference between crowdfunding and private finance strictly spoken is the largely aleatory, social-media-type creation of relations between investors and entrepreneurs. Crowdfunding happens essentially via digital platforms, where entrepreneurs auction their ventures and try to attract whoever is interested in them. Those digital platforms in themselves are marketing engines, essentially. On the other hand, the basic difference between public financial markets and crowdfunding is that the latter does not really allow tradability in financial positions. When I invest my money through crowdfunding, it is much more of a long-term commitment than investment via stock market. Less liquidity in my financial assets means more exposure to long-term risks, and yet less exposure to short-term volatility in market value.

In my own big picture of social reality, I put the emergence of crowdfunding in the same phenomenological bag as I put cryptocurrencies, progressively increasing supply of money in relation to real output in the economy (thus decreasing velocity of money), and increasingly cash-furnished corporate balance sheets. As a civilisation, we are building up a growing base of financial liquidity, and that means we are facing a quickening pace of depreciation in technological assets, and thus we are in the middle of accelerated technological change. Now, a little word is due about the way I understand accelerated technological change. I have encountered quite well-articulated views that technological change is currently disappointingly slow as compared to what we need. Well, maybe, but in strictly spoken business terms, when a piece of technology which I purchased last year ages morally twice as fast as those which I purchased 5 years ago, because new generations of the same equipment pop up faster and faster, this is accelerated technological change, and, as a businessperson, I need to figure out a strategy to cope with that change.

Here, my own point of view of that phenomenon called ‘financialization’ differs significantly from a lot of other researchers. The mainstream doctrine says that increased financialization is a bad thing, it destabilizes the economic system, and it contributes to social inequalities. I think that financialization is the by-product of something else. It is an otherwise rational coping mechanism to smooth and amortize quick social change which, without financialization, could take very nasty forms, like global wars, massive disappearance of human settlements and much greater damage to natural environment than what we use to bitch and moan about today. Just imagine that somewhere in Europe, 5 million people in a post-industrial spot cannot afford to pay for electricity anymore and they start burning wood and coal in stoves instead. This is what could happen in the presence of quick technological change and in the absence of that horrible financialization.     

Crowdfunding is essentially attached to new ideas and new business structures. It is seed capital or early development capital. When I invest my money through crowdfunding, I am opening a long-term position in something essentially young, burgeoning and full of uncertainty. One hundred years ago, mustering capital for such a venture would take an entrepreneur years of patient contacts with potential investors. Now, it can take months or even weeks, and this is the tangible gain of time through the use of digital platforms.   

That introduction kept in mind, I get closer to the main thread of that project in crowdfunding, namely to the new regulations thereof, likely to enter into force in Poland this autumn, based on recent regulations of the European Union as a whole. I am passing in review the REGULATION (EU) 2020/1503 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 7 October 2020 on European crowdfunding service providers for business, and amending Regulation (EU) 2017/1129 and Directive (EU) 2019/1937, to find at https://eur-lex.europa.eu/legal-content/PL/TXT/?uri=CELEX:32020R1503 . As I usually do, I start from the end, more specifically from Annex II, titled SOPHISTICATED INVESTORS FOR THE PURPOSE OF THIS REGULATION.

A sophisticated investor is an investor who possesses the awareness of the risks associated with investing in capital markets and adequate resources to undertake those risks without exposing itself to excessive financial consequences. Sophisticated investors may be categorised as such if they meet identification criteria, which, in turn, differ according to the legal personality of the entity. Legal persons (like a bunch of folks in a business partnership), are assumed to be sophisticated in their investments if they meet at least one of the following criteria: (a) own funds of at least EUR 100 000 (b) net turnover of at least EUR 2 000 000 (c) balance sheet of at least EUR 1 000 000.

On the other hand, natural persons can call themselves sophisticated investors when the meet at least two of the following criteria:

>> (a) personal gross income of at least EUR 60 000 per fiscal year, or a financial instrument portfolio, defined as including cash deposits and financial assets, that exceeds EUR 100 000;

>> (b) the investor works or has worked in the financial sector for at least one year in a professional position which requires knowledge of the transactions or services envisaged, or the investor has held an executive position for at least 12 months in a legal person considered as sophisticated investor;

>> (c) the investor has carried out transactions of a significant size on the capital markets at an average frequency of 10 per quarter, over the previous four quarters.

The whole distinction between ordinary investors and the sophisticated ones is in the degree of legal protection they are provided with. That distinction essentially taps into an older one, contained in the DIRECTIVE 2014/65/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 15 May 2014 on markets in financial instruments and amending Directive 2002/92/EC and Directive 2011/61/EU (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32014L0065 ). As it happens sometimes, protection turns out to be a limitation actually. Non-sophisticated investors are generally limited in the amounts of money they can invest, and the repertoire of financial instruments which they can invest in. If one wants not to be treated like a child, they have to make a special, written request to be treated as sophisticated investor, and whatever operator of financial platform is that request addressed to can accept or reject said request.    

The Polish prospective regulations on crowdfunding approach things from a different angle. By the way, they are just prospective regulations, and the only official version of that will which I could get my hands on is in Polish. For those who speak the beautiful language of my home country – distinctive, among others, by a record-level density of consonants in one word – I placed the current bill of this regulation in the archives of my blog, just here: https://discoversocialsciences.com/wp-content/uploads/2021/05/Projekt-crowdfunding-.docx . Polish regulators focus mostly on the concept of ‘key investment information sheet’, which I will allow myself to call KIIS in what follows, is present in the European regulations as well. The KIIS should warn prospective investors that the investing environment they have entered into entails risks that are covered neither by deposit guarantee schemes, nor by investor compensation schemes. The KIIS should reflect the specific features of lending-based and investment-based crowdfunding. To that end, specific and relevant indicators should be required. The KIIS should also take into account, where available, the specific features and risks associated with project owners, and should focus on material information about the project owners, the investors’ rights and fees, and the type of transferable securities, admitted instruments for crowdfunding purposes and loans offered. The KIIS should be drawn up by the project owners, because the project owners are in the best position to provide the information required to be included therein. However, since it is the crowdfunding service providers that are responsible for providing the KIIS to prospective investors, it is the crowd­funding service providers that should ensure that the KIIS is clear, correct and complete.

The specificity of the Polish regulations as regards the KIIS is largely in the addressees of that information. In the general European regulations, the KIIS is addressed to prospective and actual investors. In Polish regulations, it is strongly stressed that crowdfunding operators should communicate all their KIIS’s to the Financial Supervision Commission (PL: Komisja Nadzoru Finansowego, https://www.knf.gov.pl/en/ ), not later than 7 days before making the same KIIS available to prospective investors. On the other hand, the owner of the project subject to crowdfunding can publish the KIIS on their own platform only after the provider of crowdfunding does in on their own one. We have a sequence of KIISes. The first KIIS goes from the crowdfunding provider to the Financial Supervision Commission, which has at least 7 days to consider (what exactly?). The next KIIS goes from the crowdfunding provider to prospective investors, who also receive the last KIIS from the owner of the crowdfunded project in question.

In a general manner, those Polish regulations give a lot of discretionary prerogatives to the Financial Supervision Commission as regards crowdfunding providers. They can halt a crowdfunding project immediately, and for an essentially indefinite period of time, on the grounds of a simple suspicion. I don’t like it. Someone in charge with the Financial Supervision Commission is the first to know about a crowdfunded project, they can request any information about that project, they can halt the project whenever they want. That smells bad. That smells insider trading. That smells uncontrolled pressure on the owners of crowdfunded projects. Imagine: you start such a project, and then you have a phone call, I mean THE phone call. Someone tells you they know about your crowdfunding campaign, and they would willingly take 60% of your business for 50% of its book value. You refuse, and the next thing you know is your crowdfunding campaign being suspended for an unknown period of time. I know the scheme, I saw it play out, and when it plays out, it looks nasty, believe me. That means people close to the government taking over entire swaths of small business, and the kind of small business, which is particularly exposed to adverse actions, the emerging one.  

The type of riddle I like

Once again, I had quite a break in blogging. I spend a lot of time putting together research projects, in a network of many organisations, which I am supposed to bring to working together. I give it a lot of time and personal energy. It drains me a bit, and I like that drain. I like the thrill of putting together a team, agreeing about goals and possible openings. Since 2005, when I stopped running my own business and I settled for a quite academic career, I haven’t experienced that special kind of personal drive. I sincerely believe that every teacher should apply his or her own teaching in the everyday life of theirs, just to see if their teaching still corresponds to reality.

This is one of the reasons why I have made it a regular activity of mine to invest in the stock market. I teach economics, and the stock market is very much like the pulse of economics, in all its grades and shades, ranging from hardcore macroeconomic cycles, passing through the microeconomics of specific industries I am currently focusing on with my investment portfolio, and all the way down the path of behavioural economics. I teach management, as well, and putting together new projects in research is the closest I can come, currently, to management science being applied in real life.

Still, besides trying to apply my teaching in real life, I still do science. I do research, and I write about the things I think I have found out, on that research path of mine. I do a lot of research as regards the economics of energy. Currently, I am still revising a paper of mine, titled ‘Climbing the right hill – an evolutionary approach to the European market of electricity’. Around the topic of energy economics, I have built more general a method of studying quantitative socio-economic data, with the technical hypothesis that said data manifests collective intelligence in human social structures. It means that whenever I deal with a collection of quantitative socio-economic variables, I study the dataset at hand by assuming that each multivariate record line in the database is the local instance of an otherwise coherent social structure, which experimentins with many such specific instances of itself and selects those offering the best adaptation to the current external stressors. Yes, there is a distinct sound of evolutionary method in that approach.

Over the last three months, I have been slowly ruminating my theoretical foundations for the revision of that paper. Now, I am doing what I love doing: I am disrupting the gently predictable flow of theory with some incongruous facts. Yes, facts don’t know how to behave themselves, like really. Here is an interesting fact about energy: between 1999 and 2016, at the planetary scale, there had been more and more new cars produced per each new human being born. This is visualised in the composite picture below. Data about cars comes from https://www.worldometers.info/cars/ , whilst data about the headcount of population comes from the World Bank (https://data.worldbank.org/indicator/SP.POP.TOTL ).

Now, the meaning of all that. I mean, not ALL THAT (i.e. reality and life in general), just all that data about cars and population. Why do we consistently make more and more physical substance of cars per each new human born? Two explanations come to my mind. One politically correct and nicely environmentalist: we are collectively dumb as f**k and we keep overshooting the output of cars over and above the incremental change in population. The latter, when translated into a rate of growth, tends to settle down (https://data.worldbank.org/indicator/SP.POP.GROW ). Yeah, those capitalists who own car factories just want to make more and more money, and therefore they make more and more cars. Yeah, those corrupt politicians want to conserve jobs in the automotive industry, and they support it. Yeah, f**k them all! Yeah, cars destroy the planet!

I checked. The first door I knocked at was General Motors (https://investor.gm.com/sec-filings ). What I can see is that they actually make more and more operational money by making less and less cars. Their business used to be overshot in terms of volume, and now they are slowly making sense and money out of making less cars. Then I checked with Toyota (https://global.toyota/en/ir/library/sec/ ). These guys looks as if they were struggling to maintain their capacity to make approximately the same operational surplus each year, and they seem to be experimenting with the number of cars they need to put out in order to stay in good financial shape. When I say ‘experimenting’, it means experimenting upwards or downwards.

As a matter of fact, the only player who seems to be unequivocally making more operational money out of making more cars is Tesla (https://ir.tesla.com/#tab-quarterly-disclosure). In There comes another explanation – much less politically correct, if at all – for there being more cars made per each new human, and it says that we, humans, are collectively intelligent, and we have a good reason for making more and more cars per each new human coming to this realm of tears, and the reason is to store energy in a movable, possibly auto-movable a form. Yes, each car has a fuel tank or a set of batteries, in the case of them Teslas or other electric f**kers. Each car is a moving reservoir of chemical energy, immediately converted into kinetic energy, which, in turn, has economic utility. Making more cars with batteries pays off better than making more cars with combustible fuel in their tanks: a new generation of movable reservoirs in chemical energy is replacing an older generation thereof. 

Let’s hypothesise that this is precisely the point of each new human being coupled with more and more of a new car being made: the point is more chemical energy convertible into kinetic energy. Do we need to move around more, as time passes? Maybe, although I am a bit doubtful. Technically, with more and more humans being around in a constant space, there is more and more humans per square kilometre, and that incremental growth in the density of population happens mostly in cities. I described that phenomenon in a paper of mine, titled ‘The Puzzle of Urban Density And Energy Consumption’. That means that space available for travelling and needed to be covered, per individual capita of each human being, is actually decreasing. Less space to travel in means less need for means of transportation. 

Thus, what are we after, collectively? We might be preparing for having to move around more in the future, or for having to restructure the geography of our settlements. That’s possible, although the research I did for that paper about urban density indicates that geographical patterns of urbanization are quite durable. Anyway, those two cases sum up to some kind of zombie apocalypse. On the other hand, the fact of developing the amount of dispersed, temporarily stored energy (in cars) might be a manifestation of us learning how to build and maintain large, dispersed networks of energy reservoirs.

Isn’t it dumb to hypothesise that we go out of our way, as a civilisation, just to learn the best ways of developing what we are developing? Well, take the medieval cathedrals. Them medieval folks would keep building them for decades or even centuries. The Notre Dame cathedral in Paris, France, seems to be the record holder, with a construction period stretching from 1160 to 1245 (Bruzelius 1987[1]). Still, the same people who were so appallingly slow when building a cathedral could accomplish lightning-fast construction of quite complex military fortifications. When building cathedrals, the masters of stone masonry would do something apparently idiotic: they would build, then demolish, and then build again the same portion of the edifice, many times. WTF? Why slowing down something we can do quickly? In order to experiment with the process and with the technologies involved, sir. Cathedrals were experimental labs of physics, mathematics and management, long before these scientific disciplines even emerged. Yes, there was the official rationale of getting closer to God, to accomplish God’s will, and, honestly, it came handy. There was an entire culture – the medieval Christianity – which was learning how to learn by experimentation. The concept of fulfilling God’s will through perseverant pursuit, whilst being stoic as regards exogenous risks, was excellent a cultural vehicle to that purpose.

We move a few hundreds of years in time, to the 17th century. The cutting edge of technology is to find in textile and garments (Braudel 1992[2]), and the peculiarity of the European culture consisted in quickly changing fashions, geographically idiosyncratic and strongly enforced through social peer pressure. The industry of garments and textile was a giant experimental lab of business and management, developing the division of labour, the management of supply chains, quick study of subtle shades in customers’ tastes and just as quick adaptation thereto. This is how we, Europeans, prepared for the much later introduction of mechanized industry, which, in turn, gave birth to what we are today: a species controlling something like 30% of all energy on the surface of our planet.       

Maybe we are experimenting with dispersed, highly mobile and coordinated networks of small energy reservoirs – the automotive fleet – just for the sake of learning how to develop such networks? Some other facts, which, once again, are impolitely disturbing, come to the fore. I had a look at the data published by United Nations, as regards the total installed capacity of generation in electricity (https://unstats.un.org/unsd/energystats/ ). I calculated the average electrical capacity per capita, at the global scale. Turns out in 2014 the average human capita on Earth had around 60% more power capacity to tap from, as compared to a similarly human capita in 1999.

Interesting. It looks even more interesting when taken as the first moment of a process. When I divide the annual incremental change in the installed electrical capacity on the planet, and I divide it by the absolute demographic increment, thus when I go ‘Delta capacity / delta population’, that coefficient of elasticity grows like hell. In 2014, it was almost three times more than in 1999. We, humans, keep developing denser a network of cars, as compared to our population, and, at the same time, we keep increasing the relative power capacity which every human can tap into.    

Someone could say it is because we simply consume more and more energy per capita. Cool, I check with the World Bank: https://data.worldbank.org/indicator/EG.USE.PCAP.KG.OE . Yes, we increase our average annual consumption of energy per one human being, and yet this is a very gentle increment: barely 18% from 1999 through 2014. Nothing to do with the quick accumulation of generative capacity. We accumulate densifying a global fleet of cars, and growing a reserve of power capacity. What are we doing it for?

This is a deep question, and I calculated two additional elasticities with the data at hand. Firstly, I denominated incremental change in the number of new cars per each new human born over the average consumption of energy per capita. In the visual below, this is the coefficient ‘Elasticity of cars per capita to energy per capita’. Between 1999 and 2014, this elasticity had passed from 0,49 to 0,79. We keep accumulating something like an overhead of incremental car fleet, as compared to the amount of energy we consume.

Secondly, I formalized the comparison between individual consumption of energy and average power capacity per capita. This is the ‘Elasticity of capacity per capita to energy per capita’ column in the visual below.  Once again, it is a growing trend.   

At the planetary scale, we keep beefing up our collective reserves of energy, and we seriously mean business about dispersing those reserves into networks of small reservoirs, possibly on wheels.

Increased propensity to store is a historically known collective response to anticipated shortage. Do we, the human race, collectively and not quite consciously anticipate a shortage of energy? How could that happen? Our biology should suggest it just the opposite. With the climate change being around, we technically have more energy in the ambient environment, not less. What exact kind of shortage in energy are we collectively anticipating? This is the type of riddle I like.

[1] Bruzelius, C. (1987). The Construction of Notre-Dame in Paris. The Art Bulletin, 69(4), 540-569. https://doi.org/10.1080/00043079.1987.10788458

[2] Braudel, F. (1992). Civilization and capitalism, 15th-18th century, vol. II: The wheels of commerce (Vol. 2). Univ of California Press.

Unintentional, and yet powerful a reductor

As usually, I work on many things at the same time. I mean, not exactly at the same time, just in a tight alternate sequence. I am doing my own science, and I am doing collective science with other people. Right now, I feel like restating and reframing the main lines of my own science, with the intention to both reframe my own research, and be a better scientific partner to other researchers.

Such as I see it now, my own science is mostly methodological, and consists in studying human social structures as collectively intelligent ones. I assume that collectively we have a different type of intelligence from the individual one, and most of what we experience as social life is constant learning through experimentation with alternative versions of our collective way of being together. I use artificial neural networks as simulators of collective intelligence, and my essential process of simulation consists in creating multiple artificial realities and comparing them.

I deliberately use very simple, if not simplistic neural networks, namely those oriented on optimizing just one attribute of theirs, among the many available. I take a dataset, representative for the social structure I study, I take just one variable in the dataset as the optimized output, and I consider the remaining variables as instrumental input. Such a neural network simulates an artificial reality where the social structure studied pursues just one, narrow orientation. I create as many such narrow-minded, artificial societies as I have variables in my dataset. I assess the Euclidean distance between the original empirical dataset, and each of those artificial societies. 

It is just now that I realize what kind of implicit assumptions I make when doing so. I assume the actual social reality, manifested in the empirical dataset I study, is a concurrence of different, single-variable-oriented collective pursuits, which remain in some sort of dynamic interaction with each other. The path of social change we take, at the end of the day, manifests the relative prevalence of some among those narrow-minded pursuits, with others being pushed to the second rank of importance.

As I am pondering those generalities, I reconsider the actual scientific writings that I should hatch. Publish or perish, as they say in my profession. With that general method of collective intelligence being assumed in human societies, I focus more specifically on two empirical topics: the market of energy and the transition away from fossil fuels make one stream of my research, whilst the civilisational role of cities, especially in the context of the COVID-19 pandemic, is another stream of me trying to sound smart in my writing.

For now, I focus on issues connected to energy, and I return to revising my manuscript ‘Climbing the right hill – an evolutionary approach to the European market of electricity’, as a resubmission to Applied Energy . According to the guidelines of Applied Energy , I am supposed to structure my paper into the following parts: Introduction, Material and Methods, Theory, Calculations, Results, Discussion, and, as sort of a summary pitch, I need to prepare a cover letter where I shortly introduce the reasons why should the editor of Applied Energy bother about my paper at all. On the top of all these formally expressed requirements, there is something I noticed about the general style of articles published in Applied Energy : they all demonstrate and discuss strong, sharp-cutting hypotheses, with a pronounced theoretical edge in them. If I want my paper to be accepted by that journal, I need to give it that special style.  

That special style requires two things which, honestly, I am not really accustomed to doing. First of all, it requires, precisely, to phrase out very sharp claims. What I like the most is to show people material and methods which I work with and sort of provoke a discussion around it. When I have to formulate very sharp claims around that basic empirical stuff, I feel a bit awkward. Still, I understand that many people are willing to discuss only when they are truly pissed by the topic at hand, and sharply cut hypotheses serve to fuel that flame.

Second of all, making sharp claims of my own requires passing in thorough review the claims which other researchers phrase out. It requires doing my homework thoroughly in the review-of-literature. Once again, not really a fan of it, on my part, but well, life is brutal, as my parents used to teach me and as I have learnt in my own life. In other words, real life starts when I get out of my comfort zone.

The first body of literature I want to refer to in my revised article is the so-called MuSIASEM framework AKA Multi-scale Integrated Analysis of Societal and Ecosystem Metabolism’. Human societies are assumed to be giant organisms, and transformation of energy is a metabolic function of theirs (e.g. Andreoni 2020[1], Al-Tamimi & Al-Ghamdi 2020[2] or Velasco-Fernández et al. 2020[3]). The MuSIASEM framework is centred around an evolutionary assumption, which I used to find perfectly sound, and which I have come to consider as highly arguable, namely that the best possible state for both a living organism and a human society is that of the highest possible energy efficiency. As regards social structures, energy efficiency is the coefficient of real output per unit of energy consumption, or, in other words, the amount of real output we can produce with 1 kilogram of oil equivalent in energy. My theoretical departure from that assumption started with my own empirical research, published in my article ‘Energy efficiency as manifestation of collective intelligence in human societies’ (Energy, Volume 191, 15 January 2020, 116500, https://doi.org/10.1016/j.energy.2019.116500 ). As I applied my method of computation with a neural network as simulator of social change, I found out that human societies do not really seem to max out on energy efficiency. Maybe they should but they don’t. It was the first realization, on my part, that we, humans, orient our collective intelligence on optimizing the social structure as such, and whatever comes out of that in terms of energy efficiency, is an unintended by-product rather than a purpose. That general impression has been subsequently reinforced by other empirical findings of mine, precisely those which I introduce in that manuscript ‘Climbing the right hill – an evolutionary approach to the European market of electricity’, which I am currently revising for resubmission with Applied Energy . According to the guidelines of Applied Energy.

In practical terms, it means that when a public policy states that ‘we should maximize our energy efficiency’, it is a declarative goal which human societies actually do not strive for. It is a little as if a public policy imposed the absolute necessity of being nice to each other and punished any deviation from that imperative. People are nice to each other to the extent of current needs in social coordination, period. The absolute imperative of being nice is frequently the correlate of intense rivalry, e.g. as it was the case with traditional aristocracy. The French have even an expression, which I find profoundly true, namely ‘trop gentil pour être honnête’, which means ‘too nice to be honest’. My personal experience makes me kick into an alert state when somebody is that sort of intensely nice to me.

Passing from metaphors to the actual subject matter of energy management, it is a known fact that highly innovative technologies are usually truly inefficient. Optimization of efficiency, would it be energy efficiency or any other aspect thereof, is actually a late stage in the lifecycle of a technology. Deep technological change is usually marked by a temporary slump in efficiency. Imposing energy efficiency as chief goal of technology-related policies means systematically privileging and promoting technologies with the highest energy efficiency, thus, by metaphorical comparison to humans, technologies in their 40ies, past and over the excesses of youth.

The MuSIASEM framework has two other traits which I find arguable, namely the concept of evolutionary purpose, and the imperative of equality between countries in terms of energy efficiency. Researchers who lean towards and into the MuSIASEM methodology claim that it is an evolutionary purpose of every living organism to maximize energy efficiency, and therefore human societies have the same evolutionary purpose. It further implies that species displaying marked evolutionary success, i.e. significant growth in headcount (sometimes in mandibulae-count, should the head be not really what we mean it to be), achieve that success by being particularly energy efficient. I even went into some reading in life sciences and that claim is not grounded in any science. It seems that energy efficiency, and any denomination of efficiency, as a matter of fact, are very crude proportions we apply to complex a balance of flows which we have to learn a lot about. Niebel et al. (2019[4]) phrase it out as follows: ‘The principles governing cellular metabolic operation are poorly understood. Because diverse organisms show similar metabolic flux patterns, we hypothesized that a fundamental thermodynamic constraint might shape cellular metabolism. Here, we develop a constraint-based model for Saccharomyces cerevisiae with a comprehensive description of biochemical thermodynamics including a Gibbs energy balance. Non-linear regression analyses of quantitative metabolome and physiology data reveal the existence of an upper rate limit for cellular Gibbs energy dissipation. By applying this limit in flux balance analyses with growth maximization as the objective function, our model correctly predicts the physiology and intracellular metabolic fluxes for different glucose uptake rates as well as the maximal growth rate. We find that cells arrange their intracellular metabolic fluxes in such a way that, with increasing glucose uptake rates, they can accomplish optimal growth rates but stay below the critical rate limit on Gibbs energy dissipation. Once all possibilities for intracellular flux redistribution are exhausted, cells reach their maximal growth rate. This principle also holds for Escherichia coli and different carbon sources. Our work proposes that metabolic reaction stoichiometry, a limit on the cellular Gibbs energy dissipation rate, and the objective of growth maximization shape metabolism across organisms and conditions’. 

I feel like restating the very concept of evolutionary purpose as such. Evolution is a mechanism of change through selection. Selection in itself is largely a random process, based on the principle that whatever works for now can keep working until something else works even better. There is hardly any purpose in that. My take on the thing is that living species strive to maximize their intake of energy from environment rather than their energy efficiency. I even hatched an article about it (Wasniewski 2017[5]).

Now, I pass to the second postulate of the MuSIASEM methodology, namely to the alleged necessity of closing gaps between countries as for their energy efficiency. Professor Andreoni expresses this view quite vigorously in a recent article (Andreoni 2020[6]). I think this postulate doesn’t hold both inside the MuSIASEM framework, and outside of it. As for the purely external perspective, I think I have just laid out the main reasons for discarding the assumption that our civilisation should prioritize energy efficiency above other orientations and values. From the internal perspective of MuSIASEM, i.e. if we assume that energy efficiency is a true priority, we need to give that energy efficiency a boost, right? Now, the last time I checked, the only way we, humans, can get better at whatever we want to get better at is to create positive outliers, i.e. situations when we like really nail it better than in other situations. With a bit of luck, those positive outliers become a workable pattern of doing things. In management science, it is known as the principle of best practices. The only way of having positive outliers is to have a hierarchy of outcomes according to the given criterion. When everybody is at the same level, nobody is an outlier, and there is no way we can give ourselves a boost forward.

Good. Those six paragraphs above, they pretty much summarize my theoretical stance as regards the MuSIASEM framework in research about energy economics. Please, note that I respect that stream of research and the scientists involved in it. I think that representing energy management in human social structures as a metabolism is a great idea: it is one of those metaphors which can be fruitfully turned into a quantitative model. Still, I have my reserves.

I go further. A little more review of literature. Here comes a paper by Halbrügge et al. (2021[7]), titled ‘How did the German and other European electricity systems react to the COVID-19 pandemic?’. It points at an interesting point as regards energy economics: the pandemic has induced a new type of risk, namely short-term fluctuations in local demand for electricity. That, in turn, leads to deeper troughs and higher peaks in both the quantity and the price of energy in the market. More risk requires more liquidity: this is a known principle in business. As regards energy, liquidity can be achieved both through inventories, i.e. by developing storage capacity for energy, and through financial instruments. Halbrügge et al. come to the conclusion that such circumstances in the German market have led to the reinforcement of RES (Renewable Energy Sources). RES installations are typically more dispersed, more local in their reach, and more flexible than large power plants. It is much easier to modulate the output of a windfarm or a solar farm, as compared to a large fossil-fuel-based installation. 

Keeping an eye on the impact of the pandemic upon the market of energy, I pass to the article titled ‘Revisiting oil-stock nexus during COVID-19 pandemic: Some preliminary results’, by Salisu, Ebuh & Usman (2020[8]). First of all, a few words of general explanation as for what the hell is the oil-stock nexus. This is a phenomenon, which I saw any research about in 2017, which consists in a diversification of financial investment portfolios from pure financial stock into various mixes of stock and oil. Somehow around 2015, people who used to hold their liquid investments just in financial stock (e.g. as I do currently) started to build investment positions in various types of contracts based on the floating inventory of oil: futures, options and whatnot. When I say ‘floating’, it is quite literal: that inventory of oil really actually floats, stored on board of super-tanker ships, sailing gently through international waters, with proper gravitas (i.e. not too fast).

Long story short, crude oil has been increasingly becoming a financial asset, something like a buffer to hedge against risks encountered in other assets. Whilst the paper by Salisu, Ebuh & Usman is quite technical, without much theoretical generalisation, an interesting observation comes out of it, namely that short-term shocks, during the pandemic in financial markets had adversely impacted the price of oil more than the prices of stock. That, in turn, could indicate that crude oil was good as hedging asset just for a certain range of risks, and in the presence of price shocks induced by the pandemic, the role of oil could diminish.     

Those two papers point at a factor which we almost forgot as regards the market of energy, namely the role of short-term shocks. Until recently, i.e. until COVID-19 hit us hard, the textbook business model in the sector of energy had been that of very predictable demand, nearly constant in the long-perspective and varying in a sinusoidal manner in the short-term. The very disputable concept of LCOE AKA Levelized Cost of Energy, where investment outlays are treated as if they were a current cost, is based on those assumptions. The pandemic has shown a different aspect of energy systems, namely the need for buffering capacity. That, in turn, leads to the issue of adaptability, which, gently but surely leads further into the realm of adaptive changes, and that, ladies and gentlemen, is my beloved landscape of evolutionary, collectively intelligent change.

Cool. I move forward, and, by the same occasion, I move back. Back to the concept of energy efficiency. Halvorsen & Larsen study the so-called rebound effect as regards energy efficiency (Halvorsen & Larsen 2021[9]). Their paper is interesting for three reasons, the general topic of energy efficiency being the first one. The second one is methodological focus on phenomena which we cannot observe directly, and therefore we observe them through mediating variables, which is theoretically close to my own method of research. Finally, the phenomenon of rebound effect, namely the fact that, in the presence of temporarily increased energy efficiency, the consumers of energy tend to use more of those locally more energy-efficient goods, is essentially a short-term disturbance being transformed into long-term habits. This is adaptive change.

The model construed by Halvorsen & Larsen is a theoretical delight, just something my internal happy bulldog can bite into. They introduce the general assumption that consumption of energy in households is a build-up of different technologies, which can substitute each other under some conditions, and complementary under different conditions. Households maximize something called ‘energy services’, i.e. everything they can purposefully derive from energy carriers. Halvorsen & Larsen build and test a model where they derive demand for energy services from a whole range of quite practical variables, which all sums up to the following: energy efficiency is indirectly derived from the way that social structures work, and it is highly doubtful whether we can purposefully optimize energy efficiency as such.       

Now, here comes the question: what are the practical implications of all those different theoretical stances, I mean mine and those by other scientists? What does it change, and does it change anything at all, if policy makers follow the theoretical line of the MuSIASEM framework, or, alternatively, my approach? I am guessing differences at the level of both the goals, and the real outcomes of energy-oriented policies, and I am trying to wrap my mind around that guessing. Such as I see it, the MuSIASEM approach advocates for putting energy-efficiency of the whole global economy at the top of any political agenda, as a strategic goal. On the path towards achieving that strategic goal, there seems to be an intermediate one, namely that to narrow down significantly two types of discrepancies:

>> firstly, it is about discrepancies between countries in terms of energy efficiency, with a special focus on helping the poorest developing countries in ramping up their efficiency in using energy

>> secondly, there should be a priority to privilege technologies with the highest possible energy efficiency, whilst kicking out those which perform the least efficiently in that respect.    

If I saw a real policy based on those assumptions, I would have a few critical points to make. Firstly, I firmly believe that large human societies just don’t have the institutions to enforce energy efficiency as chief collective purpose. On the other hand, we have institutions oriented on other goals, which are able to ramp up energy efficiency as instrumental change. One institution, highly informal and yet highly efficient, is there, right in front of our eyes: markets and value chains. Each product and each service contain an input of energy, which manifests as a cost. In the presence of reasonably competitive markets, that cost is under pressure from market prices. Yes, we, humans are greedy, and we like accumulating profits, and therefore we squeeze our costs. Whenever energy comes into play as significant a cost, we figure out ways of diminishing its consumption per unit of real output. Competitive markets, both domestic and international, thus including free trade, act as an unintentional, and yet powerful a reductor of energy consumption, and, under a different angle, they remind us to find cheap sources of energy.

[1] Andreoni, V. (2020). The energy metabolism of countries: Energy efficiency and use in the period that followed the global financial crisis. Energy Policy, 139, 111304. https://doi.org/10.1016/j.enpol.2020.111304

[2] Al-Tamimi and Al-Ghamdi (2020), ‘Multiscale integrated analysis of societal and ecosystem metabolism of Qatar’ Energy Reports, 6, 521-527, https://doi.org/10.1016/j.egyr.2019.09.019

[3] Velasco-Fernández, R., Pérez-Sánchez, L., Chen, L., & Giampietro, M. (2020), A becoming China and the assisted maturity of the EU: Assessing the factors determining their energy metabolic patterns. Energy Strategy Reviews, 32, 100562.  https://doi.org/10.1016/j.esr.2020.100562

[4] Niebel, B., Leupold, S. & Heinemann, M. An upper limit on Gibbs energy dissipation governs cellular metabolism. Nat Metab 1, 125–132 (2019). https://doi.org/10.1038/s42255-018-0006-7

[5] Waśniewski, K. (2017). Technological change as intelligent, energy-maximizing adaptation. Energy-Maximizing Adaptation (August 30, 2017). http://dx.doi.org/10.1453/jest.v4i3.1410

[6] Andreoni, V. (2020). The energy metabolism of countries: Energy efficiency and use in the period that followed the global financial crisis. Energy Policy, 139, 111304. https://doi.org/10.1016/j.enpol.2020.111304

[7] Halbrügge, S., Schott, P., Weibelzahl, M., Buhl, H. U., Fridgen, G., & Schöpf, M. (2021). How did the German and other European electricity systems react to the COVID-19 pandemic?. Applied Energy, 285, 116370. https://doi.org/10.1016/j.apenergy.2020.116370

[8] Salisu, A. A., Ebuh, G. U., & Usman, N. (2020). Revisiting oil-stock nexus during COVID-19 pandemic: Some preliminary results. International Review of Economics & Finance, 69, 280-294. https://doi.org/10.1016/j.iref.2020.06.023

[9] Halvorsen, B., & Larsen, B. M. (2021). Identifying drivers for the direct rebound when energy efficiency is unknown. The importance of substitution and scale effects. Energy, 222, 119879. https://doi.org/10.1016/j.energy.2021.119879

Alois in the middle


I am returning to my syllabuses for the next academic year. I am focusing more specifically on microeconomics. Next year, I am supposed to give lectures in Microeconomics at both the Undergraduate, and the Master’s level. I feel like asking fundamental questions. My fundamental question, as it comes to teaching any curriculum, is the same: what can my students do with it? What is the function and the purpose of microeconomics? Please, notice that I am not asking that frequently stated, rhetorical question ‘What are microeconomics about?’. Well, buddy, microeconomics are about the things you are going to lecture about. Stands to reason. I want to know, and communicate, what is the practical utility, in one’s life, of those things that microeconomics are about.

The basic claim I am focusing on is the following: microeconomics are the accountancy of social structures. They serve exactly the same purpose that any kind of bookkeeping has ever served: to find and exploit patterns in human behaviour, by the means of accurately applied measures. Them ancients, who built those impressive pyramids (who builds a structure without windows and so little free space inside?), very quickly gathered that in order to have one decent pyramid, you need an army of clerks who do the accounting. They used to count stone, people, food, water etc. This is microeconomics, basically.

Thus, you can do with microeconomics if you want to build an ancient pyramid. Now, I am dividing the construction of said ancient pyramid in two stages: Undergraduate, and Master’s. An Undergraduate ancient pyramid requires the understanding of what do you need to keep the accounts of if you don’t want to be thrown to crocodiles. At the Master’s level, you will want to know what are the odds that you find yourself in a social structure, where inaccurate accounting, in connection with a pyramid, will have you thrown to crocodiles.

Good, now some literature, and a little turn by my current scientific work on the EneFin concept (see « Which salesman am I? » and « Sans une once d’utopisme » for sort of a current account of that research). I have just read that sort of transitional form of science, between an article and a book, basically a report, by Bleich and Guimaraes 2016[1]. It regards investment in renewable energies, mostly from the strictly spoken view of investment logic. Return on investment, net present value – that kind of thing. As I was making my notes out of that reading, my mind made a jump, and it landed on the cover of the quite-well-known book by Joseph Schumpeter: ‘Business Cycles’.

Joseph Schumpeter is an intriguing classic, so to say. Born in 1883, he published ‘Business Cycles’ in 1939, being 56 year-old, after the hell of a ride both for him and for the world, and right at the beginning of another ride (for the world). He was studying economics in Austria, in the early 1900, when social sciences in general were sort of different from their today’s version. They were the living account of a world that used to be changing at a breath-taking pace. Young Joseph (well, Alois in the middle) Schumpeter witnessed the rise of Marxism, World War I, the dissolution of his homeland, the Austro-Hungarian Empire, the rise of the German Reich. He moved from academia to banking, and from European banking to American academia.

I deeply believe that whatever kind of story I am telling, whether I am lecturing about economics, discussing a business concept, or chatting about philosophy, at the bottom line I am telling the story of my own existence. I also deeply believe that the same is true for anyone who goes to any lengths in telling a story. We tell stories in order to rationalize that crazy, exciting, unique and deadly something called ‘life’. To me, those ‘Business Cycles’ by Joseph Schumpeter look very much like a rationalized story of quite turbulent a life.

So, here come a few insights I have out of re-reading ‘Business Cycles’ for the n-th time, in the context of research on my EneFin business concept. Any technological change takes place in a chain of value added. Innovation in one tier of the chain needs to overcome the status quo both upstream and downstream of the chain, but once this happens, the whole chain of technologies and goods changes. I wonder how it can apply specifically to EneFin, which is essentially an institutional scheme. In terms of value added, this scheme is situated somewhere between the classical financial markets, and typical social entrepreneurship. It is social to the extent that it creates that quasi-cooperative connexion between the consumers of energy, and its suppliers. Still, as my idea assumes a financial market for those complex contracts « energy + shares in the supplier’s equity », there is a strong capitalist component.

I guess that the resistance this innovation would have to overcome would consist, on one end, in distrust from the part of those hardcore activists of social entrepreneurship, like ‘Anything that has anything to do with money is bad!’, and, on the other hand, there can be resistance from the classical financial market, namely the willingness to forcibly squeeze the EneFin scheme into some kind of established structure, like the stock market.

The second insight that Joseph has just given me is the following: there is a special type of business model and business action, the entrepreneurial one, centred on innovation rather than on capitalizing on the status quo. This is deep, really. What I could notice, so far, in my research, is that in every industry there are business models which just work, and others which just don’t. However innovative you think you are, most of the times either you follow the field-tested patterns or you simply fail. The real, deep technological change starts when this established order gets a wedge stuffed up its ass, and the wedge is, precisely, that entrepreneurial business model. I wonder how entrepreneurial is the business model of EneFin. Is it really as innovative as I think it is?

In the broad theoretical picture, which comes handy as it comes to science, the incidence of that entrepreneurial business model can be measured and assessed as a probability, and that probability, in turn, is a factor of change. My favourite mathematical approach to structural change is that particular mutation that Paul Krugman[2] made out of the classical production function, as initially formulated by Prof Charles W. Cobb and Prof Paul H. Douglas, in their common work from 1928[3]. We have some output generated by two factors, one of which changes slowly, whilst the other changes quickly. In other words, we have one quite conservative factor, and another one that takes on the crazy ride of creative destruction.

That second factor is innovation, or, if you want, the entrepreneurial business model. If it is to be powerful, then, mathematically, incremental change in that innovative factor should bring much greater a result on the side of output than numerically identical an increment in the conservative factor. The classical notation by Cobb and Douglas fits the bill. We have Y = A*F1a*F21-a and a > 0,5. Any change in F1 automatically brings more Y than the identical change in F2. Now, the big claim by Paul Krugman is that if F1 changes functionally, i.e. if its changes really increase the overall Y, resources will flow from F2 to F1, and a self-reinforcing spiral of change forms: F1 induces faster a change than F2, therefore resources are being transferred to F1, and it induces even more incremental change in F1, which, in turn, makes the Y jump even higher etc.

I can apply this logic to my scientific approach of the EneFin concept. I assume that introducing the institutional scheme of EneFin can improve the access to electricity in remote, rural locations, in the developing countries, and, consequently, it can contribute to creating whole new markets and social structures. Those local power systems organized in the lines of EneFin are the factor of innovation, the one with the a > 0,5 exponent in the Y = A*F1a*F21-a function. The empirical application of this logic requires to approximate the value of ‘a’, somehow. In my research on the fundamental link between population and access to energy, I had those exponents nailed down pretty accurately for many countries in the world. I wonder to what extent I can recycle them intellectually for the purposes of my present research.

As I am thinking on this issue, I will keep talking on something else, and the something else in question is the creation of new markets. I go back to the Venerable Man of microeconomics, the Source of All Wisdom, who used to live with his mother when writing the wisdom which he is so reputed for, today. In other words, I am referring to Adam Smith. Still, just to look original, I will quote his ‘Lectures on Justice’ first, rather than going directly to his staple book, namely ‘The Inquiry Into The Nature And Causes of The Wealth of Nations’.

So, in the ‘Lectures on Justice’, Adam Smith presents his basic considerations about contracts (page 130 and on): « That obligation to performance which arises from contract is founded on the reasonable expectation produced by a promise, which considerably differs from a mere declaration of intention. Though I say I have a mind to do such thing for you, yet on account of some occurrences I do not do it, I am not guilty of breach of promise. A promise is a declaration of your desire that the person for whom you promise should depend on you for the performance of it. Of consequence the promise produces an obligation, and the breach of it is an injury. Breach of contract is naturally the slightest of all injuries, because we naturally depend more on what we possess that what is in the hands of others. A man robbed of five pounds thinks himself much more injured than if he had lost five pounds by a contract ».

People make markets, and markets are made of contracts. A contract implies that two or more people want to do some exchange of value, and they want to perform the exchange without coercion. A contract contains a value that one party engages to transfer on the other party, and, possibly, in the case of mutual contracts, another value will be transferred the other way round. There is one thing about contracts and markets, a paradox as for the role of the state. Private contracts don’t like the government to meddle, but they need the government in order to have any actual force and enforceability. This is one of the central thoughts by another classic, Jean-Jacques Rousseau, in his ‘Social Contract’: if we want enforceable contracts, which can make the intervention of the government superfluous, we need a strong government to back up the enforceability of contracts.

If I want my EneFin scheme to be a game-changer in developing countries, it can work only in countries with relatively well-functioning legal systems. I am thinking about using the metric published by the World Bank, the CPIA property rights and rule-based governance rating.

Still another insight that I have found in Joseph Schumpeter’s ‘Business Cycles’ is that when the entrepreneur, introducing a new technology, struggles against the first inertia of the market, that struggle in itself is a sequence of adaptation, and the strategy(ies) applied in the phases of growth and maturity in the new technology, later on, are the outcome of patterns developed during that early struggle. There is some sort of paradox in that struggle. When the early entrepreneur is progressively building his or her presence in the market, they operate under high uncertainty, and, almost inevitably, do a lot of trial and error, i.e. a lot of adjustments to the initially inaccurate prediction of the future. The developed, more mature version of the newly introduced technology is the outcome of that somehow unique sequence of trials, errors, and adjustments.

Scientifically, that insight means a fundamental uncertainty: once the actual implementation of an entrepreneurial business model, such as EneFin, gets inside that tunnel of learning and struggle, it can take on so many different mutations, and the response of the social environment to those mutations can be so idiosyncratic that we get into really serious economic modelling here.

I am consistently delivering good, almost new science to my readers, and love doing it, and I am working on crowdfunding this activity of mine. As we talk business plans, I remind you that you can download, from the library of my blog, the business plan I prepared for my semi-scientific project Befund  (and you can access the French version as well). You can also get a free e-copy of my book ‘Capitalism and Political Power’ You can support my research by donating directly, any amount you consider appropriate, to my PayPal account. You can also consider going to my Patreon page and become my patron. If you decide so, I will be grateful for suggesting me two things that Patreon suggests me to suggest you. Firstly, what kind of reward would you expect in exchange of supporting me? Secondly, what kind of phases would you like to see in the development of my research, and of the corresponding educational tools?

Support this blog


[1] Bleich, K., & Guimaraes, R. D. (2016). Renewable Infrastructure Investment Handbook: A Guide for Institutional Investors. In World Economic Forum, Geneva.

[2] Krugman, P. (1991). Increasing returns and economic geography. Journal of political economy, 99(3), 483-499.

[3] Charles W. Cobb, Paul H. Douglas, 1928, A Theory of Production, The American Economic Review, Volume 18, Issue 1, Supplement, Papers and Proceedings of the Fortieth Annual Meeting of the American Economic Association (March 1928), pp. 139 – 165