Optimisation environnementale de la gestion des consommations électriques en temps réel

Smart technologies such as smart metering provide new alternatives to mitigate the impacts of electricity generation. For instance, demand-side management (DSM) programs enhance the integration of intermittent renewable energy sources through changes in consumption patterns. However, changes are led by economic incentives and environmental benefits are not maximized. Life cycle assessment (LCA) is a standardized methodology that aims to assess the environmental impacts of a product or a service over its entire life cycle. The primary objective of our project is to develop environmental indicators to assess the environmental benefits of DSM programs and to provide new alternatives of changes in consumption patterns. Two distinct approaches were used in order to develop the indicators. The first one, called attributional, aimed to assess the potential impacts of real-time electricity use in France including generation, imports and exports with a life cycle approach. The second approach, called consequential, aimed to assess the environmental consequences of electricity use through changes in the technologies producing the energy. Then, based on both approaches, we evaluated the environmental performances of simple DSM strategies using environmental and economic indicators based on historic and day-ahead data. The real-time LCA of electricity generation in France, Germany, Great Britain, Italy, Belgium and Spain were conducted every 30 minutes between 2012 and 2014 in order to compute in France the real-time LCA of electricity consumption considering imports and exports. We found that the use of French generation impact factors compared to the use of French consumption impact factors underestimated on average by 21.8 %, 3.6 %, 4.3 % and 1.7 % the impacts of electricity consumption on the four environmental indicators: Climate Change, Human Health, Ecosystem Quality and Resources respectively. A use of 1 kWh was optimized every day between 2012 and 2014 based on the four environmental indicators previously developed and the price of the electricity. Each optimum scenario minimized every day the impacts or the price of the electricity use. We compared the five optimum scenarios with the help of the five indicators to five worst case scenarios. Each worst case scenario depicted the maximum of impacts or price achievable every day. The environmental indicator Climate Change had the widest range of mitigation compared to the worst case scenario, i.e. up to 38.5 %. If each environmental indicator was considered equally important, the optimizations based on the Climate Change and Human Health indicators offered the most effective solutions on all the environmental indicators. In addition, the optimization based on the price of the electricity was compared to the environmental optimization based on the Climate Change indicator. The environmental optimization cost 1.4 € per kg CO2 eq. saved for a mitigation of 12 % of greenhouse gases (GHG) emissions compared to the economic optimization. If a carbon price of 100€ per tonne of CO2 was applied, the new economic optimization would cost 0.04 € per kg CO2 eq. saved for a mitigation of 2.5 % of the GHG emissions compared to the previous economic optimization. Finally, a day-ahead optimization model was developed based on forecasts of European generation factors. Despite uncertainties related to electricity generation and consumption, we achieved a mitigation of 27 % of GHG emissions compared to the worst case scenario with the use of a realistic optimization model a day-ahead. For the consequential approach, two different models were used in order to define the dynamic technologies affected by the consumption from global data. The first one considered that all the controllable technologies, i.e. all the technologies except those from intermittent sources, were affected by the consumption and the second considered that only the controllable technologies which fluctuated in the same direction than the global consumption were affected by the consumption. Between 2012 and 2014, in France, Italy and Spain, hydroelectric plants were the main affected technologies for the two models. The DSM programs based on the indicators using the consequential approach reduced up to 98 % the GHG emissions of the technologies affected by the consumption. However, the two models had different optimal solutions. Hence, the assumptions made on the global data affected the results of the model. The consequential approach suffered from a lack of local considerations. To become environmentally efficient, DMS programs need to work with system operators of electrical network in order to assess the real local consequences of changes in use patterns.