Optimal path to a high share of distributed PV in low-voltage distribution grids

To limit global warming, the energy transition from fossil fuels to renewables is one of the main challenges of this century. This transition will help decrease our CO2 emissions, create jobs, contribute to a broader revenues distribution, improve energy system efficiency, bring electricity to developing countries and contribute to improving water quality, health care and education. It is made possible thanks to the steady decline of renewable energy costs in particular for photovoltaics (PV) whose electricity cost has decreased by a factor 5 during the last decade. This technology is now cost-effective at different scales of deployment, from GW size power plants to a few kW rooftop systems for prosumers. In Switzerland, due to the free land scarcity, distributed PV systems is one the pillar of the energy transition. However, low-voltage distribution networks were built with centralized generation in mind and were therefore not designed to accommodate a high level of distributed generation. The grid operation requires to maintain, at all time, a balance between consumption and production. An excess of PV energy could require significant curtailment of the generation to cope with the network operating constraints. This technical challenge affects PV system profitability and adoption rate, which is critical to achieving objectives in terms of PV capacity. Moreover, current electricity tariff does not guarantee sustainable revenue for grid operation in the event of an increase in distributed generation and raises issues of equity between consumers and prosumers. Demand-side management can be used to limit PV injection peaks and to reduce the mismatch between production and consumption. By disaggregating households electricity load profiles into appliance categories, we can quantify the share of flexible loads. The electricity consumption flexibility from the behavioural change could have the potential to increase by 20% the PV hosting of a distribution grid. This measure by itself is not sufficient to achieve high PV penetration but should be considered together with technical measures such as PV curtailment or distributed storage. The current typical flat tariff does not take advantage of the flexibility that could be provided by these two measures. To harvest this flexibility, novel electricity pricing mechanisms are presented in this thesis, and their economic and operational performances are assessed using a combination of design and operation optimization, and power flow simulation. We show that unlike time-of-use tariffs which tend to increase the stress on the grid, capacity-based and block rate tariffs reduce the grid usage while ensuring low levelized cost of electricity for prosumers. To fully comply with the grid constraints, an operation regulation could be introduced, but this would lead to a less efficient system with high additional costs. However, a moderate feed-in limit is an efficient solution to reduce peak load on the transformer with only a little impact on system profitability. We finally present a possible path, from 2020 to 2050, toward high PV hosting in a low voltage grid, ensuring at the same time a safe grid operation without grid reinforcement thanks to a progressive feed-in limit, high profitability for the prosumers and constant revenues for grid operation and maintenance. This thesis ends on recommendations for distribution system operators to better anticipate and become part of the energy transition.