Mitigating Production Uncertainties of Renewables by Adjustable Load Scheduling

Abstract

In the upcoming years, electricity markets in Europe face numerous challenges both on production and demand side. Due to decreasing investment costs for renewable energy sources (RES) and continuing government support for a transition towards green energy technologies, the market share of uncertain and volatile generation capacity will grow. Other than thermal power plants, wind or solar based generation capacity cannot be controlled at will to satisfy the demand for electricity. Furthermore, actual production output for these RES is difficult to predict, because of the limited forecast accuracy of current weather forecast models. On the demand side, household consumption patterns are expected to change due to an increasing share of electric vehicles (EVs). Charging EVs will further increase electricity demand. The additional load has to be provided by renewable energy sources for EVs to be more environmental friendly than fuel powered vehicles. Instead of looking at EVs as an additional burden for the electricity system, this thesis proposes the use of flexible controlled EV charging as a method to mitigate production uncertainty. Introducing the concept of a hybrid aggregator with own photovoltaic (PV) generation capacity and the responsibility to charge EVs, the effects of combining flexible load and RES production are explored. The objective of this research is to determine the value of such flexibility to tackle the challenges imposed by forecast uncertainty of RES production capacity when being traded on the electricity market. The methodology used in this thesis is that of a stochastic linear multistage optimization approach. Based on the analysis of electricity market designs and the stakeholder environment, a conceptual model of the hybrid aggregator is developed and translated into a linear optimization problem. Solving for a maximization of the hybrid aggregator’s revenue, day-ahead market bidding, imbalance requirements and charging schedules can be determined. The optimization problem is implemented in the form of a MATLAB model. Tested for its validity, this model is used to explore the expected market behaviour of the hybrid aggregator given a case study and possible changes of the sociotechnical environment. The possible market effects of the hybrid aggregator are discussed given that he becomes a price maker. Following insights have been gained about the hybrid aggregator: 1.) Given that the hybrid aggregator acts as a party on the liberalized electricity market, his decision making problem can be described as revenue optimizing. 2.) For the hybrid aggregator, scheduling of EV charging is in first place determined by day-ahead and imbalance prices. Instances of EV charging and PV production don’t necessarily coincide. 3.) The market behaviour of the hybrid aggregator is influenced by the market design and can be steered, for example, by imposing limitations on trading capacity on the day-ahead market. Further research is required to explore the market impact of a hybrid aggregator. A recommendation is to determine the aggregator’s behaviour for changes in the sociotechnical environment and its interaction with other electricity market parties. Doing so, input data generated by valid scenario generation techniques should be used.