AMIRIS – An Agent-Based Simulation Model for the Analysis of Market Integration of Renewable Energies under Various Energy Policy Frameworks

Introduction We present an agent-based model for the analysis of market integration of renewable electricity. It is under development since 2008 and designed to answer three main research questions, namely: • How can the market integration of wind, solar and biomass energy be assessed? • How can suitable policy instruments for the integration be evaluated? • What kind of market design is appropriate for systems with a high share of renewables? In contrast to other modelling approaches in this particular field of research, this agent-based model is able to map the impact of changes in policy and market design on the various and heterogeneous actors including their interdependencies and interactions. Methods Validation of key parameters of the model is derived from document analysis, semi-structured interviews and expert workshops. Additional data input stems from time series analyses in the renewables sector. Key actors like power plant operators or direct marketers for electricity are parametrised differently, interacting on an hourly basis per simulation year. Policy instruments like the market premium are modelled. The programming language used is Java, using Repast. Results Model runs show that regulatory changes, i.e. changes in the market premium, in particular emphasize differences in the various actors. For example, forecast quality, size and technological composition of their respective portfolio (i.e. the attributes of agents, here: direct marketers) have a strong influence on their economic performance. A reduced market premium leads to small margins for all intermediaries. Together with portfolio size based compensating effects this may be one main reason for a market concentration of actors, which is an ongoing process in today’s market. Conclusion and Outlook We plan to represent other critical components of the electricity market in future versions of the model, like the mapping of an intraday market as well as load and demand response. It is also planned to base the simulation on GIS-data and integrate further storage technologies.