Discrete Events based Simulated Trading and Research System

Abstract

High penetration of variable renewable generation burdens Transmission System Operators (TSO) with operational and market challenges, such as electricity price spikes, negative electricity prices, and even branch outages. It is imperative for market players to acquire knowledge about the impacts and influences of their behaviors and unexpected events on electricity markets through simulated trading and predictions of price fluctuation trends and possible outages. TSOs have published their market data to benefit market players for their studies on market trends. However, a further step for TSOs to take is more meaningful to promote market efficiencies, which is designing and operating a Simulated Trading and Research System (STRS). We propose a Discrete Event Dynamic System (DEDS) based modeling and simulation approach to design STRS. STRS can process such events as fluctuations of high penetration variable renewable, market players’ expectations, weather and temperatures, sporting events, etc. The predictions that STRS simulates include impacts on electricity prices, electricity volumes, market powers, social welfare distributions, etc. caused by the above events. STRS implements reinforce learning architecture as a critical component. By looking ahead predictions of electricity market states, STRS helps the electricity market adapt to a forward observable closed-loop system by feeding a Model Predictive Control (MPC) actor component. As a sensitivity analysis framework, STRS connects history market data with electricity market efficiency. A case study shows the pathways of promotion of electricity market efficiency by using STRS to achieve the prediction and perturbation analysis of event-driven price and volume levels, Pareto efficiencies and social welfares.