Interpretable Hybrid Experiment Learning-Based Simulation Analysis of Power System Planning under the Spot Market Environment

Abstract

The electricity spot market plays a significant role in promoting the self-improvement of the overall resource utilization efficiency of the power system and advancing energy conservation and emission reduction. This paper analyzes and compares the potential impacts of spot market operations on system planning, considering the differences between planning methods in traditional and spot market environments through theoretical analysis and model comparison. Furthermore, we conduct research and analysis on grid planning methods under the spot market environment with the goal of maximizing social benefits. Unlike the pricing approach based on historical price data in traditional market simulation processes, a data-driven approach that combines experimental economics and machine learning is proposed, specifically using mixed empirical learning to simulate unit bidding strategies in market transactions. A simulation model for electricity spot market trading is constructed to analyze the performance of the planning results in the spot market environment. The case study results indicate that the proposed planning methods can enable the grid to operate well in the spot market environment, maintain relatively stable nodal prices, and ensure the integration of a high proportion of clean energy.