Adaptive look-ahead dispatch of power system based on deep reinforcement learning with knowledge embedded

With the large-scale integration of renewable energy sources and the continuous growth in electricity demand, the operating environment of power grids has become increasingly complex and variable. This change has introduced numerous uncertainties, making it difficult for traditional model-based or purely data-driven grid scheduling methods to formulate accurate and reliable scheduling plans. To address this challenge, this paper proposes an adaptive cruising method for grid scheduling that integrates data and knowledge. Initially, we construct a mathematical model for the adaptive cruising problem of grid scheduling and introduce the concept of a look-ahead dispatch horizon. Subsequently, we model the adaptive look-ahead dispatch problem as a Markov decision process and establish a deep reinforcement learning training framework. Furthermore, this paper introduces an improved proximal policy optimization algorithm that incorporates scheduling knowledge for offline training, guiding the dispatch agent in optimal learning. Finally, an in-depth analysis of simulation examples validates the effectiveness and superiority of the proposed method.