Deep reinforcement learning-based bidding strategy for PV-ESS system considering carbon emission reduction

In this paper, we investigate a bidding strategy for a photovoltaic-energy storage (PV-ESS) system in the electricity market, taking into account the benefits of carbon emission reduction. First, we analyze the operational characteristics of the PV-ESS system and its revenue streams in the electricity market. Next, we formulate a bidding model that incorporates the value of carbon emission reductions by considering China Certified Emission Reduction (CCER) trading. The objective is to maximize the overall gain of the PV-ESS system, accounting for both electricity market revenues and CCER benefits. To address the uncertainty inherent in electricity prices and renewable output, we propose a deep reinforcement learning (DRL)-based approach to derive optimal bidding decisions. Finally, we validate the proposed strategy through case studies, demonstrating not only its effectiveness but also its feasibility. The results indicate that energy storage effectively smooths photovoltaic power fluctuations, captures additional revenue from both the electricity and CCER markets, and substantially enhances the profitability of the PV-ESS system.