Price signal forecasting for day-ahead offering strategy of price-maker renewable energy producers considering different risk preferences

Market prices play the core role in establishing the optimal offering strategy in the day-ahead electricity market. However, how to precisely predict market prices and incorporate price forecasts in offering strategies considering risk preference remains unresolved. To address this challenge, this paper proposes an offering strategy for renewable energy producers in the single-price imbalance settlement under the price-maker assumption. The proposed offering strategy highlights that, apart from stochastic generation forecasting, the only required forecast is the difference between balancing prices and day-ahead prices, referred to as price signals. Then, three forecasting models and two corresponding training strategies are proposed for predicting price signals, including 1) Bounded Model which introduces upper and lower bounds in price signal forecasts, 2) Symbol Model which predicts the sign and value of price signals sequentially, and 3) Decomposition Model which first decomposes the price signal into three components and then predicts each component independently. Numerical results on publicly available real-world datasets verify that the proposed models outperform baseline models in price signal forecasting accuracy. Moreover, the proposed offering strategy based on price signal forecasts achieves substantial profit and reduces risk exposure in renewable energy trading.