Algorithm Design of Day Ahead Market Marginal Price Forecasting Considering New Energy Absorptive Capacity

On the premise of considering the absorptive capacity of new energy, the day ahead market marginal price forecasting algorithm is designed to ensure the price forecasting effect and facilitate the accurate decision-making of both parties in the power market. To analyze the correlation between the day-ahead market marginal price and the new energy consumption and load, the singular spectrum analysis (SSA) algorithm is used to decompose the original data sequence of the three, obtain several SSA subsequences, construct the long short memory network (LSTM) model for each subsequence, and optimize the key parameters of such a model through the crossover (COS) algorithm. The COS-LSTM prediction model for each subsequence is obtained. After the prediction results of each subsequence are superimposed, the final day-ahead market marginal unit price prediction value is obtained. In the experiment, the algorithm can effectively decompose various original data sequences, and predict the day-ahead market marginal price under different new energy consumptions based on the decomposition results. The prediction results show that the higher the new energy consumption, the lower the day-ahead market marginal price, and the prediction results are accurate and reliable.