Research on Electric Load Forecasting and User Benefit Maximization Under Demand-Side Response

In this paper, the real-time changes of demand-side response factors are accurately considered. First, CNN is combined with BiLSTM network to extract the spatio-temporal features of load data; then an attention mechanism is introduced to automatically assign the corresponding weights to the hidden layer states of BiLSTM. In the optimization part of the network parameters, the PSO algorithm is combined to obtain better model parameters. Then, considering the average reduction rate of various users under energy efficiency resources and the average load rate under load resources on the original forecast load and the original forecast load, the original load is superimposed with the response load considering demand-side resources to achieve accurate load forecast. Finally, "price-based" time-of-use tariff and "incentive-based" emergency demand response are selected to build a load response model based on the principle of maximizing customer benefits. The results show that demand-side response can reduce the frequency and magnitude of price fluctuations in the wholesale market.