基于LSTM-Transformer的钢铁工业用户调节潜力预测与优化OA

As one of the main bodies of urban electricity consumption,industrial users have complex and variable loads that are highly affected by the user regulation potential.Traditional prediction methods have difficulty accurately estimating the regulation capability of iron and steel industrial users.To balance the uncertainty of load fluctuations and the regular characteristics of power consumption behaviors of iron and steel industry users,a prediction method for the regulation potential of iron and steel industry users based on long short-term memory(LSTM)-Transformer is proposed.This method uses the LSTM network to capture the long-term dependencies of sequences such as adjustable industrial load equipment,maintenance schedules,and user regulation potential samples for feature extraction.Meanwhile,the Transformer module is adopted for position encoding,and a dual-layer multi-head self-attention mechanism is used to capture the relationships among different data attributes and concatenate features,thereby obtaining the regulation potential of industrial users under the influence of multiple factors.The actual operation data of an iron and steel plant in Tianjin,China,are selected to compare the potential values calculated by four models.The experimental results show that,compared with other models,the average error of the proposed model is reduced by approximately 40%,achieving higher accuracy and effectively reflecting the regulation potential of the iron and steel industry users.The proposed model provides strong support for optimal scheduling.