大工业用户短期电力负荷与电价预测方法

Short-term Electricity Load and Price Prediction Method for of Large Industrial User

为了进一步约束大工业企业用户用电行为,在降低总成本的同时减少碳排放,提出一种采用随机森林筛选特征,基于量子粒子群优化算法改进双向门控循环单元的大工业用户短期电力负荷与电价预测方法。这一预测方法考虑温度、湿度、日期类型等外部特征因素,最大化还原实际运行场景,通过输入历史负荷数据与历史电价数据对未来24 h负荷及电价情况进行预测。试验结果表明,这一预测方法与其它主流预测方法相比具有优越性。

In order to further constrain the electricity consumption behavior of large industrial enterprise user and reduce carbon emission while decreasing total cost,a short-term electricity load and price prediction method for large industrial user was proposed,which uses random forest screening feature and improves the bidirectional gated recurrent unit based on quantum particle swarm optimization algorithm.This prediction method considers external characteristic factors such as temperature,humidity,date type and so on,to maximize the restoration of actual operating scenario.By inputting historical load data and historical electricity price data,the future 24-hour load and electricity price situation can be predicted.The experimental result shows that this prediction method has superiority compared to other mainstream prediction methods.