基于移动平均及LSTM组合模型的火电厂日发电量预测研究

Research on the Forecast of Daily Power Generation Capacity of Thermal Power Plants Based on the Combined Model of Moving Average and LSTM

为了解决火电企业在制定发电计划过程中缺乏科学指导的问题,提出一种基于趋势移动平均法及长短期记忆网络(LSTM)组合模型的发电量预测方法。首先分别建立移动平均预测模型和LSTM预测模型,然后使用最小二乘法取不同权重将两者预测模型相结合得到一个新的预测模型。以上海某电厂2018年〜2020年每日的发电量数据作为实例验证。结果表明,所提出的预测方法相对于传统发电量预测方法、简单移动平均和标准LSTM预测方法具有更高的精确度。

In order to solve the problem that thermal power companies lack scientific guidance in the process of making power generation plans,a power generation forecasting method based on trend moving average method and long short-term memory network(LSTM) combined model is proposed.Firstly,the moving average prediction model and the LSTM prediction model are established respectively,and then the least square method is used to take different weights to combine the two prediction models to obtain a new prediction model.Take the daily power generation data of a power plant in Shanghai from 2018 to 2020 as an example verification.The results show that the proposed prediction method has higher accuracy than traditional power generation prediction methods、simple moving average and standard LSTM prediction methods.