基于深度学习的火电厂机组负荷调度自动控制方法

Automatic Control Method of Thermal Power Plant Unit Load Dispatching Based on Deep Learning

传统火电厂机组负荷调度自动控制方法无法准确预测多用户目标负荷,导致负荷调度误差较大,提出基于深度学习的火电厂机组负荷调度自动控制方法。通过火电厂的历史运行数据构建火电厂机组负荷分配模型,采用深度学习中的长短期记忆网络(Long Short-Term Memory,LSTM)神经网络进行火电机组负荷预测,建立机组系统能耗与负荷频率的相关模型,输出负荷调度结果,引入比例、积分(Proportion Integral,PI)控制器实现火电厂机组负荷调度自动控制。实验结果表明,文章设计方法能够准确预测火电厂机组运行负荷,在升负荷过程及降负荷过程中的调度负荷和调度自动控制准确性均较好。

The automatic control method of unit load dispatching in Tonghua Thermal Power Plant can not accurately predict the multi-user target load, resulting in large error in load dispatching. A deep learning based automatic control method of unit load dispatching in thermal power plant is proposed. Based on the historical operation data of the thermal power plant,the load distribution model of the thermal power plant unit is constructed. The Long and Short Term Memory(LSTM) neural network in deep learning is used to predict the load of the thermal power unit, establish the correlation model between the energy consumption of the unit system and the load frequency, output the load dispatching results, and introduce the Proportion Integral(PI) controller to realize the automatic control of the thermal power plant unit load dispatching. The experimental results show that the design method in this paper can accurately predict the operating load of thermal power plant units, and the automatic control accuracy of dispatching load dispatching in the process of load rise and load drop is good.