基于LSTM的发电机组污染物排放预测研究

为了利用电力公司积累的海量历史污染物排放数据,形成可以减少污染物排放的调度框架。采用递归神经网络,结合发电机组输出功率与污染物排放量之间的关系,并使用批规范化等深度学习技术,对数据和模型进行学习和训练。实验结果表明,可以有效预测发电机组污染物排放量,解决传统回归分析方法无法适用的难以提取有效特征的问题。

A Study on the Prediction of Generator Set Pollutant Emissions Based on LSTM

To take advantage of the massive historical pollutant emission data accumulated by power companies to build up a scheduling framework for reduction of pollutant emission,this paper adopts a recurrent neural network,as well as deep learning technologies like batch normalization,combined with the relationship between the generator set output power and pollutant emissions,to implement learning and training of data and models.Experimental results show that this method can effectively predict pollutant emissions from generator sets to solve the problem of difficult extraction of effective characteristics through traditional regression analysis.