摘要
电力企业排污数据量较为庞大,且来源广泛,如果不能从这些数据中提取出有价值的信息,则会导致切页污染物排放测算结果出现较大误差,如何通过电力大数据,运用最合适的方法动态测算电炉企业污染排放量是一个难点问题。因此,开展基于电力大数据的企业污染物排放智能测算方法精度对比研究。以测算重庆某铸造厂污染物排放量为例,运用AI智能识别技术,并结合传统的RBF神经网络模型、BP神经网络模型及多元逐步回归模型,对2021年12月—2022年6月各时段的排污量进行训练测算。基于MAE、MAPE和RMSE对比分析,测算精度结果显示:RBF神经网络模型>BP神经网络模型>多元逐步回归模型。研究结果表明,基于电力大数据配合合适的算法可以实时测算电炉企业污染物排放量,为精准治污、科学治污提供科技支撑。
The amount of pollution discharge data from power enterprises is relatively large and comes from a wide range of sources.If valuable information cannot be extracted from these data,it will lead to significant errors in the calculation results of pollutant discharge from page cutting.How to dynamically calculate the pollution discharge of electric furnace enterprises using the most suitable method through power big data is a difficult problem.Therefore,a comparative study on the accuracy of intelligent calculation methods for enterprise pollutant emissions based on power big data will be conducted.Taking the calculation of pollutant emissions from a foundry in Chongqing as an example,AI intelligent recognition technology was used,combined with traditional RBF neural network models,BP neural network models,and multiple stepwise regression models,to train and calculate the pollutant emissions for each period from December 2021 to June 2022.Based on the comparative analysis of MAE,MAPE,and RMSE,the calculation accuracy results show that RBF neural network model>BP neural network model>multiple stepwise regression model.The research results indicate that using electricity big data combined with appropriate algorithms can calculate the pollutant emissions of electric furnace enterprises in real-time,providing technological support for precise and scientific pollution control.
作者
张峻豪
杨小林
韩雪梅
欧阳蓝
刘小春
ZHANG Junhao;YANG Xiaolin;HAN Xuemei;OUYANG Lan;LIU Xiaochun(School of Electrical Engineering,Chongqing University,Chongqing 400044,China;Chongqing Guangruida Technology Co.Ltd.,Chongqing 401121,China)
出处
《工业加热》
CAS
2024年第1期52-58,共7页
Industrial Heating
关键词
电力大数据
模型测算
RBF神经网络
BP神经网络
电炉企业动态排污
power big data
model calculation
RBF neural network
BP neural network
dynamic blowdown of electric furnace enterprises
