Research On Real-time Monitoring Of Enterprise Pollution Control Based On Power Data

At present,the relevant regulatory departments have problems such as high regulatory cost and difficulty in the supervision of enterprises,especially for the large number of small and medium-sized enterprises,the supervision is difficult to cover in place.As the essential power to support the smooth production of enterprises,the power data has the characteristics of large data volume,multiple types and high value.Moreover,the power data is still in the state of mild mining at present,which contains extremely high application value.If enterprise power data can be applied in the process of enterprise pollution control supervision to realize remote supervision of enterprise pollution control,it can not only improve the intensity and efficiency of pollution control supervision,but also reduce the supervision cost of the supervision department,which has strong practical significance.Based on the idea of Non-Intrusive Load Identification,this paper applies power data to the field of pollution control supervision of enterprises.As the ultimate purpose is to judge whether the enterprise is carrying out normal production activities,the criterion of judgment is whether the enterprise starts the pollution control equipment when the operation state.Therefore,this paper firstly finds out the possible open and shut down time in enterprise history according to the change point detection theory,then locates the accurate open and shut down time in enterprise history according to the method of minimizing cost function,and finally determines the day’s open and shut down power threshold according to the weighted average of the enterprise history’s open and shut down power threshold,so as to judge the enterprise’s open and shut down status in real time.Secondly,based on the characteristics of the power data used in this paper,90 features were screened according to the feature importance score of the RF model,and the LSTM model based on feature fusion was used as the state recognition model of the enterprise pollution control equipment.The model combined the timing information and the original information of the data,and achieved a good recognition effect with an accuracy of 88%,and the recall rate reached 85%.Finally,based on the results of the first two parts,this paper summarizes the ideas and strategies for real-time monitoring of pollution control in enterprises.As long as the operating state of the pollution control equipment is identified as on,the enterprises are considered to be carrying out normal production activities,and the pollution control monitoring example is given,which has certain practical value for the pollution control supervision of enterprises.