Research On Energy Efficiency Model Of The Process Of Cigarette Production System

Due to the rise and development of modern Internet of Things and information technology,cigarette enterprises pay more and more attention to the construction and improvement of energy management system,expecting to through new technical means above to achieve fine management of energy utilization of enterprises,and so as to provide guidance and help for the work of energy conservation.Energy consumption prediction is an important component of enterprise energy management system.At present,there are few researches on energy consumption prediction of cigarette production process system,and the error of energy consumption prediction is large.Therefore,it is an important task to find an accurate and reliable energy consumption prediction model.In order to solve the problems above,based on the actual production process and energy characteristics of a domestic cigarette enterprise,and according to the whole process of data mining,this paper sets up different energy consumption prediction models for each part of the cigarette production process system,and strives to find an accurate and reasonable energy consumption prediction model for each part of the energy consumption.The main research work and conclusions in this paper are as follows:1.According to the research objective and the actual historical data accumulated by the enterprise,the data needed for the research have been collected,processed and sorted out.The collected data mainly includes the historical time series data of various energy consumption indicators and the meteorological parameter data.In the data preparation section,this paper completed the processing of missing and outlier values,data normalization and correlation analysis,so as to provide accurate,complete and consistent data sets for the establishment of energy consumption prediction models in each part..2.A univariate energy consumption prediction model was established based on the divided time series data of univariate energy consumption in cigarette production process.In this paper,the energy consumption prediction models based on ARIMA algorithm and LSTM neural network algorithm are respectively established to predict,and the prediction results of the two models are compared and analyzed.It is found that the energy consumption prediction model based on LSTM neural network algorithm can be used to better predict this part of energy consumption.The performance evaluation indexes MAE and RMSE of the model are the lowest.Therefore,this model can be applied to the univariate energy consumption prediction of the cigarette production process.3.A multivariate energy consumption prediction model was established for the divided air conditioning steam consumption in the cigarette production process.The model mainly took the historical air conditioning steam consumption time series data and the meteorological parameter data as the input.In this paper,the energy consumption prediction models based on BP neural network algorithm and CNN-LSTM neural network algorithm are established for prediction,and the prediction results are compared and analyzed.It is found that the energy consumption prediction model based on CNN-LSTM neural network algorithm has better prediction effect,and the calculated performance evaluation indexes MAE and RMSE of the model are the lowest.Therefore,this model can be applied to the multivariate energy consumption prediction of the cigarette production process.This research provides a more accurate and reasonable model choice for predicting the energy consumption of each part of the cigarette production process system,which is helpful to the construction of the energy management system of cigarette enterprises and the implementation of energy conservation.