Research On The Energy-saving Operation Optimization Of Paper Machine Vacuum System Based On Production Data Mining

China is a major paper and paper board production and consumption country.Our country has made significant progress in reducing energy consumption of papermaking in recent years.However,as energy-intensive industries,Papermaking still has a large energy saving potential.Paper machine vacuum system is the large electricity and water consumption on the paper machine with the features of high energy consumption and high energy-saving potential.The setting of vacuum system not only influences the energy consumption of vacuum system,but also influences the energy consumption of other parts in paper machine.For the research of vacuum system,it is of great significance in no matter the successful process of paper machine production,or the energy conservation of entire paper machine.This thesis takes the optimizing of paper machine vacuum system as the object,puts forward a way to analysis the production data,a modeling method of the vacuum system and a way to set up the online optimization platform.This thesis aimed at solving the lack basis problem of vacuum set point in production,and improving energy utilization efficiency of paper machine.This paper,taking a coated white cardboard factory in Guangdong as an example,acquire data from the Energy Management System built by PIO-TECH of SCUT.EMS can integrate and operate the production data every seconds,improves the plant's energy efficiency.In 2014,for example,EMS in the plant equivalent energy have saved 1048 t standard coal,that is equivalent to save energy costs 1.3164 million Yuan.When mining the data of vacuum system,3? and moving average filter methods were used for data pre-processing,and partial least squares were adopted to make analysis of relevant data of the vacuum system,and K-means clustering procedure was adopted to classify the data.The data mining results of coating white cardboard factory showed that in addition to steam consumption of dryer section,energy consumption of the vacuum system also had close relation to transmission energy consumption,and the correlation coefficient of them was 0.8017.In addition,the vehicle speed and the quantitative was an important factor that influenced the energy consumption of the vacuum system,which were ranked at the first and second place of priority,with the VIP numerical values reaching 1.2195 and 1.2057.Both of them should be mainly paid attention to during modeling.Based on the difference of the vehicle speed and the quantitative,K-means divided the data into high weight,middle weight and low weight categories,with the overall average distance of 38.97.Based on the result of data mining,the thesis puts forward a method which establishes the vacuum system and optimized model.Paper-making vacuum system faces the complex and dynamic situation which is difficult to attain optimized efforts from the respect of mechanism.This method setting out from the respect of data,based on partial least squares,artificial neural network and many other methods,to establish energy cost model of the vacuum system,and then according to genetic algorithm,the array of most energy-frugal vacuum degree under different operating conditions were found among these relations.The data tests were made at different time periods,and the coefficient of determination of energy cost and dewater ability by partial least squares was between 0.6 and 0.7,with the coefficient of determination of artificial neural network reaching above 0.80.The model was finally determined to use neural network to establish relation.Comprehensive considerations were given to the model of partial least squares and artificial neural network when seeking the solution with genetic algorithm,through comparison with the average value of production at site,it was obtained by the model that the optimization of setting value saved more than 2000 Yuan of energy cost per hour.Finally,on the basis of the efforts of vacuum system and optimized model,developed the vacuum system optimized platform in EMS.The design of framework of the platform was made,and the platform screen code was optimized combined with specific conditions and practical application demands at the site.The platform not only could achieve the real time monitoring for vacuum system,but also could provide the most optimized and real time vacuum degree set point to guide production.