Design And Implementation Of Intelligent Algorithm For Temperature Prediction In Industrial Control

With the increasing development of machine learning,more and more researchers start to focus on temperature predicting.In addition,with the upgrading of hardware,data has been gradually seen as an important resource for production.Therefore,most factories start to pay more attention on data,more and more sensors are being adopted during production.A large number of data produced from industrial production process gave a concrete base for temperature prediction in industrial production.Various algorithms adopted in industrial production for temperature prediction largely improved the accuracy for temperature predicting in industrial automatic control,thereby,products' quality received a great improvement.In recent years,a rising number of researchers have taken their eyes on temperature control.By adopting machine-learning methods or mathematical modeling,algorithm for temperature prediction will analyze raw data so that the trend of temperature change will be predicted.In most factories,machine has gradually took place of manual workforce,automatic production begins to dominate the whole producing process.In automatic production process,in order to monitor the whole production process and correctly diagnose any malfunctions happened in the process,any data collected from the process need to be carefully analyzed and appropriate action need to be taken in accordance of the data.In other hands,when production process involved with chemical reaction,the situation becomes complicated.Usually,chemical reaction has low tolerance for temperature,in other words,when comes to produce any chemical related product,temperature need to be strictly controlled.In order to satisfy this requirement,knowing the temperature change in advance,in short,temperature prediction is highly recommended in industrial production.Data collected from multiple sensors in industrial field contributes as a base for temperature production.With the background of temperature prediction with intelligent algorithms in industrial automatic control,this paper proposed an intelligent algorithm for temperature prediction in polycarboxylate superplasticizer production with the use of neural network.Additionally,a temperature predicting system for industrial automatic control is designed and implemented.To avoid manual mistakes caused by inexperienced or careless,this algorithm was designed to give more support for automatic control over industrial production.Hence,the efficiency will be accelerate and the quality of products will be improved.There are three parts in this algorithm: data mining,temperature predicting,and error correction.The environment in industrial production is complicated,there are numbers of features that will have impact on temperature change,for example,the heat emission or absorption in chemical reaction,indoor temperature,materials' temperature etc...In first step,this algorithm picked out those features who contributed most for temperature change through data mining;Secondly,A BP neural network was constructed,and this neural network was trained with those related features picked out in first step;Last but not least,to reach a better accuracy,error correction was applied to this BP neural network.In accordance of demand analysis,combine with the complex situation of industrial field and the algorithm proposed in this paper,a temperature predicting system is proposed.This system was verified with large amounts of data analyzing and comparison with various experiences.This system was adopted for production of polycarboxylate superplasticizer in a company in Shantou.The result showed that after this temperature prediction system was adopted,temperature had a better control,the quality of products have been largely improved.In other words,the temperature prediction system proposed in this paper is able to not only improved products' quality,but also decrease the production cost.The algorithm proposed in this paper was proved to be qualified to satisfy the requirements in industrial production.This algorithm is able to enhance the stability in industrial production,and the efficiency of the production will be increased.