Research On Working Condition Monitoring And Tracing Method In Cement Pre-decomposition Process

The cement precalcination process is one of the key links in cement production,and its coal consumption accounts for about 60% of the entire production.“The comprehensive work plan for energy conservation and emission reduction during the 14 th Five Year Plan” requires high energy consuming industries such as cement to reduce specific energy consumption by13.5% during the 14 th Five Year Plan period.Intelligent manufacturing is an important direction for energy consumption reduction in the current cement industry.However,the complex and variable working conditions of the cement precalcination process restrict the implementation of its intelligent control.The research on working condition monitoring and traceability methods is the key to solving the complex and variable working conditions in the precalcination process and the foundation for achieving intelligent control,which is of great significance for energy consumption reduction in the cement industry.Currently,the methods for the working condition monitoring of cement precalcination process mainly include expert system and data-driven.Expert system is difficult to establish a complete expert knowledge base,and data-driven recognition efficiency for simple working conditions is low.The working condition traceability of cement precalcination process is still in the exploratory stage.To solve them,this article conducted research on monitoring and tracing methods for the working conditions of cement precalcination process.The main content is as follows:(1)Research on working condition monitoring methods for cement precalcination process.In response to the complex and variable working conditions of the cement precalcination process,this paper proposes a working condition monitoring method of the cement precalcination process based on expert rules,One Class Support Vector Machine(OCSVM),and Denoising Autoencoder(DAE)integration.Using expert rules to monitor the dual variable(variable = 2)working condition of the cement precalcination process,using OCSVM to monitor the raw material composition working conditions,using DAE to monitor the multi variable(variable ≥ 3)working condition of the cement precalcination process.Finally,the above models are integrated to accurately monitor the working conditions of the cement precalcination process.As an example,the above method was validated using the cleaning and scaling condition at the kiln end.The results showed that 877 sets were correctly identified and had high monitoring accuracy in 882 sets of test samples.(2)Research on the working condition traceability method of cement precalcination process based on Long Short Term Memory(LSTM)and case-based reasoning.Working condition traceability can provide important information for the stable operation of the cement precalcination process.This article proposes a working condition traceability method based on LSTM and case-based reasoning.On the basis of working condition monitoring,this article uses LSTM to extract abnormal time characteristics of variables,combines with Mann Kendall to extract abnormal trend characteristics of variables,and uses case-based reasoning to trace the causes of abnormal working conditions,which provides important information for intelligent control of cement precalcination process.The above method was validated using the kiln end cleaning and scaling condition as an example,and the results showed that LSTM and case-based reasoning can accurately trace the kiln end cleaning and scaling condition.(3)Development of software for monitoring and tracing the working condition of cement precalcination process.Designed and developed a software for monitoring and tracing the working condition of cement precalcination process based on Matlab and Python.The software mainly includes three parts: data collection,working condition monitoring,and working condition traceability.Finally,a 5000 t/d clinker production line has been put into trial operation,and the results show that the working condition monitoring and traceability method proposed in this paper can effectively monitor and trace the working condition abnormalities in the cement precalcination process.