Research On Process Fault Diagnosis Of Cement Clinker Device

The new dry cement production technology is the world's most advanced cement production technology.Its production process involves complex physical and chemical reactions,with the characteristics of non-linearity,pure lag,and large inertia,so real-time monitoring and fault diagnosis of its production process is cement production is an important guarantee for safety and quality.Due to the large number of process parameters to be monitored in the cement production line and the coupling between the parameters,this results in difficulties in sample feature extraction and unclear fault diagnosis.The manual diagnosis method alone cannot meet the requirements for accurate fault diagnosis.Based on this,this paper combines the cement production line of 2000t/d of Yatai Group of Jilin Province as the research background,proposes the diagnosis model of cement process fault based on principal component analysis(PCA)and support vector machine(SVM),and optimizes the key parameters of the model iteratively,the results show that the model has a high diagnostic accuracy.The main research work is as follows:Firstly,the cement clinker production line is described in detail.The main process parameters and common process failures in the process of cement operation are introduced.The causes of failures and treatment measures are analyzed.Emphasis is placed on the many parameters that cause the phenomenon of failure in the cement production process and the coupling between parameters,which makes the decoupling of the data parameters necessary.Secondly,the basic theory of SVM and data processing are introduced in the process of cement fault diagnosis.In the process of high-dimensional data processing,the principle of principal component analysis is introduced to establish a classification original data sets,and tested with Wine and Vehide data sets.It shows that the data after dimension reduction through principal component analysis has a good classification effect and improves the classification efficiency.Based on this idea,the classification model was applied to the process fault diagnosis of cement clinker equipment.46 process parameters were selected to classify the 6 process failures.The original parameters were extracted using PCA and the fault diagnosis model was established.However,due to the arbitrariness of the key parameters of the fault diagnosis model,the actual classification effect is not ideal.Therefore,the optimization of the SVM key parameter penalty that factor c,the kernel function width?~2 and the loss factor?becomes a problem to be solved.Thirdly,the hybrid optimization algorithm(IGWO)was introduced to optimize the key parameters of the SVM model,and the basic theories of the DE algorithm and the GWO algorithm were introduced.Then the feasibility of the hybrid optimization algorithm was verified,and finally the dimension was reduced by the PCA.The data was simulated and verified based on the IGWO-SVM cement clinker fault diagnosis model,and compared with other classification models.The results showed that the model classification accuracy and efficiency were significantly improved.Finally,the design of active safety control scheme for cement clinker installation was completed,and the architecture of DCS control system was introduced.According to the actual production line of Jilin Yatai Cement,Honeywell produced the DCS system PlantScapeSCADA/hc900 to design the software and hardware structure.The configuration flow chart of the system design is given.Finally,the man-machine interface of the DCS system in the central control room is developed to realize the real-time display and control of the field equipment operation status.