Research On Big Data Mining Method For Monitoring The Operation Of Bag Filter Products

With the rapid development and popularization of information technology,all walks of life have entered the era of "big data".The data reserves of the whole life cycle of complex mechanical and electrical products are increasing day by day.How to analyze the internal relationship between the big data of the product life cycle and the operation characteristics,so as to guide the coordination and optimization of different life cycle stages of products in a reverse way,is the common face of the current global manufacturing industry challenges.As one of the most important technologies of intelligent manufacturing,big data mining can discover the hidden knowledge in big data.Applying big data mining technology to life cycle big data knowledge discovery is of great significance to product life cycle management.As an important industrial environmental protection equipment,bag filter is widely used in coal,steel,power generation and other industries.Due to the complexity of internal structure and external operation conditions,the performance of the operation characteristics of the dust collector is often closely related to the mutual coupling and feature correlation between various parameters.However,the existing operation monitoring methods usually only consider the influence of a single target parameter or a class of target parameters on the operation characteristics,which makes the monitoring results unilateral and lag.Therefore,this paper proposes a big data mining method for the operation monitoring of the bag filter,the purpose is to comprehensively consider the influence of relevant parameters of each life cycle on product operation characteristics.The research content of this paper is as follows:1.The big data and operation characteristics of the life cycle of the bag filter are analyzed.Firstly,the working mechanism of the bag filter is introduced.Secondly,the source,composition and characteristics of big data of bag filter are described in detail.Finally,the operation characteristics of the bag filter are analyzed,and the inherent relationship between mining the big data of the product and its operation characteristics importance is explained,and the task of big data mining is determined.2.The whole life cycle big data mining framework and the process of big data mining of bag filter are studied,and the data mining algorithm is analyzed and compared.XGBoost algorithm is selected to build the big data mining model of this paper.Aiming at the problem that the convergence speed of XGBoost model’s multi-parameters optimization is slow,and it is easy to fall into the local optimal solution and the accuracy fluctuates greatly,using ant colony algorithm to optimize the important parameters of XGBoost model;using XGBoost model after parameter optimization to sort and screen the importance of relevant characteristics of bag damage;according to the optimal parameters and the important characteristics after screening,retraining to get XGBoost optimization model for on-line monitoring of bag damage of bag filter.3.On the comprehensive experimental platform of bag filter,the experimental module of bag breaking monitoring is expanded and the development of software and hardware system is completed;The experimental scheme of ash cleaning is designed,and the field experiment of bag breaking monitoring method of bag filter based on XGBoost optimization model is carried out to verify the feasibility and effectiveness of the big data mining model.Research and application results show that XGBoost model,as a relatively new algorithm in the field of integrated learning,can effectively mine the relationship between product big data and operation characteristics when applied to big data mining of bag filter.The XGBoost model based on the parameters optimized by ant colony algorithm has a good recognition effect of bag damage state of bag filter,and it has a very fast running time,high recognition efficiency and good practical value.