The Research On Fault Diagnosis Of Tread Extrusion Linkage Production Line Based On Genetic Algorithm

The fault diagnosis technology of equipment has developed rapidly in recent years.It has received more and more attention.Mechanical equipment gradually tends to be large-scaled,automated,speeded up,and complicated.Ensuring the safe and stable operation of the equipment is of great significance to the safety of enterprise production and personnel.Fault diagnosis technology is a cross discipline.In the era of massive data,it is a very important method for fault diagnosis by using the input and output data of the device.This paper takes the tread extrusion linkage production line as the research background.The genetic algorithm is introduced into the fault diagnosis work.The fitness function has been designed properly.The initial population generation method and crossover operator of the algorithm has been improved.The effectiveness of the data classification is verified by comparison test.The main research work is as follows:(1)The technological process of tread extrusion linkage production line,the main equipment including mixer,extruder,calender and etc,as well as the electrical control technology index and temperature index are introduced.(2)The production,development and main technical means of fault diagnosis are introduced.The types,range and causes of the faults of the tread extrusion linkage production line are analyzed.The frequency-domain filtering method is studied.The data collection system and the method of filtering are also designed and selected.(3)The research status and common methods of clustering algorithm are discussed.The AP clustering algorithm based on K near neighbor and the improved density peak clustering algorithm are proposed.The improved density peak clustering algorithm is selected for clustering experiments.The experimental results are compared with some common clustering algorithms,such as K-means,FCM,etc.,which shows that the algorithm has certain advantages in clustering accuracy and outlier detection.The cluster test was carried out on the historical data of the tread extrusion linkage production line.(4)The genetic algorithm is deeply researched.Aiming at the problem of slow convergence rate and local optimum of classical genetic algorithm,an improvedgenetic algorithm is designed and described in detail.The comparison experiments of Iris data set,Breast-Cancer data set and Dermatology data set show that the algorithm has faster convergence speed and higher precision than classical genetic algorithm.Finally,the algorithm is applied to the fault diagnosis of the tread extrusion linkage production line,and the rule set corresponding to the fault category is obtained.The result shows that the improved genetic algorithm is effective for data classification problems.