Design And Experiment Of Abnormal Analysis System For Angle Grinder Production Process

With the continuous promotion of Industry 4.0,the angle grinder industry is developing towards intelligence.For the angle grinder industry,the stability of the production process plays a crucial role in product quality,and can be characterized by the test results of the product.When quality problems frequently occur in the testing process of the angle grinder,it is necessary to investigate its production process.However,the existing testing methods are manual testing,and the current method for identifying abnormal links in the production process in enterprises is for employees to trace the source of abnormal links in the production process through abnormal components and based on their own experience.However,in the face of multiple components of the angle grinder and its complex production process,this greatly increases the consumption of manpower and time,making it difficult to meet the needs of efficient production in enterprises.Therefore,the implementation of intelligent analysis of abnormal links in the production process of angle grinder is of great significance for ensuring product quality and safe production.The angle grinder and its production process are taken as research objects to complete the design of an abnormal analysis system for the production process,in order to achieve automation and intelligence in detection and analysis.By using machine learning instead of manual detection methods,problems such as low manual detection efficiency and inconsistent detection standards have been solved;And by adopting a probability graph model,intelligent analysis of abnormal links in the production process has been achieved,improving the efficiency of troubleshooting.The research content is as follows:(1)Based on historical data,the main production reasons for abnormal components of the angle grinder are identified,and four types of frequently occurring component abnormalities are selected for subsequent research.By analyzing the manual detection mode,a testing data collection platform is established to simulate actual working conditions.The data collection is completed and its feature parameters are extracted.The key characteristic parameters of the angle grinder are selected through the Spearman correlation coefficient method.Data support is provided for subsequent detection of abnormal components in the angle grinder(2)In order to overcome the problems of solve the problem of poor solving time and easy falling into local optima,a chaotic population and dual strategy differential evolution algorithm is proposed.In response to the problems of outdated technology and low detection efficiency in existing detection methods,intelligent detection of abnormal components is achieved by combining the improved differential evolution algorithm and support vector machine.Detection experiments are carried out,and the detection accuracy is 96.74%,the single run time is 16.11 s,and the optimal model is obtained.Based on the optimal model,validation experiments are conducted and the detection accuracy is 96%.The accurate types of abnormal components are provided for subsequent abnormal analysis of the production process of the angle grinder.(3)On the basis of determining the types of abnormal parts,aiming at the problem that the existing production process data is missing and the analysis process depends on the experience of employees,a fuzzy Bayesian network based on analysis method for the abnormal production process of angle grinder is proposed.As the most critical component in the angle grinder,the rotor is taken as the research object,and the abnormal production process of the rotor is analyzed.Based on the rotor production process,the main production process that causes the rotor abnormality is determined.The fault tree analysis and Bayesian network for abnormal analysis of rotor production process are established through the connection between the production processes.Due to the lack of historical data,fuzzy theory and DUOWA operator are introduced to synthesize the evaluation information of multiple experts,and the assignment of the prior probability of the root node is completed,and a fuzzy Bayesian network is built.And abnormal analysis of the production process of the angle grinder rotor is completed through Netica software.The theoretical support for the analysis of the production process of other components is provided.(4)A production process anomaly analysis system based on the PyQt framework is developed.A test bench is built based on the system and testing data collection platform.The abnormal component detection experiments and production process abnormal analysis experiments are conducted.The experimental results show that the accuracy rate of the system for detecting abnormal component types is 95.33%.An analysis of abnormal production processes is conducted on abnormal components of the angle grinder,and possible abnormal production processes are obtained.Compared with the investigation results of employees,the feasibility of the production process abnormal analysis system is proved.