Research And Development Of Intelligent Testing Platform For Protector Production Line

Intelligent detection promotes computer automation develop continuously in the information age,modern manufacturing enterprise continues to explore and pursue more intelligent detection as a goal.Protector production line must pay attention to the improvement of test efficiency and qualification rate due to its’ high productivity and detection load.Therefore,this paper studies the optimization and transformation of the testing process in protector production line,designs an intelligent detection platform.This platform can do protector online detection,record and monitor real-time detection data,analyze the influence of 5M1 E factors on protector detection yield,and predict the detection yield based on the factor of production 5M1 E,direct production and increase profit.Firstly,analyzes the status of current protector testing process in this paper,collects problems of the original test procedure,make intelligent detection platform design target clear,separates the platform into detection data acquisition & monitoring sub-platform and detection data integrated management sub-platform,and make the function design respectively.Secondly,designs the hardware and software of the detection data acquisition &monitoring sub-platform in this paper.the hardware design consists of three layers: monitoring management layer,process control layer and detection control layer,each layer has been configured and selected.Software design parts imports SPC quality control software to realize process control in Mass production,designs the architecture and function of the software,and make some examples.Then,designs the data warehouse and detection data integrated management sub-platform database in this paper.Define the topic of data warehouse of product detection yield.Establish the conceptual model and physical model of data warehouse,the data warehouse extract,transform,load data from SCADA database and the production management information system database which contain corresponding 5M1 E statistics.Gives conceptual model examples and physical model examples of data management sub-platform database according to the functional requirements.In addition,studies the yield prediction function in this paper.Use Genetic Algorithm Optimized BP Neural Network as yield prediction method,studies the prediction model and algorithm of protector detection yield,defines framework and algorithm parameters,then use historical data to train the model,and analyze the predicted results.The prediction results show the model can meet the prediction accuracy requirements of protector detection yield.Finally,designs the program of the detection data integrated management sub-platform in this paper.Use C/S system development mode,C# programming language,gives the program design principle,make program design examples for data statistical analysis,detection yield prediction and detection instrument management module,and tests the system,the results show that all corresponding functions of the platform system are normal and meet the platform design requirements.