Research On Online Parameter Detection System And Quality Prediction Method Of Relay

The time of "Industry 4.0" has arrived and the traditional industry will gradually realize the goal of intelligent manufacturing.Relay is the core of the electrical control switch element which will still play an important role in intelligent manufacturing.Then relay manufacturers need to complete the new round of manufacturing reforms in the first step.At present,the degree of intelligence,information,and networking of relay testing equipment is not high,resulting in data dispersion and increasing the difficulty of enterprise management.In the aspect of quality inspection of relays,enterprises generally use sampling methods to detect and use traditional statistical process control methods to control the quality of relays,which has time delay.In order to solve these problems,this paper studies the relay online parameter detection system and quality prediction method.The research on the online parameter detection system of the relay is carried out from both the upper computer and the lower computer respectively.The lower computer takes STM32F407 as the core,and developes on the base of uC/OS-? real-time operating system.The hardware part is designed and realized in modular method,and the parameters of the relay can be completely measured.At meantime,this article has proposed the improvement method to measure the coil resistance and the two times pick-up voltage.The upper computer is developed on the based of Qt program framework and Oracle database.It implements functions such as data management and parameter setting,and has a friendly user graphical interface.Using the TCP/IP protocol,the upper computer and the lower computer can communicate with each other through Ethernet.At last,a number of random measurement experiments are carried out on the relay online parameter detection system.The relative standard deviation and relative error of the measurement results are both less than 5%,which shows the performance is stable and achieves the expected design requirement.On the basis of the online parameter detection system,this article also studies the quality prediction method of the relay.After further study of deep learning algorithm,this article uses the keras platform which written in python to build a short-term relay quality prediction algorithm based on Long Short Term Memory Network.This algorithm can be used to predict short-term failure of the relay production line according to various relay parameters.The prediction of the failure rate can provide reference for the quality control personnel of the enterprise.