Research On Quality Data Acquisition System Of The Electronic Control Valve And The Intelligent Prediction Method Of Pass Rate

With the advent of the "Industry 4.0" era,traditional manufacturing industry is facing an industrial upgrade aiming at achieving intelligent manufacturing.As an important control component in automobile air-conditioning compressor,the electronic control valve still plays an important role,and its manufacturers need to complete the industrial upgrading in the new round of industrial transformation.At present,the production enterprises for the quality control of the electronic control valve mainly rely on the traditional way,such a way will often be due to the control is not timely and produce a large number of non-conforming products.In this paper,the product quality data acquisition system of electronic control valve is designed,and on this basis,using the deep learning method,a product pass rate intelligent prediction method is explored.The main contents are as following:1.On the basis of the analysis of the demand for quality data acquisition in the manufacturing process of electronic control valve products,the overall structure of the electronic control valve quality data acquisition system is established,and the general idea of the intelligent prediction method of product pass rate is put forward.2.The electronic control valve quality data acquisition system mainly expounds the specific design and implementation of the system from the perspective of client and server.The client section mainly introduces the scheme of implementing product quality data collection from the terminal equipment,the server section introduces the network connection,interface design and software development,and the database section introduces the database design,including the selection of the database and the design of the data table.Finally,several methods of data preprocessing are introduced,which can be used to provide choice and reference for the data processing of intelligent prediction of product pass rate.3.An intelligent prediction method of product pass rate based on deep learning is proposed.First of all,the principle of RNN is studied,and the shortcomings of RNN are analyzed from the practical application point of view.On this basis,the deformation form LSTM is proposed,and the optimization scheme of LSTM relative to RNN is expounded,so as to determine the implementation scheme of LSTM and verify its rationality;The deep learning optimization algorithm is studied,including activation function,overfitting optimization algorithm and learning rate optimization algorithm.Finally,the qualification rate prediction model of the electronic control valve based on Auto Encoder-LSTM is determined.4.To build a test environment,the two kinds of pass rate prediction models based on LSTM and Auto Encoder-LSTM are compared and analyzed,and the effectiveness of the prediction function is verified by making predictions about the product pass rate in the short term of the electronic control valve production line.