Research On Quality Control And Prediction Method Of Product Assembly

With the upgrading of China’s manufacturing industry to the direction of informatization,digitalization and intelligence,it is an inevitable trend to ensure the development of high quality products.Only by doing a good job in the control of product assembly production and assembly quality can we ensure the quality and reliability of products and better play its due role.There are many influencing factors in the product assembly process,the assembly process is complex,and the assembly dimension constraints between parts are complex,which will have a great impact on the assembly quality.In order to control the product in advance and take corresponding intervention measures in advance,it is necessary to use intelligent methods to control and predict the product assembly quality.In this paper,the assembly of a product is taken as the research object,and the following control and prediction methods are studied for the product assembly quality.In this paper,the assembly of a product is taken as the research object,and the following control and prediction methods for product assembly quality are studied:First,analyze the processes in the actual product assembly process,analyze the factors that affect each process from the perspective of 5M1 E,establish an evaluation index system for the importance of the assembly process,use AHP to assign weight to the evaluation index,use the fuzzy evaluation method as the modeling method,establish a fuzzy comprehensive evaluation model for the importance of the assembly process based on AHP,and evaluate the assembly process with the highest importance through comprehensive analysis,so as to achieve control.Secondly,in view of the lack of sample information and the existence of a large number of redundant and high-dimensional assembly quality characteristic datasets,the key quality characteristics are extracted by combining gray entropy correlation analysis and PCA to reduce the dimensions of the data set.On the basis of retaining useful and important information,irrelevant quality characteristics are eliminated to the greatest extent,and the structural complexity of the prediction model is simplified,which is conducive to improving the efficiency of algorithm operation.Thirdly,assembly quality prediction is conducted based on machine learning,LSSVM prediction model,PSO-LSSVM prediction model,improved PSO-LSSVM prediction model,and BP neural network prediction model are established,and different data sets are used as input vectors of prediction models for training prediction.Finally,by comparing the prediction results of each prediction model through experiments,select the prediction model that is most suitable for solving the assembly quality data set in this paper,improve the stability of assembly quality and assembly efficiency,find out the existing error problems in time,cause unnecessary time and cost consumption,and provide a reference idea for advance control and prediction,which has practical application value.