Research On Data-driven Quality Control Of Cold Rolling Manufacturing Process

At present,the cold rolling industry is developing rapidly,which makes the competition among enterprises become more and more intense.The quality of cold rolled products is one of the key factors to determine the competitiveness of enterprises.With the rapid development of computer technology,network technology and artificial intelligence,data-driven methods are widely used in the field of quality control.In this paper,data-driven quality control technology is studied in the context of cold rolling manufacturing process,and the main research contents and its specific work are as follows:Firstly,the characteristics of cold rolling process are introduced,and the quality control framework of cold rolling manufacturing process is constructed for the problems such as irregular data collection and processing methods and less data analysis means in cold rolling process.The framework specifies the methods and processes for cold rolling data collection,processing and analysis,and provides guidance for applying technical means such as quality prediction and mining.Secondly,a cold rolling product quality prediction model is proposed to address the problem that it is difficult to set reasonable process parameters for the cold rolling process.By analyzing the cold-rolled data,we use the feature importance ranking and principal component analysis methods to realize the dimensionality reduction of the data and obtain the key factors affecting the cold-rolled product quality.The cold-rolled product quality prediction model was constructed based on random forest algorithm,and the main hyperparameters of the model were optimized using genetic algorithm to improve the model performance.The model was validated using production data from a cold rolling enterprise,and the model accuracy was demonstrated by comparing the prediction performance of this model with three other control models.Then,a cold rolling quality data mining model is constructed for the phenomenon that cold rolling enterprises accumulate a large amount of cold rolling data but cannot make full use of it.The K-means method was used to discrete the continuous attributes in the cold-rolled data to suit the mining needs.Data mining is performed using FP-Growth algorithm to obtain and illustrate the correlation between production factors and product quality.For the update problem of the cold rolling database,an incremental mining model based on the FP-Growth algorithm is constructed to improve the mining efficiency in order to ensure the timeliness of the data and rules.Finally,based on the previously described research,a cold-rolled product manufacturing process quality analysis system was designed and developed using IntelliJ IDEA development platform and Mysql database to realize the application of machine learning model to analyze cold-rolled production data.