Reasearch And Implementation Of Quality Diagnostic Method For Iron And Steel Smelting Process

With the development of iron and steel smelting technology, the quality of iron and steel products has gradually become the top priority of people’s attention. How to improve the quality of steel products, reduce the failure rate of the product, reduce unnecessary energy consumption and waste of resources, has become one of the important means of iron and steel enterprises to enhance their core competitiveness in the fierce competitive environment, has become the basis for the long-term survival and development for enterprises.In the actual production process, the appearance of the unqualified products is inevitable. Through the analysis of the production process of unqualified products, to find out the reasons for production of substandard products and to adjust to avoid the recurrence of the problem,this process is called quality diagnosis. The traditional quality diagnosis mostly constructs the SPC control chart using the production process data,and carries on the pattern recognition to the control chart. However, iron and steel smelting is a multi-stage, multi-parameter, complex process of production process, it is difficult to build control chart through production data. Therefore, in this paper, we use the method of data mining, combine the quality diagnosis and data mining, reaserch and implement a quality diagnosis method for steel smelting process data.Based on the analysis of data characteristics of iron and steel smelting process, this paper presents two quality diagnosis algorithms on iron and steel smelting process data, to analyzes anomaly data in the process of iron and steel smelting. The algorithm aims to find the stage and parameters of abnormal data in the process of iron and steel smelting,and to modify them. In order to adapt to the large amount of data in the actual steelmaking process, we use the Spark parallel computing framework to parallelize the two quality diagnosis algorithms. In this way,the algorithm efficiency can be improved and the algorithm will adapt to the more complicated application scenarios. In order to verify the effectiveness of the quality diagnosis algorithm, this paper based on the classification algorithm, using BP neural network to train the existing iron and steel smelting process data,building a classification model to judge wether the unknown quality data can produce qualified products. On the basis of researching and implementing the quality diagnosis algorithm and quality evaluation method,we will design and implement a quality diagnosis system which can analyze and diagnose the iron and steel smelting process data, and evaluate the quality of the unknown data.Provide a friendly interface to the user, and display the results of diagnosis and evaluation in a visual way.