Development Of Dynamic Data Mining Platform For Production Quality Control

Production quality control is a very important issue to the improvement of product quality in the manufacture enterprises. Large amount of data is acquired by numerous sensors and there is surely much information related to quality. This thesis expounds on the use of dynamic data mining in quality control based on process properties in manufacture enterprises. The dynamic data mining technique is discussed in this thesis to solve a quality prediction problem and a quality analysis problem. The main contributions of this thesis are as follows:1. Introducing the basic concepts and methods of data mining, and then giving a brief view of the future development of data mining. Furthermore, the criterion of data mining tools is introduced.2. The quality prediction problem and the quality analysis problem are proposed. Based on the process properties in manufacture enterprises, the thesis presents the method of how to solve the two problems using dynamic data mining technique.3. Applying the method to the continuous casting process in order to improve the quality of the slab. The experimental results show the validity of the method.4. A data mining tool: DMPlatform is developed for production quality control in manufacture enterprises. 4 different data mining schemes are used for the quality control of 4 metallurgical processes. This software tool is applied in the data analysis section of a steelmaking enterprise and gains good results.