A knowledge resources fusion method based on rough set theory for quality prediction

Abstract

Quality prediction of the complex surface parts manufacturing process can be automated by designing a knowledge fusion model based on rough set algorithm. Such a model can be constructed by combining qualitative and quantitative knowledge resources modeling, and quality data mining. In this procedure, knowledge fusion model is used to merge knowledge resources in production process into data mining models, which could mine more abundant decision rules for various production conditions. This work exposes the correlation of knowledge resources that fuses in such indiscernibility correlation relation and the function relations. Quality prediction is shown for an operational evaluation processes and a set of prediction rules with varying precision and different structural forms. Results on quality prediction of multi axis milling for large hydraulic turbine blades show that knowledge fusion model alleviates the uncertainty of sample data and improves product quality conversely. Moreover, the prediction method is indeed beneficial to high quality and precision manufacture.