Research And Application Of Quality Control System For Pellet

After the realization of automation, there are needs to improve quality and increase production while decrease the cost. The development of advanced control for complex process of producing is come up. Advanced control's target is the quality of products. The products' quality is always examined after they were finished. These quality values are not realtime and they are not useful for quality control system. At the same time, the defects discovered at that time always can't be changed. Large losses will come into being and this will infect the factory's benefits inevitably. In order to change this status, a system to predict the quality of products needs to be built. Based on this system, the quality can be predicted on line and the producing parameters can be adjusted to appropriate values to produce better products. At the same time, the quality contral system is impossible to be built based.Most of the recent systems for predicting and controlling quality adjust the parameters basing on the experiences of workers and the producing flow. However the relationship between the quality and the factors that affect the quality is very complex. The knowledge of this kind of relationship is not perfect. It is not reasonable and scientific to direct the producing of factory basing on this inperfect knowledge so that a new system needs to be brought up. In this paper a new system to predict and control the quality is introduced which bases on data mining and neutral network.Data mining is a new technique, which come into the world with the development and the application of database. It is applied successfully in decision support system and prediction system. Many scientific fields have used this technique. There are many methods which can be used in data mining. As far as pellet factory is concerned, using decision tree gets a certain rule for quality prediction. At the same time, a quality model and a quality-control model are built using neutral network. With these models, the quality control system for pellet is built.