-
About
Contact UsA large group company integrating R&D, production and sales of mechanical and electrical equipment. -
Products
Contact UsA large group company integrating R&D, production and sales of mechanical and electrical equipment. -
Contact UsA large group company integrating R&D, production and sales of mechanical and electrical equipment. -
Contact UsA large group company integrating R&D, production and sales of mechanical and electrical equipment.
Metallurgical equipment manufacturers talk about how to effectively improve the finished product rate of castings.
Release time:
26 Sep,2024
Metallurgical equipment manufacturers on the casting defect diagnosis expert system is based on this situation and the establishment. For example, the expert system of casting defect analysis can help foundry producers to identify defects, analyze the causes of defects, and finally give a solution, which is mainly done through the reasoning mechanism of the expert system. However, the expert system also has shortcomings, that is, how to describe the experience of experts very accurately, how to make accurate judgments under fuzzy production conditions, how to foresee defects before defects occur, and then solve defects.
Metallurgical equipment manufacturers talk about the emergence of neural networks to provide the possibility of defect prediction. Through the forward calculation of neural network, the casting defects can be predicted, which has been realized, such as "neural network model for defect diagnosis of ductile iron castings". The system summarizes the common defects and influencing factors of ball iron parts, according to the process parameters in the production process of ball iron, uses the method of neuron network to establish the prediction system of ball iron defects, and puts forward preventive measures to predict the possible defects, so as to effectively improve the finished product rate of castings. When using neural networks to predict defects, attention should be paid to the reasonable description of process measures, and fuzzy description is an effective method, such as the rationality of gate and riser settings.
In addition, if the predicted types of defects are many and the network is large, the genetic algorithm should be used to train the neural network, which can greatly reduce the number of samples and improve the speed of network training, and it is believed that these methods will soon be applied to this field. In addition, some studies are still derived under specific experimental conditions, which limits their versatility.
Previous Page:
Related News
A modern high-end mechanical equipment manufacturing enterprise with design and development, processing and manufacturing, engineering installation services and international trade
Company factory area
No.6, Yeyu Street, Jinyuan District, Taiyuan City, Shanxi Province
Technology Center
No.1, Qianfeng South Road, Wanbailin District, Taiyuan City, Shanxi Province
TAIYUAN NORTH HEAVY INDUSTRY MACHINERY CO., LTD
Powered by 300.cn SEO Privacy Policy