Research And Design Of Steel Components Quality Management System Based On Improved RBF Neural Network

The steel structure building has many advantages,such as low pollution,time-saving,land-saving,light weight and recyclable.Recently,it is more and more widely used at airports,stations,pavilions,shopping malls and other public facilities.However,steel structure building related accidents,which are affected by the quality of steel structure directly,have happened frequently at home and abroad,thus the quality problem of the steel structure has drawn much attention from the public.The traditional artificial quality test can hardly meet the safety requirement due to the complication of steel members manufacture.With the development of information technology,it is necessary to establish a scientific and efficient quality management system.Taking a large steel structure company in Hebei Province as research object,a scientific quality management system is designed combined with the enterprise work habits.Because of the complex manufacture of steel member,the proposed quality management system takes the RBF neutral network and expert system to test the quality.In the system,neutral uses RBF,which has the advantages of strong nonlinear mapping,strong self-learning and local approximation.As genetic algorithms having the ability to search the global optimal solution,the parameters of the RBF neural network can be optimized by it.Expert system can handle the complex quality information in the quality detection.In the combination of improved RBF neural network and expert system,the self-learning ability of neural network can make up the shortcoming of traditional expert system whose knowledge base needs manual update.Based on the business process and requirement,this paper designs a quality management system starting from quality inspection and quality tracking.The proposed system strictly control quality information input and ensure the information flow between different modules to solve the “information island” problem in traditional quality management system.The system is implemented by choosing reasonable hardware and software,designing database according to demand and writing program.At last,sample data is adopted to verify the function of the system.