Font Size: a A A

Research On The Casting Slab Surface Quality Prediction Expert System

Posted on:2006-08-01Degree:MasterType:Thesis
Country:ChinaCandidate:G S JiangFull Text:PDF
GTID:2121360155468265Subject:Iron and steel metallurgy
Abstract/Summary:PDF Full Text Request
In the continuous casting process, it is important to predict the surface quality of the casting slab in real time, for ensuring process continuity, improving product quality and reducing production costs. Famous corporations all over the world pay attention to it, and have developed the casting slab quality prediction system, which evaluates the quality, classifies the products and tracks the process. In fact, most of the quality defects origin of the dynamics and thermodynamics process in the mold. The temperature information from thermocouples embedded in the mold can reflect the slab behaviors directly.The Casting Slab Surface Quality Prediction Expert System has been developed in Visual Basic 6.0,by using two kinds of artificial intelligence technique, i.e. neural network and expert system. The system conducts temperature pattern recognition of the thermocouples embedded in the mold by neural network module and logic module, and collects equipment and technological parameters of the caster to estimate the possibility of surface quality question by means of metallurgical theories and expert experience. The expert system has been verified off-line using historical data, and the results indicated it might predict parts of the quality defects, according to the causes in the process.This system adopts the research findings at home and abroad and our experience; it puts forward a number of standpoints, as follow:(1) It introduces the concept of artificial neural network , and designs new type of neural network expert system;(2) When it trains BP (back propagation) network ,it adopts the method of piece division and error subsection;(3) It regards both pulling rate and the temperatures data as a node, therefore enlarges the functions of the network.(4) It develops logic module ,as the comparison to the neural network module;(5) Using temperature ,mold level, mold powder, steel grades, and carbon content, sulfur content and phosphorus content of liquid steel, it predicts the surface crack of the casting slab .
Keywords/Search Tags:continuous casting, surface crack, neural network, surface quality prediction, expert system
PDF Full Text Request
Related items