Fault Diagnosis For Tension System Of Continuous Annealing Line Based On Ant Colony Optimization

Annealing is an important process for metallurgical industry in order to improve the steel tensile strength, and the mechanical properties of cold-rolled sheet product. At present, the most widely used annealing method is the continuous annealing process line (CAPL). As CAPL is always been used to produce thin and soft steel sheet under high speed, the strip tension of CAPL is the key factor to ensure the stable quality and output of product. Referring a lot of related literatures at home and abroad, a method of fault diagnosis for CAPL tension system is proposed in this thesis based on ant colony optimization (ACO) which has distributed computing, the positive feedback mechanism and a greedy searching way.By analysising the formulation mechanism and the features of the typical faults of tension system, the fault sample database for the tension system is established based on the field fault data in this thesis. On this basis, the fault diagnosis rule extraction solution is designed by using ant colony optimization algorithm for optimal classification in fault sample database, and the diagnosis rule can be extracted. At the same time, in order to reduce the redundant information of fault feature and simplify the diagnosis rule, the Sensitivity and the Accuracy are defined to evaluate the classification performance of diagnosis rules searched by ACO. In addition, the Path Visbility is defined to provide expert guidance for ant colony searching in order to avoid obtaining the local optimal solution.According to the obtained fault diagnosis rules, a fault diagnosis method based on ACO is proposed in this thesis. Futhermore, the structure of the process monitoring and fault diagnosis system for CPAL tension system is designed. And the simulation result not only shows that the ACO has good search capabilities and the new method can obtain accurate diagnosis result, but also shows the feasibility and validity of the fault diagnosis system for CAPL tention system based on ACO.