Attribute Reduction Based On Quantum-behaved Particle Swarm Optimization With Multi-swarm And Immunity Algorithm And Their Application In The Fault Diagnosis

With the progress of modern industries in large-scale and integration, processes have become more and more complex and contain a large lumber of measure variables, so once faults occur in the production process, they will lead to the huge economic losses, and even casualties. Therefore, timely and accurately detecting and diagnosising the process fault has a very practical significance.To implement fault diagnosis rapadly and reliable, the thorough research has been conducted on Particle Swarm Optimization (PSO). Then the quantum system, Immunity algorithm and Multi-swarm algorithm are introduced. The test results of function show these operations provide better optimization capabilities.On this basis, MIQPSO is proposed. In this algorithm, it divides the whole particle swarm into different groups and searches in different phases, and vaccination can guide the particles to evolve towards a much better direction. It is validated experimentally that this algorithm achieved much better result both in convergence speed and optimization capabilities in comparison with other algorithms.Finally, the MIQPSO-based attribute reduction algorithm is applied to solve fault feature selection of fault diagnosis. The simulation results show that feature selection based on this algorithm and SVM has an excellent fault diagnosis performance for fault diagnosis of TE process.