Research On Modeling And Optimization Design Of The Permanent Magnet Drive

The Permanent Magnet Drive is a kind of advanced energy-saving product. Except high transfer efficiency, this device has many other advantages compared with other speed control equipments, such as high security, high reliability, low failure rate, long working life, low maintenance costs and operation in harsh conditions.On the basis of maintaining its basic electromagnetic performance, it is of great practical value to optimize the structural parameters and improve cost performance of the product. It is a complicated nonlinear multidimensional space problem to optimize the structural parameters of the Permanent Magnet Drive. The objective of this paper is to research modeling and optimization design of the Permanent Magnet Drive based on intelligent algorithm.The electromagnetic analysis of the Permanent Magnet Drive is carried out by the Finite Element Method. A3-dimensional Finite Element Method model of the Permanent Magnet Drive is first created by using the Finite Element Method software ANSYS, and the distribution of magnetic field, eddy current, heat loss power and output force are analysed. Then the influence of the structural parameters’ alteration on the performance of the Permanent Magnet Drive is also deeply analyzed.A large scale iterative calculation is needed for optimizing the structure parameters, and the intelligent nonlinear regression modeling algorithms are proposed in this paper. Based on the electromagnetic characteristics obtained by Finite Element Method analysis, back propagation neural network, adaptive neuro-fuzzy inference system and support vector machine are utilized to establish the nonlinear regression models of the Permanent Magnet Drive’s output characteristics respectively, and then the results from the nonlinear regression methods and the Finite Element Method are compared to verify the accuracy, validity and forecasting capability of these three modeling methods. Results indicate that support vector machine method can make best use of the training samples’ distributed character of the three methods, and it is best on the stability and generalization ability, too. The nonlinear models built by support vector machine provide effective tools for the optimization design of the Permanent Magnet Drive.Intelligent optimization algorithms, particle swarm optimization and chaos search, are proposed in this paper to optimize the structural parameters. Results indicate that it is has more advantages to optimize the structural parameters with chaos search. After optimization, the output performance of the new designs is calculated by using the Finite Element Method software ANSYS and compared with that of the prototype to verify the rationality and validity of the optimization result obtained by support vector machine and chaos search.