永磁调速器建模与优化设计研究

发布时间：2018-02-12 19:38

本文关键词： 永磁调速器三维有限元支持向量机混沌搜索优化　出处：《东北大学》2011年硕士论文　论文类型：学位论文

【摘要】：永磁调速器是一种先进的调速节能产品。除传递效率高外,还具备许多其他调速设备不具备的优点,如高安全性和高可靠性、低故障率、长寿命、维护成本低、可在严酷条件下运行等。在保证永磁调速器电磁性能的基础上,优化结构参数、改进磁路及生产工艺、进一步提高产品性能具有重要的实际意义。永磁调速器的结构参数优化是一个复杂的非线性多维空间优化问题,本文重点研究了基于智能算法的永磁调速器建模和优化设计问题。利用有限元软件ANSYS建立了永磁调速器的三维有限元模型,对永磁调速器的电磁场分布进行了三维有限元仿真,分析了永磁调速器的磁场分布、涡流分布、热分布,同时计算了永磁调速器在额定转速下的输出扭矩和损耗功率,并深入分析了永磁调速器结构参数变化对其性能的影响。为了降低智能优化算法的计算成本,本文提出了基于智能非线性回归算法的永磁调速器建模方法。利用BP神经网络、模糊自适应神经网络和支持向量机建立了永磁调速器的非线性回归模型,并将基于这三种非线性回归模型的结果与有限元分析结果相比较,对三种方法的结构、建模效率、训练精度、预测性能及在参数分析中的应用作了对比分析。实验结果表明,用支持向量回归算法进行预测能够取得比其他方法更好的效果。在此基础上,应用支持向量机为永磁调速器建立了数学模型,为永磁调速器的优化设计奠定了基础。提出了基于智能优化算法的永磁调速器优化设计方法,分别应用粒子群算法和混沌搜索算法对永磁调速器主要结构参数进行了优化设计,实验结果表明,应用混沌搜索算法优化永磁调速器结构参数更具有优越性。最后在有限元软件ANSYS环境下对优化后的永磁调速器进行了电磁场仿真,并与样机的仿真结果进行了对比分析,验证了支持向量机建模和应用混沌搜索算法优化永磁调速器结构参数的合理性与有效性。
[Abstract]:Permanent magnet speed governor is an advanced speed control and energy saving product. Besides the high transmission efficiency, it has many other advantages such as high safety and high reliability, low failure rate, long life and low maintenance cost, which can be run under harsh conditions.
Based on guaranteeing the performance of the permanent magnet electromagnetic governor, and optimize the structure parameters of magnetic circuit and improved production technology, improve product performance has important practical significance. The optimization of structure parameters of permanent magnet speed regulator is a complex nonlinear multidimensional optimization problem, this paper focuses on the research of permanent magnet governor modeling and optimization design of intelligent algorithm based on.
The establishment of a three-dimensional finite element model of permanent magnet speed by using finite element software ANSYS, the three-dimensional finite element simulation of the electromagnetic field distribution of the permanent magnet governor, analyzes the magnetic field distribution of permanent magnet eddy current governor distribution, heat distribution, permanent magnet speed in the rated speed of the output torque and power loss were calculated, and further analysis of the effect of permanent magnet governor structure parameters on its performance.
In order to reduce the computational cost of intelligent optimization algorithm, this paper proposed the permanent magnet governor modeling method based on nonlinear regression algorithm. Using BP neural network, fuzzy neural network and support vector machine to establish nonlinear permanent magnet governor regression model and regression model based on the three kinds of nonlinear finite element analysis results and the results in comparison, the structure of the three methods, modeling efficiency, precision of training, and compares the prediction performance and application in parameter analysis. The experimental results show that using support vector regression algorithm to forecast can be achieved better results than other methods. On this basis, the application of support vector machine to establish the mathematical model for the permanent magnet speed governor, which laid the foundation for the optimization design of permanent magnet governor.
The intelligent optimization algorithm of permanent magnet speed optimization design method based on the applied particle swarm algorithm and chaotic search algorithm is used to optimize the main structure parameters of permanent magnet speed, the experimental results show that the optimization of structure parameters of permanent magnet speed regulator has more advantages in application of chaotic search algorithm. Finally the electromagnetic simulation of permanent magnet speed governor after optimization in finite element software ANSYS, and simulation and prototype results were compared to verify the rationality and effectiveness of the support vector machine modeling and application of chaotic search algorithm to optimize the structure parameters of the permanent magnet governor.

【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH139

【参考文献】