基于有限元和神经网络方法的电磁斥力机构结构优化设计  

Structural Optimization Design of Electromagnetic Repulsive Mechanism with Finite Element Method and Neural Network Method

作　　者：孟洲恬 淡淑恒 MENG Zhoutian;DAN Shuheng(College of Electrical Engineering,Shanghai University of Electric Power,Shanghai 200090,China)

机构地区：[1]上海电力大学电气工程学院,上海200090

出　　处：《高压电器》2021年第6期196-202,共7页High Voltage Apparatus

基　　金：国家自然科学基金项目(50577040)。

摘　　要：针对高压直流断路器的电磁斥力机构,提出了一种优化机构结构参数的方法。首先运用场路耦合方法对电磁斥力机构建立模型,利用有限元方法进行电磁场计算,分析了斥力盘结构参数、线圈结构参数、放电电容电容值和预充电压变化对动态特性的影响。以电流峰值、电磁斥力峰值、2 ms时刻位移值和速度值作为目标函数,利用神经网络算法,拟合了斥力机构结构参数与目标函数之间的关系,从而对斥力机构的结构进行了优化设计。优化结果表明,在不增大2 ms时刻运动速度条件下,初始运动速度和2 ms时刻位移均得到提高,改善了机构的动态性能。文中提出的优化方法克服了单一变量法计算量大、计算时间长的缺点,可为电磁斥力机构的设计提供新思路。Aiming at the electromagnetic repulsion mechanism of high voltage DC circuit breaker,a method for optimizing the structural parameters of the mechanism is proposed.Firstly,the field-circuit coupling method is used to establish the model of the electromagnetic repulsion mechanism and the electromagnetic field is calculated by finite element method.Next the effects of discharge capacitance value,pre-charge voltage,structure parameters of the coil and of the repulsion disk on dynamic characteristics are analyzed.Taking the current peak,electromagnetic force peak,the velocity and the displacement value of 2 ms as the objective function,the neural network algorithm is used to fit the relationship between the structural parameters of the repulsion mechanism and the objective function,so that the structure of the repulsion mechanism is optimized.The results show that the initial motion speed and the displacement at 2 ms are improved without increasing the motion speed at 2 ms,which improves the dynamic performance of the mechanism.This optimization method overcomes the shortcomings of large calculation and long calculation time,providing a new idea for the design of electromagnetic repulsion mechanism.

关 键 词：电磁斥力机构 有限元法 结构优化 神经网络算法 

分 类 号：TM561[电气工程—电器] TP183[自动化与计算机技术—控制理论与控制工程]