超磁致伸缩作动器非线性模型辨识研究

Parameter identification of nonlinear model of giant magnetostrictive actuator

准确辨识超磁致伸缩作动器非线性模型参数是位移精确控制的必要条件,针对标准粒子群(PSO)算法存在早熟收敛及迭代后期易陷入局部最优的不足,提出一种可动态调整惯性权重、学习因子及带遗传变异的改进型粒子群(IPSO)辨识算法,该算法可平衡全局和局部搜索能力,提高收敛速度和辨识精度,并将该算法应用于超磁致伸缩作动器非线性模型的参数辨识研究。结果表明:该算法能有效可靠地辨识超磁致伸缩作动器非线性模型参数,计算值和实验的吻合程度较高,并且具有一定的抑噪能力。

Accurate identification of nonlinear model parameters is a prerequisite to precisely control the displacement of giant magnetostrictive actuator. Aiming at the shortcomings of standard particle swarm optimization （PSO） algorithm such as existing premature convergence and easily falling into local optimum in later iteration, an improved PSO identification algorithm was proposed, which can dynamically regulate the inertia weighting, study factors and genetic variation, and so, balance the global and local search capability to improve the convergence speed and identification accuracy. Moreover, it was applied to the parameters identification of nonlinear model of giant magnetostrictive actuator. The results show that： the improved algorithm can effectively identify the nonlinear model parameters of giant magnetostrictive actuator. There is a higher degree of agreement between the results of calculations and experiments and the algorithm also has a better anti-interference ability.