基于粒子群优化算法的异步电动机动态模型参数估计方法研究

利用粒子群优化算法实现了对异步电动机动态模型的参数辨识和转矩估计。通过MATLAB建模仿真。结果表明电动机动态模型的辨识易受干扰,对于噪声敏感度较大;但在噪声适度范围内,能够比较有效的搜索到真实值,且搜索范围广,精度较高;通过多次辨识,并进行加权融合避免了某个参数陷于局部极值,增强了辨识结果的稳定性,有效地实现了转矩估计。

Particle Swarm Optimization based Dynamic Model Parameter Estimation of Asynchronous Motor

In this paper, the parameter identification and torque estimation for dynamic model were realized by using particle swarm optimization (PSO) algorithm. In the model established by MATLAB, the simulation results showed that the dynamic model of three-phase induction motor was sensitive to noise. Under moderate range of noise, the PSO algorithm could find the true value with a wider range and high precision. Moreover, through multiple identification and weighted fusion method, the stability of the identification results was enhanced and torque estimate was effectively achieved for avoiding local extremum.