基于LMS的加速收敛步长选择原理与自适应三相电网谐波检测

Principle of step-size selection for accelerating convergence and adaptive current detection of three-phase power system

基于最小二乘(LMS)算法的自适应电网电流的检测系统中,为克服检测的快速性与稳定性的矛盾,当偏离最佳值较远时系统方均误差(MSE)较大,要提高检测算法迭代收敛速度的策略是选择较大的迭代步长;而在偏离最佳值较近时系统MSE较小,为降低波动保证系统检测的稳态精度,应该选择较小的迭代步长。通过理论分析阐明了提高自适应电网电流检测LMS算法收敛速度的步长选择原理,并且采取选择迭代过程最佳变步长的策略提高自适应三相电网电流检测的性能。仿真分析验证了该自适应检测算法具有更快的动态性能与更好的稳态精度。

When the convergence conditions of LMS（Least Mean Square） algorithm in the LMS-based adaptive detection system of power grid current are satisfied,the system MSE（Mean Square Error） is larger when the value is far from the optimal value,larger step-size should be selected to accelerate the convergence speed of algorithm,while the system MSE is smaller when the value is near the optimal value,smaller step-size should be selected to ensure the steady accuracy of detection system. The principle of step-size selection for accelerating the convergence speed of LMS algorithm is introduced and the optimal variable step-size strategy is applied to improve the system performance of three-phase grid current detection. Simulative analysis validates the better dynamic performance and steady accuracy of the proposed adaptive detection algorithm.