双自适应神经网络和快速TLS-ESPRIT的谐波检测方法及仿真

针对BP神经网络应用于谐波分析时收敛速度慢、初始值选取不当等问题,为实现谐波的准确检测,提出双自适应BP神经网络和快速TLS-ESPRIT(总体最小二乘法—旋转矢量不变技术)相结合的检测方法。该方法利用快速TLS-ESPRIT算法得到频率和信号源个数,将频率作为BP神经网络的初始值,信号源数作为中间节点个数,经双自适应BP神经网络得到网络权值,进而完成谐波的幅值和相位检测。仿真试验结果表明,该算法在检测谐波时,检测速度更快,且具有较高的检测精度。

Double Adaptive Neural Network and Fast TLS-ESPRIT Harmonic Detection Method and Simulation

Aiming at the problems of slow convergence speed and improper initial value selection in harmonic analysis with BP neural network, a detection method combining double adaptive BP neural network and fast TLS-ESPRIT (Total Least Squares-Estimation of Signal Parameters via Rotational Invariance Techniques) is proposed to realize the accurate detection of harmonics. In the method, the frequency and the number of signal sources are obtained by using the fast TLS -ESPRIT algorithm. The frequency is taken as the initial value of BP neural network and the number of signal sources as the number of intermediate nodes. The network weight is obtained through the double adaptive BP neural network to complete the amplitude and phase angle detection of harmonics. Simulation results show that the algorithm can effectively detect harmonics with faster operation speed and higher detection accuracy.