基于稀疏度优化的SA-OMP的谐波和间谐波检测方法OA

In view of the problems in the denoising of the traditional matching tracking class algorithms,a harmonic and interharmonic signal detection method which combines simulated annealing(SA)algorithm and orthogonal matching pursuit(OMP)algorithm to optimize the sparsity k is proposed.The SA algorithm is introduced into the OMP to optimize the sparsity search process.By the random search and annealing mechanism of SA,the optimal sparsity is searched dynamically,which makes the sparse representation reach a balance between reconstruction error and sparsity under a specific sparse transform base and avoids the over-fitting(retaining too much noise)or under-fitting(losing the key components of the signal)caused by fixing the k value.The simulation shows that the proposed algorithm can accurately detect each harmonic and interharmonic component under the condition of noise,and the maximum detection errors of amplitude and phase are 0.8%and 0.021 4 rad,respectively.The signal reconstructed by SA-OMP has a high degree of fitting with the original signal.In comparison with those of the traditional OMP algorithm,its signal-to-noise ratio(SNR)is increased significantly,both its(mean square error)MSE and relative error are improved somewhat,and its MSE and relative error can reach 0.041 and 0.18,respectively,even in the case of low SNR of 10 dB.The proposed algorithm has obvious denoising effect and is effective for harmonic and interharmonic detection.