中文摘要：

针对间谐波不易检测的特点,提出一种新的间谐波参数分析方法。首先利用加权信号子空间投影算法优化的多重信号分类算法（MUSIC,multiple signal classification）对间谐波信号频率进行估计。在分析过程中,利用欧拉公式将信号转化为空域信号,运用改进后的谱函数对谱峰进行搜索,实现信号的频率估计。最后利用蜜蜂算法结合收敛速度较快的自适应最小均方（LMS,least mean square）算法以2种群协同进化的方式,实现间谐波信号的幅值与相角估计。研究结果表明：改进后的MUSIC算法提高了非理想情况下间谐波参数估计的精度;采用协同进化蜜蜂算法减小了算法的迭代次数以及陷入局部极值的概率,同时也提高了工蜂位置向量的准确度。

英文摘要：

Considering that the inter-harmonics can hardly be detected, a new analyzed method of inter-harmonics parameter was proposed. Firstly, by using MUSIC（multiple signal classification） optimized RWSP（random weighted signal-subspace projection） algorithm, the frequency of inter-harmonics signal was estimated. In the process of analysis, the signal to spatial signal was transformed to estimate the singal’s frequency with the Euler formula and the spectral function. Then the bees algorithm was applied combined with adaptive LMS（least mean square） and the approach of the two population co-evolutionaries to estimate the amplitude and phase of the inter-harmonics. The results show that the proposed RWSP-MUSIC can improve the estimation of non-ideal circumstances interharmonic parameter accuracy, the proposed coevolutionary bees algorithm can decrease the number of iterative algorithms and the probability of trapped in the local extremum, and can improve the accuracy of the worker’s position.