【摘要】 为了精确分析整数次谐波和非整数次谐波,讨论了基于参数固定的三角基函数的人工神经网络算法,利用该算法可一次性获得电力系统基波及各整数次谐波的频率、幅值和相位;同时提出了改进的三角基函数的人工神经网络算法,即变参数三角基函数的人工神经网络算法,把改进的模型和FFT结合起来,能实现精确的整数次和非整数次谐波的分析。仿真结果表明了两种算法的正确性和易实现性;同时也验证了改进的算法进一步提高了谐波分析的精度,为分析间谐波提供了依据。更多还原

【Abstract】 In order to obtain precise analysis of integer harmonics and non-integer harmonics in power system,firstly,neural network based on a triangle base function with fixed parameters is discussed,by which the frequencies,amplitudes and phases of fundamental and various integer orders of harmonics could be obtained once only.Secondly,an improved triangle base function based neural network is presented——neural network based on a triangle base function with adjusted parameters,precise analysis of integer harmonics and non-integer harmonics is achieved by combining fast Fourier transform(FFT) algorithm with the improved model.The simulation results have verified the effectiveness and practicability of the two algorithms,the higher accuracy of the improved algorithm is verified.It provides a reliable basis for analysis of non-integer harmonics.更多还原