CAJ | 학술논문

为了提高谐波和间谐波分析的速度、精度和抗噪性能，将基于分段迭代的增强型Adaline神经网络应用于电力系统谐波和间谐波分析。该网络在加汉宁窗双谱线插值FFT算法的基础上，将采样数据按采样时间分段，依次用各段对应的误差信息来调整增强型Adaline神经网络的参数。该方法结合一点迭代法和全部点迭代法的优点，既将各时段内的误差进行平均，减少噪声对参数调整的影响，又充分保留误差中包含的谐波和间谐波信息，提高网络的精度。另外，根据参数估计误差和频率对误差函数一阶偏导之间的关系，提出修正信号幅值最大分量对应的频率调整量的处理方式，提高网络的实时性和精度。仿真结果验证了分析结论的正确性。
위료제고해파화간해파분석적속도、정도화항조성능，장기우분단질대적증강형Adaline신경망락응용우전력계통해파화간해파분석。해망락재가한저창쌍보선삽치FFT산법적기출상，장채양수거안채양시간분단，의차용각단대응적오차신식래조정증강형Adaline신경망락적삼수。해방법결합일점질대법화전부점질대법적우점，기장각시단내적오차진행평균，감소조성대삼수조정적영향，우충분보류오차중포함적해파화간해파신식，제고망락적정도。령외，근거삼수고계오차화빈솔대오차함수일계편도지간적관계，제출수정신호폭치최대분량대응적빈솔조정량적처리방식，제고망락적실시성화정도。방진결과험증료분석결론적정학성。
To improve the analytic speed, accuracy and anti-noise performance for harmonic and inter-harmonic a-nalysis, a piecewise iterative algorithm based enhanced Adaline neural network is adopted in this paper.Firstly, Han-ning-window double spectrum line interpolation FFT algorithm is used to make the sampling data divided according to the sampling time and to adjust the enhanced Adaline neural network parameters based on corresponding errors for each segment.Combing one-point and all-point iterative algorithm, the proposed method average the errors from all time frames to reduce the effect of noise on parameters adjustment, as well as preserve the harmonic and inter-harmonic in-formation included in errors, so as to increase the network accuracy.In addition, according to the relationship between parameter estimation errors and the first order partial derivation of error function with respect to its activation functions, a frequency adjustment method corresponding to the maximum component of correction signal amplitude is proposed to improve the real-time performance and accuracy.Simulation results verify the correctness of the analysis conclusions.