基于高频噪声主频估计的自适应最小二乘算法

进一步提高电力系统继电保护的动作速度，要求寻找速度更快的保护滤波算法。传统的最小二乘算法与卡尔曼滤波算法虽然速度较快，但准确度难以提高。通过对电力系统故障信号高频噪声分量的理论分析与仿真研究，发现高频噪声分量存在一个主频，并提出一种使用遗传算法的主频频率估计器。根据高频噪声主频的存在性与自适应滤波原理，提出了一种基于高频噪声主频估计的自适应最小二乘算法，它利用高频噪声主频估计器计算输入信号的高频噪声主频频率，根据主频频率修改最小二乘算法的模型参数，实时形成最小二乘算法的计算式。仿真试验表明，该算法显著地提高了保护滤波算法的滤波速度与准确性，即使在噪声非常严重的情况下，也能取得很好效果。

HIGH FREQUENCY NOISE MAIN FREQUENCY EVALUATION BASED ADAPTIVE LEAST ERROR SQUARE ALGORITHM

Improving the action speed of power system relay protection further demands pursuing faster relay protection filter algorithm. Traditional least error square algorithm and Kalman algorithm has high speed, but their accurateness is not sufficient. Though theoretical Analysis and simulation research on high frequency noise component of power system signal during fault, the paper fmd there is a main frequency in high frequency noise component, and propose a main frequency evaluator based on generic algorithm. According to the existence of main frequency and adaptive filter theory, the paper propose a high frequency noise main frequency evaluator based adaptive least error square algorithm. The algorithm use main frequency evaluator evaluate main frequency of input signal, and adjust filter parameter accordingly, form the formulation of least error square algorithm in runtime. It is proved by computer simulation that the algorithm proposed in the paper improves the speed and performance of relay filter algorithm greatly, and it performs well even in heavy noise situation.