基于加权一阶局域理论的综合电力滤波器谐波电流预测方法

HARMONIC CURRENT PREDICTION METHOD OF COMBINED POWER FILTER SYSTEM BASED ON WEIGHTED FIRST-ORDER LOCAL AREA THEORY

受硬件和软件条件的限制,现有的各种谐波电流检测万法都存在延时问题,一股约为一个周期。检测出的结果实际上是一个周期前的谐波电流,因而会产生较大误差。对谐波电流进行准确的跟踪和预测,然后根据预测值进行补偿是解决这一问题的有效方法。该文提出了基于加权一阶局域理论的谐波电流预测方法,该方法在t时刻预测出t+2时刻的谐波电流与其理想值的偏差值,通过选择t+1时刻的控制策略,利用加权最小二乘法使得在t+2时刻该偏差值为最小,从而实现了两步预测的无差拍控制。将中心点的空间距离作为一个拟合参数引入预测过程,提高了预测精度和消噪能力。该方法具有原理简单和实时性好等特点,仿真和实验结果验证了它的有效性和实用性。

Resulting from the performance limitations of both hardware and software,the time delay exists in all the present harmonic current detection methods.It is generally one cycle.The detecting results are,in fact,the harmonic currents of the cycle proceeding to the present one.Therefore,it will produce a greater error. It is considered as an effective way to solve the problem by compensating the harmonic currents based on the tracing and predicting of the harmonic currents. This paper presents a predicting method of harmonic current based on weighted first-order local area theory.In order to achieve deadbeat control,this method predicts the deviation of t+2 moment at t moment, by choosing control strategy of t+1 moment and minimizing the deviation at t+2 moment using the weighted least square method.The prediction precision and the ability of noise elimination are improved by introducing space distance of central point as fit parameter into predicting process. It is simple in principle and good in real time use.The simulation and experiment results demonstrate the practicability and validity of this method.