A Novel Predictive Control for Active Power Filter using a Variable-Step LMS Algorithm

In order to eliminate the impact of time delay to active power filter(APF), a novel predictive control based on variable-step LMS is proposed in this paper. Taking into account the power weights of errors and influence of previous state, the step size varies with both forgetting factor and follow-up factor. Forgetting factor can change the step size according to tracing error and follow-up factor can enhance the capacity of resisting disturbance. As introducing the proposed LMS with variable-step, the performance of the novel predictive control is improved both under steady and dynamic condition. The simulation between the proposed method and traditional LMS based predictive method is carried out and the results demonstrate the effectiveness of the proposed method.