Active filter current compensation method based on the minimum current error model prediction

Abstract

The invention discloses a harmonic current compensation method of an active filter based on the minimum current error model prediction, comprising the following steps: establishing a mathematical model of the active filter, deriving the rate of change of the control amount versus time with the modulation voltage as a variable expression; collection k The input side voltage and harmonic compensation current at all times are used as reference values; according to the expression in step 1) get k after the moment n The predicted value of the control quantity for a cycle; put it into the minimum current error value function J in, yes J Calculate the extreme value to obtain the modulation voltage of the minimum current error; send it to the PWM modulator to obtain the switching signal, and realize the minimum current dead-beat control of the active filter; the present invention reduces its calculation amount, has high dynamic performance, and does not need to introduce feedback link, the multi-step prediction is eliminated, the sampling delay difference is eliminated, and the versatility is strong.

Description

technical field [0001] The invention relates to the field of active filter control, in particular to an active filter harmonic current compensation method based on minimum current error model prediction. Background technique [0002] The active power filter is a device that can dynamically compensate multiple harmonic currents. It has been widely used in the industrial field in recent years; whether it can accurately track the given command current determines the quality of the active filter APF harmonic compensation. Key; active filter control technology usually adopts PID control method; PID controller will delay the dynamic performance of the system due to its inherent hysteresis; active filter using model predictive control, because there is no The use of PID linear control has good dynamic performance, and it has been deeply researched and applied in active filter control technology in recent years. [0003] Traditional model predictive control technology usually adopt...

AI Technical Summary

Problems solved by technology

When the traditional algorithm is used, since the model prediction algorithm only outputs one switch state in each control cycle, the frequency of the switching device of the active filter will not be fixed; at the same time, the harmonic compensation current cannot track the given value more accurately due to the sampling delay. current, so a large error will be generated when the harmonic current change rate is relatively large; in addition, the calculation amount predicted by the traditional model is relatively large, which will increase as the number of switch states increases

Therefore, there are the following technical problems in the prior art: since it is necessary to substitute all switch states into the value function and scroll to optimize and screen out the minimum value, the online computation of the prior art is relatively large; due to the direct output method of the existing control technology, the There is only one switch state output in each control cycle, and the switching frequency is not fixed; after technical improvement, the error between the output value of the prediction model and the actual output value of the controlled object is reduced, but because the feedback link is a linear control link, there are Inherent hysteresis, limited error elimination ability; due to sampling delay and algorithm execution time delay, the output command voltage of the control system lags behind the actual command voltage

Images