Matlab automatic control-Adrc active disturbance rejection control parameter adjustment
See the previous article: https://blog.csdn.net/qingfengxd1/article/details/111830762
Take the simplest linear combination method (1) as an example, probably the following parameters need to be adjusted:
TD: delta h
ESO: B01, B02, B03 and observer bandwidth w0
Non-linear feedback: (beta1, beta2) replace with kp and kd , Alpha
For TD, the general simulation model delta can be as large as possible, basically the same in the range of 100~500, even if the effect is large, there will be no big improvement. h is the simulation step length in the simulation model.
ESO three parameters and observations about bandwidth, set in sequence. 3 w0,3 W0 ^ 2, ^ W0. 3 to meet the requirements.
So there are only four parameters that need to be adjusted in the end: kp kd w0 alpha. Now you can control the variables.
The basic rule is: the
smaller the alpha, the shorter the adjustment time, but too small will cause oscillation.
The smaller the w0, the longer the adjustment time, and the smaller the oscillation amplitude.
The greater the Kp, the shorter the adjustment time and the greater the oscillation.
The kd effect is not obvious, and can be fine-tuned after stabilization.
The experience is:
1. Make sure that there are no errors in the ADRC modeling process
2. Make sure that the amplitude of the input test signal is reasonable for your controlled object
3. Slowly adjust parameters
