Pi Control plant cannot be linearized

Question

Muhamed Ehab 2024-10-24

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2161840-pi-control-plant-cannot-be-linearized

编辑： MULI 2024-12-5

I try to tune my PI controller but I always get the error message "Subscript no. 2 is out of range" and "Plant cannot be linearized"

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Sumukh 2024-10-30

Can you please share the Simulink model for better understanding the issue?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

MULI 2024-12-4

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2161840-pi-control-plant-cannot-be-linearized#answer_1553408

编辑：MULI 2024-12-5

Hi @Muhamed Ehab

The errors "Subscript no. 2 is out of range" and "Plant cannot be linearized" occur because your model includes switching components (e.g., the Universal Bridge and PWM signals), which can introduce discontinuities that prevent successful linearization.

You can follow the three suggestions below to resolve the issue:

Identify the Plant Using System Identification:

Since direct linearization fails, you can use system identification with a perturbation signal (e.g., step or sine wave) to estimate the plant dynamics.
You can refer to the below link for detailed implementation related to this: https://www.mathworks.com/help/slcontrol/ug/design-controller-for-power-electronics-model-using-simulated-io-data.html

Frequency Response Estimation:

Apply a sinusoidal signal to the system and measure the frequency response.
Use the frequency response data to identify the plant model.
You can refer to this link more detailed implementation of frequency response estimation: https://www.mathworks.com/help/slcontrol/ug/design-controller-for-power-electronics-model-using-frequency-response-data.html

Simplify the Model for Linearization:

Replace the switching components (e.g., the Universal Bridge) with averaged switching models or linear approximations.
This will smooth out discontinuities and allow the model to be linearized.
You can refer to the below link for any further information on linearisation of electrical components: https://www.mathworks.com/help/sps/linearization_techniques.html

After identifying the plant model, load it into the PID Tuner App and then use the app to design and tune the PID controller for desired performance. You can refer to below link for more information on this:

https://www.mathworks.com/help/control/ref/pidtuner-app.html

You may refer to the below answer where similar query is addressed

https://www.mathworks.com/matlabcentral/answers/2161840-pi-control-plant-cannot-be-linearized