Novel auto-coupling PI cooperative control theory method

Abstract

In order to solve problems of poor gain robustness and poor anti-disturbance robustness of the PI and various improved PIs, the invention relates to a novel auto-coupling PI (ACPI) cooperative controltheory method independent of a controlled object model. According to the method, all unknown internal and external complex factors of a system are defined as total disturbance and thus an unknown nonlinear complex system is mapped into an equivalent unknown linear system, so that a controlled error system under total disturbance reverse excitation is constructed; according to dynamic characteristics of an unknown controlled object, testing is performed to obtain a value range of the transition process time and a minimum speed factor model and a self-adaptive speed factor model are designed according to the value range. The global robust stability of the ACPI cooperative control system is theoretically analyzed. According to the invention, a scientific theoretical basis and technical guarantee are provided for technical evaluation and technical upgrading of various PI controllers in existing operation; and the method has a great application value in the field of control over a first-order unknown nonlinear system and a multiple-input-multiple-output nonlinear system.

Description

technical field [0001] The invention relates to the control of a first-order unknown system or a multiple-input multiple-output (MIMO) unknown system, in particular to an auto-coupling PI (Auto-Coupling PI, ACPI) cooperative control. Background technique [0002] The PI controller is the earliest classic controller in the PID control family, which is mainly used in the control of first-order linear or nonlinear systems and multiple-input and multiple-output systems. For the known first-order linear system, the tuning of PI can be realized by using the phase margin tuning rule, and its proportional gain k p and integral gain k i Both depend on the characteristic angular frequency of the controlled object. Facing the varied controlled objects, the two gains of the PI controller are also very different, indicating that the two gains of the PI controller are closely related to the controlled object; for nonlinear complex objects In other words, the phase margin tuning rules ca...

AI Technical Summary

Problems solved by technology

[0015] ① Proportional gain lacks physical properties

[0017] ②PI control force only has generalized displacement dimension

[0027] ① There is a dimensional mismatch between the PI control force and the control input of the controlled object.

However, the control force input u of any first-order system requires a generalized "velocity" dimension. Therefore, if a PI controller is used to control any first-order system, it will cause the PI control force u and the controlled object control force input u There is a principle error of dimensional mismatch, or in other words, using low-order PI control force to control the controlled object with high-order dimensional control input will make it difficult for the PI control ability to exert a good control effect;

[0029] ②Uncoordinated control mechanism reduces dynamic quality and steady-state performance

This uncoordinated control mechanism makes it difficult for the PI control system to obtain good dynamic quality and steady-state performance, and it is also true

[0031] In summary, if k p is an independent variable with no physical properties, T i is an independent time variable, not only will there be a principle error in the mismatch of the control force class, which will reduce the control ability of PI, but also will make the two different physical links of PI show an uncoordinated control mechanism in the control process, so it is difficult to use PI control system obtains good dynamic quality and steady-state performance

[0032] In addition, the uncoordinated control mechanism can only ensure that the PI control system has a local transient state. Once there is a sudden change in the expected trajectory, or a sudden change in the working condition, or when the model parameters change with time or there is an external disturbance, the PI gain must be readjusted. This is the root cause of poor PI gain robustness, poor time-varying robustness, and poor anti-disturbance robustness

Images