Self-adapting state feedback forecasting control method based on noise computing

Abstract

The invention discloses a feedback prediction controlling method of self-adaptation condition based on noise estimation, which comprises the following steps: A. utilizing Singular Pencil Model to represent real controlled course; B. adopting recurrence expanded Kalman filter parameter estimation method to estimate the real controlled condition and model reference through Singular Pencil Model; C. utilizing the estimated condition and model parameter to calculate the present control action; adding the calculated control action into the controlled course; D. repeating the step B and C in the next executing period; realizing the feedback prediction controlling method of self-adaptation condition in the real controlled course. The invention improves the noise inhibiting ability through controlling method greatly, which reduces the calculating quantity for industrial application conveniently.

Description

technical field [0001] The invention relates to the technical field of automatic control, in particular to an adaptive state feedback predictive control method based on noise estimation. Background technique [0002] In the field of industrial automation control, the application of predictive control technology to control the controlled process has a wide range of applications. [0003] Most of the existing applied predictive control technologies are based on offline identification models, or models derived from physical and chemical reaction mechanisms, such as model algorithm control (Model Algorithm Control), dynamic matrix control (Dynamic Matrix Control) and generalized predictive control ( General Predictive Control), etc. In addition, existing applied predictive control technologies also include adaptive predictive control methods that combine online identification with the above-mentioned predictive control, such as adaptive generalized predictive control, model ref...

AI Technical Summary

Problems solved by technology

[0004] In the above-mentioned various existing application predictive control technologies, due to the difficulty of determining the noise information, especially the unmeasurable noise information contained in the industrial controlled process, the existing control methods cannot fully consider the noise contained in the controlled process information

Even if some existing control methods consider the noise information contained in the controlled process, most of the considered noise information is measurable noise information, while in the industry the noise information contained in the controlled process is mostly unmeasurable noise information, so the existing The control method does not fully consider the unmeasurable noise information contained in the controlled process

[0005] In addition, since most of the applied predictive control techniques do not introduce state feedback, the existing control methods do not fully consider the state information contained in the controlled process.

Although individual existing methods introduce state feedback, such as state feedback predictive control, they are limited by whether the state is measurable. If the state is measurable, the measured state information can be directly used for state feedback predictive control; if If the state is unmeasurable, a state observer must be designed separately, and the state information can be observed by using the designed state observer, and then the state feedback predictive control can be performed by using the measured state information. The implementation process is very complicated and the implementation cost is also very high. high

[0006] Therefore, based on the above two reasons, that is, the existing control method does not fully consider the noise information and state information contained in the controlled process, resulting in the poor ability of the existing control method to suppress noise, which is the reason why the current application of predictive control technology Important issues that need to be resolved urgently

Images