Method for identifying system containing saturation nonlinearity

Abstract

The invention provides a method for identifying a nonlinear system with saturation, and mainly provides a solution to the problem that conventional system identification methods cannot be directly used for identification of a nonlinear system with saturation. The invention designs the input signal from two aspects of frequency and amplitude to fully excite the system. A method of adding window and scanning to process each state variable of the system is proposed. By screening the variance and average value of each window data, the saturation limit of the system can be identified. According to the obtained saturation limit, the input and output data of the system are screened, and the system model parameters are obtained by using the relevant system identification method. The invention has the advantages of designing the input signal to fully excite the system for a saturated nonlinear system, correctly and effectively identifying the system model parameters, and simultaneously obtaining the system saturation limit value, which has better engineering practicability.

Description

technical field [0001] The invention belongs to the field of system identification, and in particular relates to a method for identifying a saturated nonlinear system. Background technique [0002] In practical applications, most control systems are affected by factors such as limited input energy and limited output amplitude, and usually have saturated nonlinear characteristics. In this case, if the conventional system identification method is directly adopted, the following problems will mainly exist: [0003] First, when the system is saturated, the output response is affected by the saturation characteristics of the system and cannot fully reflect its true closed-loop characteristics. At this time, if the input command data and output response data of the system are still used for conventional identification, the correct system model parameters will not be obtained. [0004] Second, the conventional system identification method cannot obtain the saturation limit, and t...

AI Technical Summary

Problems solved by technology

[0003] First, when the system is saturated, the output response is affected by the saturation characteristics of the system and cannot fully reflect its true closed-loop characteristics

At this time, if the system’s input command data and output response data are still used directly for routine identification, the correct system model parameters will not be obtained

[0004] Second, the conventional system identification method cannot obtain the saturation limit, and the acquisition of the saturation limit is an important prerequisite for correctly identifying the model parameters of the nonlinear system with saturation

[0005] Third, the traditional frequency sweep excitation only changes the frequency of the input signal but has no requirement on the amplitude. When the system has saturated nonlinearity, it is difficult for the traditional frequency sweep excitation to ensure that the system is fully excited.

Images