Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Parameter identification of nonlinear system based on improved differential evolution algorithm

A nonlinear system and evolutionary algorithm technology, applied in the field of thermal automation, can solve problems such as algorithm stagnation, and achieve the effects of improved accuracy, easy understanding and implementation, and good adaptability

Pending Publication Date: 2019-01-29
SOUTHEAST UNIV
View PDF0 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In order to solve the shortcomings of the differential evolution algorithm, which tends to converge to local optimum or algorithm stagnation with the increase of the evolution algebra, the present invention provides a nonlinear system parameter identification method based on the improved differential evolution algorithm, which dynamically adjusts the scaling factor and crossover rate Methods combined to propose new improved methods

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Parameter identification of nonlinear system based on improved differential evolution algorithm
  • Parameter identification of nonlinear system based on improved differential evolution algorithm
  • Parameter identification of nonlinear system based on improved differential evolution algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0062] In order to verify the effectiveness of the algorithm, several benchmark functions are used for testing. The test functions are Schwefel, Ackley, Rastrigin, Alpine, and Griewank functions, and the optimal solutions are all 0. The number of populations is set to be 10 times the dimension, and other parameters are set as follows: basic DE: F=0.6, CR=0.9, parameters of DEAFCR are: Fmax=1, Fmin=0.2, CRmax=1, CRmin=0.8. Table 1 shows the optimization results of these two algorithms for the 30-dimensional test function, each function is run 30 times, and the maximum number of iterations is 1000 times.

[0063] Table 1 Algorithm test results

[0064]

[0065] It can be seen from the test data in Table 1 that while maintaining the stability of the algorithm, the DEAFCR algorithm is smaller than the optimal solution and the average value of the DE algorithm. Therefore, the optimal solution and convergence accuracy searched by the DEAFCR algorithm are better than the basic DE...

Embodiment 2

[0067] In order to test the identification effect of the improved differential evolution algorithm in nonlinear fixed-length systems, this patent selects the transfer function model and the Hammerstein model for identification and simulation research. The results are compared with the genetic (GA) algorithm and the adaptive mutation rate and dynamic crossover mutation rate evolutionary difference algorithm (MCDE) to verify the effectiveness of the algorithm.

[0068] The transfer function model of the second-order inertia link plus pure time delay is shown in formula (8).

[0069]

[0070] The parameters to be identified are gain coefficient k, time constant T 1 , T 2 and delay time τ. The search range of parameters is 0≤k≤4,0≤T 1 ≤4,0≤T 2 ≤4,0≤τ≤2, the sampling period is 0.1s, the sampling data is 100, the input is random input u, the sampling time is 0≤t≤9.9, and the output signal y is calculated by the lsim function, y=lism(G, u,t).

[0071] The parameters in the s...

Embodiment 3

[0076] The Hammerstein model is composed of a memoryless nonlinear gain link and a linear subsystem in series. The model under the interference of colored noise can be described as:

[0077] A(q -1 )y(k)=B(q -1 )x(k)+C(q -1 )w(k) (9)

[0078] Considering the Hammerstein model in formula (9), the noise w(k) is Gaussian white noise with mean value 0 and variance 0.01. The input u(k) is Gaussian white noise with a mean of 0 and a variance of 0.2. Take: A(q -1 )=1-1.5q -1 +0.7q -2 , B(q -1 )=q -1 +0.5q -2 , C(q -1 )=1+1.5q -1 , x(k)=u(k)+0.5u 2 (k)+0.3u 3 (k)+0.1u 4 (k).

[0079] Therefore, the model has 8 parameters to be identified, a 1 、a 2 , b 1 , b 2 、c 1 、r 2 、r 3 、r 4 , the dimension is 8, the population size NP is 80, the number of iterations is 1000, the parameter search range is [-2,2], and the other parameters are set as above. The simulation results are shown in Table 3.

[0080] Table 3 Comparison of Hammerstein model parameter identification ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

Aiming at the difficulty and inaccuracy of parameter identification of nonlinear model, this invention proposes an identification algorithm based on improved differential evolution algorithm. By establishing a lifetime mechanism, the scaling factor and crossover rate can be adjusted dynamically according to the lifetime value, so as to avoid premature convergence at the beginning of the algorithmand retain high-quality solution at the later stage, so as to accelerate the convergence rate. In order to verify the performance and practicability of the improved algorithm, the typical test functions are used to compare and test, and the nonlinear transfer function model and Hammerstein model are identified. The experimental results show that the improved algorithm has fast convergence speed and high identification accuracy, and it is effective and feasible for parameter identification of nonlinear system.

Description

technical field [0001] The invention relates to the technical field of thermal automation, in particular to a nonlinear system parameter identification method based on an improved differential evolution algorithm. Background technique [0002] In the industrial field, the research object is usually very complex, the internal mechanism is vague, it is difficult to have a theory to directly obtain the corresponding mathematical model, and most of them are nonlinear models, so it is proposed to use the known observation data to identify the mathematical model of the research object. model and its parameters. With the emergence of intelligent optimization algorithms, nonlinear system model parameters can be identified using algorithms such as neural networks, genetic algorithms, and particle swarm optimization algorithms. Differential Evolution (DE) algorithm is an optimization algorithm that uses floating-point vector coding to perform random search in continuous space, which ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F17/50G06N3/00
CPCG06F30/20G06N3/006
Inventor 张雨飞段崇崇
Owner SOUTHEAST UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com