Received: 24 April 2019 Revised: 26 November 2019 Accepted: 10 February 2020
DOI: 10.1002/nme.6338
RESEARCH ARTICLE
Finite element model updating using crow search
algorithm with Levy flight
Hao Wu
1,2
Pingbo Wu
1
Kai Xu
2,3
Fansong Li
1
1
State Key Laboratory of Traction Power,
Southwest Jiaotong University, Chengdu,
P.R. China
2
Mechanical Science and Engineering,
University of Illinois at
Urbana-Champaign, Illinois,
3
School of Mechanical Engineering,
Southwest Jiaotong University, Chengdu,
P.R. China
Correspondence
Fansong Li, State Key Laboratory of
Traction Power, Southwest Jiaotong
University, Chengdu, P.R. China.
Email: lifansong2013@163.com
Funding information
Independent Research and Development
Project of the State Key Laboratory of
Traction Power, 2020TPL-T09; National
Natural Science Foundation of China,
11790282
Summary
The accuracy of the finite element model is essential for fatigue analysis and
vibration of a flexible body. This paper presents a finite element model updat-
ing (FEMU) algorithm based on crows search algorithm incorporated with Levy
flight (LFCSA). The proposed algorithm is tested on the updating of two differ-
ent cases: a simple structure (beam) and a complex structure (gearbox housing).
To verify the performance of LFCSA, two optimization algorithm, which is the
standard CSA and the particle swarm optimization algorithm with Levy flight
(LFPSO), are applied to FEMU of the simple beam. The result of the comparison
shows that the LFCSA has sufficient convergence speed and global search abil-
ity in the implementation of FEMU. In the case of gearbox housing, the error of
frequencies and mode shapes are analyzed to indicate that the LFCSA gives a
satisfactory result in a complex structure.
KEYWORDS
crow search algorithm, finite element model updating, gearbox housing, Levy flight
1 INTRODUCTION
Finite element (FE) model, which usually gives results that are not the same as the results presented by an experiment, is
a numerical method used to provide approximate solutions in engineering.
1,2
The inaccuracy of the FE model is generally
caused by structure error, order error, and parameter error. Finite element model updating (FEMU) is a procedure aimed at
minimizing the discrepancy between the numerical and the experiment results to ensure that the FE model better reflects
the physical properties of the structure. FEMU is fundamentally a constrained optimization approach, and the objective
function of the optimization problem is generally an equation that characterizes the error between experimental and
numerical quantities, such as the structure's natural frequencies and mode shapes. FEMU can be performed by manual or
automatic. Altunisik
3
used manual model updating procedure to update a long span reinforced concrete highway bridge
model based on the ambient vibration.
A lot of researches
4-7
have been done on FEMU and various computational methods, which include Bayesian
theory
8-10
and particle search algorithm, have been proposed. Ren et al
11
presented a response surface-based method
for FEMU based on the measured static response of structures, which has the advantages of easy implementation, high
cost-efficiency, and sufficient updating accuracy. Boulkaibet et al
12
applied the innovative volitive operator of fish school
search which is incorporated into the particle swarm optimization (PSO) algorithm to FEMU. Marwala
13
systematically
studied the application of intelligent algorithms in FEMU. The algorithms considered include the genetic algorithm,
2916 © 2020 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/nme Int J Numer Methods Eng. 2020;121:2916–2928.