See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279220714

Ride comfort optimization of a multi-axle heavy motorized wheel dump truck

based on virtual and real prototype experiment integrated Kriging model

Article  in  Advances in Mechanical Engineering · June 2015

DOI: 10.1177/1687814015584257

CITATIONS READS

7 1,685

4 authors, including:

Sanbao hu Lin Xu
Wuhan University of Technology Wuhan University of Technology
10 PUBLICATIONS   133 CITATIONS    49 PUBLICATIONS   1,145 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The theoretical and experimental study of a parallel hydraulic-electric energy-harvesting suspension system View project

Energy Harvesting View project

All content following this page was uploaded by Lin Xu on 22 December 2018.

The user has requested enhancement of the downloaded file.

Research Article

Advances in Mechanical Engineering

2015, Vol. 7(6) 1–15
Ó The Author(s) 2015
Ride comfort optimization of a DOI: 10.1177/1687814015584257
aime.sagepub.com
multi-axle heavy motorized wheel
dump truck based on virtual and real
prototype experiment integrated
Kriging model

Bian Gong1, Xuexun Guo1, Sanbao Hu1 and Lin Xu2

Abstract
The optimization of hydro-pneumatic suspension parameters of a multi-axle heavy motorized wheel dump truck is car-
ried out based on virtual and real prototype experiment integrated Kriging model in this article. The root mean square
of vertical vibration acceleration, in the center of sprung mass, is assigned as the optimization objective. The constraints
are the natural frequency, the working stroke, and the dynamic load of wheels. The suspension structure for the truck is
the adjustable hydro-pneumatic suspension with ideal vehicle nonlinear characteristics, integrated with elastic and damp-
ing elements. Also, the hydraulic systems of two adjacent hydro-pneumatic suspension are interconnected. Considering
the high complexity of the engineering model, a novel kind of meta-model called virtual and real prototype experiment
integrated Kriging is proposed in this article. The interpolation principle and the construction of virtual and real proto-
type experiment integrated Kriging model were elucidated. Being different from traditional Kriging, virtual and real
prototype experiment integrated Kriging combines the respective advantages of actual test and Computer Aided
Engineering simulation. Based on the virtual and real prototype experiment integrated Kriging model, the optimization
results, obtained by experimental verification, showed significant improvement in the ride comfort by 12.48% for front
suspension and 11.79% for rear suspension. Compared with traditional Kriging, the optimization effect was improved by
3.05% and 3.38% respectively. Virtual and real prototype experiment integrated Kriging provides an effective way to
approach the optimal solution for the optimization of high-complexity engineering problems.

Keywords
Ride comfort, optimization, the multi-axle heavy motorized wheel dump truck, virtual and real prototype experiment
integrated Kriging

Date received: 15 December 2014; accepted: 17 March 2015

Academic Editor: Jia-Jang Wu 1

Introduction
The ride comfort of a vehicle is significantly influenced
by its suspension system, the road surface roughness,
and the vehicle speed. Many research works have been
done to optimize the suspension to improve the ride
comfort. Jayachandran and Krishnapillai1 optimized
the values of stiffness and damping coefficient for a
School of Automotive Engineering, Components Key Laboratory of
Advanced Technology of Automotive Components, Hubei Collaborative
Innovation Center for Automotive Components Technology, Wuhan
University of Technology, Wuhan, China
2
Department of Mechanical Engineering, Virginia Polytechnic Institute and
State University, Blacksburg, VA, USA
Corresponding author:
Sanbao Hu, School of Automotive Engineering, Components Key
Laboratory of Advanced Technology of Automotive Components, Hubei
Collaborative Innovation Center for Automotive Components
Technology, Wuhan University of Technology, Wuhan 430070, China.
Email: husanbao@whut.edu.cn

Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 3.0 License
(http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (http://www.uk.sagepub.com/aboutus/
openaccess.htm).
Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015
2 Advances in Mechanical Engineering
passive suspension system and introduced a semi-active
control system to improve the ride comfort for a pas-
senger car. Kaldas et al.2 proposed a new optimization
methodology for optimizing the damper top mount
characteristics to improve the vehicle ride comfort.
Eriksson and Friberg3 calculated the response of the
bus by using a finite element model describing the com-
plete vehicle. The body of the bus was condensed using
a substructuring technique to reduce the optimization
time. Finally, the optimized stiffness and related damp-
ing of each power unit rubber insulator were found to
achieve the optimization of ride comfort. Yang et al.4
proposed a systematic methodology to evaluate and
improve the ride comfort. In the study, a new index
called annoyance rate was presented to indicate the
quantitative correlation between objective method and
subjective comment. But the precondition of further
investigation and optimization is an accurate model.
He et al.5 established a four-axle heavy truck and road
coupling system and carried out a multi-objective opti-
mization based on modified skyhook damping control
to improve the ride comfort, while the traditional opti-
mization method could not reduce the computational
burden.
Even in the field of construction machinery such as
earth moving equipment, ride comfort has also become
an important competitive parameter for manufacturers
of mobile machines. It means the challenge to improve
the ride comfort of the machinery, namely, the compo-
nents for suspension system should be chosen properly.
Often the design would be confined by commercially
available components.6 Evaluation and improvement of
ride comfort are made efficient by model-based proto-
typing.7 In order to find the proper parameters, most
engineers carried out the optimization based on the
computer model8,9 because it is possible to change para-
meters and evaluate the performance in a cost-efficient
way compared to changing and testing a full-scale pro-
totype.10 Just as in the research of Yong Yang et al.,4
the high-precision simulation model is needed.
In the practical complex engineering application, the
high precision of the model more likely means it will
cost more time to make the optimization. Therefore,
the computational expense would be increased. In order
to alleviate the computational burden of engineering
analyses, the Kriging surrogate model has been used
for a variety of applications, especially in the design of
complex mechanical system and aerospace engineering.
Kriging was inspired by the work of Danie G Krige,
a South African geologist, for mining estimation, but
he never formalized the method. Then, Georges
Matheron11 developed the theory of regionalized vari-
ables based on the work done by Krige and called the
method Kriging in his honor. Noel Cressie12 summar-
ized the historical origins of Kriging in detail. The stan-
dard approach, first applied to deterministic computer
Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015
simulations by Sacks et al.13 called ordinary Kriging
(OK), is a variant of Design and Analysis of Computer
Experiments (DACE). Since then, it has been widely
used in approximate computation models. Along with
the development of Computer Aided Engineering
(CAE) technique, the modeling and optimization of
Kriging based on CAE analysis have offered perfect
performance. Martin and Simpson14,15 conducted a
study on using Kriging models to approximate determi-
nistic computer models and discussed the applicability
of various Kriging variants. Duchaine et al.16 intro-
duced the Kriging meta-model with a design of experi-
ments (DOE) approach to limit the impact of many
evaluations of flow computations which are CPU- and
time-consuming. The application of this method was
proved to be feasible with available computing power
and yields manageable response time. Based on the
simulated samples, Lim et al.17 developed a surrogate
model using Kriging for the fast computation of hybrid
sandwich panel in the design optimization. Li et al.18
proposed an approach Kriging-assisted multi-objective
genetic algorithm (K-MOGA) for multi-objective
design optimization. Kriging-based meta-modeling was
embedded within a MOGA. Examples were used to
illustrate that K-MOGA reduced the number of simula-
tion calls by evaluating some individuals in the popula-
tion by Kriging meta-models instead of the simulation.
In the work of Shinkyu Jeong et al.,19 the genetic algo-
rithm using the Kriging model for objective function
evaluation was introduced. A criterion ‘‘expected
improvement (EI)’’ was used for the selection of the
additional sample point. This made it possible not only
to improve the accuracy of the response surface but
also to explore the global optimum efficiently. Yung-
Chang Cheng and Cheng-Kang Lee20 integrated the
uniform design and Kriging interpolation, took the crit-
ical hunting speed of the vehicle as the optimization
function, and created an input–output segregate model
based on the Kriging interpolation method to optimize
the suspension parameters of a high speed railway vehi-
cle, and obtained an ideal result. Wang Hao et al.21 put
forward a modified sequential Kriging optimization to
change the twice optimization problem into once by
adding more than one point at the same time. In their
work, the proposed method could approach the global
optimization quickly and accurately in the application
of active suspension.
However, in some particularly complex project, espe-
cially with strong nonlinearity, the order of Kriging sur-
rogate model obtained from the pure CAE analysis
could be lowered. In that case, it would lead to the loss
of the extremum or the optimal solution.
In this article, the hydro-pneumatic suspension
design and optimization for a multi-axle heavy motor-
ized wheel dump truck (Figure 1) is presented based on
approximate model.
Gong et al.
Figure 1. The multi-axle heavy motorized wheel dump truck.
To find the proper design, the actual test results are
introduced to modify the meta-model to remedy the
above-mentioned shortage. Hence, a novel meta-model
called virtual and real prototype experiment integrated
Kriging (VRPEI-Kriging) is proposed to perform the
ride comfort optimization of the multi-axle heavy
motorized wheel dump truck.
Modeling
The vehicle dynamic model
With the continuous development of national economy
construction, heavy motorized wheel dump truck has
been used in national road construction, mining and
other kinds of engineering, and construction of essen-
tial transport infrastructure. The cost of the design
research and development is huge, and the test cost also
accounts for a large proportion in the process of design
optimization. To make the design and optimization
convenient, the dynamic model is built up in modules
using batch files and the model can easily be parame-
terized, as shown in Figure 2.
The suspension structure for the vehicle is adjustable
hydro-pneumatic suspension with ideal vehicle non-
linear characteristics, which integrated elastic element
and damping element. It can improve not only the ride
comfort of the vehicle but also a comparatively perfect
regulating function. For instance, no matter in no load
or full load, the suspension can keep the vehicle at a
constant height from the ground and then make the
suspension stiffness adapt to the changes of no load or
full load. In addition to these, the adjustable hydro-
pneumatic suspension can realize the switch of elastic
suspension to adjust the attitude of the vehicle, such as
leaning forward and backward, rising and falling, and
so on.
There are eight axles in the dump truck. Each axle
with the same structure is equipped with a hydro-
Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015
3
Figure 2. Dynamic model of dump truck.
Figure 3. The vehicle plane diagram.
pneumatic spring. Because the ride comfort of the vehi-
cle is mainly considered here, the steering system is not
taken into account. In the suspension system, the rub-
ber bushing between connectors is given approximately
limited to the lack of model parameters. The tire
response is crucial to evaluate ride comfort realistically
on off-road conditions. A tire model developed for an
off-road vehicle with big tires is applied in the dynamic
model. The tire model needs only few modeling para-
meters to handle short-wave terrain with obstacles.22
The characteristics of the hydraulic system play a
pivotal role in the vehicle ride comfort. The hydraulic
system of two adjacent hydro-pneumatic suspension is
interconnected. In the half vehicle model, as shown in
Figure 3, the pressure of the same hydraulic system in I
and II is always the same, so does III and IV, as shown
in Figure 4. This structure helps keep the traveling vehi-
cle’s body balanced and avoid inhomogeneous stress.
For the vehicle suspension structure, the correspond-
ing simulation model of the hydro-pneumatic spring
was established in AMESim, as shown in Figure 5.
From the hydraulic simulation model, the varia-
tional piston speed was obtained and force changed
4 Advances in Mechanical Engineering
Figure 4. Hydro-pneumatic spring structure diagram.
Figure 5. The simulation model of hydro-pneumatic spring.
with the piston displacement. Then, the hydro-
pneumatic spring characteristic was introduced into
ADAMS to get the complete CAE simulation model,
as shown in Figure 6.
The working environment of the mine car is bad and
the road traffic is poor. Considering the large mass of
the vehicle, the mine road was constructed referring to
the standard three-dimensional (3D) random road and
maintained regularly. The 3D pavement model was
built based on sinusoidal superposition principle, as
shown in Figure 7.
Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015
Design variables and constraints
Design variables. For hydro-pneumatic suspension, the
initial volume and initial pressure of accumulator, and
the diameter of damping orifice are the main factors
that affect the stiffness and damping of hydro-
pneumatic spring. These parameters were adjusted to
achieve the adjustable attitude of the vehicle. For the
convenience of adjusting the parameters of hydro-
pneumatic spring in actual experiments, based on the
known hydraulic cylinder structure (Figure 4) and the
mass distribution coefficient, select the initial pressure
P0 and initial volume V0 of accumulator, and the dia-
meter of damping orifice d0 in front hydro-pneumatic
spring and rear hydro-pneumatic spring as the design
variables to be optimized. The range of variables value
is shown in Table 1.
According to the requirement of the mine car and
characteristics of the structure, the design variable val-
ues should be constrained by the offset frequency of the
front and rear suspension, the working stroke of the
hydraulic cylinder, and the dynamic load of the wheel.
Offset frequency. The sprung mass is very large in full
load, so the front and rear suspension offset frequency
is relatively low. But low offset frequency may reduce
the operation stability of the vehicle; therefore, the off-
set frequency of front and rear suspension should be
constrained. Offset frequency is defined as
Gong et al. 5

Figure 6. The simulation flow chart in modeFRONTIER.

Figure 7. The 3D pavement model.

Table 1. Range of variables value.

Initial Minimum Maximum

Front hydro-pneumatic spring Initial pressure of accumulator P0_f (Pa) 5.5 3 106 3 3 106 1 3 107
Volume of accumulator V0_f (L) 5 3 10
Damping orifice d0_f (m) 1.6 3 1022 1.2 3 1022 3.5 3 1022
Rear hydro-pneumatic spring Initial pressure of accumulator P0_r (Pa) 1 3 107 5 3 106 1.5 3 107
Volume of accumulator V0_r (L) 4 3 10
Damping orifice d0_r (m) 9 3 1023 6 3 1023 1.2 3 1022

rffiffiffiffiffiffi
1 K Working stroke. Hydro-pneumatic spring is installed
n= ð1Þ
2p Mb between the frame and the axle, so the working stroke
of hydro-pneumatic spring, from no load to full load,
where n stands for the suspension offset frequency, K is subjected to the conditions of the vehicle structure.
for the suspension stiffness, and Mb for the suspension Thus, it is necessary to constrain the dynamic
sprung mass.

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

6 Advances in Mechanical Engineering

displacement of the suspension. The working stroke is In all, the optimization model can be obtained
defined as
find X = ½P0 f , V 0 f , d0 f , P0 r, V 0 r, d0 r
(Pm  Pk )Vk minimize F(X) = J
8
Skm = ð2Þ
Pm A1 >
> Xmin  X  Xmax
>
>
>
where Skm stands for the amount of compression for < nf min  nf  nf max
>
hydro-pneumatic spring from no load to full load, Pm subject to nr min  nr  nr max
>
>
for the pressure of the gas chamber in full load, Pk for >
> Skm  300
>
>
the pressure of the gas chamber in no load, Vk for the :
jf  13 , jr  13
volume of the gas chamber in no load, and A1 for the
cross-sectional area of the piston rod. where Xmin and Xmax represent the minimum and maxi-
mum values of design variables, respectively; nf and nr
Dynamic load of wheels. According to the characteristics
stand for the offset frequency of front and rear suspen-
of hydro-pneumatic spring’s variable stiffness, the stiff-
sion, respectively; and jf and jr for the pre and post
ness of the suspension is larger when the vehicle is at
tires’ relative dynamic load, respectively.
full load. Therefore, the working stroke and the
dynamic load of the wheel are smaller than those of
other engineering vehicles. The dynamic load of the VRPEI-Kriging meta-model
wheel impacts the adhesion between the wheel and the
The VRPEI-Kriging model is based on the VRPEI-
road, associated with the grip of the tires, and plays a
Kriging interpolation as the core. In the case of virtual
crucial role in handling stability and driving safety of
simulation results, it combines the actual test results to
the vehicle. Thus, the dynamic load of front and rear
construct the approximate meta-model. The main dif-
wheels should be constrained in the optimization. The
ference with the traditional Kriging model lies in that
dynamic load coefficient is defined as
VRPEI-Kriging considers the analysis results with high
 12 reliability from the actual tests, which have cost thou-
ÐT
1
Fd2 dt sands of dollars and countless hours.
T
0
j= ð3Þ
G
Principle of VRPEI-Kriging interpolation
where j stands for the dynamic load coefficient, T for
Take x1 , x2 , x3 , . . . , xn 2 O as the known points from
the analysis of vibration time, Fd for the dynamic loads
the computer, where O is the domain of the computer
of the wheels, and G for the static load of the wheels.
model. The corresponding responses are
Z = ½Z(x1 ), Z(x2 ), . . . , Z(xn ). And m points from the
Optimization objective. In the research of vehicle ride actual test model be t1 , t2 , . . . , tm . The resulting outputs
comfort, the sprung mass’ centroid of the whole vehicle are T = ½T (t1 ), T (t2 ), . . . , T(tm ). Note that xi 6¼ tk .
model includes three directions of vibration accelera- Mingling Z with T as the final response domain, con-
tion—vertical, longitudinal, and lateral. In this article, sider a linear predictor of the output; thus, the predic-
the root mean square (RMS) value of vertical vibration tor is
acceleration in the center of the sprung mass is taken as
the objective function J . The expression is X
n X
m
Z * (x) = ai Z(xi ) + bk T (tk ) ð6Þ
2 80 312 i=1 k =1
ð
J = aw = 4 W 2 (f )Ga (f )df 5 ð4Þ ai and bk as the weight coefficient, respectively,
describe the contribution degree of sample value and
0:5
experimental value on the estimate value. In statistics
where aw stands for the RMS value of vertical vibration
acceleration in the center of the sprung mass, Ga (f ) for E½Z(xi ) = E½T (tk ) = u ð7Þ
the acceleration power spectral density function, and
In accordance with the unbiased estimation
W (f ) for the frequency weighting function which can
be expressed as " #
  X
n X
m
*
8 E Z (x) = E ai Z(xi ) + bk T (tk ) = u ð8Þ
>
> 0:5 (0:5\f \2) i=1 k=1
<
f =4 (2\f \4)
W (f ) = ð5Þ X
n X
m
>
> 1 (4\f \12:5) ai + bk = 1 ð9Þ
:
12:5=f (12:5\f \80) i=1 k=1

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

Gong et al. 7

0 1
Considering the proportion of the number of test g(t1 , t1 ) g(t1 , t2 )  g(t1 , tm )
points, the environment of the test, and so on, the test B g(t , t ) g(t , t )    g(t2 , tm ) C
B 2 1 2 2 C
weight ratio Tw is introduced D=B
B .. .. .. .. C
C ,
@ . . . . A
X
n
g(tm , t1 ) g(tm , t2 )    g(tm , tm ) m 3 m
a i = 1  Tw ð10Þ 0 1 0 1
i=1 1 0 0 1
B C B C
X
m B1 0C B 0 1C
B
G=B . . C , H=B .. .. C
bk = Tw ð11Þ C B C
@ .. .. A @ . . A
k =1
1 0 n32 0 1 m32
The RMS of estimation error is
K1 = ½g(x1 , x), g(x2 , x), . . . , g(xn , x)T
 
s2E = D Z(x)  Z * (x) = D½Z(x)
K2 = ½g(t1 , x), g(t2 , x), . . . , g(tm , x)T
 
+ D Z * (x)  2C(Z(x), Z * (x))
K3 = ½1  Tw , Tw T
X
n X
n
= C(x, x) + ai aj C(xi , xj ) a = ½a1 , a2 , . . . , an T
i=1 j=1
b = ½b1 , b2 , . . . , bm T
X
m X
m
+ bk bl C(tk , tl ) ð12Þ
x = ½e, lT for the Lagrange factor.
k =1 l=1
The obtained minimum RMS is expressed as
X
n X
m
+ 2 ai bk C(xi , tk ) X
n X
m
i=1 k =1 s2E min = ai g(xi , x) + bk g(tk , x)
! i=1 k =1 ð14Þ
X
n X
m
+ (1  Tw )e + Tw  l  g(x, x)
2 ai C(xi , x) + bk C(tk , x)
i=1 k =1
In the VRPEI-Kriging interpolation model, the
So we want to get a = ½a1 , a2 , . . . , an T and introduction of test weight ratio Tw is the key to distin-
b = ½b1 , b2 , . . . , bm T to obtain the minimum s2E . guish from other interpolations, such as Simple Kriging
When the random function does not meet the sec- (SK), OK, Universal Kriging (UK), CoKriging (CK),
ond order smoothly, and satisfy the intrinsic hypoth- Bayesian Kriging (BK), indicator Kriging (IK), and so
esis, g(h) = C(0)  C(h). on. Tw, a complex variable, is determined by the pro-
Using the Lagrange method, we obtained that portion of experimental date, the condition of actual
test, the distribution of the test date, and so on.
0 10 1 0 1
A B G a K1 In the actual construction of the VRPEI-Kriging
B T CB C B C approximation model, the choice of experimental fac-
BB HC B C B C ð13Þ
@ D A@ b A = @ K 2 A tor depends on
GT H T
0 x K3
find Tw 2 (0, 1)
where s:t: min s2E min
0 1 From (13) and (14), it can be obtained that
g(x1 , x1 ) g(x1 , x2 )    g(x1 , xn )
B C
B g(x2 , x1 ) g(x2 , x2 )  g(x2 , xn ) C X
n X
m
B C s2E min = ai g(xi , x) + bk g(tk , x)
A=B
B .. .. .. ..
C
C
B . . . . C i=1 k =1
@ A + (1  Tw )e + Tw  l  g(x, x)
g(xn , x1 ) g(xn , x2 )  g(xn , xn ) n 3 n 0 1
K1
0 1   B C
g(x1 , t1 ) g(x1 , t2 )    g(x1 , tm ) = aT bT xT @ K2 A  g(x, x)
B g(x2 , t1 ) g(x2 , t2 )    g(x2 , tm ) C K3
B C 0 10 1
B=B .. .. .. .. C A B G a
@ . . . . A  
T B BT CB C
g(xn , t1 ) g(xn , t2 )    g(xn , tm ) n 3 m = aT bT x @ D H A@ b A  g(x, x)
GT HT 0 x
ð15Þ

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

8 Advances in Mechanical Engineering

The function is F = ½f(x1 ), f(x2 ), ., f(xn ), f(xn + 1 ), f(xn + 2 ), ., f(xn + m )T

0 10 1 ð23Þ
  A B G a
f = aT bT xT @ BT D H A@ b A ð16Þ The correlation matrix, R, is composed of the spatial
GT HT 0 x correlation function evaluated at each possible combi-
nation of known points
The optimum Tw is found to minimize f .

A B
R= ð24Þ
BT D
Construction of VRPEI-Kriging approximation model (n + m) 3 (n + m)

A VRPEI-Kriging model could be given in equation
(17). The model consists of two parts. The first part,
f(x)T m, is a simple linear regression of the data. It indi-
cates a global model of the design space. The second
part models the systematic lack-of-fit or deviations
from the linear model so that the whole model interpo-
lates the experimental points generated according to
one of the designs of experiment approach
T
^y(x) = f(x) m + e(x)
This matrix, R, by definition of the spatial correla-
tion function, g(  ), being positive semi-definite, is a
positive semi-definite matrix. It is also symmetric since
g(x1 , x2 ) = g(x2 , x1 ), and the diagonal consists of all
ones since g(x1 , x1 ) = 1. Set a vector that represents the
correlation between an unknown point and all known
points
r(x) = ½g(x1 , x), g(x2 , x), . . . , g(xn , x),
ð17Þ g(xn + 1 , x), g(xn + 2 , x), . . . , g(xn + m , x)T ð25Þ

The random process e(x) is assumed to be a If lT (x) solves this constrained minimization prob-
Gaussian stationary process with zero mean and lem, then lT (x)Y is called a best linear unbiased predic-
covariance tor for ^y(x). Solving lT (x) and substituting into (17), we
obtain
V (xi , xj ) = s2 g(xi , xj ) ð18Þ
^ + rT (x)R1 (Y  Fm)
^y(x) = fT (x)m ^ ð26Þ
The spatial correlation function, g(  ), affects the
smoothness of the model and differentiability of the where the regression coefficient m
^ is
response surface.
For the convenience of description, we named the ^ = (FT R1 F)1 FT R1 Y
m ð27Þ
known points from the computer model as
The selection of the correlation function is chosen by
x1 , x2 , x3 , . . . , xn and m points from the actual test
the user when generating the VRPEI-Kriging model. In
model as xn + 1 , xn + 2 , xn + 3 , . . . , xn + m ; the resulting
the statistical and engineering literature, the Gaussian
outputs are Y = ½y(x1 ), y(x2 ), . . . , y(xn ), y(xn + 1 ),
function12 is by far the most popular and is also used in
y(xn + 2 ), . . . , y(xn + m ). Thus, at any point x, a linear
this work. It is defined as
predictor of the output can be expressed
!
X
p  2
^y(x) = l (x)Y T
ð19Þ  
g(xi , xj ) = exp  uk xki  xkj  ð28Þ
k=1
The VRPEI-Kriging approach treats ^y(x) as a ran-
dom variable and finds the best linear unbiased predic- where uk (0  uk  ‘) is the vector of unknown correla-
tor, lT (x)Y, which minimizes the mean square error tion parameters which is of length p, the number of
(MSE) of the prediction design variables. Also xki , xkj are the kth element of the
sample points xi , xj .
MSE½^y(x) = E½lT (x)Y  y(x)2 ð20Þ Under the assumption that e(x) in equation (17) fol-
lows a Gaussian process, the likelihood function for it
subject to the unbiasedness constraint is directly related to its probability distribution function
given by
E½lT (x)Y = E½y(x) ð21Þ
1
UK is defined with a set of regression functions ^ s2 , RjY = pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L½m,
(2ps2 )n jRj
  ð29Þ
f(x) = ½f1 (x), f2 (x), ., fp (x) ð22Þ 1
exp  2 (Y  Fm)R^ 1 (Y  Fm) ^
2s
Define a vector F that is the value of f(x) evaluated
at each of the known locations The negative of the Napierian Logarithm (LN)-like-
lihood is given by

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

Gong et al. 9

n   1 evaluate the VRPEI-Kriging with the optimization

LN = ln 2ps2 + ln½jRj
2 2 ð30Þ
1 results could be used to prove the superiority of
+ ^ 1 (Y  Fm)
(Y  Fm)R ^
2s2
From equation (30), the maximum likelihood esti-
mation of s2 could be obtained
1
(Y  Fm)R
^ (Y  Fm)
^ The optimization was carried out following the proce-
^2 =
s ð31Þ
n dure in Figure 8. First, the engineering problem was
By plugging it into equation (30), we obtain the con-
centrated log-likelihood function which is only related
to the parameter u
n   1 n In the numerical simulation of VRPEI-Kriging, a
LN(u) = ln 2ps2 + ln½jRj + ð32Þ
2 2 2
So the value of u could be obtained by solving the mini-
mization problem
find uk .0
n   1 n ð33Þ
min ln 2ps2 + ln½jRj +
2 2 2 large, it will reduce the structural response of the over-
Therefore, if R is known, the maximum likelihood esti-
^ and s2 are easily calculated.
mates of m
Assessment of VRPEI-Kriging meta-model
Due to the fact that the interpolation model goes
exactly through the observations, it is meaningless to
assess the capability of the meta-model to reproduce
the observed data. Thus, the approach of cross-
validation becomes the most prevalent method to
examine the accuracy of the interpolation model.23
Apparently, based on unbiased estimation, the
smaller the root mean square error (RMSE) between
the approximate model and actual model, the higher
the accuracy. Therefore, the RMSE over the domain of
the model is a crucial measurement. The maximum
absolute error represents the maximum deviation
between the estimated value and the actual. So, it
should be taken into consideration to quantify the qual-
ity of the VRPEI-Kriging model. In the construction of
VRPEI-Kriging, the minimized RMSE is acquired dur-
ing the solution of optimum Tw.
As an interpolation model based on OK, the assess-
ment of VRPEI-Kriging should contain the evaluation
of OK. The first is to leave out one or several sample
data and to fit meta-models alternatively based on the
rest of the sample data. The next step is to evaluate the
predicted values of surrogate models on those left-out
data by means of actual test. The accuracy of the
approximate model could be ensured based on the
above steps. Then, the part of the real prototype experi-
ments results should be considered. Without a large
number of actual test data, the intuitive way is to
results. The positive effect on actual optimization
VRPEI-Kriging.
Optimization and result analysis
optimized based on the traditional Kriging model. By
integrating the actual test results from prototype experi-
ments, the VRPEI-Kriging meta-model was constructed
for further optimization.
certain amount of information should be known
through the deterministic virtual test. Therefore, an
appropriate number of sample points should be chosen
in the design of space within a certain range to perform
the numerical experiments.
The choice of sample points in the design space is
very vital. If the selected range of sampling space is too
all simulation accuracy. On the contrary, the greatest
point of impact on reliability may be excluded from the
sample space and then affect the calculation accuracy.
Considering the need of actual engineering, sampling is
performed in the sampling space by means of Latin
Hypercube Sampling (LHS):23
1. Sample 100 times for the six selected design vari-
ables by means of LHS.
2. Calculate the sampling values based on the com-
bined simulation model to obtain the RMS
value of vertical vibration acceleration of front
and rear suspension, which could be taken as the
corresponding response of the Kriging model.
3. Use the calculated value obtained above, in
modeFRONTIER, to establish the approximation
model of the vertical vibration of weighted accel-
eration of the front and rear axle sprung mass.
4. Select five sampling points in the design space
and complete the calculation using the simula-
tion model and the approximation model,
respectively, and thus record the error between
them. Set the model accuracy as 0.05. If the
error meets the requirement, then stop the itera-
tion to obtain the specific parameters of the
Kriging approximation model; otherwise, add
these computed values to the response domain
and then go to step (3) to reconstruct the
approximate model.
According to the above steps, the correlation para-
meters u can be estimated by minimizing LN(u) in
equation (32). Based on the known u,^ the regression

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

10 Advances in Mechanical Engineering

Figure 8. The frame of optimization process.

coefficients m could also be estimated according to Different from the building of the traditional
equation (27). u ^ and m ^ are obtained as Kriging meta-model, the construction of VRPEI-
^
u = (6:431, 2:776, 1:877, 0:682, 7:116, 5:242) and Kriging cannot be separated from the sample data of
^ = (0:269, 0:121, 0:092, 0:035, 0:327, 0:241).
m The actual test. Besides the above results data obtained
Kriging approximation model was obtained as follows through the virtual simulation model, in reference to
(Figures 9 and 10). the constraints and under the premise of considering

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

Gong et al. 11

Figure 9. Approximation model of the RMS value of vertical vibration acceleration of the front axle sprung mass: (a) the changed
aw_f with V0_f and P0_f, (b) the changed aw_f with V0_f and d0_f, and (c) the changed aw_f with P0_f and d0_f.

actual test conditions, 10 samples were randomly
selected in the sampling interval. According to the sam-
ple data, the parameters of hydro-pneumatic suspen-
sion were adjusted and the actual test was performed.
(Figure 11). The corresponding experimental results
were recorded and analyzed, as shown in Table 2.
Taking the Kriging meta-model as the basis and
introducing the actual experimental date based on vir-
tual simulation, the minimized RMS of the error
between the estimated values and test values was taken
as the optimization target to solve the optimal test
weight factor Tw. Thus, the approximation model of
VRPEI-Kriging was constructed in the same way as
the Kriging meta-model. Genetic algorithm, after 200
generation algorithm iterations, was used to solve the
optimal solution based on Kriging and VRPEI-Kriging
meta-model, respectively. The results are shown in
Table 3.
According to the obtained optimal parameters,
actual experiments were carried out. We compare the
test results as follows (Table 4).
From the correlation tables, as shown in Figure 12,
it is obvious that the optimization results based on
Kriging and VRPEI-Kriging meta-model improved the
ride comfort of the vehicle to the same extent.
Just from the virtual optimization, the results,
obtained based on the Kriging surrogate model, showed
greater degree of improvement than on VRPEI-
Kriging. For example, the acceleration RMS of front
suspension was improved by 34.12% (from 1.3555 to
0.893 m/s2) according to the Kriging model, while it
only improved by 16.68% (from 1.3555 to 1.1294 m/s2)
based on VRPEI-Kriging. A similar conclusion can be
obtained from the curve of rear suspension.
Through the verification by the actual test, we found
that the acceleration RMS of front suspension was just

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

12 Advances in Mechanical Engineering

Figure 10. Approximation model of the RMS value of vertical vibration acceleration of the rear axle sprung mass: (a) the changed
aw_r with V0_r and P0_r, (b) the changed aw_r with V0_r and d0_r, and (c) the changed aw_r with P0_r and d0_r.

Table 2. Actual experimental data.

Front hydro-pneumatic spring Rear hydro-pneumatic spring The test results

P0_f (Pa) V0_f (L) d_f (m) P0_r (Pa) V0_r (L) d_r (m) aw_f (m/s2) aw_r (m/s2)

1 8.50E + 06 10.0 0.018 6.50E + 06 3.0 0.008 1.4971 1.1805

2 1.00E + 07 6.0 0.020 1.50E + 07 5.0 0.012 2.4307 1.2892
3 7.50E + 06 8.0 0.014 1.20E + 07 3.5 0.009 1.7531 1.3027
4 8.00E + 06 3.0 0.015 1.20E + 07 6.5 0.007 2.6844 1.1542
5 6.50E + 06 7.0 0.018 7.50E + 06 4.5 0.006 1.4686 1.3161
6 5.00E + 06 3.5 0.016 8.00E + 06 6.5 0.007 2.3126 1.1291
7 6.00E + 06 6.0 0.020 9.50E + 06 4.5 0.008 1.4182 1.4425
8 7.00E + 06 4.0 0.019 1.00E + 07 8.0 0.006 2.5234 1.5374
9 6.00E + 06 5.0 0.020 7.50E + 06 6.0 0.007 2.1326 1.2557
10 7.00E + 06 6.5 0.016 1.50E + 07 6.5 0.012 1.9307 1.2090

1.2236 m/s2, actually improved by only 9.73% based on suspension. The result, based on VRPEI-Kriging, was
Kriging. While based on VRPEI-Kriging, the result improved by 11.79%, better than 8.69% from the opti-
was improved by 12.48%. It is similar for the rear mization based on the Kriging model. Compared with

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

Gong et al. 13

Figure 11. The layout of sensors and data acquisition (left: the layout of front acceleration sensor; middle: the layout of rear
acceleration sensor; right: data acquisition system).

Table 3. Results of optimization.

Front hydro-pneumatic spring Rear hydro-pneumatic spring The acceleration RMS

P0_f (Pa) V0_f (L) d_f (m) P0_r (Pa) V0_r (L) d_r (m) aw_f (m/s2) aw_r (m/s2)

Before optimization 5.5E + 06 6.0 0.016 1.0E + 07 4.0 0.009 1.3555 1.2285
Kriging 6.5E + 06 9.0 0.012 9.5E + 06 6.0 0.006 0.8931 0.7457
VRPEI-Kriging 6.5E + 06 8.0 0.012 9.0E + 06 8.0 0.006 1.1294 0.8968

RMS: root mean square; VRPEI-Kriging: virtual and real prototype experiment integrated Kriging.

Table 4. Comparison of optimization.

Meta-model Front suspension (m/s2) Rear suspension (m/s2)

Before optimization – 1.3555 1.2285

Virtual optimization Kriging 0.8931 0.7457
VRPEI-Kriging 1.1294 0.8968
Actual test Kriging 1.2236 1.1217
VRPEI-Kriging 1.1863 1.0837

VRPEI-Kriging: virtual and real prototype experiment integrated Kriging.

traditional Kriging, the data analysis showed that the
optimization effect was improved by 3.05% and 3.38%,
respectively, based on VRPEI-Kriging.
What is more, there is every reason to think that the
optimization based on VRPEI-Kriging is more efficient
by mutual comparison of the test data.
Conclusion
The ride comfort optimization of a multi-axle heavy
motorized wheel dump truck was done completely
based on the VRPEI-Kriging model. The complete
CAE simulation model was established, including the
dynamics model, the hydro-pneumatic spring model,
and 3D pavement model. Under the constraints of nat-
ural frequency, the working stroke, and the dynamic
load of the wheel, the vibration acceleration RMS of
the sprung mass center was optimized. Considering the
complexity of the engineering and the optimization
cost, the optimization was executed based on the
VRPEI-Kriging meta-model to find the optimal para-
meters for the front and rear hydro-pneumatic suspen-
sion. The results showed that the acceleration RMS
was actually improved by 12.48% for front suspension

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

14
Figure 12. Comparison of optimization.
and 11.79% for rear suspension. It is meaningful to
relieve the fatigue of the frame and the suspension.
The VRPEI-Kriging meta-model proposed in this
article was built based on traditional Kriging. The inter-
polation principle was deduced in detail. The VRPEI-
Kriging model considers the actual test results with high
reliability, while the traditional did not. However, in
some particularly complex project with strong nonli-
nearity, the order of Kriging surrogate model obtained
from the pure CAE analysis could be lowered. In that
case, it would lead to the loss of the extremum or the
optimal solution. Then, we need the actual test results
to modify the approximate model to remedy the short-
age. VRPEI-Kriging precisely contributes to cover the
potential defect, especially in an intricate engineering
problem whose CAE simulation model is almost impos-
sible to be established with high accuracy. Maybe there
are still some deficiencies in the VRPEI-Kriging, such
as the arithmetic solution of the test weight ratio Tw. It
will be further studied in later work. Nonetheless, the
optimization in this article illustrates the efficiency of
the VRPEI-Kriging meta-model. The ride comfort of
the multi-axle heavy motorized wheel dump truck was
improved effectively based on the VRPEI-Kriging
model.
Declaration of conflicting interests
The authors declare that there is no conflict of interest.
Funding
This research is supported by the National Natural Science
Foundation of China (Grant No. 51305314) and the
Fundamental Research Funds for the Central Universities
(Grant No. 2014-VII-005).
Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015
Advances in Mechanical Engineering
References
1. Jayachandran R and Krishnapillai S. Modeling and
optimization of passive and semi-active suspension
systems for passenger cars to improve ride comfort and
isolate engine vibration. J Vib Contr 2013; 19:
1471–1479.
2. Kaldas M, Cxalısxkan K, Henze R, et al. Optimization of
damper top mount characteristics to improve vehicle ride
comfort and harshness. Shock Vib 2014; 2014: 1–15.
DOI: 10.1155/2014/248129.
3. Eriksson P and Friberg O. Ride comfort optimization of
a city bus. Struct Multidiscip O 2000; 20: 67–75.
4. Yang Y, Ren W, Chen L, et al. Study on ride comfort of
tractor with tandem suspension based on multi-body sys-
tem dynamics. Appl Math Model 2009; 33: 11–33.
5. He J, Chen Y, Zhao C, et al. Heavy truck suspension
optimisation based on modified skyhook damping con-
trol. Int J Heavy Veh Syst 2011; 18: 161–178.
6. European Union Directive and General Provisions.
Directive 2002/44/EC of the European Parliament and
the Council of 25 June 2002 on the minimum health and
safety requirements regarding the exposure of workers to
the risks arising from physical agents (vibration) (six-
teenth individual Directive within the meaning of Article
16(1) of Directive 89/391/EEC). Off J Eur Communities
Legis 2002; 117: 6–7.
7. Filla R and Palmberg JO. Using dynamic simulation
in the development of construction machinery (arXiv
preprint cs/0305036), 2003, http://arxiv.org/abs/cs/0305036
8. Uys PE, Els PS and Thoresson M. Suspension settings
for optimal ride comfort of off-road vehicles travelling
on roads with different roughness and speeds. J Terrame-
chanics 2007; 44: 163–175.
9. Shirahatti A, Prasad PSS, Panzade P, et al. Optimal
design of passenger car suspension for ride and road
holding. J Braz Soc Mech Sci 2008; 30: 66–76.
10. Langer TH, Christensen BB, Mouritsen OØ, et al. Opti-
mization of front axle suspension system of articulated
 Gong et al. 15

dump truck. In: 1st joint international conference on multi- Appendix 1

body system dynamics: CD-ROM, Lappeenranta, Fin-
land, 25–27 May 2010. Notation
11. Matheron G. The theory of regionalized variables and its
applications, vol. 5. France: Ecole nationale supérieure aw root mean square (RMS) value of vertical
des mines de Paris, 1971, p. 218. vibration acceleration in the center of the
12. Cressie N. The origins of kriging. Math Geol 1990; 22: sprung mass
239–252. A1 cross-sectional area of the piston rod
13. Sacks J, Welch WJ, Mitchell TJ, et al. Design and d0 f damping orifice in front hydro-pneumatic
analysis of computer experiments. Stat Sci 1989; 4: spring
409–423. d0 r damping orifice in rear hydro-pneumatic
14. Martin JD and Simpson TW. A study on the use of kri- spring
ging models to approximate deterministic computer mod- e(x) Gaussian stationary process
els. In: ASME 2003 international design engineering Fd dynamic loads of the wheels
technical conferences and computers and information in
G static load of the wheels
engineering conference, Chicago, IL, 2–6 September 2003,
Ga (f ) acceleration power spectral density
pp.567–576. New York: American Society of Mechanical
Engineers.
function
15. Martin JD and Simpson TW. On the use of kriging mod- J objective function
els to approximate deterministic computer models. In: K suspension stiffness
ASME 2004 international design engineering technical Mb suspension sprung mass
conferences and computers and information in engineering n suspension offset frequency
conference, Salt Lake City, UT, 28 September–2 October p number of design variables
2004, pp.481–492. New York: American Society of Pk pressure of the gas chamber in no load
Mechanical Engineers. Pm pressure of the gas chamber in full load
16. Duchaine F, Morel T and Gicquel LYM. Computa- P0 f initial pressure of accumulator in front
tional-fluid-dynamics-based kriging optimization tool for hydro-pneumatic spring
aeronautical combustion chambers. AIAA J 2009; 47: P0 r initial pressure of accumulator in rear
631–645.
hydro-pneumatic spring
17. Lim YW, Choi HJ and Idapalapati S. Design of Alporas
R correlation matrix
aluminum alloy foam cored hybrid sandwich plates using
Kriging optimization. Compos Struct 2013; 96: 17–28. Skm amount of compression for hydro-
18. Li M, Li G and Azarm S. A kriging metamodel assisted pneumatic spring from no load to full
multi-objective genetic algorithm for design optimization. load
J Mech Des 2008; 130: 031401. T analysis of vibration time
19. Jeong S, Murayama M and Yamamoto K. Efficient opti- T (tk ) actual test result at point tk
mization design method using kriging model. J Aircraft Tw test weight ratio
2005; 42: 413–420. Vk volume of the gas chamber in no load
20. Cheng YC and Lee CK. Integration of uniform design V0 f initial volume of accumulator in front
and Kriging interpolation to the optimization of suspen- hydro-pneumatic spring
sion parameters of a high speed railway vehicle. In: The V0 r initial volume of accumulator in rear
2010 international conference on modelling, identification hydro-pneumatic spring
and control (ICMIC), Okayama, Japan, 17–19 July 2010,
W (f ) frequency weighting function
pp.388–393. New York: IEEE.
21. Hao W, Shaoping W and Tomovic MM. Modified
X design variables
sequential Kriging optimization for multidisciplinary xki kth element of the sample points xi
complex product simulation. Chin J Aeronaut 2010; 23: Z(x) actual value at point x
616–622. Z(xi ) simulation result at point xi
22. Langer TH, Hansen MR, Ebbesen MK, et al. A tire Z * (x) predictor at point x
model for off-highway vehicle simulation and comfort
evaluation. In: ASME 2009 international design engineer- ai weight coefficient of sample value
ing technical conferences and computers and information in bk weight coefficient of experimental value
engineering conference, San Diego, CA, 30 August–2 Sep- g(  ) spatial correlation function
tember 2009, pp.2057–2065. New York: American Soci- u coefficients in the Gaussian correlation
ety of Mechanical Engineers. function
23. Helton JC and Davis FJ. Latin hypercube sampling and m regression coefficients
the propagation of uncertainty in analyses of complex j dynamic load coefficient
systems. Reliab Eng Syst Safe 2003; 81: 23–69. s2E RMS of estimation error
O domain of the computer model

Downloaded from ade.sagepub.com at Wuhan Institute of Technology on September 17, 2015

View publication stats