Finite Element Analysis of Cylinder Piston Impact

Based on ANSYS/LS-DYNA

Mi Yan
School of Automation
Beijing Institute of Technology
Beijing, China, 15011058244
[email protected]
Wang Tao
School of Automation
Beijing Institute of Technology
Beijing, China, 010-68944054
[email protected]


AbstractNonlinear dynamic finite element analysis system
ANSYS/LS-DYNA is used to analyze the characteristics of
cylinder piston impacts to meet the requirements of safety and
stability for cylinder. Based on the fundamental theories of
explicit dynamics, a finite element impact model of cylinder
piston was established. The impact force variation, the piston
surface stress distribution and the relationship between
maximum impact force and different piston initial impact
velocity were estimated by numerical simulation. The results
indicate that the impact force on the force transducer is about
81% of that on the cylinder piston owing to the collar between the
piston and the force transducer, and the stress, close to the
surface center of the piston, is larger than that of other areas. The
results also show that different initial impact velocity dont affect
impact time but only the impact force, and the maximum impact
force is proportional to the initial impact velocity.
Keywords- cylinder piston; explicit dynamics; finite element;
impact force; stress
I. INTRODUCTION
Cylinder is widely used as actuator in a variety of
pneumatic transmission systems, which can convert the air
pressure energy into mechanical energy.
[1]
The movements of
reciprocating, swing and spinning were made through the load
of drive mechanism. If the working parts moved by air cylinder
are sufficiently big and fast, a great deal of impact and
vibration are generated when cylinder suddenly get stopped or
reversed. So the security and stability of cylinder will be
influenced and the service life will be shorted.
[2]
To satisfy the
functional requirements, it is needed to analyze the impact
characteristics of the cylinder piston.
Recently, research on cylinder mainly focus on crawling
features, the designs of the buffering device and mounting
structures, but few is on the impact properties of the cylinder
piston. A rough estimate based on experience rather than theory
was given by instruction manual of pneumatic company, they
selected insurance coefficient and verified the value falls within
a range of safety in production. Therefore, the analysis of the
impact properties of cylinder piston is great importance for
both theory and practice.
During the analyzing and testing of the position impact
properties, How to obtain the accurate and real-time parameters
of the impact is the key point. In this paper, a numerical
simulation analysis is combined with dynamics and finite
element method, to solve the problem of impact dynamics of
complex structures.
In recent years, the explicit dynamic finite element method
has been widely used in the field of collision impact and has
showed a strong advantage in dealing with large-scale contact
problems. ANSYS/LS-DYNA is a general explicit nonlinear
dynamic finite element system software, it can solve a variety
of nonlinear problems, such as the nonlinear contact of high-
speed collision, explosion, and metal forming in three-
dimensional nonlinear structure, the impact of load nonlinearity
and material nonlinear problems
[3]
. This paper will use the
non-linear dynamic finite element analysis system,
ANSYS/LS-DYNA, to simulate and analyze the impact
properties of the cylinder piston, and use the finite element
method to build Cylinder piston collision model, therefore, the
impact properties of the piston, the general rules and
characteristics of the impact response is quantitatively analyzed.
The explicit dynamic finite element theory of collision
analysis is introduced in section II. The building of simulation
model and the settings of system parameters are introduced in
section III. In impact process the impact force changes of every
collision parts are detailed analyzed in section IV, as well as
the piston surface stress and the stress distribution, the
maximum impact force changes under different cylinder piston
initial collision velocity and etc. The whole article is
summarized in section V.
II. THE EXPLICIT DYNAMIC FINITE ELEMENT THEORY OF
COLLISION ANALYSIS
Collision is a transient complex physical process, which
belongs to Non-linear dynamics problems. The explicit
dynamics finite element method in LS-DYNA is mainly used
in solving the problems of collision
[4]
. Recent years, the
explicit dynamic finite element method has been widely used in
the field of collision impact and shows a strong advantage in
dealing with large-scale contact problems. Unlike the static
implicit one, the explicit dynamic finite element method does
not need to establish the stiffness matrix and calculate the
inverse operation, but using the explicit central difference
method for solving finite element equations can effectively
avoid the convergence problems which caused due to nonlinear
problems. The speed of solving problems will be elevated, via
Proceedings of 2012 International Conference on Mechanical Engineering and Material Science (MEMS 2012)
2012. The authors - Published by Atlantis Press 107


single-point Gaussian integration and centralized quality,
which has advantage of saving the computation time and
storage space, as well as easy to deal with.
[5]

Specific explicit dynamic finite element analysis theory is as
follows:
1) Dynamics equation, as in (1)
F KU U C U M = + +

(1)
Where, M is the mass matrix of the structure, C is the
damping matrix, K is the stiffness matrix, F is external applied
load vector, and U is the structure of the displacement vector.
2) If Displacement, velocity and acceleration of 0,t

,t
2
,t
n

are known, and structural response at the time t
n+1
(t+t) is to
be determined.
The central difference is used to replace the acceleration,
velocity derivative central difference, which represented as:
{ }
t t t t t
U U
t
U
A + A
+
A
=
-
-
2
1

(2)
{ }
t t t t t
U U
t
U
A + A
+
A
=
-
-
2
1

(3)
3) Replace equation (2) & (3) into equation (1), and we get
as follows:

t t t t t
t t t
U C
t
M
t
U M
t
K F R
C
t
M
t
M
R U M
A
A +
|
.
|

\
|
A

A
|
.
|

\
|
A
=
A
+
A
=
=
2
1 1 2

2
1 1

2 2
2
(4)
Where,
t
R

is payload vector, M

is effective mass matrix,

t
F is structural load vector.
4) Solving linear equations (4), may get the displacement
vector U
t+t
at the time of t+t. Substituding U
t+t
into the
elastodynamics balance equation (5) & (6), we may get the cell
stress and strain at the time of t+t.
Geometric equation:
( )
ji ij ij
u u + =
2
1
c (5)
Physical equation:

( )
kk kk kk
ij ij kk ij
Kc c o
c o c o
3 2 3
2
= + =
+ =
(6)
5) Solving equations with the central difference method by
iterative solution, until the closing conditions to meet the
computing.
The central difference method is an explicit algorithm, but
its condition is stable, which means that when using this
method to solve problems, the size of time step must be less
than a threshold determined by the nature of the problem
solving equation. Otherwise, the algorithm is unstable, and
time step t contented with:

max
2
e
= A s A
crif
t t (7)
Where,
max
is maximum natural angular frequency.
For the transient problem, explicit analysis of the time step
size is small. Therefore, the problem of collisions in the
simulation of piston impact test is feasible via using the explicit
dynamic finite element method.
III. SIMULATION MODELING AND SOLVING
According to the finite element theory, establishing a
simulation model is divided into solid modeling and finite
element model to establish, among which, the finite element
model includes the choice of unit type, the set of material
properties and the division of finite element meshing. Finally,
contact type and contact algorithm are defined, and the
boundary conditions, initial loads and constraints are set to
finish the solution to the model. Each section will be followed
by details.
A. Solid modeling
Model uses solid modeling, in accordance with the actual
size of the impact of various parts of the experimental
apparatus, to build the geometric model in solidworks, with
ANSYS/LS-DYNA imported, as shown in Figure 1.

Figure 1. Model diagram of the experimental device and its corresponding.
According to the computing needs, as much as possible to
keep the actual structure, details of delivery which do not affect
the force between the parts are simplified. The simplifications
include: 1) Joints of screw in the actual structure are simplified
as direct stick together and the threads are ignored. 2) To
simplify the analysis, one fourths of the model was taken in
simulation. The units of cm-g-s are used when the accuracy of
the analysis is taken into account.
B. The establishment of finite element model
1) The choice of unit type
The type of unit uses solid elements solid164, which is a
hexahedral element with eight nodes, each of which includes
the X, Y, Z, such three directions on the state of displacement,
velocity and acceleration. This low-level unit of computing has
an advantage of fast speed, high accuracy and saving the
machine-hour
[6]
. In this model, because it is a single point of
force of the collision, calculation results can be serious
108


hourglass problem. In order to prevent the hourglass problem
affecting the accuracy of calculation, the full integration
algorithm is adopted.
2) The set of the material properties
Various components of the experimental device materials
just as follows. Cylinder piston is made of 7075 aluminum
alloy, collar and experiment platform is 45 steel and force
transducer is stainless steel. Three kinds of materials are all of
No.15 material model *MAT_JOHNSON_COOK (Johnson-
Cook constitutive model
[6]
) in LS-DYNA.
3) The division of finite element meshs
The number of grid directly affects the calculation accuracy
and scale. For the convergence problem, the increase in the
number of grid will improve the precision of the calculation,
but also will increase the size of the calculation. When the grid
number achieves to a certain extent, if continue to increase,
calculation accuracy improved rarely, but calculation scale is
big many. Therefore, in order to analyze the cylinder piston
impact characteristics more accurately, the gridding of
important structure is minished, especially for piston, collar and
force transducer, as shown in Figure 2.

Figure 2. Local finite element model.
C. The definition of contact type and contact algorithm
In the cylinder piston impact process, the contact between
different parts is face-to-face contact, so using the Surface to
Surface contact type
[7]
of ANSYS/LS-DYNA. In the possible
contact effect between different parts define the master surface
and slave surface. Contact between the parts interaction
through the contact algorithm to complete, and this paper
adopted penalty method
[8]
.
D. The set of the boundary, initial loads and constraints
From mathematical perspective, each kind of boundary
conditions and load corresponding to the only solution, so as
much as possible to fit the model boundary conditions and load
with actual situation and to get the correct solution. Therefore,
in the simulation, constraint the bench bottom all freedom
degrees and the cylinder piston all freedom degrees except
collision direction.
In the experiment, cylinder piston does a free-falling
movement in the height of h. According to equation,
gh v 2 = , initial impact velocity is applied to the cylinder
piston and gravity is added to the whole model.
IV. CYLINDER PISTON IMPACT PROCESS SIMULATION
RESULTS AND ANALYSIS
This chapter mainly analyzes the impact properties of
cylinder piston, includes: in the impact process the change of
impact force, the piston surface stress and the stress distribution,
the maximum impact force changes under different cylinder
piston initial collision velocity and etc.
In the simulation experiment, the simulation conditions are
as follows: suppose the cylinder piston falls from the height of
80mm, i.e. h=80mm, consequently, piston impact collar at the
speed of 1.25m/s, i.e. v=1.25m/s. The distance between piston
and collar is 0.05mm.
A. The analysis of impact force in the impact process
The simulation results of impact force on the contact
interface between piston and collar, and impact force on the
contact interface between collar and force transducer are shown
in Figure 3. Where, curve A is the impact force changes
between piston and collar, curve B is the impact force changes
between collar and force transducer.

Figure 3. Impact force on the contact interface between piston and collar,
and between collar and force transducer
As the pulse shapes, curve A and B seen in Figure 3, (a)
Pulse shapes are quite irregular, impact force is rising rapidly
and reaching its maximum after collision occurred and then
decreasing, mostly due to high frequency response. (b) When
t=40s, piston impacted collar, and when t=72s they detached,
so the impact time is 32s, and the impact peak is 20.94kN. (c)
When impact was delivered to force transducer, due to the
damp effect of collar, the time delay is 2s. While, the impact
time is 42s, and impact peak is 17.08kN. These show that the
impact force of force transducer is smaller than the impact
force of piston, and the former is about 81 percent of the latter.
B. The analysis of the piston surface stress
The simulation analysis of impact force is the average of
piston surface force. In the practical application, there is stress
distribution in the piston surface. This chapter will introduce
the stress and deformation of the piston through analyzing the
stress and stress distribution in simulation, so as to provide
reference for the optimization design of the piston.
1) The simulation get unit stress changing when radius were
taken 11.5 mm, 12.6 mm, 17.6 mm, 22.6 mm and 31.2 mm, the
curves are presented in Figure 4. When r=11.5mm, the
maximum stress is 44.5MPa. When r=11.5mm, the maximum
stress is 44.5MPa. When r=12.6mm, the maximum stress is
38.6MPa. When r=17.6mm, the maximum stress is 33.3MPa.
When r=31.2mm, the maximum stress is 36.8MPa.
As shown in Figure 4, it is known that the stress change
time is same to impact happened time in Figure 3, but pulse
shapes are less regular than these in Figure 3. This is mainly
109


due to the particularity of individual units, impact force
increased rapidly after collision occurred, and then decreased.
Stress has a smaller concussion near 0.

Figure 4. Z-stress comparison of different radius.
2) In the simulation experiments, selecting units of piston
surface and the radius from 10 mm to 31.5 mm. Taking these
radius units to do stress analysis, maximum stress comparison
of different radius is presented in Figure 5.

Figure 5. The maximum stress comparison of different radius.
As shown in Figure 5, a) Radius range from 10.5mm to
14mm, maximum stress is much larger. When r=11.5mm, the
maximum stress is largest, it is 44.5MPa. b) With the increase
of the radius, maximum stress tend to be stable. When r=30mm,
the maximum stress is smallest, it is 32.9MPa. c) Radius range
from 30mm to 31.5mm, maximum stress gradually increasing,
and at the piston edge maximum stress reach 36.6MPa.
3) According to the above analysis results, at t=47s, the
impact force is largest, and the von-mises stress map is
presented in Figure 6.

Figure 6. Von-mises stress map,when t=47s.
As shown in Figure 6, stress wave is diffusion, maximum
stress appears in internal position of piston surface, and these
arethe same to the analysis of stress distribution.
C. The maximum impact force changes with different
cylinder piston initial collision velocity
In addition to the piston initial impact velocity v=1.25m/s,
the simulation experiment also compared impact force under
different velocities. The comparison results are presented in
Figure 7.

Figure 7. Comparison of impact forces under different velocities.
In Figure 7, the impact force curves under different
velocities are similarity, pulse width is 32s, and the maximum
impact force is proportional to initial collision velocity. In other
words, different initial impact velocity, would not affect impact
time, it can only influence the size of the impact forceand the
maximum impact force is proportional to initial piston initial
collision velocity in a certain range.
V. CONCLUSION
In this paper, based on the fundamental theories of explicit
dynamics and finite element analysis system ANSYS/LS-
DYNA, the research and analysis of the cylinder piston impact
characteristics is developed.
Theoretical analysis and simulation results show that: a)
Due to the buffer function of collar, the impact force on the
force transducer is about 81% of which on the cylinder piston.
b) Piston surface stress is different, and the inside position of
piston surface stress is larger, which can reduce the influence of
impact by adding buffer. With the increasing of radius, the
stress changes smoothly. While the outside edge stress
increases, it can be strengthened by the material. c) Different
initial impact velocity, dont affect impact time but only the
impact forceand the maximum impact force is proportional
to the initial impact velocity in a certain range. These
conclusions provide strong foundation for the coming impact
experiment and the theoretical basis for the optimization design,
maintenance and damage evaluation of cylinder piston.
REFERENCES
[1] SMCChinaCo.Ltd, Modern pratical pneumatic technology, Beijing:
Machine Industry Press, 1998.
[2] Liu Hanjun, Factors in fluencing the dynamic characteristics of the
cushion cylinder, Hydraulics and Pneumatics, vol. 1, pp. 33-37, 1987.
[3] Shang Xiaojiang, and Su Jianyu, Ansys/ls-dyna dynamic analysis
method and engineering examples, Beijing: China WaterPower Press,
2008.
[4] Bai Jinze, LS - DYNA3D theory foundation and case analysis, Beijing:
Science Press, 2005.
[5] Huang Qiao, Li Zhonglong, and Sha Xuejun, Strain sensor-based
wireless measurement system for bridge, vol. 10, pp. 1322-1325, 2007.
[6] J.O. Hallquist, LS-DYNA Theory Manual, 2006.
[7] Shi Dangyong, Analysis of explicit dynamic based on ANSYS/LS-
DYNA 8.1, Beijing: Tsinghua University Press, 2004.
[8] Zhong Yang, Zhong Zhihua, and Li Guangyao, Review on Contact
Algorithms Calculating the Contact-impact Interface in Mechanical
System with Explicit FEM, vol. 47, pp. 44-58, 2011.
110