Chapter 24: Friction Between Belt and Pulley

Friction Between Belt

24 and Pulley


Summary 439

Introduction 440

Requested Solutions 440

Analytical Solution 440

FEM Solutions 441

Modeling Tip 444

Input File(s) 446

Video 446
 CHAPTER 24 439
Friction Between Belt and Pulley

Summary
Title Chapter 24: Friction Between Belt and Pulley
Contact features • (Slightly) changing contact area
• Curved contact surfaces
• Deformable-deformable and deformable-rigid contact
• Friction between deformable bodies
Geometry 3-D (units: mm)
R
• Pulley outer radius = 0.55 ϕ
• Pulley inner radius = 0.25 r2
r1
• Out of plane pulley thickness = 0.3
• In plane belt thickness = 0.05
• Out of plane belt thickness = 0.2 y
t
• Initial angle spanned = /2 rad
z x
F

Material properties 13 10
E pulley = 1.0 10 Pa E belt = 1.0 10 Pa  pulley =  belt = 0.3
Linear elastic material
Analysis type Quasi-static analysis
Boundary conditions An 180o section of the pulley is modeled, which is clamped along the inner radius using
“glued” contact conditions. On both ends of the belt, load-controlled rigid bodies are
defined and connected to the belt using “glued” contact conditions. The forces F and R
are external and reaction forces on the control nodes. On the loaded control node we have
u x = u y = 0 , while on the other control node u x = u y = u z = 0 .

Applied loads Point load F y = – 1.0 105 N

Element type 3-D 20-node hexahedral solid elements
Contact properties Different coefficients of friction between belt and pulley:  = 0.05 ,  = 0.15 and
 = 0.25

FE results Reaction force for each value of the friction coefficient

440 MD Demonstration Problems
CHAPTER 24

Introduction
A belt is positioned around a pulley such that a 90o section of the pulley is contacted. One end of the belt is fixed; the
other end is loaded by a tensile force with magnitude F = 1.0 105 . It is assumed that the material behavior for both the
belt and the pulley is linear elastic. Although this problem can be solved by a 2-D approximation, a full 3-D model is
chosen here in order to show the characteristic behavior of 3-D parabolic hexahedral elements in a contact analysis
involving friction. An analytical solution for the case with Coulomb friction is known.

Requested Solutions
Analyses will be carried out for three different values of the friction coefficient:  = 0.05 ,  = 0.15 , and  = 0.25 .
With a constant value of the applied load, the reaction force will decrease for increasing values of the friction
coefficient. This reaction force is the primary requested quantity, as this can be easily compared with an analytical
solution.

Analytical Solution
Assuming Coulomb friction between the belt and the pulley, the principle of rope friction according to the Euler-
Eytelwein formula provides a relation between the magnitude F of the applied force, the magnitude R of the reaction
force, the angle  spanned by the belt and the friction coefficient  between the belt and the pulley:

F
R = --------

e

With F = 1.0 105 and  = --- , the theoretical value of the magnitude of the reaction force R is listed in Table 24-1 for
2
various values of the friction coefficient  .

Table 24-1 Reaction Force for Various Values of the Friction Coefficient (Theory)
Friction Coefficient  Reaction Force R
0.05 9.2447x104
0.15 7.9008x104
0.25 6.7523x104
 CHAPTER 24 441
Friction Between Belt and Pulley

FEM Solutions
Numerical solutions have been obtained with MD Nastran’s SOL 400 for the element mesh shown in Figure 24-1 using
3-D 20-node hexahedral elements. Assuming that the deformations of the pulley are small and localized around the
contact area, only an 180o section has been modeled. In total, there are five contact bodies: two deformable and three
rigid. The rigid bodies will be used to easily apply the boundary conditions (single point constraints and forces).

load controlled
rigid body

fixed rigid body;

glued contact

load controlled
rigid body

Figure 24-1 Element Mesh applied in MD Nastran Simulation

The first deformable body consists of all elements of the belt, where the second deformable body consists of all
elements of the pulley. The body number ID’s of the belt and the pulley are 1 and 2, respectively. These deformable
contact bodies are identified as 3-D bodies referring to the BSURF IDs 1 and 2:
BCBODY 1 3D DEFORM 1
BSURF 1 1 2 3 4 5 6 7
8 9 10 11 12 13 14 15
16 17 18 19 20 21 22 23
24 25 26 27 28 29 30 31
32 33 34 35 36 37 38 39
40 41 42 43 44 45 46 47
48 49 50 51 52 53 54 55
56 57 58 59 60 61 62 63
64 65 66 67 68 69 70 71
72 73 74
BCBODY 2 3D DEFORM 2
BSURF 2 75 76 77 78 79 80 81
82 83 84 85 86 87 88 89
90 91 92 93 94 95 96 97
98 99 100 101 102 103 104 105
106 107 108 109 110 111 112 113
114 115 116 117 118 119 120 121
122 123 124 125 126 127 128 129
130 131 132 133 134
442 MD Demonstration Problems
CHAPTER 24

The first rigid body is a half cylinder described as a NURBS surface and will be used to clamp the grids on the inner
radius of the pulley. Its body ID number is 3 and it is identified as:
BCBODY 3 3D RIGID 0 1 0
0 0. 0. 0. 1. 0. 0. 0.
RIGID 0 1 RIG-INNER
NURBS -7 13 4 4 50 50 0
.176777 -.176777 0. .324015 -.029538 0.
.237263 .222631 0. .0306021.24812 0.
...

The second and the third rigid bodies are load controlled rigid bodies. A load controlled rigid body is associated with
a control grid, which can be used to apply forces and/or single point constraints. In the current analysis, two flat load
controlled rigid bodies are used. They will be glued to both ends of the belt and their control grids will be used to
prevent a rigid body motion in the basic z-direction, to apply the external force on the belt and to transfer the belt load
to the fixed control grid. The load controlled rigid bodies are identified as:
BCBODY 4 3D RIGID 0 1 526
0 0. 0. 0. 1. 0. 0. 0.
RIGID 526 1 RIG-R
NURBS -2 2 2 2 50 50 4
-.2 .6 .05 -.2 .55 .05
-.2 .6 .25 -.2 .55 .25
...
BCBODY 5 3D RIGID 0 1 527
0 0. 0. 0. 1. 0. 0. 0.
RIGID 527 1 RIG-F
NURBS -2 2 2 2 50 50 4
.55 -.2 .05 .6 -.2 .05
.55 -.2 .25 .6 -.2 .25
...

Note that the control grids have the IDs 526 and 527.
The BCTABLE option will be used to indicate:
• which grids are to be treated as slave nodes and which as master grids in the multipoint constraints for
deformable-deformable contact;
• the friction coefficient between the belt and the pulley;
• glued contact between the pulley and the half cylinder;
• glued contact between the load controlled rigid bodies and the belt.
The entries of the BCTABLE option are defined as:
BCTABLE 1 4
SLAVE 1 0. 0. .05 0. 0 0.
1 0
MASTERS 2
SLAVE 1 0. 0. 0. 0. 1 0.
0 1 0
MASTERS 5
SLAVE 1 0. 0. 0. 0. 1 0.
0 1 0
MASTERS 4
SLAVE 2 0. 0. 0. 0. 1 0.
0 1 0
MASTERS 3
 CHAPTER 24 443
Friction Between Belt and Pulley

The first SLAVE MASTERS combination indicates that the grids of deformable body 1 are treated as slave grids when
contact is established with body 2. The friction coefficient is set to 0.05.
The other SLAVE MASTERS combinations activate glued contact between the bodies with body ID numbers 1 and 5,
1 and 4, and 2 and 3, respectively.
The bilinear Coulomb friction model will be activated using the BCPARA option (FTYPE = 6); this option is also used
to indicate that the separation behavior is based on stresses (IBSEP = 4), which is necessary in a contact analysis
involving quadratic elements:
BCPARA 0 NBODIES 5 IBSEP 4 FTYPE 6

In order to activate the full nonlinear formulation of the 20 node hexahedral elements, the nonlinear property extension
of the PSOLID entry is used. For the materials defining the belt (material ID number 1) and the pulley (material ID
number 2), this results in:
MAT1 1 1.+9 .3 1.
MAT1 2 1.+13 .3 1.
PSOLID 1 1 0
PSLDN1 1
PSOLID 2 2 0
PSLDN1 2

The nonlinear procedure used is:

NLPARM 1 1 FNT 1 25 UPW YES
1.e-4 1.e-4 1.e-4 10

Here the FNT option is selected to update the stiffness matrix during every recycle using the full Newton-Raphson
iteration strategy. Convergence checking is performed based on displacements, forces and work. The error tolerance
is set to 10-4 for all criteria. Note that the MAXDIV field is set to 10 to avoid that bisections occur, since too many
bisections may increase the overall solution time.
The obtained values of the reaction forces are listed in Table 24-2, together with the relative error compared to the
analytical solution. The numerical and analytical solutions turn out to be in good agreement.

Table 24-2 Numerical Solutions and Relative Errors

Friction Coefficient  Reaction Force R Error (%)
0.05 9.2314x104 0.14
0.15 7.9476x104 0.59
0.25 6.8448x104 1.37
444 MD Demonstration Problems
CHAPTER 24

Modeling Tip

Convergence Behavior
A nonlinear analysis involving contact and friction may need several iterations to fulfil the convergence requirements.
In such inherently nonlinear analyses, it may be advantageous to increase the number of criteria needed to force a
bisection. As discussed above, this number (MAXDIV on the NLPARM option) has been set to 10 instead of the default
value 3. The tables below show the convergence behavior with the increased value (Table 24-3) and the default value
(Table 24-4). The increased value clearly reduces the overall number of Newton-Raphson iterations and thus the
analysis wall time. When looking at Table 24-3, iteration 9 reaches displacement, load and work errors which are
within the required tolerances. The extra iterations needed are caused by the fact that some grids of the belt which are
initially in contact with the pulley, separate because of tensile contact stresses. After separation of these grids, a new
solution with a smaller number of contact constraints has to be found.

Table 24-3 Convergence Behavior with MAXDIV=10 (

Load Factor Step Iteration Disp. Error Load Error Work Error
1.000 1 1 1.00E+00 1.70E-01 1.70E-01
1.000 1 2 7.76E+00 3.54E-01 1.58E+00
1.000 1 3 6.61E+02 2.31E+01 6.17E+02
1.000 1 4 2.12E+02 1.80E+02 1.30E+04
1.000 1 5 8.61E-02 2.78E+01 7.33E+00
1.000 1 6 3.12E-03 1.70E-01 4.67E-02
1.000 1 7 2.60E-04 4.31E-03 3.50E-03
1.000 1 8 7.87E-06 4.09E-05 1.34E-04
1.000 1 9 3.92E-06 9.30E-07 5.09E-05
1.000 1 10 3.39E+00 1.41E-02 4.30E+00
1.000 1 11 4.26E-02 2.05E-03 6.67E-01
1.000 1 12 2.42E-03 3.31E-02 3.33E-02
1.000 1 13 8.19E-06 2.26E-05 1.30E-04
1.000 1 14 4.93E-06 1.61E-06 6.57E-05
 CHAPTER 24 445
Friction Between Belt and Pulley

Table 24-4 Convergence Behavior with MAXDIV=3 (

Load Factor Step Iteration Disp. Error Load Error Work Error
1.0000 1 1 1.00E+00 1.70E-01 1.70E-01
1.0000 1 2 7.76E+00 3.54E-01 1.58E+00
1.0000 1 3 6.61E+02 2.31E+01 6.17E+02
1.0000 1 4 2.12E+02 1.80E+02 1.30E+04
0.5000 1 1 1.00E+00 9.36E-02 9.36E-02
0.5000 1 2 8.06E+02 2.96E-01 3.12E+02
0.5000 1 3 5.62E+02 3.36E+01 6.19E+02
0.5000 1 4 8.37E+01 8.70E+01 1.92E+02
0.5000 1 5 3.27E-02 1.91E+00 8.84E-02
0.5000 1 6 8.88E-04 2.22E-02 2.19E-03
0.5000 1 7 1.27E-04 2.24E-04 2.84E-04
0.5000 1 8 2.93E-06 6.83E-06 8.15E-06
0.5000 1 9 1.94E+00 1.02E-02 2.71E-01
0.5000 1 10 2.89E-02 1.31E-03 6.47E-02
0.5000 1 11 3.25E-04 7.79E-03 5.95E-04
0.5000 1 12 2.44E-05 8.00E-06 5.31E-05
1.0000 2 1 5.60E-01 2.26E-01 1.27E-01
1.0000 2 2 1.25E+02 2.32E+02 7.04E+03
0.7500 2 1 1.25E+02 2.32E+02 7.04E+03
0.6250 2 1 1.25E+02 2.32E+02 7.04E+03
0.5625 2 1 1.25E+02 2.32E+02 7.04E+03
0.5312 2 1 3.86E-01 6.06E-01 3.32E-01
... ... ... ... ... ...
... ... ... ... ... ...
0.9688 16 3 4.10E-03 1.92E-02 6.62E-03
0.9688 16 4 7.84E-05 4.16E-04 1.37E-04
0.9688 16 5 9.70E-06 4.13E-06 1.67E-05
1.0000 17 1 3.58E-02 5.91E-03 2.16E-04
1.0000 17 2 4.49E+00 7.24E-01 6.56E+00
1.0000 17 3 3.37E-03 1.27E-02 5.40E-03
1.0000 17 4 6.27E-05 2.93E-04 1.08E-04
1.0000 17 5 7.94E-06 2.83E-06 1.34E-05
446 MD Demonstration Problems
CHAPTER 24

Input File(s)

File Description
nug_24_1.dat Friction coefficient 0.05
nug_24_2.dat Friction coefficient 0.15
nug_24_3.dat Friction coefficient 0.25

Video
Click on the image or caption below to view a streaming video of this problem; it lasts about 25 minutes and explains
how the steps are performed.

Figure 24-2 Video of the steps above