Plastic Flow in a Punch Problem 13 - 1
13 Plastic Flow in a Punch Problem
13.1 Problem Statement
Difficulties are sometimes reported in the modeling of plastic flow where large velocity gradients
exist. The use of interfaces embedded within a continuum finite-element mesh [e.g., van Langen and
Vermeer (1991)] has been suggested at locations where singular behavior is expected. The velocity
field at the corners of a punch driven into a cohesive, frictionless soil is shown to be discontinuous.
Figure 13.1 shows the problem geometry and boundary conditions. This example is used to
demonstrate that FLAC can produce accurate results without introducing interfaces, provided that
the expected singular point is not located at a gridpoint. One reason for avoiding interfaces is that,
for some cases, the internal rupture surfaces may be unknown in advance.
free surface
imposed
velocity
1 unit
1 unit
rough, rigid
punch
2 units
1 unit
Figure 13.1 Boundary conditions and dimensions for the numerical simula-
tion of a punch problem
FLAC Version 4.0
13 - 2 Verification Problems
13.2 FLAC Model
Figure 13.2 shows the grid used for the simulation. There are eight zones and nine nodes under the
punch. The properties of the material are:
bulk modulus (K) 1.66667 MPa
shear modulus (G) 1.0 MPa
density (p) 1000 kg/m3
cohesion (c) 10 kPa
friction angle (φ) 0◦
A velocity loading condition, shown in Figure 13.3, is applied. The gradual application of the
boundary velocity reduces the tendency for initial oscillation in the loading curve, but does not
affect the collapse load.
X = 1.0
Controlled Nodes
Figure 13.2 FLAC grid for 8-zone punch
FLAC Version 4.0
Plastic Flow in a Punch Problem 13 - 3
JOB TITLE : Punch Problem
FLAC (Version 4.00)
-06
(10 )
LEGEND
8-Apr- 0 7:26 4.000
step 25000
HISTORY PLOT 3.500
Y-axis :
vel_his (FISH)
X-axis : 3.000
Number of steps
2.500
2.000
1.500
1.000
4 8 12 16 20 24
+03
(10 )
Itasca Consulting Group, Inc.
Minneapolis, Minnesota USA
Figure 13.3 Applied punch velocity
13.3 Results and Discussion
The FISH function load calculates the numerical and analytical values of pressure beneath the
punch. Note that for the numerical simulation, the total pressure is taken as the sum of vertical
forces on the velocity-controlled nodes, divided by the width of the punch (unity, in this case). The
width of the punch extends to one-half the zone at which the velocity jump occurs. The load is
normalized by dividing by the cohesion, c, and the displacement is normalized by multiplying by
the factor G/c. The resulting normalized load/displacement curve is given in Figure 13.4, and the
steady-state velocity field is given in Figure 13.5. The numerical value calculated for the steady-
state load is 5.148c, which is only 0.12% in error of the exact load of (2 + π )c. The pattern of
shear strain rate is illustrated in Figure 13.6. The observed collapse mechanism is defined quite
well, even for a material that does not soften. Table 13.1 records the steady-state loads for various
mesh densities, expressed in terms of the number of FLAC zones under the punch. The accuracy is
within 3.5% when only one element represents the punch. (Recall that FLAC uses constant-strain
elements.) The overestimate of load for the finer discretizations can be reduced if the simulation is
run with zero damping or at a slower velocity.
FLAC Version 4.0
13 - 4 Verification Problems
JOB TITLE : Punch Problem
FLAC (Version 4.00)
LEGEND
8-Apr- 0 7:26
step 25000 5.000
HISTORY PLOT
Y-axis :
load (FISH) 4.000
anal (FISH)
X-axis :
disp (FISH) 3.000
2.000
1.000
1 2 3 4 5 6 7 8 9
Itasca Consulting Group, Inc.
Minneapolis, Minnesota USA
Figure 13.4 Normalized load/displacement for 8-zone punch
JOB TITLE : Punch Problem
FLAC (Version 4.00) 0.250
LEGEND
-0.250
8-Apr- 0 7:26
step 25000
-1.000E+00 <x< 3.000E+00
-3.500E+00 <y< 5.000E-01 -0.750
Boundary plot
0 1E 0 -1.250
Velocity vectors
Scaled to Max = 1.000E-05
Max Vector = 4.006E-06
-1.750
0 2E -5
-2.250
-2.750
-3.250
Itasca Consulting Group, Inc.
Minneapolis, Minnesota USA
-0.750 -0.250 0.250 0.750 1.250 1.750 2.250 2.750
Figure 13.5 Steady-state velocity field for 8-zone punch
FLAC Version 4.0
Plastic Flow in a Punch Problem 13 - 5
JOB TITLE : Punch Problem
FLAC (Version 4.00) 0.250
LEGEND
-0.250
4-May- 0 14:45
step 25000
-1.000E+00 <x< 3.000E+00
-3.500E+00 <y< 5.000E-01 -0.750
Boundary plot
0 1E 0 -1.250
Max. shear strain-rate
0.00E+00
2.00E-06
-1.750
4.00E-06
6.00E-06
8.00E-06
1.00E-05
-2.250
1.20E-05
1.40E-05
Contour interval= 2.00E-06 -2.750
-3.250
Itasca Consulting Group, Inc.
Minneapolis, Minnesota USA
-0.750 -0.250 0.250 0.750 1.250 1.750 2.250 2.750
Figure 13.6 Contours of maximum shear strain rate for 8-zone punch
Table 13.1 Steady-state punch pressures
(normalized) for various
discretizations
Punch Zones Pressure Error
1 4.969 - 3.4%
2 5.067 - 1.5%
4 5.121 - 0.4%
8 5.148 + 0.12%
16 5.156 + 0.28%
13.4 Reference
van Langen, H., and P. A. Vermeer. “Interface Elements for Singular Plasticity Points,” Int. J. Num.
Anal. Methods Geomech., 15, 301-305 (1991).
FLAC Version 4.0
13 - 6 Verification Problems
13.5 Data File “M9.DAT”
titl
Punch Problem
; measure load on moving plate
def load
sum = 0.0
loop i (1,9)
sum = sum + yforce(i,17)
end loop
load = sum / (cohesion(1,1)*(x(9,17)+x(10,17))/2.)
disp = -ydisp(1,17)*shear mod(1,1)/cohesion(1,1)
anal = (2.0+pi)
end
; gradual increase in starting velocity
def ramp
while stepping
if step <= 3000 then
ud app = 4e-7 + step *3.6e-6/3000
vel his = ud app
loop i (1,9)
yvel(i,17) = - ud app
end loop
end if
end
grid 17,24
model mohr
gen (0,-3) (0,0) (2,0) (2,-3)
prop dens=1000 bulk=1.66667e6 shear=1e6 coh=1e4 tens=1e10
mod null i=1,9 j=17,24
; boundary condition
fix x i=1
fix x i=18
fix x,y j=1
fix x i=10 j=17,25
fix x,y i=1,9 j=17
; histories
his load
his anal
his disp
his vel his
his sratio
his nstep 100
FLAC Version 4.0
Plastic Flow in a Punch Problem 13 - 7
step 25000
save m9.sav
plot hold his 4
plot hold his 1 2 vs 3
plot hold bou vel max 1e-5
plot hold bou ssr min 0 max 1.4e-5 int 2e-6 fill
ret
FLAC Version 4.0
13 - 8 Verification Problems
FLAC Version 4.0
