Académique Documents
Professionnel Documents
Culture Documents
Advanced CAE
All contents © Copyright Ahmad A. Abbas , All rights reserved.
Table of Contents
Introduction ......................................................................................................................... 4
Analysis Information .......................................................................................................... 5
Original Model Geometry ............................................................................................... 5
Material Properties .......................................................................................................... 6
Boundary Condition ........................................................................................................ 7
Vibration Profile ............................................................................................................. 8
Original Model Results and Analysis ................................................................................. 9
Stress Results .................................................................................................................. 9
Fatigue Analysis.............................................................................................................. 9
Optimized model 1............................................................................................................ 13
First optimized model Results and Analysis ..................................................................... 14
Stress Results ................................................................................................................ 14
Fatigue Analysis............................................................................................................ 15
Optimized model 2............................................................................................................ 16
Second optimized model Results and Analysis ................................................................ 17
Stress Results ................................................................................................................ 17
Fatigue Analysis............................................................................................................ 18
Conclusion ........................................................................................................................ 19
Index of Tables
The objective of the study is to evaluate the response of a circuit board to a harsh vibration situation,
and determine the root cause of reported failures and suggest new model with suitable capability.
The circuit board under study shown in Figure 1 is part of a ground vehicle engine control box and it’s
subjected to an acceleration PSD (Power Spectral Density) profile, the aim of the study is to test the
hardware for harsh road condition qualification.
The circuit board is subjected to an intense vibration environment and the durability failures have been
reported about the screw holes. There are many overlapping vibration waves that are applied to this
component, therefore and because of the mathematical complexity of working with these overlapping
vibrations statistical random vibration was used.
A random vibration was considered since the movement of this vehicle component was a random
motion with erratic manner which contained many frequencies in a particular frequency band; with
motion nature that was not repeatable.
Statistical random vibration method is a more efficient way of dealing with random vibrations to
determine the probability of the occurrence of particular amplitudes of stresses for fatigue analysis.
The random vibration can be characterized using a mean, the standard deviation and a probability
distribution. Individual vibration amplitudes are not determined. Rather, the amplitudes are averaged
over a large number of cycles and the cumulative effect determined for this time period. This provides
a more practical process for characterizing random vibrations than analyzing an unimaginably large
set of time–history data for many different vibration profiles.
The results of this analysis the represented by Gaussian process, which are described in terms of
standard deviation of the distribution. The instantaneous acceleration will be between the +1σ and the
-1σ value 68.3 percent of the time. It will be between the +2σ and the -2σ values 95.4 percent of the
time. It will be between the +3σ and the -3σ values 99.73 percent of the time. The Gaussian
probability distribution does not indicate the random signal’s frequency content. That is the function
of the power spectral density analysis.
The original model shown in Figure 2 and Figure 3 is a small circuit board with the main thickness of
.01 m. This circuit board consist of an insulator, with threads of conductive material serving as wires
on the base of the board. The insulator may consist of one or numerous layers of material glued into a
single entity. These additional layers may serve a number of purposes, including providing grounding
to the board.
For simplification the circuit board was modeled using the two main isotropic materials in the
component assembly, the main to materials are Copper Alloy and General Purpose PBT Plastic. In
Figure 4 the material assignment of the assembly is illustrated, the Copper Alloy materials are marked
with 1 and PBT Plastic components are indicated with number 2.
1 1
2
2
2
Copper Alloy 1
The complete assembly will be assembled in the engine control box using 4 screws, as shown below:
The boundary condition is fixed, that would mean there are zero degrees of freedom at the screws
mounting locations (Surfaces).
For more advance analysis spring B.C model could be used to account for a small elasticity affect of
the screws, in this case-study the fixed support will be considered for simplification.
This system has an overall damping ratio was assumed to be 5 percent. Due to the geometrical
influence the assembly will have a uniform bases excitation restricted to only the z-axis direction.
The assembly must be capable of operating in a white-noise random vibration environment with an
input PSD level of describes in Table 3 and Figure 5 for a period of 20.0 hours.
Stress Results
Now the challenge is to determine the approximate dynamic stress and the expected fatigue life of the
assembly.
Analysis of the assembly under the given vibration profile will results in a stress contour plot shown in
Figure 6, which shows a maximum 1σ stress of 4.63 MPa and the full results is presented in Table 4.
Fatigue Analysis
For fatigue life calculation in the sample problem, root mean square (RMS) stress quantities are used
in conjunction with the standard fatigue analysis procedure. The Three-Band Technique using Miner’s
Cumulative Damage Ratio will be used for this fatigue analysis.
The first step is to determine the number of stress cycles needed to produce a fatigue failure. Since we
have 4 screw holes near to the edge of the bored, the computed alternating stress has to account for
stress concentration effects. The stress concentration factor K can be used in the stress equation or in
defining the slope b of the S-N fatigue curve for alternating stresses. For this sample problem, a stress
concentration factor K = 3 will be used in the S-N fatigue curve as it was estimated from Figure 7 and
Figure 8.
The approximate number of stress cycles N required to produce a fatigue failure in the component for
the 1σ, 2σ and 3σ stresses can be obtained from the following equation:
𝑆2
𝑁1 = 𝑁2 ( )𝑏
𝑆1
Where:
𝑆2 = 49.9 MPa (stress to fail at S1000 reference point)
𝑁2 = 1000 (𝑆1000 reference point)
𝑆1 = 4.63 (1σ RMS stress)
b (Slope of fatigue line with stress concentration K = 3 as shown in figure 9 )
Figure 9 S-N curve for PBT Plastic with a stress concentration of 1, 2 and 3
4.856
49.9
1𝜎 𝑁1 = 1000 = 1.03 ∗ 108
4.63
4.856
49.9
2𝜎 𝑁2 = 1000 = 3.6 ∗ 106
9.26
4.856
49.9
3𝜎 𝑁3 = 1000 = 5.02 ∗ 105
13.89
Node at root having maximum stress at the system’s first natural frequency of about 120 Hz thus, the
actual number of fatigue cycles (n) accumulated during 20 hours of vibration testing can be obtained
from the percent of time exposure for the 1, 2and 3values:
𝑐𝑦𝑐𝑙𝑒𝑠 3600𝑆𝑒𝑐
1𝜎 𝑛1 = 120 ∗ 20ℎ𝑟 ∗ .683 = 5.90∗ 106 𝑐𝑦𝑐𝑙𝑒𝑠
𝑆𝑒𝑐 ℎ𝑟
𝑐𝑦𝑐𝑙𝑒𝑠 3600𝑆𝑒𝑐
2𝜎 𝑛2 = 120 ∗ 20ℎ𝑟 ∗ .271 = 2.34∗ 106 𝑐𝑦𝑐𝑙𝑒𝑠
𝑆𝑒𝑐 ℎ𝑟
𝑐𝑦𝑐𝑙𝑒𝑠 3600𝑆𝑒𝑐
3𝜎 𝑛2 = 120 ∗ 20ℎ𝑟 ∗ .0433 = .376∗ 106 𝑐𝑦𝑐𝑙𝑒𝑠
𝑆𝑒𝑐 ℎ𝑟
Thus it is clear why high rate of failure were occurring in the component.
Stress Results
Analysis of optimized the assembly under the given vibration profile will results in a stress contour
plot shown in Figure 12, which shows a maximum 1σ stress of 3.11 MPa and the full results is
presented in Table 5.
The approximate number of stress cycles N required to produce a fatigue failure in the first optimized
model for the 1σ, 2σ and 3σ stresses will be:
4.856
49.9
1𝜎 𝑁1 = 1000 = 7.19 ∗ 108
3.11
4.856
49.9
2𝜎 𝑁2 = 1000 = 24.8 ∗ 106
6.22
4.856
49.9
3𝜎 𝑁3 = 1000 = 3.47 ∗ 106
9.33
Therefore for the first optimized model the damage will be:
Stress Results
Analysis of optimized the assembly under the given vibration profile will results in a stress contour
plot shown in Figure 15, which shows a maximum 1σ stress of 2.42 MPa, aslo the full results is
presented in Table 6 .
The approximate number of stress cycles N required to produce a fatigue failure in the first optimized
model for the 1σ, 2σ and 3σ stresses will be:
4.856
49.9
1𝜎 𝑁1 = 1000 = 24.3 ∗ 108
2.42
4.856
49.9
2𝜎 𝑁2 = 1000 = 86.0 ∗ 106
4.84
4.856
49.9
3𝜎 𝑁3 = 1000 = 11.7 ∗ 106
7.26
Therefore for the first optimized model the damage will be:
Thus the damage to the component will be much lower than both cases.
While fatigue life evaluation under a random process is highly complicated, Miner’s Rule provides a
reasonably good prediction. In the case-study, the safety factor of 2 calculated from structural stress
values is not adequate to ensure fatigue life of the component for the chosen environment.
When it comes to design for manufacturing, it would be recommended that the circuit board design be
changed to provide a fatigue life of approximately 40 hours, amounting to a safety factor of 2 on the
fatigue life.
Therefore, it is highly recommended to adopt the second optimization for engineering design change
purposes.
Advanced CAE
All contents © Copyright Ahmad A. Abbas , All rights reserved.