SOLIDWORKS, Simulation

Real Story of Design Failure and Redesign Using SOLIDWORKS Simulation

By TriMech on March 20, 2015

Laird, a provider of systems, components and solutions that protect electronics from electromagnetic interference and heat, and that enable connectivity in mission critical systems through wireless applications and antenna systems, presented TriMech with a challenge to test a design and a redesign in SOLIDWORKS Simulation.  

Simulation Verification for Laird Technologies

LairdImage1

The above design failed in vibration testing due to fatigue at around 20,000 cycles; it needed to be rated for 100,000+. This particular design was to be mounted on a roadway and thus was subjected to complicated vibration profiles from traffic.

LairdImage2

To test this with SOLIDWORKS Simulation, we used linear dynamic harmonic tests to obtain the maximum possible acceleration due to natural resonance. You may remember that a Frequency study in Simulation Professional can obtain resonance frequencies, however it cannot obtain accelerations. To do a stress test, we needed the G loading on the body as it was shaking.
LairdImage3

LairdImage4Super-deformed acceleration plot

LairdImage5

Our test with harmonic analysis showed a maximum peak acceleration of over 1000 Gs. This was the peak for one node, however multiple nodes showed accelerations from 900-1100 Gs, so 1000 was picked as an average and worst case scenario. Although it is unlikely for a design to be in perfect resonance (and thus constantly weigh 1000 times normal weight in this case), it is still possible for a significant portion of the design’s life. For this reason, a worst case scenario with 1000 G loading in Y was used. Like the laboratory test above, one side was fixed, the other side was cantilever with a small weight attached to represent the light fixture the lever was holding. Note that accelerations in X and Z were very low, so they were omitted.

LairdImage6

LairdImage7  Worst case stress plot from 1000 G in the Y direction

LairdImage8
 Iso Plot (all sections above yield strength) of Yield on Worst Case in Y
 
LairdImage9
Section Plot of Yield on Worst Case Standard in Y

The worst case plot shows yield through a significant thickness of the part. However this is not enough evidence to support failure. Some materials can go into or above yield many times before failing. For this reason, the final check is a Fatigue Study.

LairdImage10

LairdImage11

LairdImage12 Total life plot

LairdImage13Section clip of total life on Standard in Y 

LairdImage14 Iso Clip (all sections below 20,000 cycles) on fatigue life of Standard in Y

Our test shows the extreme flange sections will fail at or below 20,000 cycles. Once this occurs, the structural strength of those will be compromised, and the I-shaped cross section will turn into a thin, flat section, which will fail much faster than expected. This confirms our test data. The location of the failure is in the same place at the same number of cycles as the laboratory test.

Next, Laird had us test their suggested redesign, which added more material for stability. Time was really of the essence, here. Since the prototype had failed, a new redesign needed to be drawn, tested, and then ordered at a foundry in a space of just a few days. Without SOLIDWORKS and Simulation, this would not be possible.

LairdImage25b
The redesigned part

LairdImage15

LairdImage16 
Redesign super-deformed acceleration plot

LairdImage17

The redesign shows significantly less accelerations in the previously measured location. However, for an absolute worst case scenario, the decision was made to test both designs with the same G loading to get a direct comparison of strength increase.

LairdImage18

LairdImage19Redesign worst case stress plot from 1000 G in the Y direction

LairdImage20Redesign Iso Clip of Yield on Worst Case C3 in Y

LairdImage21Redesign Section View of Yield on Worst Case C3 in Y

The stress is much better with the new design, far below yield. We will verify it can withstand 100,000 cycles with a fatigue test.

LairdImage22

LairdImage23Redesign total life plot

LairdImage24Redesign Section Clip of Fatigue Life on C3 in Y

LairdImage25 Redesign Iso Clip of Fatigue Life on C3 in Y

The redesign, even with far greater G loading than is predicted, shows less than .1% of the design will fail below 100,000 cycles. This new design is very safe, and the addition in mass is extremely low, only a few grams difference. This report was sent to Laird, and they immediately purchased the castings and tested them. The redesign went beyond 100,000 cycles with no failures.

“After we redesigned and ran simulation showing good factor of safety, we had no issues. The new unit passed the test on the first try.”

- Karl Searl, Laird

This was an excellent test as it showed not only the need for simulation tools to detect and correct potential design failure, but also a variety of simulations that perfectly replicate real life conditions. The original part showed failure exactly as in real life, and the redesign passed with an excellent factor of safety just as the software predicted.

Learn how SOLIDWORKS flow simulation helped Techsburg Inc. 
Techsburg Inc.