When Ford Motor Company experienced a part failure in its testing lab and needed to confidently ensure the same situation wouldn’t happen in the real world, it turned to Vibration Research Corporation for help
Lab technicians at a Ford Motor Company testing facility were facing a serious dilemma they couldn’t seem to solve: a fuel-rail on their 5.0-liter and 6.2-liter BOSS engines experienced several lab testing failures on the dynamometer that had never been observed in the field. These failures naturally created some concern. Was the company simply over-testing the fuel-rail or was it accurately predicting a real life future and a potentially dangerous situation?
Ford was concerned about repeated lab failures even though the product seemingly operated problem-free in the field. The company’s technicians knew this was a puzzle that needed to be solved properly analyzing failures could be the difference between a costly recall or no recall.
One of two things was flawed: either the fuel-rail or the test. In order to restore confidence in the fuel-rail and gain peace of mind, the Ford technicians needed to analyze their test to see if the issue was over-testing or a faulty fuel-rail. Accordingly, Ford contacted Vibration Research Corporation to see what resources and/or advice were available to help tackle this tough challenge.
The Vibration Research team studied the situation and ultimately recommended the Fatigue Damage Spectrum (FDS) vibration testing module, which is used specifically for testing a product for its expected/desired lifetime using the concepts of fatigue damage.
S-N Curve: the solution Part I
In order to determine whether Ford’s vibration tests were actually over-testing the fuel-rail, Vibration Research engineers first analyzed the test data. After reviewing all of the data compiled from various parts of the engine, the engineers decided to concentrate specifically on the engine head. This area provided the largest Grms values in response to the test and appeared to be putting the energy into the “crossover” component of the fuel-rail system.
However, the data from Ford had a broad range spectrum of 10-3200Hz, and using the fatigue damage import function required an 18.9 Grms test too large for the available shaker.
So, the team band-limited the data to 243-423Hz. This range produced the crossover’s most damaging resonances in response to the test 37 Grms, while only requiring a control of 2.2 Grms. Focusing on this range, the team proceeded.
The team observed the time and level in which the failure occurred and used the fuel-rail system test data to develop an S-N curve. This curve shows product failures in a plot of stress levels versus time to failure at those respective stress levels, and facilitates extrapolation and prediction. The S-N curve showed that the lab tests were significantly over-testing the fuel-rail system.
After extrapolating the S-N curve to the 2.1 Grms range, it became apparent that if the fuel-rail system vibrated at 2.1 Grms for the 243-423Hz range (which was Ford’s minimum Grms standard for survival of the fuel-rail system), some 10,000 hours of testing would be required before failure.
But the Ford lab technicians were observing failures in the lab during the 240-hour durability standard test which ran at the 2.1 Grms level. Something was amiss, and the puzzle wasn’t quite solved. The question became, what test would more realistically and efficiently determine the product’s longevity?
FDS: the solution Part II
The dynamometer that was used by Ford to test the fuel-rail system significantly over-tested the product. And it tested the fuel-rail system in a completely different kind of environment to that which it would experience in the real world. In their search for a better test than the one that over-tested the product, the technicians wondered whether to run a series of sine tests at various known resonances or whether to run a random test across a known frequency range.
Due to the significant resonances at a number of key locations on the fuel-rail, the technicians considered the possibility of testing with a series of sine tones with large amplitudes at the key resonant frequencies that had been observed in the engine and fuel-rail system.
The problem with this proposed solution was that the engineer was saddled with several sticky questions, including deciding which sine tones should be used and with what amplitudes. In addition, the possibility of the resonances on the fuel-rail shifting and changing when mounted to the engine would make it difficult to test the true resonances of the fuel-rail system. In light of these concerns, the sine test option became less favorable and was not used.
Enter Vibration Research’s Fatigue Damage Spectrum (FDS), a test module now offered under the company’s VibrationVIEW software that tests products by applying the amount of damage that they would experience during their lifetimes.
The team already had a waveform containing the engine head vibration patterns which caused significant resonances in the fuel-rail. Furthermore, the team had already developed an S-N curve, which afforded an approximation of the m value (the critical parameter in fatigue damage). With these two items, they only needed to import the band-limited waveform, set the target life to Ford’s expected/desired lifetime for the fuel-rail, and then enter their desired test duration.
Ford’s imported waveform had a Grms value of 2.2 rms, which was similar to its durability standard for the fuel-rail system (one only has to import a waveform that represents the product’s expected environmentwith an accurate m valuein order to utilize fatigue damage). The target life and test duration were set to 120 hours.
In importing a waveform that accurately represents a product’s expected vibration environmentusing the Fatigue Damage method and an accurate m based on the S-N curveand then setting the target life value to the product’s expected/desired lifetime, VibrationVIEW carries out several functions, including: calculating the damage present in the imported waveform, determining how many imported waveforms “fit” in the target lifespan, increasing the damage amount by the same factor, and calculating the PSD that corresponds with the final damage amount. Hence, the resultant test applies a product’s life-dose of damage.
With respect to the fuel-rail system, the resultant test gave the fuel-rails a life-dose of fatigue damage that they would experience in the field. This meant if the fuel-rails survived the lab test, they could be expected to survive in the field. In addition, with an accurate m value, the engineers could shorten the test duration while keeping target life the same. The FDS test would apply the same amount of damage but in a shorter amount of time without over-testing the product.
The engineers could also utilize Vibration Research’s Kurtosion Time Compression bringing the damaging, lifelike peaks back into the test which would decrease the Grms of the accelerated test without changing fatigue damage. Less Grms, same damage.
The S-N curve for the fuel-railbuilt from data and then extrapolatedindicated that the fatigue life of the fuel-rail was around 10,000 hours for a vibration test at 2.1 Grms. Ford utilized a standard that required the fuel-rail system to survive a disastrous waveform test for a minimum of 120 hours, based on the kind of environments that the engine was expected to encounter in real life. Ford also utilized a durability standard that required a product to survive 240 hours of similar testing. This is 10,000 hours compared to 240 hours, a life expectancy more than 40 times longer. The Ford laboratory tests were significantly over-testing the fuel-rail systems.
Vibration Research’s FDS facilitated the design of a new test for the fuel-rail, one that would apply a realistic life-dose of fatigue to the system. The resultant test gave engineers a realistic, 120-hour test that applied a life-dose of damage. They didn’t stop there.
They accelerated the test by reasonably reducing the test duration, which increased the Grms of the PSD while maintaining the same amount of damage. At the same time, the engineers utilized Kurtosion Time Compression, which decreased the Grms without changing the fatigue damage level.
Below from top:
Fuel-rail system;
Fuel-rail system leak after test;
Fuel-rail system leak;
Grms values for components of fuel-rail system at frequency range 243-423Hz;
S-N curve for fuel-rail system actual data
November 26, 2015
