Tuesday 16 October – Division Progress Meeting

The minutes of the division’s progress meeting and planning session held 3-4 April 2001 were approved. As was noted for all divisions, anyone who did not receive the minutes via E-mail (Al Conle’s distribution list) or did not look at the minutes at the “fatigue.org” site did not see the minutes, as they were not posted at SAE’s web site.

ATV Status – Ric Mousseau

Ric noted that a few of the presentations in Wednesday’s technical session would provide details about the progress made in the ATV project in the last six months.

Presentation: “Vibration Theory and Practice of Failure Mode Verification Testing” – Alex Porter

FMVT is based on testing for any condition that can break the design. All load sources are input as random profiles, not profiles based on road load data or nominal operating conditions. Nominal loading tends to be a narrow representation of the potential failure sources, leading to tests that generate only the expected (biased) failure types. FMVT uses six-axis random vibration. Multiple failure modes can be found when inputting all identified stress sources such as vibration, temperature, and humidity. The presentation showed an example of testing on an instrument panel/cockpit system, using 6 axes of loading plus temperature, humidity, simulated sunlight, and actuation of duty cycle features like pedals and switches. 14 hours of testing on this system found 62 potential failure modes, of which 8 have known warranty history, and 41 probably would not have been found through conventional methods. FMVT needs a large spectrum; this need required an evaluation of different types of equipment for the loading. Capabilities in frequency and displacement had to be considered. FMVT uses chaos theory-based recursive control to achieve response spectra in frequencies far above those of the relative low “operating” control frequencies. The chaos equations of the control achieve very drastic changes in behavior from relatively small changes in coefficients, and the equations have symmetry of scale. The behavior can eventually bifurcate into non-repeating randomness. The presentation demonstrated this rapidly changing behavior via application of the equations to the creation of music. The FMVT system controls the base frequency, the rate of iteration, and the energy level.

Presentation: “Accelerating a Durability Test with Fatigue Editing” – Kurt Munson

The goals of the test-building project were to create a test schedule capable that reproduces a heavy truck’s service life in a short time, and to run a multi-axis test based on that schedule. Data collection covered 15 events and over 12 hours of customer use, utilizing 16 strain channels for durability analysis and 32 acceleration channels for test control and modeling. The actual time for the vehicle’s service life is about 12,000 hours, but the target time for the lab test was <400 hours. The data channels included vehicle accelerations for test control and strain gages for correlation. The test was to use several hydraulic actuators attached to the vehicle frame, and was to be time-based to reproduce variable amplitudes and vibration modes. The long, unedited strain histories were from 15 field events. These were the histories subjected to fatigue-based editing. The fatigue calculation parameters for editing included strain-life data provided by the steel supplier, as well as a K_f determined by back-calculation from component test data. Damage histories were determined at each critical location. The time histories were divided into frames of a size equal to the inverse of the lowest frequency of interest. The determination of frames to retain was via a Boolean OR function, i.e. if notable damage occurred at one or more locations during a given time slice or frame, the frame was retained for all channels. The retention of all channels is necessary to maintain phasing. In order to meet the test length target of 400 hours or less, a damage retention target of 40% was used, and the retained test content was repeated 2.5 times in order to return to 100% of field damage. An exceedence plot for one of the critical strain channels showed the cycles distribution of the edited data converging with that of the unedited data for strain ranges of ~700 me and higher. The frequency content before and after editing was compared to verify that proper structural excitation was maintained. The resulting lab test was successful at reproducing several field incidents.

Wednesday 17 October – Division Planning Session

Ideas and proposed studies were discussed for assessing variability in the loads of the ATV dynamic model as a function of the following variables: payload, springs, shocks, steering, speed, and the test track. What creates the worst case?

Attendees of the session re-stated the availability of Caterpillar Component Technology’s facility and equipment for bolt fatigue testing. This availability had been noted previously in a Tuesday presentation about bolt fatigue results. Please contact Dan Lingenfelser or Jeff Nash if interested in more information.

Time limitations in the session prevented discussion of additional possible project initiatives in the division. However, ideas for such initiatives, which were listed on overhead sheets during the session, are listed here:

Variability and uncertainty studies

Accelerated testing (new methods, how much acceleration is too much)

Thermal fatigue testing (and acceleration thereof)

Projects of potential cooperation with the Structural Analysis division.

-----------------------------------------

THE DIVISION IS STILL SEEKING A VOLUNTEER FOR VICE-CHAIRPERSON!

Respectfully submitted,

Paul Lubinski

Chairperson, Component Testing division

812 341 2362

paul.lubinski(at)arvinmeritor.com