A design-of-experiments approach to quantifying test-to-test variability for a modal test

Design and development decisions are often based on the predicted and/or measured values of the frequencies of the vibration modes of a vehicle. It is known that test variability can affect our ability to experimentally measure these frequencies. Consequently, it is important to understand and quantify typical test-to-test variations. One method to isolate test variability effects is to use a formal Design of Experiments (DOE) approach. This approach prescribes the number of tests which must be conducted and the number of vehicles that need to be tested and is based on some initial estimates of variation between tests and between vehicles. It was found that in order to practically implement this approach for vehicle modal studies, a test procedure which minimizes test variability needed to be developed and used. A study was conducted to determine the test-to-test variability in vibration response for a typical mid-sized automobile. The purpose was to validate the DOE approach along with the test procedure. Based on the outcome of the DOE, seven different vehicles were tested in a laboratory setting using the test procedure that was developed to minimize test-to-test variability. Each vehicle was then tested nine different times. The results of this study are presented in the following ways: 1) Average frequency response functions (FRF) are examined for the total vehicle as well as for individual components.; 2) Four vehicle modes of vibration were identified between 19 and 25 Hz. Typical values for the test-to-test variability were calculated for the modal parameters (frequencies and damping ratios) corresponding to these modes. It is observed that the magnitude of the test-to-test variability depends on the vehicle being tested and the mode of vibration.