Time domain detection of shocks and impacts in whole-body vibration.

A method for detecting shocks and impacts in whole-body vibration time histories has been developed that is suitable for implementation as a computer algorithm. The procedure consists of comparing the magnitudes of a higher-order mean value and the impulsiveness calculated for successive time segments of the acceleration-time history. The indicators were the ratio of the 12th-order root mean value to the root mean square RMT/RMS, and the impulsiveness corresponding to a cumulative probability value of 0.97, I(0.97) (i.e., the magnitude of the positive and negative excursions exceeded 3% of the time divided by 2RMS). Both indicators have a value of 2.16 for random vibration with a Gaussian amplitude distribution, and deviate from this value when the motion possesses other characteristics. For seat motion in the Z-direction analyzed using frequency weighting W(b), and time segments of ~20 s, shocks and impacts could be identified when RMT/RMS ≥ 2.5, and I(0.97) ≤ 2.6. A subjective visual classification of 160 exposures to vibration recorded in a range of military vehicles operating under different conditions was performed by a jury of two observers. The subjective classification agreed with computer identification of shocks and impacts in 94% of the cases.