| ||Format||Pages||Price|| |
|PDF (316K)||22||$25||  ADD TO CART|
|Complete Source PDF (2.6M)||174||$55||  ADD TO CART|
A way of applying on-line compensation for system dynamic response has been used on servohydraulic and electromagnetic single-input loading devices. While continuously measuring rig response by fast Fourier transform methods, a computer monitors the demand signal and applies to the rig a version of this, modified by the inverse transfer function. Signals generated by conventional white-noise filtering, by a digital technique which allows power spectrum and amplitude probability distribution to be chosen independently, and by random selection from joint peak-trough probability distributions have been handled by the process. A by-product of the system is the detection of changes in specimen response which may coincide with the start of macrocrack growth. Acceleration of tests by deleting small amplitude cycles from the demand history is shown to have weaknesses if elements in the joint peak-trough probability matrix are simply put to zero, and an alternative way using variable width dead zones in an off-line simulation is proposed.
fatigue (materials), computers, statistical analysis, mathematical models, fatigue tests, random inputs, simulation, digital techniques
Senior lecturer, University of Warwick, Coventry, Warwickshire
Research fellow, University of Saskatoon, Saskatoon,