Using Digital Signal Processing (DSP) to Significantly Improve the Interpretation of ABAQUS/Explicit Results
Ted Diehl, Doug Carroll, Ben Nagaraj · 1999
A long-standing issue with Explicit Dynamics is the method’s propensity to calculate noisy solutions caused by unwanted high-frequency ringing. The relatively recent application of this technique to modeling drop and impact of personal electronic devices such as laptop computers, two-way radios, portable phones, and pagers has only exacerbated the problem. Quantities such as acceleration and contact force have been the most susceptible to noise, and to a lesser extent, velocity, strain, and stress. Traditional FEA post-processing approaches to reduce solution noise have been based on various curve smoothing/averaging methods and/or digital filtering (often applied improperly). These approaches have had limited success for the general class of elastically-dominated impact problems. As a result, the FEA community has struggled to develop meaningful correlation between transient quantities from simulations and experiments. The primary cause has been the improper recording of FEA data, which frequently results in corruption by aliasing. This paper presents proper techniques to avoid aliasing of the data output from transient solutions. Also presented are methods to accurately reduce (decimate) transient data sets to more manageable sizes, and to controllably remove solution noise utilizing digital filters. The success of these DSP techniques will be demonstrated on a very challenging “portable phone display” impact problem, correlating analytical and experimental results.