ID #1032

Digital Signal Processing using CFD-VIEW

Signals (stream of data) obtained from physical or numerical experiments often require additional processing to understand/interpret the physics they represent.  The processing of signals through digital means is termed Digital Signal Processing (DSP).  Signals might be available in a multiplicity of domains (time, frequency, spatial…) and dimensions (1D, 2D…).  One of the most common analyses in engineering applications involves processing time domain signals (filtering, smoothing…) and transforming to frequency domain.  This note discusses some signal processing capabilities available in CFD-VIEW for time history data.

The general procedure involves 1) supplying time history data to CFD-VIEW, 2) plotting/choosing variable(s) of interest and 3) applying DSP functions.

Time history data can be made available to CFD-VIEW in multiple ways. Some options are:

  1. Using Record Point History in CFD-VIEW (typically from DTF files of a transient run)
  2. Through PAMFLOW Time History Data
  3. Import *.MON File directly into CFD-VIEW (plan in advance to setup monitor points in CFD-ACE-GUI before starting a run)


Figure 1. Providing time history data to CFD-VIEW and available DSP operations 

The DSP operation window is active when time history data are plotted (with time as the independent variable).  Available DSP operations can be seen by clicking the add (“+”) button.  For Fast Fourier Transform (FFT) operations, the signal is re-sampled by linear interpolation to provide the nearest 2n number of points.  Users also have the option to zero pad the signal to the nearest 2n number of points.  The Power Spectral Density (PSD) describes how the power of a signal is distributed within the frequencies of the spectrum and expressed in (physical units)2 / Hz.  The PSD can be normalized and expressed in dB through the use of a suitable reference value.  For example, the commonly used reference value for acoustic pressure is 2e-5 Pa.  Other time domain operations are also available to filter a signal, calculate derivatives/integrals, smooth and perform auto-correlation of a signal.

A simple example is provided in the Figure 2 to demonstrate PSD operation on a periodic signal. The signal is a sum of two waves (a sine wave of 7Hz and a cosine wave of 14 Hz) sampled at 100 Hz for 10.24 seconds providing a total of 1024 points.


Figure 2. Example showing PSD applied on a periodic signal sampled at uniform time interval

Tags: -

Related entries:

Last update: 2009-07-13 14:49
Author: ESI-CFD Support Team
Revision: 1.2

Digg it! Print this record Send to a friend Show this as PDF file
Propose a translation for Propose a translation for
Please rate this entry:

Average rating: 1 out of 5 (6 Votes )

completely useless 1 2 3 4 5 most valuable

You cannot comment on this entry