filtering data by removing a specified frequency (noise)

3 views (last 30 days)
Hi, I have Pressure-crank angle degree data for an Internal Combustion Engine, before combustion starts, the curve is realistic and smooth, after the combustion starts, a pressure wave, of known frequency, travelling through the combustion chamber resulting in disturbing the original pressure increase curve, how can I remove this known frequency from my data ? I tried with fft, butter, and filter but I couldn't done, any kind of help will be greatly appreciated, thanks in advance,
  2 Comments
Image Analyst
Image Analyst on 7 Feb 2014
Some screenshots would help. Like, what does your data look like normally, and with the pressure wave rippling though it? And what do the spectra look like for those two cases?
Mamdouh
Mamdouh on 7 Feb 2014
Edited: Mamdouh on 7 Feb 2014
thanks a lot for your concern and reply, I attached my case, the horizontal axis is the crank angle degrees (theta), while the vertical is the pressure in (bar), the engine rotates with 5000 revolutions per min., the disturbance frequency is 252 kHz, starts with the ignition, causing the saw teeth shape, and I wanna remove it, the regular shape is almost similar but without this oscillation,

Sign in to comment.

Accepted Answer

Image Analyst
Image Analyst on 7 Feb 2014
Edited: Image Analyst on 28 Feb 2014
Take the Fourier Transform. Look for the spike at that frequency (252 kHz) and set it to zero. Then inverse transform.
Or, alternatively, you could try a Savitkzy-Golay filter (sgolayfilt in the Signal Processing Toolbox) for a one-liner solution. The SG filter is a sliding polynomial filter - it replaces the element at the center of the sliding window by the polynomial fit, at that location, of the data inside the window. (Demo available upon request.)
  6 Comments

Sign in to comment.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!