0
\$\begingroup\$

I have a fluid pressure sensor (blue line) whose measurements are affected by outdoor temperature (yellow line). So there is a 12 hours cyclic component (day/night) in my pressure measurements. How can I use the temperature variable to compensate (or "decycle") the pressure readings accordingly? I'd appreciate if you could direct me in the right direction, what are the mathematics concepts to look at?

enter image description here

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Unless you know what pressure you were actually supposed to be measuring, that graph isn't too useful because other things could be changing the pressure reading (like the actual pressure or whatever phenomena the pressure is reflecting) along with temperature. So you would need to connect the pressure sensor to a known, constant, quantity and measure how the reading responds as the temperature changes. Then you can start compensating. \$\endgroup\$ – DKNguyen Sep 24 '20 at 20:10
  • \$\begingroup\$ Assuming you can separate real pressure variation from sensor response variation as DKNguyen mentioned, the first thing I would do would be to plot these two variables using a scatter plot. \$\endgroup\$ – Justin Sep 24 '20 at 20:17
0
\$\begingroup\$

You can start with a regression analysis, but there's a lot of noise in the signals and you'd be better off with larger datasets and/or cleaner data.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Thank you, with about 40000 data points I have a nice quadratic fit. I use a one week moving average of temperature to calculate delta with actual temp and use the fitting equation. Seems correct so far. \$\endgroup\$ – Vinlar Sep 25 '20 at 21:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.