Flow Rate Change

[jlada]

The client wants to capture flow rate over pressure data for a pump under test. Pressure is increased steadily over time. At some point the pump's relief valve opens up and the flow rate increases proportionately. The resulting graph shows an abrupt 'dogleg' bend of approximately 20 degrees resulting from the increased flow. The 'legs' of data both before and after are relatively straight lines.

The problem is to identify the point at which this dogleg occurs. Any ideas?

Thanks,

Jeff

[Michael Aivaliotis]

Can you include a screencapture of the graph? That would help a lot in figuring this out.

[jlada]

Michael;

I cannot present any screens yet as the system hasn't yet been commissioned. The image I'm describing is from the end user's hand drawn graph. Their people ran a few manual tests and recorded the data by hand just to verify that it was doable.

I met with my client today who described how he thought the data should be tested. The idea is to run it through a filter first to remove the noise and then through a differentiation function to identify the bends (turns out there are two 'doglegs' per test, not one). I guess the only remaining question is which filter to use.

Thanks for the response though.

Jeff :thumbup: