Flow rate drift with time

[omarat]

Dear All, I have a quench loop for jet impingement project. The pump is oversized to the required flow so that I have a recirculating valve to return most of the flow back to the main water tank. The rest goes vertically upward to a turbine low flow meter and then to a planar 1mm*8 mm. As I read the flow rate by LabView I notice a drift (increase in value) of the flow rate for the same valves settings with the time. I may adjust the flow rate to have nozzle velocity of 0.65 m/s and end up with 0.85 or 0.9 m/s after 1 or 2 hrs. which is affecting my results.
Does anybody have suggestion why this is happening and how to resolve it?

 

[Jalink]

Omarat,
It could be that the water temperature increases (different physical properties) and additionally the pump, nozzle and control valeve increase in temperature(different system characteristics.

Your pump will put heat into the liquid. Slowly this will heat up the pump and parts downstream of the pump.

Normally of little effect, but as this appears to be a small delicate system the result could be large.

Secondly you recirculate, which brings the heated up liquid to the front of the pump again.