In this article we look at how flow meter repeatability is determined at Titan.
Flow measurement performance critically depends on flow meter repeatability. Without a repeatable sensor you cannot have good performance. The reference that your flowmeter is calibrated against must also have excellent repeatability. The Titan Enterprises flow meter test rigs are based upon piston provers. These are similar to hydraulic rams with a linear encoder on the piston. Air is used to push the piston and the fluid is displaced through the flow meter back to the storage tank. The number of counts on the encoder for a given volume of fluid represents the repeatability of the rig – or the flow meter repeatability in layman’s terms.
In the calibration certificate the repeatability of the water draw mass and the scales together is ±0.03204453%. The manufacturer specifies the rig – or flow meter repeatability at ±0.05%. This is therefore the primary uncertainty for all other repeatability runs.
A flow meter repeatability cannot be recoded any better than this even if it is perfect. For a production flowmeter calibration we set the flow rig to only accept successive results which agree within ±0.1%. If the reading is not accepted the meter is rejected and rebuilt.
Below is a series of results showing runs to check repeatability over a number of points and demonstrating the importance of number of pulses taken.
At 500 pulses we are approaching the achievable repeatability of the flow rig. The flow rate varies a little from 0.596 to 0.604 Litres per minute but as we are recording pulses per litre this is a second order affect on a linear sensor.
This example was from an Atrato® ultrasonic sensor with a sample time of around 25 ms so trying to measure just 6 pulses will give poor repeatability as each successive measurement is likely to be slightly different.