Perhaps the most perfect of all; flow technologies is that of positive displacement.
The only way for flow to pass through the meter is by causing registration, the movement of the elements and there is no open flow path between inlet and outlet. The swept volume is dictated by the geometry which does not vary with flow rate nor any other effect. Theoretically they are potentially perfectly accurate and perfectly linear at any achievable flow rate from zero to infinity.
In practise most, but not all have working clearances that allow a proportion of the flow to pass through the clearances without causing registration.
This “slip flow” is variable dependent on the flow path and flow rate or pressure drop. It is only through the impact on slip flow that anything can affect the performance (bar temperature and pressure, easily correct for).
Sliding vane and oval gear meters, perhaps the least well performing PD meters are expected to deliver (+/-)0,1% of reading over 80/100:1 turndowns.
The Brooks bi-rotor is one of the most accurate due to the sequencial clearances which limit slip flow significanlt and I have seen claims that they can achieve accuracies of (+/-)0,01% of reading, though it was not said over what flowrange.
The PD principle is used to calibrate all flow technologies, Ball and Piston provers for fiscal applications calibrating PD meters, multi chord US, coriolis, Turbine etc. In gas metering it is the bell prover and for low flows of liquids and gases the flying meniscus and flying bubble methods.
Another very accurate meter is the reciprocating piston meter which has flexible seals between the pistons and the cylinders virtually eliminating slip flow. These are used, for example when filling vehicles with petrol or diesel and have proven resistant to all attempts by modern flow meters such as coriolis to replace them because of their accuracy and robust calibration i.e. lack of calibration drift over time with wear.