With a PID controlling the valve repeatability, as far as the valve being in a certain position for a given output is somewhat moot as the position will adjust to obtain the desired flow due to the integral derivative calc, as do fouling and the cv of the block being tested. Repeatability will depend on only the accurcay of the flow meter and loop tuning. An older card which I might be able to "procure" should be able to use a digital pot as a setpoint. Using one of these controllers you could output the flow to another output 0f 0- 10v (easy to tailor output) and monitor it w/ LabView. Resolution is only 8 bit though (meaning only 255 counts), control is via PI...not as nice as PID but they control well. Pband as well as integral are adjustable and the loop can be tuned "live". The base and gain of the inputs are adjustable as well making the inputs pretty much universal, accepting 0-5v, 4-20ma (2-5v w/ a 250Ώ resitor across the inputs) or a resistance input. Wigits you might be interested in exist
here in calouge/ section/ output transducers they are fairly resonable consiring the prices asked for this type of stuff. The rtd's / thermistors etc are probaly not acurate enough for your aplication though.