Sounds like you’ve got magnetic eddy current sensors or “proximitiors” in there. We used them in our turbine at RPI.
The sensor is linear over a very small range of distances. The gap voltage is at the center of the linear range. Ideally, you would put the sensor on a calibrating stand and measure the output voltage over a range of distances, spaced out a few thousandths between samples. Or you could measure the gap with feeler gauges (removing them before you make the measurement of course); it just takes longer.
Remember that the sensors are designed to work with a certain alloy of ferromagnetic metal, and they require an area on the surface for eddy currents to flow in. In other words, you can’t use a piece of aluminum, or the point of a needle, as your calibration standard. If you have a calibrated sensor already you can of course look at the graph and use the midpoint.
Practically, you can “fake” it in a turbine. Screw the sensor in as far as you can (gently), and then back it out very slowly while watching the output voltage. It will hover near its minimum value (close to 0V on a Bently Nevada sensor, I think). At some point it will begin to increase (or become more negative) as you turn it. Write that voltage down. Eventually you will reach a point where turning it out causes little or no change in voltage (near -24V again on a BN sensor). Find the midpoint between this and the other voltage and use that as your gap voltage.