110306-1601 EST
panthripu:
You have indicated you are an engineer. Somewhere in the classes you have taken there should have been presented some basic discussion of the basic electrical elements --- resistance, capacitance, and inductance.
A basic capacitor is composed of two separated conductive elements (often times called plates). The media between these conductive elements relative to capacitance has a characteristic called dielectric constant. This is a number that relates to the media compared to air as a reference. So if you had a capacitor with air as the dielectric and in place of the air substituted Mylar, and made no other changes to the capacitor, then the capacitance of the capacitor would increase by an amount determined by the dielectric constant of Mylar.
A real world capacitor may and does have other characteristics that are represented by inductance and resistance in an equivalent circuit. There are no real world capacitors that only have the theoretical characteristics of a capacitor.
Some capacitors are more stable than others, and more closely approximate an ideal capacitor than others. Polystyrene and mica capacitors are quite close to ideal in certain frequency ranges. Electrolytic capacitors are mostly not very ideal, but still are very useful.
The measurements you are making are probably not very useful unless you note some very large disparity from one time to another.
Suppose you have a ceramic insulator. At 60 Hz and room temperature this will have some value of capacitance and shunt resistance. The capacitance won't be large, and the shunt resistance (leakage resistance) should be very large. Surface contamination should have negligible effect on capacitance, but possibly a very large effect on shunt resistance.
I believe your real concern is how can you do PM checks with the system in operation. How to predict potential failure? I do not think that in general in your field of interest you will have a failure relating to the capacitance of a component. Rather somehow there is a failure of the dielectric media or some sort of surface contamination. Dielectrics fail as a result of high electric field intensity. In an air capacitor when you reach a certain voltage the air breaks down and there is an arc. Remove the current and the arc is extinguished. Reapply a lower voltage and the capacitor dielectric has repaired itself (became deionized), and it is as good as before the breakdown if the arc did not damage the plates. Most other insulators are not self healing. High voltage over the surface or sharp points may cause radio noise that would be a useful detector.
I think your real question needs to be what are the best techniques to use to predict potential failure.
.