Computing the distances and voltages for poling, unfortunately, is configuration dependent. I have performed lots of poling operations and it is always case specific. Besides the poling itself, you must first crystallize the material properly. If you don't crystallize properly, then no amount of poling will work. There is a very good and fairly complete treatment of the poling operation contained in the text= https://www.amazon.com/Ultrasound-Piezoelectric-Physics-Engineers-Applications/dp/1793995389/ref=sr_1_3?keywords=henry+savage+piezoelectric&qid=1561056752&s=gateway&sr=8-3
You should do a hysteresis poling, monitoring the voltage and current. When the dipoles flip over, you should see a peek in the current flow, as your voltage source has to deliver charges to keep up the electric field. This should happen at a filed of 50V/µm. For a goog saturation usually a poling field of 100 V/µm has to be applied. The hysteresis and remanent displacment can then be calculated by integrating the current.
The poling voltage should be as high as possible without exceeding the breakdown voltage of the material. You know you are doing it right if you occasionally burn a part. Also, when you are using a corona poling triode, made sure that you are generating the corona at the corona electrode and not the grid. The corona generating electrode creates the charged ions and the grid insures that they reach the polymer uniformly.
There is one other item to consider. Once you reach the d33 specification for the material, you really can't get any higher. Higher voltage and higher temperatures will get you there faster, but d33 is a material property and it will not get higher. Once you achieve the d33 associated with the material, you can continue poling for days and it will not get any higher either.