I was thinking maybe this is caused by a bad line reactor, so after some googling I found this. As you know a GFI is looking for current leakage, so if the components test out ok, I think you need to look at how the motor is grounded or something along these lines.
PWM Drives — Measurement 5
Leakage currents (common mode noise)
Leakage currents
Leakage currents (common mode noise) capacitively coupled between the stator winding and frame ground will increase with PWM drives as the capacitive reactance of the winding insulation is reduced with the high frequency output of the drive. Another leakage current path may exist in the capacitance created when the motor cables are placed in a grounded metal conduit. Therefore, faster rise times and higher switching frequencies will only make the problem worse.
It should also be noted the potential increase in leakage currents should warrant close attention to established and safe grounding practices for the motor frame. The increase in leakage currents can also cause nuisance tripping of ground fault protection relays, override 4 to 20 mA control signals, and interfere with PLC communications lines.
Measure common mode noise by placing the current clamp around all three motor conductors. The resultant signal will be the leakage current.
A common mode choke can be used to reduce leakage currents (see Figure 5A). Also, special EMI suppression cables can be used between the drive output and the motor terminals. The copper conductors of the cable are covered with ferrite granules, which absorb the RF energy and convert it to heat. Isolation transformers on the ac inputs will also reduce common mode noise.
http://www.davis.com/techinfo/techin...rement-ASD.htm