Jraef,
Waross's calculations are correct. He took the nameplate rated volts and rated amps to determine rated kva and then divided rated kw/rated kva to determine a factor of 0.75 that represents the motor's rated efficiency and power factor. Since efficiency and power factor only vary slightly when operating in the range of +/- 20% of rated power, this can be used to determine a good estimate of power output. Using Waross's calculated factor of 0.75 and the load amps of the motor:
82A * 460V * 1.73 = 62.256KVA
65.256KVA * 0.75 = 48.942KW
48.942KW / 0.746 = 65.6HP
Of course, the method you used achieves similar results and is much easier. But, I want to make sure that Waross gets credit for being right and also does not lose much time trying to figure out where he went wrong. In either case, the OP is overloading his motor and will have a reduced service life.
As a general rule, each 10C increase in operating temperature cuts the insulation life in half. If you allow an increased temperature rise of at least 10C from operating at 1.09 service factor plus the additional heating from the pwm waveform of at least another 20C and then add in a 10C fudge factor, I'd give it about a year, maybe two if he's lucky. This is why, as you stated, that no manufacturer rates motors to operate above 1.0 service factor on a VFD.
Of course, this is assuming that the motor has inverter duty winding insulation (Class F with 1000V p-p insulation). If not, all bets are off, he'll be lucky to get a few months even without running above rated load.
Waross offered a good suggestion of lowering the motor speed by about 3-4%. This should get the load amps in the range of the nameplate value. However, assuming that the motor is presently running at the rated speed that is required for the heat exchanger fan to provide proper cooling, reducing the speed will reduce the cooling and therefore reduce the effectiveness of the heat exchanger.