Take out the BMS trip device, (since as you describe, its tripping.)
Then add a current clamp instead.
This could either be a controlled pass transistor.
But more efficient would be a SMPS controlled to be a current clamp.
It could be eg a LT8705 based 'buck-or-boost converter.
Or maybe some vicor module could do the trick, with some circuitry added to make it act like a current clamp.
Ill have a deeper think for you....i am now thinking of a constant off time SMPS with a fixed peak current limit, and then whenever the inductor current hits 50A, the FET turns off for an "off time", and then back on again...current rises back up to 50A...then back off again........repeat....
...This last method could be good because it would be very fast and the current clamp trip level would be bang on right away, instead of relying on an error amp output to slew into "the zone".
Actually, all that has to be respected is the 50A discharge current....so i think i would put in a parallel bank of "n" Constant-off-time peak current boost converters...with source current limit set to ~50A/n each....they woudl boost up to just 40V...so as to give a "shallow" inductor discharge. This way i woudl stop the discharge current ever going above 50A......then from the 40V otuput...i would have a Constant off time buck converter feeding 36V to the motor...(possibly a paralleled bank of such bucks)..... i woudl arrange it so that the transient response was very quick....may even do some kind of on/off control type thing, as the motor/drive will suddenly draw the 50A. (or whatever the motor/drive starting current is).
Constant off time control works well with on/off type converters (hysteretic ones) since it doesnt suffer sub harmonic oscillations.
Than again, maybe a simple dropping resistor (Big one) and a relay or FET to quickly short it out when the motor starts turning could be ok.