Classic regen overvoltage. The braking resistor testing fine with an ohmmeter does not mean it is working under load. I have seen resistors that measure correct resistance cold but go open circuit at temperature due to a cracked ceramic element. Put a clamp ammeter on the resistor leads and watch it during the return stroke. If you see zero current during decel, the resistor or its IGBT chopper has failed even though the resistance reads OK.
On the ACS880 specifically, check parameter 97.01 (DC bus voltage) in real time. If it spikes above 820V during the return stroke, the braking circuit is not absorbing enough energy. Normal operating range during regen should be 680-750V.
Before adding hardware: try reducing the return stroke speed by 10-15%. On most press applications nobody cares if the return takes an extra 50ms, and it cuts the peak regen power roughly in proportion. We had a 400T press throwing 7010 every 200 cycles and just dropping the return speed from 500mm/s to 420mm/s eliminated it completely.
If you do need to upgrade the braking resistor, size it for the actual regen energy, not just the motor rating. Calculate: E = 0.5 * J * (omega1^2 - omega2^2) + m*g*stroke. On a big press with a heavy slide the gravitational component is significant and most people undersize the resistor because they only account for the rotational energy.