A ventilation fan company needed a direct drive motor for a new highway tunnel. The state government specifications required that the motor must continue to run in the event of a fire for at least one hour (it was long ago, and it might have been two hours). The circumstances of which were that the tunnel would have doors at each end that would close at the onset of the fire. It was certain that some of the fuel was gasoline or diesel, but I don't remember if there were any others (ie: a propane truck or chemicals).

Reactors will not really help and will increase the throughput losses in the soft starter, I would not waste time on that. Starting a spinning motor is not an issue with soft starter either. Both of these are potential issues with variable frequency drive, totally different animal.

As the Settings of Differential Protection are smaller or too smaller than the "Magnetic, or Short-time Protection", we can use smaller section of cables, because smaller section means bigger cable's impedance and smaller Earth Fault Current that should be enough to trip the Differential Protection, but personally and for many reasons, I don't prefer to use a Differential Protection with motors.

In many situations we deliberately reduce the starting voltage in order to better control the actual motor and machinery start, and there is rarely any legislation telling us how to achieve this. This could be by series-reactors, auto-transformers, or the popular, but expensive (virtual reduced voltage) star/delta start, so in some cases you can simply rely on the built-in series impedance of the supply cable.

Whether the differential protection is provided for the Feeder or for the Motor windings or both. In all cases usage of differential protection or not will not result increase/reduction in the size of the cable and/or the load current that have to pass through.

Voltage transient occurs whenever there is a sudden change in current in an inductive device. Inductors resist a sudden current change.
In electric motors this occurs at start up when the contactors close and shut down when the contactors open. Soft starts reduce the start up transient, but, not the shutdown transient. This also occurs with variable speed drives which switch the current rapidly and repeatedly.

Whenever we want to start up the motor, it will trip from 4 to 6 times. There is a need to redesign the electrical system start from MCC and motor selection. Our field (mechanical) engineer found out that, there is not enough torque which is the main reason it tripping.

In starting up the motor, it will trip from 4 to 6 times. everybody believe it should be related to viscosity of the product. But, i believe, it's down to electrical problem. Some has suggested that electrical heat tracing is the solution. Motor specification is 415V, Full load current 389A, and maximum current is 2800 A. Maybe it's down to electrical cable fault. It used for Bunkering Pump. Suction and Discharge of fuel oil. We used soft starter for method of starting, and at 1st, we've setting overcurrent setting from 3.0,3.2 until 4.0, but then, it's still tripping.

I would suggest that you look in to different ways of starting the pump motors including Star/Delta or some form of reduced voltage starter (e.g soft starter). Alternatively you could look at starting the motors on no-load, this doesn't reduce the starting current but it does reduce the time the motor takes to reach full speed.

What need to be concerned for MV motor installed in Hazardous Area (classified according to NEC 500) and to be started by soft start device? Is there any specific electrical data/information that needs to be given to motor vendor apart from the Plant Electrical system?
The motor will start as softer as lower the initial voltage is. However, the lower the initial voltage is, the lower the starting torque is. Therefore, you have to check against your driven load torque curve how low starting torque you can allow.