I am not aware of any limit defined in law or by IET regulations as to the maximum size of motor to be started direct on line. There is a rather vague requirement no to use "equipment that would interfere with the  due supply of electricity to others" that could include voltage drop due to starting currents, but no actual limit is given.

Large motors should be soft started, but the actual size that constitutes "large" is not stated.

A lot depends on the size of the supply, the frequency of starting, and the nature of the other loads. Four starts an hour is a lot worse than once a month.

DOL starting on a 60 amp service will be far more obtrusive than the same motor on a 400 amp supply.

DOL starting on a service "from passing main" is more likely to be a problem than on a dedicated service to a pumping station.

This is as much to do with the supply capacity as it is to do with the motor.  In principle, the largest motor can be started DOL, but in the extreme case you have to phone up the national grid in advance of doing so.

(I'm only half joking, some of the test sites for high energy physics have big motor generator flywheels, that they spin up to provide some of the dip when it is time to fire the pulse. and such things are indeed only operated at pre- agreed times such that generation is not going to be over-stressed.)

I suspect that does not apply in your case. The varying house rules then are set by the appetite to suffer the lights going dim and pulling down the HV feed to an unacceptable degree in cases where the supply to site is at 11 or 33kV. Or the substation transformer regulation /droop.

If the customer knows that some of their sites have a rather droopy supply, they may require a slow start option. But there is no harm in asking.

Mike.

DOL starting on a 60 amp service will be far more obtrusive than the same motor on a 400 amp supply.

Indeed - I think it's often the DNO that impose specific limits - and those can vary depending on your particular supply situation, The last time I had anything to do with asking about a new (domestic) supply (quite a few years ago now) they were very interested in "disturbing" loads and ended up specifying various maximums for things like heat pumps and other motors.

    - Andy,

My experience entirely when I had my 3-phase supply put in. The thought of large motors made their man's ears prick up until I pointed out that a 3-phase motor draws 1/3 of the current of a single phase one. In the end, the conclusion seemed to be that on a domestic single phase supply, they don't really want more than 3 kW motors.

If you scale up to 100 A per phase in a light industrial unit, the OP's 7.5 kW seems perfectly reasonable.

A practical "rule of thumb" is something like the motor inrush should not take the supply over its maximum rating, with any other loads operating. This means that something like 3kW on a 100A sp supply with minimal other loads (domestic), about 90A maximum inrush, or that 22kW motor on a 200A 3ph supply as the only load. A star delta starter approximately halves the inrush, but a VSD is the kindest way to treat the motor and load nicely, as well as limit the supply voltage dip. DNOs would not like heat pumps but that is another story yet to raise its ugly head!

There is then the further complexity that not all motors of a given wattage have the same inrush, the effect varies with the mechanical load,  not necessarily a much higher  peak inrush current, but certainly the duration of the lights dimming etc so the effect is more problematic.  Some start and stop under full load, but most machines that cut or crush or indeed in a workshop generally , spin up and only then the mechanical load comes on. Other things like compressors and certain designs of pump actually have a rather higher mechanical load at switch on than once spinning, and there the inrush energy is much higher,

Then there is the no of starts and stops - a lift is more serious nuisance than  something that runs for hours at a time.

Mike.

In terms of specified / quantified limits and rules, I believe I'm right in saying that the limiting factor will be voltage dips on the supply and the effect that has on other (possibly future) customers*.To that end DNO's will require conformity to EREC P28/2. It is not a particularly user-friendly document, but it does quantitively address things like inrush magnitude, frequency of operation and supply impedance.

(*What happens in the customer's own network is their own shout, and what is acceptable will of course depend on the nature of their loads)

There are varying levels of detail required, which can be as little as a statement from the manufacturer to the effect that equipment conforms BS EN 61000-3-3 or BS EN 61000-3-11 and flicker isn't an issue with supplies of at most XYZ impedance (which you can then ascertain by testing), or as much as detailed network modelling, with levels of detail in between.

It is a requirement of the Distribution Code and hence implicit in supply agreements, and it is what DNOs refer to (in theory) when assessing, say, heat pumps or other disturbing loads when consent is requested to connect them. Certainly it is the stick that'll be used to beat you if there are complaints.