It's ok. The other answers are sufficient and make sense. It was just a general question!
Thanks guys. I figured it would probably be an engineer's job. I also figured a plant would never want to shut down for a service upgrade either so that would drive them to go big.
As an engineer you start with the loads, both total and the larger ones. Also a good idea of layout, existing or planned. This step generally involves working with process engineers or sometimes equipment manufacturers. This gets into the soft side if engineering. There are some “guidelines”. For instance if a chemical engineer sizes a pump, plan on a +/-50% error in their calculations. Look at the overall design and leave between 10 and 50% extra capacity, again depending on how reliable the design data is. Be prepared to pivot multiple times. Typically they will want to put the electrical gear in a stupid and bad spot as well as change their minds twice a week. It helps to have a very high tolerance for a lack of commitment and people coming up with all kinds of craziness. You have to learn how to adapt quickly. Don’t get too hung up on perfecting any design. Flexible, modular designs are key. But as you move into construction planning the designers will want to tinker constantly (that’s what they do). That’s when you need to be firm in moving forward.
Once you get through a rough design (roughly a load list), there are certain “limits” that you should respect. For instance above roughly 200-250 HP, 208 or 230 V becomes more difficult to support. With 480 of 600 that increases to 500 HP with starters, 1000 HP with drives although I’ve had issues at 700+ HP drives. That pushes you into MV. Line losses may also be a determining factor. For total power I try to split out another transformer at 1000 kVA. Why so small? For one thing arc flash and maintaining under 40 cal in some cases and 8 cals for “every day” is easily achieved. For another we can use cheap panelboards and ILines and reduce the amount and cost of wiring to the loads. Similarly at 4160 things become a lot more challenging above about 5000 HP per load and 10 MVA overall, pushing you into the 15 kV class (12,470 or 13,500) or possibly jumping to 35,400. At this point for practical reasons you sort of have to adopt the local utility distribution designs. If you don’t and there are issues on their end, getting back online can be a long drawn out process.
You CAN buy much higher rated equipment but there are consequences. When you jump from 35 kA ratings to 65 kA ratings on breakers for instance the prices triple not just at the switchgear but on every control panel and most downstream devices. Similarly at MV staying under 10 kA short circuit and under 500 A loads (1000 A at most) is reasonably priced. Costs explode as you go above those thresholds. In years past it was very easy to figure this out because we had printed catalogs where you could look right down the Colin. Today getting pricing is like pulling teeth.
At one time all these general design guidelines were published in the IEEE Red Book for industrial plants. Similar ones are available for commercial. Now IEEE is finally revising them after decades of collecting dust but the new versions are scattered.
I tend to just do everything in HP and kVA. It’s just easier. With motors use 5x the largest plus the total HP and assume 1 kVA=1 HP (85% efficiency, a reasonable fudge factor). With under 5 loads you may want to use VD estimates and don’t use the 5x number so it doesn’t blow up sizing. This is where soft starts make good economic sense.
I’m saying this much because it gives you an idea of what goes into the decision making because it is very different from just calculating cable ampacity or making stamped drawings for permitting.
You need experience (10+ years) to be successful in herding cats and dealing with people that struggle with decision making. You sort of have to be both coach and human calculator. You have to recognize that this is all a “process” with an overall design procedure. It often drives field guys that are results oriented crazy. It takes training in coaching and project management and experience to do this.