One cannot fault a user or even the electrical system designer (for fear of starting a cycle of intermittent trips) wanting to calculate the total load various bits of equipment would place on a supply point. It is not just a simple case of using the VA markings on all pieces of equipment, and coming up with a total. As is, this total may well be beyond the limits of the circuit.
The simple rule is; There is no accurate way to add line currents as the load characteristics need to be taken into account.
The trick is to measure the whole load and not just a parts of it (either individual or sections). The reason is the power factor of various pieces varies dramatically. Direct rectified (such as a PC) can have anything from 0.6 leading while transformer based power supplies could have up to the same but lagging. When connected to the same supply they could cancel resulting in a pretty convenient unity power factor (and, as a result, a lower current than the computed maximum).
The easiest explaination is using a motor with capacitive power factor correction. A 2.3kVA motor taking 100A may do so with a displacement angle of 45 deg (i.e. DPF of 0.7). The PF correction caps will be taking 30A. The total line current could then be assumed to be 130A. In practice this is not so. In fact the total line current is only 70A (so where did 60A magically disappear to!?). This is further explained here.
The same occurs in real life where there is a large inductive load (e.g. air-con units) coupled with a capacitive like load such as energy efficient lighting. The overall current is likely to be less than the large inductive load.
There is no simple means of calculating combined loading. When it comes to domestic loads, it would be far easier to test either at the plug sockets or, better still, the mains distribution board (as this would encompass the complete load). I have yet to find a "plug in and test" device to measure the load which is why I made one for myself - see here. This is used in conjunction with my clamp-on amp meters that have both RMS and Crest Factor measurements.
When it comes to industrial situations, the only thing one can do is to add all the line currents together (in which ever manner you chose) and use this a "worst case scenario" for design. It's not the best if one is looking for the most economical transformer for the job, but at least one is catering for "future expansion".
Then, let's not forget about diversity (i.e. not all loads are drawing current simultaneously)! Using a PC as an example; The loading on the power supply can vary dramatically based on what is being done. High disk intensive activity can be up to twice as high a loading as when the PC is in screen save doing nothing (and the more modern machines could be even higher as energy saving becomes more efficient).
Then you get devices that have small no-load to full-load ratios such as little adapters with no markings; These are an unknown entity as it all depends on the efficiency of the transformers used. Some have very high magnetism losses (ever wondered why they are warm even with no load on them?) making measurement of their load characteristics the more obvious choice.
The only way to win on this one is to know the loads, their characteristics, and diversities, and then do proper waveform addition (note, not Fourier!) and come up with a proper kVA loading (not forgetting K-factor for hi-tech portions).
I still prefer measurement!
© 12.04.04 / 14.09.06