# Does it matter where we measure the Voltage in a large-scale, industrial installation?

So, let's say we have a big shop floor with several 3-phase, high power consumption machines being far apart from each other. If we measure the mains voltage at one of the plugs, could this be used to accurately calculate the consumption of each of the other machines, just by measuring their current?

What I am trying to understand is whether we can have a non-intrusive distributed power consumption monitoring system like the Open Energy Monitor, where the instantaneous voltage is measured at a single point only and then multiplied with each of the currents to get the instant power consumption.

I have the belief that in an industrial environment, this would not be accurate due to longer distances between the appliances, bigger voltage spikes and higher inductances of the transfer lines. Am I right?

It depends on the level of accuracy you require.

Off the cuff, I would expect single-point voltage measurement to introduce about 10% error on power measurements at remote points.

A properly designed LV distribution system (Australia: 230 VAC single phase, 415 VAC three-phase) should keep the steady-state voltage drop within about 10%. That means there should be no more than 10% voltage difference between the supply and the load terminals. (10% isn't a mandatory limit - more of a rule of thumb.)

Phase shift doesn't seem like it would be an issue. If in doubt, look up some typical cable impedances (as found in a standard like AS/NZS 3008, or a cable manufacturer's catalog i.e. Nexans / Olex or Prysmian). You can use the cable impedances to calculate the voltage drop and phase shift for your given load current.

At high voltage (3.3 kV, 11 kV) there really isn't much voltage drop on cable runs. I would expect single-point voltage measurement to work well here. (This is, in fact, how it's done. See below.)

With regards to voltage spikes affecting your metering: don't worry about it. You are measuring power consumption (kilowatt-hours) not instantaneous power (kW). The spikes average out over time.

For perspective, voltage transients from motor starting typically disappear within 10 seconds. There are 86,400 seconds in a day. Voltage transients won't affect your power measurements.

Note: In practice, power measurements are normally taken at the upstream end of the circuit - at the switchboard/motor control centre.

For low-voltage MCC's, each meter gets a direct connection to the phase conductors. Meters can be built to measure 240 VAC directly with no problems.

For high-voltage MCC's, the voltage is too high to measure directly. A single voltage transformer is provided (i.e. 3.3kV/110V) to scale the voltages down to a measurable level. That scaled 110V voltage is then distributed to all the meters on the switchboard.

Metering isn't done at the downstream end (where the plugs are) because that's typically an exposed environment, subject to lots of abuse and damage.

Well current will be the same at any point in the system, but voltage drop between the service and the load will affect the kW reading you take at the service vs at the load. So it depends upon the PURPOSE you have in mind for this metering. If the purpose is to determine the effect any particular load has on the financial aspect of your power consumption, then the only thing that matters is measurement taken at the same point as the utility measurement, which will be the service entrance. If you measure kW at the load, AFTER the voltage drop, then the number taken there will appear lower, mostly because it is not going to be inclusive of the ohmic heating losses in the cables (i.e. the voltage drop). But from a revenue standpoint, that machine is causing that ohmic heating loss, so it SHOULD be included in the cost of power for that machine. if the cables are sized correctly though, that number may be very small not all that significant.

• Your last sentence is my point. If it is that insignificant to be negligible in the calculations, that's perfect. I am also interested in any phase-shifting that might appear too, not solely about potential voltage drop. To answer the first bit, my purpose is to measure consumption of each individual load.
– DimP
Oct 17, 2016 at 23:07

If you want an accurate measurement of the "instant power consumption" for your scenario ("high power consumption machines being far apart from each other"), I recommend that you measure both the voltage and the current for each machine. This way you will see clearly how much each machine consumes. Also, any differences in the voltage measured at different machines might give you an indication of wire losses and potential wire improvements.

On top of that, you can also measure the voltage and current at the "entry" point. This will allow you to assess losses and do some further analysis and cross-check your meters.

• The OP wants to measure power at various points on an AC system. Simply measuring voltage and current won't allow calculation of power due to power factor and non-sinusoidal waveforms. The way to do it is by using proper energy meters which integrate the product of voltage and current over time. Remember that you can edit your post at any time to improve it. Jun 24, 2021 at 22:19
• I was assuming that the OP is measuring instantaneous values of voltage and current, not averages (RMS) as "Transistor" assumed in his comment. With instantaneous values of voltage and current, one can calculate the power at any given time, hence the energy consumption over time. The OP never mentioned non-sinusoidal waves - although the concepts discussed here (where to measure) would still apply. Since the OP question was not about HOW to measure Power given a specific harmonic regime, my answer focused only on: 1. What to measure; 2. Where to measure. Jun 25, 2021 at 22:23