Three Phase system question

[mariano54]

Hi everyone,

I'm not an electrical engeneering nor do I have a significant background in the field. Right now I'm trying to monitor current consumption from three phase systems. My question is: is it possible to obtain the current consumption of the whole system as a function of the current consumption of the single phases?

I should note that I have both how much L1, L2 and L3 consume and the trigonometric sum.

Thanks,
M.

 

[FOIWATER]

If the system is perfectly balanced there is information about all three phases that you could get from one single phase.

Practically the phases are never perfectly balanced, and to find the current in each phase it is necessary to use a current transformer arrangement. Zero sequence arrangements are used for fault detection or any imbalance situation which is not what you want here. A zero sequence CT is a single CT that passes all three phases through it. There should be no current induced in the CT if all three phases are balanced (hence the name)

You're looking for a residual CT arrangement. Where each of the three phases has it's own separate CT, which monitors each phase separately.

 

[Baluncore]

At any point in time the three line currents should sum to zero.
Those currents are not really meaningful without phase information.

Why are you wanting to do this?
Would you not be better monitoring the energy flow as power rather than the current?
Any circulating reactive current will corrupt your current measurements.

 

[mariano54]

Hi both, thanks for answering. So what you're saying, Baluncore, is that attempting to measure the "current consumption" of a three phase system would be meaningless and would not indicate how much the system is consuming. Instead, a more meaningful value would be the power consumption of the system?

 

[Baluncore]

Imagine you had a reactive load such as an induction motor or a capacitor. That would result in a reactive current in quadrature with the in-phase current on the phases. When you measure the current on a phase you measure the vector sum of the real and reactive current components.

To monitor power in the real world, you need to multiply the voltage on a phase by the current in that same phase. Summing, or integrating, the three products for all phases will give you the real power.

The phase voltage multiplied by the in-phase current is the product of two sine waves. That is a rectification and produces an offset proportional to the real power flowing. The sign of the offset also identifies the direction of energy flow.

The reactive component is known as VAR = Volts * Amps(Reactive). It is not real energy or power. The product of a sine wave by a cosine wave is symmetrical about zero and so has no DC offset. The reactive component of the current is therefore rejected by the multiplication process.

 

[Windadct]

Hello M... well to simplify and answer your question, as I read it...Yes. You can "obtain the current consumption (flow really ) of the whole system as a function of the current consumption of the single phases?"

If you can describe further what you are trying to do I think you get a more meaningful answer - rather than speculation.

 

[Baluncore]

Yes. You can "obtain the current consumption (flow really ) of the whole system as a function of the current consumption of the single phases?"

“Current consumption” really only has meaning in a circuit that uses a common return conductor.

Without a neutral current, the sum of the currents in the three phases will always be zero.
The relative phase of those currents to the voltage is quite unknown.

It is the direction and the magnitude of the energy flow that is important.

 

[mariano54]

Thanks to everyone for the answers. I'm working with powerbars that measure current and energy consumption from datacenters (in particular, Janitza powerabars, three-phase system).

My client would like to switch from current (which is currently measure as single phase or trigonometric sum) to power, and use that single value to represent what the device is consuming. Before making this transition though, I wanted to make sure there wasn't any other way to extract a single current value representing what the device in consuming.

 

[Windadct]

TO be clear - you are trying to measure Energy ( POWER over time) - in you OP you say current... Since power = Voltage x current - it is difficult to meaningless to measure current and try to convert to power.

If it is a three phase system - and they want accurate data, they need a system that measures V and I.

This all ties back to an accurate statement of what the client wants - the specification. Then the question needs to be can I meet this specification with x-y-z... My SWAG is this data center is using a lot of power/energy i.e. MONEY - and they do not want to spend money to measure money.

 

[russ_watters]

TO be clear - you are trying to measure Energy ( POWER over time) - in you OP you say current... Since power = Voltage x current - it is difficult to meaningless to measure current and try to convert to power.

It is very common to use a single amperage reading as a proxy for power by assuming a constant voltage and power factor. The simple reason is cost. You do a fine job of explaining why that is problematic.

If the phases are balanced and PF constant, it can work fine, but for a data center, neither of those assumptions are likely true.

 

[mariano54]

Ok, maybe I haven't been very clear. Right now, we are measuring current. The problem is, three-phase systems are giving either the single-phase amperage or the trigonometric sum, so we do not have a real value of how much current is actually going through. At this point, an idea came up: what if, instead of current, we measure power. And it can be done, that's not the issue.

My point is, before going from current to power, which we know won't be a problem per se, isn't there a way to measure a single amperage value from a three-phase system? What russ_watters is suggesting is that this option would be problematic (indeed, in almost every device, the phases are not balanced).

 

[Baluncore]

There is no single measurement you can do on three phase and neutral to measure total current.

 

[flandersflyer]

If the system is perfectly balanced there is information about all three phases that you could get from one single phase.

Practically the phases are never perfectly balanced, and to find the current in each phase it is necessary to use a current transformer arrangement. Zero sequence arrangements are used for fault detection or any imbalance situation which is not what you want here. A zero sequence CT is a single CT that passes all three phases through it. There should be no current induced in the CT if all three phases are balanced (hence the name)

You're looking for a residual CT arrangement. Where each of the three phases has it's own separate CT, which monitors each phase separately.

motors?

 

[mariano54]

Perfect Baluncore, that's what I wanted to confirm.

The problem I was facing is that some of the devices are three-phase while others aren't, so when I had to aggregate them (for example, see how much 2 single phase and 1 three phase were consuming), I couldn't just "sum" the current consumed by single and three phase systems.

That is something I can do with power.

 

[Baluncore]

The simplest single current reading would be to measure the currents in all three phases, (and neutral where used), sum them and then divide the answer by two since the current is metered both coming and going.

That takes three or four current transformers, each with a bridge rectifier on it's output to produce identically scaled DC currents. The four rectified DC outputs are then paralleled. One DC current meter can then be connected between the rectifiers paralleled + and – terminals.

Because this is insensitive to phase, it will give a total current which will include reactive circulating currents. This is the same reactive current situation as the single phase measurements you are now making.

 

[Windadct]

Do you have a budget for this? Have you looked at any of the products on the market? A real power / energy meter is hands down the best solution.

 

[mariano54]

That's exactly what I have, in particular, the meters are Janitza.

 

[Windadct]

I guess I do not understand what you are asking then. Yes you have current sensors, but it is a power meter. Perhaps a sketch of what you have installed and then which meter you want to use. Of course this is mains voltage - so we should not be advising unqualified people on what to do in this forum.

 

[jim hardy]

Mariano -
You do understand that instantaneous power is product of volts X amps, and that real power is time average of that product not the product of those averages...

your little power meter probably measures at least two phases and does a real time multiplication to give you true power, and from the sophistication of their website i'd wager it gives you power factor and Q as well..
I am accustomed to analog multiplier based meters , but these days it's probably all done in software.

So if your meter is anything like the UMG103 on Janitza's home page, study its instruction manual thoroughly and get with your electricians to accomplish a safe hookup. Maintenance men are happy to share their knowledge with anyone who asks honest questions and comes across humble & wanting to learn .

 

[mariano54]

Guys, thanks to everyone for your help. I think, however, that you are giving me more information than what I really needed.

The only thing I needed to know was wheter I could get a single amplerage value representing how much a three-phase system was consuming, and if this could be done as a function of L1, L2, L3 and L4. Just amperage, not power. This is aparently not possible.

The Janitza meters already give me voltage, power and current. Current, however, is given on the single phases and as a trigonometric sum. That's why I asked my original question. That's all I needed to know.

Again, thanks everyone for your input on this.