Three phase imbalance detection

Go To Last Post
6 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hey chaps:

 

How you doing? I'm designing a three phase voltage imbalance detector

My problem is that it must work with a delta wired power line, so, no neutral available to take measurements

 

I have resolved that problem for a Wye configuration with neutral, but this config is giving me some

headaches.

 

I figured out this solution, that I consider somewhat "bloated", And I can't figure if it's the correct solution for the analog frontend of such a device.

In this circuit an optocoupler is used for re-referencing the three leg voltages instead a transformer. R11, D1, C1 and D4 are there to act as V5 does (voltage source)

then C12, R12 and R1 act as a modulator and bias source for U1. The output voltage is picked from R2 and re-referenced for a phase sum (not in this schematic)

Ignore the U2 and U3 circuits, I left it there for testing purposes.

 

Do you know of any other solution for voltage detection without having to resort to voltage transformers or a delta-wye transformer?

What is your advice?

 

 

Best regards

Ignacio

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

How many amps will be drawn by the 300 ohm resistors at 220V? You'll need big resistors. It pays to keep your simulations based on reality.

If we picture the three phases as a triangle, run a resistor from each point to the middle. This is your notional neutral or 0V point. You can measure the relative phase voltages from that point.

Optocouplers aren't real good at transferring analog signals. Their current transfer ratio varies from device to device, with temperature and aging.

One solution might be to use a microcontroller to measure the three phases and report via an optocoupler using uart data.
In the analog domain you could sum the three phases and measure the ripple. As the three phases become imbalanced, the ripple will increase.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
How many amps will be drawn by the 300 ohm resistors at 220V? You'll need big resistors. It pays to keep your simulations based on reality. If we picture the three phases as a triangle, run a resistor from each point to the middle. This is your notional neutral or 0V point. You can measure the relative phase voltages from that point. Optocouplers aren't real good at transferring analog signals. Their current transfer ratio varies from device to device, with temperature and aging. One solution might be to use a microcontroller to measure the three phases and report via an optocoupler using uart data. In the analog domain you could sum the three phases and measure the ripple. As the three phases become imbalanced, the ripple will increase.

 

Hello Kartman

 

thanks for your quick response!

Look, the 300 ohm resistors are there to test the generator placement in ltspice. They're not for part of the design.

I know the issues of ageing and current transfer of optocouplers, but, as I use them with very low current I considered them as an option. Didn't considered temperature variation though

 

I've just simulated the topology you suggested and it worked like a charm! In fact it was one of the first arrangements that I've tested

but it failed. Now I think that I messed with something in Ltspice related to the generators placement or internal resistance.

D1 to D3 are 1N4007. Too lazy to download and insert the model

 

best regards

Nachus

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The trick will be to tune the value of C4 so that you get an amount of AC ripple when you get an inbalance. You could use something like a cmos 555 as a comparator and monostable to give a pulse stretched output.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
The trick will be to tune the value of C4 so that you get an amount of AC ripple when you get an inbalance. You could use something like a cmos 555 as a comparator and monostable to give a pulse stretched output.

 

Hey Kartman

 

In fact I plan to rectify the points at Rsense, Ssense, Tsense and check the crossing point of the three phases with a

comparator. I do this because in absence of a phase (typically from a faulty contactor) the circuit receives a signal induced fro the

motor's remaining coils that are energized, but wit a notorious phase shift. With this method the protection is for voltage and phase imbalance.

 

I have to test how the circuit behaves thermally

 

regards

Ignacio

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If it is for motor protection, then the usual technique is to use two current transformers. That way you can determine overload and phase mismatch.

Last Edited: Fri. Nov 13, 2015 - 09:57 PM