Electricity Meters, how they work

[bernardgreen]

I strongly recall that the common electricity meters were affected by the power factor ( phase angle between current and voltage ) and with mechanical meters the electicity companies required that the power factor was as close to unity as possible. An example of this was a large officle block in London where the escalators were kept running day and night even when the building was closed. The power factor of the motors cancelled out the power factor of the majority of the flourescent lighting ,.I also recall chart recordors used to record difference in phase angle as proof that the consumer ( large company ) was maintaining a power factor close to unity and thus not making the meter read less than was being used.

( To save me typing I have copied this from Wikipedia )

Electromechanical meters
The most common type of electricity meter is the electromechanical induction watt-hour meter.[14][15]

The electromechanical induction meter operates by counting the revolutions of a non-magnetic, but electrically conductive, metal disc which is made to rotate at a speed proportional to the power passing through the meter. The number of revolutions is thus proportional to the energy usage. The voltage coil consumes a small and relatively constant amount of power, typically around 2 watts which is not registered on the meter. The current coil similarly consumes a small amount of power in proportion to the square of the current flowing through it, typically up to a couple of watts at full load, which is registered on the meter.

The disc is acted upon by two sets of coils, which form, in effect, a two phase induction motor. One coil is connected in such a way that it produces a magnetic flux in proportion to the voltage and the other produces a magnetic flux in proportion to the current. The field of the voltage coil is delayed by 90 degrees, due to the coil's inductive nature, and calibrated using a lag coil.[16] This produces eddy currents in the disc and the effect is such that a force is exerted on the disc in proportion to the product of the instantaneous current, voltage and phase angle (power factor) between them. A permanent magnet acts as an eddy current brake, exerting an opposing force proportional to the speed of rotation of the disc. The equilibrium between these two opposing forces results in the disc rotating at a speed proportional to the power or rate of energy usage. The disc drives a register mechanism which counts revolutions, much like the odometer in a car, in order to render a measurement of the total energy used.

The type of meter described above is used on a single-phase AC supply. Different phase configurations use additional voltage and current coils.



Three-phase electromechanical induction meter, metering 100 A 240/415 V supply. Horizontal aluminum rotor disc is visible in center of meter
The disc is supported by a spindle which has a worm gear which drives the register. The register is a series of dials which record the amount of energy used. The dials may be of the cyclometer type, an odometer-like display that is easy to read where for each dial a single digit is shown through a window in the face of the meter, or of the pointer type where a pointer indicates each digit. With the dial pointer type, adjacent pointers generally rotate in opposite directions due to the gearing mechanism.

The amount of energy represented by one revolution of the disc is denoted by the symbol Kh which is given in units of watt-hours per revolution. The value 7.2 is commonly seen. Using the value of Kh one can determine their power consumption at any given time by timing the disc with a stopwatch.

P = 3600 ⋅ K h t {\displaystyle P={{3600\cdot Kh} \over t}}

.

 

[JohnW2]

I strongly recall that the common electricity meters were affected by the power factor ( phase angle between current and voltage ) and with mechanical meters the electicity companies required that the power factor was as close to unity as possible. ... I also recall chart recordors used to record difference in phase angle as proof that the consumer ( large company ) was maintaining a power factor close to unity and thus not making the meter read less than was being used.

As I said, I don't think that is correct. I believe that you are right that monitoring of PF is required in commercial/industrial installations. However, my understanding is that the reason is not because of 'inaccuracies' in the recorded kWh (which IS correct), but that a PF appreciably deviant from unity (for any true kWh) attracts a surcharge because of the increased demand in terms of current flowing in the network As your quote from the Wikipedia says ...

... the effect is such that a force is exerted on the disc in proportion to the product of the instantaneous current, voltage and phase angle (power factor) between them.

Kind Regards, John

 

[Jackrae]

I was taught that a domestic 'watt-hour' meter is just that, and hence PF (reactive current) is not considered as far as the consumer is concerned. PF and peak load current do come into effect in industrial metering charges; hence why we used to avoid starting multiple heavy loads at the same time (to limit the peak current demand)

 

[JohnW2]

I was taught that a domestic 'watt-hour' meter is just that, and hence PF (reactive current) is not considered as far as the consumer is concerned. PF and peak load current do come into effect in industrial metering charges; hence why we used to avoid starting multiple heavy loads at the same time (to limit the peak current demand)

Indeed.

Of course, even in terms of domestic metering, the 'modern' (electronic) meters do effectively record 'both'. My meter, which I presume is typical of modern ones, records both kWh and kvarh, so that one can calculate kVAh (and hence average PF) if one wishes to. However, at least in terms of human meter readers, the supplier only records kWh, since domestic charging is solely on the basis of kWh, regardless of PF.

It's unusual, but I think that, on this occasion, bernard is just plain wrong in his beliefs.

Kind Regards, John

 

[AdrianUK]

I'm afraid my recollection is similar. The magnetic type meter integrates the in-phase components of the voltage & current over time thus automatically only registering real power. The power factor of the load is irrelevant.

This is, however, only true in domestic situations. As others state, commercial supplies have other meters which do take pf into consideration and tariffs that penalise accordingly.

The reason for the penalty is that, as the power factor becomes worse, more current is needed to do the same work. But the supply needs to be rated to take the 'apparent' current & the network losses are a function of the apparent current not just the component of the current which actually provides the Watts ('real current). Adding pf correction at the load end will improve the power factor & decrease the 'apparent' current to be closer to the 'real' current. Thus it is in the network operators interest to encourage the user to have a pf close to unity to reduces their losses. It is not because the Watt meter will underread ... it is a Wattmeter, not a current meter.

 

[SUNRAY]

Years ago we did some tests on 8ft fluo fittings and an electromechanical, we found the meter recorded considerably less without the pf cap connected.

 

[JohnW2]

Years ago we did some tests on 8ft fluo fittings and an electromechanical, we found the meter recorded considerably less without the pf cap connected.

Even if the meter were recording kVAh (which I very much doubt) what you describe is surely 'the wrong way round'? Without a PF capacitor, current (and hence VA) would be higher.

Kind Regards, John

 

[AdrianUK]

Years ago we did some tests on 8ft fluo fittings and an electromechanical, we found the meter recorded considerably less without the pf cap connected.

It might. One effect of a low, lagging, pf is that it can lower the voltage at the load. Adding capacitance to increase the pf will also act to bring up the voltage at the load terminals. The greater voltage is also seen by the 'real' component of the load which will hence pull a higher current & thus consume more real power in the process.

This is a 'classic' PF theory learning experiment with RLC loadbanks in parallel - seeing how improving the pf also leads to the R element of the load consuming more power.

 

[Jackrae]

Years ago we did some tests on 8ft fluo fittings and an electromechanical, we found the meter recorded considerably less without the pf cap connected.

That would be perfectly normal for a kW meter as most fluorescent lamp systems were highly inductive. Since the XL component was being ignored the metered watts would be less then the consumed VA. Adding capacitance would reduce the XL component and hence the metered VA would tend towards the metered watts. Or to put it another way, the metered watts would increase towards the consumed VA

 

[winston1]

That would be perfectly normal for a kW meter as most fluorescent lamp systems were highly inductive. Since the XL component was being ignored the metered watts would be less then the consumed VA. Adding capacitance would reduce the XL component and hence the metered VA would tend towards the metered watts. Or to put it another way, the metered watts would increase towards the consumed VA

Don't follow that at all. Putting a capacitor in parallel with the load will not affect the load at all. It will affect the total current as the XL current will cancel out the load current. But the metered, and actual, watts will remain the same.

 

[JohnW2]

Deleted ...pressed send prematurely

 

[JohnW2]

That would be perfectly normal for a kW meter as most fluorescent lamp systems were highly inductive.

There were/are, but that does not explain what he says that he observed.

Since the XL component was being ignored the metered watts would be less then the consumed VA.

Indeed.

Adding capacitance would reduce the XL component and hence the metered VA would tend towards the metered watts.

Yes, but VA was not (presumably) what was being metered. Watts (as kWh) would not change when the capacitor was added, since the reactive component of the current was never being metered.

Or to put it another way, the metered watts would increase towards the consumed VA

As above, metered (true) Watts would not change, no matter what one did to reactive components of the load.

As I wrote, even if he had had a kVAh meter (which I veryt much doubt), what he reported as having observed was 'the wrong way around'. Running an inductive load without a PF correction capacitor would increase, not decrease, the VA (hence kVAh).

Kind Regards, John

 

[Jackrae]

Spent a few moments drawing vector diagrams and now I have to agree. If Xl and XC in parallel circuit were identical, although current flows through both the L and C component, the consumed power, reactive and/or resistive, will be zero (assuming no R within the circuit)

 

[JohnW2]

Spent a few moments drawing vector diagrams and now I have to agree. If Xl and XC in parallel circuit were identical, although current flows through both the L and C component, the consumed power, reactive and/or resistive, will be zero (assuming no R within the circuit)

Indeed. In fact, in terms of true Watts (or kWh), any XL and/or XC are totally irrelevant, whether equal or not. True power is simply determined by R (and, of course, voltage), regardless of any L and/or C.

Kind Regards, John