Open LV Trial - Making local substation data available to community groups

Look what I found, @Transparent - who knows the latest with this OpenLV project?

Just reading up on this exciting-sounding trial. Looks like they’ve made the end-of project documents available on their website here.

Had a chance to check them out yet, @Transparent

Well that’s the “glossy” end-of-project report @Jess_OVO !

About half of my own data can’t be released publicly because it either identified properties on maps belonging to Western Power or else it highlighted system weaknesses.

There were three groups in the OpenLV Project whilst funded by Ofgem. The academic studies (group-1) and all but two of the commercial studies (group-2) finished when the project ended. That was the extent of the certification for the prototype monitoring equipment across the 80 transformers.

However, Western Power have retained two monitored substations, and all fourteen of those monitored by Community Groups (group-3) are still operating. These use new equipment installed by EA Technology with fresh certification.

For interest, here’s some data from yesterday (Fri 8th Oct) and overnight to this morning:

The graph shows the current supplied by a substation from just one of its 3-phase feeds at 10min intervals.

L2 takes much less current during daytime, possibly due to more houses on that feed having PV Solar Panels. This is a clear case of phase-imbalance, resulting in losses at the transformer.

The phases gradually come back into alignment during the evening and enter balance just before 11pm. They remain in that state until 6:20 this morning (Saturday) when L3 starts rising again.

Also of interest here are regular sinusoidal current-demands. I’ve highlighted an area around lunchtime yesterday when L2 has a series of 4 cycles with a period of 25 minutes.

Such low-frequency oscillations are clearly man-made, and most likely due to hysteresis on a heating system. To have such a pronounced effect suggests either a public building or a community-run heat-pump is the cause.

Both L1 and L3 also exhibit similar cyclic current-demand, but there is little synchronisation between the phases. These are therefore most likely also due to single-phase appliances.

Losses are of interest to Distribution Network Operators who are required to reduce them under their RIIO agreements with Ofgem. In this case the losses could readily be addressed by applying storage control techniques using the Flex Platform and/or supplying the heat-pumps from off-grid batteries.

By way of comparison, here’s a graph from another local substation which also shows phase-imbalance. However, in this case Phase-L3 is permanently under-loaded compared with the other two.

These losses due to phase-imbalances are very common. Adding Low Carbon Technology such as Smart chargers and Heat pumps is currently making matters worse.

Without addressing this issue, it is unlikely that Distribution Network Operators (DNOs) will be able to meet efficiency targets under their forthcoming RIIO-ED2 agreement with Ofgem.

This is an important factor in the fight against Climate Change. Energy is a valuable commodity and we cannot afford to treat it with such profligacy.

Without an overall intelligent load-control mechanism, such as Flex or Octopus’ Kraken-flex, we cannot readily combat this needless wastage.

By way of comparison, here’s a graph from another local substation which also shows phase-imbalance. However, in this case Phase-L3 is permanently under-loaded compared with the other two.

These losses due to phase-imbalances are very common. Adding Low Carbon Technology such as Smart chargers and Heat pumps is currently making matters worse.

Without addressing this issue, it is unlikely that Distribution Network Operators (DNOs) will be able to meet efficiency targets under their forthcoming RIIO-ED2 agreement with Ofgem.

This is an important factor in the fight against Climate Change. Energy is a valuable commodity and we cannot afford to treat it with such profligacy.

Without an overall intelligent load-control mechanism, such as Flex or Octopus’ Kraken-flex, we cannot readily combat this needless wastage.

Using this as an example, can we tell roughly how many kWh are being "lost" at this substation and roughly what percentage is the loss?

Is it difficult to switch individual homes or groups of homes from one phase to another, appreciating this isn't a long term fix?

That’s a tricky question @Jeffus 

At the start of the OpenLV Trial, the prototype LV-CAP monitors had three temperature sensors, one of which recorded the transformer cooling-oil. The units installed now have no temperature monitoring.

The rise in temperature provides a reasonable estimation of the amount of energy being lost as heat for transformers operating well inside their maximum specification. Those running at 80% or more of max-load will obviously also experience a measurable temperature increase due to the throughput current.

The transformer is a multi-material device which includes the magnetic core, copper coils, coolant and outer-casing. A typical unit might have a mass in the region of 1500Kg to 3000Kg, of which half will be the mass of the windings and core.

I’ve monitored a 500kVA transformer (mass 2000Kg) enclosed within an unheated brick building, showing a temperature difference of  +15°C compared with the outside air.

At 8pm the coolant oil was a maximum of 32°C caused mainly by phase-imbalance from solar-panel operation during a sunny day, plus a small amount for the period of evening demand. At this moment the coolant temperature sensor was dislodged from its position and fell onto the concrete floor, about 1m below the sensor for the interior air temp.

The red line on the graph shows a 10°C temperature drop from the cooling oil and a further 2°C due to the floor being colder than the position of the air-temp sensor.

At an average daytime current of around 100A, this transformer was operating at about 10% of its maximum specification. The temperature rise due to the early-evening peak demand is therefore very low.

Reverse the problem for a moment and consider how much energy you would need to put into that 2-tonne transformer to raise its temperature by 10°C. That’s an indication of the losses we’re considering.

As I’m not a mathematician, I’ll also just tag @Simon1D and see if his skill-set can give us a reasonable approximation.

Thanks, @Transparent.

donning my metrologist’s hat ...

I’d go with your approach, which starts from the undeniable assertion that energy “lost” somewhere accumulates as heat, either right where it’s lost or wherever it might be transported to, e.g. via a cooling system. My knowledge of substation transformers is close to zero, but they are evidently massive (tonnes) and benefit from active cooling. That cooling system will have a rating (specified to remove heat at a rate of x kW from a source without the source temperature rising above some value T) and that coolant temperature sensor could be calibrated (provided it hasn’t fallen off) in terms of the power of that heat transfer. Making due allowance for the effects of smoothing (a few tonnes of iron should be fairly good at smoothing out fluctuations in heating power from these losses), the last element of the estimation would involve allowing for heat that would be dissipated when the phases are balanced (transformers have less than 100% efficiency at the best of times) in order to see what is the excess heat that could be attributed to dealing with phase imbalances.

I have a sneaking suspicion that I’ve added next to nothing to what @Transparent would have said anyway, but am happy to confirm. (I think this is more of a physics/measurement problem than a maths one, though I wouldn’t dodge the calculation if I’d arrived at one. I should add that, although heat and its measurement is actually what I’ve specialised in, professionally, that was on a scale many orders of magnitude smaller (mW and uW instead of tens of kW).

PS On the specific point:

Reverse the problem for a moment and consider how much energy you would need to put into that 2-tonne transformer to raise its temperature by 10°C. That’s an indication of the losses we’re considering.

That might be thought to hint at a lot of heat being needed, which is true, but once the transformer is warm, that’s it. What counts is the ongoing extraction of heat by that cooling system. And, for the question at hand, the key question is the difference between the run of the mill warming (with balanced phases) and the extra warming (due to the imbalance). Maybe worth comparing temperature logs from days of otherwise similar demand on the substation, days which differ significantly in the solar PV power fed into the grid.

Just looking in more detail at those plots, I suppose that on a sunny day, the “interior air” temperature would go up in the afternoon anyway, and that is fairly clear from the yellow line. but there is a hint of a variation in the cooling system temperature that slightly lags that variation, just as one might expect from the warming effect of an afternoon peak in the phase imbalance losses coming from solar PV fed into the grid. The fact that that’s a barely perceptible variation on an excess of the order 10 K may be a hint that it’s not toooooo bad after all, because that means that the ordinary losses are still dominant. I think.

Except, looking back further, I see that this poor transformer faces a persistent phase imbalance, so that constant background is already more than normal transformer losses.

I’m getting further and further out of my depth, here, so I’ll be off now - but thanks for the diversion from coding :-)

(At risk of boring through repetition:

metrology = the science of (usually physical) measurement

and I am a theoretical physicist who ended up as a metrologist)

Thanks for the prompt replies. 

The imbalance i can see in the graphs. Is all the imbalance lost as heat, ie the delta between the lines on the graphs? Can we see how much power is being lost simply by analysing the data in the graphs?

Do all 3 phases ideally need to be in perfect balance so there is no loss?

My questions may be showing my lack of knowledge 😊