Most distribution networks run almost blind: there are hardly any sensors. Onto a network designed for power to flow one way have come rooftop solar, batteries and electric cars. This video summarises six technical webinars on what that has changed and on what can actually be measured at the bottom of the network.
Key points
- A low-voltage monitor at the distribution transformer answers four questions at once, which is what makes it affordable.
- Distribution transformers fail in the same ways as large ones; what is worth monitoring depends on the size of the unit.
- Tank pressure is the earliest indicator of an internal arc.
- Power quality has stopped being a specialist subject and is now part of asset management.
- The economics are not universal: they depend on each network.
The network nobody can see
The starting point of every session is uncomfortable. There is little or no real-time data: few sensors, and often no smart meters. Faults and outages cannot be located quickly, so they cost more. Assets fail without warning, and some of those failures reach the public. Hidden capacity cannot be found, so it cannot be used. And power theft cannot be identified at all.
The load stopped behaving
The planning figure for an average house is about 2.5 kW, and an average home electric-car charger draws about the same again. Public chargers run from 2.5 to 7 kW, and fast chargers up to 350 kW. Meanwhile solar and storage push power the other way through the transformer. The result is overloading, overvoltage and power quality complaints.
Behind it sits a physical problem. The grid is a constant balance between demand and generation, and renewable output follows the weather, not a dispatch instruction. Traditional plant ramps at tens of megawatts a minute; diversity across regions firms up the average, but it pushes the burden onto the network that has to move the energy, and storage and demand management cover the short dips. The sessions also cover virtual power plants and new forms of frequency control: private aggregation of what customers already own.
One box, several answers
A low-voltage monitor at the distribution transformer measures three voltages and four currents, power quality and harmonics, event waveforms captured when something happens, and the temperature at the transformer. From that it answers questions about:
- Safety: failing assets near people, and loss of neutral.
- Reliability: locating faults fast and finding incipient ones.
- Planning: real capacity, and where the constraints are.
- Hosting capacity: for chargers, solar and storage.
- Revenue: theft and illegal connections.
Answering several questions with one device is what makes it affordable.
What utilities actually asked for
The requests collected in the sessions are more practical than any specification: know which customer is on which phase before switching one off; know when a line is down before a customer rings to say so; find out how much rooftop solar the network can take; understand pillar box fires and see them coming; and find the capacity for chargers and spread them so nothing overloads. One of them is a life-safety matter: customers on life support were disconnected because nobody knew which phase they were on.
It is not the technology
Distribution transformers fail in much the same ways as large ones, but they cannot be monitored in the same way. The limit is access points, space, power, communications and mounting — in other words, the physical design of the unit itself. Failure modes may be equal; visibility is not. So what is realistic depends on the size of the unit:
| Measurement | Pole mount | Pad mount | Large |
|---|---|---|---|
| Oil level | rare | yes | yes |
| Overheating | basic | yes | yes |
| Moisture ingress | rarely practical | possible | possible |
| Internal arcing | indirect | two-gas | 3 or 5-gas DGA |
| Partial discharge | no | rarely | possible |
| Pressure build-up | rare | possible | standard |
The best strategy is not measuring everything; it is measuring what matters on that particular transformer.
Pressure and fire
Pressure inside the tank is the earliest indicator of an internal arc. Arc energy is arc voltage times arc current times time, and the time is usually the fault clearing time, four to five cycles. Arc voltage depends on the arc length and the electrode shape: an inter-turn fault is low energy, while a line-to-tank fault is the highest. The main source of oxygen for a fire is the air that enters after the tank ruptures. Real-time pressure data mitigates the risk of fire; it does not prevent it.
Power quality is no longer a specialist topic
It has become part of ordinary asset management, across four areas: voltage (swells, sags, over- and undervoltage), loading (current, kVA, peak demand), harmonics (overheating, losses, nuisance tripping) and imbalance (uneven loading, neutral current).
Is it worth it? Not the same answer everywhere
Where units are cheap and locally made, a failure is replaced, not prevented. Where they are imported, the replacement cost and the delay change the sum. Failure rates differ by an order of magnitude between regions, regulatory penalties and incentives often decide it rather than the repair bill, and the financial impact of an outage goes well beyond the replacement. Where the consequences are low, monitoring is rare, and reasonably so. The case has to be made asset by asset.
What the six sessions agree on
- The distribution network is close to blind, and that is the root problem.
- Distributed generation and charging have made the old assumptions wrong.
- The obstacle to monitoring is physical access, not technology.
- Measure what matters on that unit, rather than everything everywhere.
- Justify it on penalties, safety and capacity, not on the transformer’s price.
Related equipment
For power quality at the transformer and in the network, see our power quality analysers, including the APQM series. For the transformers themselves, see transformer testing and monitoring. To talk through a monitoring case for your network, contact our engineers.