Part 8 · Chapter 29 · Finale

Asset Management and Smart HV Systems

The course began with a single electric field stressing a sliver of insulation. It ends here, with a continent of equipment — transformers, lines, cables, switchgear — that must be kept alive and economical for decades. This closing chapter zooms all the way out: from the physics of one breakdown to the stewardship of a whole ageing fleet, and to the digital revolution now transforming it. Sensors, data, machine learning and the digital twin are turning the dark art of "when do we replace it?" into a measured, risk-based science — the modern face of everything this course has taught.

High-Voltage Engineering Prof. Mithun Mondal Reading time ≈ 45 min
i What you'll learn
  • The shift from component to fleet thinking, the asset lifecycle and the bathtub failure curve.
  • The spectrum of maintenance strategies from reactive through time- and condition-based to risk-based.
  • The digital substation — IEC 61850, the process bus, merging units and non-conventional instrument transformers.
  • Data analytics, machine learning and the digital twin, with prognostics and remaining-useful-life estimation.
  • The asset health index and the risk picture \(R = \text{PoF}\times\text{CoF}\) that drives run / refurbish / replace decisions.
  • The wider smart grid, the sustainable future of HV engineering, and a synthesis of the whole course.
Section 29-1

From Component to Fleet

Every chapter so far has examined a thing — a gap, a dielectric, an arrester, a cable. Asset management looks instead at the population: the thousands of transformers, hundreds of thousands of kilometres of line, and uncounted items of switchgear that a utility must run as one fleet, extracting the most reliability and value from them across their whole lives at acceptable cost and risk. The pressure to do this well has never been greater, because much of the world's grid was built in a post-war boom and is now at or beyond its design life — a vast inventory of ageing assets that cannot all be replaced at once.

So asset management answers three linked questions for every item and for the fleet as a whole: what is its condition? (the diagnostics of Part 6), what is the risk it poses? (how likely is failure, and how bad would it be?), and when and how should we act? (run on, refurbish, or replace). The discipline turns the physics of the earlier chapters into decisions about money, reliability and safety — the point where high-voltage engineering meets management.

Section 29-2

The Asset Lifecycle

An asset is managed from cradle to grave. It is specified and designed, manufactured and tested (the routine and type tests of Part 6), installed and commissioned, then enters the long phase of operation and maintenance, before eventually being refurbished to extend its life or retired and recycled. Decisions taken early — the insulation level, the quality bought — echo through the whole life, and the cheapest purchase is rarely the cheapest over forty years, so good practice weighs the whole-life cost rather than the purchase price alone.

specify& design make& test install& commission operate& maintain refurbish retire& recycle life extension
The asset lifecycle — from specification through manufacture, installation and the long operate-and-maintain phase to refurbishment (which loops back to extend life) and finally retirement and recycling; decisions are judged on whole-life cost

Across that life the failure rate follows the famous bathtub curve. Early on, manufacturing defects cause infant-mortality failures that fall away as weak units are weeded out; then comes a long useful life of low, roughly constant random failures; and finally the failure rate climbs as wear-out sets in — insulation ageing, contacts eroding, paper depolymerising. The whole purpose of condition monitoring and asset management is to see that wear-out climb coming and act on the right units before it turns into outages.

λ age → infant mortality useful life wear-out
The bathtub curve — early defects, a long flat plateau of random failures, then a rising wear-out tail; asset management exists to detect the wear-out climb and intervene on the right units before failures occur
Section 29-3

Maintenance Strategies

How an owner maintains the fleet has evolved along a clear spectrum of sophistication, each step using more information to time intervention better.

increasing data, sophistication & value → reactive (run to fail) time-based (preventive) condition-based (predictive) risk-based (RCM)
The maintenance spectrum — from reactive run-to-failure, through calendar-based preventive and condition-based predictive maintenance, to reliability- and risk-based strategies that direct effort where it most reduces risk

Reactive maintenance fixes things when they break — cheap until the break is catastrophic. Time-based preventive maintenance services everything on a calendar — safer, but it disturbs healthy units and may still miss a sick one. Condition-based maintenance, the heart of Part 6, acts on the evidence of the diagnostics, intervening only when an asset's measured condition warrants. The most refined approaches — reliability-centred and risk-based maintenance — go further, directing scarce maintenance effort to the assets where it most reduces overall risk, accepting run-to-failure for cheap, non-critical items while lavishing attention on the critical few. The trend across the industry is steadily up this spectrum, from the calendar toward the sensor and the risk model.

Section 29-4

The Digital Substation

Condition- and risk-based maintenance run on data, and the modern grid is being rebuilt to supply it. The digital substation replaces the copper wiring of conventional control with a fibre-optic network governed by the standard IEC 61850. Measurements are taken by non-conventional instrument transformers (optical current and voltage sensors that are compact, safe and wide-band), digitised at the equipment by merging units, and streamed as sampled values over a process bus to protection and control devices, which exchange fast status messages over the network as well.

field sensors: optical CT/VT · DGA · PD · temperature (non-conventional instrument transformers) merging units · process bus IEC 61850 sampled values · GOOSE protection · SCADA · analytics · digital twin station bus maintenance & operating decisions
The digital substation — optical sensors and condition monitors feed merging units that publish IEC 61850 sampled values on a process bus; the station level fuses protection, control and analytics, and a digital twin, into decisions

The pay-off for asset management is that the condition data of Chapter 21 — dissolved gas, bushing tan δ, partial discharge, temperature — now flows continuously and natively across the same digital backbone, rather than being collected by hand on periodic visits. The substation becomes a source of rich, real-time data about its own health, the raw material for everything that follows.

Section 29-5

Analytics and the Digital Twin

Data alone is not insight. The streams from a digital substation are turned into decisions by analytics: simple trending to watch a quantity creep toward a limit, anomaly detection to flag a departure from normal behaviour, and increasingly machine learning trained on fleets of historical failures to recognise the early signature of an emerging fault. The goal is prognostics and health management (PHM) — not just diagnosing a present fault but predicting a future one and estimating the remaining useful life (RUL) of the asset, so intervention can be scheduled before failure rather than after.

The most ambitious expression of this is the digital twin: a living virtual model of a real asset or substation, continuously updated with its sensor data, that mirrors the real thing's state and can be used to ask "what if?". Run the twin forward under a forecast loading and it predicts the hot-spot temperature and the insulation ageing that loading will cause; push a hypothetical fault through it and it shows the consequence — all without touching the physical asset. The twin fuses the physics of the whole course with live data into a tool for foreseeing, rather than merely observing, an asset's future.

Section 29-6

The Health Index and Risk

To act across a fleet, the many condition indicators must collapse into something comparable between assets. The asset health index (HI) does this, aggregating the scored diagnostics — DGA, furan, tan δ, partial discharge, oil quality, age — into a single weighted number from healthy to end-of-life. But condition is only half the decision. A health index says how likely an asset is to fail (its probability of failure, PoF); the other half is how much it would matter (its consequence of failure, CoF — the cost, the outage, the safety and environmental impact). The two combine into risk:

Asset risk
\[ R = \text{PoF}\times\text{CoF} \]
PoF → CoF → low medium high A B C risk = PoF × CoF — act on the upper-right first
The risk matrix — plotting each asset by its probability and consequence of failure sorts the fleet into low, medium and high-risk zones; investment flows first to the high-risk upper-right, even if those units are not the most likely to fail

The risk matrix is what makes asset management rational. A unit may be quite likely to fail yet sit in a low-risk corner because its failure is cheap and easily covered; another may be unlikely to fail but command urgent attention because its failure would black out a city. Ranking the whole fleet by \(R = \text{PoF}\times\text{CoF}\) tells the owner where the next pound of maintenance or replacement budget does the most good — turning the diagnostics of Part 6 and the health index into a defensible, fleet-wide investment plan.

Section 29-7

The Smart Grid

Asset management is one facet of a broader digitalisation that is reshaping the whole power system into a smart grid. Beyond the substation, wide-area monitoring using phasor measurement units (synchronised across the network by GPS) watches the dynamic state of the grid in real time, catching instabilities that local measurements miss. Dynamic line rating lets a line carry more power when weather allows by sensing its real thermal state rather than assuming the worst. The flexible AC and DC devices of Chapter 26 steer power flows actively, integrating volatile renewables and HVDC links. All of this rests on pervasive sensing, fast communication and computation — and so brings a new vulnerability: cybersecurity. A grid that is controlled through its data network can also be attacked through it, so securing the digital infrastructure has become as much a part of high-voltage engineering as securing the insulation.

Section 29-8

The Future and a Synthesis

The road ahead is already visible. Artificial intelligence will increasingly read the flood of condition data, spotting faults earlier and estimating life more accurately than human analysts can. Robots and drones will inspect lines and substations that are dangerous or tedious for people to reach. And sustainability is reshaping the materials themselves: the search for replacements for \(\mathrm{SF_6}\) — an excellent insulating gas but an extraordinarily potent greenhouse gas — and for recyclable, eco-designed insulation, is now a central concern, driven by the same energy transition that is filling the grid with renewables, HVDC and the pulsed-power and PEF technologies of this part.

And so the course closes where it is fitting to close — by looking back. Every chapter, however different its subject, has been a variation on a single theme: the unrelenting contest between electric stress and the strength of an insulating medium. We met that contest first as a field concentrating on a sharp edge; then as the breakdown of gases, liquids and solids when the stress wins; as the generation and measurement of the high voltages that apply the stress; as the testing and diagnostics that judge the insulation's strength; as the overvoltages that threaten it and the coordination and protection that defend it; and finally as the modern materials, the direct-current and pulsed systems, and the smart, data-driven management that carry the whole enterprise into its future. From a single electric field to a continent of intelligently managed apparatus, it is one story — the lifelong duel between stress and strength, and the engineering that keeps strength, just, ahead.

Section 29-9

Worked Examples

1 Ranking by risk

Problem. Asset A has a failure probability of \(0.10/\mathrm{yr}\) and a failure consequence of \(\$2\,\mathrm{M}\); asset B has \(0.02/\mathrm{yr}\) and \(\$20\,\mathrm{M}\). Which carries the greater risk?

Solution. Risk is \(R = \text{PoF}\times\text{CoF}\):

Working
\[ R_A = 0.10\times 2 = \$0.2\,\mathrm{M/yr}, \qquad R_B = 0.02\times 20 = \$0.4\,\mathrm{M/yr} \]

Asset B carries twice the risk, despite being five times less likely to fail — its severe consequence dominates. Budget should favour B, the lesson the risk matrix exists to teach.

2 A weighted health index

Problem. A transformer scores (out of 10) DGA \(8\), furan \(5\), bushing tan δ \(9\), with weights \(0.4, 0.3, 0.3\). Find the health index.

Solution. The HI is the weighted sum of the indicator scores:

Working
\[ \mathrm{HI} = 0.4(8) + 0.3(5) + 0.3(9) = 3.2 + 1.5 + 2.7 = 7.4/10 \]

A health index of 7.4/10 — broadly healthy, but the low furan score (paper ageing) is the weak point to watch and trend.

3 Remaining useful life

Problem. A transformer's paper has a degree of polymerization \(\mathrm{DP} = 450\) now, falling at \(25\) per year. Taking end-of-life as \(\mathrm{DP} = 200\), estimate the remaining useful life.

Solution. Divide the DP margin by the rate of decline:

Working
\[ \mathrm{RUL} = \frac{450 - 200}{25} = 10~\mathrm{years} \]

About 10 years on the current trend — a prognostic estimate that lets the replacement be planned and budgeted rather than forced by a failure.

4 Thermal ageing acceleration

Problem. Using the IEEE ageing-acceleration factor \(F_{AA} = \exp\!\big[\frac{15000}{383} - \frac{15000}{\theta_h + 273}\big]\), find how much faster the insulation ages at a hot-spot of \(120^{\circ}\mathrm{C}\) than at the \(110^{\circ}\mathrm{C}\) rated value.

Solution. Evaluate at \(\theta_h = 120^{\circ}\mathrm{C}\) (\(F_{AA}=1\) at \(110^{\circ}\mathrm{C}\)):

Working
\[ F_{AA} = \exp\!\Big[\tfrac{15000}{383} - \tfrac{15000}{393}\Big] = \exp[39.16 - 38.17] \approx e^{0.99} \approx 2.7 \]

The insulation ages about 2.7 times faster for a \(10^{\circ}\mathrm{C}\) hot-spot rise — the quantitative form of the rule that overloading a transformer steals years from its life.

5 Availability

Problem. A breaker has a mean time between failures \(\mathrm{MTBF} = 20~\mathrm{yr}\) and a mean time to repair \(\mathrm{MTTR} = 0.05~\mathrm{yr}\). Find its availability.

Solution. Availability is uptime over total time, \(A = \mathrm{MTBF}/(\mathrm{MTBF}+\mathrm{MTTR})\):

Working
\[ A = \frac{20}{20 + 0.05} = \frac{20}{20.05} \approx 0.9975 = 99.75\% \]

About 99.75% available. Cutting the repair time (faster spares, better diagnostics) or extending the MTBF (better maintenance) both push this toward the very high availability a transmission asset demands.

Review

Chapter Summary

Component → fleet

Asset management runs a whole ageing fleet, asking each asset's condition, risk and the right time to act.

Lifecycle

Cradle-to-grave, judged on whole-life cost; failures follow the bathtub curve of infant, random and wear-out.

Strategies

From reactive through preventive and condition-based to reliability- and risk-based maintenance.

Digital substation

IEC 61850, optical sensors, merging units and the process bus stream condition data natively and continuously.

Analytics & twin

Trending, ML and the digital twin give prognostics and remaining-useful-life, foreseeing faults not just observing them.

Health & risk

The health index plus \(R=\text{PoF}\times\text{CoF}\) rank the fleet; the smart grid and sustainability shape the future.

Practice

Problems

For each item, first identify what it tests — the fleet view, the lifecycle and bathtub curve, the maintenance strategies, the digital substation, analytics, or the health-index and risk picture — then apply it. Difficulty rises down the list.

  1. Explain the difference between managing a component and managing a fleet, and the three questions asset management asks.
  2. Sketch the bathtub curve and name the three regions and their causes.
  3. Place the four maintenance strategies in order of sophistication and say what information each one uses.
  4. Two assets have PoF/CoF of \(0.05/\mathrm{yr}\)/\(\$5\,\mathrm{M}\) and \(0.01/\mathrm{yr}\)/\(\$40\,\mathrm{M}\). Rank them by risk.
  5. A transformer scores DGA \(6\), furan \(8\), tan δ \(7\) (out of 10) with weights \(0.5, 0.3, 0.2\). Find the health index.
  6. Paper DP is \(500\) now and falling \(20\)/yr; end-of-life is \(200\). Estimate the remaining useful life.
  7. Using \(F_{AA} = \exp[\frac{15000}{383} - \frac{15000}{\theta_h+273}]\), find the ageing factor at \(\theta_h = 116^{\circ}\mathrm{C}\).
  8. A line has MTBF \(15~\mathrm{yr}\) and MTTR \(0.04~\mathrm{yr}\). Find its availability.
  9. Explain what IEC 61850, merging units and the process bus contribute to asset management.
  10. Explain what a digital twin is and how it supports prognostics and remaining-useful-life estimation.
Tip — and a farewell: this final chapter rests on one equation, \(R=\text{PoF}\times\text{CoF}\), and one shift of scale: from caring for a component to stewarding a fleet, using sensors, the digital twin and the health index to act on the right asset at the right time. But step back further. The whole course has been a single contest — electric stress against insulation strength — fought across fields, gases, liquids, solids, the generation and measurement of high voltage, its testing and diagnosis, the overvoltages that attack it and the coordination that defends it, and the modern materials, DC and pulsed systems that extend it. Asset management is simply that contest watched over a lifetime and a fleet. Keep strength a margin ahead of stress, everywhere and always, and the lights stay on. That is high-voltage engineering.