The Future Power Grid
A technical answer to a political problem

The grid is the constraint. Risk is the key to unlocking it.

The IEA finds that 1,200–1,600 GW of grid capacity could be unlocked worldwide on infrastructure that already exists — no new lines required. The barrier is not steel and copper. It is a decision framework. We make the case for operating the grid by quantified risk instead of a 1950s pass/fail heuristic.

Download the 2-page brief ↓ Read the white papers
● LIVE  EENS — system risk (MWh/h) Møre region, Norway · 5–6 Dec 2014
0.0258 baseline FAULT 19:32

19 hours of steadily rising risk, fully visible — then a blackout that left 150,000 people without power. The signal was there. The decision framework was not.

Unlocking grid capacity — a system-level approach

The limitation is not physical. It is operational.

Grid capacity is increasingly the main bottleneck for electrification, industrial growth and the integration of new demand. But existing technologies can already unlock significant additional capacity — only when combined within a consistent system-level framework.

That framework is built on three layers, applied in order. Each maps directly to a distinct function in how the grid is operated:

01DecisionReveal the risk
02PhysicalIntegrate capacity
03SecurityDeploy flexibility
Capacity gains are non-additive — without risk-based operation as a foundation, other solutions cannot scale at system level.

The missing piece

Today, grid-enhancing technologies, flexibility and digital solutions are deployed in isolation. As a result:

  • capacity gains remain fragmented
  • improvements are non-additive
  • system utilisation stays below real potential

The answer is not another technology. It is a structure that makes the ones we already have work together.

LAYER 01
Reveal
Decision layer

Risk-based operation (PRA / DSA) makes system margin visible and actionable — the foundation everything else builds on.

LAYER 02
Integrate
Physical layer

Grid-enhancing technologies increase available capacity — governed by the risk layer rather than static, worst-case assumptions.

LAYER 03
Deploy
Security layer

Flexibility ensures secure operation close to system limits — turning consumers into security assets under quantified risk.

GRID³  =  Reveal  –  Integrate  –  Deploy

Implications for policy and system development

  • No single technology can solve grid congestion alone.
  • Capacity gains require coordinated system operation.
  • Risk must be made explicit and measurable to unlock existing margins.
  • Digitalisation and AI enable this — but only embedded in operational decision frameworks.

Unlocking grid capacity is not a technology challenge — it is a system-integration challenge.

Technologies already exist — what is missing is a structure to make them work together.
01

The grid became the bottleneck of the energy transition.

Generation is being built faster than the network can absorb it. New lines take a decade. The queue cannot wait that long — and most of the answer is already in the ground.

2,500 GW
In connection queues globally
+50%
Grid investment needed to keep pace
1,200–1,600 GW
Unlockable on existing infrastructure
€43.5B
Annual European system value at stake

For seventy years, grids have been operated to a single rule: N-1 — survive the loss of any one component. It was a sound heuristic for the 1950s, built for a system with limited computing power, predictable load, and dispatchable generation.

That world is gone. Weather-driven renewables, ageing assets, and sudden new loads from electrification and data centres have made the grid a system of probabilities, not certainties. A binary pass/fail test cannot see the difference between a margin that is genuinely needed and one that is simply being held in reserve out of habit.

The result: networks run conservatively far below their real capability, while billions in renewable capacity wait in the queue for lines that may never need to be built. This is not due to a lack of technology, but a lack of coordinated system operation.

No single technology can relieve system-wide congestion — the gains of grid-enhancing technologies are only realised alongside calculated, risk-based approaches to operation, backed by robust forecasting. — IEA, Electricity 2026, p.67 (paraphrased)

Two independent, authoritative conclusions point the same way: the missing piece is not more hardware. It is a way to measure and act on risk in real time.

Grid capacity is constrained as much by operational rules as by physical limits.
02

It's happening on our watch.

The challenge shifts from “which technology” to “how to combine them at system level”. A practical framework for operating the grid by quantified risk, in the right order — three layers, where the first makes the other two possible.

Capacity is not unlocked by stacking gadgets. It is unlocked by knowing, at every moment, how much risk you are actually carrying — and operating accordingly.

LAYER 01 Decision Reveal →
LAYER 02 Physical Integrate →
LAYER 03 Security Deploy
LAYER 01 · REVEAL

Decision — Technical Rigour

Probabilistic Risk Assessment and Dynamic Security Assessment as the foundation. Measure expected unserved energy across every credible state, in real time. Without this layer, the gains below cannot be realised safely.

LAYER 02 · INTEGRATE

Physical — Implementation

Grid-enhancing technologies — dynamic line rating, topology optimisation, advanced power-flow control — deployed and risk-gated. Physical headroom, governed by the risk layer rather than static assumptions.

LAYER 03 · DEPLOY

Security — Policy & Flexibility

Conditional connection and demand-side flexibility as operating principles. Large flexible loads connect under interruptible terms — turning consumers into security assets and aligning regulation with reality.

03

Harmonics: a capacity problem first.

Harmonic distortion is usually filed under power quality. That ordering is backwards. Its first and largest effect is on grid capacity — which is precisely the bottleneck the IEA's Electricity 2026 identifies. Read in the right order, harmonics are a structural capacity constraint that the three-layer framework can measure and manage. Harmonics is an extension of the three-layer framework.

What should arrive — the clean wave

A smooth 50 Hz sine — the rhythm the entire grid runs on. Every transformer, motor and protection relay is built assuming the wave arrives in exactly this shape.

What actually arrives — the distorted wave

Non-linear loads pull power in sudden bursts. That bends the wave out of shape — and the distortion travels back through the grid, where it heats transformers, ages motors and loads up cables.

The usual conversation starts at power quality. The causality runs the other way ↓
First · the structural constraint

Grid capacity

MW / MVA envelope

Harmonics inject non-fundamental currents that raise the true RMS load on conductors, transformers and switchgear. This triggers thermal derating well before nameplate limits, shrinking the available MW/MVA envelope. Everything else follows from this.

Second · the consequence

Congestion & queues

Dispatch & interconnection

Once headroom is compressed at a node or corridor, dispatch margin tightens and connection requests pile up behind it. Harmonic-induced derating closes a node earlier than its nominal rating suggests — extending interconnection queues and distorting redispatch costs.

Third · the symptom

Power quality

THD / voltage distortion

THD, voltage distortion and equipment interference are real — but they are downstream symptoms. They show up on power-quality analysers and drive compliance costs, which is why they dominate operational talk. Fixing THD at the load side doesn't recover the capacity lost upstream.

Why data centres make this urgent

A data centre is the archetypal non-linear load: enormous, switched-mode, and growing fast. At scale it doesn't just consume power — it injects harmonic currents that derate the very infrastructure feeding everyone else. As regions electrify at once (EVs, heat pumps, solar, AI compute), this compounds, and the lost capacity is invisible on a nameplate.

Read through the hierarchy above, a new large load is not just a power-quality question at the point of connection — it is a capacity-and-congestion question for the whole corridor. The IEA's framing applies directly: congestion and instability cannot be managed without measuring the risk they create.

This is why harmonics belong inside the framework, not beside it. The distortion injected at one node propagates back into the wider system — a system-level risk that the same probabilistic, real-time risk evaluation is built to quantify, rank and act on.

Risk evaluation, extended

Each harmonic can be measured continuously and ranked against its safe limit — green, amber, red — at every connection point, then folded into the same capacity-risk number as everything else.

SAFEAPPROACHINGBEYOND LIMIT

Historical data is a forecasting asset. Records of rising harmonics show how derating and instability have propagated back into the grid before — turning past distortion into a forward-looking signal for where capacity risk emerges next.

01

Distorted wave

The messy signal a non-linear load sends back into the network.

→
02

Decompose it

A 200-year-old idea — the Fourier transform — splits the mess into the clean waves hiding inside.

→
03

Meet the harmonics

The fundamental (50 Hz) plus 3rd, 5th, 7th… each running at an exact multiple, each with its own damage signature.

→
04

Assign a risk level

Keep every harmonic within its safe limit — and assess any new large load before it connects.

Treated as power quality, harmonics get filtered locally and the lost capacity is never recovered. Treated as capacity risk — measured, ranked and managed within the three-layer framework — they become one more margin the grid can safely unlock. This is the natural extension developed in the white-paper series.

04

How we got here.

Two decades turning probabilistic risk analysis from a master thesis into real-time grid operation — across Norway, Iceland and the UK. The work behind the framework didn't start with a white paper. It started in 2001.

● Track record & ambition — deployments & capability over time 2001 → Europe-wide
Cumulative deployments & capability → 2004 Troll Power founded 2010 Goodtech acquires 2016 Promaps Technology founded 2021 Vysus Group 2023 Independent 2025 · RGI Award ★ THE GOAL Real-time risk, Europe-wide 2001 2008 2014 2018 2025 future
Company milestone Deployment / analysis project R&D / key moment Møre 2014 — the proof point The goal — Europe-wide real-time
  1. 2004Troll Power founded
  2. 2010Goodtech acquires
  3. 2016Promaps Technology founded
  4. 2021Vysus Group
  5. 2023Independent
  6. 2025RGI Grid Award ★
  7. GoalReal-time risk, Europe-wide
2001–2004 · Origin

A new reliability method

A user need defined at BKK Nett became a master thesis at NTNU on the reliability of protection and control equipment. The Markov–Kronecker method behind it was later included in System Reliability Theory (Rausand), taught at universities worldwide.

2009–2016 · Real time

From study tool to control room

A multi-year Statnett R&D programme took the method online. By the mid-2010s, Promaps Realtime was running in live system operation at TSOs and DSOs across Norway and Iceland — computing risk continuously, not just in retrospect.

Dec 2014 · The proof point

Why this is not theoretical

The Møre blackout showed the risk signal climbing for 19 hours before the fault. With the probabilistic framework in the control room, that trajectory would have triggered reserve alerts and preventive switching long before the cascade. The cost was real. So was the warning.

Furthermore, the risk and reliability capabilities provided snapshot analysis for maintenance- and preparedness planning. In total the reliability and risk assessment capability has a combined operational run-time of over 20 years (through several TSOs and DSOs in Norway and Iceland). Operational implementations of probabilistic risk assessment in Nordic systems have demonstrated the ability to identify elevated system risk several hours ahead of disturbances, providing operators with actionable decision support.

2025–2026 · The case

From practice to policy

Two decades of deployment now meet the IEA's Electricity 2026 finding: a thousand-plus gigawatts unlockable on existing infrastructure — if the decision framework exists. Recognised by the 2025 RGI Grid Award, the method is now the basis of this published white-paper series.

The goal · Europe-wide

Real-time risk on the large grids

The ambition is to take what has been proven on national systems and deploy it in real time across Europe's interconnected grid and other large power systems — so every operator can see, rank and act on risk continuously, and safely unlock the capacity the transition needs.

The graph above highlights the milestones. The full record — every deployment, R&D programme and key moment behind the methodology — is below.

Key moments in time · 2001–2018▸
2001User need for a new reliability methodology for protection & control units definedYngve Aabo · BKK Nett
2002Master thesis: reliability analysis of protection & control equipment in transformer stationsNTNU · A. B. Svendsen
2002Markov–Kronecker method solved by supervisor Tørris DigernesAker Elektro
2004Method included in System Reliability Theory (M. Rausand) — used at universities worldwide, incl. MIT curriculumAcademic
2009SOW for real-time risk management of power systems defined (2009–2013)J.O. Gjerde & S. Løvlund · Statnett
2013Promaps Online prototype running on Statnett's Norwegian model in near real-time (2013–2015)Statnett
2015Promaps Realtime in system operation (2015–2022)Elvia (Hafslund)
2016Promaps Realtime in system operation (2016–2021)Landsnet
2018Promaps Realtime in system operation (2018–)Tensio (NTE)
2018Promaps Realtime in system operation (2018–2022)Lede (Skagerak Nett)
2018Promaps Realtime in system operation (2018–2021)Dalane Nett
Analysis projects · 2006–2020▸
2006Lütelandet – Gjøa platformSFE
2007TFD electrification of Troll A – KollsnesStatoil Hydro
2008TFD alternative load point to Troll AStatoil Hydro
2008Electrification of Goliat FPSO, HammerfestEni
2012Risk management of Statnett PAS55Statnett
2013Reliability analysis, Steinsland transformer stationBKK Nett
2014Regularity analysis of Møre power systemStatnett
2020Regularity analysis of Haugaland K power systemHaugaland Kraft
R&D projects · 2005–2024▸
2005SognenettetStatnett SF
2005East Icelandic 132 kV — KarahnjukarLandsnet
2006West Icelandic 220 kV power systemLandsnet
2007North Icelandic 132 kV power systemLandsnet
2007West coast 300 kV power systemStatnett SF
2008Optimal risk-based substation designNational Grid UK
2009Optimal risk-based substation designLandsnet Iceland
2009–13Risk management of operation — R&D programme (multiple phases)Statnett
2011Risk analysis of flare system (phases 1–3, 2011–2013)Gassco
2013Risk management with real-time weather dataStormGeo / RFF
2014Online risk management with weather influence (phases 1–2, 2014–2015)StormGeo / RFF
2014Gassco OPRA phase 4 — overpressure risk analyserGassco
2018Promaps OPRA Kårstø simulation (phases 1–2, 2018–2019)Equinor
2021Internal development — Promaps Maintenance moduleVysus Group
2024Kollsnes multi-party projectEquinor · Gassco · Statnett
05

Publications.

A series building the technical and policy case, piece by piece. Open access.

IEA

This white paper series was submitted to the IEA in response to their findings in Electricity 2026.

Brief · 2 pages

The Future Power Grid — Decision Brief

The whole argument in two pages, written for grid and energy-policy leadership. The problem, the framework, the proof point, and a clear ask. Start here.

Download PDF ↓ WP1 · Published

Unlocking Grid Capacity Through Risk-Based Operation

A practical framework for the coordinated deployment of PRA, DSA, and grid-enhancing technologies. Svendsen, Nyiredy & Milenkovic, May 2026.

Open PDF ↗ WP2 · Published

From N-1 to Virtual N-1

Once risk-based operation is established, a natural next step is to move from implicit to explicit management of system margins — extending the framework through conditional connection. Svendsen, Nyiredy & Milenkovic, 2026.

Open PDF ↗ Reference · IEA

IEA — Electricity 2026

The underlying global research. The source of the capacity figures and the call for risk-based grid operation.

Open report ↗
Academic series · Forthcoming

Peer-reviewed series

A formal academic series extending the framework is in preparation. Details to follow.

Coming soon
Foundational record

Two decades of peer-reviewed and conference work underpinning the methodology — from the original reliability theory to online, weather-aware risk assessment.

2003
Economic benefits of maintenance methodology and probabilistic methods in high-voltage installationsCIRED, Paris
2004
Analysis including reliability, income and cost for power systemsPMAPS, Iowa
2007
Analyses of delivery reliability in power systemsESREL, Stavanger
2007
Maintenance planning based on simulation of power delivery reliability and economic consequencesDoble, Germany
2009
Power system regularity challenges connected to electrification of large-scale offshore installations from landESREL, Prague
2009
Reliability analysis including load flow and power demand in power systemsAR2TS, Loughborough
2012
Online reliability assessment of power systemsPMAPS, Istanbul
2014
3D representation of geographical power system network as a function of regularity propertiesESREL, Wrocław
2015
Online reliability calculations of power systems with forecasted and real-time weather influenceESREL, Zurich
2017
Modelling weather dependence in online reliability assessment of power systemsJournal of Risk and Reliability
2018
Digitalization of the power business: how to make this work?ESREL, Trondheim
06

Recognised by Europe's grid community.

The approach behind this work was named the 2025 Prize for Technological Innovation.

RGI Grid Awards 2025 · PCI Energy Days, Brussels

Prize for Technological Innovation

Renewables Grid Initiative

Exactly what Europe's power system needs today — a transparent tool for real-time prediction of grid risk, safely unlocking at least 25% more capacity and accelerating the transition without costly new infrastructure.

Presented by Anna Stürkgh (EU Commissioner) · Lars Aagaard (Danish Minister for Climate, Energy & Utilities)
Paraphrased from the jury citation.
Read the citation ↗
25%+
Additional capacity unlocked safely
12–48h
Risk prediction ahead of the fault
Europe-wide
Transferable across power systems
Awards & nominations

A decade of recognition for the methodology behind the framework.

2025
Winner RGI Grid Award — Good Practice of the Year Renewables Grid Initiative · European Commission PCI Energy Days, Brussels
2017
Nominated S.P.I.R Award for Promaps Realtime National prize for climate technology and renewable energy
2017
Nominated Norway's Smartest Industrial Enterprise — Promaps Realtime Norsk Industri
2014
Nominated Norway's Smartest Industrial Enterprise — Promaps Online Norsk Industri
2014
Nominated Norwegian Technology Award Tekna
07

The authors.

Three independent voices in power-system reliability and risk.

A · S

Arne Brufladt Svendsen

Power-system reliability pioneer; developer of real-time probabilistic risk methodology for complex grids.

LinkedIn ↗ arne@thefuturepowergrid.org
R · N

Robert Nyiredy

Commercial and industry lead in power-system reliability and security of supply.

LinkedIn ↗ robert@thefuturepowergrid.org
M · M

Mathieu Milenkovic

Physicist working at the intersection of power systems, probabilistic risk, and grid analytics.

LinkedIn ↗ mathieu@thefuturepowergrid.org
General enquiries info@thefuturepowergrid.org
Risk-based operation is not an option — it is a prerequisite for scaling grid optimisation.