Grid-enhancing technologies

GETs are digitally enabled tools that increase the capacity, flexibility and efficiency of existing networks without waiting for major new infrastructure.1 As shown in the examples in Chapter 1, they can deliver substantial gains when they address a specific constraint. Their safe and scalable use, however, depends on co-ordination across planning, operational planning and real-time operations, and integration within network companies’ processes.

In practice, GETs and AI are increasingly complementary. GETs create new controllable options for operators, for example by revealing dynamic thermal headroom, redirecting flows or identifying alternative topologies, while AI can improve the forecasts, screening and decision support needed to use those options more consistently.

A decade of deployment shows that GETs can reliably unlock substantial capacity from existing networks. Typical installation is measured in months and costs an order of magnitude less than building new lines, making GETs an important complement to grid buildout, able to act on a shorter timescale. Industry coalitions and specialist initiatives, such as the Global Power System Transformation Consortium (G-PST), CurrENT in Europe, the Utilize Coalition in North America and the WATT Coalition, are also helping build the evidence base, standardise use cases and share deployment experience across operators, regulators and technology providers.

GETs deployment remains geographically concentrated. In the case of DLR, for example, three-quarters of deployment is concentrated in Europe and North America. There are several reasons for this pattern, including congestion levels, the prevalence of older grid assets that make optimisation more valuable relative to new build, and the maturity of the regulatory frameworks needed to credit these technologies in planning and markets. However, the relationship is not purely a function of grid age: some fast-growing systems face similar congestion pressures and deploy these technologies for different reasons.

Dynamic line rating and dynamic transformer rating

Dynamic ratings reveal capacity that conservative, fixed assumptions leave unused. The capacity available between any two points on the network is set by the most limiting piece of equipment along that path – lines, transformers or substation equipment – so unlocking capacity generally requires rating across the whole path. Utilities have long used seasonal ratings set to normal ambient conditions, leaving real headroom unused when actual conditions are more favourable.2

Dynamic rating replaces fixed or seasonal assumptions with time-varying thermal limits for lines and transformers, using weather, loading and asset-condition data to reveal usable headroom. For overhead lines (dynamic line rating [DLR]), capacity depends on how well the conductor sheds heat, so it rises in cool or windy conditions and falls in still, hot ones. For transformers (dynamic transformer rating [DTR]), the limit is set by winding and oil temperature. Utilities have long varied transformer loading with operating conditions; what is newer is the automation and integration of this into real-time systems. Ratings can be estimated from weather data alone, measured directly with line or transformer sensors, or both combined for greater confidence; the trend for lines is towards forecast-led methods, reserving sensors for the most constrained corridors.

Advanced power-flow control

A newer class of modular devices is making power-flow control faster and cheaper to deploy. Operators have been aiming to redirect power from congested circuits towards underused paths: phase-shifting transformers and flexible alternating-current transmission system (FACTS) devices (including TCSC, STATCOMs, static VAR compensators, series compensation and unified power-flow controllers) have done this for decades, and high-voltage direct-current (HVDC) links have long moved bulk power over long distances and subsea. What has changed is the emergence of modular, series power-flow controllers that clamp onto existing lines without new civil works or substation rebuilds, making deployment a matter of months rather than the years a full FACTS installation or new phase-shifting transformer typically requires.

HVDC is also taking on a broader role. Embedded and multi-terminal links let operators adjust power flows independently of the surrounding alternating-current network, and, with grid-forming control, contribute to system strength and stability, not only bulk transfer. As these schemes multiply into multi-terminal grids and offshore hubs, the co‑ordinated control of these assets becomes both more valuable and more complex, which is what makes them a natural fit for AI-assisted co‑ordinated control, optimising setpoints across multiple links and devices at once, something not practical to do manually at this scale.

Topology optimisation

Reconfiguring the network to relieve congestion is an old idea made newly practical by faster computation. Grid reconfiguration has traditionally been a rudimentary process, decided manually by operators based on past experience and operational guidelines. Using mathematical modelling to identify the best switching action instead of relying on operator judgement is not itself a new idea; what has changed is tractability. Finding the network configuration that best relieves congestion is a genuinely hard computational problem, and faster computing and modern optimisation methods, increasingly AI-based, now solve it fast enough to be useful in operations, seconds to minutes, rather than only as an offline planning exercise. This is what is moving topology optimisation from research into real deployment.

Topology optimisation now supports long-term planning, operational planning and real-time decision making in several US system operator regions, with case studies concluding that recommended reconfigurations can typically be identified in seconds to minutes, expanding effective grid capability by around 5-25% while meeting reliability requirements and delivering material congestion and curtailment reductions in operational settings.

Real-time remedial action

Real-time remedial action allows operators to comply with operational security criteria without holding large headroom in reserve for every contingency. A pre-approved automatic action, such as tripping a specific line or adjusting a setpoint, executes within milliseconds once a defined trigger condition is detected. Traditionally, this kind of rapid predefined response was reserved for rare, high-impact contingencies; contingencies were instead typically handled preventively, by holding headroom in reserve. Greater data availability and computing power now make it possible to address likely contingencies in real time, either correctively after the event, or preventively through continuous risk-based assessment.

Multiplying benefits by combining solutions

Combining GETs can multiply their benefits when each addresses a different constraint on the same network. The benefits of advanced technologies are greatest when they are deployed in co‑ordination. Some gains overlap, where technologies draw on the same underlying headroom, while others compound, since capacity and stability are mutually dependent: extra headroom is only useful if voltage, protection, dynamic stability and restoration risk remain within secure limits, so line ratings, power-flow devices, topology actions and automation need to be assessed together.

For every GET discussed in this chapter, revealing or creating headroom is not the same as being able to use it. How much can be credited in planning, operations or markets depends on the operational security layer discussed in Chapter 1: validated safeguards, operator trust and integration into control-room processes. GETs can also shift constraints rather than remove them: dynamic rating can move a bottleneck from a line to a transformer, while topology optimisation may require updated protection settings and switching procedures. Capacity unlocking therefore needs to be assessed against the network’s full operating limits, not only against local constraints. This is where AI can become most valuable for GET deployment: identifying when a technology creates usable headroom, or when it simply moves the constraint elsewhere.  

Stability limits also need to be considered as grid capacity is unlocked. Higher utilisation is only useful if the system can still withstand disturbances, maintain voltage and frequency, and recover securely after faults. Wide-area monitoring, drawing on PMUs and other high-speed measurements, gives operators visibility of how the system behaves dynamically, not just how loaded it is. Converting that visibility into secure headroom depends on assets that can actively support stability, including devices that provide rapid voltage support and oscillation damping. Increasingly, it will also depend on converter-connected resources such as batteries, wind and solar operating in grid-forming mode, contributing the voltage- and frequency-setting behaviour traditionally provided by synchronous generators. As more of the system's capacity comes from converter-connected resources rather than conventional generation, this stability-support function becomes an increasingly necessary complement to technologies that unlock thermal capacity or redirect flows.

GET priorities differ by network segment. In transmission, the main value is usually relieving congestion, increasing transfer capacity and reducing curtailment, by revealing dynamic headroom, redirecting flows or automating corrective actions. In distribution, the priority is often to connect more demand and host more distributed resources while maintaining voltage quality, protection co‑ordination and reliability. The underlying functions are therefore similar, but the operational problems, data requirements and relevant technologies differ by network level and local constraint. Deployment also varies considerably even among technically mature GETs. Chapter 3 examines why this gap between proven capability and actual deployment persists, and what would close it.

Replication challenge

The survey results show that technical maturity does not automatically translate into widespread deployment. Although many of the underlying engineering principles are broadly common across countries, deployment remains highly uneven. This suggests that successful replication depends not only on technology performance but also on local conditions and the institutional environment. At the same time, each technology has physical preconditions that decide whether it is suitable for a particular site, and these are not barriers that policy can remove.

DLR, which estimates how much power a line can safely carry under actual weather and operating conditions, illustrates the point. Because a circuit’s capacity is determined by its most restrictive element, DLR creates value only where the conductor is the main constraint and cannot address limitations elsewhere in the circuit, such as substation equipment or stability limit. Corridors with limited variation in wind speed and direction may offer little additional headroom, while under less favourable conditions the dynamic rating can temporarily fall below the static limit it replaces.

DTR applies the same principle to transformers, using their thermal behaviour to determine when short-term loading can safely exceed a static rating. Transformers have enough thermal mass that their temperature lags the current flowing through them, and that lag can allow a short peak to be carried before the transform cools again. The load profile therefore has to be cyclical for the method to work, as under a daytime solar profile. A flat load at full capacity leaves no opportunity for cooling.

Advanced power-flow control redirects electricity away from constrained lines and towards parts of the network with spare capacity. It therefore needs somewhere to send the power: a parallel path with enough headroom to absorb the redirected flow. On a radial line there is no alternative route, and where the parallel circuits are themselves near their limits, the diverted flow simply creates a new overload.

Topology optimisation changes the electrical configuration of the network by opening and closing breakers to route flows more efficiently. It depends on the same meshed structure and adds a further demand: network model accuracy. The model has to represent what happens inside substations, rather than treat each substation as a single point. Topology optimisation also requires live breaker status and remotely operable switchgear, since a reconfiguration identified in the control room is of little use if it cannot be executed from there.

The transferable lesson for replication is therefore not the headline figure but the conditions that produced it. Operators first need to establish whether the site suits the core function the technology is meant to perform: revealing dynamic headroom, redirecting flows, automating corrective actions or supporting stability. They then need to test whether that technical fit translates into usable value: whether congestion economics justify the investment, whether data and forecasting are good enough, whether protection and control procedures can accommodate the change, and whether the regulatory framework lets the operator capture the benefit. Chapter 3 examines the conditions that enable this kind of replication: regulatory frameworks, data foundations, skills and organisation, and trust and security.