The modern world relies on access to affordable, dependable power, delivered by a secure and robust electricity grid network. Grids have traditionally been based on large, dispatchable point power generation sources. Globally there are more than 80mn km of grid network, and it is expanding by 2.3mn km/y with repair, replacement and upgrade required as well as expansion.

Electricity supply and demand must always be balanced. System stability requires constant equilibrium between power generation and consumption, even as demand fluctuates. When demand exceeds supply, such as when a power plant fails, the associated drop in system frequency can result in widespread problems if not managed appropriately. If system frequency is not restored within a limited time window (typically nine minutes) power plants operational on the system can enter controlled shutdown as a protection mechanism. If system frequency falls further, system collapse may occur, with associated loss of energy to the entire system.

During the 2021 Winter Storm Uri, the US’ Electric Reliability Council of Texas (ERCOT) was four minutes from complete system collapse. Almost half its power generation was lost, narrowly avoiding complete system failure with associated catastrophic impact on the region and its population, which would have taken weeks to restore. As it was, there was power failure to five million people with a further 11 million impacted by power interruptions. There were 246 attributed deaths and an economic impact of around $200bn.

Texas is just one example; others include Argentina, Bangladesh, India, South Australia and the UK.

Grid decarbonisation 
Power sector decarbonisation is transforming electric power grids, as the share of variable renewable energy (VRE), mainly wind and solar, feeding into the grid is expanding, while demand grows with increasing electrification. Meanwhile, large dispatchable generating capacity (such as coal-fired power) is being phased out. New VRE capacity are smaller units, in greater numbers, and often located in remote areas both on land and offshore.

Deployment of more VRE-based generating capacity and the associated intermittent nature of power generation, plus the age of much of the grid, extreme weather events and system congestion, mean that substantial investment is required to upgrade and expand electricity grids.

The less-predictable nature of VRE makes it harder to continuously match electricity supply and demand. The more VRE on the system, the larger the challenge.

Therefore, it is important to understand the supply-related risks, such as ageing infrastructure and increasing power demand. Some regions have simultaneous demand increases related to economic development or population growth. Many areas are likely to see power demand double over the next 20 years.

There are issues due to grid congestion, and reinforcement requirements to handle the higher loading. Extreme weather events are more frequent and severe; and increasing dependence on grids increases the risk of terror and cyber-attack. Additional risks associated with change, affordability, speed of deployment, supply chains and materials availability, as well as planning, policies and market reforms, alter how systems are operated, governed and rewarded. Furthermore, there is potential risk from an over-reliance on VRE in maintaining a stable, dependable and affordable system, as efforts to decarbonise are accelerated.

VRE integration requires grid adaptation. Many grid systems have some VRE capacity but also rely on legacy dispatchable power assets and interconnection to maintain security and reliability of supply. In addition, integration of VRE requires energy storage, curtailment, re-dispatch, grid modification, grid-forming inverters and synchronous compensation, consumer flexibility, wholesale price volatility and market reform.

Even when only 25% of peak capacity is installed as VRE, issues become apparent. This means there is a need for potentially expensive system-wide mitigations to integrate the higher VRE share being demanded.

Balancing the grid 
VRE is largely enabled by taking advantage of the flexibility of pre-existing systems, but there is a paradox. As system flexibility needs increase, flexible power assets that have enabled VRE are being decommissioned. If this continues, there may no longer be adequate flexible dispatchable capacity.

Fig 1 illustrates a UK 2050 scenario with huge increase in flexibility demand evident, but also an almost complete elimination of existing flexibility capacity. Yet it is not proven that reliance on consumer demand management, storage and interconnectors will be adequate, dependable or cost-effective to satisfy needs at national scale in the future.

Fig 1: A huge increase in UK electricity system flexibility (GW) is forecast by 2050
Source: National Grid ESO, 2023

Is surplus VRE capacity the answer? 
Building excess VRE capacity is an option, but may not be a solution in still, dark winter periods for example. Building more interconnectors to share surplus power may not be sufficient, as similar weather patterns can be continent-wide. The cost of overbuilding is high, and materials are resource-intensive; also, the more VRE, the lower the load factor under average conditions.

Alternatives include shifting demand or introducing new industries which can absorb the surplus power, such as production of electrolytic hydrogen. However, the generation production costs would still need to be covered by energy sales, and new industries would need to be viable businesses, even accounting for the variability of the renewable energy surplus.

Dispatchable technologies are required not only for smoothing out variations in VRE output but also, critically, to bridge extreme system events – especially those of significant duration. There are clean dispatchable power options available. A resilient system is typically one with a broad technology portfolio able to adapt to a variety of different circumstances.

Calculating total system cost 
Calculating cost is complex. Costs differ in time and place for the same thing, and are continually distorted by taxes, incentives, project and market conditions. Only cost ranges exist, and they can overlap widely for a range of technology options, meaning that any of the overlapping options could be the least cost in particular circumstances.

The standard levelised cost of electricity (LCOE) can only be used for comparison where the impact of options considered are identical outside the boundary of the cost calculation. However, system-wide impacts are generally not included in plant LCOE calculations, and so LCOE cannot be used directly to compare the economics of dispatchable and non-dispatchable power sources.

For VRE technologies this means firming, networks, storage and system stability costs fundamentally change the messages portrayed by most LCOE figures (see Fig 2). The ‘cost’ a consumer sees is not the project or plant cost, but the total system cost impact, reflected in consumer prices. This price includes a large share of costs for transmission and distribution networks but also government costs, taxes, levies and incentives. Assuming a correlation between LCOE and consumer price is not correct.

Fig 2: Grid-level system integration costs of technologies in the US, OECD. The myth that renewables are cheap persists in part due to the flawed use of LCOE.
Source: Watt-Logic 2023

Technology installation takes time. High-capacity options typically take 6–12 years to reach commercial operation, and grid developments can take 5–15 years to implement. In the US for example, capacity awaiting permitting in 2022 was almost twice the total currently-installed generation fleet.

Globally, over 3,000 GW of renewable capacity awaited permitting in 2023. Thus, for any technology option, decisions need to be made far enough in advance (15 years or more) to enable planning and deployment before that option becomes critical to the system.

Standard levelised cost of electricity (LCOE) can only be used for comparison where the impact of options considered are identical outside the boundary of the cost calculation. However, system-wide impacts are generally not included in plant LCOE calculations, and so LCOE cannot be used directly to compare the economics of dispatchable and non-dispatchable power sources.

VRE-heavy systems around the world rely on fossil fuel-fired plants and interconnectors. Extreme flexibility demands are being made on these fossil plants with associated consequences.

Regulators and operators are raising concerns about system reliability.

In the Australian National Energy Market, system operator AEMO has forecast that system reliability standards could be breached in some areas by 2025 on the current path, due to inadequate consideration of the wider system implications of power generation policy choices.

As yet, no major grid system has demonstrated a reliable, affordable transition pathway to net zero. Electricity grids must provide dependable, affordable, stable and secure power for our societies and economies, and this should continue to be a priority throughout the low-carbon energy transition.

*This article is based on the report Maintaining a stable electricity grid in the energy transition, by Mike Garwood, ICSC/CIAB, January 2024.

• Further reading: ‘Synchronous condensers deliver vital inertia for grid stability’. Find out how synchronous condensers can restore the balance for both urban networks and remote islands.
• Only a combination of decarbonisation approaches applied across the value chain will help industries and countries reach ambitious net zero targets. Electrification and improved energy efficiency will be the key approaches.