Working paper
Transition Risk: Investment signals in a decarbonising electricity system
Apr 2023
Development of low carbon electricity generation capacity on a level never seen before is required over the next two decades both to decrease the emissions intensity of electricity production and meet growing demand. Although a fair amount of new generation will be connected within the distribution networks, hugely increased flows on the transmission network can be expected. In this extended blog, Simon Gill and Keith Bell discuss the challenge of identifying quite how much – and where – the electricity transmission network in Britain needs to be expanded.
In the late 1950s and early 1960s the Central Electricity Generation Board (CEGB) carried out a process to decide where to build new generation capacity and design the transmission network to accommodate it[1]. This led to the construction of Britain’s first 400 kV electricity transmission network. At the time, the national electricity system in England and Wales operated on a relatively limited 275 kV network developed since the Second World War to provide some interconnection between regions. However, by the 1960s it was clear that a strategic approach was needed, one that reflected the rapid growth in demand for electricity focused on the cities. At the time, the main options for how to meet that demand were from burning coal or oil, or from nuclear fission. A major focus was on the potential for economies of scale in large power stations to reduce the total cost of electricity. No longer was it efficient for regions of the country to supply their own demand from relatively local generation. Given where the demand was and where different primary sources of energy could be accessed – coal mines in the Midlands, north east England, south Wales and Kent; ports for imports of oil; or coastal sites for nuclear power stations – and the potential for moving energy via a national, high voltage electricity transmission network more cheaply than moving coal or oil via road or rail, what would be the least cost mix of power sources, their location and the cost of transmission?
Today we are faced with a network planning challenge at least equal in magnitude to that of the 1960s. Delivering a net zero economy by 2050 (or 2045 in Scotland) means significantly increasing electricity demand as large parts of the energy needed for heat and transport are electrified. However, the best locations for, in particular, wind farms are not in the same places as the old fossil fuelled power stations they’re replacing meaning that we need to build new sections of network to access them, and a much bigger electricity system as demand grows means we need a bigger network pretty much everywhere. As many stakeholders are now starting to realise – for example, the Electricity System Operator (ESO), The UK Government and the Climate Change Committee – this represents a major challenge, both for the supply chain and in terms of getting public acceptance for the building of upscaled or brand new lines.
Just how much electricity transmission do we need to build? According to the owner of today’s electricity transmission network in England and Wales, National Grid, by 2035 – just 12 years away – we need to build 5 times more transmission overhead or underground lines than we’ve built in the last 30 years and around 4 times more subsea high voltage cables than are currently installed around the British coast. However, that all depends on precisely where the new sources of power will be located.
As the ESO sets out in its “Pathway to 2030”, enabling of access to sources of low carbon electricity that have already been built – most notably in Scotland – or are under construction already gives rise to a huge need for new transmission that, many observers argue, is already late in being delivered. However, other observers – mostly economists – argue that developers of generation, even of wind and solar, have a choice in where they can build generation and that market signals should be defined to direct those decisions towards places that result in less need for transmission[2] (It’s impossible to avoid at least some).
A new project being undertaken by the University of Strathclyde on behalf of the Offshore Renewable Energy Catapult (as part of our Electrical Infrastructure Research Hub (EIRH) collaboration) is exploring transmission planning for net zero. The project is developing a high-level model of the GB electricity system and will use it to explore different generation scenarios – in particular, how much of what type is built where – and their implications for the scale of transmission network infrastructure investment needed to enable use of that generation. The project will enable insights into questions such as the balance between transmission capacity, the provision of flexibility such as energy storage, and the need for output from wind farms to be curtailed for the transmission of power across the country to stay within the network’s limits. It also aims to support the sector in understanding the often opaque process of quantifying costs and benefits associated with new transmission infrastructure, and reaching well evidenced decisions on the best investments to support decarbonisation.
So, how do you plan the transmission network for the kind of seismic change that we’re now in the middle of? In 1962, the shape of the whole electricity system was under the control of the CEGB and a strategic ‘systems’ approach could be taken across the development of both generation and networks. Today, transmission planning takes place in a liberalised environment with the ‘what, where and when’ of generation investment left to the market. Given where we’re trying to get to in terms of low carbon energy, is it enough to follow the approach of the last 30 years, since privatisation, in which the development of the network follows the building of generation? Or, given how long it takes to approve and build transmission, should its need be anticipated, and network capacity built for the generation investment to follow?
A typical offshore wind farm can take a decade to go from conception to operation and most of a project’s lifetime costs are sunk before generation starts. The upfront nature of the costs puts pressure on a wind farm’s final investment decision and means that, to make the decision, investors need high levels of confidence that their project can be connected on time.
Major transmission network infrastructure can require similar timescales, as illustrated by the re-building of the 220 km Beauly Denny line in Scotland from 132 kV to 400 kV which was first submitted to the planning system in 2005, given initial approval in 2010 and fully commissioned in 2015. The cost of investment in the transmission network is recovered from users of the network, both generators and energy users, with Ofgem wanting a high degree of confidence that the money will be well spent, and the assets are unlikely to be stranded.
How much transmission capacity should be used to connect generators and demand? It’s not simply a case of adding up the total generation capacity in a region and building an amount of transmission export capability equal to that total generation capacity. For one thing, we need to take account of the demand for power within that region which will use whatever is produced there, reducing the amount that is exported. Furthermore, the level of demand varies, through a day and through the year. So, too, does the amount of generation that is available to be used and how much the wholesale electricity market chooses to dispatch.
Obviously, the power available from wind and solar farms depends on the weather. Individual farms only produce at full capacity when the wind is blowing, or the sun is shining. Across a large region it is rare that all wind or solar farms can produce at full capacity simultaneously. When the power available from wind and solar is low, something else will need to be used in order that demand for electricity is met. In Britain, a capacity market exists to ensure that there is usually enough ‘schedulable’ generation to meet demand when wind and solar output is low. ‘Low merit’ generation used to fill in gaps in renewable production might be located in the same regions as wind or solar capacity. However, because it is rarely, if ever, used at the same time as high output from wind and solar, it doesn’t – or shouldn’t – add to the need to build transmission export capability from that region.
The optimal amount of transmission will be one at which the incremental cost of transmission equals the incremental value, i.e., the cost of the next MW of transmission capacity equals the value it brings. That sounds reasonable, but how can the value be quantified? We can work that out by comparison with the costs associated with not having transmission.
There are two main consequences of lack of transmission: the market can’t use the cheapest source of power available at that moment; or, for regions that depend on imports to meet demand, not all demand can be met. Quite how those impacts are quantified in monetary terms depends on the precise market arrangements. For example, in Britain’s current wholesale market, it is left to participants to decide how to operate their generation plant half-hour by half-hour. They tell the ESO what they’re intending to do an hour in advance and the ESO takes action to curtail output or bring on more production to ensure that total generation matches demand. In so doing, they also need to make sure that network limits are respected. The total cost of those actions to respect network constraints reveals one part the value of network capacity.
The other part – the cost of not meeting demand – would, for an economist, be revealed through energy users’ willingness to pay for extra power, to be paid for receiving less than they want to use, or for shifting their use to a different time when more power is available, and the network is not congested. However, in spite of the relative success of the ESO’s new “demand flexibility service” in the winter of 2022-23[3], there are, as yet, few actions taken by the ESO on the demand side and a limited number of example cases where turning down demand has been required, particularly due to lack of network capacity.
Action taken by the ESO to re-dispatch generation, demand, or storage in order to balance the system and ensure that power flows stay within the network’s limits are purchased and dispatched through the Balancing Mechanism. This is done by the acceptance of market participants’ Offers to increase generation or decrease demand, and Bids to decrease generation or increase demand. The price component of Offers generally covers the cost of any extra fuel burned to produce the increased output. Bid prices tend to reflect, for fossil fuelled generation, the owner’s savings in terms of reduced fuel use, and market participants tend to pay the ESO for the privilege of turning down (they still receive their wholesale market revenues as if they had generated). However, unlike wholesale revenues, payments for support mechanisms such as Renewable Obligation Certificates (ROCs) and low carbon Contract for Difference (CfD) uplift payments are paid based on a generator’s metered output. Bid acceptances, i.e., reductions of output, would cause renewable generators to lose those support mechanism payments. To compensate for lost income, Bid prices for renewables are therefore negative meaning that money flows from the ESO to the generator.
In spite of the cost of curtailment – the ESO paying both for Bids in renewables dominated areas for generators to turn down and Offers to replace that output on the other side of a constrained network boundary – the cost of building extra transmission means that the optimal amount of transmission results in at least some curtailment. The challenge for the ESO and the network owners is to work out how much is the right amount given the way generation availability and demand vary through a year and into future years, and how much is likely to need to be paid for curtailment actions. Given the different options for building extra transmission capacity, uncertainties in the supply chain due to limited capacity, the effects of inflation, and the potential for big delays in getting planning permission causing extra constraint costs, the numbers on the transmission investment side of the equation aren’t easy to work out, either.
The biggest difference between wind and solar power and other forms of generation is the inability to store the primary energy associated with the former. If it’s windy or sunny, we can flex production from wind and solar by turning the output down (and, if it’s still windy and sunny, turn it back up again) but we lose the electrical energy that would have been produced. In contrast, we can store liquid, gaseous or nuclear fuels or, given a decently sized reservoir, the water for hydro power.
Until recently, the transmission equation for wind-dominated regions was based on a two-sided balance: build more transmission capacity or accept more curtailment. Today, the maturation of battery technology, the growing potential for flexible demand, and the opportunity to develop a green hydrogen sector potentially capable of linking to the electricity system in a highly flexible way, add a third dimension to that balance.
These other sources of ‘flexibility’ depend on investment and availability provided by actors other than the ESO or the network owners. This, in turn, depends on market signals. In spite of apparent analogies between network capacity and replenishable energy storage – the network moves energy from one place to another; storage moves it from one time to another – the ESO and network owners are forbidden from owning storage. Today’s transmission planners need to anticipate not just the relative locations of generation and demand and the cost of networks to link the two, but also the potential for flexibility to reduce curtailment and increase the utilisation of the transmission capacity that is built.
The new project aims to explore the potential future locations of generation and the growth of demand. It will show what these changes would mean for how much energy the transmission network would need to carry over what distances over the course of a year, and what the peak power flows would be if the network had infinite capacity. In so doing, we will shine a light on the way that transmission planning is done and the scale of the challenge in developing the network to accommodate different future energy transition pathways.
In recent years, the Government’s main focus has been on encouraging investment in renewable generation capacity. However, there is a now a recognition of the importance of transmission and, belatedly, the challenge of delivering it. Major wholesale market reforms are being advocated by some, partly as way of encouraging ‘flexibility’ and partly to reduce the need for new transmission. But how realistic is that? And, if low carbon energy ambitions are to be realised, and realised with demand reliably met at least cost, what is the minimum amount of transmission we really must build, and how quickly?
Simon Gill runs the independent consultancy The Energy Landscape and is an associate of Regen.
Keith Bell holds the ScottishPower Chair in Future Power Systems at the University of Strathclyde, is a Co-Director of the UK Energy Research Centre (UKERC), leads the UKERC research theme on Energy Infrastructure Transition, and, along with Ian Cotton, is Scientific Director of the Electrical Infrastructure Research Hub, a collaboration between the Offshore Renewable Energy (ORE) Catapult, University of Strathclyde and University of Manchester.
The blog has been published on the websites of both UKERC and the ORE Catapult.
[1] See Booth. E. S. et al, The 400kV Grid system for England and Wales, The Institution of Electrical Engineers, Paper No. 3883 S, March 1962.
[2] The big idea that many economists put forward in, for example, consultancy reports for the energy regulator, Ofgem, or for the ESO, is ‘locational marginal pricing’ (LMP). However, as others have pointed out – see, for example, a report that we published recently along with our colleague, Callum MacIver – adoption of LMP would have a range of consequences and practical implementation is not without its challenges.
[3] Lots of willingness to flex was revealed but it doesn’t seem to have been an unqualified success: there have been some suggestions of game-playing and there is an ongoing challenge for all forms of demand response associated with ‘baselining’ what the demand profile would have been in the absence of the flexibility request.