Pumped hydroelectric energy storage is a perfect fit for Ireland’s path to zero emissions electricity generation, writes Chris Bakkala. It is a case of feast and famine: more electricity than we can use and not enough when we need it!

On February 23 last, the not-for-profit EnergyCloud Ireland announced a pilot initiative to provide free hot water to 1,000 Clúid Housing homes. 

Working in collaboration with Amazon Web Services, immersion heaters are powered using “surplus renewable energy, which would otherwise be wasted” via smart devices that link people’s domestic hot water tanks to the internet and to the electricity distribution network. A wonderful initiative, to be sure, and timely, in a period of high inflation.

Less than a week later, under the headline 'Ireland rues mistakes of the past as it struggles to keep the lights on', the Telegraph reported the arrival in Ireland of what will amount to 450MW of “temporary” generation capacity in the form of open cycle gas turbine generators [OCGT], which we had learnt in October of 2022 that the government had purchased as a measure of “last resort”. This generation capacity is scheduled to be up and running by the end of 2023.

The two events reflect the growing presence of renewable wind power supplying electricity into the Irish grid alongside our continued dependence on fossil fuel for 'peaking plant' and for 'baseload', a fuel basis in which the market is increasingly unwilling to invest. 

Peaking plant and baseload

Peaking plant provides dispatchable power, or power that can be rapidly switched on in response to spikes in demand such as the morning rush and the evening dinner. Baseload is the steady generation capacity that is always on in the background, providing continuous power supply capable of responding to the surges and waning of wind power on the transmission network and to sudden variations in demand.

Battery storage increasingly provides bridges between peaks and troughs in consumption, helping to flatten the demand curve. The fast response time of batteries also helps to maintain voltage and frequency at the required levels, to cope with the sudden arrival or departure of a dose of wind power onto the grid or to surges or drop-off in demand. 

And batteries serve to resolve bottlenecks, or 'constraints', in the transmission network. All of this has aided EirGrid, the transmission system operator (TSO) for Ireland, to admit a higher proportion of wind onto the system.

The continual challenge is to reliably match generation capacity to electricity demand at every moment of every day. In the first instance we need to maintain a steady power supply when the wind eases. 

At the other extreme, if we cannot consume all of the electricity being generated by the wind, and if we cannot store it, then we need to switch the wind power off to avoid overloading our high voltage network. 

'Dispatch-down'

'Dispatch-down' is an order from our grid operator to a wind generator to disconnect the power supply when energy production exceeds demand. And such orders have been increasing alongside our increasing use of wind to generate electricity. EirGrid’s annual constraint and curtailment report quantifies just how much.

The task of maintaining perfect balance in capacity and demand is at a scale one or two orders of magnitude greater than current battery technologies can reasonably provide, something like using a team of electric vehicles to haul a loaded semi-trailer. 

And with a discharge time typically less than two hours, grid-scale batteries suffer from the same sort of range anxiety as EVs. The transmission system needs long-term storage deployed at scale to dislodge the grip that fossil fuels have on it, and batteries are not going to provide this in the near future.

Reliability of supply is the reason we continue to burn coal, heavy fuel oil, and gas to meet our electricity needs. Maintaining the balance between supply and demand is assisted by our interconnectors linking the Irish and Northern Irish high voltage networks to the UK, and soon to France, and hopefully intra-island via the north-south interconnector (now 19 years in the planning stage); but all of the planned interconnectors will not come close to supplying sufficient imported power to fill the gap between what the wind can supply at any moment, and our demand at any instant.

Our electricity generation mix has begun to resemble the “two-legged stool” forecast more than 14 years ago by Dermot Byrne, the then chief executive of EirGrid. Commenting on the impending closure of the coal-fired station in Moneypoint, Co Clare, (which closure has now been postponed indefinitely) he said at the time that (with the closing of Moneypoint) “we will then have – under the business-as-usual scenario – 40% renewables, and unless some decision is taken, we will be reliant for 60% on gas. The question then is instead of three legs of the stool [coal, gas, wind], in terms of diversity, you're down to two legs of the stool. And there will be times when the wind is not blowing so you'll be very reliant on gas for significant periods of the year." – Eolas Nov 9, 2009.

The reality we now face is very like the scenario he foretold. And some of the risks of an over-reliance upon clean-burning gas were made a little bit clearer last winter. 

Referring to the figure from an EirGrid report from May of 2023, in 2022 we derived about 40% of our electricity from wind, and about 60% from the burning of fossil fuels, mainly gas, in the production of more than 34TWh of energy in the year, the same proportions that had been predicted 13 years earlier.

The limits of wind power

The 60/40 proportion of fossil fuels to renewables achieved in our electricity generation in 2022 is a considerable improvement on the 80/20 mix of the year 2014, when 26TWh of energy was generated in Ireland, but it represents a current upper limit to the wind power that can be accepted by our high voltage network. 

Asynchronous wind power still needs to be balanced with synchronous electricity generation to maintain stability in our high voltage network and to respond to surges in demand.

It is a routine matter for a fossil fuel burning plant on our transmission network to synchronise electricity generation with the electricity phase frequencies of the distribution network, and to maintain a steady and reliable voltage to the transmission network. 

Not so for wind generation, which is non-synchronous, and subject to the continuous variation, or absence, of wind. The inclusion of wind generators in the high voltage transmission network introduces a variety of risks to grid stability, which can cause the system to crash if they are not controlled. 

Historically, the variations in frequency, in voltage, and other non-synchronous anomalies have been resolved by maintaining stable, synchronous spinning reserve with sufficient 'inertia' to correct small anomalies before they can grow into instabilities.

 And while batteries are now providing stabilising services to admit more wind power onto the network, a fossil fuel generating plant provides inertia at scale, and remains crucial for safeguarding the system from the effects of non-synchronous wind generation coming in and out of the system throughout the course of each day, and, crucially, to maintaining sufficient supply to meet demand all day long, every day.

Recognising the value of interconnectors to facilitate more renewables, but neglecting for the moment their relatively small contribution, the SNSP is the maximum proportion of non-synchronous power to total demand that is allowed to be present on the high voltage network at any moment. 

When it was established, EirGrid set the non-synchronous limit at 50%, so the minimum synchronous plant requirement in the absence of interconnector contributions was also 50%, beyond which it deemed the risks of instability became too great to be able to reliably operate the high voltage transmission network. 

In the years since 2011, as improvements to the network have been made, and as the operator has become more familiar with the increasing proportion of wind on the network, EirGrid has incrementally increased the SNSP. 

Today, again thanks in part to batteries, the system operates with an SNSP of 75%, and EirGrid has set a target of 95% for the years ahead. Effectively, the TSO now operates the transmission system with as little as 25% of the power derived from synchronous generators at any moment in time. 

Yet, with all of that wind potential, the need for operational reserve during periods of low wind, and the need for a timely response to peaks in demand, means that we still derive about 60% of our electricity from the burning of fossil fuels, just as Dermot Byrne predicted 14 years ago.

Doubling down on wind

As our wind energy sector has expanded, so too has market competition. At EirGrid’s first offshore wind capacity auction of last May, the generating capacity of each lot was of sufficient scale to attract interest from overseas, and all of the successful bidders were backed by capital from abroad. 

As reported by breakingnews.ie on the day the results were published, the successful bidders’ financial backing comes from Norway’s Statkraft (North Sea Irish Array 500MW), Germany’s RWE (Dublin Array 824MW), France’s EDF Renewables and Norway’s Fred Olsen Seawind (Codling Wind Park 1300MW), and Australian Bank Macquarie’s Green Investment Group (Sceirde Rocks 450MW). 

At an average price of €86.05/MWh, the auction secures a wholesale price for electricity lower than the latest price of onshore wind in Ireland (€98/MWh), and fixes it for 20 years. With a 'Deemed Energy Quantity' of 12,117GWh per annum, it represents a near doubling of the 13,676GWh of wind energy generated in Ireland in 2022, which is a huge step-up in generating capacity.

While market competition served consumers well in securing a better price than was on offer from incumbent market participants, part of the attraction to the new investors was undoubtedly the risk reduction for them in the form of 'unrealised available energy compensation'. 

This is a payment mechanism that allows the offshore wind generator to be fully reimbursed, at the agreed wholesale price, for any electricity that it is able to produce but which we are unable to consume. The compensation is to be paid directly by consumers, via the TSO levy, to the operator of the offshore wind farms, all of whom are based in Europe and outside of Ireland. It is impossible to know in advance what the hidden cost to the consumer might eventually be, but history may provide a guide to a reasonable estimate.

To try to estimate the amount of wind energy that might go unutilised, but for which the consumer will nonetheless be required to pay, we can look again at EirGrid’s Renewable Energy Constraint and Curtailment report for 2022, a year in which 1,279GWh of electrical energy went to waste, enough energy to supply about 300,000 homes for a year, and about 8.5% of total wind energy production for the year, as indicated in the table below. 

An upward trend in dispatch down orders in the 10 years from 2013 is evident from the table, with somewhat lower values in the relatively ‘low-wind’ years of 2021 and 2022. 

During the years from 2013 to 2022 total installed wind capacity more than doubled, from 1923MW to 4527MW, and ‘dispatch down’ orders rose to two-and-a-half times during the same period, despite the 50% rise in non-synchronous penetration of the SNSP. 

If we assume dispatch down orders in 2030 will remain constant, then something like 10% of 12,000GWh at €86.05 represents a recurring cost to consumers of about €100m per annum, for 20 years, to be paid to the companies outside of Ireland who own the assets. And if we do not provide the necessary grid infrastructure and consumption capacity to cater for a doubling of our wind energy, the figure could range considerably higher.

It seems reasonable, therefore, to seek ways to ensure consumption of energy that we have now guaranteed to purchase. Pumped hydroelectric energy storage [PHES] is a technology that can be brought to bear to do just that, and to store the energy consumed for dispatch when needed, using long-established technology, that can be realistically deployed by 2030, in time for the sea-change arrival of big wind.

Pumped Hydroelectric Energy Storage (PHES)

PHES uses electricity to power pumps that shift water uphill, converting the electrical energy into the potential energy of water residing at height, and storing it there. When the energy is needed, the water is released downhill through turbines to generate electricity. 

It does what it says on the tin, energy storage, and it can do so for a very long time without loss of potential energy, and with a total round-trip efficiency in excess of 82%. 

PHES provides an excellent sink for the absorption of wind power when the wind is blowing hard, and a capacity for significant contribution to demand when the wind does not blow. 

PHES provides an inertial heft which cannot be obtained from batteries and it relies on proven, long-established technologies that can be procured in the EU and deployed with confidence. PHES can be deployed at a scale that batteries cannot, and provide a discharge time measured in days, weeks or months, rather than mere hours.

PHES need a hilly or mountainous topography and sufficient water to fill a lake. It is the most established technology for energy storage at scale and constitutes more than 90% of the world’s grid energy storage (Source: Grid-Scale Storage, IEA, IEA (2022), Paris https://www.iea.org/reports/grid-scale-storage).

Put simply, when electricity is abundant, a PHES system (PHESS) consumes electricity to power pumps to lift water from a low reservoir to a high reservoir, so that when electricity is scarce, the water can fall by gravity from high reservoir to the low one through turbines. 

The pump can be a reversible unit, such as the Francis pump-turbines at Turlough Hill (the only PHESS unit currently deployed on the Irish high voltage network), or the pumps and turbines may be separate machines, connected to a single drive shaft fitted with a clutch. 

Such a configuration, known as a ternary set, allows pumps and turbines to spin in the same direction, and simultaneously if desired, enabling a very fast response time to variations in demand, variations in voltage, and variations in frequency of supply. 

The rapid response time and the potential for simultaneous pumping and synchronous generating makes it possible to provide a range of services to the transmission system operator that increase the resilience and stability of the high voltage distribution network. 

It provides inertia that looks, to the grid, like any fossil fuel generator of comparable size, displacing fossil fuel from the network and enabling growth in the proportion of wind energy. 

And it stores substantial energy for long periods of time to provide dispatchable electrical power when the wind is not blowing. Ternary PHES becomes an increasingly compelling proposition as the proportion of non-synchronous generation on the network increases, such as now. 

The geometry and arrangement of traditional PHES installations (non-ternary), are illustrated in the figure below.

An illustration of a hydraulic short circuit for simultaneous pumping and generating in a ternary set is reproduced below from Modelling Ternary Pumped Storage Units by Argonne National Laboratory.

PHES has been developed all over the world because it can store substantial energy over long time periods at a cost that has proven to be affordable for both advanced and developing economies. 

While capital cost may be high, it more probably will be low in relation to comparable alternatives when total life cycle costs are considered. Operational cost of PHES is quite low, and the life cycle of plant, at 50 to 100 years or more, is several times greater than comparable thermal plant, and an order of magnitude greater than that of batteries, which, as noted previously, cannot at present provide comparable performance.

PHES offer capacity several orders of magnitude greater than batteries can achieve, with storage duration measurable in days rather than hours. 

Batteries undoubtedly have a bright future in their contribution to the resilience of our high voltage network, though their relative new-ness means we have not yet had to confront the challenge of replacement of the hundreds of thousands of batteries now planned for grid-deployment when they come to the end of their 10-year design lives. 

There is an as-yet untapped potential in the 'behind-the-meter' batteries that will be crowd-funded and crowd-sourced in the electric vehicles that are now finding their way to people’s homes. 

In time, they will form a substantial reserve of stored electricity, in a manner similar to the hot water cylinders that EnergyCloud and Amazon Web Services are now keeping hot for Clúid Housing, but the energy will move in two directions, from the distribution grid to cars and from people’s cars across their meter back to the distribution grid.

For our high voltage transmission network, PHES installations, like batteries, provide a means to bridge the gap between the time of day when wind power is being wasted, or given away, and the times of day when electricity is most expensive – the morning and evening peak demand. 

As well, when distributed around the grid (and there are potential sites in every county in Ireland) they will also reduce constraints and curtailments so that less energy will be wasted, just as batteries do to admit greater penetration of non-synchronous generation into our network. But unlike batteries, the scale of PHES, and the ability to store energy for extended periods, will enable the displacement, ie removal of fossil fuels from our high voltage distribution network.

Looking again at EirGrid’s constraints and curtailments report from 2022, the figure below shows that it is the small hours of the morning, when most of us are asleep, that most of our excess wind energy goes to waste. 

We can heat our immersion tanks, we can charge our electric vehicles, we can run our heat pumps, and we can pump water uphill from 11pm until 7am to make ready for the predictably diurnal patterns of human activity, all using energy that will otherwise go to waste.

PHES is already here, and should expand

Pumped Hydroelectric Energy Storage is currently deployed to good effect in Scotland and Wales and in Ireland. Here we have one such installation, at Turlough Hill in Co Wicklow with a rated capacity of 292MW. It is now in its 50th year of operation, and may very well continue for another 50 years.

A second PHES project with 360MW rated capacity is currently being planned for Silvermines, Co Tipperary. It has been deemed a project of common interest by the European Commission, and has won several million euro in grants to fund feasibility and environmental studies. 

The project does not yet feature in EirGrid capacity forecasts because it has yet to run the gauntlet that is the planning process. In its Capacity Outlook 2022-2031, EirGrid simply states: “This project has not been included in adequacy assessments, development of this project will be followed and included when appropriate.”

Absent any other proposed hydro power developments in EirGrid’s radar, its 10-year planning horizon includes no new hydropower installations. We need to discover and develop a new perspective for assessing PHES. 

It will not be developed in gigantic grand schemes, but it can be developed incrementally, where site conditions are favourable, in installations that are scaled in proportion to particular site conditions.

The Capacity Outlook report is reproduced in a modified form in the figure below. In the image on the left, the original figure is redrawn with the relative size of operational thermal plant and hydro plant drawn at the same scale. 

It is notable that thermal plant in excess of 30 years old is scheduled for retirement, and that there is no retirement in sight for hydro installations ranging from 59 to 87 plus years in age. Longevity is a telling differentiator between turbines that burn stuff and turbines that don’t. 

In the image, new PHES installations with rated capacity totalling 910MW are included in the hydro fraction in 2030, with thermal plant displaced at a 1:1 capacity ratio. The PHES projections assume the Silvermines project becoming operational in 2027, and a project at Lough Salt, Co Donegal, becoming operational in 2030. 

It is not an unrealistic expectation to wean ourselves off of fossil fuels; it is a requirement. What is needed is the development of the alternatives, and their integration into a network of technologies that does not yet exist.

H₂ from H₂O and the wind

One significant consumer for the offshore wind that is scheduled to start arriving in 2030 is the hydrogen industry. The National Hydrogen Strategy, published by the Department of Environment, Climate and Communications in July 2023, forecasts 19,000GWh of dispatchable hydrogen powered electricity generation in Ireland by 2050, a small fraction of the 106,000GWh hydrogen demand forecast for all end users, including industry, aviation and road and rail freight. If all of that hydrogen is derived from zero emission wind energy, ie 'green hydrogen', we will have a zero emissions energy system for all of our energy needs.

If we can obtain hydrogen by electrolysis at 95% efficiency, and transport it at 95% to liquefied storage at 60% efficiency, it can be burnt in an engine operating at perhaps 35% efficiency, we will have a total efficiency of 19.9%, so about 560,000GWh of wind energy will be needed to meet hydrogen demand. 

If we can avoid the energy intensive process of liquefying and storing liquid hydrogen, and instead convert hydrogen gas to electricity in fuel cells operating at 60% efficiency, we will have a process that is 54% efficient, requiring 195,000GWh of wind energy. Either way, the 25,000GWh of wind energy currently in the pipeline for 2030 will require significant ramping up to realise the zero emissions future. 

The promise of green hydrogen is a zero emissions means of storing the energy we will need. The challenge is that the industry is in its infancy. We have hydrogen-ready boilers coming into our homes now, and we have gas pipelines capable of transporting blends of hydrogen and gas. But the complete infrastructure does not yet exist, and its invention and deployment will take decades. 

Undoubtedly, during that time there will be technological advances that will improve the energy efficiency of green hydrogen. We don’t know what they will be, but history shows us that technology which exists will be improved upon so by 2050 we will surely have ways of using hydrogen at better than 20% or 54% efficiency.

In the meantime, at 82% round-trip efficiency, PHES will provide an important and efficient means for displacing fossil fuels from our energy mix. It is something that we can start to deploy right now. 

And we can take substantial strides to implement PHES at scale in the years leading up to 2030 and the years beyond. Given the inherent efficiency of the technology, the complete absence of emissions, the system longevity, and the very low maintenance requirement, we can also be confident that the technology will continue to make economic and environmental sense in the decades to 2050, and in the decades beyond.

Author: Chris Bakkala, CEng MIEI is a leading multidisciplinary engineer who has delivered projects valued at hundreds of millions of euro. Chris is now developing Ireland’s most advanced Pumped Hydro Energy Storage System, to support the flexibility, cost efficiency and reliability of the Irish grid in its transition away from fossil fuels.