6 Energy Storage Technologies applicable to Scotland
6.1 Scenario Impacts
In order to determine the role that energy storage measures and potentially demand side management may be able to play in Scotland, we have assessed the impact of Scenarios 1 and 2 on the network flows in Scotland. From this we can determine whether the generation mixes outlined and resulting TWh output present problems that may require energy storage/ DSM measures.
Using the detailed generation profiles created for Scenarios 1 and 2, the power flows across the Scotland to England boundary have been determined using a half-hourly GB dispatch based on merit order. These flows have been calculated for years 2015, 2020 and 2030.
The results have been collated as a duration curve that shows the amount of time that power flows are above a certain value. In the charts presented below it is assumed that positive values correspond to power flows from Scotland to England (exports) and negative values show power flows from England to Scotland (imports).
In the charts:
- The X axis is black if the chart shows all generation
- The X axis is green if the chart shows renewable generation only
- The Y axis in the charts is shown in red for the imports.
The charts also show the transfer limits for export and import with the main planned and projected network reinforcements over the period to 2020 (2010 interconnector upgrades, 2012 series capacitors, 2018 West Coast HVDC interconnector and, by around 2020, the East Coast HVDC interconnector) 12, details of these are provided later in this Section.
Finally the results are presented with all the generation as assumed in the study Scenarios and with the flows that would arise with only renewable generation output.
Modelling total generation output and that of only renewables allows us to determine the potential role of energy storage. For renewable generation we assume that, given 'priority access' (and the low marginal cost of renewables) then renewables are dispatched ahead of thermal generation. When assessing the impact of total generation, the dispatch of thermal generation in Scotland is determined by dispatch on the GB system.
When assessing the role of energy storage we implicitly assume it makes little economic sense to 'store' thermal generation - a more economically optimum approach would be to 'store' thermal generation as unused fuel and thus not to generate from thermal plant. However, as renewable generation is very low cost, then it may make some economic sense to 'store' this generation. As a result, the analysis of only renewable generation allows us to determine the role that energy storage may play, assuming thermal plant is ultimately flexible.
6.1.1 Scenario 1
Figure 6.1.1 shows the cumulative power flows from Scotland to England from 2015 to 2030 resulting from the generation portfolio considered in Scenario 1. The equivalent power flow considering only renewable generation in Scotland is presented in Figure 6.1.2
Figure 6.1.1: Scenario 1 - Scotland to England Power Flows 2015 to 2030-All plant
Figure 6.1.2: Scenario 1 - Scotland to England Power Flows without conventional plant 2015 to 2030
The results of our analysis indicate that, under Scenario 1, given the transmission upgrades currently proposed between Scotland and England to accommodate the increasing power flows resulting from additional Scottish generation, there may be only limited requirement for energy storage. Even with thermal and renewable plant operating in tandem and economically dispatched the transmission upgrades outlined are sufficient to export Scottish generation.
In simple terms the electricity generated in Scotland from both thermal and renewable plant can be exported to England providing the transmission upgrades identified are constructed in the time frames outlined. As a result the construction of additional energy storage is not required in Scenario 1 - a Scenario that meets the Scottish Government's existing renewables target.
If the transmission upgrades are delayed or not undertaken, then the output of thermal plant in Scotland may need to be constrained. Constraining thermal plant in Scotland and replacing with output from plant in England is likely to incur costs of around £60-70/ MWh in the long term. Constraining renewable plant is more costly due to costs incurred from incremental generation in the South, but also the lost Renewable Order Certificate ( ROC) income incurred by the constrained renewable generator - currently around £50/ MWh - giving total constraint costs of around £100/ MWh. As a result conventional plant will always be constrained ahead of renewable generation. In addition renewable generation should have 'priority access' over and above conventional thermal generation.
6.1.2 Scenario 2
Figure 6.1.3 shows the cumulative power flows from Scotland to England from 2015 to 2030 resulting from the generation portfolio considered in Scenario 2. The equivalent power flows considering only renewable generation in Scotland are presented in Figure 6.1.4.
Figure 6.1.3: Scenario 2 - Scotland to England Power Flows 2015 to 2030-All plants
Figure 6.1.3 shows that, under the 'all generation' option in Scotland, then despite the construction of the transmission upgrades currently proposed, 'surplus' Scottish generation occurs from 2015 onwards (shown by the area above the 2020 Eastern HVDS line. As a result a number of system management options could be employed:
- Scottish generation must be constrained off and/or
- Energy storage/demand side measures are employed
- Additional transmission capacity constructed.
The results of our analysis indicate that, by 2020 up to 1.5 GW may need to be constrained off the system if energy storage or additional transmission capacity is not available. As a result constraints could occur for around 10% of the time.
By 2030 up to 5 GW may need to be constrained off the system if no storage/ DSM or additional transmission upgrades are instigated, leading to constraints occurring for 28% of the time.
For comparison, figure 6.1.4 shows the power flows that might result if all thermal plant is either constrained off the system, or is not dispatched. The results show that, even if Scottish thermal generation is not operating, then by 2020 power flows are beginning to be constrained. Beyond 2020 the constraints rise. As a result, given the generation mix in Scenario 2, if no energy storage/ DSM measures or additional transmission upgrades are instigated, then beyond 2020 renewable generation may need to be also constrained off the system in addition to thermal.
The results of our analysis indicate that, beyond 2020, as the contribution of renewables in Scenario 2 increases, then without energy storage/ DSM measures or additional transmission upgrades, up to 4 GW of renewable output may be constrained off the system. As a result constraints could occur for up to 15% of the time, in addition to fully constraining all output from thermal plant.
Figure 6.1.4: Scenario 2 - Scotland to England Power Flows without conventional plant 2015 to 2030
6.1.3 Role of the interconnectors
Our analysis above highlights the key role of the proposed interconnector upgrades to support additional renewable generation in Scotland.
In its 2009 report the Energy Networks Strategy Group ( ENSG) concluded that, to support the development of 12 GW of renewable generation in Scotland major reinforcements to the transmission network would be required. The ENSG identified three main upgrade requirements that have been included in our analysis:
- The 'Incremental' upgrade, which includes reconductoring and reinsulation work on existing onshore transmission routes, along with the development of new and existing substations and the installation of series compensation thus making maximum use of existing transmission routes. The total cost of the reinforcements was calculated at £625m.
- The Western subsea HVDC Link. A 1.8 GWHVDC link between Hunterston and Deeside providing additional capacity across the 'interconnector' circuits and additional capacity across the upper North of England. The total cost of the reinforcement was estimated at £760m
- The Eastern subsea HVDC Link. A 1.8 GWHVDC between Peterhead and Hawthorne Pit providing additional capacity across the key transmission boundaries between central Scotland and the north of England and limited additional capacity across the upper North of England. The total cost of the reinforcements was estimated at £700m
The results of our analysis concur with the ENSG work. In Scenario 1 some 10.2 GW of renewables is commissioned in Scotland by 2020, rising to potentially 13.4 GW by 2030. The generation mix, outlined in Scenario 1, including some replacement thermal generation, can be accommodated within the proposed transmission upgrades over the period to 2030.
But the more ambitious renewable development in Scenario 2, with renewable capacity of over 15 GW by 2020 and 21 GW by 2030, results in the transmission upgrades currently identified as unable to accommodate this level of renewable generation without constraints. As a result further transmission upgrades may be required, potentially including the development of an offshore grid - ENSG are considering the impact of additional potential renewable generation in Scotland.
6.2 Scope of storage to manage transmission constraints
A key result emerging from our analysis so far is that, given the level of renewable generation in Scotland in Scenario 2, constraints will occur. We now go on to determine, at a high level, the potential contribution storage may make to helping manage transmission constraints.
Our study has been undertaken by assessing the impact of introducing additional storage to manage constraints in the interconnection between England and Scotland. The results and conclusions are only indicative as the study has not considered the costs of providing storage versus the costs of other alternatives to reduce constraints such as; the provision of additional transmission capacity or the curtailment of wind energy.
Our analysis is based on the Scenarios presented above. However, as the generation output in Scenario 1 did not result in significant transmission constraints post-2015 we focus only on Scenario 2 where significant constraints occurred.
6.2.2 Why does constrained energy not equal storage requirements?
Figure 6.3 showed the energy flows from Scotland that may be constrained under Scenario 2 due to lack of transmission capacity currently proposed. The chart showed that, by 2030, power flows could be constrained around 30% of the time (and about 10% by 2020). The potential constrained energy may provide the economic justification for further reinforcement or the introduction of alternative mitigation measures, such as energy storage. As a result we concluded that some 2.8 GW may be required to remove the peak constraint in 2020 and some 7.5 GW by 2030.
However, although Figure 6.3 shows the peak storage capacity that may be required to remove the peak constraint, the constrained energy is not an indication of the total volume ( GWh) of storage that may be required and the duration of time over which it may be needed. Figure 6.3 shows cumulative flows and so cannot indicate the volume and duration of storage due to the sequential nature of the storage cycles and excess wind production. The volume of storage required would depend on the amount of time when excess energy flows occur. Put simply the same amount of constrained energy could be accrued by short spells of high wind output followed by longer spells of low output or vice versa. Any excess energy will have to be stored and released before the storage capacity becomes available again. If the storage capacity is 'full' when an excess energy period occurs, then wind output may have to be curtailed.
In order to examine the effects of variability of wind output on energy flows, storage requirements and network reinforcement requirements it is necessary to examine energy flows (and wind output in particular) on a time series basis. 13
A high level model has been established to examine the benefits of additional storage - in particular pumped storage. The model uses 3 year half hourly wind generation series based on the generation capacities in Scenario 2. The resulting output is combined with half hourly Scottish demand values to arrive at flows of wind generation output net of demand. These flows are then offset by the MW to determine the flows shown in Figure 6.3 from the economic dispatch.
When the unconstrained flows resulting from the model exceed the transmission capacity then pumped storage is used to manage the constraint until the storage is full. Storage is discharged at the earliest opportunity whilst maintaining the resulting flows within the boundary limits. The storage capacity required is dictated by the maximum excess generation over the transmission capacity limit. The amount of energy storage has also been varied to assess the sensitivity to the amount of storage required taking into account the variability of wind generation.
6.2.2 Results - 2030
Figure 6.1.3 indicated that constraints could occur for over 25% of the time by 2030 in Scenario 2. Figure 6.2.1 below shows the power flows from Scenario 2 with both unconstrained power flows (blue line) and the impact of utilising some 7 GW of pumped storage (red line) 14. The flat section in the power flow (red line) corresponds to the maximum transmission capacity that storage is attempting to manage to keep flows within the boundary transmission capacity (some 7.8 GW, assuming all the reinforcements identified by the ENSG by 2020).
Figure 6.2.1: Scenario 2, 2030 - Effect of storage on flows and constraints.
Figure 6..2.1 shows that storage can assist the management of power flows in Scenario 2 with constraints (energy flows above 7.8 GW) falling from around 30% of the time to less than 10%. The constraints generally occur due to longer periods of high wind output where the storage capacity of 10 hours used in this example is insufficient to manage the maximum flow limits.
In order to investigate the impact of storage capacity (in hours) on the management of constraints a further study was undertaken to analyse the sensitivity of constraints to the amount of energy storage available.
Figure 6.2.2 shows the effect on constraints (measured as proportion of time when they occur) of varying the hours of energy storage capacity. The results show that introducing 4 hours of storage leads to a 50% reduction in constraints. Thereafter diminishing returns set in, with 16 hours of storage needed to reduce constraints by 80% but close to 80 hours needed to virtually remove constraints.
Figure 6.2.2: Scenario 2, 2030 - Effect of storage energy capacity on constraints.
Figure 6.2.3 shows the impact of 3.5 GW of storage capacity - around half the level analysed in Figure 6.2.1. Halving the storage capacity from around 7 GW to 3.5 GW leads to an increase in constrained periods from some 10% of the time to over 15% of the time.
Figure 6.2.3: Scenario 2, 2030 - Effect of storage capacity (3.5 GW) on flows and constraints.
Figure 6.2.4 shows that the sensitivity to the amount of energy storage follows a similar pattern, with a rapid reduction of constraints followed by a "plateauing" of benefits in constraint reduction with the amount of stored energy. However, with 3.5 GW of storage, constraints cannot be fully relieved and are reduced by a maximum of 75% regardless of energy storage volumes.
Figure 6.2.4: Scenario 2, 2030- Effect of storage capacity (h) on flows and constraints for 3.5 GW.
6.2.3 Results 2020
Similar conclusions are obtained from the analysis of 2020 capacity and power flows for Scenario 2, but on a reduced scale given the smaller constraints that must be overcome by 2020. In Scenario 2 by 2020 constraints without storage could amount to about 10% of the time and peak at about 2.8 GW.
Figures 6.2.5 and 6.2.6 show the flows with and without storage using the methodology described above, and the sensitivity to the amount of stored energy.
Figure 6.2.5: Scenario 2 - Effect of storage capacity (2.9 GW) on flows and constraints. 2020.
6.2.6: Scenario 2 - Effect of storage capacity (h) on flows and constraints for 2.9 GW.2020
The results show that storage again can help alleviate constraints, but that a relatively significant amount is required.
Ultimately the merits of providing storage must be assessed against the alternatives, such as including provision of further network reinforcements, interconnections, seasonality aspects and scope for use of storage heaters etc. This more detailed analysis is outside the scope of this study but the findings above could be used in such assessment.
It is also worth noting that the analysis of the impact of storage above focuses on all generation output in Scenario 2. Figure 6.1.4 showed that constrained energy reduces if renewables are given priority access over thermal plant - in short thermal plant flexes down to accommodate only renewables. If thermal plant was constrained off then no constraints would appear by 2020, thus reducing potential storage requirements. As a result a further investigation should explore the costs of constraining thermal plant versus the construction of energy storage. However, some indicative analysis is provided in Section 6.
Our analysis indicates that, given the transmission upgrades currently outlined by ENSG, the increase in Scottish generation outlined in Scenario 1 will not trigger the requirement for substantial energy storage or demand side management measures. The role of existing pumped storage in Scotland is likely to provide sufficient energy management capacity.
However, the increase in generation, particularly renewables, outlined in the more ambitious Scenario 2 leads to network flow problems despite the currently proposed transmission upgrades. Put simply, the volume of generation outlined in the Scenario cannot flow on the system as it is economically dispatched. Even if thermal plant could be fully constrained off (with the associated costs) the transmission upgrades identified are insufficient to export the remaining renewable generation.
As a result there may be a role for energy storage and demand management in addition to further transmission upgrades. Installing energy storage capacity of some 7 GW by 2030 would alleviate the constraints identified, with most of the constraints relieved with less than 10 hours of storage. Installing 3.5 GW of storage will alleviate 75% of constraints.
While energy storage could play a role in alleviating the transmission constraints, the costs of additional storage must be set against the cost of both transmission upgrades and the costs of constraining thermal plant off the system. Is it economically rational to store electricity generated from thermal plant with a positive marginal cost?
Furthermore, connecting offshore wind in Scotland directly to England via a subsea cable may help alleviate transmission flows and thus reduce the need for storage.
Under the operation of the present Security and Quality of Supply Standard it is likely that the generation mix and resulting flows in Scenario 2 will trigger the requirement for transmission upgrades.
The results of our 2020 analysis concur with the findings of the ENSG (Electricity Networks Sub Group) who identified the need for the Western subsea link as 'very strong' and 'strong' for the Eastern subsea link. The results also indicate that, if Scottish renewable aspirations increase above the current 2020 target, then measures may be required, including potentially additional transmission capacity may be required and/or energy storage and DSM.
6.3 Technology Assessment
The various energy storage technologies were rated on a matrix scoring their overall technology status and their suitability for application to Scotland. The scoring matrix is provided in Appendix 1.
A high storage requirement was identified by SKM's modelling work under stage 1. To take this factor into account a weighting was placed upon the technologies ability to store significant quantities of power as this will be the greatest challenge.
The assessment criteria and weighting used were:
Table 6.3.1 Assessment Criteria
Efficiency of recovery
Future Potential advances
The technology with the highest score was Pumped Storage. While not being the cheapest technology pumped storage scored very highly against Storage Capacity, Technical Maturity and Scottish Infrastructure.
Figure 6.3.1: Technology suitability to Scotland, green representing energy storage technologies and pink representing power quality technologies.
6.4 Top Three Energy Storage Technologies
6.4.1 Energy storage technologies
Stage 1 of this study identified the need for significant volumes of energy storage under Scenarios of high wind penetration. As such, a key criteria of energy storage technologies for Scotland is that they can store significant amounts of energy. Three technologies have been highlighted as those that have the greatest applicability to addressing the problem of increasing intermittency in Scotland. Of these three pumped hydro had the greatest overall compatibility confirming why it is considered as the number one energy storage option worldwide.
Pumped Hydro: a proven technology for dealing with fluctuations in energy demand or production. The ability to store large amounts of power (100s of MW) and also respond quickly to changing demand makes it a versatile technology. Scotland has a track record of using this technology and therefore it presents the least risk option to the network operators of all the technologies. As part of the stakeholder consultation limited risk was an important criteria. The main barrier to wide scale deployment is suitable geography and environmental constraints with significant local alteration occurring to landscapes and wildlife.
CAES: the only other large scale energy storage technology than pumped hydro. The technology offers flexibility in functioning as a conventional fossil fuel plant if long periods of low wind production occur. A major constraint upon this technology is suitable geography of which there are no natural caverns in Scotland. Investigating the potential of disused coal mines is recommended to determine whether this exists as an option or not.
Flow Batteries: this technology is much smaller in scale than the two above although future technological advances may increase the scale of the battery storage. The major current barrier is technology cost. If this is addressed flow batteries would represent an ideal decentralised energy storage solution particularly to remote areas of the grid and the islands of Scotland. At present this technology is being trialled in reducing the need for line upgrades. Operational trials of larger devices functioning primarily as storage devices are required in Scotland before the net economic benefits can be evaluated. In order to overcome the present economic barriers a number of revenue streams for investigation have been highlighted in the regulatory section. Flow batteries can also be deployed as power quality devices, the discussion under power quality highlights the multiple applications that flow batteries could be used for.
6.4.1 Power quality technologies
These technologies are suited to deal with excess wind generation and/or alleviate distribution lines from being overloaded. Many of the UK utility companies are already investigating using power quality technologies for energy storage along certain distribution lines. Installing smaller energy storage technologies such as advanced batteries can mitigate the need to upgrade specific electricity lines and thereby save the utility companies significant costs. Increasing the proportion of intermittent generation and in particular wind can make distribution lines particularly vulnerable to overloading. It is beyond the scope of this project to consider which specific lines will need support from storage technologies as this is line specific and will be better understood by the utility companies. We have however listed those power quality technologies that are likely to have an important role either being installed as part of a wind farm and hence improving the winds dispatchability or negating the need to significant line upgrades in areas with high levels of wind capacity.
The most promising power quality technologies in the Scottish context are:
Redox Flow batteries: whilst also being a energy storage technology the flexible nature in which flow batteries are able to store and release electricity makes them suited to situations such as improving line stability. Vanadium flow batteries can operate thousands of cycles without needing replacing or suffering from performance degradation, this represents a significant advantage over other advanced batteries. The barrier to extensive development over recent years has been the high cost associated with the technology, this has hindered flow batteries from being considered as large scale storage devices of tens of MW.
Flywheels: this technology has the ability to smooth out short term (0-30 minutes) frequency variation in wind output. This ultimately improves the quality being supplied to the grid and can enable a greater amount of power to be exported from the wind farm.
Both redox flow batteries and flywheels are flexible in their deployment and therefore can be considered as decentralised solutions that would also be applicable to remote parts of the Scottish grid network.
As previously reported in the regulatory section Scotland has the opportunity to become a 'first-mover' in this sector providing support towards smaller decentralised technologies. These technologies have potentially multiple different revenue streams and wider benefits such as enabling further connections of distributed generation.
The main revenue streams can be summarised as:
- Voltage fluctuation regulation
- Avoiding or deferring the need for line upgrades
- Potential black start capability, depending upon size
- Arbitrage trading
Within the scope of this project the total value is not quantifiable as the current revenue streams are not clearly defined and in some cases such as avoiding the need for line upgrades are very site specific and require a detailed site specific assessment. The Scottish Government have an opportunity to investigate ways of combining the benefits so that storage/power quality devices are more economic.
6.5 Energy Storage requirement
6.5.1 Capacity of storage required
Under Scenario 2 it was found that there is a significant storage requirement needed in 2030. Considering the case of Scenario 2 Figure 6.5 showed that in order to reduce the amount of time constraints are required upon generation to 10% then approximately 7,000 MW of storage is required.
To put this figure into context the current size of pumped storage schemes in Scotland and those planned is around 400 MW. Installing 7,000 MW of storage would be a massive engineering undertaking and would equate to over seventeen 400 MW plants. Installing 17 storage plants of this size could equate to a cost of £11.2 billion. It is highly unlikely that there are as many as 17 further sites in Scotland that would be suitable for 400 MW pumped storage schemes. Previous investigations have examined sites such as one on the east of Loch Lomond. Hence there may be potential for some further development. Establishing the potential would require a detailed assessment of the potential sites
Hence irrespective of any economic assessment, it is likely that pumped storage could only provide a fraction of the storage needed in 2030 under Scenario 2.
The other alternative bulk storage technology, CAES, also suffers from uncertainty over the location, size and practicality of potential storage sites. No study has been undertaken of this in Scotland, hence there is no certainty that any scheme could be developed.
In summary bulk storage technologies cannot provide the full level of storage required under Scenario 2 for 2030. Hence other solutions such as transmission upgrades will need to be deployed along with, or instead of, bulk storage technologies.
6.5.2 Demand side management potential
An alternative to reduce the amount of energy storage and time that constraining generation is required would be to increase electrical demand in a flexible manner. This is most likely to occur through the uptake of electric vehicles and electric heating. Both of these demand side management options would require an extensive roll out of smart metering and smart grids as an enabling technology for them to work in conjunction with power production. During the 2020s it looks likely that a large proportion of Scotland may have both smart meters and grids.
In light of the above it is therefore useful to consider the scale of the contribution that electric vehicles could make. In 2008 there were about 2.2 million cars in Scotland (DfT 2009). The CCC has presented the projections of electric vehicle growth up to 2022. For consideration of the scale of the contribution that this could make in 2030 we extrapolated the CCC growth trajectories to 2030. On this basis it has been assumed that that by 2030 electric vehicles may represent 42% of all new car sales in Scotland. Assuming the sales of new cars stays constant into the future (currently 7.6% for Scotland) then by 2030 there could be 560,000 electric vehicles in Scotland by 2030.
Advances in charging and battery technology for electric vehicles are expected over the next 10-15 years. Fast or variable speeds of charging are likely to be best suited to providing a balance to a grid comprising of high levels of intermittent generation. To test the maximum charging demand we have assumed that car owners will not have empty batteries in their cars, therefore we have assumed that average non-charged battery levels may be 50%. Taking 75kW as the potential level of fast charging capacity in 2030 the instantaneous energy demand would be 21,000 MW for 20 minutes or 7,000 MW spread over 1 hour, 700 MW for 10 hours.
Electric resistance heating of domestic hot water represents another technology to increase demand and as previously reported in section 4 the load could represent 7,000 MW for 20 minutes or 2,333 MW over an hour.
Similarly, electric space heating of domestic properties could provide a switchable load in section 4 of the report it has been estimated that this could provide around 1,000 MW of additional demand.
The above estimates of additional electrical demand are extreme estimates, for example all vehicles will not need charging regularly. Increasing the electric load into a functioning smart grid of the above options could represent significant contribution towards reducing short term sporadic peaks. Over a 10 hour period of excess generation the contribution becomes more modest. Combined they would represent 933 MW per hour. This represents 12% of the total storage required meaning that other solutions are still required.
The third solution is further interconnection links beyond those that were envisaged by the ENSG. This is a well proven solution and unlike pumped storage or CAES there are no geographical or geological constraints to their use or capacity.
In addition the capital costs of HVDC interconnection are lower than pumped storage.
6.5.4 The Economic case for pumped storage
There are numerous studies on the relative costs of different conventional and renewable energy technologies. Despite many of these being long established technologies, different views on the relative costs of the different generation types remain.
In comparison there are far fewer studies on the economics of energy storage systems. This is perhaps because this requires a much greater level of detailed data and analysis. A number of studies were found during this project, however none were specific to the UK or Scotland.
In this section we look at two different high level approaches to assessing the costs for pumped storage.
Because these approaches draw on international analysis and data they should be used with care. As an example a Swiss study ( ETH Zurich 2010) quotes capital costs of â'¬600 to 3,000 â'¬/kW for a range of projects in Austria and Switzerland. This study also uses a example project of 1,400 MW project with a head of 600 metres. The larger of the 2 Scottish schemes has a capacity of 400 MW and a head of just under 400 metres.
Hence these studies may not represent the site specific costs that would be seen in Scotland. However as no pumped storage scheme has been built in recent times there are no Scottish data available. The closest example is the natural flow Glendoe hydro scheme which cost £150 million and has a capacity of 100 MW. This is £1,500/kW - i.e. in the mid range quoted above, but without the extra cost of the reversible turbine to pump water back to the reservoir.
With no definitive data or previous UK policy studies to draw upon the following sections use two different methods to estimate the costs of pumped storage.
Assessment Method 1
For this study a search was under taken for evidence and results, One of the few recent studies was undertaken for the US Dept. of Energy in 2007 (Sandia National Laboratories, 2007).
This assesses the levelised cost of storage capacity - in terms of $/kW per year. This analysis assumes that each storage device discharges once per day and that the system operates 250 days per year (i.e., 5 days/week, 50 weeks/year). Hence based on this level of duty a 10 MW device with a cost of $400/kW per year would need to earn income of 10,000 x $400 = $40 million per year of operation to break even.
In this US study the technologies and analysis results were grouped as:
- Bulk Storage Technologies - with 8 hour discharge time.
- Distributed Generation - with a 4 hour discharge time.
- Power Quality Technologies - with a 20 second discharge time.
The results of the Scenarios show that in 2030 Scotland would require Bulk Storage Technologies, the results for this are shown below:
Figure 6.5.1 Levelised Costs of Bulk Storage Technologies
While the results are US based costs, the relative ranking of the results are likely to be valid. Hence this suggests that:
- Compressed Air Energy Storage ( CAES) - has the lowest costs at around $250/kW per year
- Pumped Hydro has the next least costs - around $300/kW per year
- The remaining technologies have much higher costs from $500 to $1,300/kW per year.
For Scotland the deployment of CAES is uncertain as the availability of suitable underground caverns is unknown. In addition CAES uses gas, which is not consistent with the renewable energy targets in Scotland.
Hence pumped hydro (with or without variable speed) is the Bulk Storage Technology which offers the lowest cost solution for Scotland, the rival technologies have significantly higher costs.
Using this US data a 400 MW pumped storage scheme would need to earn around 400,000 x $400 = $160 million pa. Using an exchange rate of £1 = $1.44 this is £111 million pa. A 400 MW pumped storage scheme would not cover all of the storage needs identified in the analysis of Scenario 2 - this shows a peak gap of 5 GW in 2030, implying a need for up to 12 similar 400 MW pumped storage schemes and hence annual costs of many £100 millions.
Assessment Method 2
As summarised by SEI (2004) from Kroon (2002) there are three main categories of revenue stream for energy storage investments, these revenue streams are not exclusive.
Table 6.5.1: Pumped Hydro Revenue Options
Category 1: Power quality and power management
a) Preventing voltage dip
b) Prevents cascading grid failures
Category 2: Tariff trading
c) Peak shaving arbitrage
d) Coverage of high value power imbalance requirements
Category 3: Energy storage for the integration of renewable and distributed generation
e) Improves dispatchability of renewable energy
f) Allows curtailment of base load conventional generation capacity
g) Reduces cost and complexity of integrating renewable energy to transmission, distribution systems as well as incorporating renewables into the new Market Arrangements for Electricity
Energy storage technologies typically earn revenue from streams (c), (d), (e), (f) and (g). Pumped storage has primarily functioned on point (c), peak shaving arbitrage. Increasingly it is enabling greater volumes of renewables onto the network.
Hence this assessment focuses on the potential revenue from (c). To assess this requires an estimate of the load factor for pumped storage.
Table 6.5.2: Scottish pumped Hydro generation figures (source of generation figures DUKES)
Generation ( GWh)
The installed capacity of pumped hydro in Scotland is 700 MW. This is from two sites Cruachan (400 MW) and Foyers (300 MW). The load factor in 2006 and 2007 was almost double that from previous years. As reported by MacLeod et al (2005) the historic load factor for pumped hydro from 1996-2002 in Scotland was 9%. This recent increase in usage indicates a greater requirement for storage triggered by the increasing intermittent capacity installed in Scotland over this period and perhaps changes in market structure.
Pumped storage, and other energy storage technologies can generate revenue from consuming energy at time of surplus and cheap prices and subsequently selling on during peak demand periods. The economics of pumped storage for arbitrage trading is that the pumping price has to be at least 25%-30% cheaper than the selling price (Deane et al 2010).
The average market price for the UK was £37/ MWh in 2009. Assuming that the selling price must be at least 30% higher than the average then the average market selling price must exceed £48/ MWh. During 2009 this occurred in 14.1% of the half hour periods during the year.
As previously reported the average cost of pumped hydro is £1,650/kW, this is higher than for conventional hydro due to the additional cost of a reversible turbine and often that an upper reservoir is required. This average figure is potentially lower than a complete new construction as it includes the cost for new pumped hydro projects plus costs for retrofitting reversible turbines to existing natural flow hydro projects. However, using the £1,650/kW value, a 400 MW project would cost in the region of £660 million. Using a load factor of 20% annual generation would be 700,600 MWh per annum. If a payback period of 10 years is required a net generation revenue of £66 million a year would be needed, added to this would be the energy purchased cost of £27 million (@ £30 MWh). This would mean that the average price received for every MWh sold would need to be £133 MWh. This is a high value which is currently only experienced in only a few time periods, see figure below.
Figure 6.5.2 GB Electricity Market Prices
In future prices in the electricity market may become more dynamic, with greater variation in price levels as the market reacts to the increasing levels of renewable generation. However this is uncertain which means the economic returns for storage are less certain.
Table 2.5.3 Assumptions of pumped hydro costs
Generation for sale (20% output)
Energy input (77 % efficiency)
Cost of energy purchased
Annual revenue required
£93.3 million (assuming 10 year pay back)
As assessment method 1 highlighted the requirement for pumped hydro would be more than one plant, therefore the total revenue generated from peak shaving would need to be many times more than the annual revenue requirement stated above. The results from the modelling indicate, by 2030, over 15% of the time renewable only generation will exceed the demand and export capacity. This could result in significant constraining costs. The above analysis shows that the cost of constraining wind however is of a similar value to the annual revenue required to make pumped hydro economic. Therefore in future markets where wind may need to be constrained the value of storage will be much greater to the generation companies. In addition there are a number of other revenue streams that will contribute to the economics of storage, notably (d), (e), (f) and (g).
Assessment Method 1 suggests that a 400 MW pumped storage scheme would need to earn around £111 million a year to cover the costs of capital and operation. Actual earnings would need to be higher than this to provide a profit on top of the breakeven costs.
Assessment Method 2 suggests that a 400 MW pumped storage would need around £93 million of income a year.
A 400 MW pumped storage scheme would use 911,000 MWh of electricity a year for pumping to re-charge the top reservoir. If this electricity were from onshore wind and costs £100/ MWh to constrain off, the cost of this level of constrained generation is £91.1 million a year.
These cost estimates are similar in scale - indicating that there is not a clear cut opportunity for pumped storage. However there will be differences in the capital and operational costs of actual pumped hydro projects.
More generally there are a number of cost issues that could change these results:
- As renewable energy increase there may be larges changes in minimum, maximum and average electricity prices. Because renewable energy receives ROCs and has low operational cost, wind and other renewable generation can operate at times of low electricity prices. Hence the cost of electricity for the pumping phase could fall. Similarly when wind output is low, electricity prices could peak at higher levels than seem now - offered greater income for pumped hydro. These are complex market issues which are not part of this study.
- Average capital costs have been uses some sites may have lower costs if the geography allows smaller dams and less pipework.
- Installing pump-back systems on existing hydro-electric facilities. This was identified as a popular trend in Europe as many of the economically attractive sites have been taken (Deane et al 2010). The costs of dams are avoided. In Scotland there are likely to exist a number of opportunities to install pump back systems on existing dams. In a study undertaken by Strathclyde the pumped hydro potential was identified as 494 GWh from 13 sites (Sloy not included, as this is already being investigated).
- Constraining conventional hydro-electricity before intermittent renewables.
Pumped storage is the most cost effective of the large scale energy storage technologies that can be used in Scotland. A high level assessment of pumped storage economics suggests that new pumped storage would be more expensive than constraining renewable energy generation off the network. This assessment does not include other potential revenues that a pumped storage device may be able to access but these are unlikely to revise the economic case radically. The conclusion reinforces the view that accommodating the forecast generating plant mix under the highest renewable energy Scenarios will require a wide range of solutions of which the various demand side response measures enabled by the smart grid and smart meters will be an important and cost effective component.
6.5.5 The Economic case for distributed storage
This study has focused on the need for large scale bulk energy storage technologies, comparing these with the results from the Scenarios. There may also be a role for smaller scale storage technologies that are distributed across the network and connected at lower voltage levels. This is particularly true of remote areas especially the Scottish islands. In these areas renewable energy generation potential is high, but the local demand for energy is low due to the sparse population. Hence local energy storage is a potential solution.
The results of an economic analysis of these technologies is shown below:
Figure 6.5.3 Levelised Costs of Distributed Storage Technologies
Again CAES appears the most cost effective - however this technology requires gas as a supplementary fuel. Gas is not normally available in the remoter parts of Scotland where these technologies may be of most value.
Hydrogen engine is the next cheapest technology, followed by Na/S batteries and hydrogen fuel cells. Of these hydrogen has some supply chain in Scotland and a number of pilot/demonstration projects.