Beyond the spin of energy storage
- Posted on June 26, 2018
- 174 views
The Irish electricity grid combining the Republic of Ireland and Northern Ireland is under-taking a unique experiment as it seeks to integrate ever greater amounts of renewable energy – primarily wind - on a relatively isolated, island grid.
Figure 1: Ireland’s target of 40% renewable electricity – primarily from wind – is the highest in the EU for an isolated, synchronous system
The Irish grid serves a population of circa 6.6 million and had a demand of ~38TWh in 2017. Gas supplies somewhat more than 50% of generation – primarily via combined cycle plants - with coal and peat providing around 20% and the balance being mostly onshore wind. There is a small pumped hydro unit of 290 MW/1600 MWh and 2 HVDC inter-connectors to the UK of 950 MW. A further inter-connector to France of similar capacity is planned for the 2020’s
In April 2018 the transmission system operators (Eirgrid in the Republic of Ireland & SONI in Northern Ireland) reported that the instantaneous system non-synchronous penetration – or SNSP - had reached >60% on more than one occasion. The goal now is to move from a design basis of integrating 60% SNSP at present to 75% by 2020, as ever increasing amounts of wind and other non-synchronous renewables are added to the grid.
In order to support such a high percentage of SNSP, the system operators (SO) have developed a suite of enhancements that includes system tools as well as grid codes and policies. This programme is referred to as ‘DS3’.
The aim of the DS3 Programme is to meet the challenges of operating the electricity system in a secure manner while achieving EU-directed 2020 renewable electricity targets.
DS3 is a multi-year programme begun in 2011 and has seen various enhancements added as SNSP has increased. A key component of DS3 is the necessity to provide ancillary services.
It is an imperative of DS3 to facilitate greater volumes of renewable energy as well as to minimise curtailment of wind farms. The cost of DS3 should also be less than the reduction in wholesale electricity costs, so as not to burden consumers.
Conventional ancillary system services on the Irish Grid are listed in the table below
Table 1 : The 7 ‘conventional’ existing system service products – all of which can be supplied by
thermal plant, pumped hydro or inter-connectors
The DS3 programme identified 7 additional ancillary services as below :-
Table 2; the 7 additional services being introduced from 2016 onwards
4 new services (SIR and the 3 ramping margin services RM1,3 & 8) were introduced in Oct 2016 and a further 3 “fast acting “ services (DRR, FFR & FPFAPR) are to be introduced on a staggered basis from Oct 2018, making 7 new services in addition to the 7 conventional services to be in place by 2020.
Service requirements/ system conditions
All ancillary services are required to operate under the following system conditions:
- Demand range 2000 to 7000 MW
- Wind ranging from 0 to 4600MW
- Full import to full export on interconnectors
- Largest single infeed/outfeed 200 to 530 MW
- Transmission infrastructure build-out and outages
Sources of system services
Ancillary services are not a new thing on any AC grid & they have traditionally been provided by conventional generators as indicated in table 3
Table 3 : Current sources of ancillary system services . Only OCGT can supply the key
ramping services RM1 & 3, as well as replacement reserve RRD
Key new services – SIR, FFR & DRR
One of the most important ancillary services is Synchronous Inertial Response or SIR.
SIR is the response in terms of active power output and synchronising torque that a spinning unit can provide following disturbances. It is a response that is immediately available – in less than 10 milliseconds - from synchronous generators, condensers and some demand loads. It is a key determinant of the strength and stability of the power system, and also has significant implications for rate of change of frequency during power imbalances and for transmission protection devices and philosophy. With increasing non-synchronous generation this response becomes scarce.
Technically SIR is the stored kinetic energy term in MW.seconds multiplied by the SIR factor , also in seconds - giving units of . The SO s estimate that a volume of 15,473 million units of this service will be required annually by 2020.
SIR runs in conjunction with Fast Frequency Response and both have implications for the Rate of Change of Frequency or RoCoF.
Fast frequency response service – FFR
FFR can supplement any inherent inertial response and with the appropriate control systems, can be supplied by both synchronous & non-synchronous generators.
FFR is a power service measured in MW. It is required to respond in 2 to 10 seconds- ahead of the Primary Operating Reserve. As such it can slow RoCoF and lessen any impact of a frequency disturbance. This service is complimentary to SIR.
Figure 2 : Both inertial response and fast frequency service impact significantly on RoCoF.
Dynamic Reactive Response DRR
At high levels of SNSP there are relatively few conventional synchronous generators on the system, and there is increased distance between them. The synchronous torque holding these units together is therefore weakened. This can be mitigated by an increase in the reactive response which can be supplied by e.g. CCGT. DRR can also potentially be supplied by windfarms. This service is measured in the same units as reactive power – MVaR.
Rate of Change of Frequency - RoCoF
RoCoF is a grid standard and the regulator has made a decision in principle to increase this from 0.5Hz/Sec to 1Hz/Sec (measured over 500 millseconds) under the DS3 programme, in order to Increase the system’s resilience. Some utilities have indicated that this may cause problems for them and it is not clear at this point if this standard can be implemented.
Inertia and frequency response services from various technologies are listed below in merit order :
Above : Ranking of inertia and frequency response technologies in the order of best providers with respect to delivering control over RoCoF. Synchronous technologies are in the top 6. Technologies 7-13 can only provide a digital inertia response – i.e they have to be instructed by means of control systems. The Irish grid does not have 1 or 3 available. HVDC interconnectors are de-rated by 50% from an availability perspective.
Table 4 : System services technical explanation and required response times as provided by the Irish SO s
The cost of ancillary services on the Irish grid are capped, via a ‘glide path’ to 2020, as indicated in the graph below.
Figure 3 : Cost projection for DS3 ancillary services is capped with a ceiling of €235 million p.a. by 2020
As previously stated a requirement of DS3 is that the cost of ancillary services should be less than the savings in wholesale electricity prices resulting from the greater volumes of renewable energy facilitated by it.
However a reduction in the wholesale price has a knock on effect in that it increases the Government subsidy paid to renewable generators and recovered from consumers via a separate levy - the PSO. This levy amounted to €400 million in 2017.
Furthermore whereas the savings are intended to reduce curtailment of windfarms, there is another factor – namely constraints imposed on renewable generators at the distribution level. In 2017 these were a further cost of €129 million to be borne by the consumer.
There is a further important issue that needs to be considered and that is the impact of higher levels of renewables penetration on CO2 and other emissions.
As non-dispatchable renewables penetration increases, conventional thermal plant operate considerably less efficiently. For example a modern CCGT plant operating in optimum conditions emits ~ 350gCO2/kWh However, operating on partial load with frequent ramp up/down as a compliment to stochastic renewables , the carbon intensity can rise dramatically. Other emissions such as SOx and NOx rise accordingly.
Recent research has shown that operating in this manner, Ireland’s CCGT fleet had an average carbon intensity of 575g/kWh in 2014-15 with 23% wind generation when allowable SNSP was set at 55%. The same authors predict that by 2020 this is expected to reach ~720g/kWh – comparable with new build coal generation.
Indeed in 2016 overall carbon emissions from power generation in the Republic of Ireland - with wind penetration at 23% - were on an upward trend at 483g/kWh.
Batteries are seen as one possible solution to address the inefficiencies and higher emissions intensity of the gas turbine fleet in respect of providing fast-acting grid services.
Lumcloon battery storage
Lumcloon energy proposes to build 2 battery energy storage systems (BESS) of 100 MW capacity and 150 MWh energy each on 2 different sites. The units will use South Korea’s HanWha Energy technology, most likely based on their ESS 2000-6000 ‘smart energy solution’ which is a battery rated at 2 MW capacity and 3 MWh of storage. 50 of these modules stacked would yield an array with 100 MW of capacity and 150 MWh of storage. These units use Li-Ion batteries.
Lumcloon & HanWha will work with LSIS, another South Korean supplier of smart power solutions in transmission, distribution and automation including switchgear and smart grids
Table 5 : list of services proposed to be provided by LES in green. Those services in high demand by the SO are marked with an asterisk . The DDR service here actually refers to DRR.
The services proposed to be provided by Lumcloon are listed in table 5. It can be seen that these include all the 7 new services as well as 2 conventional services.
As a battery cannot supply synchronous inertia, the SIR service indicated actually relates to provision of a future “synthetic inertia” service.
Synthetic or digital inertia is the new sexy on smart grids. At high levels of SNSP there is a scarcity of synchronous generators. Synthetic inertia could potentially be supplied by batteries with the appropriate fast-reacting power control systems. Although batteries do not provide spinning mass, digital inertia can potentially provide the same benefits – or better - as conventional inertia.
However, there are issues around the capability of digital units to detect a frequency excursion and meet the appropriate response time of less than 200 milliseconds from an event.
Furthermore, RoCoF is the 2nd derivative (F where F = phase), so it is very sensitive to noise. Accurately capturing a frequency excursion in real time electronically from a high voltage AC system when you most need it – i.e during an earth fault, generator trip or power surge, when the grid is ringing with harmonics, disturbances and inertial responses from other generators – well, It’s a handful.
Accordingly, the SO s have been running trials that include a mix of synchronous and synthetic inertia. While the conclusions have been generally positive - provided that assets can provide partial response within 0.1 secs and full power delivery within 0.2 secs - further work is required to see whether synthetic inertia can provide the same or better level of system integrity as provided by conventional rotating plant.
Figure 4 While the response of synchronous inertial devices is automatic and immediate [under 10 milliseconds], synthetic or digital inertia requires to detect, measure and respond to a RoCoF signal in less than 200 milliseconds in order to be effective.
However batteries can potentially supply almost 10 times the power rating of conventional generators in inertial terms – as indicated by the relative scale of the different services.
Lumcloon is not the first BESS on the Irish grid. Since Jan 2016 US-based AES Corporation has been trialling a 10 MW/5 MWh array at Kilroot power station in Northern Ireland, which is seen as a fore-runner to an eventual 100 MW system. It is currently the largest such BESS in the UK. AES has been working with Queens University Belfast on developing and demonstrating this system , & this work is supported by a grant from the UK Government via Innovate UK
Above : An AES battery array. AES has used batteries supplied by Samsung SDI
AES have yet to make an announcement as to whether they intend to proceed with their larger 100 MW array.
AES is a company with considerable experience in the battery storage field where it recently partnered with Siemens in a joint venture to develop this area. In 2017 it delivered a 30 MW/120 MWh plant as part of the Aliso Canyon BESS array in Southern California alongside Tesla. California has a mandate to build 1,300 MW of storage power by 2020.
Summary – Beyond the spin
The green media has jumped on the recent announcement of this project, claiming that it is a grid-scale battery whose primary purpose is to store surplus renewable power produced at night time and release it to the grid during the day.
A battery can of course store electrical energy and can provide energy services i.e Primary Operating Reserve (POR) and the 8 hour ramping margin RM8.
Neither POR nor RM8 can be considered as significantly contributing to energy storage. POR has a duration of 5 to 15 seconds and ramping margin is designed to assist other generators in integrating fast-changing variations in supply and demand. The RM8 service at Lumcloon would at best deliver ~20MW per hour over an 8 hour period, which equates to less than 0.5% of hourly average demand from each battery plant.
Even if Lumcloon could theoretically provide full reserve energy – at 100MW each plant could only provide less than 2.5 % of demand for 1 ½ hours – hardly the kind of service that is going to enable Ireland’s electrical grid to run on renewables and storage – which is the message that is been marketed.
One area where a BESS could score is inertial response. Batteries can potentially supply this service with 10 times the power rating of synchronous generators (see figure 4)
Clearly LES aims to target the more exotic enhanced services where competition is from a limited number of OCGT plants or indeed where there is no other provider online or able to meet the response times/revised RoCoF standard of DS3 going forward.
While its commercial imperative will be to buy off-peak electricity at a low price, it’s DS3 ancillary services contract will require it to be available to deliver specific services at a specific time and in the required volumes as determined by the SO. Accordingly Lumcloon and similar batteries around the world are being deployed to deal with the problems that stochastic renewables are causing on the grid.
Whether the exotic Lumcloon Energy Storage project can be financed on the back of a DS3 contract, and how much of the ancillary services market it can capture, will be interesting to watch.
About the author
Diarmuid Foley is a research engineer and entrepreneur with over 30 years experience at senior level in industry. He is graduate of University College Cork and holds an MSc in Nuclear Engineering from Imperial College London. Current areas of research include the economics of power & heat generation systems.
Other key components include integrating large amounts of Demand Side generators and a faster RoCoF requirement of 1 Hz/sec
Via a reduction in the SMP – the system marginal price. See http://www.sem-o.com/Pages/default.aspx
Article 16 of EU Directive 2009/EC/28 requires member states to minimise the curtailment of renewable energy.
SEM-13-098 DS3 System services technical definitions decision paper – final. Available at https://www.semcommittee.com/publication/sem-13-098-ds3-system-services-technical-definitions-decision-paper-final-0
Economic appraisal of DS3 system services for the RAs – IPA report, July 2014 , available at https://www.semcommittee.com/sites/semcommittee.com/files/media-files/SEM-14-059b%20IPA%20Report%20-%20Final.pdf
The SIRF is defined as the ratio of kinetic energy to the exported power i.e MW.Seconds/MW so yielding units of seconds. The Irish grid uses a SIR factor between 15 - 45 seconds.
2 Hz/Sec is proposed in Northern Ireland
Report by Maria Tsagkaraki and Riccardo Carollo of Incoteco available at http://euanmearns.com/co2-emissions-variations-in-ccgts-used-to-balance-wind-in-ireland/
See Alternative solutions project for more details. http://www.eirgridgroup.com/site-files/library/EirGrid/RoCoF-Alternative-Solutions-Project-Phase-2-Report-Final.pdf
Annual usage of ~ 38 TWh in 2017 equates to an average hourly usage of 4,340 MWh .[all-island basis]