There are growing noises suggesting that the UK’s only large seasonal gas storage facility, Rough, which closed in 2017 may re-open for the winter. Centrica’s storage licence was renewed late last month, and on 3 August, Ofgem granted Centrica an exemption from the Third Party Access Regime for Rough, allowing the company to operate a capacity of 28 bcf (0.8 bcm) for winter 2022/23 and 59 bcf (1.6 bcm) for winter 2023/24. Without this exemption, Centrica had indicated it would not be willing to go ahead with the investment needed to bring the facility back into use. In its heyday, its 3.3 bcm capacity was 70% of all the gas storage capacity across the UK market, and it had a withdrawal rate of around 44 mcm /d. This all sounds marvellous, but I have many questions about how realistic it is to expect Rough to operate at even one quarter of its capacity this winter.
Background to the Rough gas storage facility
The Rough reservoir is a depleted gas field located approximately 29 km (18 miles) off the east coast of Yorkshire, in Rotliegendes sandstone, 2.7 km beneath the sea bed. The reservoir is approximately 10 km (6 miles) long by 3 km (1.8 miles) wide and varies from 24 metres to 36 metres in depth. The reservoir is relatively rare geologically, being formed of rock with broadly uniform porosity which enables gas to flow through it. (Most other reservoirs tend to be more “compartmentalised”.) The porous rock that forms the Rough reservoir is surrounded by non-porous rock, meaning that gas can be pumped into the rock formation and held there under pressure.
In 1980 when British Gas bought the field, about a third of the reserves has been depleted leaving about 7.2 bcm of recoverable reserves, and the reservoir pressure had declined from 4,535 psia (31 MPa) to around 2,800 psia (19 MPa). With the declining reservoir pressure, had normal production continued, compressors would have been required on the production platform. Since it was deemed impractical to use the same compression modules for both injection and production purposes, the storage facility was designed so that compression would only be required on injection.
A configuration of wells, cushion gas and pipelines was selected so that no production compression was needed. A reservoir pressure roughly 1,000 psia higher than the existing pressure at the time was required – under normal storage operations, the maximum operating pressure was 3,500 psia. At the time of closure the cushion gas was estimated to be 1.8 billion therms (5 bcm).
The Rough storage facility comprised two offshore installations (47/3 Bravo and 47/8 Alpha), and a terminal at Easington, which was used for the injection and withdrawal of gas to and from the reservoir:
Gas was injected and withdrawn via the 47/3 Bravo (47/3B) platform, the main offshore complex, operating 24 wells across 3 linked platforms.
47/8 Alpha (47/8A) had 6 wells and was used to maintain deliverability of gas from the field during peak demand days. It was permanently withdrawn from service in September 2016.
Closure of Rough due to extensive well integrity problems
In 2017, the Competition & Market Authority published its preliminary decision to allow the suspension of the Rough Undertakings ending the operation of Rough as a storage facility and allowing the production of all recoverable gas in the facility (the cushion gas). This document described the reasons behind the decision to close the facility.
In March 2015, following technical reports on the integrity of its wells, Centrica Storage Limited (“CSL”) announced a decision to limit the maximum operating pressure in Rough from 3,500 psia to 3,000 psia while it conducted tests to establish whether the wells could be safely operated at the higher pressure. An independent well testing program (conducted between March 2015 and June 2017) demonstrated that the Rough wells were susceptible to a range of unpredictable age-related failures and any return to injection operations would pose an unacceptable health and safety risk. The offshore platforms and onshore Easington terminal were also showing substantial age-related deterioration.
The only technically viable option for reducing the risk associated with injection operations using the existing Rough wells and infrastructure to an acceptable level was considered to be abandoning the existing Rough wells, drilling new wells, and substantially rebuilding the both offshore and onshore infrastructure. The required investment was deemed to be uneconomic.
In November 2015, CSL had tested well B11 and identified a failure in the production tubing (the primary barrier). In June 2016, the secondary barrier of well C6 failed during the testing procedure. Well C6 was an operational well (ie was not plugged) and its secondary barrier failed at a pressure below 500 psia, well below the then prevailing reservoir pressure of around 2,200 psia.
These results were of particular concern because had the failures occurred in the same well, the B-annulus (tertiary barrier) would have been exposed to the full pressure of the reservoir, higher than the maximum allowable annual surface pressure of the B-annulus of 1,500 psia.
On 15 July 2016 CSL announced that it would cease all injection operations until the well testing program was completed. By 20 June 2017 (the date on which CSL announced that it could not return Rough to storage operations) it had conducted calliper surveys on 22 wells and pressure tests on 21 of the 47/3B wells with the following results:
Eight secondary containment failures:
Four failed a 3,000 psi test (B9z, C1z, C3, C10), three failed a 2,000 psi test (B1, B5, C5) and one failed to hold 500 psi (C6);
Of the five that failed a 3,600 psi test, two were tie-back wells, where the tie-back seals (located at the seabed) were the most likely cause of the failure;
Because two of the three tie-back wells failed, the third well, C2, was also taken out of service even though it passed the 3,600 psi A-annulus test.
One primary containment failure:
The first well tested was B11, with severe production tubing corrosion observed; this was deduced to be on the outside of the production tubing (in the A-annulus). This well passed its 3,600 psi secondary containment test;
B3 appeared to have had seal failure previously, however, the leak could not be replicated during testing;
C9 appeared to have a primary containment failure through a suspected tubing hanger seal failure, however, once plugged, this failure could not be replicated with a hydraulic test (this well’s secondary containment had not been tested). A tubing hanger is a component used in the completion of oil and gas production wells. It is set in the tree or the wellhead and suspends the production tubing and/or casing.
This equated to:
One confirmed primary containment envelope impairment (well B1) out of 21 wells (pressure tested) = 5% (well B3 not confirmed)
Eight confirmed secondary containment envelope impairments out of 21 wells = 38%
These results demonstrated that around one in every two of the 47/3B wells had some form of identified failure (or an unacceptable likelihood of failure in the case of well C2). It also demonstrated that there were several different forms of well integrity issues.
Stresses imposed on wells during injection are more likely to result in loss of primary containment than stresses imposed during normal production because the higher surface pressures required to push the gas into the reservoir result in the production tubing “ballooning” slightly (expanding radially under pressure). This, together with the cooling effect of the colder grid gas being pumped into the wells, causes the production tubing to contract axially (the bottom of the well tries to move upwards). To counteract the stresses imposed by gas injection, all the Rough wells were fitted with telescopic joints at the bottom of the wells which contained elastomeric seals that eventually fail over time. By 2017 they were 30 years old, beyond their 25 year design life.
According to the CMA report, CSL had considered various remedial measures which could potentially allow it to continue to inject gas through the in-service wells. These included: continuous monitoring of all well annuli; installation of a remote actuated emergency venting system (controlled release of any gas reaching the A-annulus to atmosphere through a platform vent stack); and a remote actuated emergency kill system (push gas reaching the A-annulus back into the production tubing by pumping treated seawater into the well).
These options all involved reacting to rather than resolving the well integrity issues, and since further elastomeric seals in the telescopic joints would fail over time, the risks of this were not considered to be as low as reasonably possible.
“In light of the provisional conclusion on change of circumstance, that the age and degradation of the wells and other facilities at Rough (including likely future degradation) mean that the Rough is no longer capable of safe injection operations without a substantial refurbishment which is not economically viable,” – Competition & Markets Authority
Upon closure, Centrica began to produce the cushion gas, a process it expected would take 4-5 years. According to Centrica’s accounts, cushion gas production was:
2016: 9 bcf (0.26 bcm)
2017: 56 bcf (1.59 bcm)
2018: 67 bcf (1.90 bcm)
2019: 40 bcf (1.13 bcm)
2020: 23 bcf (0.65 bcm)
2021: 16 bcf (0.45 bcm)
2022: the interim results indicate a 71% increase in production versus 2021 which would be a further 0.78 bcm
Altogether 6.8 bcm of gas has been produced from Rough since the start of the cushion gas extraction, which is more than the amount that was estimated to be held in the facility at the time of closure (in other words, some of the previously un-recovered reserves have also been produced). The statements in the 2022 interim report (Centrica has a 30 June year end) show that CSL has continued to produce the remaining gas in the reservoir – there are no indications that any new gas injections have taken place.
Re-opening Rough: questions over safety and access to cushion gas
There are two main questions associated with the planned re-opening of Rough:
As there is no indication that any remedial works have been carried out to the wells (there is no mention of any such investments in any of the annual reports, and cushion gas withdrawal has continued), can Rough be operated safely?
Since up to the end of June gas had continued to be extracted from Rough, how much cushion gas must be re-injected before Rough could achieve consistent withdrawal rates?
The TPA application indicated that CSL plans to operate one quarter of Rough’s previous capacity this winter. Of course, it is now early August, with just under two months of the gas summer season remaining, which is little enough time to inject working gas volumes, let alone cushion gas.
Given the apparent lack of investment in upgrading Rough’s wells, and the extensive withdrawal of gas since closure, I can suggest two theories for the prospective operations. Firstly, the expectation in the 1980s had been that compressors would be required in order to fully extract the remaining reserves. It’s possible that such compressors have since been installed given that the amount of gas withdrawn exceeds the levels at which such compression would be needed. This might reduce the cushion gas requirement. Secondly, CSL could adopt the mitigation measures described in the CMA report, implementing safety protocols to be used in case of well failures. It is likely that CSL would choose to operate the facility at lower pressures in order to further reduce the operational risks.
However, I have heard from industry contacts that the compression equipment for injection has been removed, and presumably other injection infrastructure that remains has not been maintained in the past 5 years, so even reaching 25% of capacity might be ambitious. There must surely be question marks over the reliability of the facility if it does manage to open this winter, and surely deliverability rates would be lower than the 44 mcm/d that was previously achieved.
I did wonder whether Centrica would seek to operate the facility itself, and not offer the capacity to the market. This would help with security of supply (possisbly only marginally – see below) and would remove the need to provide a consistent level of service. Variable withdrawal rates would be less of a problem if the capacity was not being offered on a commercial basis. However the TPA exemption states that CSL will offer storage capacity to the market on commercial terms, so presumably the minimum performance levels will need to be met.
What difference could Rough make this winter
Even at a quarter of capacity, re-opening Rough would increase Britain’s gas storage capacity by just over 1.5 times. National Grid’s early gas winter outlook for 2022/23 indicates gas demand ranging from 363 mcm/d in mild weather to 465 mcm/d in extreme cold weather. At a withdrawal rate of 44 mcm/d, Rough could meet between 9.5 – 12% of peak daily gas demand, and when full, the facility could operate at this level for 75 days, just under half of the winter period. At the old withdrawal rates, 25% of Rough’s capacity would take under just 19 days to deplete – in other words, the best we could hope for would be for Rough to meet 12% of demand over a 19-day period (or 9.5% during extreme cold spells when the gas would be most needed).
Re-gaining its licence and the TPA exemption are only the first steps CSL must undertake before Rough can re-open: the North Sea Transition Authority must also determine that it can be safely operated. If this hurdle is cleared, we will learn more about how the facility will be run, what sort of injection and withdrawal rates will be expected, and how much cushion gas will be needed to support reliable operations. But even if everything goes smoothly – and this seems like a big “if” – there is little time to inject working gas into the reservoir ahead of the heating season. At best we can hope that Rough will provide some limited support to the market on the coldest days…let’s hope there are not too many of those.