Projections of Gas and Electricity Used in LNG Public Version

Transcription

1 Projections of Gas and Electricity Used in LNG Public Version AEMO SH43585 Revision th June 2014

2 Project Name Project no: Document title: Revision: 4 SH43585 Date: 24 June 2014 Client name: Client no: Project manager: Authors: Projections of Gas & Electricity Used in LNG AEMO Richard Lewis Richard Lewis, Rohan Zauner File name: Report to AEMO on Gas & Electricity Use in LNG Jacobs SKM Final Public Jacobs SKM ABN Floor 11, 452 Flinders Street Melbourne VIC 3000 PO Box 312, Flinders Lane T F SKM.com Document history and status Revision Date Description By Review Approved 1 28/02/14 Outline of scenarios & methodology R Lewis N/a N/a 2 16/04/14 Complete Draft R Lewis R Zauner N/a 3 21/05/14 Final report incorporating AEMO comments R Lewis R Zauner N/a 4 24/06/14 Final report for publication R Lewis R Zauner N Falcon i

3 Contents Executive summary... iv Terms of reference... iv This report... v Summary of findings... v Scenarios... v Methodology... vi Projections... vii 1. Introduction LNG exports from Gladstone Information cut-off date Scenarios Determining factors AEMO planning and forecasting scenarios LNG projects under construction Further trains for these projects Planned projects Gas resource availability Global LNG demand Competition from other suppliers Scenario selection Methodology Overview Historical data Gas exported Energy used in LNG production (liquefaction) Energy sources Testing gas usage Liquefaction gas usage Energy used in transmission Energy used in gas storage Energy used in gas supply Gas supply Supply model Gas field and processing plant energy usage Operated gas Non-operated gas Total energy usage Estimates of peak gas and electricity demand Peak gas ii

4 3.7.2 Peak electricity Sensitivity of gas and electricity demand to gas and electricity prices Gas demand Electricity demand Confidence in the base scenario projections Calculating monthly estimates Projections Annual projections Monthly projections Appendix A. Energy usage for gas production Appendix B. Abbreviations Disclaimer This report has been prepared solely for the Australian Energy Market Operator for the purpose of assessing gas and electricity use in liquefied natural gas production. Jacobs SKM shall have no liability to any party (other than specifically provided for in contract) for any representations or information contained in or omissions from the report or any written or oral communications transmitted in the course of the project. iii

5 Executive summary Terms of reference The Australian Energy Market Operator (AEMO) has engaged Jacobs SKM to provide the following consultancy services: 1) Development of a project plan clearly setting out the key tasks, deliverables and timing for both a primary report and mid-year update. 2) Liquefied natural gas for export (LNG) industry stakeholder consultation: a) Working with AEMO to develop templates and/or questionnaires for industry stakeholders. b) Where applicable, identifying key stakeholders beyond those identified by AEMO. c) Assessing stakeholder responses to questionnaires and working with AEMO on any required follow-up. d) Work-shopping/presenting forecasts with industry stakeholders prior to publication. 3) Internal AEMO stakeholder consultation through the engagement: a) Working with key AEMO stakeholders to respond to questions on the deliverables and making appropriate revisions to draft deliverables based on AEMO feedback. b) A transfer of knowledge to members of AEMO's Energy Forecasting team and other teams as appropriate regarding LNG sector consumption modelling. 4) Development of a report and mid-year update, suitable for publication on the AEMO website, detailing forecasts of LNG production (in million tonnes per annum (Mtpa)) from eastern 1 Australia as well as the gas and electricity consumption associated with this production. 5) Provision of an Excel database(s) proving data underpinning any chart, figure and/or forecasts presented in the report and the mid-year update, which at AEMO s option shall be published on the AEMO website. 1 Eastern Australia refers to New South Wales (including the Australian Capital Territory), Victoria, South Australia, Tasmania and Queensland. iv

6 This report This report fulfils the requirements of item 4) above. Items 1), 2), 3) and 5) have been or will be complied with separately. It is noted however that the required forecasts of LNG production and the gas and electricity usage associated with this production have not been derived from information provided by stakeholders via the questionnaires, as anticipated in the terms of reference, but are the result of modelling undertaken by Jacobs SKM based largely on information in the public domain, much of it provided by the stakeholders on their websites. Use of confidential material provided by the stakeholders is limited to details of production facility capacities and this material and results derived from it have been redacted from this report. This approach was taken because: Jacobs SKM had been in possession of relevant data for the model and had intended to build the model to assist in items 2) and 3); when the consultation timeframe became extended, AEMO requested Jacobs SKM to adopt this approach. Summary of findings Three large projects planning to export LNG from Curtis Island, near Gladstone, made final investment commitments in 2011 and 2012: Queensland Curtis LNG (QCLNG); Gladstone LNG (GLNG); and Australia Pacific LNG (APLNG). Their six committed LNG trains, each capable of delivering about 3.9 to 4.5 million tonnes (Mt) of LNG per year, are under construction, with the first deliveries scheduled from the final quarter of 2014 to A fourth major project, that of Arrow Energy, remains in the planning phase and may combine with one or more of the existing projects to improve its economics and their gas reserves positions. The purpose of this study is to provide AEMO with consistent estimates of the gas supply required for export, including gas used in the supply chain, and grid-supplied electricity usage in the supply chain. These estimates will be used in the preparation of AEMO s 2014 National Electricity Forecast Report (NEFR) and the inaugural 2014 National Gas Forecast Report (NGFR), ensuring consistency in regard to LNG assumptions in these two reports. In contrast, the preceding 2013 NEFR and 2013 Gas Statement of Opportunities (GSOO) demand forecasts used LNG electricity and gas usage forecasts prepared independently of one another. The key elements of the study are: 1. Scenarios concerning the overall levels of exports. 2. Methodology for estimating electricity and gas used in the LNG supply chain 3. Projections of electricity and gas used in LNG export based on applying the methodology to the scenarios. Scenarios Since the final investment decisions on the six trains were made, the prospects of further trains being committed have diminished significantly. The scenarios selected for this study therefore have relatively limited variation: Base Scenario the six trains operating at their contracted levels of capacity and reaching that level on the most recent ramp-up schedules released by LNG project operators. Low Scenario the six trains operating at 10% below contract, with slower ramp-up High Scenario the six trains operating at boilerplate capacity (above contract) with faster ramp-up, plus a seventh train Jacobs SKM has not formed any view as to the probabilities that might be attached to the low and high scenarios. v

7 Methodology Jacobs SKM has developed a model to estimate gas and power usage to supply LNG exports based on information in the public domain. The model works backwards from the volume of LNG exported through the liquefaction, transmission and production components of the supply chain. It does not take into account gas used in shipping or energy used in drilling, which we understand to be mainly diesel rather than gas or electricity. Key assumptions and parameters for each component are summarised below: Liquefaction it is understood that all three plants use gas for their power and compression requirements. 8% of gas input to the plant is estimated to be used in modern plants. Transmission the large diameter pipelines used by each project have sufficient capacity to ship daily quantities for two trains without mid-point compression. In the High Scenario, the use of one pipeline by the seventh train will necessitate installation of mid-point compression. Gas Supply gas is largely sourced from each projects coal seam gas (CSG) reserves in the Surat and Bowen Basins and although the legacy CSG fields are gas-powered, the new developments are all to be grid-connected. QCLNG and GLNG have also purchased third party gas for up to 25% of their requirements and as the third party sources are either not known or known to be well outside the National Electricity Market (NEM), all are assumed to be gas powered. Field and gas processing plant energy use is calculated using parameters estimated by Jacobs SKM, which express the energy used as a percentage of the gas produced (refer to Appendix A for details). Table E 1 Energy usage for gas supply (% of gas produced) Gas Electric Field 3.3% 1.2% Processing Plant 2.5% 0.9% Auxiliaries 0.6% 0.2% Total 6.4% 2.4% To illustrate this approach, for every 1,000 petajoules (PJ) exported, the gas used in liquefaction is estimated as 8% of gas entering the plant, or 8.7% of exports, equivalent to 87 PJ. The 1,087 PJ transported uses no gas for transmission compression in the base case, hence the total production output required is 1,087 PJ. If all the field and processing compression is gas-powered, gas used for this compression is estimated as 6.4% of output, equivalent to 70 PJ. Total gas withdrawn at the wellhead would therefore be 1,157 PJ. vi

8 Projections Total LNG export projections in each scenario are presented in Figure E 1 together with equivalent projections from the 2013 GSOO (dashed lines). Please note that all annual projections refer to financial years Jacobs SKM converted the calendar year GSOO estimates to financial year estimates by averaging. It appears that while the plateau values of the two sets of projections are very similar, the 2013 GSOO projections grow more slowly through 2015 to 2017, in line with Jacobs SKM s low scenario. Figure E 1 Total LNG export projections Mtpa High Base Low High Sensitivity 2013 Planning 2013 Low Sensitivity 2013 Note: 2013 GSOO projections extend to 2033 Figure E 2, Figure E 3 and Figure E 4 show the total gas used in liquefaction and production, total gas exported plus gas used and total grid power usage respectively. Figure E 2 and Figure E 3 also illustrate the equivalent projections from the 2013 GSOO (dashed lines). Energy usage figures include estimates of energy usage in third party gas production. Note that the 2013 NEFR did not publish disaggregated LNG usage projections GSOO gas usage projections assumed that production compression would be 100% gas-powered, whereas the current projections assume a majority of production compression is powered by electricity from the grid. This accounts for the material differences between 2013 GSOO projections and Jacobs SKM s projections of gas usage. The long-term decline in gas usage and the corresponding increase in grid power usage in Jacobs SKM s projections is due to the assumption that third party production is gas-driven and is replaced by electrically-driven equity gas at the end of third party contracts. vii

9 Figure E 2 Total gas used in liquefaction and production PJ High Sensitivity 2013 Planning 2013 Low Sensitivity 2013 High Base Low Figure E 3 Total gas exported and used in liquefaction and production PJ High Sensitivity 2013 Planning 2013 Low Sensitivity 2013 High Base Low viii

10 Figure E 4 Total grid power usage GWh 6000 High Base Low ix

11 PJ Projections of Gas & Electricity Used in LNG 1. Introduction 1.1 LNG exports from Gladstone From the discovery of natural gas in eastern Australia in the 1960s until the mid-2000s, gas reserves were sufficient only for the domestic market. However by 2007, following the steady development of economic extraction technologies, reserves of coal seam gas (CSG) reached a critical point at which further growth would have outstripped local demand, and other market options were sought. Worldwide, the preferred technology for monetising excess gas 2 is currently LNG production, which is a global commodity that saw rapid growth and high prices during the oil price surge from 2003 to Since proposals have been put forward to export LNG from liquefaction plants in eastern Australia with eight proposed for the Queensland coast and one each in New South Wales and South Australia. From 2008 to 2011 CSG reserves surged to meet anticipated export demand (Figure 1-1). Figure 1-1 Eastern Australian CSG reserves development 70,000 60,000 50,000 40,000 3P 2P 1P 30,000 20,000 10,000 0 Sources: Company data; Queensland Department of Natural Resources and Mines; Jacobs SKM analysis. Reserve categories reported refer to: proved (1P); proved and probable (2P); and proved, probable and possible (3P). 2 Excess gas is defined as gas that cannot reach a market by pipeline, either because the pipeline market is fully supplied or because a suitable pipeline cannot be constructed. LNG is currently preferred to conversion technologies such as Gas-To-Liquids, on the grounds that the conversion cost is lower and for as long as LNG prices remain tied to oil prices it will be more profitable. 1

12 Three large projects which are planning to export from Curtis Island, near Gladstone made final investment commitments in 2011 and 2012: QCLNG; GLNG; and APLNG. Their six committed LNG trains, each capable of delivering about 3.9 to 4.5 Mt of LNG per year, are under construction, with first deliveries scheduled from the final quarter of 2014 to The six trains will use approximately 3,750 PJ each over 20 years, or 28,500 PJ in total. A fourth major project, that of Arrow Energy, remains in the planning phase and industry analysts have suggested it may combine with one or more of the existing projects to improve its economics and their reserves positions. All three projects have planning approval for more than two trains. The other Queensland export projects and the sole New South Wales project have made limited progress since their first announcements, mainly because of difficulty securing gas supply. The South Australian project, promoted by Beach Petroleum, would be based on Cooper Basin shale gas, which is still in the early stages of development. The purpose of this study is to provide AEMO with consistent estimates of the gas supply required for export, including gas used in the supply chain, and grid-supplied electricity usage in the LNG supply chain. These estimates will be used in the preparation of AEMO s 2014 NEFR and NGFR, which have previously used electricity and gas forecasts for LNG prepared independently. The key elements of the study presented in the following sections are: 1. Scenarios concerning the overall levels of exports, focussing primarily on: the numbers and timing of trains constructed; full production levels of exports for each train; and timing of ramp-up to full production. 2. Methodology for estimating electricity and gas used in the LNG supply chain 3. Projections of electricity and gas used in LNG export based on applying the methodology to the scenarios. 1.2 Information cut-off date The modelling documented in this report incorporates information available as at 21 st March Since that date the following potentially material information has become available: GLNG has entered a 20-year contract to purchase gas from Westside s Bowen Basin CSG reserves (announced on 27 th March 2014). This agreement may reduce GLNG s electricity demand if the Westside gas development does not use grid-connected electric drives for compression. Each of the projects has made further public statements confirming their confidence in LNG delivery targets used in this report. 2

13 2. Scenarios 2.1 Determining factors Jacobs SKM has constructed three scenarios for LNG exports from eastern Australia based on near-term, midterm and long-term considerations including: AEMO planning and forecasting scenarios; LNG projects under construction and planned; gas resource availability; global LNG demand and competition from other suppliers AEMO planning and forecasting scenarios AEMO and the industry-based Scenarios Working Group (SWG) have developed three scenarios representing high, medium, and low energy consumption for both gas and electricity from centralised sources 3, for use in AEMO s planning studies in The three scenarios are broadly defined in Table 2-1. Scenario assumptions related to domestic gas and LNG exports are set out in Table 2-2. Table 2-1 AEMO 2014 Scenarios High energy consumption from centralised sources Medium energy consumption from centralised sources Low energy consumption from centralised sources Energy consumption High Medium Low Type of consumer Low engagement Highly engaged Highly engaged Economic activity High Medium Low Table 2-2 Gas and LNG assumptions Scenario Gas and LNG Assumptions Medium Energy Domestic gas production and global LNG markets continue as per market expectations (central estimate) and are consistent with current growth in Australia s production levels. High Energy Domestic gas production is higher than expected (exceeding production required for LNG). Australia s international competitiveness for gas exports is higher than the medium scenario. o Hurdles for fuel substitution are expected to discourage switching between gas and electricity usage. This means, for example, if gas prices were low, industry would not be able to easily switch from electricity to gas due to either prohibitive infrastructure costs or specific production processes which will not allow it. Global demand for LNG is stronger than in the medium scenario, encouraging higher exports. 3 A centralised source refers to the national electricity transmission grid for electricity and the national gas transmission pipeline for gas Planning and Forecasting Scenarios, AEMO 11 February

14 Scenario Gas and LNG Assumptions Low Energy Domestic gas production is more difficult than expected. Australia s international competitiveness is lower than the medium scenario. Fuel substitution hurdles discourage fuel switching between electricity and gas. Global LNG demand is weaker than the medium scenario and there is low penetration of gas as a transport fuel LNG projects under construction The three LNG export projects under construction in Queensland have the following features: 1) Each project has sold gas under long-term contracts with two or more buyers. QCLNG will also sell to BG Group portfolio customers in Chile, China and Singapore. The terms of all contracts are 20 years or longer, though it is not known whether the contract periods start at first supply or when full offtake (plateau production) is reached. 2) Each project is constructing 2 LNG processing trains, each of 3.9 to 4.5 Mtpa capacity, on Curtis Island off the coast near Gladstone. 3) Each project is developing gas supply capacity based on equity CSG reserves in the Bowen and Surat Basins. GLNG and QCLNG are also sourcing gas supplies from third-party owned CSG and conventional gas resources in the Cooper Basin. Table 2-3 Gladstone LNG Project Parameters Project Partners Nameplate Capacity (Mtpa) Contracts Party Volume (Mtpa) QCLNG 5 BG Group (73.75%) CNOOC (25%) Tokyo Gas (1.25%) 8.5 CNOOC Tokyo Gas Chubu Electric Total GLNG 6 Santos (30%) Petronas (27.5%) Total (27.5%) Kogas (15%) 7.8 Petronas Kogas Total bg databook 2013 Australia 6 Santos 2013 Investor Seminar, 4 December

15 Project Partners Nameplate Capacity (Mtpa) Contracts Party Volume (Mtpa) APLNG 7 Origin Energy (37.5%) Conoco Phillips (37.5%) Sinopec (25%) 9.0 Sinopec Kansai Electric Total ) Each project has constructed its own transmission pipeline to deliver gas to Curtis Island. The pipelines will be interconnected at their upstream and downstream ends to facilitate operational gas management and trading. 5) In the above table project ownership is stated on an aggregate basis across production and liquefaction. In some projects percentages are different in each component Further trains for these projects APLNG 8 has environmental approval for 4 trains and QCLNG 9 and GLNG 10 have approval for 3 trains, however none of the projects currently has sufficient reserves to support a third train (refer to section 3.6 for details on each projects gas reserves). BG Group has stated that it does not anticipate making a decision on a third train at QCLNG in the near future Planned projects A fourth project, that of Arrow Energy, has cleared environmental approvals but a final investment decision (FID) has been deferred until after 2014 because it does not meet shareholder investment hurdles 12. The project may be costlier than the other three because it requires construction of two pipelines to link to reserves in the Surat and Northern Bowen Basins to supply two LNG trains. A single train independent project will lack economies of scale, however Arrow Energy s Surat Basin gas reserves could be used to supply a third train at one of the projects under construction, using their pipeline(s). This could be several billion dollars cheaper than an independent project. On 21 st May 2014 the Australian Financial Review newspaper reported that Arrow Energy had been in discussion in March with each of the established projects regarding these options and that a direct sale of gas appeared most likely, though the rumoured gas price of $8.50/GJ would give the LNG project a low rate of return. The report indicated that Arrow Energy was likely to have more intense discussions with just one party before a definitive outcome is achieved. 7 Origin Energy Asia Roadshow, November Queensland Co-ordinator General website 9 ibid 10 GLNG Project Environmental Impact statement - Executive Summary. 11 BG Group 2013 fourth quarter and full-year results transcript (4th February 2014) 12 Sydney Moring Herald, 1 February

16 A number of other LNG export projects have been put forward but none has access to demonstrated gas resources and none is considered likely to proceed in the period to Projects that may proceed after 2020 are unlikely to be at the planning stage yet and such projects are not considered in the projections by Jacobs SKM owing to their very high levels of uncertainty at this point in time Gas resource availability During the period from 2008 to 2011 when CSG reserves were growing rapidly (Figure 1-1), based on published contingent and prospective resources it was projected that this growth rate could continue and may support considerably more than six or eight trains. However since 2011 that reserves growth has reversed, for a number of reasons, and the prospect of further trains is not as bright: 1) During project construction appraisal drilling to prove reserves has given way to development drilling of production wells. 2) Criteria for classifying contingent CSG resources have been tightened by the Society of Petroleum Engineers (SPE). 3) Opposition to CSG development has intensified, particularly in New South Wales, with the result that in the Hunter and Sydney Basins reserves have been declassified and development in other New South Wales basins has slowed down or stopped. 4) Productivity of some fields appears to be below expectations 5) Exploration for CSG in the Galilee Basin has faltered after a number of poor drilling results 6) As a result of the above and the shale gas success story in the United States, the focus of exploration in eastern Australia has moved from CSG to shale gas, with interest mainly in the Cooper Basin because of its well-developed gas gathering and processing infrastructure. However the shale industry is five to ten years away from demonstrating the economics and reserves necessary to support large-scale production. In particular, the Australian industry does not yet support the competitive drilling and fraccing sub-sector needed to reduce costs and at the $A8.30/GJ cost assumed in the 2013 GSOO LNG exports via Gladstone would not be economic Global LNG demand Global LNG demand is widely expected to continue to grow over the next decade, as further countries join the 27 current importers. Figure 2-1 illustrates projections assembled by BG Group and Santos from independent industry forecasters. Supply from existing plants and plants under construction are expected to peak by 2020 and then decline, creating an opening for 130 to 160 Mtpa of new capacity by This outlook suggests there is no shortage of opportunities for further capacity to be constructed in Queensland or elsewhere in eastern Australia. 6

17 Figure 2-1 Medium-term LNG demand-supply gap Mtpa BG Demand BG Supply STO Demand STO Supply Sources: Are the dynamics of the LNG industry changing? BG Group (BG), 5 March 2014; Santos (STO) 2013 Investor Seminar, 4 December Competition from other suppliers During the past five years Australia has represented a significant proportion of new LNG capacity that has commenced construction. As well as the 25 Mtpa capacity in Gladstone, the following projects are under construction in Western Australia and the Northern Territory: Gorgon (15 Mtpa); Wheatstone (8.9 Mtpa); Ichthys (8.4 Mtpa); and Prelude (3 Mtpa). Australian exports are scheduled to reach approximately 80 Mtpa by 2018, which will position Australia as the leading gas exporter slightly ahead of the current leading exporter Qatar on 77 Mtpa. The reasons for Australia s market dominance during this period include the availability of gas resources, an absence of export constraints, the absence of a national oil and gas company with mandatory project participation, manageable regulatory systems and limited sovereign risk. These factors continue to be present, particularly in Western Australia where the Browse and Scarborough projects have to date narrowly failed to reach committed status. However all the Gladstone projects and Gorgon have suffered significant cost overruns, in part because of their nature (Gorgon is located on Barrow Island, a Class A nature reserve) and in part due to competition for construction resources among these projects and with other Australian resource projects. Future greenfield Australian projects now appear uncompetitive with new entrants that have emerged in North America and east Africa (among others). Using cost data prepared by McKinsey & Company 13, Jacobs SKM has estimated that greenfield Australian LNG projects would cost 15% to 20% more than comparable projects in the United States, Canada and Mozambique. However, brownfield Australian projects making use of existing liquefaction sites and sharing pipeline infrastructure could be competitive and from 2018 there will be eight operating brownfield options to choose from. 13 Extending the LNG Boom, Improving Australian LNG productivity and competitiveness, McKinsey& Company May

18 It is also likely that the preferred status of competing projects will erode over time. Australian project costs are now almost fixed whereas the cost of projects elsewhere continue to be subject to uncertainty, with a strong bias towards overrun rather than underrun, noting that some Canadian projects are already experiencing cost overruns due to their remote locations. Apart from the impact of opposition to CSG developments, the non-cost risks in Australia are also well understood, whereas projects in countries with limited histories of petroleum production, such as in east Africa, will most likely encounter delays due to the absence of, or uncertainty in, petroleum legislation and regulation. These factors suggest that Australian projects, including those in eastern Australia, may return to contention over the next five years to meet post 2020 opportunities. 2.2 Scenario selection The above considerations lead to the following scenarios: 1) It is desirable to have a scenario which reflects current commitments, i.e. the six trains under construction, all ultimately producing at their contracted output levels. Jacobs SKM considers this scenario represents the most likely outcomes of the proposed stakeholder consultation process. Since production could be lower than expected levels, the current commitments must form the basis of the Base 14 Scenario. QCLNG s contracted output is assumed to be 8 Mtpa, allowing a similar margin between contracts and capacity as the other two projects. 2) The Low Scenario must then be the six trains with lower production due to either buyers not opting to take their full contract cargoes or due to gas supply shortfalls (refer to next section). A 10% reduction on contracted levels has been assumed, as this is viewed as likely to be consistent with LNG contract take-orpay. Slower start-up has also been assumed. Owing to the low marginal cost of LNG production up to the capacity of the LNG plants, the projects have strong incentives not to allow the plants to run below capacity for prolonged periods. Consequently Jacobs SKM considers this scenario to have low probability in the long term. 3) To test the upper boundaries of feasible gas and electricity usage, the High Scenario assumes a seventh train that would be supplied from Arrow Energy s Surat Basin gas with no additional pipelines required. It would be approved no sooner than 2015 and in production no earlier than The study refers to this as the Arrow Energy project but the liquefaction plant could be a third train on any one of the other projects sites. It is also assumed that the other six trains ultimately operate at full nameplate capacity rather than at contracted capacity, following a period during which plant operation is optimised. This is a feature observed in many LNG projects: most recently the Pluto project in Western Australia sold an additional 0.7 Mtpa to Korea Gas. The transition point to full capacity is set at the start of 2019 for all three projects. The High Scenario could involve further trains supplied by shale gas but the economics have yet to be demonstrated. The timeframe is highly uncertain and this possibility has not been considered. Other scenarios are also feasible, including a mixed Low/High scenario in which LNG ramp-up is slower than in the Base Scenario but production ultimately reaches full capacity. Jacobs SKM considers the above scenarios to be consistent with the AEMO scenario definitions set out in section Jacobs SKM has used the term Base Scenario to denote the most likely outcomes. It is intended that this scenario should correspond to AEMO s Planning or Medium scenarios. 8

19 3. Methodology 3.1 Overview The gas supply chain from wellhead to export and the relevant components of each component of the chain are illustrated in Figure 3-1. Note that this supply chain representation excludes the shipping component, as it is understood the contracted export quantities are free on board (FoB) volumes, and excludes energy used in drilling wells, which we understand is mainly diesel rather than gas or electricity. Figure 3-1 The LNG supply chain Jacobs SKM has developed a model to estimate gas and power usage to supply LNG exports based primarily on information in the public domain. In this context modelling of the supply chain is most readily undertaken backwards from the volume of gas exported, for which contracted quantities are available in the public domain, to the numerous production facilities whose individual outputs can only be estimated.. The calculations are essentially straightforward, if not rudimentary. The modelling effort lies in obtaining the data and estimating energy use parameters. To illustrate, for every 1,000 PJ exported, the gas used in liquefaction is estimated as 8% of gas entering the plant, or 8.7% of exports, equivalent to 87 PJ. The 1,087 PJ transported uses no gas for transmission compression in the base scenario, hence the total production output required is 1,087 PJ. If all the field and processing compression is gas-powered, gas used for this compression is estimated as 6.4% of output, equivalent to 70PJ. Total gas withdrawn at the wellhead would therefore be 1,157 PJ. The model assumes that export demand will be met and is not constrained by supply capacity, so Jacobs SKM has not considered the projects progress with drilling wells, likely well production capacity or constraints on production ramp-up. It is noted however that AGL has recently released a report 15 which suggests that LNG demand may not be met after 2016 but may suffer an average supply deficit of 2% under some supply conditions. 15 AGL Working Paper No 40 Solving for x. March

20 The underlying model is based on quarterly intervals, to achieve greater precision than achievable using annual intervals but retaining greater computational manageability than possible using a monthly model. Annual and six-monthly results are simple summations of quarterly results, while monthly results have been derived from quarterly results using the algorithm described in section Table 3-1 The LNG supply chain Process Energy use Gas volume Grid supplied power Export Nil, export volumes assumed to be on an FoB basis Gas exported Nil Liquefaction Direct drive compressors, electrical power Gas delivered by pipeline Possible but not selected for current Gladstone projects 16 Gas transmission Mid-point compression Gas delivered by processing plant Possible 17 Gas production Compression, auxiliaries Gas extracted from reservoirs A selected option for all projects Historical data Although the study terms of reference contemplate consideration of historical data, LNG has yet to be produced by any of the projects. Upstream power consumption associated with gas for testing has commenced however corresponding gas data is not available and no relationships can therefore be determined. 3.2 Gas exported The assumed plateau export levels in each scenario are summarised in Table 3-2. The Base Scenario export levels are set at the levels of the foundation LNG contracts. Corresponding to the 20-year contract terms that form the basis of the Base Scenario, the exports in each scenario can be assumed to extend to at least LNG plants have serviceable lives well in excess of 20 years and with low marginal costs are likely to remain competitive in the global LNG market provided competitively-priced feed-gas is available. Each scenario is therefore assumed to extend to Future Curtis Island LNG facilities could be connected to the NEM via a cable connection 17 The QCLNG and APLNG pipelines are sufficiently proximate to the Queensland EHV electricity grid for grid electrically driven compression to be credible. The mid-point of the GLNG pipeline is not close to the EHV electricity grid. 10

21 Table 3-2 Plateau export levels (Mtpa) Scenario QCLNG GLNG APLNG Arrow Energy High Base Not applicable Low Not applicable Start-up and ramp timing assumptions are presented in Table 3-3: The Base Scenario is Jacobs SKM s interpretation of the most recent timing statements by projects: o QCLNG 18 : Train 1 (T1) starts in Quarter 4 (Q4) 2014; T2 six months after T1; ramp-up over 2015 o GLNG 19 : T1 in 2015; T2 six to nine months after T1; ramp-up T1, six months; T2, two to three years o APLNG 20 : T1 mid-2015; T2 late 2015; full revenue mid-2017 FY High Scenario First LNG and plateau are accelerated relative to the Base Scenario for T2 only, as acceleration of T1 appears unlikely. Greater acceleration is applied to GLNG because the Base plateau is further out. For the High Scenario Plateau refers to the contract level with the increase to full plant capacity occurring in Low Scenario First LNG and plateau are delayed relative to Base Scenario for all elements. This would be consistent with minor technical problems being experienced during start-up. Less delay is applied to GLNG because the Base plateau is further out. Table 3-3 Start-up and ramp-up timing QCLNG GLNG APLNG Arrow T1 T2 Plat T1 T2 Plat T1 T2 Plat T1 T2 Plat High Q4 14 Q2 15 Q4 15 Q2 15 Q4 15 Q4 17 Q3 15 Q4 15 Q3 16 Q1 20 n/a Q4 20 Base Q4 14 Q3 15 Q1 16 Q2 15 Q1 16 Q2 18 Q3 15 Q1 16 Q4 16 Low Q1 15 Q4 15 Q1 17 Q3 15 Q2 16 Q4 18 Q4 15 Q2 16 Q4 17 Notes: T1 = Train 1 first gas exported; T2 = Train 2 first gas exported; Plat = Plateau production reached for both trains; Q1 = first quarter of the calendar year. 18 BG Group 2013 fourth quarter and full year results transcript (4 th February 2014) 19 Santos 2013 Investor Seminar (4th December 2013) 20 ORG US Roadshow 2014 (3 rd March 2014) 11

22 The first quarter of production is assumed to be at 50% of contract level and thereafter production ramps up smoothly to the plateau level. Jacobs SKM understands that many liquefaction trains cannot operate below 50% capacity levels, though lower output levels can be achieved by on-off operation. This figure does not relate specifically to any of the Curtis Island trains. Jacobs SKM has assumed that gas supply could be ramped up to match these scenarios. In all scenarios each project s plateau production is assumed to be shared equally between the two trains, although an unequal sharing would not affect energy usage outcomes. 3.3 Energy used in LNG production (liquefaction) Energy sources Gas liquefaction plants use energy mostly for compression of the gas and of the refrigerants. The compressors can be powered directly by gas turbines, typically aero-derivative 21 turbines of around 25 megawatts (MW), or by electric drives with power sourced internally from 100 MW power generation turbines or from the grid. The last option is not available to most LNG plants globally as they are located remotely from large grids. Consequently standard LNG plant designs are fully gas-powered. The electric drive option with internal power has higher capital cost than gas turbine (GT) driven plants, though the cost of the electric drives is offset by the economy of the larger GT used in this case. All of the Gladstone plants are being built by Bechtel using the Conoco Philips Optimised Cascade design, first used in 1969 and most recently used in Australia in the Darwin LNG project, in operation since Jacobs SKM understands that all of these plants to date have used GT-driven compression and based on a diagram presented by APLNG (Figure 3-2) the same applies to the Gladstone plants. Consequently the plants are believed to be gas powered and require no grid power. 21 Refrigerant compression can be powered by more efficient power GTs. 12

23 Figure 3-2 APLNG liquefaction plant outline Source: APLNG Project Overview, 20 Sept Testing gas usage Gas is introduced to the liquefaction plants some six to nine months before first LNG production, to enable all elements of the plant to be thoroughly tested and for the storage tanks to be cooled down to -161 Celsius. Jacobs SKM understands that during the testing period gas consumption ranges up to 20 terajoules per day (TJ/d) per train. A figure of 10 TJ/d average during testing has therefore been assumed Liquefaction gas usage The key to gas usage in liquefaction is the overall efficiency of the plant. Jacobs SKM has not been able to source efficiency figures specific to the Gladstone plants, but has applied figures estimated by Conoco Philips for its Optimised Cascade process 22. This study reports efficiencies under a range of operating conditions from 89% - 93% for plants with industrial gas turbines, i.e. 7% - 11% of gas used in the process, to 92% - 94% for aero-derivative gas turbines. It is not clear which operating conditions might be closest to those at Gladstone so Jacobs SKM has selected the lowest efficiency level associated with aero-derivative plants, namely 92% or 8% of gas used in processing. This is most likely to be appropriate in Gladstone because of the relatively low heating value of CSG compared to conventional gas normally used for LNG. 22 A fresh look at LNG process efficiency, Weldon Ransbarger, Conoco Philips, reprinted from LNG industry Spring

24 The 8% figure has been applied under all conditions modelled. The only variable condition modelled is the percentage utilisation of the plant. 3.4 Energy used in transmission Each of the three LNG projects has constructed a transmission pipeline to convey gas from their CSG processing plants in the Surat and Bowen Basins to Gladstone. The pipeline routes are depicted in Figure 3-3 and major parameters are presented in Table 3-4. Each project also has a network of smaller diameter pipelines connecting the gas processing plants (GPPs) with the export pipelines and operating at the same pressures. Table 3-4 LNG export pipeline parameters QCLNG GLNG 23 APLNG 24 Length Main Export PL (kilometre) Internal Diameter (millimetre) 25 1,040 1,040 1,050 Maximum Operating Pressure (kilopascals) 10,200 10,200 13,500 Jacobs SKM has assessed the energy usage in the pipelines under the following assumptions: Each project transports its own gas requirements for liquefaction Gas is compressed up to operating pressure at the GPPs, or elsewhere for third-party gas. The energy used at GPPs is part of the processing energy use Further compression may be required at the approximate mid-points of the export pipelines for the gas to reach Gladstone. The requirement for further compression under each scenario has been tested using a Jacobs SKM gas flow model. The model computes the pressure loss along the pipeline at any given flow rate, using steady state flow pressure loss calculations. If the pressure at the pipeline delivery point falls below the target pressure for delivery into the LNG plants, assumed to be 3,000 kpa, then further compression is required at an intermediate point along the pipeline. The amount of compression in MW required to lift delivery point pressure to the required level is similarly calculated and compression fuel usage is computed from the power requirement. 23 GLNG Gas Transmission Pipeline Description 24 Constructing the pipeline, available at 25 All pipes are stated to be 42 inches in diameter. Stated diameters in mm vary. 14

25 Figure 3-3 LNG pipelines map Source: Jacobs SKM produced under licence from Google Earth The flow modelling suggests that no mid-line compression is required in the Low and Base Scenarios where the pipelines are carrying gas for two trains and consequently there is no transmission energy requirement in the Base and Low Scenarios. This is consistent with all reports sighted, for example, Desktop review of QCLNG Looping Scenario prepared by OSD Pipelines in December In the High Scenario however, whichever pipeline is used to supply the seventh train will be carrying gas for three trains, assuming that the Arrow Energy processing plants providing gas for this train are connected to one pipeline only. In this case the flow modelling suggests that mid-line compression is required for this pipeline. 15

26 As with the liquefaction plant, this compression can be driven by a gas turbine or electrically, via local generation or grid connection. The latter is available at low cost near the midpoints of the APLNG and QCLNG pipelines but not at the midpoint of the GLNG pipeline, which lies further west. To standardise the calculation of compression energy independent of the pipeline used by the Arrow Energy project Jacobs SKM has therefore assumed pipeline compression is gas-driven. Estimated gas compression usage due to the Arrow Energy project is 37TJ/d for 700TJ/d incremental load i.e.5.3% of load. This factor is applied at all Arrow Energy project loads. In the modelling and reporting all intermediate compression usage is allocated to the Arrow Energy project. The project also uses energy for liquefaction and upstream in the same proportions as the other projects. 3.5 Energy used in gas storage Two of the LNG projects, QCLNG and GLNG, plan to make use of underground gas storage to augment gas supply during the initial ramp-up phases and possibly to assist in gas management during LNG plant shutdowns or periods of declining gas production. Underground storage uses energy to compress gas into the storage field and again to extract it. The storage projects at the Silver Springs field (developed by AGL and contracted to QCLNG) and the Roma field (developed by Santos for GLNG), have been completed and filling has been initiated however the volumes of gas stored are not published. For this reason and because storage contributions are likely to be small under steady state operation, Jacobs SKM has excluded storage from the modelling. This will result in a minor understatement of total energy use but which is likely to be less than the uncertainty in other factors such as liquefaction use. Taking storage into account would also accelerate the relevant energy usage. 3.6 Energy used in gas supply It is understood that each LNG project will use electrically driven compression at its GPPs with power taken off the Queensland electricity grid. Some projects will also use gas from third-party plants that cannot be gridconnected. To determine energy used in gas supply Jacobs SKM has therefore disaggregated supply to the levels necessary to capture gas versus electric compression and, for the latter, disaggregation by Powerlink grid connection point Gas supply Gas supply to each project will be provided mainly from their equity reserves in Queensland CSG. However, owing to the joint ownership of some fields, some equity gas will be developed and operated by other projects. Projects have also entered contracts with one another to purchase operated gas they don t own and have entered contracts with third parties for gas from other sources, to assist project start-up and to compensate for current reserve shortfalls. Table 3-5 shows each project s operated and equity gas reserves and Table 3-6 documents third party contracts. Note the discrepancy between total equity and operated reserve figures is largely accounted for by AGL s equity in Arrow Energy operated acreage in the Northern Bowen Basin. Regarding the contracts, as the buyers in the last two contracts in the table are not known, Jacobs SKM has not used them in this analysis. 16

27 Table 3-5 LNG project equity and operated gas reserves as at 30 June 2013 (PJ) Equity Operated QCLNG 8,708 9,785 GLNG 5,373 6,305 APLNG 12,951 10,960 Arrow 7,891 8,961 34,923 36,011 Source: Operated, Queensland Department of Natural Resources and Mines; Equity, Jacobs SKM analysis based on company reports Table 3-6 LNG project contracts with third party suppliers Seller Operator Buyer Source Delivery point Start date End date APLNG QCLNG QCLNG Surat CSG Field 1/10/2014 2/10/2034 Annual volume (PJ) 95 falling to 25 in 2016 Santos Santos GLNG Cooper primarily Wallumbilla? 1/01/ /12/ AGL QCLNG QCLNG Surat CSG Field 1/01/ /12/ Origin Energy Unknown GLNG Origin Energy Portfolio Wallumbilla 1/01/ /12/ Origin Energy Unknown QCLNG Origin Energy Portfolio Wallumbilla 1/01/ /12/ Origin Energy Unknown GLNG Origin Energy Portfolio Wallumbilla 1/01/ /12/ Stanwell Unknown Unknown Wallumbilla? Wallumbilla? 1/10/ /09/ AGL QCLNG? Unknown Surat CSG Wallumbilla? 1/07/2014 Unknown 15 Sources: Company media statements. A question mark indicates that the relevant information has not been published. The value in the table is Jacobs SKM s best estimate. 17

28 Complementary to the first contract, it is assumed that APLNG will take 70 PJ of its equity share in the QCLNGoperated fields, when it reaches plateau LNG production levels in There is also an arrangement between GLNG and APLNG for GLNG to take approximately 9 PJ per annum of equity gas at Combabula Supply model For each LNG project, the contracts are separated into operated (contracts 1 and 3 above) and non-operated (all other contracts). The non-operated contracts are assumed to be used to their maximum subject to the LNG plant s gas requirements, as it is reasonable to assume the contracts all have high take-or-pay provisions. The operated gas requirement is then the LNG plant requirement, less the relevant non-operated contract volume, plus supply obligations to other projects. It is also assumed that contracts are not recontracted on termination but are replaced by additional equity gas. Figure 3-4 illustrates the application of this approach to GLNG in the Base Scenario. The operated gas requirement must be supplied from the projects CSG fields, which Jacobs SKM has grouped according to projects connections to the Powerlink grid. None of the projects has indicated publicly how supply will be drawn from the fields and Jacobs SKM has therefore based supply from each group on estimated capacities and timing. Note that these allocations are approximate in the short-term as the estimated capacities allow for more gas to be sourced from each group and will inevitably be incorrect in the long term as production in the first developed wells declines and is replaced by production in new wells and new fields. Figure 3-4 GLNG Base Scenario gas allocation to operated and non-operated average daily supply 1,400 1,200 1,000 TJ/d Operated Non-operated