Wairakei Ring Investment Proposal

Transcription

1 Wairakei Ring Investment Proposal Project Reference: CTNI_TRAN-DEV-01 Attachment B Analysis assumptions and methodology December 2008

2 Document Revision Control Document Number/Version Description Date 001/Rev Wairakei Ring Investment Proposal Attachment B

3 Contents 1 Introduction Assessment of proposals under the Rules Long-list and Short-list options Document Structure Outline of this document Approach to the application of the GIT The market development scenarios The modelling approach The base case Net Present Value versus Real Options analysis Staged development plans Discount rates Analysis in current or future dollars Analysis period Market costs Transmission costs Statutory compliance costs and benefits Inflation Community care costs Interest During Construction Exchange rate approach Market benefits Capital benefits Fuel and transmission loss benefits Greenhouse gas emission, spill and load shedding benefits Ancillary services benefits Competition benefits Sensitivities Forecast demand Variations in the size, timing, location, and operating and maintenance costs Capital cost Timing of decommissioned assets Value of expected unserved energy Discount rate Variation in hydrological inflow sequences Generator and demand side bidding strategies

4 5.9 Competition benefits Carbon Charges Probability of occurrence of market development scenarios Exchange rate sensitivity using 10 year average exchange rates Additional Sensitivity - Property Cost Appendix A Summary of Assumptions...21 Appendix B Appendix C Appendix D Glossary...23 GEM Modifications and tests...24 Generation fixed in initial years

5 1 Introduction This document describes the assumptions and methodologies used in the GIT analysis for the Wairakei Ring Investment Proposal (the Proposal), the results of which are presented in Attachment A. This document incorporates the feedback Transpower received from the consultation it has undertaken in accordance with the agreed process with the Electricity Commission (the Commission). Under rule 14.2 of Section III, Part F of the Rules, Transpower must, in relation to any proposed economic investment, consult on whether it has applied the GIT reasonably. The assumptions and methodologies described in this document form a critical part of the consultation. 1.1 Assessment of proposals under the Rules To determine the expected net market benefit of an option, Transpower has adopted certain input assumptions, parameters and market scenarios in its application of the GIT. A sensitivity analysis has also been undertaken to test the robustness of applying the GIT to the options. There are some areas where the Rules constrain Transpower s input or approach. Transpower has identified where this is the case in the relevant sections below. 1.2 Long-list and Short-list options To develop a list of appropriate options for the potential investment proposal, Transpower consulted on a long list of options in phase 1 of the GIT consultation in May The final long list is set out in Attachment E. Following this consultation Transpower, upon considering the feedback, has developed a short list of options to which the GIT is applied. This short list of options, and the short listing process applied by Transpower is set out in Attachment C. This was consulted on in October Document Structure This document forms part of the Proposal. The documentation is structured according to the following diagram: 5

6 Investment Proposal Attachment A GIT Results Attachment B Assumptions and Approach Attachment C Power System Analysis Attachment D Costing Report Attachment E Final Long-list and Criteria Attachment F Models and Data (in separate files) Attachment G Submissions 1.4 Outline of this document This document follows the structure, and provides detail as to: Section 2 - approach for applying the GIT; Section 3 - market costs used in the GIT; Section 4 - market benefits calculated; Section 5 - sensitivity analysis; Section 6 - approval costs; Section 7 - summary; and Appendices A - D. A glossary of terms and acronyms used in this document is included in Appendix B. All references to the Rules in this document refer to Section III of Part F of the Electricity Governance Rules 2003 unless otherwise specified. 6

7 2 Approach to the application of the GIT This section describes the economic approach Transpower has used for assessing the short-listed options in the GIT, including: the market development scenarios; the modelling approach; use of NPV analysis; assessing a staged development plan and inclusion of modelled projects; approach to calculating market costs and benefits; and the analysis period. The GIT involves assessing the expected value of each short-listed option 1 against a Base Case. The Proposal is the option: with a net market benefit greater than zero (when assessed against the Base Case); and that maximises the expected net market benefit compared with the other options. The remaining options form the alternative projects to be considered in relation to the GIT 2. Should the Commission grant approval for the Proposal, Transpower will be able to recover the cost through the transmission pricing methodology set out in the Electricity Governance Rules. There is uncertainty inherent in the assumptions to be used in this analysis. In applying the GIT, Transpower has attempted to account for these uncertainties in a way that is appropriate and reasonable for the scale of analysis commensurate with the estimated capital expenditure required for the proposed investment. 2.1 The market development scenarios The GIT requires Transpower to develop market development scenarios and analyse the base case and each short list option against each scenario. 3 The market development scenarios utilise a set of generation drivers as inputs. These drivers include details of generation cost, timing, location, carbon charges, demand etc. To date, the Commission has established the scenarios by using the drivers within its generation expansion model (GEM). These scenarios are set out in the Commission s Statement of Opportunities (SoO) published in accordance with Rule 9. There are five scenarios within the SoO. Transpower has used the five scenarios as a starting point for the analysis, subject to the changes described below. For this Proposal, Transpower has reviewed and updated the generation details in the current operative SoO. In doing so, Transpower has also sought feedback from stakeholders. As a result of the feedback received, Transpower has incorporated a number of changes to the drivers to reflect what it considers to be reasonable assumptions for the Proposal. The generation drivers are set out as part of the detailed input data in Attachment F. The changes Transpower has made to the SoO scenarios consist of: 1 Refer to Attachment C for a detailed list of short listed options. 2 Refer to the definition of alternative projects at clause 19, Schedule F4 of the Rules. 3 Refer to definition of market development scenarios at clause 28 of Schedule F4, Part F of the Rules. 7

8 1) an update of specific generation project details with the latest available information; 2) an extension to the data to indicate the implementation status of all the projects in the list; and 3) a change of approach for the early years of each scenario (2009 to 2016) by incorporating actual and pending generation investments, independent of perceived generation cost. This change is considered appropriate by Transpower so as to provide a link between observed behaviour of generation investors and the longer term scenarios presented in the SoO. The generation information has been sourced from either recent public announcements or generation companies themselves where they are happy for the details to be released. The method for deriving these changes is set out below Updates to the generation data Through 2008, Transpower held a number of workshops with generation companies with the aim of: gaining a better understanding of the generation investment process; and gathering a more complete set of information on pending generation investments. In parallel, public data sources have been monitored to ensure that any updates not captured through the process with generation investors themselves could be incorporated into the analysis. This also assisted in verifying the information provided directly by the generation investors. Additionally, in May 2008, an initial draft list was published for consultation, and some changes were made as a result of the submissions received. A further update was provided by Contact Energy as part of the GIT consultation in October These have been incorporated into the data used for the GIT. As a result of this extensive process, there are now some differences between the information used by the Commission in the SoO and that used by Transpower in the Wairakei Ring analysis. Many of these changes are relatively minor. As part of the information gathering exercise, Transpower has also extended the data set to include the status of the generation projects included in the data set. These have been classified as either: committed; consented; in the consenting process; investigating only; or resource only. The generation project information is now complete for the purposes of this Proposal. The detailed list of generation changes is included in Attachment F Change in the approach for the early years of each scenario Some criticism has been expressed by stakeholders as to the apparent gap between the longer term scenarios in the SoO and the generation projects that are actually progressing. Transpower has been considering for some time how best to objectively reconcile these factors with the scenarios used for the Wairakei Ring. Additionally, 8

9 accounting for locational variations in generation projects is particularly important for the assessment of regional transmission projects such as the Wairakei Ring. There appears to be three basic approaches that could be taken to the analysis: 1) Use the SoO scenarios without modification, except for the updated generation details. Unfortunately this approach does not address the issues noted above; 2) Only use generation projects that have been committed to or consented as a basis for determining the benefits of transmission investments. This however does not consider the long term trends, impacts of technology shifts etc; or 3) Take account of the investments made by generation investors in generation projects that are committed, consented or are in the consenting process and integrate a representative sample of these into the early stages of each of the longer term scenarios. This is easily achieved by fixing the commissioning years of certain projects in the initial years of each scenario, rather than have them dependant on the cost drivers used in the medium and longer term modelling. Transpower considers that the third approach is the most appropriate to use for the Wairakei Ring analysis. The approach provides several advantages over 1 and 2 as it: recognises that while there is no assurance that a generation project is going to be built, some generation projects are more advanced and more likely to be built than others; recognises that generation projects are, at least in the short run, subject to other commercial drivers that are difficult to model in a cost based approach; provides a logical and unbiased transition between current day activities and the longer term scenarios; and recognises the geographical and technology diversity of generation projects around the country that are actually being invested in. Transpower considers that the key is to ensure that the approach is implemented without any bias towards any particular transmission investment or generation project. To achieve this, Transpower has applied the following constraints: Only generation projects that have been committed to publicly by the investor should be considered for fixing in the early years of the scenarios i.e. the generation project is classed as committed, is already consented or is in the process of being consented. Generation projects are only fixed in the first 7 years of the scenarios (between 2009 and 2016). Beyond this, any number of projects that are only now being investigated, or are currently known as a potential resource, could be developed. Only a representative selection of generation projects should be used as some will remain simply as options that will never be exercised. It is worth noting that the total quantity of generation that is publicly committed to, as defined above, is enough to supply demand growth within the 8 year timeframe several times over. All publicly committed generation projects should be treated equally and there should be no bias towards any one region or developer. The selection of generation projects with commissioning dates to be fixed should be consistent with the supporting SoO scenario. 9

10 A stable demand/supply balance should be maintained over the time period as far as possible. The generation from the fixed projects should be an underestimate of the total required over the time period, so that GEM can fill in the gaps. Too many fixed projects would amount to transmission planning based on generation over-build; too few would mean that the final scenario bears less relationship to current build plans. To satisfy these criteria, Transpower has undertaken a process of determining a representative mix of generation projects for each scenario. In summary, the process followed consists of the following steps: 1) Generation projects already committed to are identified and included in each scenario. There are eight such projects. 2) Each generation project that is either consented or currently in the consenting process (of which there are 32) is assigned a first build year as a likely first possible commissioning date, based on the status of the project and recent experience of similar projects. 3) 5000 random samples are then created. This ensures an unbiased final selection of generation projects. 4) Five representative samples are chosen out of the 5000 random samples which align with the technology descriptions used to define each SoO scenario. As a result of this analysis, a list of generation projects and fixed commissioning dates has been established. This is set out in Appendix D Carbon costs, technology trends, fuel costs and other inputs Along with the generation list itself, there are a number of other inputs used within the GIT analysis. These have largely been taken from the 2008 SoO. The following is a list of those assumptions, the detail of which is included in Attachment F: generation details (technology, costs, heat rates, location, MW, forced outage rate, annual GWh, earliest build year, retirement year, fixed build year); technology trends; fuel costs and trends; carbon sequestration factors and costs; fuel resource limits (limits the availability of gas and LNG in the modelling); LNG prices; plant efficiency adjustments; renewables capacity targets; carbon charge data; emission factors; demand data (GWH by load block and peak MW); adjustments to NZ peak demand; wind states (correlation between wind sites); 10

11 hydro output adjusters; and demand data Commission approval of alternative scenarios Transpower considers the updated market development scenarios are more appropriate than the possible future scenarios outlined in the SoO. Transpower seeks a determination from the Commission for the purposes of clause 6.1 of the GIT that the scenarios are more appropriate than the use of the SoO scenarios Generation scenario weightings Transpower has used equal weighting for the 5 scenarios (the same as the weighting for the scenarios set out in the SoO). 2.2 The modelling approach Transpower has modelled the Wairakei Ring using two different methods: The GEM Model 1. a modified GEM model; and 2. SDDP. Transpower has utilised the Commission s GEM model as the primary tool for investigating the impacts of a possible upgrade to the Wairakei Ring. In order to do this, Transpower had to incorporate a number of changes. The two most significant are: inclusion of a meshed transmission network model; and the addition of multiple wind states to reflect the variability of wind. To assess the impact of regional investments such as a Wairakei Ring upgrade, the original two node GEM model was insufficient. As such, Transpower had to modify the Commission s GEM model to include a meshed regional transmission network. This involved incorporating a DC power flow into the model using a simplified 36 node network. Associated changes were made to the model include: Incorporation of transmission constraints that span multiple transmission links. Disaggregation of the demand data into GXP s. The change to a regional network required a mapping of GXP demand to the regional network representation. The introduction of multiple winds states allowed for assessment of the impact that wind has on peak transmission flows. For each load block there is now the ability to model up to 10 wind states per load block. The Wairakei Ring only utilises two wind states per load block. Other changes made include simplifying the formulation to increase solution times and the ability to handle multiple wind states, changes to the input data to allow for batch running of scenarios and changes to the output data format to allow for easier analysis of the outcomes. As a result of the increased complexity in the transmission modelling, the GEM model has been run in a relaxed model. This means Transpower has run all the analysis using linear solutions. Testing of the impact of this has shown that the overall impact of this simplification is not significant and it allowed for faster solution times that enabled appropriate scenario analysis to be undertaken. Transpower also considers 11

12 this to be a reasonable approach given the incremental nature of most transmission investments being studied for the Wairakei Ring. The tests undertaken to demonstrate the impact of the wind states, the linear approximations and the verification of the power flows benchmarked against Digisilent are set out in Appendix C of this report. Transpower considers that these changes are appropriate for the Wairakei Ring SDDP Therefore, additions to the assumption list set out in section 2.1 above are: Network parameters for the simplified network; Constraint equations for each transmission configuration; and Wind states for each wind generator. Transpower has utilised SDDP to verify the operational costs derived from the GEM model. The results of the SDDP analysis are shown in Attachment A the GIT results. The inputs into SDDP were aligned with the inputs used for the scenario modelling in GEM. The generation build schedule which is output from GEM for each scenario was also used as an input to SDDP to specify the build date for each generation station. 2.3 The base case The GIT requires the net market benefits of each short list option for each market development scenario be compared with the corresponding scenario developed for the Base Case. In the case of an economic investment, Transpower considers a reasonable Base Case would be to do nothing. That is, the Base Case takes account of the situation that is likely to arise if Transpower did not make any investment. The use of a do-nothing base case fits with the concept of an economic investment, which can only be justified when the expected net market benefit is positive. In this case, Transpower has interpreted do nothing to mean continuing with the current transmission configuration in the Wairakei Ring region. Transpower has assumed that the capacity in the region will continue throughout the analysis period and therefore, by implication, any asset replacement that might occur during the period would continue the current capacity limits. 2.4 Net Present Value versus Real Options analysis Clause 13 of Schedule F4 requires that: Either standard net present value analysis or real options analysis must be applied in assessing the expected net market benefit of a proposed investment or alternative project.. Transpower considers that a real options approach is not practicable, and is at a level of complexity unnecessary for this Proposal. Power system analysis is complex and time consuming, and the integration of the power systems analysis required to compare the alternative development plans with the economic modelling of costs and benefits in a real option setting would not be practicable. As such, Transpower has utilised an NPV approach for this Proposal. 12

13 2.5 Staged development plans The options assessed incorporate a staged investment plan. That is, each option consists of a series of investments rather than a simple single investment at one point in time. The staging reflects the increasing need for transmission capacity over time. The details of how the staged development plans have been developed are set out in Attachment A, the GIT Results report. Of note, for most of the short-list options second stage investments are uneconomic when assessed against the SoO scenarios and therefore not included at this time. 2.6 Discount rates Clause 14 of Schedule F4 of the Rules requires that the discount value used in present value calculation be: the discount rate determined by the Board, from time to time, for the purposes of this grid investment test; or if the Board has not determined a discount rate for the purposes of clause 14.1, a discount rate of, or equivalent to, a pre-tax real rate of 7% Accordingly, Transpower uses the required pre-tax, real rate of 7%, with sensitivities of 4% and 10%, in its GIT analysis. 2.7 Analysis in current or future dollars The economic analysis has been undertaken in 2008 dollars. Results from the economic analysis are reported in commissioning year dollars as this is the amount Transpower is seeking approval for. 2.8 Analysis period Clause 27 of Schedule F4 of the Rules requires that in applying the GIT, the market benefit of a proposed investment or alternative project be assessed: over a period of 20 years from the commissioning date (unless significant market benefits or costs are expected to arise from the proposed investment or alternative project after that time, in which case the then present-value of any future benefits may also be included ) The generally accepted approach in cost benefit analysis is used to assess the cash flows arising from costs and benefits over the full economic/useful life of a proposed investment. Transmission lines have an expected life of 30 or more years and it is expected that there will be significant benefits arising from any upgraded transmission circuits during the period between 20 years and the end of its expected life. Transpower therefore has used a 35 year analysis period in its application of the GIT. However, the results of the analysis have been reported using a terminal value approach, with the GIT results being reported for a period of 20 years after commissioning and future benefits after 20 years being reported as a terminal value. 3 Market costs The definition of cost for the purpose of the GIT analysis is set out in clause 23 of Schedule F4 of the Rules. 13

14 The costs used in the short-list options are expected costs. Expected costs represent the estimated cost (referred to as the P50 cost) plus a contingency for accuracy in the scope of the proposal or alternative options (referred to as scope allowance). This allows for unexpected variations in the design scope and a standard allowance, for items not considered in the design. In this respect, expected costs are more than a P50 estimate and represent the maximum cost of the short list option, excluding financial contingencies such as exchange and commodity cost variations, and interest during construction. For the purposes of a GIT application, Transpower considers that costing accuracy needs to be sufficient to be able to: (1) distinguish between the net market benefits of the options being considered. If the difference in the net market benefits between two options is small, then it may be necessary to further refine the costs of one or both of them (or the benefits); and (2) determine that the benefits are greater than the costs involved. However, for the purpose of determining whether an option has satisfied the GIT there is no merit in refining costs beyond the margin of error in the overall net market benefit calculation. Consequently, for the purposes of applying the GIT, Transpower has costed options to a level of accuracy that: allows for the difference in the net benefit of each option to be identified; and is sufficiently robust for a decision to be made on which option should be put forward as the proposal. Although it is slightly inconsistent to use expected costs for transmission and estimated costs for generation and all market benefits, it is a conservative approach which will tend to underestimate the value of the expected net market benefit of any short-listed options. Transpower considers that the rigour and comprehensiveness of this analysis is commensurate with the estimated capital expenditure required for the Proposal. No contingency or equivalent is included in the market benefit calculation for scope allowance. The cost categories considered in this analysis are: transmission costs for each transmission option; statutory compliance costs; inflation costs; community care costs; interest during construction; exchange rates; and value of lost load. Each of these are discussed below 3.1 Transmission costs The process of costing a transmission solution is a process of refining the cost estimates over time, recognising there is a trade-off between cost accuracy and timely and efficient investment decisions. This is similar to any cost estimation process where, as more detailed design is undertaken, costs become more refined. 14

15 The costs used and the costing approach for the Proposal are set out in detail in Attachment D, the cost report. As noted, the approach to the Wairakei Ring Project has been to analyse the costs to an accuracy required to robustly differentiate between the options. It is Transpower s expectation these costs have been presented at a Preliminary Study level. The transmission costs used in this Proprosal have been summarised into the following components: line capital costs; substation costs; property and easement costs; consenting costs project management; and operating and maintenance costs. Each of these factors is described below. a. Line Capital costs Transpower has considered the capital costs of the equipment that would be incurred prior to the commissioning of each option, including line works and any cable works required for each option. New transmission line costs are based on indicative costing corridors only. b. Substation Capital costs Transpower has considered the capital costs of the substation equipment that would be incurred for each option. c. Property and easement costs Transpower has taken into account the probable range of property and easement costs as relevant for each short-listed option. d. Consent costs The costs associated with satisfying the requirements of the RMA and other relevant legislation have been estimated and taken into account in the cost analysis. These costs include those incurred in obtaining designations, resource consents and other permissions necessary to undertake works contemplated in the short list of options. e. Project management costs Transpower has allowed for project management costs to reflect the cost of managing the design, construction and commissioning of works to be undertaken for each short list option. f. Operating and maintenance costs Operating and maintenance costs over the operating life of each short-listed option are included in the analysis. 15

16 3.2 Statutory compliance costs and benefits The relevant statutory costs have been identified and are included as necessary. 3.3 Inflation The expected cost of each short-list option has been derived in real terms and the results are presented in dollars relative to the current year. However, property costs have been inflated by 3% p.a. to reflect the real average increase in property value. 3.4 Community care costs In 2006, Transpower established its CommunityCare Fund, which recognises that Transpower needs to offset the impact on communities of major grid investment projects and contribute to communities where we operate, particularly those communities which accommodate Transpower assets which benefit the whole country. The Fund has been developed along similar models used by other power companies in New Zealand and overseas and aims to ensure that the mitigation or benefit the community receives reflects the impact of Transpower s works in that community. Transpower's community funding approach is aligned with project approval. It is based on a funding formula derived from an impact analysis and the number of kilometres of line in an impacted area, multiplied by the cost of the line per kilometre. The impact scale is determined on a number of factors, including the size of the line required, the impact of the line within the region it traverses, the population of the communities it crosses, and the visual impact of the line. The impact scale is capped at 1% of the project line cost. The funding formula is: 1% x number of kms x cost per km of overhead line In accordance with this approach $600,000 has been added to each option. This is the cap for community mitigation and the costs will be recovered as they are incurred up to a maximum of $600,000. When an investment proposal is approved by the Commission, community organisations in the affected area may apply to the CommunityCare Fund for nominated projects that meet specific funding criteria. The CommunityCare Fund is managed in accordance with the guidelines established by the Office of the Auditor General (OAG). 3.5 Interest During Construction Interest during construction has not been included in the GIT analysis, but is reflected in the approval cost calculation for the Proposal. The reason for this approach is that financing costs are not added within the GIT and it is incorrect to add them to discounted cashflow analysis. In addition, IDC does not impact on the GIT results in that: it does not change the ranking of the options, and it does not impact on whether the expected net market benefits of the Proposal are greater than zero, as the market benefits of the Proposal substantially outweigh the market costs. 16

17 3.6 Exchange rate approach Where costs have a foreign exchange component, Transpower has used an exchange rate determined by taking an average rate calculated over +/- 20 business days from 1 September As the foreign exchange component can in some cases differ significantly between short-listed options, the exchange rate assumption may be significant and so Transpower has also undertaken a sensitivity run on the exchange rates using the 10 year average rates. 4 Market benefits The GIT requires Transpower to analyse the market benefits of the options which will make up its short list. Market benefit for the purposes of the GIT is defined at clause 27 of Schedule F4 of the Rules. To evaluate the economic benefits of the short-list options versus the Base Case the key measures used are: capital benefits; fuel costs and transmission losses; greenhouse gas emission, spill and load shedding; ancillary services; instantaneous reserves; competition benefits; consumer benefits; terminal benefits; and clause 9 benefits. Each of these measures is discussed in more detail below. 4.1 Capital benefits These are benefits in capital reductions or deferrals arising from reductions in the need for, or deferral of, new generation (modelled projects), or new transmission investment (either alternative projects or modelled projects). Some of the capital expenditure savings result from a changing mix of new generation required. 4.2 Fuel and transmission loss benefits Fuel cost benefits arise when a proposed transmission investment enables generation plants with lower fuel costs to be dispatched. There are benefits from differences in fuel costs and reductions in transmission losses (which reduce the amount of fuel consumed) associated with various options. 17

18 4.3 Greenhouse gas emission, spill and load shedding benefits These are benefits from reductions in greenhouse gas emissions (valued at the carbon charge appropriate to the generation scenario), reductions in the amount of hydro spill and reductions in the amount of load shedding. 4.4 Ancillary services benefits Ancillary service benefits have not been quantified as part of the Proposal. Given the scale and the impact that each option may have on ancillary services, Transpower does not consider that it is appropriate. However, benefits from Ancillary Services are likely to be favoured by the higher capacity options. 4.5 Competition benefits Competition benefits can be included in the GIT under the Rules, 4 but due to the difficulty in calculating these and the resultant uncertainty over their magnitude, Transpower has not quantified competition benefits other than those already reflected in the generation expansion modelling. However, competition benefits are likely to favour the higher capacity options Terminal benefits Terminal benefits are included in the analysis. They have been calculated as being the benefits that would accrue assuming an asset life of 35 years i.e. from 2035 to This is consistent with the analysis period as discussed in Section Other benefits The GIT allows for the situation where a material market benefit (or cost) cannot be quantified. In such a situation, the direction of the market benefit and likely magnitude of the market benefit must be identified 5. In this context Transpower has included those benefits above that have not been quantified along with a qualitative assessment of consumer benefits, market efficiency benefits and longevity related benefits. 5 Sensitivities The GIT provides that sensitivity analysis may be conducted to determine if the preferred investment is sufficiently robust. Transpower has conducted a sensitivity analysis for the Proposal in accordance with clause 17 of Schedule F4 of the Rules. The treatment by Transpower of each of the items for sensitivity analysis as listed in clause 17 is considered below. Notably, the sensitivity calculations also need to be considered in light of the modelling approach that has been adopted. 5.1 Forecast demand A sensitivity is included in the GIT analysis for each of a low, medium and high demand forecast. Transpower has used the Commission s high, medium and low forecasts for this investigation. 4 Clause 10 of Schedule F4, Part F of the Rules. 5 Clause 9 of Schedule F4, Part F of the Rules. 18

19 5.2 Variations in the size, timing, location, and operating and maintenance costs Variations in the: size, are reflected in the choice of alternatives; timing, are considered within the timing analysis; location, are not relevant for either investment as the need for the investment has been identified specifically in the Wairakei Ring region; and operating and maintenance costs, have not been considered as a sensitivity as the impact is not significant. 5.3 Capital cost Capital costs of the short-listed alternatives have been analysed with sensitivities that recognise the level of uncertainty in the cost estimates. 5.4 Timing of decommissioned assets Several options require assets to be decommissioned. However, the timing of the decommissioning does not have significant cost implications; therefore, it has not been considered as a sensitivity. 5.5 Value of expected unserved energy There is no unserved energy within the modelling; therefore, it has not been included as a sensitivity. 5.6 Discount rate Transpower has tested the options using a 4% and 10% discount rate. 5.7 Variation in hydrological inflow sequences Variations in hydrological inflow sequences have been considered within the base analysis. Accordingly, Transpower does not consider it reasonably necessary to include this in the sensitivity analysis. 5.8 Generator and demand side bidding strategies Different bidding strategies have not been considered in the sensitivity analysis. However, the equivalent has been considered in the base analysis by considering the cost drivers in GEM and SDDP which differ across scenario. 5.9 Competition benefits Due to the lack of agree method for determining competition benefits, Transpower has not conducted a sensitivity analysis on competition benefits Carbon Charges A sensitivity on carbon charges, and carbon tax benefits has been included. 19

20 5.11 Probability of occurrence of market development scenarios Transpower has not run a sensitivity on the scenario weightings. However, the results from each scenario is been reported individually Exchange rate sensitivity using 10 year average exchange rates For the purposes of the sensitivity, Transpower has used a 10-year average exchange rate for sensitivity analysis Additional Sensitivity - Property Cost Transpower has included a sensitivity on property costs. This is to verify the ranking of new build options relative to the reconductoring options. 20

21 Appendix A Summary of Assumptions The following table is a summary of the approach and the assumptions used in the analysis. A.1 Approach and Assumptions Approach Modelling approach and market development scenarios (MDS) GEM Model Scenario weightings Net present value or real options analysis Value 5 Scenarios based on 2008 SoO. Initial years updated, generation details updated. Incorporation of regional transmission network and wind states Equally weighted NPV with scenarios Generation assumptions Updated from SoO 2008, including fixing build years from 2008 to 2016 in scenarios. Other GIT model inputs Network parameters and constraint equations Demand Forecast Included as per 2008 SoO Included, based on reduced network model SoO. Discount rate 7% Analysis in current or future dollars $2008 Staged development plans Staging of options included Approach to calculating market costs Transmission costs Line and Substation capital costs Property and easement costs Included Included Project management Included Consenting costs Included Operating and maintenance costs Included Decommissioning costs Included Generation costs Capital Based on the Commission Scenarios and altered after consultation Fuel Included Carbon Charges Included Statutory compliance costs Included Inflation costs Included, 3% Approach to calculating market benefits Community Care costs Interest during construction Exchange rates Reduction in capital costs Reduction in fuel costs and transmission losses Greenhouse gas emission, spill and load shedding Market Services benefits (Ancillary services Included, 1% of line costs In approval amount but not in GIT. Average over +/- 20 business days, sensitivity on 10 average Included and quantified Included and quantified Included and quantified Included but not quantified 21

22 Approach and Reserves) Competition benefits Consumer benefits Market Efficiency Value Included but not quantified Included but not quantified Included but not quantified Longevity related benefits Included but not quantified Terminal benefits Included and quantified (final 8 years of benefits) Analysis period 2008 to Presented at 20 year plus terminal benefit. A.2 Sensitivities Sensitivity Forecast demand Variations in the size, timing, location, and operating and maintenance costs Value Included - High and low demand as per 2008 SoO Not included as a sensitivity Capital cost Included, Low 80%, high 120% Timing of decommissioned assets Value of expected unserved energy Discount rate 4% and 10% Variation in hydrological inflow sequences Generator and demand side bidding strategies Competition benefits Not significant therefore not included n/a Not included as a sensitivity, modelled within SDDP analysis Not included as a sensitivity, included in scenarios Not included Carbon charges Included, low 80%, high 120% Property Costs Included, 200% 22

23 Appendix B Glossary Term Alternative Project Economic Investment Electricity Commission GEM GIT GPS Grid GRS GUP MWh mds NZES Primary Transmission Equipment Reliability investment RFI Description Options which are reasonable alternatives to any investment proposal as defined in clause 19, Schedule F4 of the Rules. Investments in the grid that can be justified on the basis of the Grid Investment Test set out at Schedule F4 of the Rules and are not reliability investments. A Crown entity established under the Electricity Act 1992 to oversee New Zealand s electricity industry and markets. The Electricity Commission s Generation Expansion Model. Modified by Transpower for the purpose of the Wairakei Ring Investigation. Grid Investment Test. A test for reliability investments and economic investments in the grid developed in accordance with Rule 6. The specific rules defining the Grid Investment Test, are set out in Schedule F4 of Part F of the Rules. Government Policy Statement on Electricity Governance. Is the system of transmission lines, substations and other works used to connect grid injection points and grid exit points to convey electricity throughout the North and South Islands of New Zealand. Grid Reliability Standards are standards for reliability of the grid developed in accordance with Rule 4. The standards themselves as currently developed are detailed in Schedule F3 of Part F. Grid Upgrade Plan. A plan for grid expansions, replacements and upgrades. Megawatt hour of electrical energy Market development scenario New Zealand Energy Strategy 2050, published by Ministry of Economic Development, October Any plant or equipment forming part of the grid which enables the bulk transfer of electricity, including without limitation transmission circuits, busbars and switchgear. Investments by Transpower in the grid, or alternative arrangements by Transpower, the primary effect of which is, or would be, to reduce expected unserved energy. Request for Information RMA Resource Management Act 1991 Rules SDDP SoO Transpower USE VoLL The Electricity Governance Rules Unless otherwise specified, the reference to the Rules in this document is to the rules in Part F, Section III of the Electricity Governance Rules A model developed by Power Systems Research in Brazil. Statement of Opportunities developed by the Electricity Commission in accordance with Rule 9. Transpower New Zealand Limited, owner and operator of New Zealand s highvoltage electricity network (the national grid). Unserved Energy Value of Lost Load 23

24 Appendix C GEM Modifications and tests As discussed in Section 2.2 Transpower has used a generation expansion model to determine possible generation build scenarios. The Commission s GEM model has been chosen for this work, but various changes have been made to make it more suitable for the current purpose. GEM is a mixed integer linear program (MILP) formulated in the Generalised Algebraic Modelling System (GAMS). This section outlines the changes made to GEM. C.1 Base Version The model is based on version of GEM. C.2 Wind Generation Wind generation is represented by an average availability in the standard GEM model. This representation does not capture the inherent intermittency of wind generation resources but rather modelled wind as being perfectly reliable at its average availability. The overall effect of this is that GEM underestimates the impact of supplying the load in the absence of wind. To address this, a wind state representation was introduced within GEM in each of the load blocks to model the fact that wind could take on one of several states within each of the load blocks. 6 This wind state representation is an approximation of the output distribution of the wind generators and specifies the available capacity and amount of time the wind resource is available at this capacity. Example wind states are shown in Figure C-1. Figure C-1: Multiple wind states representation 100% Actual Multiple Wind States Average 80% 60% Output 40% 20% 0% 0% 100% Year In theory it would be possible to have separate wind states for each generator, but then the number of states overall would be exponential in the number of generators. 6 This solution is the same as that adopted by the Commission in version of GEM. 24

25 For this reason there can only be one set of states across all generators, and some assumptions must be made about the correlation of their output. A variety of wind state configurations were tested. The effect on the generation build is minor but probably greater than the errors involved, so the most accurate feasible representation is used for base runs and two extremes for sensitivities. Several different wind state representations were used to determine the impact of this representation on the wind generation build plan. The wind generation representations that were used for the tests are as follows: Average wind: This is the standard wind representation in GEM where a single wind state is assumed. Partial correlation 5 Wind States: Each wind generator is modelled as taking on 5 different states in each load block. Partial correlation implies that the diversity across the wind sites is taken into account. This was accomplished by calculating a potential wind output duration curve for the whole of New Zealand using current and potential future sites. Each individual wind generator was then allocated a proportion of this total output based on their contribution to the total output. Perfect correlation 5 Wind States: Again each wind generator is modelled using 5 capacity states but the diversity across the wind sites was ignored. This implies that all wind generators were assumed to be unison in the different states. That is there was perfect correlation of the wind states of different generators. Perfect correlation 2 Wind States: This is the same as the previous scenario except that only 2 wind states were modelled. In conjunction with the wind state representation, the cost of unserved energy (CUE) was also adjusted. It was found that this together with the wind state representation influenced the wind build in GEM. The results of these tests are shown in Figure C- 2. Figure C- 2 : Wind Build versus cost of underserved energy and wind representation 35.00% 30.00% 25.00% % W in d In sta lle d 20.00% 15.00% 10.00% Avg Wind Partial Correlation - 5 Wind States Perfect Correlation - 5 Wind States Perfect Correlation - 2 Wind States 5.00% 0.00% CUE = $500/MWh CUE = $5,000/MWh CUE = $20,000/MWh CUE = $1,000,000/MWh Cost of Unserved Enegy ($/MWh) The results of this test illustrate the impact of modelling the unavailability of wind. The proportion of installed wind capacity reduces as the intermittency in wind availability is introduced. As the representation of this intermittency becomes more extreme the proportion of installed wind reduces even further. There is also a correlation between the cost of unserved energy and proportion of installed wind generation when the multiple wind state representation is used. The reasons for these relationships are intuitive. As wind becomes more variable more generation needs to be put in place to supply the load when the wind generators are not 25

26 available. In the extreme case this is when all wind is perfectly correlated and can only take on one of two states (on or off). Furthermore, as the cost of unserved energy is increased the effective cost of wind is increased due to the unserved energy that might arise during the low wind availability states. This makes other more reliable generation sources more cost effective in relation to wind generators and thus results in a reduction in the proportion of wind generators installed. The extreme wind representation in terms if intermittency is when all the wind generators are perfectly correlated in terms of their availability (no diversity) and can only assume one of two states (on or off). This is the extreme to the average availability and together with the average availability representation provides the boundaries to wind representation in GEM. C.3 Network Representation DC Load Flow The standard GEM network representation uses a transportation model that consists of nodes and arcs and is used to represent networks in which the flows on different arcs are independent of each other. This representation can be a good approximation of power flow on radial electric networks but is not suitable for representing the power flow on electric networks which contain loops ( meshed networks ). This is because the power flows on different arcs are interdependent in meshed electric networks. Modelling the effects of the constraints and upgrade options in the Wairakei Ring requires taking account of loop flows. The standard 7 linear DC approximation to AC power flow model (loosely DC load flow model ) was used to accomplish this while still maintaining the linearity of GEM. Reduced Network The Wairakei Ring and lower South Island studies require a model of the New Zealand grid that accurately represents the flow on the relevant lines. A full model of the grid would be too slow to execute, so a reduced network is required that has sufficient detail for the study area. The model that was derived (see Figure C-3), includes 33 nodes and 46 branches. 7 See, e.g., Arillaga, J and Arnold, C P, Computer Analysis of Power Systems, Wiley,