Investment Planning for Jamaica's Electricity Sector

Investment Planning for Jamaica's Electricity Sector

Long-term Infrastructure Investment Planning for the Electricity Sector in Small Island Developing States Case Study for Jamaica Travis Atkinson Paul V. Preckel Douglas Gotham November 4, 2019 Funding support from: Jim & Neta Hicks Graduate Student Small Grant Need for New Capacity Increasing Electricity Consumption per capita World Bank (2019) 02/25/2020 Increasing Electricity Access (% of population) World Bank (2019)

2 Problems Common practice in SIDS to focus only on generation expansion planning (GEP) o Neglect (or only subsequently account for) transmission constraints and loop flow Loop flow: a phenomenon intrinsic to electricity networks that cause electricity to flow along all paths connecting two nodes, not just the shortest path Can misalign social costs and is exacerbated by complexity of network (Chao & Peck, 1996) Current frontier of research is to co-optimize generation and transmission investments (Krishnan et al., 2016) Paucity of empirical research applicable to SIDS context o (Hemmati et al., 2013; Lu, et al., 2016, Nahmmacher, & Schmid, 2015; Roh et al., 2007; Roh et al., 2009; Sauma & Oren, 2006) Most planning tools are proprietary o Freely available tools limit programming flexibility; they do not allow modeler to look under the hood 02/25/2020 3

Research Questions Is simultaneously planning for generation and transmission investments more efficient than sequential planning approaches? What is the impact of loop flow on long-term investment planning? 02/25/2020 4 Snapshot of Results Simultaneous investment planning is more efficient than sequential investment planning but, difference is less than expected. Loop flow, in the Jamaican context, does not appear to meaningfully impact infrastructure investment decisions. 02/25/2020 5

Jamaica as a case study 1/3 of SIDS found in Caribbean 3rd largest Caribbean island, largest English-speaking Caribbean island Explicit policy objectives o National Energy Policy (2009) (3 of 7 priorities) i. Enhancing security of energy supply through diversification of fuels ii. Modernizing the countrys energy infrastructure iii. Development of renewable energy sources such as solar and hydro 02/25/2020

6 Jamaica as a case study Empirical Motivation Increasing Expected Electricity Demand Recent developments in sector Empirical Motivation: o 2010 generation plan (only one publicly available) neglects transmission investments and therefore loop flow o Officials say they account for loop flow but no publicly available report o Integrated Resource Plan still outstanding Source: OUR (2016) 02/25/2020

7 Jamaicas Electricity Network Source: JPSCo Cost Table 02/25/2020 8 Methodology Dynamic Optimization Model Direct Current Optimal Power Flow (DCOPF) (Krishnan et al., 2016) Modified to include greater operational details according to long term investment model designed by Purdue Universitys Power Pool Development Group 02/25/2020 9 Inputs

Optimization Output Supply Generation Assets(EIA, 2016, OUR, PCJ) Transmissio n Assets (EIA, 2016) Demand Demand Forecast by rate class (OUR, 2016) Aggregate peak demand forecast (OUR, 2016)

Regional distribution (computed) Costs Sequenti al Cost Minimization Operating Cost Investment Cost Investment Decisions Location Timing

Generation Portfolio Generation by technology Electricity Flows Flows Shadow Prices Prices Investment Costs (EIA, 2016) Fuel Costs (EIA, 2018) O&M Costs (OUR, 2017) 02/25/2020

Engineerin g Constraints Economic 10 Assumptions Deterministic model Central planning perspective o Vertically Integrated Monopolies Often State-owned o Establish socially optimal baseline o For Jamaica (case study): Government has 20% stake in transmission monopoly/generation incumbent Capacity decision rests with the Ministry of Science, Energy & Technology (Electricity Act, 2015) 02/25/2020

11 Results Cost comparison across model specification (reference case) Units Total Cost US $ mil. Difference* US $ mil. Difference* percent Simultaneou s 2,727

Sequentia l No Loop flow 2,731 2,727 3 0 0.12% 0% * Relative to simultaneous model Network 02/25/2020 12 Results

Optimal Generation Infrastructure Investments (Simultaneous Model) Source: JPSCo (modified) Black shapes represent new generation investments; year of investment in call-out boxes; hydro capacity indicated in parenthesis; no new transmission infrastructure 02/25/2020 13 Results Optimal Generation Investments (Sequential Model) Source: JPSCo (modified) Black shapes represent new generation investments; year of investment in call-out boxes; hydro capacity indicated in parenthesis 02/25/2020 14 Results Optimal Transmission Infrastructure Investments (Sequential Model) Source: JPSCo (modified) Black shapes represent new transmission line investments; year of investment in call-out boxes

02/25/2020 15 Results Generation Portfolio (Simultaneous Model) 7000 6000 5000 wind solar NG hydro HFO ADO GWh 4000 3000

2000 1000 0 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 Years 02/25/2020 16 Sensitivity Analysis Fuel Price Uncertainty 02/25/2020 17 Sensitivity Analysis Results are robust but demonstrate that simultaneous planning is even more important in the face of fuel price uncertainty.

Model Type Baseline Fuel High Fuel Price Low Fuel Price Scenario Scenario Price Scenario Simultaneous Model Sequential Model Model excluding loop flow Total Cost (US$mil) 2,727

2,785 2,484 Total Cost (US$mil) 2,731 2,800 2,496 Difference (US$mil)* 3 15 12 Difference (%)* 0.12% 0.52%

0.47% Total Cost (US$mil) 2,727 2,785 2,484 Difference (US$mil)* 0 0 0 Difference (%)* 0% 0%

0% * Relative to simultaneous model associated with the same fuel price scenario 02/25/2020 18 Conclusion Focus: o Evaluate implications of investment planning approaches for electricity sector in SIDS using Jamaica as a case study o Assess economic significance of loop flow in a SIDS context Results o Simultaneous planning is more efficient than sequential planning and is worth the modest additional computational requirement. o Fuel price uncertainty exacerbates the differences between model specification. o Loop flow was not found to be economically significant. Contributions o Empirically compare energy modelling practices o Framework for external validation of IRP o Gain insights relevant for Small Island Developing States (SIDS), largely ignored by existing literature

02/25/2020 19 Future Work Evaluating energy policy options for SIDS using Jamaica as a case study o Renewable portfolio standard o Carbon tax o Production tax credits o Capital investment subsidies 02/25/2020 20 Thank You [email protected] 02/25/2020

21 Appendix 02/25/2020 22 Generation Investments Low Fuel Price Scenario High Fuel Price Scenario Lower investments in new thermal plants due to greater utilization of ADO & HFO; renewables satisfy (small) excess demand Renewable resources become more attractive 02/25/2020 23

Transmission Investments Low Fuel Price Scenario 02/25/2020 High Fuel Price Scenario 24 Objective Function (1) (2) 02/25/2020 25 Constraints: Binary Variables (3) (4)

(5) (6) 02/25/2020 26 Operational Constraints (7) where where appropriate Operational Operational Constraints Constraints including including KVL KVL where M is a large constant (8) (8)

(9) (9) (10) (11) (10) (11) (12) (12) 02/25/2020 27 Power Flow Balance & Budget Power flow balance (13) Budget Constraint (14) Back 02/25/2020

28 Data Gaps Source: graphed based on OUR (2017) data points) Source: graphed based on OUR (2017) data (144 data points) 02/25/2020 (17,520 data Mis-matched temporal and sectoral resolutions of demand. No geographic distribution of demand 29 Data Gaps Source: JPSCo 02/25/2020

30 Average load curve by day of week 02/25/2020 Rate Class Definitio n rate 10 Residentia l rate 20 Small Commerci al/industri

al rate 40 Large Comm./in d. (<25 kVA) rate 50 Lare Comm./In d (>25 kVA) Source: graphed based on OUR rate 60 Street data lights 31 Typical Weekday Load Curve (2010)

Source: OUR, 2010 02/25/2020 32 Average Load Curves 2014 & 2017 Source: Graphed based on OUR data 02/25/2020 33 Historical sale of electricity Electricity Sales per sector (excluding losses) (GWh) Sales (GWh) 1200

1000 800 600 400 200 0 1990 1991 1992 1993 1994

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Year

Rate 10 Rate 20 Rate 40 Rate 50 Rate 60 Other Source: graphed based on OUR data 02/25/2020 34 Annual projected electricity demand Source: graphed based on OUR data 02/25/2020 (144 data points)

35 Peak Demand (2017) MW PEAK Demand 2017 750 700 650 600 R = 0.24 550 500 450 400 350 300 250 200 150 100 50 0

17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 01 09 17 25 02 10 18 26 06 14 22 30 07 15 23 01 09 17 25 02 10 18 26 04 12 20 28 05 13 21 29 06 14 22 30 08 16 24 01 09 17 25 03 11 19 27 n n n n b b b b r r r r r r r y y y y n n n n l l l l g g g g p p p p t t t v v v v c c c c Ja Ja Ja Ja F e i Fe F e F e Ma Ma Ma Ma i Ap t Ap Ap Ma Ma Ma Ma i Ju t Ju Ju Ju e Ju d Ju u Ju ri Ju Au Au Au Au S e S e i Se S e Oc Oc Oc No No No No De De De De n n e d u r t n n n u e h n r t n n e d u r t n n e d u i t n n e d r Su Mo Tu We Th F Sa Su Mon Tue ed Thu F Sa Su on Tue ed Thu F Sa Su Mo T W T F Sa Su Mo Tu We Th F Sa Su Mo Tu We Th Fr Sa Su Mo Tu We W M W Date Source: graphed based on OUR data 02/25/2020 36

Peak Demand (2014) MW PEAK Demand 2014 800 750 700 650 600 550 500 450 400 350 300 250 200 150 100 50 0

14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 20 , 01 08 15 22 29 05 12 19 26 05 12 19 26 02 09 16 23 30 07 14 21 28 04 11 18 25 02 09 16 23 30 06 13 20 27 03 10 17 24 01 08 15 22 29 05 12 19 26 03 10 17 24 31 n an an an an eb eb eb eb ar ar ar ar pr pr pr pr pr ay ay ay ay un un un un Jul Jul Jul Jul Jul ug ug ug ug ep ep ep ep ct ct ct ct ct ov ov ov ov ec ec ec ec ec a J J J J J F F F F M M M M A A A A A M M M M J J J J d d d d d A A A A S S S S O O O O O N N N N D D D D D ed ed ed ed ed ed ed ed ed d d d d ed ed ed ed ed d d d d ed ed ed ed e e e e e d d d d ed ed ed ed ed ed ed ed ed d d d d ed ed ed ed ed W W W W W W W W W We We We We W W W W W We We We We W W W W W W W W W We We We We W W W W W W W W W We We We We W W W W W Date Source: graphed based on OUR data 02/25/2020 37 Jamaica Market Structure Generation Transmission & Distribution Consumption

JPSCo 75% Capacity, 54%Generation, Monopoly on transmission IPPs 25% capacity, 46% generation 02/25/2020 38

Recently Viewed Presentations

  • Stock Assessment Methodologies for Reef Systems: A Joint

    Stock Assessment Methodologies for Reef Systems: A Joint

    Journal du Conseil, 6:1-20) Equilibrium, Sustainability, and Maximization An equilibrium stock size has an associated sustainable equilibrium yield Maximum sustainable yield (MSY) generally occurs at an intermediate stock size and depends on the harvest pattern "MSY is the largest average...
  • A Public Health Approach to Strengthening Families and ...

    A Public Health Approach to Strengthening Families and ...

    A Public Health Approach to Strengthening Families and Promoting Social-Emotional Health in Children. Sara van Driel, Ph.D. Implementation Consultant, Triple P America
  • ATLAS ATLAS Pixel Detector Support Structure Status and

    ATLAS ATLAS Pixel Detector Support Structure Status and

    Review frame design status Recent FEA results and plans Discuss trade-off of sandwich core materials, carbon foam versus honeycomb in terms of performance definite cost impact Discuss prototypes for testing and test objectives Frame components Frame sub-assembly Current Studies Based...
  • Planning and Preparing for Emerging Infectious Diseases AN

    Planning and Preparing for Emerging Infectious Diseases AN

    "Airborne Precautions prevent transmission of infectious agents that remain infectious over long distances when suspended in the air (measles, chicken pox, TB)" Patient must be placed in an airborne isolation infection room (AIIR): this is a single-patient room equipped with...
  • Chapter 1

    Chapter 1

    Legal System Principles. 1) Equal Justice Under the Law. 2) Due Process of Law. a) substantive due process - the. substance of the law. b) procedural due process - the way. the law is administered. 3) The Adversary System. 4)...
  • 1.1 Silicon Crystal Structure - University of California ...

    1.1 Silicon Crystal Structure - University of California ...

    Semiconductor Fundamentals OUTLINE General material properties Crystal structure Crystallographic notation Read: Chapter 1 What is a Semiconductor? Low resistivity => "conductor" High resistivity => "insulator" Intermediate resistivity => "semiconductor" conductivity lies between that of conductors and insulators generally crystalline in...
  • Solving Difficult SAT Instances In The Presence of

    Solving Difficult SAT Instances In The Presence of

    Predicates Earlier work: By Crawford, Ginsberg, Roy and Luks (92,96) Not based on cycle notation for permutations Our construction is more efficient Every cycle considered separately In practice almost all cycles are 2- or 3-cycles Two types of 2-cycles: (aa')...
  • Lake County IL RACES/ARES Training Amateur Radio Emergency

    Lake County IL RACES/ARES Training Amateur Radio Emergency

    Lake County IL RACES/ARESĀ® Training Amateur Radio Emergency Communications Course Level I*