U.S. Electric Power Generation Planning under Endogenous Learning-by-Searching Technology - PowerPoint Presentation

Slide1

U.S. Electric Power Generation Planning under Endogenous Learning-by-Searching Technology Change

Tuesday, October 11, 2011Session 31: Electricity Demand Modeling and Capacity PlanningUSAEE/IAEE North American Conference, Washington DCNidhi R. Santen, Massachusetts Institute of Technology (nrsanten@mit.edu)Mort D. Webster, Massachusetts Institute of Technology David Popp, Syracuse University/National Bureau of Economic Research

2

Introduction (1 of 2)

EIA, AER 2009; EIA 2011

3

Introduction (2 of 2)

Power System Technology R&D

(Public and Private)

GovernmentMakes Environmental Policies

Electric UtilitiesBuild Power Plants Using Available Generation Technologies

Natural

Environment

1. Constraining Regulations

2. Production Support

Direct

R&D Support

New or Improved Generation Technologies

Increased Demand for Technologies

CO2 Emissions

Two main policy pathways to reduce

cumulative

power sector emissions

“Now v. Later”

“Adoption v. Innovation”

4

Research Question and Outline

Research Question:

What

is

the socially

optimal balance

of inter-temporal regulatory policy and

technology-specific

R

&D

expenditures for the

U.S. electricity generation

sector, given a

specific cumulative climate target?”

Outline for Today’s Presentation

Overview of existing electricity sector planning models’ capabilities

Introduction of the current modeling framework

Snapshots from first results

Future research

Summary

5

1. Overview of Existing Numerical Power Generation Expansion Models (1 of 2)

Top-Down v. Bottom-up ModelsTop-Down: Use Average Costs and Assume Capacity FactorsBottom-Up: Use Specific Costs (e.g., Capital, O&M, Fuel) and Solve for Capacity FactorsRigorously studying emissions potentials from the power sector requires modeling operational details of the physical system (more easily resolved in bottom-up models).

Including Operational Realism Matters!

Results Preview – Less Detail

Results Preview – More Detail

6

1. Overview of Existing Numerical Power Generation Expansion Models (2 of 2)

Common Methods to Model Technology Change and Learning DynamicsDecision Variables: Capacity Additions1. (Exogenous) Fixed Trend:CapCostt,g = CapCostt-1,g*(1+ α)2. (Endogenous) Learning-by-Doing:CapCostt,g = InitialCapCostg / (CapitalStockt,g)LBDCoeffDecision Variables: Capacity Additions + R&D Investments3. (Endogenous) Learning-by-Searching:CapCostt,g = InitialCapCostg / (KnowledgeStockt,g)LBSCoeff KnowledgeStockt,g = δΣ1:t-1R&D$t,g + R&D$t,g4. 2-Factor Learning Curves (2FLC):CapCostt,g = InitialCapCostg / [(CapitalStockt,g)LBDCoeff2 * (KnowledgeStockt,g)LBSCoeff2] KnowledgeStockt,g = δΣ1:t-1R&D$t,g + R&D$t,g

Numerical Models

Representation of

Technology

Improvement

Fixed

Time Trends (Exogenous)

Learning by Doing (Endogenous)

Learning by Researching (Endogenous)

NREL ReEDS

X

X

Gen

Star Lite

(Ramos et. al./IIT)

X

SMART

(Powell et. al)

X

MIT EPPA

(Electricity

Sector)

X

(Modified)

IIASA ERIS (Electricity Sector)

X

(Limited)

EPA/MIT

MARKAL

(Electricity)

X

7

Knowledge Stock (H)

R&D$

New Knowledge (

h)

Generation Planning Inputs

Generation Technology Costs ($/MWh)

Electricity Demand (MW/time)

Generation Technology Availability (Year)

Learning by Experience

Technology Change Module “Innovation Possibilities Frontier”

h

t

= aRD$bHΦ

Environmental Policy

New Power Plant Additions

(GW)

Production (

GWh)

Learning by Researching

2. Modeling Framework for this Research

Generation

Planning Model

CO2 Emissions (Million Metric Tons)

Generation

Planning Model

H

t

,g

=

(

1-δ

)

H

t

-1,g

+

h

t,g

8

2. Modeling Framework for this Research

Structural DetailsCentralized, social planning (decision-support model)Representative technologies of the U.S. systemRepresentative U.S. load duration curve50-year planning horizon, 10-year time stepsObjective Decision Variables (per period) (1) R&D $ (by Technology) (2) Carbon Cap (3) Generation Expansion (4) Generation OperationKey Constraints (1) All traditional generation expansion constraints (e.g., demand balance, reliability, non-cycling nuclear technology, etc.) (2) Cumulative carbon cap (3) Cumulative R&D funding account balance

Generation Technologies

Coal

Steam

GasWind

Advanced CoalGas CCNuclearSolar

Coal w CCSGas CTHydroOther

9

3. First Results: With and Without Learning-by-Searching (under a Medium Cumulative Emissions Target)

No LBS

With LBS (NPVLBS < NPVNoLBS)

10

3. First Results: Medium v. Strong Cumulative Emissions Target

Medium Target

Strong Target

11

3. First Results: Sensitivity of Innovation Possibilities Parameters (Strong CCS Possibilities under a Medium Emissions Target)

Base Case Innovation Possibilities

Strong CCS Innovation Possibilities

12

4. Future Research

Model optimal generation (carbon cap distribution) and R&D investment decisions under multiple uncertain innovation possibilities using stochastic dynamic programming

13

Summary

Studying how to balance regulatory efforts and R&D efforts for the electricity generation sector requires a decision model where the capital costs of technology change endogenously with respect to new builds (adoption)

and

new research (innovation

)

Rigorous study of emissions management from the power sector requires operational details of the physical system, embodied within bottom-up type models

.

Results

confirm both a “tradeoff” and “interaction” between adoption v. innovation for technologies with strong learning potentials (dynamics that are popular in the theoretical literature

)

More research needs to be done to 1) understand the sensitivity of innovation parameters on decisions, 2) compare these results with more traditional knowledge stock formulations, and 3) model the effect of uncertainty of returns to research on near-term regulatory and R&D decisions.

Thank You

14

Source: US EPA E-Grid Database & NPR.org

Barreto, L. and S. Kypreos. (2004). “Endogenizing R&D and market experience in the "bottom-up" energy-systems ERIS model,” Technovation, 24(8):615-629.Fischer, C. and R. G. Newell. (2008). “Environmental and technology policies for climate mitigation.” Energy Economics 55: 142-162.Hobbs, B. F. (1995). “Optimization methods for electric utility resource planning.” European Journal of Operational Research 83:1-20.Ibenholt, K. (2002). “Explaining learning curves for wind power,” Energy Policy 30: 1181-1189.Jaffe, A., and M. Trajtenberg. (2002). Patents, citations, & innovations: a window on the knowledge economy. MIT Press: Cambridge, MA, 478pp. Johnstone, N., Hascic, I, and D. Popp. (2010). “Renewable Energy Policies and Technological Innovation: Evidence Based on Patent Counts,” Environmental Resource Econ, 45: 133-155.Messner, S. (1997). “Endogenized technological learning in an energy systems model,” J Evol Econ 7: 291-313.Miketa, A. and L. Schrattenholzer. (2004). “Experiments with a methodology to model the role of R&D expenditures in energy technology learning processes.” Energy Policy, 32(15):1679-1692.Popp, D. (2002). “Induced Innovation and Energy Prices.” American Economic Review 92(1): 160-180.Popp, D. (2006). “ENTICE-BR: Backstop Technology in the ENTICE Model of Climate Change.” Energy Economics 28(2): 188-222.Popp, D. (2006b). “They Don't Invent Them Like They Used To: An Examination of Energy Patent Citations Over Time.” Economics of Innovation and New Technology 15(8): 753-776.

15

References

Title Slide Photo Credits

(from left to right): (1)

www.scientificamerican.com

(2)

http://www.pelamiswave.com

(3) Sandia National Labs (4)

http://www.metaefficient.com

(5)

http://img.dailymail.co.uk

(6)

https://inlportal.inl.gov