Technology Bundle Approach with Parameter estimated from Bottom-up model to Integrate between Top-down and Bottom-up Model

INTERNAL USE ONLYINTERNAL USE ONLY

Hiroshi Hamasaki Research Fellow, Economic Research Centre, Fujitsu Research Institute

Visiting Fellow, Centre for International Public Policy Studies

[email protected]

[email protected]

Amit Kanudia Partner, KanORS-EMR

Technology Bundle Approach with Parameter

Estimated from Bottom-up Model to Integrate

between Top-down and Bottom-up Model

66th Semi-annual ETSAP meeting, 17-21 November 2014


Contents

I. Overview of Linkage between CGE and TIMES

II. Top-down Model: CRESH & Tech Bundle Approach in CGE Model

III. Bottom-up Model: JMRT (Japan Multi-Regional Transmission Model)

IV. Test Simulations

V. Lessons

Copyright 2014 FUJITSU RESEACH INSTITUTE 1


I. Overview of Linkage between CGE and TIMES


TIMES MODEL

JMRT (Japan Multi-regional Transmission) Model

CGE MODEL Based on GTAP Model

CGE MODEL with Tech Bundle Technology Information

in Electricity Sector

Parameter


II. TOP-DOWN MODEL: CRESH & TECH BUNDLE APPROACH IN CGE MODEL



Base CGE Model & Database

GTAP Model version 6.2

2007 Based GTAP DB8


Description CHN China IND India JPN Japan KOR Korea ASIA Other Asia USA USA CAN Canada AUS Australia EU12 EU12 DEU Germany FRA France GBR UK RUS Russia CEU Central & Eastern Europe RoA1 Rest of Annex LSA Latin & South America RoW Rest of the World

Description agr Agriculture coa Coal oil Oil gas Gas p_c Petroleum & Coal Product ely Electricity i_s Iron & Steel nfm Non-ferrous Metal min Mineral Product crp Chemical, Rubber and Paper omf Other Manufactruing trp Transport ser Service

Sectors (13) Regions (17)


Step 1: Include “E” part in Conventional CGE

Copyright 2014 FUJITSU RESEARCH INSTITUTE 5

Output

VA

K L Intermediate

Output

VA

Capital-Energy

Composite

L Intermediate

Electricity Non-Electricity

Coal Non-Coal

Gas Oil Petroleum

Product

Conventional CGE CGE Reflects “E” Part

Capital Energy

• This assumes that there is only one electricity generation technology.

• This reflects just fuel-substitution, but technology substitutions.

• Some of technologies do not use fossil fuels.


Step 2: Technology Bundle in Japan Ely

Copyright 2014 FUJITSU RESEARCH INSTITUTE

E G H

E

K

L

Nuc Gas Coal PV

Ely_Nuc

Generation

VA

Capital-Energy

Composite

L Intermediate


Coal Non-Coal

Gas Oil Petroleum

Product

Capital Energy

Ely_Nuc Ely_Oil Ely_Gas Ely_Coal

Distribution

+Sales

＋

Generation Technology

Distribution & Sales

Electricity 𝝈 = 𝟎 I-O Table


• If all is equal, the production function is CES.

• If , Leontief • If , Cobb-Douglas

CRESH Production Function (Hanoh, 1971)


Ely_Nuc

Generation

Ely_Oil Ely_Gas Ely_Coal Ely_Wind

: Electricity generated by technology i

: Total Electricity Generation

: Price of Electricity Generated by Technology i

: Price of Electricity

: CRESH Parameter for technology i

Bottom-up Model

VA

Capital-Energy

Composite

L Intermediate


Coal Non-Coal

Gas Oil Petroleum

Product

Capital Energy

𝛾𝑖

𝛾𝑖 = 1

𝛾𝑖 = 0

Distribution & Sales

Electricity 𝝈 = 𝟎


III. BOTTOM-UP MODEL: JAPAN MULTI-REGIONAL TRANSMISSION (JMRT) MODEL

8 Copyright 2014 FUJITSU RESEARCH INSTITUTE


Overview of JMRT


Existing

PowerStation

Electricity New Technology

USC

Industry

Domestic

Transport

IGCC

GTCC

Nuclear

Manufacturing

Non-manufacturing

Household

Office

Small Hydro

Wind

PV

Geothermal

Existing

Pumped-Storage

USC: Ultra-super Critical

IGCC: Integrated Gasification Combined Cycle

GTCC: Gas Turbine Combined Cycle

Biomass

Oil

Conventional

Electric Car

Fuel Cell Vehicle

Hydrogen

Fuel Cell


10 Grids and Grid Connections


0.6GW

6GW

0.9GW

0.3GW

5.57GW

1.4GW

16.66GW

5.57GW

5.57GW

2.4GW


12 Time Slices

3 Time Periods

Day（8～13、16～23）

Peak（14～15）

Night（0～7）

4 Seasons

Spring（3～6）

Summer（7～9）

Autumn（10～12）

Winter（1～2）


Load Curve in Most Electricity Consumed day

Million kW

Peak Demand in each Year

Million kW

hr

month


Existing PowerStation Data

12 Copyright 2014 FUJITSU RESEACH INSTITUTE

Existing PowerStation Data include •Type of PowerStation •Latitude, Longitude •Prefecture •Start Year •Life Time •Electricity Generation Capacity •Availability Factor (AF)


Data of Renewable Potential


No. Prefecture

Code Lati-

tude Long-

itude Wind

Speed

1

2

3

Geothermal

Offshore Wind

Onshore Wind

GIS Data is from MOE Potential Survey

Huge Renewable Potential in Hokkaido Area.

Huge Electricity Consumption in Kanto Area including Tokyo.

1 km

mesh


Geological Information (e.g. offshore wind)


Wind Speed

Availability Factor

Wind Speed

(m/s) AF

(%)

5.5 15.8%

6 19.7%

6.5 23.5%

7 27.3%

7.5 31.0%

8 34.5%

8.5 37.9%

Distance

from road

Sea Depth

(Offshore)

Distance

from grid

Initial Cost

More than 20,000V

http://www.gsi.go.jp/KIDS/

map-sign-tizukigou-h07-

02-01soudensen.htm

http://www.gsi.go.jp/KIDS/map-sign-tizukigou-h07-02-01soudensen.htm













GIS to Calculate Dist. From Grid and Road


Onshore Wind

Offshore Wind

Road

Electricity Grid (>=20,000 volt)


CCS (Carbon Capture Storage) Potential


Source: Calculation based on METI

Potential (billion ton-CO2)

*Japan CO2 emission was 1.16 billion ton-CO2 in 2010

**Total CCS Potential is 32.8 billion ton-CO2.


Design of Simulations


Case

Reference Grid

Expansion Storage

Carbon Abatement (%)* 0, 20, 29.375, 38.75, 48.125, 57.5, 66.825, 76.25, 85.625, 95

Grid Expansion No Yes No

Electricity Storage No No Yes

Grid Expansion

Reduction below 2009 level by 2050

Storage


Parameter Estimation


JMRT Model

CRESH Production Function

i

CRESH Parameter


Onshore Wind

There are huge onshore wind potentials in north part of Japan, Hokkaido, but electricity demands are done in Tokyo. There is very weak grid connection between Hokkaido and Japan main land and the capacity of grid connection is mere 60GW. Grid expansion make possible to access to Hokkaido on-shore wind potential.

In addition, storage also plays a role to boost onshore wind. Onshore is intermittent generation technology and storage will charge excess generation from onshore wind and discharge when generated electricity is less than demand.


Technology

Onshore

Condition

Reference 1.65***

Storage 5.76***

GE 3.54 ***

***: 0.0001, **:0.001, *:0.01


Demand & RE Potential


Twh

High Demand Low RE Potential

Low Demand High RE Potential

Weak Connection 0.6GW


Solar


Elasticity is the biggest in storage scenario and storage will work as back-up battery for solar to match between demand and supply. The potential of solar are geologically equally distributed in Japan and geological un-matching between potential and electricity demands are not big.

Technology

Solar

Condition

Reference 1.24***

Storage 3.59***

GE 1.23***

***: 0.0001, **:0.001, *:0.01


Offshore Wind


Same as onshore-wind, offshore-wind benefits from both storage and grid expansion. However, elasticity in storage scenario, 7.57, is more than that in grid expansion scenario, 3.95, because offshore wind potential is geologically equally distributed all over Japan.

Technology

Offshore

Condition

Reference 4.27***

Storage 7.57*

GE 3.95***

***: 0.0001, **:0.001, *:0.01


Elasticities under several systems


Technology

Hydro Solar Offshore Onshore

Scenario

Reference 1.36** 1.24*** 4.27*** 1.65***

Storage 2.22 3.59*** 7.57* 5.76***

GE 1.39** 1.23*** 3.95*** 3.54 ***

***: 0.0001, **:0.001, *:0.01


IV. TEST SIMULATIONS



Application to the case of JAPAN

Japan’s fulfillment of Kyoto Obligation with no international emission

trade and under various assumptions regarding the electricity sector:

REF: Reference

GE: Grid Expansion Case

STO: Storage Case



Renewable Energy Generation Changes


(%)

Note: Deviations from the baseline


Economic Impacts


REF GE STO

C -0.65 -0.63 -0.45

I 0.04 0.04 0.05

G 0.27 0.27 0.28

X -4.46 -4.41 -3.86

M -2.94 -2.89 -2.45

GDP -0.65 -0.64 -0.5

Carbon Abatement Cost (US$/t-CO2) GDP Decomposition


V. Lessons

Conventional top-down model fails to represent substantially different technological futures.

Common deficiency of tech-bundle CGE is the lack of the real estimates for the model parameters.

Using parameter estimated by detailed bottom-up which is complex enough make top-down model reflect technology completeness.

Reflect the characteristics of each technology

Reflect system changes



V. Lessons

Copyright 2014 FUJITSU RESEARCH INSTITUTE

Source: Hourcade, Jaccard, Bataille and Ghersi (2006)

29

TOP-DOWN

Strength • Micro-economic Realisms • Macro-economic Completeness Weakness • Fails to represent substantially different

technological futures

BOTTOM-UP

Strength • Technology Explicitness

Weakness • Lack of micro-economic realisms • Lack of macro-economic completeness

• Reflect geographical character

INTERNAL USE ONLYINTERNAL USE ONLY 30

Copyright 2011 FUJITSU RESEARCH

INSTITUTE


Future Works


TIMES MODEL

JMRT (Japan Multi-regional Transmission) Model

CGE MODEL Based on GTAP Model

CGE MODEL with Tech Bundle Technology Information

in Electricity Sector

Parameter Demand

Price


Hydrogen Cycle

Copyright 2014 FUJITSU RESEARCH

INSTITUTE

H2

H2

Transport

Buildings

Hydrogen Station

Hydrogen

Electricity

Heat

Electricity


Technology Choices in Top-down


Capital

Energy

Source: Ban(2010)


Technology Choices in Bottom-up


Capital

Energy Existing Technology

Renewable Technology

Source: Ban(2010)

Technology Bundle Approach with Parameter estimated from Bottom-up model to Integrate between Top-down and Bottom-up Model

Data & Analytics

internal use onlyinternal

gtap model cge model

cge model copyright

coal ely

gas ely

oil ely

nuc ely

japan ely copyright