Economic Input-Output Life Cycle Assessment

1
Economic Input-Output Life Cycle Assessment
12-714/19-614 Life Cycle Assessment and Green
Design
2
Structure of a Process-based LCA Model
3
Criticism of LCA

•          There is lack of comprehensive data for
  LCA.
•          Data reliability is questionable.
•          Defining problem boundaries for LCA is
  controversial and arbitrary. Different boundary
  definitions will lead to different results.
•          LCA is too expensive and slow for
  application in the design process.
•          There is no single LCA method that is
  universally agreed upon and acceptable.
•          Conventional, SETAC-type LCA usually
  ignores indirect economic and environmental
  effects.
•          Published LCA studies rarely
  incorporate results on a wide range of
  environmental burdens typically only a few
  impacts are documented.
•          Equally credible analyses can produce
  qualitatively different results, so the results
  of any particular LCA cannot be defended
  scientifically.
•          Modeling a new product or process is
  difficult and expensive.
•          LCA cannot capture the dynamics of
  changing markets and technologies.
•          LCA results may be inappropriate for
  use in eco-labeling because of differences in
  interpretation of results.

4
How Research is Done

• Sitting around in an office, we were complaining
  about problems of LCA methodology.
• Realized economic input-output models could solve
  boundary and circularity problems.
• Then hard work assembling IO models, linking to
  environmental impacts and testing.
• Found out later that Leontief and Japanese
  researchers had done similar work, although not
  directly for environmental life cycle assessment.

5
Economic Input-Output Analysis

• Developed by Wassily Leontief (Nobel Prize in
  1973)
• General interdependency model quantifies the
  interrelationships among sectors of an economic
  system
• Identifies the direct and indirect economic
  inputs of purchases
• Can be extended to environmental and energy
  analysis

6
The Boundary Issue

• Where to set the boundary of the LCA?
• Conventional LCA include all processes, but at
  least the most important processes if there are
  time and financial constraints
• In EIO-LCA, the boundary is by definition the
  entire economy, recognizing interrelationships
  among industrial sectors
• In EIO LCA, the products described by a sector
  are representing an average product not a
  specific one

7
Circularity Effects

• Circularity effects in the economy must be
  accounted for cars are made from steel, steel is
  made with iron ore, coal, steel machinery, etc.
  Iron ore and coal are mined using steel
  machinery, energy, etc...

product
emissions
8
Building an IO Model

• Divide production economy into sectors (Note
  could extend to households or virtual sectors)
• Survey industries Which sectors do you purchase
  goods/services from and how much? Which sectors
  do you sell to? (Note Census of Manufacturers,
  Census of Transportation, etc. every 5 years)

9
Building an IO Model (II)

• Form Input-Output Transactions Table Flow of
  purchases between sectors.
• Constructed from Make and Use Table Data
  purchases and sales of particular sectors.
  (Note need to reconcile differing reports of
  purchases and sales...)

10
Economic Input-Output Model

• Xij Yi Xi Xi Xj using Aij Xij /
  Xj
• (AijXj) Yi Xi
• in vector/matrix notation
• AX Y X gt Y I - AX
• or X I - A-1Y

11
Building an IO Model (III)

• Sum of Value Added (non-interindustry purchases)
  and Final Demand is GDP.
• Transactions include intermediate product
  purchases and row sum to Total Demand.
• From the IO Transactions Model, form the
  Technical Requirements matrix by dividing each
  column by total sector input matrix A.
  Entries represent direct inter-industry purchases
  per dollar of output.

12
Scale Requirements to Actual Product
Engine
Conferences
Steel
. . .
20,000 Car
2500
2000
1200
800
10
Other Parts
Plastics
Electricity
. . .
2500 Engine
300
200
150
10
Steel
Aluminum
13
Example Requirements for Car and Engine
Engine
Conferences
Steel
. . .
Car
0.0005
0.125
0.1
0.06
0.04
Other Parts
Plastics
Electricity
. . .
Engine
0.12
0.08
0.06
0.004
Steel
Aluminum
14
Using a Requirements Model

• Columns are a production function or recipe for
  making 1 of good or service
• Strictly linear production relationship
  purchases scale proportionally for desired
  output.
• Similar to Mass Balance Process Model inputs
  and outputs.

15
Mass Balance and IO Model
Racing
Engine
Car Production (Motor Vehicle Assembly)
Etc.
Steel
Final Demand
Etc.
16
Supply Chains from Requirements Model

• Could simulate purchase from sector of interest
  and get direct purchases required.
• Take direct purchases and find their required
  purchases 2 level indirect purchases.
• Continue to trace out full supply chain.

17
Leontief Results

• Given a desired vector of final demand (e.g.
  purchase of a good/service), the Leontief model
  gives the vector of sector outputs needed to
  produce the final demand throughout the economy.
• For environmental impacts, can multiply the
  sector output by the average impact per unit of
  output.

18
Supply Chain Buildup

• First Level (I A)Y
• Second Level A(AY)
• Multiple Level X (I A AA AAA )Y
• Y vector of final demand (e.g. 20,000 for auto
  sector, remainder 0)
• I Identity Matrix (to add Y demand to final
  demand vector)
• A Requirements matrix, X final demand vector

19
Direct Analysis Linear Simultaneous Equations

• Production for each sector
• Xi ai1 X1 ai2 X2 . ainXn Yi
• Set of n linear equations in unknown X.
• Matrix Expression for Solution X(I
  - A) Y ltgt X (I - A)-1 Y
• Same as buildup for supply chain!

20
Effects Specified

• Direct
• Inputs needed for final production of product
  (energy, water, etc.)
• Indirect
• ALL inputs needed in supply chain
• e.g. Metal, belts, wiring for engine
• e.g. Copper, plastic to produce wires
• Calculation yields every input needed

21
EIO-LCA Implementation

• Use the 491 x 491 input-output matrix of the U.S.
  economy from 1997
• Augment with sector-level environmental impact
  coefficient matrices (R) effect/ output from
  sector
• Environmental impact calculation
• E RX RI - A-1 Y

22
In Class Exercise

• Two Sector Economy.
• Model Final Demand 100 for Sector 1.
• Haz Waste of 50 gm/ in Sector 1 and 5 gm/ in
  Sector 2.
• Transaction Flows ( billion) are

1 2 Final Dmd.
1 150 500 350
2 200 100 1700
V.A. 650 1400 1100
23
Solution

• Requirements Matrix Row 1 0.15 and 0.25, Row
  2 0.2 and 0.05
• (I-A) inverse Matrix Row 1 1.2541 and 0.33, Row
  2 0.264 and 1.1221
  
• Direct intermediate inputs 15 of 1 and 20 of 2
• Total Outputs 125.4 of 1 and 26.4 of 2
• Hazardous Waste 6402 gm.