Quick-but-Clean? A Screening LCA Tool Using PERT Estimates to Incorporate Uncertainty Debbie Kalish Program Manager, Ingersoll Rand Center for Energy Efficiency & Sustainability
Christoph Koffler, PhD Technical Director PE International, Inc.
Ingersoll Rand Overview • A $14 billion diversified industrial company • Publicly-held; NYSE:IR • More than 52,000 employees worldwide in 54 countries • Operations in every major geographic region • Strategic brands are #1 or #2 in their markets • Products and services for commercial, industrial and residential markets
Advancing the Quality of Life We advance the quality of life by creating and sustaining safe, comfortable and efficient environments addressing the world’s growing critical needs for • Clean and comfortable air • Secure homes and buildings • Safe and fresh food
• Energy efficiency • Sustainable business practices
A World of Sustainable Progress and Enduring Results
Why Create a Screening LCA? • Desire life cycle thinking in our New Product Development (NPD) process
• We need a tool that can estimate environmental burden and perform scenario analyses that is: – Easy to use – Cost effective – Inputs can be gathered quickly
– Uses minimal resources – Is robust – Reproducible
Model structure & results Christoph Koffler, PhD – PE INTERNATIONAL
Quick-but-Clean? LCA model
Product
Number
Steel
Unalloyed
Non-ferrous
Stainless
Aluminum
Castings
Copper
Ferrous
Weight [lb]
Polymers
Non-ferrous
Blow molding [lb]
Bar [lb]
Misc [lb]
Bar [lb]
Fin Stock [lb]
Iron [lb]
Aluminum [lb]
Compressio n molding [lb]
Coil [lb]
Sheet [lb]
Extrusion [lb]
Industrial tube [lb]
Steel [lb]
Brass [lb]
Extrusion [lb]
Plate [lb]
Fin stock [lb]
Technical tube [lb]
Titanium [lb]
Sheet [lb]
Sheet [lb]
Brass extrusion [lb]
Zinc [lb]
Slit coil [lb]
Tube [lb]
Power storage
Chemicals
Operating materials
Paints & Coatings
Battery
Electronics
Battery [lb]
LCD [in2]
Use phase
Energy
Consumabl es
Electrodeposition [ft2]
Lead-acid
PWB [in2]
Diesel [gal]
Lubricant [lb]
CO [lb]
R134a
Powder coating [ft2]
Li-Ion
Wiring [lb]
Gasoline [gal]
Refrigerant [lb]
HC [lb]
Injection molding [lb]
R404a
Spray coating [ft2]
Rubber [lb]
R410a
Desiccants [lb]
Lubricants [lb]
Refrigerant
Refrigerant [lb]
Natural gas [mmBTU]
NO [lb]
Power [kWh]
NO2 [lb]
LPG [gal]
NOx [lb]
Structural [lb]
Tube [lb]
Direct emissions
PM unspec. [lb]
• 7 material groups based on IR BOM taxonomy & available LCI data
PM10 [lb]
PM2.5 [lb]
• Sum of material weights is scaled to match product weight • Use phase with energy & consumables consumption & direct emissions • Calculation of GWP, AP, EP, ODP, and SFP (TRACI 2.0)
AP42 Natural gas (2-stroke lean)
CNG (4stroke lean)
CNG (4stroke rich)
Gasoline (uncontroll ed industrial)
Diesel (uncontroll ed industrial)
LPG
Quick-but-Clean? ‘Castings / Non-ferrous / Aluminum’
Quick-but-Clean? Model parameterization parameter
best case
most likely
worst case
recycled content
100%
50%
0%
material loss during manufacturing
0.2%
6.5%
20%
2%
65%(1)
80%(2)
landfilled in EoL (1) McMillan et al. 2012 (2) EPA 2010
• Total of 81 parameters / 243 individual parameter values for model calibration • All parameter values reviewed with various Ingersoll Rand brands
• Data gaps filled with literature values
Quick-but-Clean? Calculation of results
Impact category indicator result
worst case
worst case (16.67%)
PERT estimate
most likely (66.67%)
best case
best case (16.67%)
?
Quick-but-Clean? The PERT distribution • beta distribution
• first developed by the U.S. Navy • min / max / most likely • stronger tendency towards the most likely value than the triangular distribution đ?’‡ đ?’™ =đ?‘˛âˆ— đ?’™âˆ’đ?’‚
đ?œś
∗
(đ?’ƒ − đ?’™)đ?œˇ
• PERT estimate = (best + 4 x most likely + worst) / 6
Quick-but-Clean? Cradle-to-grave GWP per unit of material class
Quick-but-Clean PE Global LCIA Expert Survey 2012 • Establish weighting factors for 13 impact categories • 660 LCA experts worldwide contacted via email
• 245 online questionnaires completed (37%) APAC RER RNA RAF RLA
Quick-but-Clean? Cradle-to-grave single score per unit of material class
Quick-but-Clean? Model validation: no use phase burden Screening tool significantly overshoots ďƒ check modeling choices & data
Steelcraft Steel Door
Quick-but-Clean? Model validation: moderate use phase burden Screening tool much less detailed than benchmark ďƒ underestimation
IR Cordless Assembly Screwdriver Benchmark based on GaBi DB 2006 ďƒ higher ODP values in electricity datasets
Quick-but-Clean? Model validation: high use phase burden Dominated by copper tubes ďƒ tool assumes lower recycling rate
Trane Chiller Use phase doesn’t add any significant ODP emissions
Quick-but-Clean? Conclusions Quick, …
• customized, intuitive, non-expert tool
• minimizes data collection & analysis costs … but clean!
• PERT single score is a robust estimate • use phase energy consumption offsets uncertainty from manufacturing • display of upper & lower limits improves decision making
Quick-but-Clean! Outlook • road test tool with various brands / product lines
• review and refine model parameterization
• establish it as core DfE tool in NPD process
To be continued… Deborah Kalish – Ingersoll Rand Christoph Koffler – PE INTERNATIONAL