Comparing the Impact of Energy Model Results on Life Cycle Energy: Focus on High Performance Residential Building

Christi Saunders University of Pittsburgh, csaunders@mascaroconstruction.comAmy E. Landis Arizona State University, amy.landis@asu.eduLaura A. Schaefer University of Pittsburgh, las149@pitt.eduAlex K. Jones University of Pittsburgh, akjones@pitt.eduMelissa M. Bilec University of Pittsburgh, mbilec@pitt.edu

Abstract. Studies have shown that design-phase energy models, such as those used in Leadership in Energy and Environmental Design (LEED) calculations, have significant error rates, sometimes as much as 50%. These same energy model results are also being utilized in the field of life cycle assessment (LCA), where energy model results are being input for electricity and energy inventory calculations for the entire life cycle of the building. Error rates in energy modeling results have been well documented; however, research is lacking on the effect of this uncertainty within LCA, specifically life cycle energy. This research analyzes energy modeling results in terms of building life cycle energy use and metered energy data with a case study of a Solar Decathlon House. The life cycle assessment results indicated that the impact of energy model results is dependent on the impact category. Life cycle energy, however, appeared to be dependent on the electricity, resulting in an average error rate of about 44%. The sensitivity study analyzed these discrepancies and produced results that reduced the life cycle energy error rate to 26%.

Proceedings of the International Symposium on Sustainable Systems and Technologies (ISSN 2329-9169) is published annually by the Sustainable Conoscente Network. Melissa Bilec and Jun-ki Choi, co-editors. ISSSTNetwork@gmail.com.

Copyright 2013 by Christi Saunders, Amy E. Landis, Laura A. Shaefer, Alex K. Jones, Melissa M. Bilec. Licensed under CC-BY 3.0.

Introduction. Green building rating systems, such as Leadership in Energy and Environmental Design (LEED), heavily rely on the results from design-phase energy models to determine energy efficient building. LEED has been scrutinized for its reliance on energy models due to their well-documented error rates [1, 2]; some studies show design-phase energy model error rates around 50% [3-5]. These same design-phase energy models that are used in LEED calculations have also been utilized in life cycle assessments (LCAs). Error rates in energy modeling results have been well documented [2-5]; however, research is lacking on the effect of this uncertainty within LCA, specifically life cycle energy. Since operating energy has been modeled as about 80-90% of the life cycle energy in conventional buildings and about 50% in low energy buildings, the impact of modeling error rates on life cycle energy could (or could not be) significant [6-13]. LCA, formalized by the International Organization for Standardization (ISO) 14040, is a systematic method to quantify the life cycle energy and environmental impact of buildings, products, and activities [14]. LCA comprises the following four steps: goal and scope definition, life cycle inventory (LCI), life cycle impact assessment (LCIA), and interpretation. The goal and scope involves the delineation of the objectives of the LCA, as well as, functional unit and

Cite As:Comparing the Impact of Energy Model Results on Life Cycle Energy: Focus on High Performance Residential Building. Proc. ISSST, Christi Saunders, Amy E. Landis, Laura A. Shaefer, Alex K. Jones, Melissa M. Bilec. http://dx.doi.org/10.6084/m9.figshare.976932. (v1, 2013)

system boundary definition. For building related LCA, LCI entails the collection of all building data, such as quantities of materials and predicted energy use, and also produces data in the form of raw material usage and emissions to air and water. LCIA then aggregates this data into environmental loads, such as global warming potential and eutrophication. Overall, life cycle analyses produce results of emissions, resource depletion, and energy usage from a building in order to identify areas of improvement. A growing body of literature exists. Research has been published that utilizes LCA as a method to quantify life cycle energy and environmental impacts of whole buildings [6, 9, 10, 15-20].

Research Objectives. LCA and energy modeling results can depend on each other to predict life cycle energy and environmental impacts. The main question guiding this research was: from a life cycle perspective, how does energy models prediction impact life cycle energy use?

This research analyzes energy modeling results in terms of building life cycle energy use and metered energy data with a case study of a Solar Decathlon House. LCA and various energy modeling techniques were utilized as methods to evaluate the differences between predicted data and metered data. The objectives of this research were to:

Model life cycle energy use and environmental impacts

Evaluate the difference between predicted and metered life cycle energy use

Perform a sensitivity analysis to analyze the impact of energy model input variables

Case Study Description. The 2005 Solar House is a low energy home that was designed for the Solar Decathlon competition, which is an international competition held by the U.S. DOE to encourage affordable residential solar energy [21]. The house was originally designed as net zero by Pittsburgh Synergy, a group of students from Carnegie Mellon University (CMU), University of Pittsburgh, and the Art Institute of Pittsburgh. Today, the Solar House is utilized as office space on the campus of CMU. The house, shown in Figure 1, is about 79 m2 (850 ft2). Three different design-phase energy models (Autodesk Green Building Studio (GBS), Energy-10, EnergyPlus) were developed for the Solar House. Saunders, et al. presented a detailed description of the case study and the energy models [22].

Figure 1: Perspective of the 2005 Pittsburgh Synergy Solar House. This building model was built in Autodesk Revit Architecture 2011 for energy modeling purposes and therefore excludes the structural system and foundations.

Investigative Method. An LCA was completed of the Solar House. The system boundary was from raw materials extraction through the use phase, excluding the construction and end of life (Figure 2) [16, 17]. Similar to previous research, the life cycle of the Solar House was defined as a 25 years and the functional unit was one house at 79 m2 or 850 ft2 [23].

Data was collected for the life cycle assessment through a variety of means. Construction documents and manufacturer specifications, as well as field measurements, were used to derive the quantities of materials. Previous literature was utilized to calculate the different materials within heating, ventilation, and air conditioning (HVAC) systems [24]. The weights for HVAC materials reported by Shah et al. were scaled to match the total weight specified by the manufacturer [24]. Electricity usage predicted by all three models, as well as the metered data, which is detailed in Table 1, provided the inputs for the Electricity (Local Mix) process.

USE PHASE

CONSTRUCTION

AND ASSEMBLY

DISPOSAL

STRUCTURE ENVELOPE

ELECTRICITY HVAC SYSTEM

REPLACEMENT

LANDFILL

RECYCLE

MAINTENANCE

MEP

SYSTEMS

RAW

MATERIALS

EXTRACTION

MATERIALS

MANUFACTURE

Legend:

Unit Processes

System Boundary

Figure 2: System Boundaries for Solar House Life Cycle Assessment. The construction phase was omitted from the analysis due to the unconventional nature in which the building was built. End of life was also neglected due to

uncertainty in the disposal of the building.

Due to the significant original average error rate (59%) of the energy model results, a sensitivity analysis was performed. The sensitivity analysis presented reduced the average model error rate from 59% to 34%. This data was used to generate the LCI. The following are issues that were addressed within the sensitivity analysis:

PV efficiency:o Altered from 13% to 9%

Consistent occupancy schedule:o Workday was adjusted to 8 am to 6 pm

Consistent and adjusted lighting and plug loadso Metered lighting load calculated at 0.62 W/m2 (0.058 W/ft2)o Metered plug load calculated at 4.3 W/m2 (0.4 W/ft2)

Consistent and Adjusted Setpoints:o Winter setpoint was adjusted to 24C (75F)o Summer setpoint was adjusted to 26.7C (80F)

Table 1: Predicted and Metered Annual Electricity Usage. For a detailed analysis of the energy models and the inputs for the original models as well as the sensitivity analysis, see [22].

LCA was performed using primarily process related data. The LCI was developed using several different databases. USLCI and Franklin 98 (US databases) were selected first; then Ecoinvent, ETU-ESU, and IDEMAT (European databases) were used [25]. Table 2 details each material and process used in the LCI and the associated databases. The HVAC systems (the water heater, ERV, and heat pump) were modeled both in the materials phase of the life cycle as well as the use phase. In addition to the information listed in Table 2, some LCI data when necessary was acquired using Economic Input-Output LCA (EIO-LCA).

LCIA was completed using IMPACT 2002+ [26] . Non-renewable energy (MJ primary) was used to calculate life cycle energy. IMPACT 2002+ utilizes the cumulative energy demand (CED) method from ecoinvent to calculate non-renewable energy use [27]. The resources included within non-renewable energy are: hard coal, lignite, crude oil, natural gas, coal mining off-gas, peat, uranium, wood, and biomass from primary forests. The upper heating value of these resources is used to determine the characterization factor [26, 27]. A complete inventory of the upper heating values used in the CED method has been documented by the Swiss Center for Life Cycle Inventories [27, 28]. Three impact categories in IMPACT 2002+, ionizing radiation, land occupation, and mineral extraction, were omitted from this analysis due to data scarcity among unit processes.

Green Building Studio Energy-10 EnergyPlus Metered

Original Predictions

8511 kWh 687 kWh 5180 kWh 11788 kWh

Sensitivity Analysis

8684 kWh 4389 kWh 10250 kWh

Table 2: Life Cycle Assessment - Modeled Processes and Materials in the Solar House. All materials and processes were utilized to perform the process LCI. Square footage, densities, and weights were determined through construction documents, field measurements, and manufacturer specifications. *Refer to Table 1.

MATERIALS

Material Database Area (ft2) Weight (lb)

Manuf. Oriented Strand Board (OSB) USLCI 5782

Waste Oriented Strand Board (OSB) USLCI 729

Manuf. Expanded Polystyrene (EPS) Ecoinvent 15712

Waste Expanded Polystyrene (EPS) Ecoinvent 178

Cross-linked Polyethylene Ecoinvent 18.6

Red Maple Siding (Softwood) Ecoinvent 960 2622

Cypress Plywood Ceiling Ecoinvent 411 544

Veneer Lumber USLCI 2834

White Oak Floor (Hardwood) Ecoinvent 331

Low-E Glass Ecoinvent 196 314

Aluminum Window Frame Ecoinvent 75.8 133

Manufactured Polycarbonate Ecoinvent 596 18455

Wasted Poylcarbonate Ecoinvent 30.4 941

20% Fly Ash Concrete Ecoinvent, 42918

Total Steel (Hot rolled, Low alloy, EAC) Ecoinvent 6143

#4 Rebar (Reinforcing Steel) Ecoinvent 106

PV Panel (Monocrystalline Cells) Ecoinvent 406

Aluminum (Secondary) USLCI 92.1

Inverter Ecoinvent

Solar Tubes Ecoinvent 476

Interior Door (Wood) Ecoinvent 65.3

Exterior Door (Aluminum) Ecoinvent 131 Tankless Water Heater

Nylon 66 Ecoinvent 1.56 21.3

Polyester (Thermoplast) Ecoinvent 3.13 1.43

Copper (Secondary) Ecoinvent 0.03 0.09

Steel (Cold Rolled, BOF) Franklin 98 1.56 1.97

Ventilation Equipment Assembly Ecoinvent

Energy Recovery Ventilator

Galvanized Steel Sheet Metal USLCI 65.2

Expanded Polystyrene (EPS) Ecoinvent 0.04

Polyester (Fabric) IDEMAT 1.41

Polyvinylchloride (Suspension) Ecoinvent 0.95

Copper (Secondary) Ecoinvent 3.13

Aluminum (Secondary) USLCI 11.9

Synthetic Rubber Ecoinvent 0.55

Corrugated Cardboard Franklin 98 4.78

Packaging Paper Franklin 98 0.24

Ventilation Equipment Assembly Ecoinvent

Heat Pump

Galvanized Steel Sheet Metal USLCI 26.0

Steel (Cold Rolled, BOF) Franklin 98 90.1

Polyvinylchloride (Suspension) Ecoinvent 2.2

Copper (Secondary) Ecoinvent 14.7

Aluminum (Secondary) USLCI 12.5

Nylon 66 Ecoinvent 0.28

Brass Ecoinvent 1.38

Refrigerant (R134-A) ETH-ESU 6.61

Ventilation Equipment Assembly Ecoinvent

OPERATIONS AND MAINTENANCE

Process Database Power (kWh) Total (25 yrs)

Electricity (Local Mix) USLCI Variable* Variable

Results. LCA results for the Solar House are detailed in Figure 3. The results from each energy model were utilized to develop comparative LCA results, which were classified into the appropriate materials and use phases and then were normalized to the metered results. As illustrated in Figure 3, the disparity between the predicted LCA results and the metered results depends on the impact category. In the categories of carcinogens, non-carcinogens, ozone layer depletion, and terrestrial ecotoxicity and acidification, electricity usage does not appear to have a substantial impact on the life cycle impact results. This outcome could be due to higher impacts in those categories from other materials or activities included in the Solar House LCI. Plastics/glass and appliances/lighting consist of a large percentage of the impact in those categories. In categories of respiratory organics and inorganics, aquatic acidification, global warming, and primary energy use, the LCA results are largely dependent on the energy results. In terms of life cycle energy, the metered data indicated that the operations phase accounted for about 75% of the primary energy use, whereas GBS, Energy-10, and EnergyPlus predicted it to be about 70%, 30%, and 60% respectively. Depending on the impact category, LCA results could or could not be largely affected by predicted versus metered electricity usage.

Respiratory organics

Aquatic ecotoxicity

Terrestrial ecotoxicity

Terrestrial acid/nutri

Aquatic acidification

Aquatic eutrophication

Global warming

Non-renewable energy

0

10

20

30

40

50

60

70

80

90

100

Carc

ino

ge

ns

No

n-c

arc

ino

gen

s

Res

pir

ato

ry in

org

an

ics

Ozo

ne

la

ye

r d

ep

leti

on

Res

pir

ato

ry o

rgan

ics

Aq

ua

tic e

co

toxic

ity

Terr

es

tria

l e

co

toxic

ity

Te

rre

str

ial A

cid

ific

ati

on

/

Aq

ua

tic a

cid

ific

ati

on

Aq

uati

c e

utr

op

hic

ati

on

Glo

ba

l w

arm

ing

Pri

mary

En

erg

y

No

rmali

zed

Perc

en

tag

e (

%)

Green Building Studio O&M Energy-10 O&M EnergyPlus O&M Metered* O&M

Materials

Nu

trif

icati

on

Life Cycle Assessment Results for the Solar House for a 25-year Life Cycle. Results from each energy model and the metered data have been used to develop four different LCA results. The results are normalized to the highest metered impact in each respective category. O&M = Operations and Maintenance (use); M&C = Materials and Construction. *Metered data is 18 months of measured energy averaged for one year and then projected for 24 years.

A further analysis of the life cycle energy use of the Solar House is presented in Figure 4. The life cycle energy calculations include both the embodied energy in the materials of the Solar House and the operating energy required for 25 years. The metered data produced a life cycle energy consumption of 5.77 TJ. GBS, Energy-10, and EnergyPlus predicted a life cycle energy use of 4.58, 1.76, and 3.38 TJ, respectively, consequently, producing error rates of 20%, 70%, and 41%, respectively. The average error rate for life cycle energy use, 44%, is substantial. Since low energy buildings seem to have higher energy modeling error rates than other buildings, these results could be different for a conventional building. Overall, the results of energy models have a considerable impact on life cycle energy calculations.

0 1 2 3 4 5 6 7

Green

Building

Energy-10

EnergyPlus

Metered*

Primary Energy (TJ)

Materials Operations and Maintenance

20%

Error

70%

Error

41%

Error

4.58

TJ

1.76

TJ

3.38

TJ

5.77

TJ

Green

Building

Studio

Figure 4: 25-year Life Cycle Energy Use of the Solar House. Life cycle energy results are illustrated for each of the energy models and the metered data. Life cycle energy includes embodied energy and operating energy. *Metered data is 18 months of measured energy data averaged for one year and then projected for 24 years.

Sensitivity Analysis The sensitivity results from the energy models were input into the life cycle assessment to generate adjusted life cycle energy results. Figure 5 translates the results of the sensitivity analysis into life cycle energy. Since the net usage results for GBS were fairly similar to the original predictions, its life cycle energy results remained analogous. In accordance with the energy model results, the life cycle energy results greatly increased for EnergyPlus and Energy-10. The sensitivity analysis produced life cycle energy results that decreased the average errorrate from 44% to 26%.

0 1 2 3 4 5 6 7

Green

Building

Energy-10

EnergyPlus

Metered*

Primary Energy (TJ)

Materials Operations and Maintenance

20%

Error

47%

Error

10%

Error

4.6

TJ

3.05

TJ

5.16

TJ

5.77

TJ

Green

Building

Studio

Original Predictions

Figure 5: Sensitivity Analysis Results of the Life Cycle Energy of the Solar House. The adjusted results from the energy models were utilized to generate sensitivity results for life cycle energy. *Metered data is 18 months of

measured energy data averaged for one year and then projected for 24 years.

Conclusions. This research utilized several different energy modeling programs and life cycle assessment as methods to analyze the impact of energy modeling results on life cycle energy. The life cycle assessment results indicated that the impact of energy model results is highly dependent on the impact category. Several categories such as ozone depletion only slightly varied between each energy model program and the metered data. Life cycle energy, however, appeared to be highly dependent on the electricity unit process, resulting in an average error rate of about 44%. Several variables exist between predicted energy usage and metered usage, including occupancy densities, activities within the building, and the efficiency of systems, which could cause these substantial error rates. The sensitivity study analyzed these discrepancies and produced results that reduced the life cycle energy error rate to 26%. The accuracy of the energy models highly depends on the inputs and their reflection of the actual systems and activities within the building.

The variability of energy model results has a substantial impact within the building industry. LEED buildings have lost some credibility in terms of energy efficiency partly due to their reliance on model results [1, 2]. The prediction of the overall environmental impact of a building can also rely on energy model results to determine operational energy usage. Error rates of predictive energy models can be a considerable variant to an LCA and could be considered as a part of uncertainty within LCA. In order to mitigate these issues with design-phase energy models, buildings can sub-meter energy usage in order to better determine the differences between the predicted and metered usage. The tracked energy usage could then be used to continually update the LCA to produce more accurate life cycle energy calculations.

Acknowledgements. This work was supported by National Science Foundation under EFRI-SEED Grant #1038139 and the Mascaro Center for Sustainable Innovation at the University of Pittsburgh. The authors would like to acknowledge Carnegie Mellon University for their continued support and assistance in this research. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

References 1. Navarro, M., Some Buildings Not Living Up to Green Label, in New York Times. 2009: New

York, NY.2. Turner, C. and M. Frankel, Energy Performance of LEED for New Construction Buildings.

2008, New Buildings Institute: Vancouver, WA.3. Karlsson, F., P. Rohdin, and M.-L. Persson, Measured and predicted energy demand of a

low energy building: important aspects when using Building Energy Simulation. BuildingServices Engineering Research and Technology, 2007. 28(3): p. 223-235.

4. Neto, A.H. and F.A.S. Fiorelli, Comparison between detailed model simulation and artificialneural network for forecasting building energy consumption. Energy and Buildings, 2008.40(12): p. 2169-2176.

5. Daly, A., E. Franconi, and L. Morrison. Improving Modeling Credibility. in Greenbuild NEXT.2011. Toronto, ON.

6. Scheuer, C., G. Keoleian, and P. Reppe, Life cycle energy and environmental performanceof a new university building: modeling challenges and design implications. Energy andBuildings, 2003. 35: p. 1049-1064.

7. Ramesh, T., R. Prakash, and K.K. Shukla, Life Cycle Energy Analysis of Buildings: AnOverview. Energy and Buildings, 2010. 42(10): p. 1592-1600.

8. Cole, R.J. and P.C. Kernan, Life-Cycle Energy Use in Office Buildings. Building andEnvironment, 1996. 31(4): p. 307-317.

9. Keoleian, G.A., S. Blanchard, and P. Reppe, Life-Cycle Energy, Costs, and Strategies forImproving a Single-Family House. Journal of Industrial Ecology, 2000. 4(2): p. 135-156.

10. Ochoa, L., C. Hendrickson, and H.S. Matthews, Economic Input-output Life-cycleAssessment of U.S. Residential Buildings. Journal of Infrastructure Systems, 2002. 8(4): p.132-138.

11. Paulsen, J.H. and M. Borg, A Building Sector Related Procedure to Assess the Relevance ofthe Usage Phase. International Journal of Life Cycle Assessment, 2003. 8(3): p. 142-150.

12. Blengini, G.A. and T.D. Carlo, The changing role of life cycle phases, subsystems andmaterials in the LCA of low energy buildings. Energy and Buildings, 2010. 42(6): p. 869-880.

13. Sartori, I. and A.G. Hestnes, Energy use in the life cycle of conventional and low-energybuildings: A review article. Energy and Buildings, 2007. 39(3): p. 247-257.

14. International Organization for Standardization, Environmental management-Life cycleassessment-Principals and framework, in ISO 14040:2006. 2006: Geneva, Switzerland.

15. Sharrard, A.L., H.S. Matthews, and R.J. Ries, Estimating Construction Project EnvironmentalEffects Using an Input-Output-Based Hybrid Life-Cycle Assessment Model. Journal ofInfrastructure Systems, 2008. 14(4): p. 327-336.

16. Junnila, S. and A. Horvath, Life-Cycle Environmental Effects of an Office Building. Journal ofInfrastructure Systems, 2003. 9(4): p. 157-166.

17. Junnila, S., A. Horvath, and A.A. Guggemos, Life-Cycle Assessment of Office Buildings inEurope and the United States. Journal of Infrastructure Systems, 2006. 12(1): p. 10-17.

18. Ochoa, L., et al. Life Cycle Assessment of Residential Buildings. in Construction ResearchCongress 2005: Broadening Perspectives. 2005. San Diego, California.

19. Bilec, M., et al., Example of a Hybrid Life-Cycle Assessment of Construction Processes.Journal of Infrastructure Systems, 2006: p. 207-215.

20. Bilec, M.M., R.J. Ries, and H.S. Matthews, Life-Cycle Assessment Modeling of ConstructionProcesses for Buildings. Journal of Infrastructure Systems, 2010. 16(3): p. 199-205.

21. U.S. Department of Energy. Solar Decathlon. 2010 [cited 2012 3/28]; Available from:http://www.solardecathlon.gov/.

22. Saunders, C., et al. Utilizing Measured Energy Usage to Analyze Design Phase EnergyModels. in 2012 IEEE International Symposium on Sustainable Systems and Technology.2012. Boston: Institute of Electrical and Electronics Engineers ( IEEE ).

23. Wang, N., et al., A marketable all-electric solar house: A report of a Solar Decathlon project.Renewable Energy, 2009. 34: p. 2860-2871.

24. Shah, V.P., D.C. Debella, and R.J. Ries, Life cycle assessment of residential heating andcooling systems in four regions in the United States. Energy and Buildings, 2008. 40(4): p.503-513.

25. Product Ecology Constultants, SimaPro. 2011.26. Jolliet, O., et al., IMPACT 2002+: A new life cycle impact assessment methodology. The

International Journal of Life Cycle Assessment, 2003. 8(6): p. 324-330.27. Frischknecht, R. and N. Jungbluth, Implementation of Life Cycle Impact Assessment

Methods. 2010, Swiss Centre for Life Cycle Inventories: Dubendorf28. Dones, R., et al., Life Cycle Inventories of Energy Systems: Results for Current Systems in

Switzerland and other UCTE Countries. 2007, Swiss Centre for Life Cycle Inventories:Dubendorf.