science behind ornl’s building technology research...

50
Science Behind ORNL’s Building Technology Research Integration Center (BTRIC) Joshua New, Ph.D. Building Technologies Research & Integration Center (BTRIC) Whole Building and Community Integration Group Overview of BTRIC Visual Analytics and Computational Efforts

Upload: others

Post on 08-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Science Behind

ORNL’s Building

Technology Research

Integration Center

(BTRIC)

Joshua New, Ph.D.

Building Technologies Research & Integration Center (BTRIC)

Whole Building and Community Integration Group

Overview of BTRIC Visual Analytics and Computational Efforts

Page 2: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

2 Green Economy 1302

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 3: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

3 Green Economy 1302

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 4: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

4th

Paradigm – The Science behind the Science

• Empirical – guided by experiment/observation

– In use thousands of years ago, natural phenomena

• Theoretical – based on coherent group of principles and theorems

– In use hundreds of years ago, generalizations

• Computational – simulating complex phenomena

– In use for decades

• Data exploration (eScience) – unifies all 3

– Data capture, curation, storage, analysis, and visualization

4

Tycho Brahe

Johannes Kepler

Page 5: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

4th

Paradigm

5

Page 6: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 7: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

COMPUTER TOOL FOR SIMULATING

COOL ROOFS

INDUSTRY

COLLABORATIVE R&D

Marc LaFrance

DOE BT

R. Levinson,

H. Gilbert,

H. Akbari

Chris Scruton CEC

A. Desjarlais,

W. Miller,

J. New WBT

Joe Huang,

Ender Erdem

Roof Savings Calculator (RSC)

Page 8: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

8

Roof Savings Calculator

Replaces:

EPA Roof Comparison Calc

DOE Cool Roof Calculator

Minimal questions (<20)

Only location is required

Building America defaults

Help links for unknown information

Page 9: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

RSC = AtticSim + DOE-2.1E

AtticSim - ASTM C 1340 Standard For Estimating Heat Gain or Loss Through Ceilings Under Attics

Page 10: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Summer Operation of HVAC Duct in

ASHRAE Climate Zone 3

Page 11: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

11

Roof Savings Calculator

• Building Details

• HVAC efficiency and utility prices

• Roof and Attic Information (base vs. comp)

• Reports energy and cost savings

DOE-2.1E+AtticSim

Page 12: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Office “Big Box” Retail Warehouse

Commercial building types

Torcellini et al. 2008, “DOE Commercial Building Benchmark Models”,

NREL/CP-550-43291, National Renewable Energy Laboratory, Golden CO.

Page 13: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

AtticSim

DOE-2

Page 14: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

14

RoofCalc.com Impact

Average: ~100 visitors/day

24,100 web simulations, 156 users/feedback, 3+ million runs

Page 15: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Enhanced RSC Site

Input Parameter GUI Result Output

Database

User Hyperion RSC Engine Inputs

Savings

Simulate

Savings

Exists?

Resu

lts

Simulation

Page 16: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Testing RSC – Python Robot Framework

Page 17: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Current Results

Description Reflectance Emissivity SRI Atlanta Austin Baltimore Chicago Fairbanks Fargo Houston Kansas City Los

Angeles Miami Minneapolis New York Phoenix San

Francisco

BUR No Coating 10 90 6 -54 0 -66 -36 -125 -99 42 -47 98 75 -53 -89 39 -68

Mineral Mod Bit 25 88 25 -422 -39 -507 -325 -941 -659 103 -368 383 276 -419 -669 70 -420

Single Ply 32 90 35 -384 71 -437 -253 -901 -660 230 -320 614 441 -382 -582 154 -494

Mineral Mod Bit 33 92 35 -574 3 -655 -407 -1302 -908 197 -477 648 463 -560 -871 118 -659

Metal 35 82 35 -883 -191 -1000 -742 -2213 -1296 60 -698 293 212 -863 -1558 74 -322

Aluminum Coating over BUR 43 58 35 -9 189 -64 -46 -237 -298 279 -45 585 372 -93 -189 294 -58

Mineral Mod Bit 45 79 55 -564 84 -657 -408 -1385 -1003 291 -475 872 594 -582 -907 216 -693

Coating over BUR 49 83 55 -413 231 -461 -250 -1154 -872 433 -345 1075 742 -441 -680 348 -640

Metal 49 83 55 -1191 -126 -1231 -837 -2855 -1697 208 -857 771 576 -1102 -1891 138 -957

Aluminum Coating over BUR 55 45 48 39 174 -35 -29 -276 -367 390 -21 825 502 -90 -202 419 -51

Mineral Mod Bit 63 88 75 -909 203 -996 -571 -2372 -1661 525 -726 1473 1105 -933 -1380 300 -1419

Coating over BUR 63 86 75 -606 334 -664 -347 -1787 -1305 607 -501 1512 1102 -659 -980 452 -1104

Metal 63 84 75 -1487 -31 -1465 -919 -3600 -2151 361 -1028 1295 986 -1356 -2198 171 -1704

Single Ply 64 80 75 -637 304 -712 -386 -1850 -1345 578 -528 1480 1067 -694 -1031 408 -1105

Aluminum Coating over BUR 65 45 65 -80 272 -160 -88 -696 -655 542 -123 1230 758 -227 -399 558 -301

Metal (White) 70 85 85 -1622 14 -1592 -967 -4005 -2422 436 -1133 1522 1211 -1502 -2353 166 -2131

Coating over BUR (White) 75 90 93 -770 417 -875 -443 -2391 -1732 767 -664 1822 1460 -900 -1261 526 -1642

Single Ply (White) 76 87 94 -840 384 -962 -502 -2547 -1829 745 -722 1808 1460 -974 -1358 471 -1720

Coating over BUR (White) 79 90 100 -812 450 -928 -471 -2571 -1862 820 -710 1906 1576 -974 -1336 553 -1825

Mineral Mod Bit (White) 81 80 100 -1025 355 -1161 -642 -3006 -2131 748 -867 1876 1556 -1175 -1634 444 -2057

Single Ply (White) 82 79 100 -819 455 -949 -494 -2643 -1912 822 -722 1934 1578 -1002 -1373 554 -1847

Coating over BUR (White) 85 90 107 -873 499 -1008 -524 -2845 -2073 905 -782 2003 1761 -1097 -1454 592 -2123

Single Ply (White) 85 87 107 -936 459 -1083 -577 -2969 -2143 871 -830 1974 1736 -1156 -1536 531 -2167

Page 18: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

RSC Web Service

• SoapResults = simulate(SoapModel)

– Accepts a model and returns the RSC results

• ZipString = test(SoapModel)

– Forces the model to be evaluated by the engine (rather than checking the database) and returns a zip (as a base64-encoded string) of the DOE2/AtticSim output files

• ScenarioID = upload(SoapModel, SoapResults)

– Uploads the model and results to the database, bypassing the engine

• (SoapModel, SoapResults) = download(ScenarioID, VersionNumber)

– Downloads a model/result pair for the scenario ID and version number

Page 19: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

RSC Service Example (Python)

client = suds.client.Client('URL/TO/WEB/SERVICE/rsc.wsdl')

print(client)

sm = client.factory.create('schema:soapmodel')

load_soap_model_from_xml('../examplemodel.xml', sm)

sr = client.service.simulate(sm)

print(sr)

sm = client.factory.create('schema:soapmodel')

load_soap_model_from_xml('../examplemodel.xml', sm)

print(sm)

contents = client.service.test(sm)

with open('pytest.zip', 'wb') as outfile:

outfile.write(base64.b64decode(contents))

sm = client.factory.create('schema:soapmodel')

load_soap_model_from_xml('../examplemodel.xml', sm)

sr = client.factory.create('schema:soapresults')

load_soap_results_from_xml('../exampleresults.xml', sr)

sid = client.service.upload(sm, sr)

print(sid)

modres = client.service.download(83356208, '0.9')

print(modres['soapmodel'])

print(modres['soapresults'])

Page 20: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Millions of simulations visualized for DOE’s Roof

Savings Calculator and deployment of roof and attic

technologies through leading industry partners

Leveraging HPC resources to facilitate deployment of building energy efficiency technologies

DOE: Office of Science CEC & DOE EERE: BTO Industry & Building Owners

Roof Savings Calculator (RSC) web

site/service developed and validated

[estimates energy cost savings of

improvements to flat or sloped roofs for

any existing condition or climate]

CentiMark, the largest nation-wide

roofing contractor (installs 2500

roofs/mo), is integrating RSC into

their proposal generating system

(others expected to follow)

AtticSim

DOE-2

Engine (AtticSim/DOE-2) debugged

using HPC Science assets enabling

visual analytics on 3x(10)6 simulations

Page 21: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 22: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

22

Current Projects

• UC-Berkeley – testing, regression (quick estimation, rules of thumb) [donated effort]

CITRIS, UC-Berkeley 96 ~ HP rx2600

RSC

Simulations

Testing

Analysis

Web Server PowerEdge R510

RoofCalc.com

Page 23: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

23

Visual Analytics (demo)

• Visualization techniques (for Energy Simulation)

– City-Scape, Artificial Terrain

Page 24: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Climate Zone Map

• Climate zones (1-8) shown on map.

Page 25: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

High-Density Time Plots

Context Focus

• Each line is the energy usage for a single simulation

• High Dynamic Range rendering (HDR)

• Apply logarithmic coloring scaling to emphasis high traffic regions

• Render outlier lines separately

Page 26: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Category View

Categorical Context Mouse Hover Highlight Categorical Focus

• Bars for each category show occupancy levels

• Grouped by dimension; highlighting & focus rendering

Foundation Type HVAC Vintage

Cra

wl S

pac

e (8

0%)

Sla

b (

37%

)

Bas

emen

t (1

9%)

Page 27: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Parallel Coordinates

• One parallel axis per data dimension; One line per data item crosses every axis

Min

Max

Max X

Y

Min

Max

X Y

Scatterplot vs. Parallel Coordinates

Page 28: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

PCP - car data set

2

Page 29: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

PCP Bin Rendering

• Transfer Function Coloring:

– Occupancy or Leading Axis

Page 30: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Bug Vis Old New

11 23 3 11

Page 31: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Outliers (Heating)

• Selection of heating outliers

• Find all have box building type and in Miami

Page 32: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 33: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Image Fusion

(based on cone-fusion of mammalian retina)

Typical MRI and SPECT imagery Colorfuse Image

Page 34: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Learning Associations

Full Results DetailResults

Page 35: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 36: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Source of Input Data

• 3 Campbell Creek homes

(TVA, ORNL, EPRI)

• 100+ sensors/home, 15-minute data:

• Temperature (inside/outside)

• Plugs

• Lights

• Range

• Washer

• Radiated heat

• Dryer

• Refrigerator

• Dishwasher

• Heat pump air flow

• Shower water flow

• Etc.

Page 37: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

List of Machine Learning

Techniques to Explore

• Linear Regression

• Feedforward Neural

Network

• Support Vector Machine

Regression

• Non-Linear Regression

• K-Means with Local Models

• Gaussian Mixture Model

with Local Models

• Acknowledgment: UTK computer science Ph.D. student Richard

Edwards is doing bulk of the work; student of Dr. Lynne Parker

• Self-Organizing Map with Local

Models

• Regression Tree (using Information

Gain)

• Time Modeling with Local Models

• Recurrent Neural Networks

• Neural Network with Genetic

Algorithm

• Ensemble Learning

Page 38: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Example Results

• Robust Linear Regression Model can map current

sensor observations to energy use

House 1 (House 2 is similar) House 3 – More difficult, due to

solar energy input

Page 39: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Example Results to Date (con’t.)

• Robust Linear Regression Model for predicting energy

usage 1 hour ahead:

House 2 (House 1 is similar) House 3

(all models are Markov Order 3)

Page 40: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Performance Metrics

Page 41: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Presentation summary

• Scientific Paradigms

• Roof Savings Calculator

• Visual Analytics

• Machine Learning

• Prediction of Electrical Consumption

• Autotune

Page 42: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

The Autotune Idea

Making building energy models more useful by calibrating them to data

.

.

.

E+ Input

Model

Page 43: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Goal: Reduce Project Development Costs for

Small Building Retrofit Projects

• High performance computing applied to task of auto-tuning building energy models – Jaguar, Nautilus & Frost supercomputers all engaged (32k E+ sims in <5 mins!)

– ORNL, U of Tennessee-Knoxville, Jacksonville State U

Handful of Data Channels & Weather

Page 44: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Computational Complexity

E+ Input

Model

Problems/Opportunities:

Thousands of parameters per E+ input file

We chose to vary 156

Brute-force = 5x1052 simulations

main_Tot None_Tot(

1) None_Tot(

2) HP1_in_To

t HP1_out_

Tot HP1_back

_Tot HP1_in_fa

n_Tot HP1_comp

_Tot HP2_in_To

t HP2_out_

Tot HP2_back

_Tot HP2_in_fa

n_Tot 1172.5 0 0 6.75 18.75 0 0 0 6.75 18 0 0

E+ parameters

The Universe:

13.75 billion years?

Need 4.1x1028 of those

Page 45: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

ORNL High Performance Computing Resources

Multi-million dollar cost share and infrastructure on 6 supercomputers including the world’s fastest Currently use 128,000+ cores to run over 530,000 EnergyPlus simulations and write 45TB of data in 68 minutes

Jaguar: 224k cores, 360TB memory, 10PB of disk, 1.7 petaflops Cost: $104 million DOE BTO: 500k hours granted (CY12)

Nautilus: 1024 cores, shared-memory

DOE BTO: 30k hours granted (CY11) 200k hours granted (CY12) 150k hours (CY13)

Frost: 2048 SGI Altix; 136 nodes 200k hours granted (CY13)

Lens cluster: 77 nodes – 45x128GB, 32x 64GB with NVIDIA 880 and Tesla dual-GPU EVEREST visualization (CY13)

Gordon (12,608 cores): 250k hours (CY13)

Kraken (112,896 cores): 100k hours (CY13)

Page 46: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Titan fully utilized

Page 47: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

On-deck Circle

Combining a different way…

60

62

64

66

68

70

72

74

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47

Fo

ur-

day

SA

E

Generation

Trial 1

Trial 2

Trial 3

Trial 4

Trial 5

Trial 6

Trial 7

Trial 8

25%

Page 48: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Jibo

Sanyal

Mahabir

Bhandari Som Shrestha Joshua New Aaron

Garrett

Buzz

Karpay

Richard

Edwards

The Autotune Team

http://autotune.roofcalc.com

Page 49: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Autotune calibration of building energy models

MLSuite - HPC-enabled suite of 12+ machine

learning algorithms for large data mining

ASHRAE G14

Requires

Autotune

Results

Using Monthly

utility data

CV(RMSE) 30% 0.318%

NMBE 10% 0.059%

Using Hourly

utility data

CV(RMSE) 15% 0.483%

NMBE 5% 0.067%

Autotune could have saved 2+ man-months of

effort (over 2 calendar years) modeling 1 field

demonstration building

Within 30¢/day

(actual use

$4.97/day)

Residential Commercial

Hourly – 8%

Monthly – 15%

Average error of

each input

parameter

Page 50: Science Behind ORNL’s Building Technology Research ...web.eecs.utk.edu/~jnew1/presentations/2013_RCMA_RSC.pdf · Description Reflectance Emissivity SRI Atlanta Austin Baltimore

Discussion