HomeMy WebLinkAbout20210107Avista to Staff 25-29.pdfDemo of StampPDF by Appligent, Inc. http://www.appligent.comDemo of StampPDF by Appligent, Inc. http://www.appligent.com
AVISTA CORPORATION RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO DATE PREPARED: 01/06/2021 CASE NO: AVU-E-20-13/AVU-G-20-08 WITNESS: N/A REQUESTER: IPUC Staff RESPONDER: Randy Gnaedinger TYPE: Production Request DEPARTMENT: Transmission Services
REQUEST NO.: Staff – 25 TELEPHONE: (509) 495-2047 REQUEST:
On page 3 of its application in Case No. AVU-E-13-08, the Company defines R&D for the purposes of its proposal "to be applied R&D that could yield benefits to customers in the next one to four years." Please provide the following information: a. Examples of R&D projects that have yielded benefits to customers since the Commission authorized funding of R&D projects in Order No. 32918. For each project
cited, please include funding costs and the estimated benefits to customers. b. Any reports or cost/benefit analyses conducted by Avista regarding R&D projects since the Commission issued Order No. 32918 in 2013.
RESPONSE: a. During the Simulation-Based Commission of Energy Management Control System (2016), the research team identified a HVAC control error at the University of Idaho COBE building which resulted in an annual savings of 14 MWh or approximately $630-$650 annually based on the customer rates in effect during that time.
The actual budget for phase 1 of this project during the 2014-15 academic year was $26,696. A second phase was subsequently funded in 2016-17 academic year for $64,230.
RSVC for Near Customer Use
The Residential Static Var Compensator (RSVC) project with Boise State took place over a four-year period covering the 2014-18 academic years. The research resulted in developing intellectual property that enables a small voltage control device to be installed near the customer meter. Such a device would enable additional power savings through a highly
tuned use of conservation voltage reduction.
Along the way, this research funding also assisted in four advanced degrees. One of the graduate research students, Mr. A. Delgado is now a System Planning Engineer at Idaho Power.
The actual budget for the four-year period was:
• $56,823 for 2014-15
• $67,513 for 2015-16
• $95,935 for 2016-17
• $90,377 for 2017-18 IR Camera for HVAC Control The Infrared Red (IR) Camera project with the University of Idaho took place over a two-year period, from 2017-2019, developing initial energy analysis for Avista’s service
RECEIVED
2021 January 7, PM 12:23
IDAHO PUBLIC
UTILITIES COMMISSION
territory and identifying a commercial gap in the availability an off-the-shelf product. The research into such a product has resulted in still-pending intellectual property.
Along the way, this research funding also assisted in two advanced degrees, and provided an educational opportunity to Dr. Damon Woods who is a current professor for the University of Idaho at the Integrated Design Lab.
The actual budget for year 1 of this project during the 2017-18 academic year was $24,009. A second year was subsequently funded in the 2018-19 academic year for $48,678. b. See the attached report (Staff_PR_25 – Attachment A) from the research team quantifying the energy savings at the COBE building referenced in section a.
Realized Savings for COBE Building
Estimates prepared by the University of Idaho’s Integrated Design Lab on behalf of AVISTA Utilities.
The College of Business and Economics building is a 50,000 ft2 facility on the University of Idaho’s
Moscow campus. In 2015, the building’s main air handler had an economizer lockout of 65oF. Thus,
anytime the outdoor air was between 65oF and 75oF, this free cooling was being lost. Instead, chilled
water was being used unnecessarily to cool down the return air – which was above 75oF.
Through a sponsorship from AVISTA the University of Idaho’s Integrated Design Lab used a co-simulation
connected to the air handler controls to identify this problem. Last spring, the facilities team
implemented our recommendation by increasing the economizer lockout by 5oF on May 19th 2016.
As a result, energy has been saved although directly reporting those savings required the use of an
energy model. There is a record of monthly electricity consumption, but the building receives its chilled
water from a central plant on campus; the savings do not appear directly in the building’s electrical bills.
The savings are in the form of a reduction in the plant’s production of chilled water.
While the energy management system is capable of recording chilled water temperature and flow at the
COBE building, there are several months of missing data over the 2016 summer. Some of the data that
had been recorded was corrupt – listing strings of letters instead of numeric data. This made it
impossible to rely on recorded information alone.
Instead the UI-IDL used the energy model that had been calibrated to the building and employed in the
previous co-simulation study. We acquired historical weather based at the nearby Moscow airport and
used these 2016 readings to calculate the realized savings.
From May 19 2016 – December 31 2016, the differences of increasing the economizer lockout from 65oF
to 70oF were as follows:
Total energy saved: 14,477 kWh
Fans – increase of 3 kWh
Pumps – increase of 6 kWh
District Cooling – decrease of 14,417 kWh
District Heating – decrease of 69 kWh
IDL had estimated 56,000 kWh of savings in a 2014 report that was based on an economizer lockout
change from 60oF to 75oF. The ideal savings were estimated for a full year and at weather typical (not
actual) weather. The realized savings are less than predicted because the lockout change is smaller and
occurred only halfway through the year.
Staff_PR_25 - Attachment A
AVISTA CORPORATION RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO DATE PREPARED: 12/17/2020 CASE NO: AVU-E-20-13/AVU-G-20-08 WITNESS: N/A REQUESTER: IPUC Staff RESPONDER: Randy Gnaedinger TYPE: Production Request DEPARTMENT: Transmission Services
REQUEST NO.: Staff – 26 TELEPHONE: (509) 495-2047 REQUEST:
For the Aerogel Phase II project described in Avista's 2019 Idaho Research and Development Report, please answer the following questions: a. How does Avista plan to use the results of this project? b. What is Avista's timeline for using the results of this project?
c. The two-page report in Appendix A states that one of the deliverables is a cost
analysis. Please provide this cost analysis. d. Please provide any reports or cost/benefit analyses conducted by Avista in conjunction with this project.
RESPONSE: a. The results of the 2018-19 Aerogel Phase II project concluded that aerogel insulation and glazing is a viable technology for significantly increasing the R value of windows. The information and conclusions of the report are used by the DSM group to inform
future technologies that maybe applied to meet energy savings targets. It should be noted that while the technology is a way to increase the R-value for windows, the only application that might possibly exist into the foreseeable future is for privacy windows. b. Avista has no specified timeline for using the results of the project. As indicated in the
response above, the information and conclusions of the report are used by the DSM
group to inform future technologies that maybe applied to meet energy savings targets; the conclusions from these projects inform Avista’s overall pool of data that can be applied to future projects. As such, Avista may utilize this technology at any such time if/when there is a need for such practical application to increase the R-value for privacy
windows.
c. At the time of the 2019 Idaho Research and Development (R&D) Report, Avista’s research team was unable to complete a detailed cost analysis as anticipated, as there were no known manufacturers of a commercially-available window from which a cost estimate could be obtained. However, the R&D Report does contain analysis of energy
and heat flux savings by applying this technology to a window, so a cost/benefit
analysis could be performed in the future. d. Please see response c. above.
AVISTA R&D –All-Iron Battery
Allen Lab at the University of Idaho
Project/Task No. 03803300 / 242674
1Staff_PR_27 -Attachment A
The Problem
https://www.qic.com.au/knowledge-centre/technology-disruptions-affecting-infrastructure-20160414Queensland Investment Corporation 2Staff_PR_27 -Attachment A
Batteries
Daniel Cell
Invented 1836
3Staff_PR_27 -Attachment A
Batteries,
Simplified
Rusting Plating
electrons
V
Salt Bridge
Iron Nickel
Edison Cell
Invented 1901
4Staff_PR_27 -Attachment A
electrons
VBatteries by
Analogy
Iron metal really
wants to “deflate”
(lose electrons)
Nickel (II) is relatively
easy to “inflate”
(accept electrons)5Staff_PR_27 -Attachment A
Li-ion
6
Goodenough Cell
Invented 1980
Staff_PR_27 -Attachment A
electrons
VLi-ion
by Analogy
Li+ions want to get
out of graphite
Li+ions don’t mind
being stuck in Iron
Phosphate 7Staff_PR_27 -Attachment A
Tesla Powerwall (image by nrstor.com)GoalZero Yeti battery 8Staff_PR_27 -Attachment A
Credit: Ars technica
Advantages:
High energy density
Perfect for portable applications
Disadvantages:
Air and water sensitive chemistry
Flammable
Expensive
9Staff_PR_27 -Attachment A
Yearly Production of Metals
10
Metal World Production Total possible battery capacity
•Iron 3500 mmt 1000 TWh
•Sodium 108 mmt 108 TWh (experimental & like Li-ion)
•Aluminum 50 mmt 65 TWh (non rechargeable)
•Lithium 0.6 mmt 1 TWh
•Lead 3 mmt .15 TWh
•Vanadium .08 mmt .02 TWh
Staff_PR_27 -Attachment A
Three kinds of iron
Ferrous iron
Fe2+
Ferric iron
Fe3+
Iron Metal
Fe
11Staff_PR_27 -Attachment A
All iron
Oxidizing Reducing
electrons
V
Salt Bridge
Iron Iron (II)Iron (III) Iron (II)
12Staff_PR_27 -Attachment A
electrons
VAnalogy
Iron metal really
wants to “deflate”
(lose electrons)
Iron (III) is relatively
easy to “inflate”
(accept electrons)13Staff_PR_27 -Attachment A
Battery Fueled by Iron and Water Could Transform the Power Grid
Youtu.be/HmtI8Wat7rY14Staff_PR_27 -Attachment A
Flow Battery
https://en.wikipedia.org/wiki/Flow_battery
15
Advantages:
•Power and Energy are independent
Disadvantages:
•Pumps and plumbing add complexity and cost
•Power is limited by the size of the flow cell
Staff_PR_27 -Attachment A
Goal:
•Cheap
•Safe
•Recyclable
•Environmentally friendly/nontoxic
•1000 cycles
•Power: 20 W/liter
•Energy: 10 Wh/liter
16Staff_PR_27 -Attachment A
4.6 Joules, 100 ml
17Staff_PR_27 -Attachment A
pH 8 Trial:EDTA Sucrose Fructose Sulfate Cyanide
Compound:
Open Cell
Potential:
(Voltage)
0.5V 0.7V 0.4V 0.9V 0.8V
Discharge
to Failure:
(Coulombs)
16 5 4 290 >60
Iron Anode
Loss:
(Percent)
15.7%1.25%0.5%-15.9%13.5%
CN-
pH 7.5 pH 8 pH 8.5
Validation and Characterization
18Staff_PR_27 -Attachment A
19Staff_PR_27 -Attachment A
Gr
a
p
h
i
t
e
F
o
i
l
Fe
Fe
2+
Se
p
a
r
a
t
o
r
Fe
3+
Fe
2+
Ca
r
b
o
n
F
e
l
t
Gr
a
p
h
i
t
e
F
o
i
l
e-e-
Load
20Staff_PR_27 -Attachment A
21Staff_PR_27 -Attachment A
22Staff_PR_27 -Attachment A
20 mL
120 mL
200 mL
1000 mL
23Staff_PR_27 -Attachment A
700 mAh for
200 ml cell
24Staff_PR_27 -Attachment A
2.43m
2.59m
12.2m
76,000 L1 L version 1
•10 mA
•3 V
•7000 mAh
•21 Wh
•10 mW
•1.5 MWh
•0.76 kW
Extrapolate
25Staff_PR_27 -Attachment A
0.1 mL cell
26Staff_PR_27 -Attachment A
50 μm
4
27Staff_PR_27 -Attachment A
Fe
3+
Fe
2+
e-
Fe3+
Fe2+
As sulfate Soluble
e-
28Staff_PR_27 -Attachment A
Cu
r
r
e
n
t
(
A
)
Time (sec)
29Staff_PR_27 -Attachment A
Micro cell Lessons Learned
•Conductivity in Cathode is our major bottleneck
•We can overcome that with conductive carbon
•We don’t need carbon felt if we use conductive carbon
30
Fe
3+
Fe
2+
As sulfate
Gr
a
p
h
i
t
e
Staff_PR_27 -Attachment A
Micro cell cycling stability
31
0
0.5
1
1.5
0 100 200 300
Time (sec)
0
0.5
1
1.5
0 2000 4000 6000
Time (sec)
… 4000 cycles
Staff_PR_27 -Attachment A
2.43m
2.59m
12.2m
76,000 L1 L Version 2
•60 mA
•3 V
•7500 mAh
•13 Wh
•36 mW
•1.7 MWh
•.7 kW 32Staff_PR_27 -Attachment A
1 L Version 2, cycling stability
33Staff_PR_27 -Attachment A
+
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 1 2 3 4 5 6
Ah
Time (Hours)Graphite
Anode
chamber
Membrane
Cathode
chamber
Graphite
-
High surface area 0.006 L cell
34Staff_PR_27 -Attachment A
2.43m
2.59m
12.2m
76,000 L.006 L Version 3
•90 mA (!)
•.5 V
•70 mAh
•35 mWh
•45 mW
•0.43MWh
•820 kW 35Staff_PR_27 -Attachment A
Theoretical Max
36
-3.0
2.0
7.0
12.0
17.0
22.0
27.0
4/10/2018 7/19/2018 10/27/2018 2/4/2019 5/15/2019 8/23/2019 12/1/2019
Date
Capacity/Volume over Time
0
0.2
0.4
0.6
0.8
1
1.2
1.4
4/10/2018 7/19/2018 10/27/2018 2/4/2019 5/15/2019 8/23/2019 12/1/2019
Date
Current/Volume over Time
Staff_PR_27 -Attachment A
Conclusion:
Cheap
Safe
Recyclable
Environmentally friendly/nontoxic
1000 cycles
Power: 20 W/liter (9 W/liter)
Energy: 10 Wh/liter
Cost Estimate for 1.7 MWh container:
4.2 metric tons iron $20
7 metric tons carbon $7,000
38,000 L iron (III) chloride $8,000
38,000 L iron (II) chloride $18,000
38,000 m2 nafion-paper $150,000
38,000 m2 carbon sheet $150,000
$333,000
0.09 $/wH
37Staff_PR_27 -Attachment A
Backup
38Staff_PR_27 -Attachment A
39Staff_PR_27 -Attachment A
40Staff_PR_27 -Attachment A
AVISTA CORPORATION RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO DATE PREPARED: 12/17/2020 CASE NO: AVU-E-20-13/AVU-G-20-08 WITNESS: N/A REQUESTER: IPUC Staff RESPONDER: Randy Gnaedinger TYPE: Production Request DEPARTMENT: Transmission Service
REQUEST NO.: Staff – 27 TELEPHONE: (509) 495-2047 REQUEST:
For the All Iron Battery project described in Avista's 2019 Idaho Research and Development Report, please answer the following questions: a. How does Avista plan to use the results of this project? b. What is Avista's timeline for using the results of this project?
c. The two-page report in Appendix A states that one of the deliverables is a report on
"path to market." Please provide this report. d. Would Avista or its customers receive any tangible benefit if this project were to result in a marketable product? e. Please provide any reports or cost/benefit analyses conducted by Avista in
conjunction with this project.
RESPONSE: a. The results of the 2018-19 all-iron battery project concluded that an all-iron battery could be constructed and scaled as a viable technology for utility scale energy storage.
However, during the project it was evident that the chemistry still needs considerable work, and that scaling up in size may prove difficult. The results of this project may inform future understanding of battery chemistry technologies. b. There is no known timeline for a specific use of this project’s results.
c. The researcher’s time was focused on developing and testing the battery forms and
chemistry and they did not complete a path to market analysis. The final presentation in Staff_PR_27 – Attachment A (“20190820 Avista final presentation all-iron battery.pptx,”) does depict a hypothetical size, weight, and energy density for an all-iron battery at a small utility scale on slide 25.
d. Yes, if this research to result in a marketable product, that battery would be able to
safely store excess renewable energy at a potentially lower cost than other battery chemistries. The stored energy could then be used for higher load peaks or to fill in renewable energy production gaps. Cost effective utility scale storage has many other benefits that have been highlighted by organizations such as the National Renewable
Energy Laboratory (NREL).
e. Since there is no plan to further this research or implement the storage chemistry, no cost/benefit analysis was conducted.
AVISTA CORPORATION RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO DATE PREPARED: 12/17/2020 CASE NO: AVU-E-20-13/AVU-G-20-08 WITNESS: N/A REQUESTER: IPUC Staff RESPONDER: Randy Gnaedinger TYPE: Production Request DEPARTMENT: Transmission Services
REQUEST NO.: Staff – 28 TELEPHONE: (509) 495-2047 REQUEST:
For the Energy Trading System project described in Avista's 2019 Idaho Research and Development Report, please answer the following questions: a. How does Avista plan to use the results of this project? b. What is Avista's timeline for using the results of this project?
c. Is Avista considering an energy trading system like that conducted in this project?
d. Please provide any reports or cost/benefit analyses conducted by Avista in conjunction with this project. RESPONSE:
a. The results of this research are informing future distribution plans and possible customer engagement opportunities using markets. b. There are no specific plans currently to implement a distribution market. However, the dynamics of available energy markets are constantly changing. For example, Avista is
entering into the Western Energy Imbalance Market in March of 2022, opening the door to a west wide 5-minute locational marginal price market. In addition, FERC’s recent Order 2222 may further broader energy market participation. c. To the extent a distribution locational marginal price market was implemented on
Avista’s system, this research shows promise in a trading system that enable our
customers to participate as both consumers and prosumer of energy. A trading system like the one being researched enables transactions, price signals, and a check against reliable operations to all happen at the customer level. d. No cost/benefit analysis has been conducted in conjunction with this project.
AVISTA CORPORATION RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO DATE PREPARED: 12/17/2020 CASE NO: AVU-E-20-13/AVU-G-20-08 WITNESS: N/A REQUESTER: IPUC Staff RESPONDER: Randy Gnaedinger TYPE: Production Request DEPARTMENT: Transmission Services
REQUEST NO.: Staff – 29 TELEPHONE: (509) 495-2047 REQUEST:
For the IR Camera Phase I project described in Avista's 2019 Idaho Research and Development Report, please answer the following questions: a. How does Avista plan to use the results of this project? b. What is Avista's timeline for using the results of this project?
c. Please provide any reports or cost/benefit analyses conducted by Avista in
conjunction with this project. RESPONSE: a. The 2018-19 IR Camera HVAC control project builds on the key research findings of
the early 2017-18 Operating Temperate research, which concluded that during
occupancy cooling periods there is significant cooling energy savings of 3-8% by adjusting up the temperature control points, which will also result in increased occupancy comfort. See Staff_PR_29 – Attachment A for a copy of the 2018 Report completed by Integrated Design Lab (IDL). In the IR Camera research, IDL identified
a commercial gap where no signal device is available to control HVAC equipment based on thermal comfort, occupancy detection, and glare. The information and conclusions of these reports are used by the DSM group to evaluate and analyze potential future technologies that may be deployed to meet energy savings targets.
b. To date, there are no commercially-available devices to implement a timeline for using
the results of the IR Camera project. At such time these devices become fully commercially-available, Avista believes its customers would benefit in development of such a device primarily by having them installed in commercial buildings. The device could obtain energy savings, and also improve occupant comfort.
c. In the IR Camera final report, provided as Staff_PR_29 – Attachment B, the
researchers concluded that building such a device is feasible. Further laboratory testing concluded that a small office building in the north of Avista’s territory may save 5% annually on HVAC, while in the southern Avista territory those savings may be 7% annually (IDL Report 2019).
Report Number: 1808_01
USING IR CAMERAS IN BUILDING CONTROLS
PROJECT REPORT PREPARED FOR AVISTA UTILITIES
September 30, 2019
Prepared for:
Avista Utilities
Authors:
Damon Woods
Jubin Mathai
Ken Baker
Staff_PR_29 - Attachment B
This page left intentionally blank.
Staff_PR_29 - Attachment B
Prepared by:
University of Idaho Integrated Design Lab | Boise
306 S 6th St. Boise, ID 83702 USA
www.uidaho.edu/idl
IDL Director:
Ken Baker
Authors:
Damon Woods
Jubin Mathai
Ken Baker
Prepared for:
Avista Utilities
Contract Number: R-39872
Please cite this report as follows: Woods, D., Mathai, J., Baker, K.
(2019). Using IR Cameras in Building Controls. (1808_01).
University of Idaho Integrated Design Lab, Boise, ID.
Staff_PR_29 - Attachment B
DISCLAIMER
While the recommendations in this report have been reviewed for
technical accuracy and are believed to be reasonably accurate,
the findings are estimates and actual results may vary. All energy
savings and cost estimates included in the report are for
informational purposes only and are not to be construed as
design documents or as guarantees of energy or cost savings. The
user of this report, or any information contained in this report,
should independently evaluate any information, advice, or
direction provided in this report.
THE UNIVERSITY OF IDAHO MAKES NO REPRESENTATIONS,
EXTENDS NO WARRANTIES OF ANY KIND, EITHER EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE
WITH RESPECT TO THE INFORMATION, INCLUDING BUT NOT
LIMITED TO ANY RECOMMENDATIONS OR FINDINGS, CONTAINED
IN THIS REPORT. THE UNIVERSITY ADDITIONALLY DISCLAIMS ALL
OBLIGATIONS AND LIABILITIES ON THE PART OF UNIVERSITY FOR
DAMAGES, INCLUDING, BUT NOT LIMITED TO, DIRECT, INDIRECT,
SPECIAL AND CONSEQUENTIAL DAMAGES, ATTORNEYS’ AND
EXPERTS’ FEES AND COURT COSTS (EVEN IF THE UNIVERSITY HAS
BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES, FEES OR
COSTS), ARISING OUT OF OR IN CONNECTION WITH THE
MANUFACTURE, USE OR SALE OF THE INFORMATION, RESULT(S),
PRODUCT(S), SERVICE(S) AND PROCESSES PROVIDED BY THE
UNIVERSITY. THE USER ASSUMES ALL RESPONSIBILITY AND
LIABILITY FOR LOSS OR DAMAGE CAUSED BY THE USE, SALE, OR
OTHER DISPOSITION BY THE USER OF PRODUCT(S), SERVICE(S), OR
(PROCESSES) INCORPORATING OR MADE BY USE OF THIS REPORT,
INCLUDING BUT NOT LIMITED TO DAMAGES OF ANY KIND IN
CONNECTION WITH THIS REPORT OR THE INSTALLATION OF
RECOMMENDED MEASURES CONTAINED HEREIN.
Staff_PR_29 - Attachment B
This page left intentionally blank.
Staff_PR_29 - Attachment B
TABLE OF CONTENTS
1. Acknowledgements ..................................................................................................................... 3
2. Executive Summary ..................................................................................................................... 3
3. Research Motivation ................................................................................................................... 4
4. Project Methods ......................................................................................................................... 6
4.1 Importing Data from the IR Camera ...................................................................................... 6
4.2 Occupancy Detection Methods ............................................................................................. 8
4.3 Generating a Control Signal .................................................................................................. 9
5. Results ....................................................................................................................................... 10
5.1 Energy Modeling Methods .................................................................................................. 10
6. Discussion and Future Work ..................................................................................................... 12
7. Budget Summary ....................................................................................................................... 13
9. Appendix ................................................................................................................................... 17
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 2
Using Infrared Cameras in Building Controls (Report 1808_01)
ACRONYMS AND ABBREVIATIONS
AHU Air Handling Unit
ASHRAE American Society of Heating Refrigeration and Air conditioning Engineers
EMS Energy Management Control System
EPW Energy Plus Weather file
HVAC Heating, Ventilation and Air Conditioning
IDL Integrated Design Lab
MRT Mean Radiant Temperature
PMV Predicted Mean Vote
PNNL Pacific Northwest National Laboratory
PPD Percentage of Population Dissatisfied
TMY Typical Meteorological Year
UI University of Idaho
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 3
Using Infrared Cameras in Building Controls (Report 1808_01)
1. ACKNOWLEDGEMENTS
This research was made possible through funding support from Avista Utilities via Idaho PUC Case Number
AVU-E-13-08. The research team is very grateful for the project management from Natasha Jostad at T.O.
Engineers. Thank you to Randy Gnaedinger and Carlos Limon at Avista for their supervision, guidance, and
support of this project.
2. EXECUTIVE SUMMARY
The University of Idaho – Integrated Design Lab (UI-IDL) combined an infrared camera with a
thermostat to deliver more efficient heating and cooling signals. The team set up a camera in an
experimental chamber and verified the accuracy of the camera’s temperature readings. The team
then used an algorithm to process the camera’s measurements into a comfort index. In addition to
assessing comfort, the team processed the camera inputs through a separate machine learning
process to detect occupancy. The comfort and occupancy data is translated into a standard control
signal for a thermostat. The team used Energy models to estimate the potential savings of such a
controller and documented the results.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 4
Using Infrared Cameras in Building Controls (Report 1808_01)
3. RESEARCH MOTIVATION
Last year, the Integrated Design Lab collected data at several office buildings in the northwest. Data
collected from these sites was used to quantify the occupant comfort levels. The research team found
that all offices surveyed dipped below the ASHRAE comfort guidelines and were calculated to be
uncomfortably cold. Typical thermostat settings were found to be 70oF for heating and 74oF for cooling
with setbacks that varied by location. In every case that was monitored, instruments installed next to the
thermostats recorded higher air temperatures than instruments installed closer to the working areas of
the office employees.
Controlling a building based on both surface and air temperatures could reduce energy
consumption and promote healthier buildings. Most buildings manage their heating or cooling based
solely on zone air temperature. Even when the thermostats are kept within a narrow band of 4oF,
occupant complaints persist [1]. A large factor of human comfort is being missed in conventional
thermostats: the inclusion of the mean radiant temperature. Many previous studies have established that
the temperatures of the surfaces surrounding us have a far greater impact on our comfort than the air
temperature [2]. While operative temperature is widely used for predicting human comfort, rarely is it
directly used in building controls.
Building operations could benefit greatly from operative temperature control. Operative
temperature control would allow for wider air temperature setpoints, thus saving energy each year. A
study by the Pacific Northwest National Lab estimated the following general energy savings: a 2oF
adjustment on both heating and cooling setpoints led to a fairly uniform HVAC energy savings of 12-20%
[3]. Nationwide, such savings would be equivalent to 370-615 trillion Btu saved annually [4]. This research
focus is decidedly narrower in scope: focusing purely on small commercial office buildings in the Pacific
Northwest, where HVAC energy is typically 35% of the annual load [4]. Research has shown that a
thermostat reset could result in total annual energy savings of 4-7% of annual energy consumption per
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 5
Using Infrared Cameras in Building Controls (Report 1808_01)
building [5]. The adoption of wider setpoints based on operative temperatures by controls engineers and
consultants could quickly scale up and result in widespread savings. In addition to energy savings, the
operative temperature control could dramatically enhance occupant comfort, thus making it an attractive
service to offer to clients.
The simple shift in control from air temperature to operative temperature may have a dramatic
impact on energy consumption. Controlling for operative temperatures can be done with relatively simple
and inexpensive instruments: thermocouples. Expanding just the air cooling setpoint by 5oF can save up
to 27% of the total HVAC energy [5]. However, expanded setpoints are only viable if occupant comfort is
maintained. While establishing high cooling setpoints can save a lot of energy, if the occupant is
uncomfortable, they will either adjust the thermostat on their own or place considerable pressure on the
building operator to do so (thus nullifying any potential savings). Occupant satisfaction (or the lack
thereof) is the driving force behind thermostat adjustments. Up to $330 per employee per year in lost
productivity at work can be attributed to an office environment outside the bounds of ASHRAE 55’s
comfort setpoints [6]. Therefore, any thermostat control scheme, operative or otherwise, should focus on
comfort so that the setpoints are not immediately changed after they are implemented. An operative
thermostat control scheme focused on comfort has a far greater likelihood of remaining in place. Most
energy modeling misses this crucial interaction and instead assumes standard setpoints and setbacks.
That is why the first phase of this research was identifying typical setpoints at real buildings instead of
relying on ideal assumptions.
A control system that incorporates the operative temperature of a zone would allow for a wider
range of supply air temperatures and better meet the needs of occupants. The wider range of air
temperatures would reduce energy consumption and help to capitalize on operational features such as
natural ventilation, night-flush and optimized set-points. The outline of the work is to develop a proof of
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 6
Using Infrared Cameras in Building Controls (Report 1808_01)
concept for a thermostat that incorporates surface temperature measurements to deliver a better
comfort profile. This could save energy by reducing overcooling during the summer.
4. PROJECT METHODS
4.1 Importing Data from the IR Camera
The first step of the project was to verify that an inexpensive camera could pick up accurate
surface temperatures. The research team purchased a Flir C2 camera, which retails for $499 (the camera
specifications are available in the Appendix). The team set up the camera on a tripod in the experimental
chambers at IDL. The chamber is an empty office room with two windows and a ductless mini-split heat
pump. The room was heated and cooled at various points to test the camera at a range of temperatures.
The team measured surface temperatures in three ways: by using thermocouples attached to the wall
with insulation tape, specialized thermocouples (TMCx-HE) with thermal paste, an IR beam gun, an
expensive thermal camera, and the inexpensive Flir C2 camera.
Figure 1: Thermocouple sensor as viewed through the Flir C2 IR camera
After conducting experiments at a range of conditions, the team was able to verify that the Flir C2
could predict the surface temperature within 1oF of accuracy compared to the other methods. This
provided enough accuracy for the team to use the camera to estimate the mean radiant temperature of
the space. The camera requires inputs of reflected (air temperature), emissivity values (0.9 is the default
that worked best) and the camera’s distance from the object (set between 0 and 3 meters).
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 7
Using Infrared Cameras in Building Controls (Report 1808_01)
Once the camera settings were verified, the team used that data to calculate the average of the
surrounding surface temperatures in the room or the Mean Radiant Temperature (MRT). The mean
radiant temperature is based on view factors between a point in space and the surfaces that surround it.
P.O. Fanger [7] developed a set of formulas that can be used to predict the view factors of a seated
occupant and their surrounding surface temperatures. The team began with this approach. Fanger uses a
set of multiple integrals to calculate the radiation impact of each surface. The research team wished to
avoid using multivariable calculus, and so curve approximations were developed in Excel. The best method
was found to be recommended by the RHEVA guidebook [8]. These equations were programmed into an
Excel worksheet and are listed below:
Fp-N = Fmax (1-e-(a/c)/Ʈ ) . (1-e-(h/c)/ƴ )
Ʈ = A+B(a/c)
Ƴ = C+ D(b/c) + E(a/c)
Table 1 : Equations for calculations of the angle factors
[8]
The mean radiant temperature is calculated using the below equation in excel,
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇=�∑�𝐹𝐹𝑝𝑝−𝑁𝑁∗ 𝑇𝑇𝑁𝑁4)4
Where Tmrt is mean radiant temperature
Fmax A B C D E
Seated Person, Figure
2.2a Vertical Surfaces:
Wall, Window
0.118 1.216 0.169 0.717 0.087 0.052
Seated Person, Figure
2.2a Horizontal Surfaces:
Wall, Window
0.116 1.396 0.13 0.951 0.08 0.055
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 8
Using Infrared Cameras in Building Controls (Report 1808_01)
Fp-N is the view factor from person to surface
TN is the temperature of the surface being measured by the infrared camera
The team now had the ability to take a photo of the wall, export that temperature data into Excel,
and estimate the mean radiant temperature for an occupant. The mean radiant temperature was
imported into a separate spreadsheet that calculates thermal comfort predictions. Thermal comfort
metrics include the following: humidity, dry-bulb temperature, mean radiant temperature, air velocity,
clothing level, and activity level. The team measured humidity and dry-bulb temperature with HOBO data
loggers. Assumptions were made for air velocity (20 fpm) and occupant clothing and activity levels (seated,
typing, wearing jeans and a long-sleeve shirt).
The comfort metrics can be calculated in two ways: either as a Predictive Mean Vote (PMV) or as
a Percentage of Population Dissatisfied (PPD). The PMV ranges in its values from -3 to +3, where negative
values are associated with being too cold, and positive values corresponding to being too warm. Thermal
comfort is defined as having a PMV between -0.5 and +0.5. This value is directly related to the percent
that are predicted to be uncomfortable in that situation. For example, a PMV of 1 equals a PPD of 25%.
The minimum PPD is 5%.
4.2 Occupancy Detection Methods
One of the advantages of using an IR camera to measure surface temperatures is that it can also
be used the count the occupants in a room. This provides crucial information to the HVAC system on the
ventilation requirements in a space, which are now required in many systems under Washington’s new
energy code (WEC C403.1.6.1). The team used machine learning to detect occupants with the low
resolution camera. There are numerous open source codes available to detect humans in RGB images.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 9
Using Infrared Cameras in Building Controls (Report 1808_01)
The IDL researchers used the same codes by changing the infrared image inputs to an RGB format. The
application program interface used is the TensorFlow object detection.
TensorFlow is an open source platform from machine learning. The TensorFlow project has many
useful framework extensions, one such is Object detection API. As the name suggests, the extension
enables TensorFlow users to create powerful object detection models using TensorFlow’s directed
compute graph infrastructure. The TensorFlow Object Detection API is an open source framework built
on top of TensorFlow that makes it easy to construct, train and deploy object detection models. This API
can be used to detect, with bounding boxes, objects in images or video using either some of the pre-
trained models made available. The object detection code was downloaded and used from Github, a host
to open source projects owned by Microsoft. TensorFlow was developed by Google’s AI team and codes
are available on Github to download. It can be implemented and modified according to user’s application
needs. The repository is maintained by individual researchers on Github.
4.3 Generating a Control Signal
The purpose of using the camera is to send better heating and cooling signals to the HVAC system for
an office. The established protocol for sending control signals is the Building Automated Control Network
Protocol, or BACnet. The research team bought a wireless thermostat as well as a digital router that can
process BACnet signals. With this, the team was able to establish both wired and wireless communication
with a thermostat. Typically a thermostat tries to maintain a certain air temperature keeping the room
within a certain deadband between heating and cooling setpoints. For this project, the setpoints were
manipulated in order to maintain certain comfort levels. For example, if the PMV comfort calculated for
the room drops below -0.5, then the setpoints are increased above the current air temperature to force
the heating system on. Similarly, if the calculated PMV rises above +0.5, the thermostat is to call for
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 10
Using Infrared Cameras in Building Controls (Report 1808_01)
cooling. During unoccupied times, the setpoints are expanded dramatically and the PMV is allowed to
float anywhere between +3.
5. RESULTS
5.1 Energy Modeling Methods
The research team used a small office prototype model to estimate energy savings. The energy
model uses the DOE software program, EnergyPlus. The geometry and loads are derived from the Pacific
Northwest National Lab’s database of standard building types. These prototype models are used by
organizations to test the impacts of new energy codes.
Figure 2: PNNL Small office prototype energy model
The model can be paired with different weather files to estimate utility costs in different locations.
The American Society of Heating Refrigeration and Air conditioning Engineers (ASHRAE) partitions the US
into eight climate zones that account for minimum/maximum temperatures and humidity. All of Avista’s
service territory falls into either climate zone 5B, or 6. A sketch overlaying climate information onto the
general Avista territory is shown in Figure 1.1
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 11
Using Infrared Cameras in Building Controls (Report 1808_01)
Figure 3 A general outline of Avista’s service territory and the climate zones present in that area.
The weather files used in this project included locations in Lewiston, Spokane, and Kalispell.
Lewiston and Spokane, which are both within ASHRAE climate zone 5. The Kalispell weather file is in
climate zone 6 and was used to approximate the weather Avista’s rural customers experience in northern
Idaho and Washington.
Initially, the prototype model was simulated with a standard thermostat that uses air temperature
setpoints and setbacks. The team then adjusted the code within the model to have the thermostat trigger
heating and cooling based on PMV instead of air temperatures. During business hours, the PMV was
maintained between +0.5, while at night, the PMV was allowed to float between +3. Savings were
identified for each climate. For each of the weather files tested, the heating load increased, while the
cooling load decreased. In each situation, the decrease in cooling load was greater than the increase in
heating, leading to a net energy savings in every location. The savings were smallest for the colder climate
zone 6, with the Kalispell model predicting 3% HVAC savings annually in a typical year. In Spokane, the
small office model predicted 5% annual HVAC energy savings from using a thermostat based on PMV,
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 12
Using Infrared Cameras in Building Controls (Report 1808_01)
while in Lewiston, the building saved 7% of its annual HVAC energy. The small office in Spokane reduced
its cooling energy by over 1,500 kWh, while keeping the occupants more comfortable.
6. DISCUSSION AND FUTURE WORK
The research work this year proved that it is possible to use an inexpensive thermal camera to
predict comfort and occupancy in an office and run a thermostat based on those comfort predictions. The
energy models showed net energy savings for Avista customers in both ASHRAE climate zones 5 and 6.
The camera was able to accurately record surface temperatures, those surface temperatures were
converted into a comfort signal. The team also trained the camera to recognize occupants in low-
resolution IR images using machine learning. By processing that information through an Excel
spreadsheet, that information can be sent as a BACnet signal to a standard thermostat.
Over the next year, the research team will work to streamline and automate this process. The goal
is to commercialize this device so that Avista customers may benefit from its development. The IDL also
hopes to incorporate glare detection into the algorithm so that this device may serve multiple functions
and eliminate superfluous sensors and wiring and in buildings. So far the team has only sent and received
thermostat signals from a computer. In the next year, the team aims to connect a heating and cooling
device to the thermostat to complete the control feedback loop. This will allow for further testing and
refinement as well as providing physical data on energy savings.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 13
Using Infrared Cameras in Building Controls (Report 1808_01)
7. BUDGET SUMMARY
These hours reflect only Avista’s contribution to this project and are not reflective of total project
investment by the research team, industry sponsors, or other university staff.
Personnel Hours estimate Description
FY18/FY19
Expense
PI/Faculty Salaries (Cooper and TBD ME) $ 13,243
PI/Faculty Benefits (Cooper and TBC Woods) 3,509
Graduate Student Salaries 8,960
Student Benefits 340
Graduate Student Tuition Remission (partial) 5,387
Travel 2,250
Equipment 500
F&A / Overhead (Excludes Tuition) 13,200
Total $ 47,935
Indirect Costs
For this contract, UI-IDL was considered an on-campus unit of the University of Idaho with a federally
negotiated rate of 50.3%.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 14
Using Infrared Cameras in Building Controls (Report 1808_01)
8. REFERENCES
desirable?," Building and Environment, vol. 45, no. 1, pp. 4-10, 2010.
Clima 2007 WellBeing Indoors, Helsinki, 2007.
large office buildings," Pacific Northwest National Lab Report 21596, Richland, WA, 2012.
U.S. Energy Information Administration, Washington, D.C., 2012.
and design considerations for new and retrofit buildings," Building and Environment, 2014.
ASHRAE,
McGraw-Hill Book Company, 1970.
and cooling systems, Federation of European Heating and Air-Conditioning Associations, 2009.
https://groups.google.com/forum/#!forum/bcvtb.
Bed," Philips, Briarcliff Manor, 2011.
Whole Building Performance Assessment," Building and Environment, pp. 100-108, 2012.
Virtual Test Bed," Journal of Building Performance Simulation, pp. 185-203, 2011.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 15
Using Infrared Cameras in Building Controls (Report 1808_01)
Interfaces of the Building Controls Virtual Testbed," in Conference of International Building
Performance Simulation Association, Sydney, 2011.
American Sociaety of Heating, Refrigerating and Air-Conditioning Engineers, Inc., Atlanta, 2007.
Refrigerating, and Air-Conditioning Engineers, Atlanta, 2005.
Developing and Testing Control Algorithms Strategies and Systems," in International Building
Performance Simulation Association, Beijing, 2007.
Greenhouse Gas Emissions," Lawrence Berkeley National Lboratory, 2009.
Characteristics of Commercial Building HVAC Systems Volume III: Energy Savings Potential,"
Department of Energy, Cambridge, 2002.
HVAC Systems Volume II: Thermal Distribution, Auxiliary Equipment, and Ventilation," U.S.
Department of Energy, Cambridge, 1999.
Nonresidential sector: ID, MT, OR, and WA," Northwest Energy Efficiency Alliance, Portland, 2008.
American Society of Heating and Refrigerating Engineers, Atlanta, 2002.
[Accessed 15 January 2017].
https://github.com/hrbrmstr/darksky. [Accessed 15 January 2017].
Renewable Energy Laboratory, Golden, Colorado, 2012.
[Online]. Available: http://solardat.uoregon.edu.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 16
Using Infrared Cameras in Building Controls (Report 1808_01)
Standards, 1983.
National Passive Solar Conference, 1978.
Radiation for Chinese Locations," Journal of Asian Architecture and Building Engineering, vol. 41,
2003.
Radiation on Inclined Surfaces: Models Re-Visited," Energies, vol. 10, no. 134, 2017.
surface and prediction of insolation on tilted surfaces," Transactions, Architectural Institute of
Japan, no. 330, 1983.
real-time energy demand prediction using weather forecasting data," Energy and Buildings, vol.
57, pp. 250-260, 2013.
Diffuse Radiation Model," Albany, NY, 1988.
Occupancy," American Society of Heating Refrigerating, and Air-Conditioning Engineers, Inc.,
Atlanta, 2013.
Staff_PR_29 - Attachment B
Integrated Design Lab | Boise 17
Using Infrared Cameras in Building Controls (Report 1808_01)
9. APPENDIX
Table 2: Camera Specifications
Specification FLIR C2
USB 2.0
Operating temperature range -10°C to +50°C
Color palettes
Iron, Rainbow, Rainbow HC,
Staff_PR_29 - Attachment B
Report Number: 1708_01
MANAGING FOR EFFICIENCY BASED ON OPERATIVE
TEMPERATURES
PROJECT REPORT PREPARED FOR AVISTA UTILITIES
August 31, 2018
Prepared for:
Avista Utilities
Authors:
Damon Woods
Elizabeth Cooper
Neha Pokhrel
Ryker Belnap
Staff_PR_29 - Attachment A
This page left intentionally blank.
Staff_PR_29 - Attachment A
Prepared by:
University of Idaho Integrated Design Lab | Boise
306 S 6th St. Boise, ID 83702 USA
www.uidaho.edu/idl
IDL Director:
Elizabeth Cooper
Authors:
Damon Woods
Elizabeth Cooper
Neha Pokhrel
Ryker Belnap
Prepared for:
Avista Utilities
Contract Number: R-39872
Please cite this report as follows: Woods, D., Cooper, E., Pokhrel,
N., and Belnap, R. (2018). Managing for Efficiency Based on
Operative Temperatures. (1708_01). University of Idaho
Integrated Design Lab, Boise, ID.
Staff_PR_29 - Attachment A
DISCLAIMER
While the recommendations in this report have been reviewed for
technical accuracy and are believed to be reasonably accurate,
the findings are estimates and actual results may vary. All energy
savings and cost estimates included in the report are for
informational purposes only and are not to be construed as
design documents or as guarantees of energy or cost savings. The
user of this report, or any information contained in this report,
should independently evaluate any information, advice, or
direction provided in this report.
THE UNIVERSITY OF IDAHO MAKES NO REPRESENTATIONS,
EXTENDS NO WARRANTIES OF ANY KIND, EITHER EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE
WITH RESPECT TO THE INFORMATION, INCLUDING BUT NOT
LIMITED TO ANY RECOMMENDATIONS OR FINDINGS, CONTAINED
IN THIS REPORT. THE UNIVERSITY ADDITIONALLY DISCLAIMS ALL
OBLIGATIONS AND LIABILITIES ON THE PART OF UNIVERSITY FOR
DAMAGES, INCLUDING, BUT NOT LIMITED TO, DIRECT, INDIRECT,
SPECIAL AND CONSEQUENTIAL DAMAGES, ATTORNEYS’ AND
EXPERTS’ FEES AND COURT COSTS (EVEN IF THE UNIVERSITY HAS
BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES, FEES OR
COSTS), ARISING OUT OF OR IN CONNECTION WITH THE
MANUFACTURE, USE OR SALE OF THE INFORMATION, RESULT(S),
PRODUCT(S), SERVICE(S) AND PROCESSES PROVIDED BY THE
UNIVERSITY. THE USER ASSUMES ALL RESPONSIBILITY AND
LIABILITY FOR LOSS OR DAMAGE CAUSED BY THE USE, SALE, OR
OTHER DISPOSITION BY THE USER OF PRODUCT(S), SERVICE(S), OR
(PROCESSES) INCORPORATING OR MADE BY USE OF THIS REPORT,
INCLUDING BUT NOT LIMITED TO DAMAGES OF ANY KIND IN
CONNECTION WITH THIS REPORT OR THE INSTALLATION OF
RECOMMENDED MEASURES CONTAINED HEREIN.
Staff_PR_29 - Attachment A
This page left intentionally blank.
Staff_PR_29 - Attachment A
TABLE OF CONTENTS
1. Acknowledgements ..................................................................................................................... 3
2. Executive Summary ..................................................................................................................... 4
3. Research Motivation ................................................................................................................... 5
4. Project Methods ......................................................................................................................... 7
4.1 Establishing a Baseline .......................................................................................................... 7
4.2 Energy Modeling Methods .................................................................................................. 11
4.2.1 Creating the Weather File ............................................................................................ 12
4.2.2 Generating Solar Information from the data for other sites........................................ 13
4.2.3 Formatting the Weather Histories ............................................................................... 17
5. Monitoring Results .................................................................................................................... 19
5.1 Thermostat Reading Versus Other Instruments ................................................................. 23
5.2 Energy Modeling Results - Winter ...................................................................................... 25
5.3 Energy Modeling Results - Summer .................................................................................... 27
6. Discussion and Future Work ..................................................................................................... 29
7. Budget Summary ....................................................................................................................... 32
9. Appendix ................................................................................................................................... 36
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 2
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
ACRONYMS AND ABBREVIATIONS
AHU Air Handling Unit
API Application Programming Interface
ASHRAE American Society of Heating Refrigeration and Air conditioning Engineers
EMS Energy Management Control System
EPW Energy Plus Weather file
HVAC Heating, Ventilation and Air Conditioning
IDL Integrated Design Lab
NOAA National Oceanographic and Atmospheric Administration
NRSDB National Radiation Solar Data Base
PMV Predicted Mean Vote
PPD Percentage of Population Dissatisfied
RMSE Root Mean Squared Error
TMY Typical Meteorological Year
UI University of Idaho
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 3
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
1. ACKNOWLEDGEMENTS
This research was made possible through funding support from Avista Utilities via Idaho PUC Case Number
AVU-E-13-08. The research team expresses gratitude to Avista staff and project managers for their
support of this project. This project could not have happened without the coordination and help received
from facilities managers at each of the sites that was measured.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 4
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
2. EXECUTIVE SUMMARY
The University of Idaho – Integrated Design Lab (UI-IDL) monitored the humidity, surface
temperatures and air velocities at six different office sites. Data collected from these sites was used
to quantify the occupant comfort levels. Energy models were created for each of these sites. Weather
histories and solar models were used to generate simulations based on historical measurements. The
models were then adjusted to match the measured performance. Finally, the control settings in the
energy models were adjusted in order to contrast current performance with new control sequences
that provided better operational performance both in terms of comfort and energy.
The research team found that all offices surveyed dipped below the ASHRAE comfort guidelines
and were calculated to be uncomfortably cold. Typical thermostat settings were found to be 70oF for
heating and 74oF for cooling with setbacks that varied by location. In every case that was monitored,
instruments installed next to the thermostats recorded higher air temperatures than instruments
installed closer to the working areas of the office employees. Analysis of the data collected during the
winter indicated that increasing clothing levels helped, but did not alleviate predicted discomfort until
the heating setpoint was raised to 72oF. Raising this setpoint may still come with an economic benefit
to customers due to increased productivity from a more comfortable workforce and may cut down
on the number of electric space heaters that were present in almost every location visited. Analysis
of the summer data showed that increasing the cooling setpoint to 76oF with a setback of 80oF
resulted in a cooling load reduction of 15-40% for each site during the week observed. Including more
holistic metrics such as surface temperatures showed that raising setpoints during the summer would
both increase comfort and save energy for commercial offices in Avista Territory.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 5
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
3. RESEARCH MOTIVATION
Controlling a building based on both surface and air temperatures could reduce energy
consumption and promote healthier buildings. Most buildings manage their heating or cooling based
solely on zone air temperature. Even when the thermostats are kept within a narrow band of 4oF,
occupant complaints persist [1]. A large factor of human comfort is being missed in conventional
thermostats: the inclusion of the mean radiant temperature. Many previous studies have established that
the temperatures of the surfaces surrounding us have a far greater impact on our comfort than the air
temperature [2]. Imagine a driver getting into a car that’s been in the sun all day – the air conditioning
may be running, but the driver is still sweating for the first part of the drive until all the surfaces cool off.
A better comfort metric is the operative temperature which is a mix of air and surface temperatures.
While operative temperature is widely used for predicting human comfort, rarely is it directly used in
building controls.
Building operations could benefit greatly from operative temperature control. Operative
temperature control would allow for wider air temperature setpoints, thus saving energy each year. A
study by the Pacific Northwest National Lab estimated the following general energy savings: a 2oF
adjustment on both heating and cooling setpoints led to a fairly uniform HVAC energy savings of 12-20%
[3]. Nationwide, such savings would be equivalent to 370-615 trillion Btu saved annually [4]. This research
focus is decidedly narrower in scope: focusing purely on small commercial office buildings in the Pacific
Northwest, where HVAC energy is typically 35% of the annual load [4]. Research has shown that a
thermostat reset could result in total annual energy savings of 4-7% of annual energy consumption per
building [5]. The adoption of wider setpoints based on operative temperatures by controls engineers and
consultants could quickly scale up and result in widespread savings. In addition to energy savings, the
operative temperature control could dramatically enhance occupant comfort, thus making it an attractive
service to offer to clients.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 6
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
The simple shift in control from air temperature to operative temperature may have a dramatic
impact on energy consumption. Controlling for operative temperatures can be done with relatively simple
and inexpensive instruments: thermocouples. Expanding just the air cooling setpoint by 5oF can save up
to 27% of the total HVAC energy [5]. However, expanded setpoints are only viable if occupant comfort is
maintained. While establishing high cooling setpoints can save a lot of energy, if the occupant is
uncomfortable, they will either adjust the thermostat on their own or place considerable pressure on the
building operator to do so (thus nullifying any potential savings). Occupant satisfaction (or the lack
thereof) is the driving force behind thermostat adjustments. Up to $330 per employee per year in lost
productivity at work can be attributed to an office environment outside the bounds of ASHRAE 55’s
comfort setpoints [6]. Therefore, any thermostat control scheme, operative or otherwise, should focus on
comfort so that the setpoints are not immediately changed after they are implemented. An operative
thermostat control scheme focused on comfort has a far greater likelihood of remaining in place. Most
energy modeling misses this crucial interaction and instead assumes standard setpoints and setbacks.
That is why the first phase of this research was identifying typical setpoints at real buildings instead of
relying on ideal assumptions.
A control system that incorporates the operative temperature of a zone would allow for a wider
range of supply air temperatures and better meet the needs of occupants. The wider range of air
temperatures would reduce energy consumption and help to capitalize on operational features such as
natural ventilation, night-flush and optimized set-points. The outline of the work is to profile typical
setpoints found in operation, test alternative control methods, and estimate potential savings. No
occupant surveys were required for this study (opinions are highly variable and unique). Instead, the
research relied on established thermal comfort criteria based on operative temperatures and humidity
readings taken at the site [7]. This study, leveraged extensive data collection paired with energy modeling
and comfort standards to inform efficient control schemes. The work provided key insights on how to best
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 7
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
balance thermal comfort and energy savings for small commercial office buildings in the Pacific
Northwest.
4. PROJECT METHODS
4.1 Establishing a Baseline
This work began with the research of typical thermostat settings for buildings in Avista territory.
The baseline was established through a mix of literature reviews and contact with controls engineers,
consultants, and building operators. While there are established guidelines for thermostat setpoints and
setbacks, the goal of this research phase was to uncover what the actual settings typically are. After
reaching out to several engineering firms, control contractors, and facility managers, the general
consensus was to keep buildings at a heating setpoint of 70oF and a cooling setpoint of 74oF. All who were
surveyed gave answers within one or two degrees of this range. One controls contractor noted that
thermostats can contain either a slight delay or overshoot so that the system will not commence operation
until the setpoint is exceeded by about 1oF. During the data collection phase, independent temperature
sensors were set next to the zone thermostats to verify these settings.
IDL coordinated with Avista to select six building sites for analysis as case studies. Each site was
an office – either a private office, open office, or cubicle. Some of these areas allowed employee access
to thermostat controls and some had locked thermostats that gave only facilities managers (not
employees) control over the setpoints. Each site was located in a climate shared with Avista Service
territory customers. The American Society of Heating Refrigeration and Air conditioning Engineers
(ASHRAE) partitions the US into eight climate zones that account for minimum/maximum temperatures
and humidity. All of Avista’s service territory falls into either climate zone 5B, or 6. A sketch overlaying
climate information onto the general Avista territory is shown in Figure 1.1
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 8
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.1 A general outline of Avista’s service territory and the climate zones present in that area.
Due to IDL’s location, several of the sites surveyed were located in Boise, which is also in climate
zone 5. The buildings surveyed ranged in size and had a diverse mix of HVAC systems. At each site, the
team chose one or more spaces inside the building to study in detail. Next, an array of data loggers were
installed in those spaces to record comfort metrics at the site throughout a work week. The comfort
metrics included the following: humidity, dry-bulb temperature, air velocity, and surface temperatures of
relevant surfaces (walls, windows, floors, ceilings, and desks). An image of one of the offices and the
instruments installed is shown in Figure 1.2 while a list of data collected and the instruments used is
provided in Table 1.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 9
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.2 The temperature and velocity sensors installed at one of the offices surveyed.
1. COMFORT METRIC DATA COLLECTED
Measurement Tool
The research team also gathered thermostat readings and energy consumption data from the
energy management system (EMS) if it was possible. The IDL team only used the EMS data at one site and
this data collection was through a secondary program called SkySpark. A list of the sites selected and their
HVAC systems is shown in Figure 1.1. The dates of data collection for each site are listed in the Appendix.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 10
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.1 The six sites that were chosen for analysis and the HVAC system at each site.
With data from logger deployment, the research team was able to calculate thermal comfort
conditions without the need for occupant surveys. The comfort calculation is listed in ASHRAE Standard
55 and is provided as a code in BASIC [7]. The comfort calculation includes the indoor conditions listed in
Table 1, as well as view factors to each surface, and the clothing and activity levels of the occupants. The
activity level was set at 1.1 Mets which corresponds to someone sitting and typing. The clothing level was
set at 0.57 as that was representative of the clothing observed (shoes, socks, underwear, pants and long-
sleeved button downs). The IDL converted these comfort equations into an Excel sheet so that the comfort
index could be calculated for every data point. In most cases the data was recorded in five minute
intervals.
The comfort metrics can be calculated in two ways: either as a Predictive Mean Vote (PMV) or as
a Percentage of Population Dissatisfied (PPD). The PMV ranges in its values from -3 to +3, where negative
values are associated with being too cold, and positive values corresponding to being too warm. Thermal
comfort is defined as having a PMV between -0.5 and +0.5. This value is directly related to the percent
that are predicted to be uncomfortable in that situation. For example, a PMV of 1 equals a PPD of 25%.
The minimum PPD is 5%.
Office A
Office B
Office C
Office D
Office E
Office F
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 11
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
4.2 Energy Modeling Methods
The team developed simple energy models for the selected sites in the established, DOE software
program, EnergyPlus. Architectural drawings were obtained for each site either through public records or
supplied by the facilities team. This allowed the IDL research team to replicate the geometry in realistic
detail. The EnergyPlus models were informed by the temperature and thermostat data gathered at the
site instead of ideal assumptions. The HVAC modeling was set to ideal air loads, which meant that the
model would provide as much heating or cooling as needed at 100% efficiency to meet the setpoint. The
internal loads were set at ASHRAE defaults for medium to large offices and the construction materials
were selected based on the architectural drawings [8] [9]. By using the recordings from the site, the
research team was able to create schedules that followed the recorded setpoints in 15 minute intervals
so the models could match the recorded behavior.
In order to run the controls based on measured loads and responses, a framework was created
that can support the simulation inputs and process the outputs. Instead of using a full year of weather,
the simulation time was shortened to two weeks at most, and simply matched the time during which
instruments were installed at each site. This enabled rapid feedback and provided a window into how the
building was performing. The simulation would run different scenarios including keeping the setpoint the
same or adapting the setpoints and operation times to increase efficiency and comfort. The outputs of
the simulation included comfort and energy metrics: the predicted mean vote and the heating and cooling
energy consumption based on an ideal system. The next step of the process was to develop historical
weather files for the energy models to simulate observed conditions.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 12
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
4.2.1 Creating the Weather File
Developing the weather history file involved three specific phases:
1. Collecting a weather history for a specific time and place.
2. Generating solar radiation data.
3. Formatting the information for EnergyPlus.
Weather histories can be accessed in any number of ways. Many use data from the National
Oceanographic and Atmospheric Association (NOAA). At times, the NOAA data can have gaps or errors in
it. This research relied on a third-party weather resource: the DarkSky Application Programing Interface
(API) [10]. The Dark Sky Company is a self-funded software startup that specializes in local weather
observations and forecasts. It uses bots to access a range of weather histories and forecasts, collects
NOAA satellite information and filters it all. The service is used by Microsoft, Yelp, ConEdison, and others.
Users register for an API key that gives them access to 1,000 free queries each day. The data analyst Bob
Rudis built an R-script that leverages this API and uses a j-son wrapper in order to download this data for
a given set of coordinates on command [11]. The research team adapted Rudis’ script in order to download
a weather history specific for the study’s locations and times in an hourly format. The script then sends
this output to a comma delimited file for further operation.
EnergyPlus models require 35 weather data points to perform a complete simulation. The fields
of interest from the weather history include the timestamp, precipitation intensity, dry bulb temperature,
dewpoint, humidity, atmospheric pressure, wind speed, wind bearing, cloud cover, and visibility. Many of
these values could be directly used in the EPW for the simulation, with careful attention to the units used
in the EPW as described in the EnergyPlus Auxiliary Programs guide. However, many solar fields required
by the EPW had to be calculated. Most of these 35 weather aspects have to do with the quality of sunlight
as this has a major impact on daylighting calculations. Very few weather stations provide the level of detail
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 13
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
on solar information contained in the EPW. There are several research stations in the Pacific Northwest
that provide some of this historical information, but this detailed solar data was not available from Dark
Sky or even in the NOAA data recorded at the Boise airport. Therefore, research included the development
of this solar information using solar altitude equations and regressions based on the observed cloud cover,
temperature, and humidity.
4.2.2 Generating Solar Information from the data for other sites
EnergyPlus parses the solar radiation in the weather file into several different fields. These
include, the extraterrestrial horizontal, extraterrestrial direct normal, horizontal infrared sky, global
horizontal, direct normal, and diffuse horizontal radiation. The Department of Energy has determined
typical values for each of these radiation fields for locations across the country. These values can be found
in the Typical Meteorological Year (TMY) file that is standard for many energy simulations. The solar values
in the TMY are derived from the National Solar Radiation Database (NSRDB). While the NSRDB is derived
from observed data, most of it is modeled [12].
“Nearly all of the solar data in the original and updated versions of the NSRDB are modeled. The intent of
the modeled data is to present hourly solar radiation values that, in the aggregate, possess statistical
properties (e.g., means, standard deviations, and cumulative frequency distributions) that are as close as
possible to the statistical properties of measured solar data over the period of a month or year. These data
do not represent each specific hourly value of solar radiation to the same or equivalent accuracy as the
long-term statistics.”
The data that the NRSDB models are based on comes from about 40 stations in the United States.
The stations and data for the Pacific Northwest comes from the University of Oregon Solar Radiation
Monitoring Laboratory [13]. The NSRDB models can vary significantly from the historical observations. For
the four months of the TMY in Boise, ID when observed data were available, the measured diffuse
radiation and the modeled diffuse radiation had a correlation coefficient of only 0.67. This is in stark
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 14
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
contrast to the observed dry bulb temperatures at the same site which had a correlation coefficient of
0.98 for the same period. The histories must rely on some models to estimate solar radiation based on
sky and temperature conditions. In order to test the effectiveness of the correlations used for developing
weather data, the team compared the models to observed data rather than the TMY. This is to avoid
comparing a model to a model, and instead only compares the model to observed conditions.
The extraterrestrial horizontal radiation and extraterrestrial direct normal radiation in (Wh/m2)
are a function of the latitude and solar hour. These can be derived using clear-sky solar insolation
equations. The extraterrestrial solar radiation is listed as Io and is the total solar radiation that falls on a
spot above the atmosphere. It is based on the eccentricity of the earth’s orbit around the sun.
𝐼𝑜=1367 ∗ 𝐸𝑜
𝐸𝑜= 1.00011 + 0.034221cosΓ + 0.00128 sin Γ + 0.000719cos2Γ + 0.000077 sin2Γ
Γ = 2𝜋𝑛 − 1
365
Io = Extraterrestrial direct normal radiation (Wh/m2)
Eo = The Eccentricity of the earth’s orbit at a particular time
n = The numerical day of the year
The Horizontal Infrared Radiation Intensity from the Sky measured in (Wh/m2) is dependent on
other weather conditions. For locations where this parameter is not recorded or available (e.g. Lewsiton,
Idaho), Walton and Clark have developed an estimation method [14] [15]:
𝐻𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙𝐼𝑅=𝜖𝜎𝑇4𝑑𝑟𝑦𝑏𝑢𝑙𝑏
𝐻𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙𝐼𝑅 = The Horizontal Infrared Radiation Intensity from the sky (W/m2)
𝜖= The sky emissivity
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 15
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
𝜎 = The Stephan-Boltzman constant 5.6697e-8 (W/m2K4)
𝑇𝑑𝑟𝑦𝑏𝑢𝑙𝑏 = Outdoor dry bulb temperature (K)
𝜖 = (0.787 + 0.764𝑙𝑛(𝑇𝑑𝑒𝑤𝑝𝑜𝑖𝑛𝑡
273 ))(1 + 0.0224𝑁 + 0.0035𝑁2 + 0.00028𝑁3)
𝑇𝑑𝑒𝑤𝑝𝑜𝑖𝑛𝑡= Outdoor dew point temperature (K)
𝑁 = Opaque sky cover (in tenths)
The opaque sky cover has a minimum value of 0 and a maximum value of 10. The opaque sky
cover is slightly different than the total sky cover. Typically, only the total sky cover is reported. The
fraction of the cloud cover that reflects the solar radiation is called the opaque sky cover. The opaque sky
cover is always less than the total sky cover. The two values are close, and for Boise, ID have a correlation
of 0.91. For the estimate, the opaque sky cover was assumed to be equal to the total sky cover.
The other two solar fields of consequence include the direct normal radiation and the diffuse
horizontal radiation. These are components of the global horizontal radiation. Since this radiation data is
not contained in the DarkSky history, the global horizontal radiation was generated using the Zhang-Huang
model [16]. This correlation could then be used for sites in the area that did not have historical data.
𝐼ℎ= [𝐼𝑜∙sin(𝛽)∙{𝐶0 + 𝐶1 ∙𝐶𝐶
10 + 𝐶2 ∙ (𝐶𝐶
10)
2
+ 𝐶3 ∙(𝑇𝑑𝑏𝑛− 𝑇𝑑𝑏𝑛−3)+ 𝐶4𝜙}− 𝐶5]/𝑘
Io = The extraterrestrial solar radiation (W/m2)
Ih = Global Horizontal Radiation Intensity (W/m2)
= Sun’s altitude (Radians)
CC = Cloud cover in tenths
Tdb = Outdoor air dry bulb temperature at the current hour
Tdbn-3 = Outdoor air dry-bulb temperature at three hours previous
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 16
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
𝜙 = The relative humidity (%)
C1, C2, C3, C4, C5, and k are regression coefficients specific to the location
The estimation was originally developed for locations in China. Therefore, the team developed a
regression specific to the test sites based on historical observations for locatins in Idaho. The observed
global horizontal radiation intensity for all 8,760 hours was plotted against the estimation. Excel solver
was used with an evolutionary engine solver for non-smooth data with bounds on the constants near the
minimum and maximum observations from Zhang et al. The final regression is shown in Figure 1.2.
Figure 1.2 Results of correlation between horizontal radiation intensity estimate vs observed data for a site in Idaho based on TMY3 file.
The final correlation showed a correlation coefficient of 0.92 and a Root Mean Squared Error
(RMSE) of 78. This falls within the correlations developed by the Zhang-Huang model.
R² = 0.9171
0
100
200
300
400
500
600
700
800
900
1000
0 200 400 600 800 1000 1200
Es
t
i
m
a
t
e
d
S
o
l
a
r
R
a
d
i
a
t
i
o
n
(
W
/
m
2)
Observed Solar Radiation (W/m2)
Correlation of Global Horizontal Radiation Intensity vs
Estimate for a site in Idaho
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 17
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Table 1.1: Resulting regression coefficients for Boise, ID from Zhang-Huang model correlation
C0 0.671205
C1 0.04125
C2 -0.32346
C3 0.004766
C4 -0.0063
C5 27
k 0.91427
The global horizontal radiation is further broken down for EnergyPlus into its two components:
the direct normal and diffuse horizontal radiation. There are many models available for the decomposition
of the global horizontal radiation. The model that showed the best correlation with the recorded data for
diffuse horizontal radiation in Boise was the Watanabe model [17]. This is the same model used by Kwak
et al. [18]. With the diffuse horizontal and global horizontal known, the Perez model can be used to
determine the last remaining solar component [19].
𝐺𝑙𝑜𝑏𝑎𝑙ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛= 𝐷𝑖𝑟𝑒𝑐𝑡ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛+ 𝐷𝑖𝑓𝑓𝑢𝑠𝑒ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛
𝐷𝑖𝑟𝑒𝑐𝑡𝑛𝑜𝑟𝑚𝑎𝑙 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛=𝐷𝑖𝑟𝑒𝑐𝑡ℎ𝑜𝑟𝑖𝑧𝑜𝑛𝑡𝑎𝑙 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛
sin(𝛽)
Where is the solar altitude.
4.2.3 Formatting the Weather Histories
Once downloaded and derived, the weather history data must be manipulated to fit the file
format EnergyPlus models use. EnergyPlus models require a custom file format for weather inputs called
an EPW file or (EnergyPlus Weather file). The first seven lines of the EPW contain generic information
regarding the location and ground temperatures. After this, each line consists of 35 data points of weather
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 18
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
information for one hour. The weather data must be formatted in a very specific manner for it to be
compatible with the energy simulation. The run period field in EnergyPlus must be adapted to fit the time-
frame of interest. This run period includes the starting and ending month and day for the simulation as
well as the name of the weekday on which the simulation is to begin.
While the simulation can be started and stopped for any day, it must start and end at midnight.
The EnergyPlus simulation cannot be started at any random hour. As a consequence, the weather data
file must have a range of inputs equal to or longer than the run period. For example, if the energy model
requests a simulation from 1/25 to 1/27, the EPW must have data for at least 1/25 0:00 – 1/27 24:00. If
the weather observation is from 1/25 07:00 – 1/27 07:00, extra data must be added to the file for the
hours of 1/25 0:00 – 1/25 07:00 and from 1/27 07:00 – 24:00. This can be done by either repeating data
lines, or stitching together weather history and projections to extend the weather file to the full time
required. In this research, the weather histories were appended to older observed weather conditions to
create full data sets for the simulations. The simulation outputs were then filtered for the outputs during
the hours of interest. In order to produce this weather data set, the R-script from Rudis was adapted to
have a starting and ending date with a for loop that sends an API request for each 24-hour period during
that timeframe, appends all of the data, and downloads it to a comma delimited file. The researchers
developed an Excel workbook to automatically parse the observed data into the EPW order and derive
the solar fields based on the equations listed above. The Excel book uses macros to automatically export
a comma delimited file in the format of the EPW. Once the file extension is changed it can be used with
any EnergyPlus or OpenStudio model.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 19
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
5. MONITORING RESULTS
The research team collected data at each of the sites during both heating and cooling seasons. The
recordings were taken at each relevant surface so that the mean radiant temperature could be calculated.
The surface temperatures were relatively tightly grouped within a few degrees of each other, except for
the windows which showed very strong deviations from the rest of the surfaces.
Figure 1.3: Recorded surface temperatures at Office D
All offices had a PMV below 0, indicating a trend towards being too cool for occupants in office-wear
who are sitting and typing. Even during the fall, when the outdoor air temperature rose above the balance
point of the building, most offices showed a tendency to stay cold. One instance of this was at Office C
during October and is displayed in Figure 1.4.
65°F
70°F
75°F
80°F
85°F
90°F
6/22
0:00
6/22
12:00
6/23
0:00
6/23
12:00
6/24
0:00
6/24
12:00
6/25
0:00
6/25
12:00
6/26
0:00
6/26
12:00
6/27
0:00
6/27
12:00
6/28
0:00
6/28
12:00
SURFACE TEMPERATURES OFFICE D JUNE 2018
IDL-122 (window)IDL-127 (W.Wall)IDL-565 (ceiling)IDL-123 (N.Wall)
IDL-541 (floor)IDL-564 (central desk)IDP-005 (E.Wall)IDL-130 (S.Wall) ERROR
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 20
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.4: The comfort index (PMV) measured at an open office in October
Most measurements occurred over an extended time period. In order to identify daily trends in
the comfort, the team layered the daily calculations on top of one another as shown for a different site in
Figure 1.5.
Figure 1.5: The comfort index measured at a site with the weekday information layered in different colors
Layering the weekdays showed a general profile of chilly conditions in the mornings with
increasing comfort in the afternoon. Many offices showed trends well below the comfort standard of -
01020304050607080
-3
-2.5
-2
-1.5
-1
-0.5
0
OA
T
PM
V
OFFICE C PMV October 2017
OFFICE D PMV OCTOBER 2017
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 21
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
0.5. For example, many weekday readings at Office B dropped down below a PMV of -1 as shown in Figure
1.6.
Figure 1.6: Comfort measurements recorded in November 2017 at Office B
The trend of offices to be too cold continued even into the summer data collection period as
shown in Figure 1.7.
Figure 1.7: The Predicted Mean Vote calculated based on measurements taken at Office D
-2
-1.5
-1
-0.5
0
12:00 AM 3:00 AM 6:00 AM 9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM 12:00 AM
Office B cubicle winter PMV -weekdays
Fri - 10 Mon - 13 Tue - 14 Wed - 15 Thu - 16
Fri - 17 Mon - 20 Tue - 21 Wed - 22 Thu - 23
Fri - 24 Mon - 27 Tue - 28 Wed - 29
-2
-1.5
-1
-0.5
0
12:00 AM 3:00 AM 6:00 AM 9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM 12:00 AM
Office D PMV June 2018
Fri - 22 Sat - 23 Sun - 24 Mon - 25 Tue - 26 Wed - 27 Thu - 28
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 22
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
To present the data of Figure 1.7 in an alternative way, the same comfort information is listed in
Figure 1.8 as a calculation of the percentage of people that are predicted to be dissatisfied with the
thermal conditions in the office.
Figure 1.8: The calculated prediction of the PPD at Office D in June 2018
While the comfort index is trivial during the unoccupied hours of the building, the discomfort was
actually shown to trend higher during the occupied hours of the building from 8:00 AM – 5:00 PM. This
same trend could be seen at Office B in July. The shark-fin curves displayed in Figure 1.9 show how at
night, the temperature rises in the open office, bringing it closer to the desired PMV of 0, but as soon as
cooling starts again in the morning, it becomes uncomfortably cold.
0%
10%
20%
30%
40%
50%
60%
12:00 AM 3:00 AM 6:00 AM 9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM 12:00 AM
PERCENTAGE OF PEOPLE DISSATISFIED Office D 2018
Fri - 22 Sat - 23 Sun - 24 Mon - 25 Tue - 26 Wed - 27 Thu - 28
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 23
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.9: Measured comfort at Office B in July
5.1 Thermostat Reading Versus Other Instruments
One of the unanticipated results was that the average air temperatures recorded in the offices
deviated from the air temperatures recorded at the actual thermostat location. The ADA required
mounting height for a thermostat is between 48 and 54 inches. However the center of mass for the seated
height of most individuals in an office is significantly lower (between 23 and 25 inches) and is the height
at which most air temperature recordings were taken (excluding the floor and ceiling measurements). In
each case, the recording taken adjacent to the thermostat was higher than the average office air
temperature by at least 0.5oF, and sometimes as much as 2oF. The recordings at three different sites are
shown below in Figure 1.10, Figure 1.11, and Figure 1.12.
-2
-1.5
-1
-0.5
0
Office B July PMV
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 24
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.10: Recorded average air temperature versus recorded temperature next to thermostat at Office D
Figure 1.11: Recorded average air temperature versus recorded temperature next to thermostat at Office B
AIR TEMPERATURES AT OFFICE D JUNE 2018
AIR TEMPERATURES AT OFFICE B JUNE 2018
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 25
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.12: Recorded average air temperature versus recorded temperature next to thermostat at the Integrated Design Laboratory
5.2 Energy Modeling Results - Winter
Based on the recorded weather and setpoints, the energy models were able to re-create the
historical observations. Once verified, the research team used the energy models of the case study
buildings to test alternative control strategies. This allowed the team to essentially go back in time and
see what the effects of adjusting the HVAC operation would be by providing the models the same weather
patterns, but supply different control signals. The control strategies were developed directly in
EnergyPlus. The control strategies were simple in nature – similar to an air-based thermostat, but with a
weighting factor for the surface temperatures to better manage thermal comfort and energy savings.
While it was possible to save both energy and increase comfort by eliminating over-cooling during the
summertime, this was not the case during the winter. The office conditions indicated chilly occupants and
increasing the heating setpoint would increase energy consumption. The IDL research team looked at
changing occupant behavior by increasing the simulated clothing level to 1.0. This was effectively
AIR TEMPERATURES AT OFFICE F JUNE 2018
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 26
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
modeling everyone wearing a sweater at work. The baseline comfort based on standard office wear at
this site is shown in Figure 1.13.
Figure 1.13: Modeled comfort at Office E during observation based on typical office wear (trousers and long sleeves)
The modeled comfort at the Lewsiton office showed very cold conditions with conditions well
below the ASHRAE minimum of -0.5 when occupants were modeled as wearing conventional office outfits.
The effect of increasing the clothing level is shown in Figure 1.14.
Figure 1.14: Modeled comfort at office E during observation based on increased clothing levels (adding sweaters)
Increasing the clothing level brought the modeled PMV closer to the ASHRAE minimum of -0.5,
but many hours still dipped below this range. Therefore, within the energy model the setpoints were
OFFICE E
OFFICE E
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 27
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
adjusted from what was observed and increased to 72oF while maintining a clothing level of 1.0 (wearing
sweaters). The results are shown in Figure 1.15.
Figure 1.15: The modeled comfort at Office E based on increased clothing level and an increased heating setpoint of 72oF
Increasing the heating setpoint did increase the modeled heating load, but managed to bring the
predicted comfort metric within ASHRAE’s acceptable PMV range of -0.5 to + 0.5.
5.3 Energy Modeling Results - Summer
With an emphasis on saving energy while providing comfort, the cooling setpoints were increased
until the PMV during the occupied hours was closer to 0. In general, the ideal cooling setpoint to provide
a minimum PMV was 76oF. Setbacks were increased to 80oF to account for the surface temperatures and
comfort calculations. Increasing these setpoints improved predicted comfort and significantly reduced the
cooling load. The results of this altered control are shown in Figure 1.16.
OFFICE E MODELED HEATING SETPOINT 72oF, CLO =1.0
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 28
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Figure 1.16: The resulting comfort from increasing setpoints (black) versus observed controls (blue) at the Office A
For Office E i, the current operation showed operation in a comfortable range, but with a still slightly
negative PMV. Increasing the setpoint slightly up to 76oF was still able to save energy and keep the comfort
index closer to the ideal value of 0 as shown in Figure 1.17.
Figure 1.17: Predicted comfort modeled with proposed (black) setpoints versus observed setpoints (blue)
The energy impacts of increasing the setpoint vary based on the HVAC type and efficiency at each
site. In order to provide uniform comparisons, the models only estimated the amount of cooling load that
was reduced during the observed week. Actual energy savings from these loads may be up to four or five
OFFICE A
Office E
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 29
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
times lower depending on the EER of the cooling equipment and its part-load performance. However, the
models were able to quantify the kBtU of total cooling load that could be reduced by using the new
setpoints and setbacks and the percent reduction in load over the observed week that was simulated. The
results are shown for each site in Figure 1.18.
Figure 1.18: Cooling load reduction predicted for each site based on increasing cooling setpoint to 76oF to better meet comfort metrics.
6. DISCUSSION AND FUTURE WORK
The simulation results were used to estimate energy savings and comfort impacts of incorporating
surface temperatures into thermostat controls. The research identified the savings possible by shifting
controls from their current baseline settings to a strategy that is more in line with the demands of
occupant comfort. Based on the research findings, the development of the alternative controls could
prove useful and marketable to controls engineers, consultants, and building operators.
One of the key findings from this project was that when one considered all aspects of comfort
(including surface temperatures) all the offices studied were found to be uncomfortably chilly even during
the summer. ASHRAE mandates a minimum PMV of no lower than -0.5, but offices were found to be
regularly dipping below this and over-cooling the spaces. This finding was borne out anecdotally through
OFFICE A
OFFICE B
OFFICE C
OFFICE D
OFFICE E
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 30
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
multiple comments from the occupants and from the many electric space heaters that were observed
under employee desks.
In every situation measured, the air temperature recorded next to the thermostat was higher
than either the average or the median office air temperature measured closer to the office worker’s
location. This meant that employees were experiencing setpoints that were1-2oF colder than the
thermostat’s setpoint. This could be due to the height disparity between a seated worker’s center of mass
and a thermostat’s installed height. It may also be due to the warmth generated by the electrical current
running to the thermostat. The largest temperature disparities were found for “smart” thermostats with
colored graphics and more controls which have a higher power draw. More careful experimentation
would be needed to quantify the exact nature of this relationship, however the research did uncover
enough of a disparity that this could be one of the contributing factors to overcooling in offices.
The simulations showed that for most sites the chilly conditions could not be fully overcome by
changing occupant behavior (encouraging sweaters). Instead, most sites also required an increase in their
heating setpoints up to 72oF with the increased clothing level in order to meet the ASHRAE 55 comfort
standards. The research team did observe that the comfort increased in the offices throughout the day as
the surface temperatures in the space slowly rose from internal gains. Even in the winter, there may be
some financial gain to increasing the setpoints. For example, increasing the setpoints at Office E increased
the heating load by approximately 20% per week. For a small office of this size (10,000 ft2) this translates
to between 150 – 300 extra therms of gas per year. Even assuming an extra 300 therms of gas is required
each year to raise the setpoints, this may offset the use of many space heaters present in individual offices.
Purchasing these extra therms may cost the customer at most an extra $300 per year, but with over a
dozen employees in the office and Berkeley’s finding of a $330 productivity benefit per employee when
comfort conditions are satisfied, this increase in gas consumption should have a net financial benefit in
employee output.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 31
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
Increasing the cooling setpoint during the summer provides an even stronger financial incentive.
Increasing the cooling setpoint to 76oF could bring more office areas into compliance with ASHRAE’s
comfort standards. In addition, simulation results showed savings of 15 – 40% of the cooling load.
Assuming a generous EER of 17 for this cooling equipment, that would still result in at least 3 – 8% of
cooling energy saved at each of these sites during each of the weeks studied.
The incentive for this study was based on the effect surface temperatures have on comfort. The
research team found that by incorporating the surface temperatures to develop holistic comfort metrics,
the air temperature thermostat setpoints could be changed to increase the comfort of office employees
in the Pacific Northwest. This simple change resulted in significant savings and other studies suggest that
it would also increase employee happiness, wellbeing, and productivity. At the very least, engineers,
controls manufacturers, and building managers should re-visit the default heating and cooling setpoints
of 70 -74oF that are currently in place in most offices.
Operative controls could rely on several thermocouples as in this study which are inexpensive and
widely available sensors that can be easily integrated into a control sequence. Alternatively, infrared
cameras could also be used to map surface temperatures and provide better comfort metrics. As the
operative temperature control approach is adopted, the technology allows either building managers or
utility companies to provide incentives for this type of control if savings are verified. A simple control
scheme using operative temperatures will encourage efficient control setpoints that endure. The market
path may include the adoption of this control strategy by a controls company or as a new tool used by
consultants and building operators.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 32
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
7. BUDGET SUMMARY
These hours reflect only Avista’s contribution to this project and are not reflective of total project
investment by the research team, industry sponsors, or other university staff.
Personnel Hours estimate Description
FY17/FY18
Salaries: $11,684
Fringe: $1,052
Travel: $1,000
F&A: $6,910
Tuition $3,365
Total: $24,011
Indirect Costs
For this contract, UI-IDL was considered an on-campus unit of the University of Idaho with a federally
negotiated rate of 50.3%.
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 33
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
8. REFERENCES
Building and Environment,
Clima 2007 WellBeing Indoors
Building and Environment,
ASHRAE,
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 34
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
National Passive Solar Conference
Journal of Asian Architecture and Building Engineering,
Transactions, Architectural Institute of
Japan,
Energy and Buildings,
Building and Environment,
Journal of Building Performance Simulation,
Conference of International Building
Performance Simulation Association
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 35
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
International Building
Performance Simulation Association
Energies,
Staff_PR_29 - Attachment A
Integrated Design Lab | Boise 36
Managing for Efficiency Based on Operative Temperatures (Report 1708_01)
9. APPENDIX
2. WINTER DATA COLLECTION DATES
–
3. SUMMER DATA COLLECTION DATES
Staff_PR_29 - Attachment A