Energy End-Use Patterns in Full-Service Hotels: A Case Study

Marylynn Placet, Srinivas Katipamula, Bing Liu, James Dirks, and Yulong Xie
Pacific Northwest National Laboratory
Gregory Sullivan, Efficiency Solutions
Jim Walent, BF Saul, Inc.
Rebecca Williamson, InterContinental Hotel Group

ABSTRACT
The U.S. Department of Energy (DOE) recently initiated a program Commercial
Building Partnerships (CBP) to work with private-sector companies in the design of highlyefficient retrofit and new construction projects. Pacific Northwest National Laboratory (PNNL)
is conducting a project with a major hotel company to retrofit a full-service, large hotel with the
goal of reducing energy consumption by at least 30%. The first step of the project was an
intensive metering and monitoring effort aimed at understanding energy end use patterns in the
hotel. About 10% of the guest rooms (32), as well as circuits for most of the end uses in public
spaces (lighting, elevators, air handlers and other HVAC system components, and various
equipment), were equipped with meters. Data are being collected at 1- or 5-minute intervals and
downloaded on a monthly basis for analysis.
This paper presents results from the first four months of the monitoring effort, which
revealed energy end-use consumption patterns, variability of guest room energy use, daily load
curves, monthly variations, and other aspects of hotel energy use. Metered end-use data for
hotels at this level of detail are not available from any currently-available public sources. This
study presents unique information and insight into energy end-use patterns in the lodging sector
of commercial buildings and can also serve as a case study of a complex sub-metering project.

Background on the Project

The US DOEs CBP Program is providing technical assistance through several of DOEs
National Laboratories to support energy efficient retrofit and new-construction projects being
undertaken by about 20 private-sector companies (DOE 2010). The project goals are twofold: 1)
to design and build new buildings that are 50% more energy efficient than the ASHRAE
Standard 90.1-2004 building code requires, and 2) to retrofit existing buildings in a way that
achieves a 30% reduction in energy consumption. Lessons learned from these projects will help
the companies alter their design, construction, and building operation practices in the future.
PNNL is working with a major hotel company on both a new construction and retrofit project.
The retrofit project, which is the subject of this paper, is a twelve-story, full-service hotel with
over 300 guest rooms and gross floor area of 212,000 ft (see Figure 1). Originally constructed
in 1968, the hotel underwent a light renovation nearly 10 years ago. This minor renovation was
primarily cosmetic, and there has been very little energy focus at this hotel until this project
began. Building energy systems are somewhat outdated. The hotel has an energy management
system (EMS) that is not connected to all building energy systems (e.g., the newly-installed
chiller, exhaust fans) and is not being used in a fully-functional way. Also, single-pane windows
and uninsulated exterior masonry walls result in a thermally-inefficient building envelope, and

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-310

most of the packaged terminal heat pumps (PTHP) serving the guest rooms are 7-years old or
more (though they are gradually being replaced). There was no energy benchmarking or energy
audit completed prior to commencement of this project.
Figure 1. Aerial Photo of the Hotel

The first step of the project was to measure the baseline energy use and conduct a
detailed study of the end-use energy consumption patterns in the hotel. Some information on
energy end-use in hotels is available from sources such as the Commercial Building Energy
Consumption Survey conducted by the Energy Information Administration (EIA 2005), Energy
Star (EPA, 2010) and the California Commercial End-Use Study (CEC 2006). However, these
sources are primarily aimed at benchmarking and are typically based on whole-building energy
consumption data or modeled/estimated energy loads by end use. One published study
monitored lighting in hotel guest rooms (Page and Siminovitch, 1999). The metering effort
described here was designed to provide a detailed and accurate picture of energy consumption at
this particular property to support building energy simulation modeling and the development and
analysis of energy efficiency measures, leading to a 30% or more improvement in energy
efficiency. However, the end-use consumption data reported in this paper are also generally
useful for describing end-use patterns in the large hotel sector, because the guest room
equipment and various other features of the hotel are typical of large hotels in the United States.
The property is a full-service hotel with sit-down restaurant, coffee shop, on-site laundry,
offices, and conference/banquet rooms, making it a challenging project. Hotels have a wide
diversity of energy uses, with spaces that may or may not be occupied on any given day, and
guest rooms that are occupied by different guests with varying habits. As a result, the metering
data collection for this project has been considerably more extensive and complicated compared
to other projects in the CBP Program.
Initial analysis of the 2009 electricity interval data from the utility (Dominion Power)
shows a relatively flat daily load curve in all seasons across all days of the week (Figure 2). The
lack of variation between weekdays and weekends is not surprising, given the hotel has a very
high occupancy rate and serves both tourists (including weekend stays) and business guests

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-311

(typically during the week). From these data, it appears that this hotel has almost 400 kW of
base load. The peak cooling load is about 200 kW (see Figure 2, August) and the hotel has
overall higher electricity use in the winter due to use of electric resistance heat1 in the packaged
terminal heat pumps (PTHPs) serving the guest rooms and the duct heaters serving common
spaces.
Figure 2. Daily Electricity Consumption (kW) based on Utility Interval Data in Selected
Months

In addition, the hotel uses natural gas for service water heating (a significant hotel load),
as well as kitchen appliances and laundry equipment. Table 1 shows the monthly energy
consumption for both electricity and natural gas in the year of 2009. At the site, electricity

During the metering period, many of the PTHPs were heating in resistance mode due to a software control issue.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-312

consists of 73% of the total energy consumption. Converting to source energy use, electricity
consists of 89% of total Btus. The metering and monitoring effort was aimed at breaking down
these totals by end use.
The effort to understand the breakdown of energy by end use load involved: (1)
extensively metering various circuits and pieces of equipment in the hotel, (2) estimating loads
using information about equipment capacity and use, and (3) building energy simulation
modeling and calibration using EnergyPlus. This paper will discuss items (1) and (2); the energy
simulation modeling is underway and will be described in a future paper.
Table 1. Monthly Energy Consumption, 2009
January

Site Electricity Consumption

kWh
Billion Btu
563,995
1.92

February

394,642

March

342,010

April

258,781

May

273,499

June

318,592

July

350,831

August

391,549

September

276,165

October

280,615

November

316,914

December

506,513
4,274,106

Month

Total

Site Natural Gas Consumption

Therms
Billion Btu
4,118
0.41

1.35

4,876

0.49

1.17

6,103

0.61

0.88

5,441

0.54

0.93

5,390

0.54

1.09

4,773

0.48

1.20

4,313

0.43

1.34

3,829

0.38

0.94

3,360

0.34

0.96

4,447

0.44

1.08

3,995

0.40

1.73

4,016

0.40

14.58

54,661

5.46

Electricity Metering Approach and Results

Approach
Since guest room loads were deemed important, approximately 10% of the rooms (32
rooms) were chosen for monitoring and evaluation. PNNL and its subcontractors installed pluglogging electricity meters (Educational Electronic Devices Watts-Up) to measure electricity
consumption of plug loads and portable lighting fixtures, including the desk lamp with
convenience outlet, floor lamp, two table lamps, the refrigerator, and the TV. In addition, Onset
Computer (Hobo) light intensity data loggers were installed at the bathroom lights and wall
sconces to determine the on-hours of these hard-wired fixtures. The PTHPs were instrumented
with Hobo loggers measuring amperage draw, supply/return air temperatures, and relative
humidity. To ensure metering accuracy, a check-sum approach was used. For the guest room
plug loads the check sum metering was at the electrical panel where the individual breaker
serving each room was monitored. For the PTHPs, the check sum was at the sub-panel serving
the PTHPs. In both cases the panel-level monitoring was done using a combination of a pulseoutput watt transducer (Continental Control Watt-Node) and a pulse collecting data logger
(Madgetech Pulse 101). We also installed the same combination of watt transducers and pulse
loggers on the two chillers, rooftop exhaust fans, the elevators and elevator HVAC, make-up

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-313

rooftop unit (to provide outdoor ventilation air to guest floor hallways), the three points (preheat, re-heat, and fan motor) for the nine air handing units (AHUs), and various kitchen, laundry,
and other equipment circuits. For the lighting, we metered circuits serving various common
areas and conference/ballrooms with a combination of Hobo data loggers and current
transformers. In total, the metering devices include 96 watthour meters, 158 plug power meters,
91 temperature measurement devices, 32 relative humidity measurements, 109 current
measurement devices, and 64 lighting level measurements for a total of 550 measurement
devices. The installation was complete and tested on September 25, 2009. Data has been
downloaded on a monthly basis.
Metering Results: End Use Loads
Figure 3 shows daily consumption for October 2009 for guest rooms [chart (a)], hotel
heating, ventilation, and air conditioning (HVAC) systems [chart (b)], and other loads [chart (c)].
The charts also show average daily outdoor air temperature. Note that, for the purpose of
developing an estimate of total load in the hotel, the data for the 32 guest rooms and guest
hallway equipment (ice makers and vending machine) were scaled using guest occupancy rates
in the 32 rooms versus the hotel as a whole. The PTHP load in Figure 3 (a) is highly correlated
with average outside air temperature (R-squared is 92%). As currently configured, the balance
point for the guest rooms is approximately 55F; i.e., at 55F, the minimum PTHP consumption
occurs. Temperatures lower 55 result in significant heating loads, and at temperatures greater
than 55F, PTHP consumption increases due to cooling. The HVAC loads shown in Figure 3 (b)
for the common areas in the hotel, including duct heaters, air handling units (AHU), and rooftop
units (RTU), also indicate a high correlation between average outside air temperature and heating
and cooling consumption. Here the balance temperature appears to be closer to 60F, as the
majority of the HVAC in this portion of the building is running continuously with 100% outside
air. The other loads shown in Figure 3(c) are not correlated with outdoor temperature. Figure
4 shows monthly energy consumption by end use for the four metered months.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-314

Figure 3. Daily Metered End Use Loads versus Outdoor Temperature (T out) in October
(a) Guest Room Loads
Guest Room Daily Energy Use and T out
5,000

80

4,500

70

4,000

Average Outside Air Temp

60

Energy Use (kWh/day)

3,500
50

3,000
2,500

40

2,000

30

PTHP
Refrig
Lights/TV/Other
T out (Avg)

1,500
20
1,000
10

500
0

0
1

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Day of Month

Figure 3 (continued). Daily Metered End Use Loads versus Outdoor Temperature (T out)
in October
(b) HVAC Loads in Common Spaces
Hotel HVAC Daily Energy Use and T out
80

2,500

Energy Use (kWh/day)

60

50

1,500

40
1,000

30

20

Average Outside Air Temperature

70
2,000

AHU/Ventillation
Resistance Heat
Chiller/AC
RTU
T out (Avg)

500
10

0
1

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Day of Month

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-315

(c) Other Loads

Hotel "Other" Daily Energy Use and T out
80

700

Energy Use (kWh/day)

60
500
50
400
40
300
30
200
20
100

Average Outside Air Temperature

70

600

Lights
Kitchen
Vending/Ice
Laundry
Elevator
T out (Avg)

10

0
1

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Day of Month

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-316

Figure 4. Monthly Electricity End-Use for the Four Metered Months (Combined)
Distribution of End Use Energy (ave. kWh/day) Based on Metered Data
(Oct 2009 - Jan 2010)
Vending/Ice 218 Laundry 44

Elevator 141

Kitchen 233
Lights 467
RTU 11
Chiller/AC 240

PTHP 4,016

Resistance Heat 2,160

AHU/Ventillation 653

Refrig 196
Lights/TV/Other 193

Details on PTHP Energy Use

The PTHP energy consumption in 32 guest rooms was monitored using current
measuring devices at 1-minute intervals. The current data was converted to power by
multiplying with the one-time voltage measurement (277 volts and an assumed power factor of
0.95). The clouds of dots in Figure 5 shows the average daily power consumption (average
Wh/h) as a function of the average daily difference between the outdoor and return air
temperatures for selected rooms. The red lines in Figure 5 are the Loess2 curve fits. Some of the
32 rooms were recently upgraded with new PTHPs, which are slightly more efficient than the
older units. Figure 6 shows the comparison of Loess fits for all the 32 metered rooms. Note that
the old units (blue lines) generally have higher slope than the new units (red lines), indicting new
units are more efficiency than the old units.

Loess in a nonparametric method for estimating local regression surfaces.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-317

Figure 5. PHTP Average Power Consumption as Function of Average Daily Difference in

Outdoor and Return Air Temperatures

Figure 6. Comparison of Loess Fits for New and Old PTHP Units

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-318

Unmetered Load Analysis

Figure 7 shows total daily average electricity consumption at the whole-building meter
versus the sum of the metered loads in one month (December 2009). Over the entire period, on
average, the individual meters capture about 62% of the total electricity consumption, but the
percentage varies from 50% to 83% on any given day.
While the monitored electricity use follows the building level revenue meter very well,
the analysis shows a significant unaccounted-for electricity use. To better understand this
unmetered use, the following three analyses were conducted:

Unmetered use as a function of occupancy

Unmetered use as a function of outdoor air temperature
Regression of both revenue metered and project monitored usage with outdoor air
temperature

Independently, the unmetered loads do not show a strong correlation with either
occupancy or outdoor air temperature. Through regression analysis, we correlated both the
utility-supplied and project-monitored data with outside air temperature. The premise behind
this analysis is that if one data set has an appreciably different relationship to outdoor air
temperature, the resulting regression lines will not be parallel. As shown in Figure 8 the lines
are, by-and-large, parallel. This consistency highlights the relative constant nature of the
difference indicating this difference is not weather variant and is likely driven by miscellaneous
base loads.
Figure 7. Whole-Building Utility Data Versus Sum of Metered Data for December
Utility - Metered Data Comparison
December 2009
25,000

kWh per Day

20,000
Utility Data
15,000
Meter Data
10,000

5,000

1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031

Day of the Month

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-319

Figure 8. Regression of Utility and Metered Data with Outdoor Air Temperature

The estimation of unmetered loads began with a thorough review of audit notes, various
equipment lists, and installed metering points. Once all major loads were identified, they were
reconciled as being metered or not; those not metered were listed for inclusion in the estimation
exercise. The five main categories of unmetered loads are as follows:

Pumping. Includes all domestic cold water pumps, chilled water and condenser water
pumps, domestic hot water recirculation pumps, and the pool pump. Chiller, condenser
water, and pool pump energy use were scaled based on metered equipment run times;
domestic water pumps were calculated based on schedule.
HVAC. Includes unmetered exhaust fans and a variety of electric space heaters. It is
likely that this group has additional, yet-to-be discovered, electrical heating loads possibly including other space heat and pre/reheat coils. (Note: the pool is unheated.)
Lighting. Includes all unmetered lighting loads categorized using a detailed lighting
audit.
Kitchen. Unmetered kitchen energy use was derived based on an audit done by PNNLs
kitchen subcontractor. To prevent double counting, these values were input and then
reduced by the metered kitchen loads.
Miscellaneous Equipment. Includes miscellaneous appliance and equipment at the onsite Starbucks, the restaurant bar, the back office, and in the engineering areas.

Once identified, these loads were researched for rated energy use and schedule to
estimate whole building impact. The results of this effort are shown in Figure 9.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-320

Figure 9. Unmetered Load by End-Use Category

Note that, in this figure, the top series represents the total unmetered load, and the series
below are cumulative, such that they build up the estimated unaccounted load. The remaining
unaccounted-for load increases with time, indicating a potential for missing heating/ventilation
loads, as the outdoor temperature was generally decreasing over the analysis period.
Given the magnitude and complexity of energy use in this hotel, unaccounted-for electric
loads are expected. The above analysis reduced the average aggregate unaccounted-for loads
from 37.9% to 16.4%. The resulting breakdown of energy use is shown in Figure 10. (Note:
average over the four months was 13,285 kWh/day.) Further analysis is being conducted to
reduce the percentage of unaccounted-for energy use.

Natural Gas Consumption

Natural gas use at the hotel is confined to three areas:

The penthouse where gas is used in three boilers and one water heater to generate
domestic hot water.
The laundry area where gas is used in two water heaters to generate hot water for clothes
washers, in three clothes dryers, and in one ironer (known in industry as a flat iron).
The kitchen where gas is used in a variety of cooking and warming equipment.

The challenge of this activity was disaggregating the various gas using devices listed
above and developing these into relative shares of the whole-building use. While end-use
metering of each device was desirable, the cost, complexity, and relative intrusiveness of gas
metering made this impractical.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-321

Gas is metered at the building service entrance via a utility-owned rotary style gas meter.
Gas bills from this meter were reviewed. In November 2009, the utility meter was retrofit with a
pulse output device with full safety isolation. These pulses were now collected by a data logger
at 5-minute interval and downloaded on monthly basis.
Figure 10. Metered plus Estimated Electricity End-Use for October through January
Distribution of End Use Energy (ave. kWh/day) Based on Metered and
Estimated Un-Metered (Oct 2009 - Jan 2010)
Unaccounted For 2,124

Misc. 437

PTHP 4,016

Elevator 141
Laundry 44
Vending/Ice 218
Kitchen 997
Refrig 196
Lights/TV/Other 193

Lights 959

AHU/Ventilation 907

RTU 11
Chiller/AC 240
Resistance Heat 2,804

To estimate gas use for the domestic hot water and laundry systems, a number of proxy
measurement techniques were employed.
Predominantly, these relied on temperature
measurements of products-of-combustion either in exhaust stacks (boilers, water heaters, iron) or
combustion chambers (clothes dryers). In addition, flow (via a non-intrusive ultrasonic flow
meter) and temperature measurements were made in the Penthouse to assess gas used in
domestic hot water generation.
By its nature, proxy metering does not measure the variable of interest, rather a surrogate
for that variable. As such, the potential for inaccuracy can be significant. To minimize these
potential inaccuracies and to improve confidence in the result, secondary measurements or
calculations are always recommended. For this activity, all proxy measurements were verified
with secondary calculations typically based on manufacturer-provided energy-use intensities.
Gas use for the kitchen was calculated by a team from the Halton Inc. Figure 11 presents
the proxy measurement results by major end-use. Service hot water accounts for 51% of the gas
use, with kitchen and laundry accounting for about 25% each.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-322

A comparison was completed whereby the sum of the parts gas use (i.e., domestic hot
water, laundry, and kitchen) was compared to the whole (i.e., the average annual daily gas
use). The result was that the former was calculated to be 6.3% greater than the latter. This
difference is well within the expected accuracy of a proxy analysis.
Figure 11. Average Daily Natural Gas Use by End Use

Conclusions
The work to date in this project, i.e., to measure, monitor, and estimate end-use energy
consumption in a 1970-era full-service hotel in the Middle Atlantic region, was a challenging
undertaking and provided a rich source of information about the major energy end uses in hotels.
A few major conclusions from the study include:

Heating of the guest rooms and the public areas is by far the largest consumer of
electricity, accounting for almost 60% of total electricity consumption, in the autumn and
winter months in this large hotel in the Washington, D.C. area, and would likely the
dominant seasonal load in any hotel in a similar climate zone. For new construction
projects, more efficient approaches for heating the guest rooms, instead of PTHPs, should
be evaluated. In a retrofit situation like the project described here, replacement of PTHPs
with the most efficient products available on the market is warranted and is likely to have
a significant impact on total energy use. The fact that the PTHPs were not operating
properly in this hotel and the energy management system is not fully functional probably
resulted in a somewhat higher heating load than would be seen in newer hotels with fully
functional energy management systems. For heating of public spaces, alternatives to
electric resistance heat and better control systems will be examined in the next phase of
this project.

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-323

Not surprisingly, energy consumption in both PTHPs and the other HVAC equipment in
the hotel is correlated with outside temperature. This study shows that the correlation is
very strong.
The rest of the loads in this hotel (and most hotels) are diverse, making the development
of an energy efficiency improvement strategy complicated. The kitchen consumes about
8% of the electricity and a quarter of the natural gas, hence about 10% of total energy use
in the winter (the percentage would be higher in the summer). Unlike office buildings or
some other commercial building types, lighting loads in a large hotel account for a fairly
modest percentage of electricity consumption (about 7% in the winter and a somewhat
higher percentage in the summer). Water heating consumes somewhat less than lighting
at about 6%.
Despite a rigorous metering campaign and subsequent analysis using calculations and
proxy measures, we could account for only 84% of the electricity load in this hotel. This
underscores the need for better understanding of miscellaneous electricity loads in
commercial buildings, including hotels. This issue will be further studied in subsequent
work.

References
CEC. 2006. California Energy End-Use Survey Consultant Report. CEC-400-2006-005.
California Energy Commission, Sacramento, CA.
DOE. 2010. Commercial Building Partnerships Program, U.S. Department of Energy,
Washington,
DC.
Last
accessed
on
May
11,
2010
at
http://www1.eere.energy.gov/buildings/commercial_initiative/building_partnerships.html
EIA. 2005. Commercial Buildings Energy Consumption Survey 2003. Energy Information
Administration, U.S. Department of Energy, Washington, D.C. Last accessed in May 11,
2010 at http://www.eia.doe.gov/emeu/cbecs/contents.html.
EPA 2010. ENERGY STAR for Hospitality and Entertainment. U.S. Environmental
Protection Agency, Washington, DC.
Last accessed on May 11, 2010 at
http://www.energystar.gov/index.cfm?c=business.bus_hospitality_entertainment
Page and Siminovitch. 1999. Lighting Energy Savings Opportunities in Hotel Guestrooms.
LBNL-44448. Lawrence Berkeley National Laboratory, Berkeley, CA. Last accessed on
May 11, 2010: http://gaia.lbl.gov/btech/pubs/pubs.php?code=Lighting%20Systems

2010 ACEEE Summer Study on Energy Efficiency in Buildings

3-324