Metering Best Practices Applied in the National Renewable Energy Laboratory’s Research Support Facility: A Primer to the 2011 Measured and Modeled Energy Consumption Datasets

NREL is a national laboratory of the U.S. Department of Energy, Office of Energy

Efficiency & Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.

Contract No. DE-AC36-08GO28308

Metering Best Practices Applied

in the National Renewable

Energy Laboratory’s Research

Support Facility

A Primer to the 2011 Measured and

Modeled Energy Consumption Datasets

Michael

Sheppy, Aaron Beach,

and Shanti Pless

Technical Report

NREL/

TP-5500-57785

April 2013

NREL is a national laboratory of the U.S. Department of Energy, Office of Energy

Efficiency & Renewable Energy, operat

ed by the Alliance for Sustainable Energy, LLC.

National Renewable Energy Laboratory

15013 Denver West Parkway

Golden, Colorado 80401

303-275-3000 • www.nrel.gov

Contract No. DE-AC36-08GO28308

Metering Best Practices Applied

in the National Renewable

Energy Laboratory’s Research

Support Facility

A Primer to the 2011 Measured and

Modeled Energy Consumption Datasets

Michael Sheppy, Aaron Beach,

and Shanti Pless

Prepared under Task No. UA10.4000

Technical Report

NREL/

TP-5500-57785

April 2013

NOTICE

This report was prepared as an account of work sponsored by an agency of the United States government.

Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty,

express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of

any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately

owned rights. Reference herein to any specific commercial product, process, or service by trade name,

trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation,

or favoring by the United States government or any agency thereof. The views and opinions of authors

expressed herein do not necessarily state or reflect those of the United States government or any agency thereof.

Available electronically at http://www.osti.gov/bridge

Available for a processing fee to U.S. Department of Energy

and its contractors, in paper, from:

U.S. Department of Energy

Office of Scientific and Technical Information

P.O. Box 62

Oak Ridge, TN 37831-0062

phone: 865.576.8401

fax: 865.576.5728

email: mailto:reports@adonis.osti.gov

Available for sale to the public, in paper, from:

U.S. Department of Commerce

National Technical Information Service

5285 Port Royal Road

Springfield, VA 22161

phone: 800.553.6847

fax: 703.605.6900

email: [email protected]world.gov

online ordering: http://www.ntis.gov/help/ordermethods.aspx

Cover Photos: (left to right) PIX 16416, PIX 17423, PIX 16560, PIX 17613, PIX 17436, PIX 17721

Printed on paper containing at least 50% wastepaper, including 10% post consumer waste.

Acknowledgments

The authors would like to thank the U.S. Department of Energy for its support. This report was

prepared by the National Renewable Energy Laboratory’s Center for Electricity, Resources, and

Building Systems Integration under Task Number UA10.4000.

The authors would like to acknowledge the input of Benjamin Kroposki and Stuart McMillan

during the planning of this report. We also thank Adam Hirsh and Luigi Gentile Polese for their

reviews.

Nomenclature

DOE U.S. Department of Energy

ERE energy reuse effectiveness

EUI energy use intensity

FEMP Federal Energy Management Program

HVAC heating, ventilation, and air conditioning

M&V measurement and verification

NILM non intrusive load monitoring

NREL National Renewable Energy Laboratory

NZE net zero energy

PUE power utilization effectiveness

PV photovoltaic

RSF Research Support Facility

iii

Contents

Acknowledgments............................................................................................................................ i

Nomenclature .................................................................................................................................. ii

Contents ......................................................................................................................................... iii

1.0 Introduction ................................................................................................................... 1

2.0 Designing an Energy Metering System for a Commercial Building ............................ 2

2.1 Step 1: Establish Metering Objectives Upfront ........................................................... 2

2.2 Step 2: Decide How “Detailed” Your Metering Should Be ........................................ 3

2.3 Step 3: Select Meters ................................................................................................... 3

2.4 Step 4: Select an Appropriate Sampling Interval ......................................................... 3

2.5 Step 5: Install Meters and Verify Meter Output .......................................................... 4

2.6 Special Considerations for New Construction and Extensive Retrofits ....................... 4

2.7 Special Considerations for Existing Buildings ............................................................. 4

3.0 Analyzing Metered Data to Draw Conclusions About Energy Performance ............... 6

3.1 Step 1: Download the Data .......................................................................................... 6

3.2 Step 2: “Clean” the Data .............................................................................................. 6

3.3 Step 3: Aggregate the Data .......................................................................................... 6

3.4 Draw Conclusions About Energy Performance ............................................................ 7

3.4.1 Compare Metered Data to an Energy Model ................................................................ 7

4.0 Discussion of the Accompanying 2011 RSF Dataset ................................................... 8

5.0 References ..................................................................................................................... 9

Appendix A Link to RSF Datasets........................................................................................... 10

1.0 Introduction

Modern buildings are complex energy systems that must be controlled for energy efficiency.

The Research Support Facility (RSF) at the National Renewable Energy Laboratory (NREL) has

hundreds of controllers -- computers that communicate with the building’s various control

systems -- to control the building based on tens of thousands of variables and sensor points.

These control strategies were designed for the RSF’s systems to efficiently support research

activities. Many events that affect energy use cannot be reliably predicted, but certain decisions

(such as control strategies) must be made ahead of time. NREL researchers modeled the RSF

systems to predict how they might perform. They then monitor these systems to understand how

they are actually performing and reacting to the dynamic conditions of weather, occupancy, and

maintenance.

This report describes how the B and C wings (a total of 220,000 ft

) of the RSF were equipped

for monitoring its energy use, how its systems were modeled to forecast their energy use, and

how those systems actually performed in 2011. An accompanying dataset presents the hourly

energy use of the RSF’s systems aggregated by end use into the following categories: cooling,

heating, mechanical, lighting, plug loads, data center, and photovoltaic (PV) generation. The

forecasted values for these systems are included along with the actual performance. Finally, a

dataset matching pertinent weather information to each hourly data point is included for

reference. For more information on the RSF, visit the Research Support Facility Web site.

2.0 Designing an Energy Metering System for a Commercial Building

This section discusses metering best practices and highlights the metering approach that NREL

has taken in the RSF. A rationale is provided for each aspect of the metering system design. For

a comprehensive guide on metering in commercial buildings, please refer to the Federal Energy

Management Program’s (FEMP) Metering Best Practices guide [1]. Torcellini et al. studied six

high performance buildings, and found that metering leads to better management and improved

energy performance [2].

2.1 Step 1: Establish Metering Objectives Upfront

Making a metering plan is perhaps the most important aspect of designing a metering system,

and at the same time, the most commonly overlooked. Figure 2-1 is a graphical representation of

the metering plan that was implemented in the RSF’s data center to be able to calculate Power

Utilization Effectiveness (PUE) [3]. This graphic helped multiple parties at NREL understand

the metering plan and buy in to the energy performance goals of the RSF.

Figure 2-1 Diagram of the power meters in the RSF data that allow PUE calculation (Credit: Marjorie

Schott, NREL) [3]

When a metering system is not well planned, it leads to more laborious and computationally-

intensive post-processing of collected metered data. It can also result in insufficient metering: a

situation where there are not enough metered points to calculate key energy metrics. Insufficient

metering makes it difficult or impossible to effectively conduct continuous commissioning,

measurement and verification, and energy benchmarking. If some energy end uses (HVAC,

lighting, plug loads, heating, cooling, etc.) are incompletely metered, there may be insufficient

data to spot improvements or degradations in whole-building and building system performance

over time.

The RSF was designed to meet strict energy goals. One goal was to achieve a whole-building

demand-side energy use intensity (EUI) requirement of 25 kBtu/ft

·yr . Another goal of the

building was to be net zero energy (NZE). An NZE building consumes as much energy as it

produces (via its PV system) annually. Therefore, the first step in designing the metering system

for the RSF was to recognize that NREL would want to be able to quantify annual whole-

building EUI and NZE status with metered data.

2.2 Step 2: Decide How “Detailed” Your Metering Should Be

The more sensors that are deployed in a building, the more “detailed” a metering system is

considered to be. A metering system can be very “macro,” using only one whole-building

electricity and one whole-building gas utility meter. Conversely, a metering system can be very

“micro,” having a meter on every electrical circuit and at every piece of gas-consuming

equipment. Typically, the number of sensors that are deployed in a building is determined by

budget. One pitfall that many metering systems encounter is that they aren’t designed to monitor

energy end-uses or certain key equipment because they aren’t called out in the goals of the

project -- given the expense of a detailed system. However, facilities- and energy-managers can

use this level of metered data to investigate and track down equipment that isn’t working

properly.

The metering system in the RSF strikes a balance between macro and micro. There is sufficient

metering to be able to fully account for all of the energy end uses, and to individually monitor

the performance of some of the building’s key equipment (such as the transpired solar collector).

Plug load metering is an example of how the RSF strikes a balance between macro and micro.

The RSF was not set up to meter each plug load circuit separately (a micro level of metering).

Instead, all plug load circuits are organized onto dedicated panels and each panel is metered.

This provides a sufficient level of detail to quantify whole-building: plug loads, EUI, and NZE

status. Thus, the energy goals of the building are able to be met without over-metering or under-

metering.

2.3 Step 3: Select Meters

A vast selection of metering equipment is currently on the market. Metering systems vary

widely in their accuracy, price, ease of installation, and advanced/smart features. The energy

meters in the RSF have an accuracy of + 5% or better. The energy meters in the building are all

appropriately sized for the electrical circuit or other variable being monitored. For example, a

current transducer (CT) that is rated for 32 Amps would be inappropriate for monitoring a circuit

that draws a current of 100 Amps. Likewise, an ultrasonic flow meter that is rated for up to a 6-

in. diameter pipe would be inappropriate for monitoring the flow in an 8-in. diameter pipe.

2.4 Step 4: Select an Appropriate Sampling Interval

Sampling interval refers to how often a meter takes a reading. When measuring power and

energy consumption (typically kilowatts and kilowatt-hours, respectively), the sampling interval

must strike a balance between: (1) being fast enough to capture key peaks and valleys of system

operation throughout the day; and (2) not being so fast that data analysis and storage are

cumbersome because of the high volume of data being collected. In general, energy meters

should be able and configured to take readings at least every 15 minutes.

The electric energy meters in the RSF take readings every 5 minutes. Weather conditions are

monitored by an array of sensors that take readings every minute. The collected data are

typically downsampled (through averaging) to 1-hour intervals for ease of data analysis.

2.5 Step 5: Install Meters and Verify Meter Output

In existing buildings, the installation of meters should be coordinated so that daily building

operations are not disrupted or impeded, if possible. In new construction, the installation of

meters should be completed before the building opens. Every measure should be taken to keep

the metering system out of sight of the building occupants. Unauthorized access to the metering

system and its components should be prohibited to ensure and maintain data integrity and

preserve safety.

Once installed, the metering system in the RSF was properly calibrated according to the

manufacturer’s recommendations and the output of the system was verified and validated to

ensure it is measuring and recording accurately. The metering system is maintained and

operated according to the manufacturer’s recommendations, and will likely require periodic

calibrations.

The metering system in the RSF is designed to ensure monitoring quality. In many cases, faults

in the metering system can be identified via an energy balance. For example, the sum of the

energy end-use meters should equal the sum of the main electricity and gas meters (within the

accuracy range of the meters). The RSF metering system facilitates remote monitoring of all the

measured points back to a central data logger that enables users to remotely access the data in

real time and analyze it for faults.

2.6 Special Considerations for New Construction and Extensive Retrofits

In new construction and extensive retrofits, it is advisable to set up the electrical infrastructure

such that the energy end uses are aggregated onto their own dedicated electrical panels; this is

typically cost prohibitive in a minor retrofit. For instance, in the RSF, there are electrical panels

dedicated to only: plug loads, lighting, mechanical, and the data center. This approach makes

metering much simpler because an entire panel can be metered with just one meter and an entire

energy end use can be captured. Not taking this approach requires many meters to be used,

which may need to be added and/or subtracted to isolate each energy end use. For example, if

Panel X has all the building’s plug loads on it and one lighting circuit (say on Circuit Y), the

plug load energy end use for the building would be Meter

Panel X

– Meter

Cicuit Y

. Therefore,

organizing electrical circuits onto panels that are dedicated to just one energy end use reduces

upfront capital costs for the metering system and the required data post-processing.

2.7 Special Considerations for Existing Buildings

It is a common practice to add more electrical circuits to a building as they are needed over time.

Also, circuits may be repurposed for a completely different energy end use during a retrofit. This

adds an extra layer of complexity when designing a metering system. When designing a

metering system for an existing building, two key items must be verified: (1) meters are being

installed on the correct circuits; and (2) all the energy consumption for a given building system

or energy end use will be captured by the circuits that are metered. There are several steps to

verify these two key items:

1. Locate the electrical 1-line diagram and panel schedules for the building.

2. If possible, meet with the electrician or engineer who created the 1-line diagram to verify

whether changes have been made to the electrical infrastructure since the diagram was

created.

3. If the electrician or engineer is absolutely certain that no changes have been made, base

the metering system design on the 1-line diagram and panel schedules.

4. If changes have been made, use circuit tracing to verify, at a minimum, the new circuits

that have been added.

5. Update the 1-line diagram and panel schedules and base the metering system design on

these documents.

An alternative approach to circuit tracing is nonintrusive load monitoring (NILM). This

technology uses metering at the branch circuit level. NILM relies on a library of equipment

electrical “signatures” to identify loads that are being fed through a given branch circuit. One

advantage to this technology is that it does not require circuit tracing, and save significant time.

One disadvantage is that not every load may be in the NILM library, so some loads will be

unidentified. This makes it difficult or impossible to fully account for every energy end use in

the building.

3.0 Analyzing Metered Data to Draw Conclusions About Energy

Performance

Once the metering system is installed and calibrated, the next steps are to collect and analyze the

metered data to assess and draw conclusions about energy performance.

3.1 Step 1: Download the Data

The first step is to download the data. This step may be easy or difficult, depending on the

manufacturer of the metering and data logging equipment. Metered data from the RSF are

downloaded via a Ruby script in comma separated values (CSV) format. Other file formats will

work, but CSV is usable by many programs across multiple platforms.

3.2 Step 2: “Clean” the Data

No matter how carefully a metering system is designed and installed, the metered data will have

errors, which may include data stream cut-outs resulting in missing data, extreme meter spikes,

missing or overlapping timestamps, and erroneous values (such as lighting circuits reporting

negative power values). If a meter is regularly producing erroneous data, it should be re-

calibrated or replaced as necessary.

The 2011 RSF dataset that accompanies this report has already been “cleaned” (the errors have

been removed). Keep in mind that “clean” data are not the same as “perfect” data. Engineering

judgment was used to replace missing or erroneous data. The most commonly used data cleaning

techniques that were used on the 2011 RSF dataset follow:

• If 1 hour or less of data was missing, simple, linear interpolation was used to fill in data.

• If more than 1 hour of data was missing, data from a typical hour during that same time-

of-day or day-of-week were used to fill in the gap.

• Erroneous values were corrected as logically as possible. For example, if a meter that

should only have positive kilowatt values reported negative values, the data for those

timestamps were corrected to a value of zero.

• Extreme meter spikes were flagged based on whether they were three standard deviations

(or more) above or below the average value for that meter. Simple, linear interpolation

was used to fill extreme meter spikes.

3.3 Step 3: Aggregate the Data

Once the data have been cleaned, meters should be aggregated (if desired) resulting in key

building performance parameters, such as whole-building energy end uses, energy use by space

type, and gross or net building energy use. In the RSF, the meters are aggregated into whole-

building energy end uses, which include:

• Cooling

• Heating

• Mechanical

• Lighting

• Plug loads

• Data center

• PV production.

3.4 Draw Conclusions About Energy Performance

Once the data have reached this point, building energy performance metrics for the RSF, such as

EUI and NZE status, are calculated. For the RSF data center power utilization effectiveness and

energy reuse effectiveness are calculated (see reference [3] for more information).

Key graphs that are generated each month to assess building energy performance are:

• Outdoor air temperature versus heating load

• Outdoor air temperature versus cooling load

• Typical 24-hour weekday load profile for each energy end use

• Typical 24-hour weekend load profile for each energy end use

• Stacked bar chart with a bar for each of the 12 most recent months; all the energy end

uses are stacked in each column.

3.4.1 Compare Metered Data to an Energy Model

We can strive for perfect energy efficiency of a building during the design phase; however it is

its operation that dictates its performance. Comparing actual performance (metered) with

expected performance (modeled) is key to understand corrective actions to ensure performance

as originally intended. The modeling approach that was used in the RSF is documented in detail

by Hirsh et al. [4]. The results of the RSF energy model for 2011 are included in the dataset that

accompanies this report.

A few items need to be checked when comparing measured data to modeled data:

1. The sampling rate of both datasets must match. For example, if the modeled data are in

1-hour intervals, the measured data must be adjusted to 1-hour intervals to facilitate

comparisons.

2. The date range of interest needs to be matched between the measured and modeled data.

For example, to compare measured performance to the energy model for January 2011,

be careful to only compare data from that month in both datasets.

3. The units must match between the measured and modeled datasets. For example, if the

modeled data are in gigajoules and the measured data are in kilowatt-hours, then modeled

data should be converted into kilowatt-hours.

4.0 Discussion of the Accompanying 2011 RSF Dataset

This report provides an explanation of how the RSF was set up for metering and how metered

data are analyzed to track energy performance over time. The accompanying datasets include

hourly measured and modeled energy data as well as measured weather data for 2011. The

report and datasets can be used by energy modelers, building science students, building energy

managers, and others to further their understanding of metering and analyzing the performance

of buildings.

There are a few things to note about the 2011 RSF dataset:

• To account for the unquantified efficiency of delivered hot water from NREL’s central

plant, the 2011 RSF dataset has assumed/accounted for a fixed delivered hot water

efficiency of 90%. The hot water efficiency is based in a high efficiency natural gas

boiler.

• To account for the unquantified efficiency of delivered chilled water from NREL’s

central plant, the 2011 RSF dataset has assumed/accounted for a fixed delivered chilled

water efficiency based on a COP of 3.0. The chilled water efficiency is based on code

minimum air-cooled chillers.

• The central plant energy use is included in the overall demand side EUI requirements.

5.0 References

[1] Federal Energy Management Program. Metering Best Practices: A Guide to Achieving

Utility Resource Efficiency. Release 2.0. 2011. URL:

http://www1.eere.energy.gov/femp/pdfs/mbpg.pdf.

[2] Torcellini, P., et al. 2006. "Lessons Learned from Case Studies of Six High-Performance

Buildings." NREL Report No. TP-550-37542. Golden, Colo.: National Renewable Energy

Laboratory. www.nrel.gov/docs/fy06osti/37542.pdf.

[3] Sheppy, Michael; Lobato, Chad; Van Geet, Otto; Pless, Shanti; Donovan, Kevin; Powers,

Chuck. 2011. Reducing Data Center Loads for a Large-Scale, Low-Energy Office

Building: NREL’s Research Support Facility. NREL Report No. BK-7A40-52785. Golden,

Colo.: National Renewable Energy Laboratory.

http://www.nrel.gov/sustainable_nrel/pdfs/52785.pdf

[4] Hirsch, Adam; Pless, Shanti; Guglielmetti, Rob; Torcsellini, Paul; Okada, David; Antia,

Porus. The Role of Modeling When Designing for Absolute Energy Use Intensity

Requirement in a Design-Framework. Preprint. Presented at ASHRAE Winter

Conference. 2011. URL: http://www.nrel.gov/sustainable_nrel/pdfs/49067.pdf.

Appendix A Link to RSF Datasets

NREL RSF Measured Data 2011

http://en.openei.org/datasets/node/933

NREL RSF Weather Data 2011

http://en.openei.org/datasets/node/934

NREL RSF Energy Model 2011

http://en.openei.org/datasets/node/935