rightsizing of rooftop hvac system - · pdf fileright sizing of rooftop hvac systems,...

TE

CH

NIC

AL R

EP

OR

T

RIGHTSIZINGOF ROOFTOP HVAC SYSTEMS

Advanced Energy Efficiency 2009

Prepared For:Idaho Power Company

Authors:Djunaedy, E.

Van Den Wymelenberg, K.Acker, B.

Thimmanna, H.

March 17, 2010Date

20090208-01Report No.

Prepared By:University of Idaho, Integrated Design Lab-Boise108 N 6th St. Boise ID 83702 USA www.uidaho.edu/idl

Kevin Van Den Wymelenberg Director

Ery Djunaedy, PhDProject Manager

1-Ery Djunaedy2-Kevin Van Den Wymelenberg3-Brad Acker4-Harshanna Thimanna

C09410-PO3 Contract No.

Prepared For: Idaho Power Company

Billie Jo McWinn Project Manager

DISCLAIMER

This report was prepared as the result of work sponsored by Idaho Power Company. It does not necessarily represent the views of Idaho Power Company or its employees. Idaho Power Company, its employees, contractors and subcontractors make no warrant, express or implied, and assume no legal liability for the information in this report; nor does any party represent that the uses of this information will not infringe upon privately owned rights. This report has not been approved or disapproved by Idaho Power Company nor has Idaho Power Company passed upon the accuracy or adequacy of the information in this report.

Please cite this report as follows: Djunaedy, E., Van Den Wymelenberg, K., Acker, B., Thimmana, H., 2010. Right Sizing of Rooftop HVAC Systems, Technical Report, 20090208-01, Integrated Design Lab, University of Idaho, Boise, ID.

This page left intentionally blank.

University of Idaho, Integrated Design Lab-Boise (Report # 20090208-01) Page iii

Table of contents

Table of contents.............................................................................................................iv

Executive Summary.......................................................................................................vii

1. Introduction...............................................................................................................11.1 Background..............................................................................................................................1

1.2 Objectives................................................................................................................................1

1.3 The scope and limitation..........................................................................................................1

1.4 Methodology............................................................................................................................21.4.1 Survey and Interview................................................................................................................2

1.4.1.1 Survey....................................................................................................................2

1.4.1.2 Interview................................................................................................................3

1.4.2 Measurement.............................................................................................................................3

1.4.3 Simulation.................................................................................................................................4

1.4.4 The Buildings............................................................................................................................4

1.5 Report Structure.......................................................................................................................5

2. RTU Oversizing and Part-Load Degradation: a Literature Review..........................72.1 Introduction..............................................................................................................................7

2.2 RTU Market potential..............................................................................................................7

2.3 Previous studies on RTUs........................................................................................................8

2.4 How RTUs are sized................................................................................................................8

2.5 The Rules of Thumbs.............................................................................................................10

2.6 Basic terminologies................................................................................................................142.6.1 Run-time fraction versus Cycling rate ....................................................................................14

2.6.2 Part-load degradation..............................................................................................................16

2.7 How to quantify over-sizing..................................................................................................19

2.8 How to quantify the penalty associated to oversizing...........................................................22

3. Surveys and Interviews...........................................................................................263.1 Introduction............................................................................................................................26

3.2 The respondents.....................................................................................................................26

3.3 The survey..............................................................................................................................26

3.4 The survey result....................................................................................................................28

3.5 The interviews........................................................................................................................30

3.6 Lessons Learned: Factors contributing to oversizing............................................................313.6.1 Generous Level of Internal Load.............................................................................................31

3.6.2 Shading is not included...........................................................................................................31

3.6.3 Safety factor............................................................................................................................31

3.6.4 Communication with architect or other designers...................................................................32

Page iv Rightsizing of Rooftop HVAC Systems; Advanced Energy Efficiency 2009

4. The signature of oversizing.....................................................................................334.1 Introduction............................................................................................................................33

4.2 The Measurement...................................................................................................................334.2.1 Methodology...........................................................................................................................33

4.2.2 Data Analysis..........................................................................................................................34

4.2.2.1 The maximum cycling rate...................................................................................35

4.2.2.2 Part-load ratio (PLR) and Part-load factor (PLF)..................................................38

4.2.2.3 The penalty of oversizing.....................................................................................41

4.3 Measurement results..............................................................................................................474.3.1 The cycling rate and RTF........................................................................................................48

4.3.2 Part-load ratio.........................................................................................................................49

4.3.3 Energy Penalty........................................................................................................................50

4.3.4 Peak demand penalty...............................................................................................................50

5. Simulation work......................................................................................................525.1 Objectives..............................................................................................................................52

5.2 Methodology..........................................................................................................................525.2.1 Calibration Stage.....................................................................................................................52

5.2.2 Sizing Stage............................................................................................................................53

5.2.3 Penalty Estimation..................................................................................................................53

5.3 Building A..............................................................................................................................545.3.1 Simulation Settings.................................................................................................................54

5.3.1.1 Description...........................................................................................................54

5.3.1.2 Construction.........................................................................................................54

5.3.1.3 HVAC System......................................................................................................54

5.3.2 Calibration Stage.....................................................................................................................56

5.3.3 Sizing Stage............................................................................................................................61

5.3.3.1 Load Component Analysis...................................................................................61

5.3.3.2 Sizing Calculation................................................................................................64

5.3.4 Penalty estimates.....................................................................................................................66

5.3.4.1 Energy Penalty of Oversizing...............................................................................67

5.3.4.2 Peak Demand Penalty of Oversizing....................................................................68

5.3.4.3 Comfort Penalty of Undersizing...........................................................................69

5.4 Putting it all together..............................................................................................................75

Appendix A. Survey and interview result......................................................................80 A.1 Survey result.......................................................................................................................80

A.2 Interview result...................................................................................................................82

Appendix B. Measurement results.................................................................................89 B.1 Introduction.........................................................................................................................89

B.2 The outdoor temperature comparison.................................................................................89

B.3 Measurement results...........................................................................................................96

B.3.1 Building A...............................................................................................................96

B.3.1.1 Building and RTU Description: Building A..........................................96

B.3.1.2 Measurement result: Building A............................................................96

B.3.1.3 Data analysis: Building A......................................................................97

B.3.2 Building B...............................................................................................................99

B.3.2.1 Building and RTU Description: Building B...........................................99

University of Idaho, Integrated Design Lab-Boise (Report # 20090208-01) Page v

B.3.2.2 Measurement result: Building B............................................................99

B.3.2.3 Data analysis: Building B.....................................................................102

B.3.3 Building C.............................................................................................................102

B.3.3.1 Building and RTU Description: Building C.........................................102

B.3.3.2 Measurement result: Building C..........................................................103

B.3.3.3 Data analysis: Building C.....................................................................107

B.3.4 Building D.............................................................................................................107

B.3.4.1 Building and RTU Description: Building D........................................107

B.3.4.2 Measurement result: Building D..........................................................107

B.3.4.3 Data analysis: Building D....................................................................112

B.3.5 Building E – RTU7................................................................................................113

B.3.5.1 Building and RTU Description: Building E.........................................113

B.3.5.2 Measurement result: Building E...........................................................113

B.3.5.3 Data analysis: Building E.....................................................................116

B.3.6 Building E – RTU6................................................................................................118

B.3.6.1 Building and RTU Description: Building E – RTU6...........................118

B.3.6.2 Measurement result: Building E – RTU6.............................................118

B.3.6.3 Data analysis: Building E – RTU6.......................................................120

B.3.7 Building F..............................................................................................................121

B.3.7.1 Building and RTU Description: Building F.........................................121

B.3.7.2 Measurement result: Building F...........................................................121

B.3.7.3 Data analysis: Building F.....................................................................122

B.3.8 Building G.............................................................................................................123

B.3.8.1 Building and RTU Description: Building G........................................123

B.3.8.2 Measurement result: Building G..........................................................123

B.3.8.3 Data analysis: Building G....................................................................124

B.3.9 Building H.............................................................................................................125

B.3.9.1 Building and RTU Description: Building H........................................125

B.3.9.2 Measurement result: Building H..........................................................125

B.3.9.3 Data analysis: Building H....................................................................127

Page vi Rightsizing of Rooftop HVAC Systems; Advanced Energy Efficiency 2009

Executive Summary

This report summarizes the work and findings of the “Rightsizing of Rooftop HVAC Systems” project funded by Idaho Power Company in 2009. The main objective of the project was to identify the 'signature of oversizing' of rooftop units (RTU). This was accomplished through measurement during peak cooling conditions and estimating the penalties associated with oversized RTUs for the monitored period. We also interviewed mechanical engineering firms to better understand the design and sizing process. Finally, extensive simulation work was conducted to extend the penalties associated with oversizing throughout the whole cooling season.

Determining the 'Signature of Oversizing'

From a review of the literature, we found several previous projects that had been carried out to study the performance of RTUs. All previous studies were carried out in order to identify solutions to increase the performance of (oversized) RTUs. Even though many of the earlier studies concluded that the RTUs were oversized, there was no attempt to look back at the design stage when these RTUs were sized to identify what factors led to oversizing.

We determined that previous methods for estimating penalties associated with oversizing did not account for all penalties due to their focus o on operations and maintenance. The earlier studies typically estimated the penalties by comparing two scenarios using the same RTU: (1) the cycling scenario and (2) the steady state scenario. Both scenarios used the the original oversized RTU and did not consider the performance of a 'rightsized' unit. By contrast, our method included a calculation of the penalty assuming a rightsized RTU as the comparison baseline, thus increasing the penalty associated with oversizing. For each oversized RTU we calculated the rightsized capacity and estimated the penalties by comparing the energy consumption and the peak cooling demand of two RTUs, the existing (oversized) RTU and the rightsized RTU.

The rightsized RTU has considerably lower capacity as compared to the existing RTU.. Somewhat surprisingly, the energy penalty estimated with the new method is similar to the earlier studies, because even though the baseline capacity had decreased, the smaller unit needed to run longer. However, the new method found that the peak cooling demand penalty was considerably higher than earlier studies because the peak cooling demand was compared to the rightsized RTU with a smaller capacity.

This research proposes to use both the maximum cycling rate (Nmax) and the run-time fraction (RTF) as the signature of oversizing. The desired combination

University of Idaho, Integrated Design Lab-Boise (Report # 20090208-01) Page vii

is to have low Nmax and high RTF. These two parameters were easily calculated based on data from the multiple RTUs measured during this research.

The Signatures of the Measured RTUs

A total of nine RTUs in eight buildings were measured during peak summer conditions in July and August 2009. The measurement protocol including logging the air temperatures at various points of the air distribution system (at a minimum the outside air, return air, mixed air, supply air, and indoor air temperatures). Additionally, the electric current drawn by the compressor and the supply fan were also logged.

Out of nine RTUs measured, this study found only two that were rightsized while the rest showed different degrees of oversizing. The two rightsized RTUs had the following combination of Nmax and RTF: 0 cycle/hour and 1 (Building G) and 1.13 cycles/hour and 0.9 (Building F). The low Nmax for Building F and G meant that the RTUs rarely cycled ON and OFF, and the high RTF meant that the RTUs ran almost all the time during the peak condition.

The rest of the buildings had various degree of oversizing with relatively high Nmax and low RTF. However an RTU with a high Nmax did not necessarily have a low RTF and, vice versa, an RTU with a low RTF did not necessarily have ahigh Nmax. The highest Nmax was 8.78 cycle/hour, with an RTF of 0.31. The lowest RTF was 0.15, with an Nmax of 2.66 cycle/hour.

Penalties of Oversizing

This study also found that the signature of oversizing (the Nmax and the RTF) accurately indicated oversizing. The rightsized RTUs (with the right combination of Nmax and RTF) hada Part-Load Ratio (PLR) of 1, which meant that the RTUs ran at full capacity. The other RTUs with various degrees of oversizing ran at part of the capacity. The RTU with the highest Nmax ran at a PLR of only 0.21, which meant it used only about 20% of its capacity to meet the cooling load on a design day condition. The RTU with the lowest RTF ran at a PLR of 0.5, which meant that the RTU met the peak load with only half of its capacity.

A similar trend was also observed in the penalty estimate. The rightsized RTUs had almost no penalty at all, both in terms of energy and peak demand. On the other hand, the RTUs with the signature of oversizing described above showed both an energy penalty and a peak demand penalty. The energy penalty was up to 50%, although the range of 15%-25% was more typical depending on the degree of oversizing. The peak demand penalty was as high as 0.92 kW/ton, meaning the peak demand savings from a 5-ton RTU can be 4.6 kW.

Design and Oversizing

Out of the nine RTUs measured, we have interview data associated with five of them. Interviews were conducted with the mechanical engineering firms to determine how the design and sizing process was carried out. The main result

Page viii Rightsizing of Rooftop HVAC Systems; Advanced Energy Efficiency 2009

of the interviews was a set of assumptions that were often used in the sizing calculations by mechanical engineers. With this and other parts of the interview, we have identified a number of design related factors that can lead to oversizing: (1) excessively high internal load assumptions (2) external shading ignored in the sizing calculation (3) safety factors in the sizing calculation , and (4) lack of communication with other members of the design team.

Simulating Oversizing

One building with severe indicators of oversizing was taken as a case study for simulation. The simulations were used to calculate the energy penalty for the whole cooling season (as compared to the energy penalty estimate based on the measurement period). Furthermore, simulations were conducted under different scenarios so that a number of RTU capacities were tested and compared against the 'as-designed' RTU capacity.

The results show that the as-designed RTU capacities had energy penalties as high as 48% throughout the cooling season, with peak demand penalties up to 62%. Another important result of the simulations work was determining how energy simulation software can be used as a sizing tool during design stages. This report shows how simulation software can produce the same result as the traditional sizing tools used during the design stages (given the same assumption), thus increasing confidence of design engineers. The added value of using simulation as a sizing tool are also illustrated. Specifically, one of the extra capabilities is the ability to provide multiple sizing options along with penalties of each option in terms of energy, peak demand and occupant thermal comfort. The designer will have the option to present a more complete picture of the sizing alternatives based on a number of scenarios so that the client/owner can make an informed decision when choosing a certain scenario, and potential carry some of the risk associated to the selection.

Incentivizing Right-Sizing

Unfortunately, design engineers currently do not have any incentive to rightsize RTUs, while at the same time they will avoid a great deal of potential risk by oversizing HVAC systems. Previous studies have made recommendations on how to increase the performance of an installed RTU. However, there is no clear recommendation on how to address the issue of oversizing during the design stage for new construction projects. This report suggests an accurate and repeatable method for rightsizing RTU replacements. The issue of rightsizing for new construction will addressed briefly below.

There is sufficient data available in this report to support the development of utility-funded pilot incentive programs for RTU replacements. Considering the age of typical office buildings in the Pacific Northwest, where two-third of small commercial space are built in 1987 (NEEA 2004), the market potential for RTU replacement is significant. This study recommends using the signature of oversizing (the cycling period and the RTF) as the basis for rightsizing design and any potential incentive programs. These two parameters can accurately estimate the penalties associated with oversizing. It is reasonably

University of Idaho, Integrated Design Lab-Boise (Report # 20090208-01) Page ix

easy for any maintenance engineer or HVAC contractor to conduct the measurement and process the measurement results. This method is explicitly outlined in this report.

The question of rightsizing for new construction is more complicated because actual measurement data are not available. The simulation methods outlined in this paper are new and the proposed practices will take time to penetrate into the design engineering market. Furthermore, this paper only examined one building via simulation. The sample size should be dramatically increased to provide greater confidence in the proposed method for simulation-based new construction rightsizing. A 'one year post new construction' incentive program is a possible alternative since this would allow one year of actual field monitoring. Unfortunately, for systems that prove to be oversized, this is simply too late. There is a definite need to consider ways to incentivize rightsizing prior to new construction. While this report provides relevant background data for this effort, additional research is needed.

Page x Rightsizing of Rooftop HVAC Systems; Advanced Energy Efficiency 2009

1 Introduction

1.1 BackgroundRooftop units (RTUs) are one of the most commonly used heating ventilating and air-conditioning (HVAC) systems for small commercial buildings. In Northern California (PGE 1997), they represent more than 2.3 million tons of air conditioning capacity, covering around 70% of the commercial cooling. In the Pacific Northwest, 34% of the commercial buildings are cooled with RTUs comprising an estimated 1.3 million tons (NEEA 2004).

Multiple institutions have conducted research projects to study the performance of RTUs. In the Pacific Northwest, studies have been carried out by the Northwest Energy Efficiency Alliance (NEEA), Portland Energy Conservation Inc. (PECI), and the Pacific Northwest National Laboratory (PNNL). These studies and other research will be reviewed in Section 2. Most research has focused on the operation and maintenance issues associated with RTUs, not the design and sizing of RTUs as will be discussed in this paper.

1.2 ObjectivesSpecifically, the study aims to:

1. Identify which aspects of the design process lead to RTU oversizing

2. Identify the “signature” of RTU oversizing by monitoring several RTUs at peak summertime conditions,

3. Estimate the energy penalties associated with RTU oversizing

4. Develop educational materials based upon the results

1.3 The scope and limitationThe scope of work was limited in the following ways:

1. RTU Type: packaged single zone RTUs only; we did not examine RTUs with variable air volume (VAV)

2. Building size: small commercial applications only; buildings less than 25,000 ft2 were considered in the sample

University of Idaho, Integrated Design Lab-Boise (Report # 20090208-01) Page 1 of 142

3. Building type: office buildings only; office buildings were selected because they have relatively consistent internal gains from building to building. This facilitates comparing data between buildings, however limits the generalizability of the results to office applications. Another useful building type to consider for future study would be small retail applications, however, the internal loads at this type of facility are highly variable and therefore not addressed in this paper.

4. Age of buildings: two groups were established, 'old' and 'new'; old buildings were determined to be anything designed prior to 2001. New buildings were determined to be those that were designed during or or after 2001. This date was selected since January of 2001 marked the date when the first energy code was enforced in Idaho. The term 'designed' refers to the point at which construction documents were submitted to the building department for review.

5. RTU mode: cooling only; although RTUs usually come with both heating and cooling systems, this study examined cooling performance only.

1.4 MethodologyThis research comprises three distinct methods:

1. Surveys and interviews with engineers

2. Measurement of RTU performance during peak summer conditions

3. Simulation of RTU performance using Energy Plus version 4.0

1.4.1 Survey and Interview

1.4.1.1 Survey

A survey was developed and sent to mechanical engineering firms in the Treasure Valley, near Boise. The firms were requested to list at least ten buildings (five 'old' and and five 'new') that incorporated RTUs as the HVAC system. Specifically, the survey requested the following information:

1. Building name

2. Year of design

3. Floor area of office space

4. Capacity of the RTU for the office space

Survey results provided general information about the buildings and basic sizing and capacity data of the RTU (in terms of ft2/ton, and total tons). From this list of ten buildings, we worked with the firm to select two buildings that

Right Sizing of Rooftop HVAC Systems; Advanced Energy Efficiency 2009 Page 2 of 142

would be studied in detail. The selected buildings would be considered for on-site monitoring, follow-up interviews, and simulation. The two buildings selected for additional investigation were chosen based upon the best match to the sample criteria defined in Section 1.3.

1.4.1.2 Interview

An interview (see Section 3.5 and Appendix A.2) was carried out with each engineering firm individually to examine the two buildings selected for detailed study. The scope of the interviews included a detailed review of the design process for each building and specifically addressed the single RTU that would be measured. The interview included questions in the following categories:

1. RTU sizing process: defining the design assumptions.

2. Design/construction dynamics: interactions with owner and contractors.

Results from the interviews were used to identify factors that might lead to sizing errors and to provide general insight into the engineering process.

1.4.2 Measurement

The monitoring was carried out on the RTUs for buildings that were selected as a result of the survey. Only one RTU for each building was selected for monitoring due to budget constraints. For each selected RTU, the following parameters were logged:

1. Outdoor air temperature (OAT)

2. Supply air temperature (SAT)

3. Mixed air temperature (MAT)

4. Indoor air temperature (IAT)

5. Return air temperature (RAT)

6. Supply fan current

7. Compressor current

The field monitoring was carried out for a period long enough to encompass at least one day where the maximum OAT was above 95ºF, representing a typical design day for Boise, ID.

The field monitoring data provided critical insights to determine the 'signature of oversizing'. In particular, cycling patterns of the compressor and runtime percentages were examined.


1.4.3 Simulation

Simulation was used to estimate, with as much accuracy as presently feasible with EnergyPlus, the total annual energy penalty associated with RTU oversizing. The simulations were carried out in several stages:

1. Calibration: the simulation settings (construction materials, schedules, and capacities) were set to match the existing construction and operating conditions. Simulations were conducted under the same conditions as the field monitoring. The compressor cycling patterns were compared for the two data sets.

2. Auto-sizing: the calibrated model was re-run using the 'auto-sizing' option such that the compressor was sized using the default method in EnergyPlus. The results of this stage were then established as a new (lower) capacity level for the compressor that still met the cooling load.

Annual simulations for both models (Calibrated and Auto-sized) were conducted and the energy penalty associated with oversizing was determined as the difference between the cooling energy consumption of the two simulation results.

1.4.4 The Buildings

Table 1.1 shows the list of firms and buildings involved in this study. The total number of buildings surveyed was 23, field monitoring was conducted at eight of these. A detailed report is provided for each of these eight buildings, and they are identified in this report as Buildings A-H. We conducted interviews with design firms representing six of the eight buildings monitored. One building was simulated using EnergyPlus. Two other buildings had eQUEST models that we acquired which had been previously completed by the mechanical engineers.


Table 1.1 Survey Sample

No Building Name Measured Interviewed Simulated eQuest Model

Firm A Building 1 Yes – Building C Yes Yes

Firm A Building 2

Firm A Building 3

Firm A Building 4

Firm A Building 5

Firm A Building 6 Yes – Building D Yes Yes Yes

Firm A Building 7

Firm A Building 8

Firm A Building 9

Firm B Building 10 Yes – Building E

Firm B Building 11 Yes – Building G

Firm B Building 12 Yes – Building F Yes

Firm C Building 13

Firm C Building 14

Firm C Building 15

Firm C Building 16

Firm C Building 17 Yes – Building B Yes Yes Yes

Firm C Building 18

Firm C Building 19

Firm C Building 20

Firm C Building 21

Firm C Building 22 Yes – Building A Yes Yes

Firm D Building 23 Yes – Building H

1.5 Report StructureThis report has the following structure:

Section 1 - The introduction describes the background, the objectives, the methodology, and the reporting of the project.

Section 2 - A review of previous studies, with the primary goal to develop a theoretical basis for the quantification of part-load degradation of cooling systems.

Section 3 - The results from surveys and interviews.


Section 4 -The signature of oversizing for RTUs measured, summarizing the results of on-site monitoring, and estimating penalty.

Section 5 - Energy simulation: estimating the penalty associated to oversizing RTUs beyond the period of monitoring over the whole cooling period.

Section 6 - Appendices:

a. Detailed results of survey and interviews (coded for anonymity) (Appendix A)

b. Detailed results of field monitoring (8 locations) (Appendix B)

c. Detailed results of simulations (4 building) (Appendix C)


2 RTU Oversizing and Part-Load Degradation: a Literature Review

2.1 IntroductionThis chapter summarizes existing literature on RTU performance. The first section highlights the niche that the current study fills among previous research. In the subsequent sections, the theoretical background on part-load degradation of HVAC equipment is discussed, in particular how to quantify degradation based on measured data. The review is concluded by a discussion of penalty estimation. This study estimated the penalty in peak-demand by addressing the oversizing in designed capacity of the RTUs.

2.2 RTU Market potentialIn northern California (Felts & Bailey 2000), small commercial office and retail buildings account for 50% of commercial building floor area and HVAC system energy use. Over 75% of the building stock is less than 5,000 ft2, and almost 90% is less than 10,000 ft2 . Approximately 80% of the buildings were constructed before 1985. Annual air conditioning energy use for the buildings in the hotter inland areas is 3.64 kWh/ft2. RTUs consume 4.3 billion kWh per year, which translates into approximately $400 million/year in energy expenses.

In the Pacific Northwest (NEEA 2004; NEEA 2005), small commercial office and retail buildings account for approximately 33% of commercial building floor area and about 36% of HVAC system energy use. Around 11% of the building stock is less than 5,000 ft2 , and almost 36% is less than 20,000 ft2 . Around 67% of the buildings were constructed before 1987.

These studies suggest substantial potential for RTU replacements in California and the Pacific Northwest due to the large number of small commercial buildings that were constructed over 25 years ago. It is likely that most other areas of the United States have similar statistics.


2.3 Previous studies on RTUsWe identified six previous studies that examined the performance of RTUs since 1998:

1. Pacific Gas and Electric (PGE), 1998, surveyed 250 RTU in Northern California.

2. Eugene Water and Electric Board (EWEB), 2001, studied 30 RTUs in Eugene, Oregon.

3. Puget Sound Energy (PSE), 2003 – 2004, studied 118 RTUs in Puget Sound, Washington.

4. Northwest Energy Efficiency Alliance (NEEA) Phase I, 2002, studied 65 RTUs in Idaho, Montana and Washington.

5. Northwest Energy Efficiency Alliance (NEEA) Phase II, 2003 – 2004, studied 75 RTUs in Idaho, Montana, Oregon and Washington.

6. California Energy Commission (CEC), 2001 – 2002, studied 215 RTUs in California.

All studies found various problems with RTU installation, maintenance, and operations, and all of them recommend action programs to mitigate these problems. Based upon the cycling rates identified, these previous studies concluded that many RTUs are oversized. (Cowan 2004). Specifically, a report by Felts and Bailey (2000) regarding the 1998 PGE study listed above stated that:

In at least 40% of the cases, the unit size could be dropped by 50% or more or conversely, the floor area that the unit serves could be increased by 100%.

Nonetheless, no previous study addressed the issue from the design point of view. Furthermore, very little guidance was offered to individuals planning RTU replacements in order for them to determine rightsized replacements. Our research aimed to examine oversizing from the design perspective and to provide accurate and repeatable field monitoring protocols for individuals planning rightsized RTU replacements. Finally, our research also aimed to develop simulation-based RTU sizing methods for new construction.

2.4 How RTUs are sizedSmall commercial buildings, the focus of this study, are typically skin dominated. That means that the cooling load responds more to the climate conditions (outside air temperature and solar radiation) than to the internal gains (people, equipments, lighting). A RTU, similar to any air conditioning system, is sized based on the peak external air temperatures. The challenge in


sizing a RTU is that the peak temperatures occur for only a few hundred hours throughout the entire year.

Figure 2.1 shows the temperature distribution for Boise, ID, based on the typical year data. The air conditioning (AC) unit is sized based on design day conditions, which is 95 °F for Boise, ID. There are only about 100 hours in a typical year that exceed this design condition in Boise.

Since small buildings are typically skin dominated, the cooling load is very sensitive to changes in the outside air temperature. The lower the outside air temperature, the lower the cooling load. Figure 2.2 shows how the cooling load of a building changes as the outside air temperature changes. Note that the example given in Figure 2.2 shows a scenario where the peak cooling load does not actually occur at the peak outdoor air temperature. This scenario happens when the cooling load is very much affected by the solar radiation, which peaks several hours before the outdoor peak temperature. Without considering solar radiation load, the cooling load at the peak outdoor temperature can be considerably lower, as illustrated in Figure 2.2.

Figure 2.1 Bin hour profile for Boise, ID


Figure 2.2 Building load v.s. AC capacity

Since the air conditioning is sized based on the 95 °F design day, it will operate most of the time to handle a cooling load (much) lower than its capacity. This is referred to as part-load operation. Furthermore, when the RTU has excessive safety factors (oversizing) (see Figure 2.2),the part-load condition is even worse.

Air-conditioning units do not operate as efficiently at part-load as they do at full-capacity. The theoretical background of part-load degradation will be discussed in the following sections.

2.5 The Rules of ThumbsThe problem with sizing HVAC systems for small buildings are shown through the results of the previous survey by Jacobs and Henderson (Jacobs & Hugh Henderson 2002). At least two conclusions are relevant for the current report. The first important conclusion is the average time spent designing HVAC systems for small building projects. Figure 2.3 shows the average time spent to complete various tasks when designing small building projects. The average time for engineers to design HVAC systems for small building projects is approximately 40 hours. Although this seems like a short period of time, the HVAC system design represents a large proportion of all hours spent on the design of small commercial buildings. Furthermore,


HVAC system design involves very broad work ranging from sizing calculations to air distribution calculations to overall system selection.

Figure 2.3 Design time for small building (Jacobs & Conlon 2002)

The second conclusion relates to the tool used for sizing calculations (Figure2.4). Approximately half (51%) of the respondents use manufacturers' sizing calculation software. The next biggest proportion (17%) rely only on previous experience and the rules of thumb. The widespread use of “previous experience” and rules of thumb could be an indication of why oversizing is so prevalent in small commercial buildings.


Figure 2.4 Tools used for sizing (Jacobs & Hugh Henderson 2002)

There is nothing inherently wrong with rules of thumb. Figure 2.5 and Table2.1 show some examples of the rule of thumb method. The rules of thumb are usually presented as a range of numbers and do not by themselves cause a problem if they are used as intended. That is, as a starting point and a secondary guide to verify other calculations.

Figure 2.5 Rules of thumb (Guthrie 2003)


Table 2.1 Rules of thumb for AC unit capacity (Bell 2008)

Building Type ft2/ton

Offices, Commercial: General 300 – 400

Offices, Commercial: Large perimeter 225 – 275

Offices, Commercial: Large interior 300 – 350

Offices, Commercial: Small 325 – 375

Banks, Court Houses, Municipal Buildings, Town Halls 200 – 250

Police Stations, Fire Stations, Post Offices 250 – 350

Precision Manufacturing 50 – 300

Computer rooms 50 – 150

Restaurants 100 – 250

Medical/Dental centers, Clinics, Offices 250 – 300

As Haines and Wilson (2003) put it:

Every HVAC designer needs some handy empirical data for use in approximating loads and equipment sizes during the early conceptual stages of the design process.

And further:

Energy conserving practice in envelope construction, in lighting design, and in system design has resulted in decreased loads in many cases. But increased use of personal computers and other appliances has the opposite effect of increasing the air conditioning requirements. Designers must develop their own site-specific data if the data are to be reliable.

The above quotes represent best practices with regards to the rules of thumb. However, some engineers use rules of thumb as their primary design tool. The following paragraph represents the view that rules of thumb are “accurate” enough even if design practices have changed over the years (Bell 2008):

Many of the rules of thumb listed within this reference manual were developed many years ago. I have received many questions when conducting seminars regarding these rules of thumb. The most often asked question is “Are the


cooling and heating load rules of thumb still accurate with the mandate of energy codes and tighter and improved building envelope construction?” The answer to this question is yes. The reason the cooling rules of thumb are still accurate is that the internal loads have increased substantially and cooling loads have switched from building-envelope-dependent, to lighting-dependent, and now to people-and-equipment-dependent (more people and equipment placed in the same area). The reason the heating load rules of thumb are still reasonably accurate is that the ventilation air (outdoor air load dictated by code) has increased.

The size of a RTU depends on the how they were designed and what tool was used. The fact that many RTU's are oversized has a lot to do with the designers relying heavily on manufacturers' software and the rules-of-thumb. We are not trying to suggest there is necessarily a problem with these tools, rather we are suggesting that designers must know how and when to use these tools and to understand their advantages and limitations.

2.6 Basic terminologies

2.6.1 Run-time fraction versus Cycling rate

Figure 2.6 shows the relationship between space temperature, the AC unit status and the set-point temperature. The AC unit is ON when the space temperature reaches the maximum point in the set-point range and is OFF when the space temperature reaches the minimum point. The range around the set-point temperature is ∆tspt.

The cycle time (tcycle) is the time for the AC unit to do a complete cycle of ON and OFF. Run-time fraction is the ratio of the time when the AC unit is ON to the total cycle time. It is important to understand the difference between run-time fraction (RTF) and the cycling rate.


Figure 2.6 Space temperature and AC status (Hugh Henderson et al. 1991)

Figure 2.7 Run-time fraction comparison

Run-time fraction (X) is the fraction of time in which the unit is ON.


X =tON

t cycle

(1)

And,

t cycle=tONtOFF (2)

Figure 2.7 above illustrates the difference between run-time fraction and cycle rate. Both graphs in Figure 2.7 have a run-time fraction of 50%. However, the bottom graph shows that the AC unit is ON for 60 minutes and then OFF for 60 minutes, and the top graph shows that the AC is ON for 30 minutes and OFF for30 minutes. A run-time fraction of 50% does not say a great deal about how frequently a unit cycles. It simply states the total time over a given period that a unit is ON.

Cycling rate (N), on the other hand, is how long a unit has a complete ON-OFF cycle.

N=1

t cycle

(3)

Referring to the above examples, 60 minutes ON and then 60 minutes OFF means a cycling rate of 0.5 cycles/hour, while 30 minutes ON then 30 minutes OFF means 1 cycles/hour.

2.6.2 Part-load degradation

Part-load degradation is described by the part-load factor (PLF), which is the ratio of the part-load coefficient of performance (COP) to the steady-state COP.

PLF=COPavg

COP ss

(4)

where:COPavg = Average COP (“degraded” COP) over the cycling timeCOPss = Steady-state COP

PLF is not to be confused by the part-load ratio (PLR) or the cooling load factor (CLF), which is the ratio of the building load to the total AC capacity.


PLR=Qavg

Q ss

=Building LoadAC Capacity

(5)

where:Qavg = Average capacity (“degraded” capacity) over the cycling timeQss = Steady-state AC capacity

The efficiency degradation curves can be explained in Figure 2.8 (Jacobs 2003b). AC equipment will need a certain amount of “start up” time to reach a steady-state output.

Figure 2.8 Performance degradation due to cycling (Jacobs 2003b)

The startup losses (shaded in red above) mark the difference between the actual output and the steady state output. When the AC unit operates continuously (with a high RTF and low cycling rate) the startup losses will be negligible. However, if the unit cycles frequently the startup time will become a significant portion of the total run-time. In this scenario, the startup losses becomes significant and degrade the performance of the AC unit.

A previous study by Parken et. al. (1985) established that the part-load performance of cycling HVAC equipment depends on:

1. the response of the system at startup (defined by a time constant or dead time).


2. the cycling rate of the equipment (governed by the thermostat characteristics and a building's thermal mass)

Figure 2.9 Part-load factor as function of part-load ratio

Figure 2.9 shows the correlation between PLF and PLR which is developed by using the following equation:

PLF=1−C D1−PLR (6)

And

C D=4N max 1−e

−14N max

(7)

where:PLF = Part-load FactorPLR = Part-load RatioNmax = Maximum cycling rateτ = The HVAC system time constant


The maximum cycling rate (Nmax) is derived from the characteristics of the HVAC system and the building.

N max=14

QAC

C T spt

(8)

where:QAC = Capacity of the AC system∆Tspt = The thermostat dead-bandC = Thermal capacity of the building

2.7 How to quantify over-sizingPrior research (Felts & Bailey 2000) shows that over 60% of rooftop units surveyed had a cycling rate of at least 3 cycles/hour. The same study further concluded that more than 40% of the units studied were more than 25% oversized and about 10% are considerably greater than 50% oversized. The study only labeled RTUs as 'oversized' if they were at least 25% oversized because many HVAC engineers consider oversizing by 25% as a “safe and acceptable practice” for oversizing.

In the same study, the quantification of oversizing was determined by monitoring the RTU compressor. Figures 2.10 and 2.11 show the typical measurement result for California. The graphs show the outdoor temperature and the operating status of the compressor. The oversized RTU shows a pattern of continuous cycling (Figure 2.10) while the properly sized RTU shows no cycling during the peak-day operation (Figure 2.11).


Figure 2.10 Measurement result with the signature of oversizing (Felts & Bailey 2000)

Figure 2.11 Measurement result showing properly sized RTU (Felts & Bailey 2000)


From the above measurement, the cycling rate and the RTF of the RTU can be calculated. Based on the cycling rate and the RTF, the part-load degradation (in terms of efficiency reduction) can be estimated.

Previous studies show several ways to estimate the efficiency of RTUs:

1. By using the compressor power (Felts 1998):

The measured compressor power is used to calculate the system efficiency as shown in the following equation:

Efficiency=compressor input power

capacity=

kWton

(9)

This efficiency can then be compared with the published efficiency (in terms of EER). The ratio of the measured efficiency and the nominal efficiency is the PLF (see equation 4).

The term 'capacity' in the above equation is the actual capacity (at the actual air temperatures at the condenser and evaporator). However, Felts (1998) decided to use the nominal capacity for two reasons; (1) their monitored data did not include air flow measurement and (2) their monitoring period was not confined to the 'design day' conditions or near-design day condition.

2. By using the measured air temperatures (outside air and at evaporator) (Felts 1998):

The study uses this method not to estimate the efficiency of the RTU but to generate a benchmark of ideal operation. The method uses the linear regression equations supplied by the Air Conditioning Contractors Association (ACCA) Manual J, as follows:

Total capacity=Km1CFM m2 T ewbm3 T oa (10)

Sensiblecapacity=Qn1CFMn2 T ewbn3T edbn4 T oa (11)

Compressor power=Ro1CFMo2 T ewbo3T oa (12)

The constants (K, Q, R, m1, m2, m3, n1, n2, n3, n4, o1, o2, o3) are determined using manufacturers' data.

Using this method one can calculate the efficiency of a RTU at any combination of air temperatures (outdoor and at evaporator). Comparing


this efficiency with the nominal efficiency, one can calculate the degradation.

3. By using the measured refrigerant condition:

A Pacific Northwest National Laboratory (PNNL) study (Armstrong et al. 2006) found that the measurement of airflow is not feasible because (1) it is difficult to do with acceptable accuracy, and (2) the space inside the RTU (around the evaporator) does not provide enough room to do a proper airflow measurement. The study concluded that measuring the refrigerant is the more accurate and feasible option.

Our study, however, did not use any of the above methods to quantify the degree of oversizing. The first method is not accurate enough because it uses the nominal capacity (instead of the capacity at the time of measurement) to calculate the efficiency. The second method will provide the ideal operation, but will not help in estimating the real situation (the second method was not initially used for part-load degradation in the first place). The third method is simply beyond the scope of this project. The method used for this study will be discussed in Section 4.2.2 on page 34.

2.8 How to quantify the penalty associated to oversizingAn American Council for an Energy Efficient Economy (ACEEE) study (Neme et al. 1999) found only a few studies that reported the benefits of rightsizing (or the penalty of oversizing). This is because:

It is not possible to correct equipment sizing problems without replacing the unit. That is extremely expensive and, therefore, never done.

Neme et. al. (1999) quoted McLain and Goldberg (1984) who estimated an energy savings of 0.2% for every 1 percent reduction in oversizing. That means an energy savings of 10% for correcting an average oversizing of 50%. The savings in terms of peak demand is also estimated as “moderate”, and no number is associated to the qualitative description.

It should be noted that McLain and Goldberg (1984) focused on the residential sector. The reported energy savings assumed an average oversizing of 50% or more, which means an average of around 1 ton oversizing for the average home. The average oversizing may be different for commercial buildings, and the average oversizing in tons will typically be more than 1 ton. Furthermore, the operation mode is rarely “continuously ON” for residential sector, only in about 20% of homes. Therefore the penalty for oversizing should be be significantly higher for the commercial sector.


Felts and Bailey (2000) report that over 60% of RTUs surveyed have cycling rates of 3 cycles/hour or more. Jacobs (2003a), referring to Felts and Bailey (2000) report, estimated that the potential energy savings from mitigating this problem is around 10%. Jacobs (2003a) did not elaborate on how he arrived at the 10% savings.

Felts and Bailey (2000) reported that 40% of RTUs are more than 25% oversized. This represents about 900,000 ton or around 180,000 units in Northern California. The study also found that the power draw of an average RTU is about 1.5 kW/ton. Their study estimated a 2.5 kW reduction in the peak demand by replacing an oversized RTU with a more efficient and properly sized RTU. The reduction in the peak demand (assuming40% of the RTUs in Northern California were replaced) would be 450 MW of the 1,350 MW peak (roughly 33% savings). Assuming 1,000 hours of operation for the whole cooling season, the savings would be 450 million kWh (roughly 33% savings).The 33% savings figure also represents the penalty due to oversizing.

Another study (Hugh Henderson et al. 1991) estimated the penalty for oversizing to be roughly 11% for Nmax of 2.5 cycles/hour, the average found in their study (Table 2.2). However, this estimated energy penalty simply considered the same size unit without cycling (under steady state energy use) instead of a rightsized unit which would actually be necessary to eliminate cycling. Essentially, the fact that the unit is oversized, thus resulting in the high cycling rate, is not factored into the energy penalty published.

Table 2.2 Penalty for oversizing (Hugh Henderson et al. 1991)

Figure 2.12 illustrates what is described in Henderson et al. (1991). The first graph shows the oversized scenario, where the Qss is the steady state capacity of the AC unit that cycles with tON1 and tcycle, and, q1 is the area under the curve which represents the energy output of the system.


To estimate the energy penalty, the same RTU (with capacity of Qss) is assumed to run under steady state condition for a certain period of time, so that the energy removed from the space (q2) is the same as the energy removed by the same RTU under cycling condition (q1). This is illustrated in the bottom graph of Figure 2.12. The result is a shorter ON time for the RTU under steady state condition, and the shorter ON time will result in about 11% savings (for Nmax = 2.5 cycle/hour) as described in Table 2.2.

Figure 2.12 Quantifying the penalty of oversizing as described by Henderson et al. (1991)

It should be noted that the RTUs shown in both scenarios in Figure 2.12 are the same oversized unit. This scenario is unrealistic because the oversized unit would, by definition, never run without cycling.

We developed Figure 2.13 to illustrate our concern with the above method. It address the issue of oversized capacity and represents a more realistic estimation on the penalties of oversizing.

Figure 2.13 Quantifying the penalty of oversizing


The first graph in Figure 2.13 represents the same cycling RTU. The cycling scenario (as indicated in Table 2.2) is calculated based on the assumption of 50% PLR, which means that the building load is only half of the RTU capacity. In other words, the building could function with a RTU of half the original capacity (or 0.5Qss) as described in the bottom graph of Figure 2.13.

The rightsized unit would run for tON3 which is the same as the tcycle in the first scenario, not at part-load but at full-load, which will improve its performance beyond what was estimated by Henderson et al. (1991).

The numbers in Table 2.2 may still represent the same penalty for the rightsized unit illustrated in Figure 2.13. The decrease in efficiency will still be the same if the RTUs (the oversized and the rightsized) both have the same EER. The increase in energy use is the same even though the rightsized RTU has half of the capacity because it will need to run longer (continuously). However, what was hidden in Table 2.2 is the peak-demand penalty. The scenario described in Figure 2.13 reduced the peak-demand by half compared to the scenario described in Figure 2.12. A previous study (Felts & Bailey 2000) estimated a reduction of 2.5 kW from the average 5-ton unit drawing 7.5 kW, which equates to a 33% reduction in peak demand.


3 Surveys and Interviews

3.1 IntroductionThe survey and interviews were carried out in accordance with the University of Idaho, Human Assurances Committee (HAC) requirements. The questions and survey formats were approved by the HAC. The survey and interviews are all confidential. The names of the buildings and engineering firms that participated in these surveys and interviews were replaced with unidentifiable codes.

3.2 The respondentsThe respondents for the surveys included mechanical engineers working in firms near Boise, ID. Furthermore, the firms selected had some previous awareness of the Integrated Design Lab (IDL). This decision was made based on the nature of the information sought in the survey. The confidential information that was shared by the firms demanded a high level of confidence and trust. Therefore, no “blind” or random invitations to participate in the survey were made.

Table 3.1 shows the list of participants with a summary of their involvement. Some firms had difficulty in responding to the survey because of the scope of work of this study. For example, some of the firms did not often design office buildings with RTUs, reducing the potential buildings to choose from.

3.3 The surveyThe survey requested that respondents list the buildings that met basic study parameters (office building with RTU, or buildings with office space that were served with a separate RTU). For those buildings, the respondents were asked to provide data for five 'new' buildings and five 'old' buildings:

1. Building Name: Enter the name of the building

2. Description: Enter the description of the building. General information on shape and use


3. Owner/Contact Info: Enter the name of the owner and the contact information

4. Year: Enter the year when the building is designed

5. Total SF (office only): Enter the total SF for the office space

6. Total SF (excluding office): Enter other areas of the building (if any)

7. Total Tons (office only): Enter the total refrigeration tons that serve office zones

8. Total Tons (excluding office): Enter the total refrigeration tons for the rest of the buildings

9. Heating source: Electric or Gas

10. Heating capacity (Btu/hr): Enter the designed heating capacity

11. Glazing Area: WWR for the whole building (Low=less than 20%, High=more than 40%

The survey was sent to all of the potential respondents, however, only four firms replied with usable data as summarized below in Table 3.1. The complete response from the respondents are available in Appendix A.1.


Table 3.1 Summary of respondents

Firm Name Response

Firm A Mechanical engineering firm, responded to the survey, two of the buildings were measured, interviewed for those two buildings

Firm B Mechanical engineering firm, responded to the survey, two of the buildings were measured, interviewed for those two buildings

Firm C Mechanical engineering firm, responded to the survey although with difficulties due to the building type, responded with three buildings for the survey, all three buildings were measured, interviewed for two of the buildings (the firm did not design the third building themselves)

Firm D Mechanical engineering firm, responded to the survey although with difficulties due to the building type, responded with three buildings for the survey, all three buildings were measured, interviewed for two buildings (the firm did not design the third building)

Firm E Mechanical engineering firm, attempted to respond to the survey, but failed to respond due to the building type requirement. The firm agreed to participate in the monitoring (their office building) although they did not design the building.

Firm F Mechanical engineering firm, attempted to respond to the survey, but failed to respond due to the building type requirement.

Firm G HVAC contractor firm, agreed to participate, but since they did not do any design work, they could not provide any data.

3.4 The survey resultTable 3.2 shows the results of the survey. The range of the RTU capacity is from 275 ft2/ton to 488 ft2/ton. Depending on which rules of thumb, the numbers range from slightly oversized to rightsized. The three entries with capacity above 700 ft2/ton are not used (highlighted in grey in the table) because they were installed in a building with many unconditioned areas and the firm did not have the data on the total unconditioned area. Also excluded were buildings with missing data (highlighted in yellow). These are the entries where the firm submitted the names of the buildings but later could not find the relevant information for various reasons.


Table 3.2 Survey result

No Building Name

Description Year Total SF Total Tons

ft2/ton

Firm A Building 1 One story office building 2000 13700 41 334


Firm A Building 3 One story office/exam building 1995 9300 26 358


Firm A Building 5 Polygon Two storey office building 2000 28900 69 419

Firm A Building 6 Polygon One story office/garage building 2005 6300 21 300

Firm A Building 7 Two storey office building 2006 2800 8 350

Firm A Building 8 One story office/garage building 2005 10500 37 284


Firm B Building 10 Office 2001 28800 76.5 376

Firm B Building 11 Office 1999 11125 32 348

Firm B Building 12 Office 2006 4875 10 488

Firm C Building 13 Office 1999 10878 36 302

Firm C Building 14 club house LOST 1999

Firm C Building 15 High school 1999 91054 65 1401

Firm C Building 16 Office LOST 1999

Firm C Building 17 Medical office 2004 16527 37 447

Firm C Building 18 Industrial office 2004 46515 59.5 782

Firm C Building 19 Medical office 2004 33460 45 744


Firm C Building 21 Bank 2004 3459 11 314

Firm C Building 22 Office 2001 6194 18 344

Firm D Building 23 Office

Without the missing and ignored entries, the overall average capacity is 359 ft2/ton, which is within the range of rightsized (on the end of the range towards oversized). Again, this depends on which rules of thumb are used to determine the degree of oversizing.

The average ft2/ton for each firm does not suggest a large variation among firms (348, 378, 363 ft2/ton for Firms A, B and C). For Firm A (who submitted the survey without any missing data), there is no significant variation between old buildings and new buildings (361 and 322 ft2/ton for old and new buildings) despite increased performance standards in energy


codes. The new buildings have slightly more capacity than the old buildings, although this cannot be concluded as a trend due to the small sample size. Of the three buildings with a capacity less than 300 ft2/ton, two are 'old' buildings and one is a 'new' building.

3.5 The interviewsThe interviews were carried out with three firms for a total of five buildings (see Table 3.3). The data from three buildings (highlighted in grey) could not be collected in the interview. Firm B could not collect all the data necessary for the interview for Building E so we could not complete the interview for this building. Buildings G and H are the office spaces for Firms B and D respectively, and the firms allowed us to measure the RTU even though the firms did not design the HVAC systems for these two buildings.

Table 3.3 The Firms and the measured buildings

Firm Building

Firm C Building A

Firm C Building B

Firm A Building C

Firm A Building D

Firm B Building E(see note in the text above)

Firm B Building F

Firm B Building G(see note in the text above)

Firm D Building H(see note in the text above)

The interviews were carried out after the data from the field monitoring had been compiled so that these preliminary results could be discussed during the interview. T he interview questions and their corresponding responses are compiled in Appendix A.2. The main objective of the interviews was to


discuss the design process and try to identify the factors that lead to system sizing. The results are presented in two forms:

1. Individual responses to the questions are discussed in terms of their potential contribution to oversizing.

2. Qualitative analysis of the responses are presented in the following section.

3.6 Lessons Learned: Factors contributing to oversizing

3.6.1 Generous Level of Internal Load

Occupancy load: the default policy for the mechanical engineers interviewed is to wait for the architect to provide a schedule for occupancy. We found a great over-estimate in occupancy load for the 'old' buildings. There was certainly a trend towards “head-count” for the design occupancy load for office spaces, especially for the 'new' buildings. However, there is only one firm who set the occupancy load based on the workstation count. This firm also said that for retail areas they typically use the net area instead of the gross area to calculate the occupancy load, and then use a 50% diversity factor.

Equipment (plug) load: The equipment load was often assumed to be high, up to 1.25 W/ft2. A previous study found (Komor 1997) that even 1 W/ft2 is at the high end of the normal range, and values above 1 W/ft2 are only found in less than 5% of the total areas studied. Typically, the equipment load was determined on a per area basis, and we found only one firm who set the equipment load at a per workstation basis.

Lighting load: usually set to the code requirement. The comparison between the 'old' and the 'new' buildings are a significant contrast since the old buildings were sized prior to energy codes. One of the old buildings listed upwards of 2 W/ft2 for lighting, while new buildings were typically closer to 1 W/ft2 .

3.6.2 Shading is not included

We found that engineers typically excluded all building shading during sizing calculations, even external fixed shading. There is no argument for this practice other than a safety precaution in case the shading was taken away by the building owner at a later date.

3.6.3 Safety factor

As indicated in the literature review, safety factors are commonly the focus of the oversizing debate. Nobody has the exact number as to what safety factor


should be included on top of the peak cooling load when determining the size of a RTU.

The firms interviewed discussed this problem. The general practice is to add a certain safety factor before selecting the RTU. Often, selecting the next available unit size for purchase involves gross oversizing. One firm has a practice to select the next available smaller unit, provided that the downsizing is within a certain safety limit. The firm did not specify what safety limit was used for these downsizing selections. One firm has a policy to not oversize at the end of the calculation because assumptions involved in the calculations have already built-in a safety factor. This firm also has a policy to downsize for retail spaces because of:

1. diversity factor (of the cooling load).

2. interaction with other systems; the cooling load for the space can be satisfied by other RTUs serving the building.

3.6.4 Communication with architect or other designers

Oversizing can happen because of communication problems. The sizing of mechanical systems runs in parallel with other design work, including cost estimation and often 'value engineering'. Building shading or advanced lighting control systems could be removed and result in system under-sizing and are therefore sometimes left out of sizing calculations. Furthermore, the decision to use high-performance windows, for example, might come too late in the design process, well after systems have been sized and specified. The sizing of mechanical system has long been completed using the worst case scenario, i.e. using low performance windows and the lowest common denominator for other system choices.

Another example from our experience suggests that design engineers are reluctant to consider load reductions due to daylight harvesting systems or other load reduction measures that involve complex commissioning or user involvement. These practices will inherently lead to a significant oversizing in terms of AC capacity given proper commissioning and user operation.


4 The signature of oversizing

4.1 IntroductionThis section summarizes the field monitoring work for this study. The main goal for the monitoring is to develop a methodology to identify the signature of oversizing. The data analysis is summarized by using an example from one of the data set gathered during the field monitoring. The rest of the measurement data is presented in the summarized form, and the analyzed data is used to show the signature of oversizing.

The equations used in the data analysis section is taken from (Hugh Henderson et al. 1991).

4.2 The Measurement

4.2.1 Methodology

The measurements were carried out on eight buildings and a total of nine RTUs. Table 4.1 shows the buildings and the measurement period.

Table 4.1 Measurement period

Building Start Date End Date

Building A 10-Aug-2009 12-Aug-2009

Building B 9-Jul-2009 17-Jul-2009

Building C 19-Aug-2009 25-Aug-2009

Building D 20-Aug-2009 31-Aug-2009

Building E – RTU7 3-Aug-2009 7-Aug-2009

Building E – RTU6 29-Jul-2009 3-Aug-2009

Building F 10-Aug-2009 12-Aug-2009

Building G 23-Jul-2009 28-Jul-2009


Building H 25-Aug-2009 31-Aug-2009

The summary of the RTUs measured is shown in Table 4.2. Almost all of the RTUs were small, typically less than 10 tons. Two RTUs were 10 tons or larger but these had two smaller compressors instead of one large compressor. The EERs ranged from 8-11.

For every RTU, the measured parameters are recorded for at least one peak condition design day. The measurement was stopped once we recorded a set of data with OAT of at least 94 °F. Some buildings have data for multiple days, either because the measurement began on a relatively cool day, or simply because the next building to be logged was not available once we recorded the peak conditions required.

Table 4.2 RTU summary

No. of compressor

Capacity(tons)

EER

Building A 1 4 9.7Building B 1 4 11.1Building C 1 4Building D 1 3Building E – RTU6 2 10 9.1Building E – RTU7 1 3Building F 1 6 10.1Building G 1 6Building H 2 17.5 11

The measurement was carried out using HOBO data loggers with temperature probes for air temperature and current transformers to monitor electric current. Some of the compressors were measured using a status monitoring sensor, which reported only the ON/OFF status. The following sections describes a complete example of how the data were analyzed in order to determine the oversizing of each RTU.

4.2.2 Data Analysis

From different parameters discussed earlier (see Section 2.6), this study uses the Nmax and the PLR as the signature of oversizing. Both parameters are calculated from the measurement data.


4.2.2.1 The maximum cycling rate

Nmax is calculated based on the cycling data. Figure 4.1 shows the measurement data for Building A. The outdoor temperature during the measurement reached 100 °F, which is higher than the design day temperature for Boise. These data show that the compressor is cycling during a very hot day, a clear sign of oversizing.

Figure 4.1 Field Monitoring data

From the field monitoring data, the following data were extracted:

1. The start time of every compressor cycle

2. The end time of every compressor cycle

3. The OFF period between the compressor cycles

4. The outside air temperature at the start of the compressor cycle

From the above data the following parameters were calculated:

1. tON

tON=t start−tend (13)


2. tcycle

t cycle=tONtOFF (14)

3. Cycling rate

N=1

t cycle

(15)

4. Runtime Fraction (RTF)

RTF=tON

t cycle

(16)

Table 4.3 shows the results of the data analysis described above. Note that the last cycle is not used because the period from the last time off to the next cycle is too long.

Table 4.3 Cycling pattern for Building A

Cycle #tON

(hour)tcycle

(hour)OAT(°F)

N(cycle/hour)

RTF(-)

1 8 35 74.30 1.71 0.23

2 5 21 79.00 2.86 0.24

3 5 87 77.90 0.69 0.06

4 7 68 88.49 0.88 0.10

5 6 42 93.52 1.43 0.14

6 7 37 95.62 1.62 0.19

7 7 29 98.23 2.07 0.24

8 7 38 96.58 1.58 0.18

9 7 40 97.21 1.50 0.18

10 7 31 98.53 1.94 0.23

11 9 49 99.32 1.22 0.18

12 8 48 95.18 1.25 0.17

13 10 55 95.14 1.09 0.18

14 7 56 89.78 1.07 0.13

15 7 73 89.13 0.82 0.10


Nmax is the maximum cycling rate that is calculated using a curve fit from the combination of RTF and the cycling rate. The equation is:

N=4 N max X 1−X (17)

where:N = Cycling rateΝmax = Maximum cycling rateX = Runtime Fraction (RTF)

Figure 4.2 shows the result of the curve fit. The maximum cycling rate for Building A is 2.66 cycles/hour.

Figure 4.2 Curve fit to calculate Nmax

Table 4.4 shows the maximum cycling rate from previous studies. Henderson et. al. (1991) found that the average Nmax for the study is 2.5 cycles/hour.


4.2.2.2 Part-load ratio (PLR) and Part-load factor (PLF)

The PLR and PLF are calculated based on the following equations (Hugh Henderson et al. 1991):

PLR=tON

t cycle

−

t cycle

1−e−tON

(18)

PLF=1−

tcycle

1−e−t ON

(19)

where:tON = The time when the compressor is ONtcycle = The cycling timeτ = RTU time constant

Table 4.4 Maximum cycling rate from previous studies (Hugh Henderson et al. 1991)

The time constant (τ) is a time that shows how fast a compressor reaches the steady state output when it starts from OFF. This is empirical data that can only be found from previous studies. Henderson et. al. (1991) used 80


seconds. Another study (H Henderson et al. 2000) used 60 seconds for “typical AC” and 30 seconds for “good AC”.

Table 4.5 show the PLR for Building A. The PLR shows the degree of oversizing. The maximum PLR is 0.2 (with τ=60 sec). Considering that PLR shows the building load (see equation 4 on page 16), and that the field monitoring was carried out on a peak cooling day, then the capacity of the RTU is about 5 times greater than peak load, or 400% oversized.

Table 4.6 shows the PLF for Building A. The PLF provides the performance degradation of the RTU. The minimum PLF is 0.8 (with τ=60 sec). As described in equation 5 (on page 17), this means that the RTU was running at 80% of its nominal COP.

Table 4.5 Part-load Ratio for Building A

Cycle # tON

(min)tcycle

(min)N

(cycle/hr)RTF PLR

(t=60 sec)PLR

(t=30 sec)% diff

1 8 35 1.71 0.23 0.200 0.214 7.0%2 5 21 2.86 0.24 0.191 0.214 12.0%3 5 87 0.69 0.06 0.046 0.052 13.0%4 7 68 0.88 0.1 0.088 0.096 9.1%5 6 42 1.43 0.14 0.119 0.131 10.1%6 7 37 1.62 0.19 0.162 0.176 8.6%7 7 29 2.07 0.24 0.207 0.224 8.2%8 7 38 1.58 0.18 0.158 0.171 8.2%9 7 40 1.5 0.18 0.150 0.163 8.7%

10 7 31 1.94 0.23 0.194 0.210 8.2%11 9 49 1.22 0.18 0.163 0.173 6.1%12 8 48 1.25 0.17 0.146 0.156 6.8%13 10 55 1.09 0.18 0.164 0.173 5.5%14 7 56 1.07 0.13 0.107 0.116 8.4%15 7 73 0.82 0.1 0.082 0.089 8.5%

4.2.2.3 The penalty of oversizing

Energy penalty

The energy delivered by the RTU to cool the space throughout tON is determined by the following equation:

qcycling=QsstON−1−e−tON

(20)

where:tON = The time when the compressor is ON (in hour)


τ = RTU time constant (in hour)Qss = Steady state RTU total capacity (in W or Btu/hr)qcycling = energy over the cycling period (in kWh or Btu)

This energy is depicted as q1 in Figure 4.3 below. This amount of energy was delivered by the cycling compressor.

Table 4.6 Part-load Factor for Building A

Cycle #tON

(min)tcycle

(min)N

(cycle/hr) RTFPLF

(t=60 sec)PLF

(t=30 sec) % diff1 8 35 1.71 0.23 0.875 0.938 7.2%2 5 21 2.86 0.24 0.801 0.900 12.4%3 5 87 0.69 0.06 0.801 0.900 12.4%4 7 68 0.88 0.1 0.857 0.929 8.4%5 6 42 1.43 0.14 0.834 0.917 10.0%6 7 37 1.62 0.19 0.857 0.929 8.4%7 7 29 2.07 0.24 0.857 0.929 8.4%8 7 38 1.58 0.18 0.857 0.929 8.4%9 7 40 1.5 0.18 0.857 0.929 8.4%

10 7 31 1.94 0.23 0.857 0.929 8.4%11 9 49 1.22 0.18 0.889 0.944 6.2%12 8 48 1.25 0.17 0.875 0.938 7.2%13 10 55 1.09 0.18 0.900 0.950 5.6%14 7 56 1.07 0.13 0.857 0.929 8.4%15 7 73 0.82 0.1 0.857 0.929 8.4%


Figure 4.3 Energy penalty

The input energy of the cycling compressor can be calculated using the following equation:

ecycling=E ss tON (21)

where:ecycling = input energy of the cycling compressor over the period of one

cycle (in kWh or Btu).Ess = Steady state RTU input power (in W or Btu/hr)

To calculate the energy penalty, the same amount of energy should be delivered by the existing compressor during steady-state (the bottom graph in Figure 4.3 with q2 equals q1).

The tON2 (which is the ON time of the compressor under steady state) was calculated as follows:

tON2=qcycling

Q ss

(22)

where:qcycling = energy over the cycling period (in kWh or Btu), as defined

above (equation 20).Qss = Steady state RTU total capacity (in W or Btu/hr)


The input energy of the compressor on steady state was calculated as follows:

ess=E ss tON2 (23)

where:ess = input energy of the compressor under steady state condition

(in kWh or Btu).Qss = Steady state RTU total capacity (in W or Btu/hr)

The energy penalty was calculated as follows:

energy penalty=ecycling

ess

−1 (24)

Table 4.7 and Table 4.8 show the calculation results for the energy penalties for =60 sec and =30 sec respectively. For =60 sec, the energy penalty is around 16% while for =60 sec the energy penalty is 7.5%.

Table 4.7 The energy penalty for oversizing for Building A (=60 sec)

Cycle #tON

(hour)tcycle

(hour)qcycle

(Btu)ecycle

(Wh)tON2

(hour)ess

(Wh)Energy Penalty

1 0.13333 0.58333 5600 582 0.11667 509 14.28%2 0.08333 0.35000 3205 364 0.06678 291 24.79%3 0.08333 1.45000 3205 364 0.06678 291 24.79%4 0.11667 1.13333 4801 509 0.10002 436 16.65%5 0.10000 0.70000 4002 436 0.08337 364 19.94%6 0.11667 0.61667 4801 509 0.10002 436 16.65%7 0.11667 0.48333 4801 509 0.10002 436 16.65%8 0.11667 0.63333 4801 509 0.10002 436 16.65%9 0.11667 0.66667 4801 509 0.10002 436 16.65%

10 0.11667 0.51667 4801 509 0.10002 436 16.65%11 0.15000 0.81667 6400 655 0.13334 582 12.50%12 0.13333 0.80000 5600 582 0.11667 509 14.28%13 0.16667 0.91667 7200 727 0.15000 655 11.11%14 0.11667 0.93333 4801 509 0.10002 436 16.65%15 0.11667 1.21667 4801 509 0.10002 436 16.65%

TOTAL 73619 7782 6693 16.27%


Table 4.8 The energy penalty for oversizing for Building A (=30 sec)

Cycle #tON

(hour)tcycle

(hour)qcycle

(Btu)ecycle

(Wh)tON2

(hour)ess

(Wh)Energy Penalty

1 0.13333 0.58333 6000 582 0.11667 545 6.67%2 0.08333 0.35000 3600 364 0.06678 327 11.11%3 0.08333 1.45000 3600 364 0.06678 327 11.11%4 0.11667 1.13333 5200 509 0.10002 473 7.69%5 0.10000 0.70000 4400 436 0.08337 400 9.09%6 0.11667 0.61667 5200 509 0.10002 473 7.69%7 0.11667 0.48333 5200 509 0.10002 473 7.69%8 0.11667 0.63333 5200 509 0.10002 473 7.69%9 0.11667 0.66667 5200 509 0.10002 473 7.69%

10 0.11667 0.51667 5200 509 0.10002 473 7.69%11 0.15000 0.81667 6800 655 0.13334 618 5.88%12 0.13333 0.80000 6000 582 0.11667 545 6.67%13 0.16667 0.91667 7600 727 0.15000 691 5.26%14 0.11667 0.93333 5200 509 0.10002 473 7.69%15 0.11667 1.21667 5200 509 0.10002 473 7.69%

TOTAL 79600 7782 7236 7.54%

Peak demand penalty

In order to calculate the peak demand penalty, a “rightsized” capacity ws estimated for a particular RTU. The rightsized capacity was calculated based on the peak load. Once the rightsized capacity was calculated, the input power was alos calculated by assuming the same EER for the RTU. The rightsized input power was then compared to the installed input power to determine the peak demand penalty.

The peak cooling load of the building was calculated using the following equation:

Building Cooling Load=PLR×Q1 (25)

where:PLR = Part-load ratioQ1 = RTU total capacity (in W or Btu/hr)

Since the measurement was carried out on a peak cooling day, the measured cooling load represents the peak cooling load.


The rightsized capacity of the RTU is set to be the same as the peak cooling load.

Q2=Peak Cooling Load (26)

where:Q2 = Rightsized RTU total capacity (in W or Btu/hr)

The degree of oversizing was calculated by using the following equation:

Oversizing Factor=Q1

Q2

−1 (27)

The input power for the rightsized capacity was calculated by assuming the same EER:

E2=Q2

EER (28)

where:E2 = input power for the rightsized RTU (in W or Btu/hr)Q2 = Rightsized RTU total capacity (in W or Btu/hr)EER = Energy Efficiency Ratio of the RTU

The peak demand penalty can be expressed in terms of percentage, or it can be expressed in terms of kW per ton.

Peak demand penalty=E1−E2 (29)

Or,

Peak demand penalty=E1−E2

Q1

(30)

Where:E1 = input power for the oversized RTU (in W or Btu/hr)E2 = input power for the rightsized RTU (in W or Btu/hr)Q1 = RTU total capacity (in ton)


Table 4.9 shows the peak demand penalty calculation (assuming =60 sec). The maximum “rightsized” capacity is 12,416 Btu/h, which makes the RTU (at 4 ton capacity) around 200% oversized. The peak demand penalty is around 3.5 kW or about 0.89 kW/ton.

Table 4.10 shows the peak demand penalty calculation (assuming =30 sec). The maximum “rightsized” capacity is 13,448 Btu/h, which makes the RTU (at 4 ton capacity) around 156% oversized. The peak demand penalty is around 3.5 kW or about 0.877 kW/ton.

The peak demand penalty does not seem to be sensitive to the time constant (τ).

Table 4.9 The peak demand penalty for oversizing for Building A (=60 sec)

Cycle #tON

(hour)tcycle

(hour)CLF

Peak Load(Btu/h)

Degree Oversized

Input Power

(W)

Penalty(W)

Penalty(kW/ton)

1 8 35 0.20 9600 4.00 873 3491 0.8732 5 21 0.19 9158 4.24 833 3531 0.8833 5 87 0.05 2211 20.71 201 4163 1.0414 7 68 0.09 4236 10.33 385 3979 0.9955 6 42 0.12 5717 7.40 520 3844 0.9616 7 37 0.16 7785 5.17 708 3656 0.9147 7 29 0.21 9933 3.83 903 3461 0.8658 7 38 0.16 7580 5.33 689 3675 0.9199 7 40 0.15 7201 5.67 655 3709 0.927

10 7 31 0.19 9292 4.17 845 3519 0.88011 9 49 0.16 7837 5.12 712 3651 0.91312 8 48 0.15 7000 5.86 636 3727 0.93213 10 55 0.16 7855 5.11 714 3650 0.91214 7 56 0.11 5144 8.33 468 3896 0.97415 7 73 0.08 3946 11.16 359 4005 1.001


Table 4.10 The peak demand penalty for oversizing for Building A (=30 sec)

Cycle #tON

(hour)tcycle

(hour)CLF

Peak Load(Btu/h)

Degree Oversized

Input Power

(W)

Penalty(W)

Penalty(kW/ton)

1 8 35 0.21 10286 3.67 935 3429 0.8572 5 21 0.21 10286 3.67 935 3429 0.8573 5 87 0.05 2483 18.33 226 4138 1.0354 7 68 0.10 4588 9.46 417 3947 0.9875 6 42 0.13 6286 6.64 571 3792 0.9486 7 37 0.18 8432 4.69 767 3597 0.8997 7 29 0.22 10759 3.46 978 3386 0.8468 7 38 0.17 8211 4.85 746 3617 0.9049 7 40 0.16 7800 5.15 709 3655 0.914

10 7 31 0.21 10065 3.77 915 3449 0.86211 9 49 0.17 8327 4.76 757 3607 0.90212 8 48 0.16 7500 5.4 682 3682 0.92113 10 55 0.17 8291 4.79 754 3610 0.90314 7 56 0.12 5571 7.62 506 3857 0.96415 7 73 0.09 4274 10.23 389 3975 0.994

4.3 Measurement resultsThe measurement results are presented in terms of parameters that have been discussed in the previous section. The calculation methods are the same as the examples in the previous section, so they will be omitted for the other buildings. Only the summary of the results will be presented for the rest of the RTUs studied.

Detailed measurement results for each RTU are presented in Appendix B.3 on page 94.

The results have been summarized for all measurement days, except for Building H where daily results are presented to highlight how the staging of the two compressors work. Building E RTU-6 also has two compressors and is assumed to have only one. The RTF of the second compressor is so low that it is assumed to have never cycled during the measurement period (see Appendix B.3.6).

Table 4.11 shows how the measurement data is presented.


Table 4.11 Measurement result presentation

Compressor Stage Day

Building A 1 1 All dayBuilding B 1 1 All dayBuilding C 1 1 All dayBuilding D 1 1 All dayBuilding E – RTU6 1 1 All dayBuilding E – RTU7 1 1 All dayBuilding F 1 1 All dayBuilding G 1 1 All dayBuilding H – 2-1 Day 1 2 1 Day 1Building H – 1-1 Day 2 1 1 Day 2Building H – 1-1 Day 3 1 1 Day 3Building H – 1-2 Day 1 1 2 Day 1

4.3.1 The cycling rate and RTF

Table 4.10 shows the cycling rates – average and maximum – and the RTF. The average cycling rate is the average from the actual data. The maximum cycling rate is the result of the curve fit based on the measurement data. The RTF presented below is the maximum RTF from the measurement data. The maximum cycling rate and the maximum RTF indicate the performance of the compressor during the design day.

The maximum cycling rate is high for most buildings, too high compared to other values found in the literature. At this point we do not have any explanation on why the cycling rate is so high. The desired combination is to have a low maximum cycling rate and high RTF. The opposite combination is a sign of oversizing.

Almost all buildings show a degree of oversizing, except Buildings F and G. Buildings A has the lowest RTF and relatively high cycling rate. Buildings C, D and E have a very high cycling rate, higher than any values reported by previous studies.

Building H shows relatively good performance (low cycling rate and a RTF higher than 0.5). However, the data for Building H only represents a single compressor of a dual compressor system. The single compressor has shown


some degree of oversizing. Therefore if we consider both compressors the system is dramatically oversized.

Table 4.12 The cycling rate and run-time fraction

Number of cycles

Cycling rate (Ave)

Cycling rate (Max)

RTF

# cycle/hour cycle/hour (ratio)

Building A 15 1.27 2.66 0.15Building B 32 1.63 1.75 0.55Building C 161 5.01 6.22 0.36Building D 44 2.97 4.53 0.29Building E – RTU6 27 4.16 6.50 0.56Building E – RTU7 228 6.91 8.78 0.31Building F 3 0.32 1.13 0.9Building G 3 0.12 0.00 1Building H – 2-1 Day 1 26 1.12 2.65 0.5Building H – 1-1 Day 2 7 0.41 1.52 0.71Building H – 1-1 Day 3 7 0.41 1.49 0.88Building H – 1-2 Day 1 6 1.88 2.51 0.27

4.3.2 Part-load ratio

Table 4.13 shows the PLR and the average EER, which is the degraded EER due to cycling. The values of the maximum cycling rate and the RTF are also shown in the same table to illustrate the relationship between these two values with the PLR and the EER.

The RTUs with a signature of rightsizing (low cycling rate and high RTF) tend to have a high PLR, which means the compressors in these RTUs run at or almost at full capacity (see Buildings F, G and also B). This also means that these RTUs have very low EER degradation. On the other hand, RTUs with high cycling rates and low RTFs have low PLRs and high EER degradation.


Table 4.13 The part-load ratio and EER degradation

Cycling rate (Max)

RTF PLREERavg

EER Nominal

EER Degradation

cycle/hour (ratio) (ratio) (Btu/hr)/W (Btu/hr)/W

Building A 2.66 0.15 0.21 9.46 11.00 14.00%

Building B 1.75 0.55 0.75 12.35 13.00 5.00%

Building C 6.22 0.36 0.65 8.69 11.00 21.00%

Building D 4.53 0.29 0.78 9.13 11.00 17.00%

Building E – RTU6 6.50 0.56 0.75 7.96 9.00 11.56%

Building E – RTU7 8.78 0.31 0.50 6.02 9.00 33.11%

Building F 1.13 0.90 1.00 10.04 10.10 0.59%

Building G 0.00 1.00 1.00 8.98 9.00 0.22%

Building H – 2-1 Day 1 2.65 0.50 0.89 10.59 11.00 3.73%

Building H – 1-1 Day 2 1.52 0.71 0.97 10.89 11.00 1.00%

Building H – 1-1 Day 3 1.49 0.88 0.98 10.91 11.00 0.82%

Building H – 1-2 Day 1 2.51 0.27 0.36 9.73 11.00 11.55%

4.3.3 Energy Penalty

Table 4.14 shows the energy penalty due to oversizing. The similar pattern as before can be observed: the RTUs with high cycling rate and low RTF have high energy penalties. On the other hand, the RTUs with low cycling rate and high RTF have low energy penalties (see Buildings B, F and G).

For RTUs with two compressors, like Building H, the energy penalty is small because the penalties listed are for the first stage only (one compressor). The energy penalty would be considerably higher for the second stage where it is substantially oversized.

4.3.4 Peak demand penalty

Table 4.15 shows the peak demand penalty. Again, a similar pattern can be observed. RTUs with high cycling rate and low RTF (the signature of oversizing) have high peak demand penalties.


Table 4.14 Energy penalty

Cycling rate (Max)

RTFEnergy Penalty

cycle/hour (ratio) %

Building A 2.66 0.15 16.27Building B 1.75 0.55 5.24Building C 6.22 0.36 26.59Building D 4.53 0.29 20.50Building E – RTU6 6.50 0.56 13.06Building E – RTU7 8.78 0.31 49.62Building F 1.13 0.90 0.59Building G 0.00 1.00 0.20Building H – 2-1 Day 1 2.65 0.50 3.87Building H – 1-1 Day 2 1.52 0.71 0.97Building H – 1-1 Day 3 1.49 0.88 0.78Building H – 1-2 Day 1 2.51 0.27 13.04

Table 4.15 Peak demand penalty

Cycling rate

(Max)RTF

Oversized (peak)

Peak-load penalty

Peak-load

penalty

Peak-load penalty

cycle/hour (ratio) % W % kW/ton

Building A 2.66 0.15 383.26 3461.00 79.33 0.87Building B 1.75 0.55 34.09 3692.00 100.00 0.92Building C 6.22 0.36 53.85 1527.00 35.00 0.38Building D 4.53 0.29 27.78 711.00 21.73 0.24Building E – RTU6 6.50 0.56 33.33 3333.00 25.00 0.33Building E – RTU7 8.78 0.31 99.90 1999.00 49.98 0.67Building F 1.13 0.90 0.44 31.00 0.43 0.01Building G 0.00 1.00 0.17 13.00 0.16 0.00Building H – 2-1 Day 1 2.65 0.50 12.09 2059.00 10.79 0.12Building H – 1-1 Day 2 1.52 0.71 2.96 549.00 2.88 0.03Building H – 1-1 Day 3 1.49 0.88 2.44 455.00 2.38 0.03Building H – 1-2 Day 1 2.51 0.27 175.00 12149.00 63.64 0.69


5 Simulation work

5.1 ObjectivesThe main objectives of the simulation work for this project are:

1. To estimate the cooling energy penalty associated with oversizing for the whole cooling season instead of just the field monitoring period.

2. To highlight the capability of the energy simulation tool as an alternative sizing tool.

5.2 MethodologyThe simulation work was carried out in three stages:

1. calibration stage, where the internal gains were calibrated to the level found in field monitoring. The result of this stage is the calibrated scenario, which was a set of simulation settings that represent the field monitoring condition.

2. sizing stage, where the RTU is sized under different scenarios (including the calibrated scenario above). The result of this stage is a number of capacities that were based on different scenarios.

3. penalty estimation stage, where the energy simulations were carried out for the combination of capacities and scenarios. This stage highlights the sizing problem, i.e. the RTU (with a certain capacity) is usually operated under a scenario that is different from what is used for sizing the capacity.

5.2.1 Calibration Stage

The calibration stage focuses on the measurement period. During this stage, the simulation settings (in this case the internal gains) were calibrated so that the results closely matched the existing design conditions as found during field monitoring.

Ideally, the calibration process would include simulation of the compressor status at the same time-step as used during the field monitoring, and then compare the compressor cycling profile and the temperature profile with the measurement results. If the simulated compressor cycled with the same


profile as the field monitoring data, then the simulation setting could be determined to represent the existing condition within a specified accuracy range. However, this is not possible due to the limitations of the RTU model in Energy Plus. Alternatively, the RTF over the field monitoring period is used as the parameter to compare the simulation results against. Once the RTF of the simulation and the field monitoring matched (within a certain limit) , then the simulation settings were determined to represent the actual condition during the field monitoring. This condition represented the calibrated scenario.

The calibrated scenario is important because it represented the peak cooling load condition. The capacity that is sized using this scenario is one potential candidate for the rightsized capacity.

5.2.2 Sizing Stage

In the sizing stage, the 'autosize' function of EnergyPlus was employed to calculate the RTU capacities under multiple scenarios. There are two important questions in this stage:

1. under the as-designed scenario (using the same load assumptions as the design engineers), can we arrive at the same RTU capacity using simulation as a sizing tool ?

2. what other capacities are produced using other scenarios?

Before the actual sizing calculation was carried out, a load component analysis was performed. The load component analysis is a series of sizing simulations with a certain cooling load component activated at a given time. The cooling load component included in these analyses are (1) opaque wall, (2) glazing, (3) lighting, (4) equipment or plug-load, (5) people ( including the ventilation load), and (6) external shading.

From the load component analysis we can determine dominant load components and those that are less dominant. We also can test the sensitivity of the result to the boundary condition definition.

5.2.3 Penalty Estimation

The combination of capacities and scenarios were simulated for the whole cooling season. The penalties are calculated in terms of the cooling energy consumption and the peak electricity demand. Apart from the (energy and peak demand) penalty of oversizing, we also investigated the comfort penalty for 'undersizing'.


5.3 Building A

5.3.1 Simulation Settings

5.3.1.1 Description

The simulated space is one thermal zone served by a single RTU. The zone has three rooms: a conference room, an office space, and a storage room. The zone is located in the northwest part of the building, with its north and west walls exposed to the external environment. The zone has a flat roof. Figure5.1 shows the simulated zone modeled in EnergyPlus. The total area of the zone is 940 ft2.

Figure 5.1 The simulated zone in Building A

5.3.1.2 Construction

External walls are made up of 5/8” gypsum board on either side with R-19 insulation. The concrete floor is a 6” thick slab on grade. The roof is a concrete mass with R-30 insulation. The glazing area is located in the north wall (area=252 ft2, WWR=17%) and the west wall (area=784 ft2, WWR=9%). Double pane glass is used with the following thermal properties: SC=0.81 and U-value=0.59.

5.3.1.3 HVAC System

The RTU is modeled using the EnergyPlus template for a unitary system. There are five curves used to describe the performance of the RTU. For this simulation we used custom curves based on the manufacturer's data (see Figure 5.2 and Figure 5.3).


Figure 5.2 RTU Performance Curve (function of temperature) for Building A

Figure 5.3 RTU Performance Curve (function of airflow) for Building A


The figures above show the comparison between the manufacturer's (measured) data and the EnergyPlus default curve. Figure 5.2 shows that, the default curve (as a function of temperature) is a very good fit to the measurement data. Figure 5.3, however, shows that the default curve (as a function of airflow) is not a good fit to the measurement data. Figure 5.3 also shows the corrected curve (as a function of airflow) used in all simulations.

5.3.2 Calibration Stage

Figures 5.4 through 5.8 and Tables 5.1 and 5.2 show the results of the calibration process for Building A. Tables 5.1 and 5.2 show the RTF and the PLR for both the operating period and the period where the outside air temperature is above 90°F. The operating period (in Tables 5.1 and 5.2) is the period when the compressor is operating from the first time it cycles ON in the morning to the last time it cycles ON during the night.

Table 5.1 RTF Comparison for Building A

RTF(operating period)

RTF(above 90°F OAT)

Simulation (As Designed) 0.51 0.69Simulation (Calibrated) 0.23 0.32Simulation (No internal gains) 0.14 0.19Measurement Data 0.15 0.18

Table 5.2 PLR Comparison for Building A

PLR(operating period)

PLR(above 90°F OAT)

Simulation (As Designed) 0.45 0.69Simulation (Calibrated) 0.22 0.30Simulation (No internal gains) 0.13 0.18Measurement Data 0.16 0.21

Table 5.1 shows the difference in the RTF between the simulation and the field monitoring data. The RTF for the as-designed case is significantly higher than the measurement data, an indication that the load assumptions are higher than during measurement period. The calibrated simulation RTF is closer to the measured RTF although it is still higher.


Table 5.1 also shows that if all gains are eliminated, the RTF is further decreased. Even though the RTF for the 'no internal gains case' is closer to the measurement, we cannot take this to represent the measurement condition because we know for a fact that there were internal gains during the measurement period as listed. The fact that the calibrated case still has a RTF higher than the measurement data suggests there is a 'heat sink' that is not taken into account. As will be discussed later, the heat sink that is not defined was determined to be the heat loss from the floor.

Table 5.2 shows the PLR between the simulation cases and the measurement. The same discussion as the RTF above also applies to the PLR result.

Figure 5.4 shows the simulation result with as-designed assumptions for internal gains. Comparing this with the measurement data (Figure 4.1 on page 35) it is obvious that the as-designed assumptions are too high. The main difference (between Figure 5.4 and Figure 4.1) is that in the as-designed simulation, the compressor works continuously during the peak period between 4pm and 5pm (in fact for the whole operating period), which did not happen during the measurement. The measurement data (Figure 4.1) shows that the compressor cycled on and off during that period 16 times.

Figure 5.4 Simulation Result for Building A (As-Designed Condition)


Figure 5.5 shows another representation of the results. It shows how the RTF and PLR changes over time. The two average values are indicated in the graph as a comparison.

Figure 5.5 Simulation Results (PLR and RTF) for Building A (As Designed)

Figures 5.6 and 5.7 show the calibrated simulation result. The internal loads were set according to the observations made on site during the field monitoring period. Table 5.3 shows the difference in the internal gain assumptions between the as-designed case and the calibrated case.

Table 5.3 Internal gains for Building C

As Designed Calibrated

Occupancy 24 persons 6 personsLighting 2 W/ft2 0.7 W/ft2

Equipment 0.851 W/ft2 Same as-designed(only in office)


Figure 5.6 Simulation Result for Building A (Calibrated Condition)

Figure 5.7 Simulation Results (PLR and RTF) for Building A (Calibrated Condition)


The results in Figure 5.7 show that the calibrated condition did not result in a RTF and PLR (0.32 and 0.30 respectively for OAT above 90°F) similar to the field monitoring data (0.18 and 0.21 respectively). This result suggests that (1) the internal gains are still too high, or (2) there is a heat sink that is not accounted for.

To investigate further, another simulation was carried out under no-internal-gains condition. Figures 5.8 to 5.9 show the result of this simulation. Both the RTF and PLR values (0.19 and 0.18 respectively for OAT above 90°F) are close to the measurement values (0.18 and 0.21 respectively).

Figure 5.8 Simulation Result for Building A (No Internal Gains)

However, even though the RTF and PLR values for the no-gains condition is closer to the field monitoring data we know that this does not represent the condition during the measurement period. We know that there were internal gains during the field monitoring. As such, we took the internal gains as described in Table 5.3 as the calibrated condition. We also discerned that there was a heat sink that was not represented in the simulation. This will be discussed further in the next section.


Figure 5.9 Simulation Results (PLR and RTF) for Building A (No Internal Gains)

5.3.3 Sizing Stage

5.3.3.1 Load Component Analysis

Figure 5.10 shows the load component of the as-designed condition. The line at the bottom (red) represents the load from the opaque envelope, and the lines above it represent the inclusion of various internal gains. The comparison between the lines show the relative significance of the load component to the peak load.

There are three important notes from Figure 5.10:

1. The biggest contributors to the peak cooling load are the glazing and the people (including ventilation).

2. The shading has a minimal effect on the peak cooling load (the difference between the top two curves). Note that this is caused by the location of the zone, which is on the northwest corner of the building.


The effect of the shading will be more dominant in the zone in the opposite corner of the building.

3. The peak cooling load estimated by EnergyPlus corresponds closely with the designed peak load for the building calculated by the mechanical engineer.

Figure 5.10 Peak Load Component (Adiabatic Ground)

Even though Figure 5.10 replicates the designed peak load (as calculated by the mechanical engineer in the design stage), it should be noted that the simulation uses the assumption of an adiabatic floor, which means there is no heat loss from the floor. The same assumption was also used in the sizing calculation done by the mechanical engineer. The result of the calibration stage (see discussion in the previous section) suggests that this assumption is not correct. Under the assumption of an adiabatic floor, the simulation would produce a RTF and PLR that are significantly higher than the measurement period RTF and PLR.

A number of simulations were carried out to investigate the effect of floor heat loss to the sizing calculation. To avoid a more complicated calculation, a constant ground temperature was used for the simulations. The EnergyPlus


manual suggests to set the ground temperature at 2°C ( 3.6 °F) less than the set point temperature of the room.

Figures 5.11 and 5.12 show the sizing calculation based on the ground temperatures of 22°C (71.6 °F)and 21°C (69.8° °F) respectively.

Figure 5.11 Peak Load Component (Ground at 71.6 °F)

During the calibration stage, the first case (with 22°C ground) does not result in a RTF or PLR similar to the field monitoring data. The results of the calibration stage are all simulated using 21 °C (69.8 °F) ground temperature. The temperature 21 °C (69.8 °F) is 2.8 °C (5 °F) below the set-point temperature, slightly higher than the 2 °C suggested by EnergyPlus manual.

Even though the results from the calibration simulation (Tables 5.1 and 5.2) produce a RTF and PLR that are higher than the field monitoring, there is no attempt to decrease the ground temperature further, because the ground temperature used is already lower than what is suggested by the EnergyPlus manual (2 °C or 3.6 °F). Therefore the result from the sizing calculation (in the next section) will be relatively higher, and in turn, the estimation of the penalty due to the oversizing will be more conservative.


It is important to note that the effect of the floor heat loss in the peak load is significant. Comparing Figure 5.10 and Figure 5.12, the floor heat loss can reduce the peak cooling load by approximately 21%.


5.3.3.2 Sizing Calculation

The sizing calculations were carried out under four scenarios (Table 5.4). Scenario 1 represents the “as-designed” scenario, while Scenario 2 represents the calibrated scenario. Scenario 2, however, has only 5 occupants in the conference room, which is true during the field monitoring period, but may not represent the peak occupancy of the conference room. Note that Scenario 2 has one more person compared to Table 5.2. This is because during the measurement there was nobody in the file room, while for sizing calculation (Scenario 2) we needed to assume at least one person is in the room.

Scenarios 3 and 4 are introduced to test exactly the same scenario as Scenario 2, but with higher occupancy in the conference room.


Table 5.4 RTU Sizing for Different Scenario

Scenario 1 Scenario 2 Scenario 3 Scenario 4

Ground Adiabatic at 69.8F at 69.8F at 69.8FShading No Yes Yes Yes

OccupancyConference 20 5 10 20Office 2 1 1 1File 2 1 1 1

LightingConference 2 0.7 0.7 0.7Office 2 0.7 0.7 0.7File 2 0.5 0.5 0.5

EquipmentConference 0.87 0.8 0.8 0.8Office 0.87 0.5 0.5 0.5File 0.87 0.1 0.1 0.1

Figure 5.13 and Table 5.5 show the sizing calculation for the four scenarios. Scenario 1 has exactly the same RTU capacity as what was designed by the mechanical engineer. Scenario 2 (the calibrated scenario) has a much smaller capacity compared to Scenario 1. Note that even if it is smaller than Scenario 1, it is considerably higher than the estimated peak cooling load during the measurement period. There is no attempt to further reduce the capacity of Scenario 1 to the estimated peak load during the measurement period because we did not do the detailed analysis on the floor heat loss and because the higher capacity will result in more conservative estimates of the energy penalty.

Table 5.5 RTU Sizing for Different Scenario

Capacity Scenario Btu/hr ton CFM

Capacity 1 Scenario 1 38,424 3.20 1,553Capacity 2 Scenario 2 16,753 1.40 684Capacity 3 Scenario 3 18,663 1.56 745Capacity 4 Scenario 4 22,475 1.87 867

The capacities for Scenario 2 and 4 are higher than Scenario 1 because of the higher occupancy in the conference room. However, even though Scenario 4 has the same occupancy in the conference room as Scenario 1, it has a capacity that is much less than Scenario 1 (41% less). The big difference between the capacity of Scenarios 1 and 4 is caused by (1) the use of external


shading in the calculation, (2) a lower lighting level, (3) slightly fewer occupants, and (4) the floor heat loss.


5.3.4 Penalty estimates

To estimate the penalty associated with RTU oversizing, all four calculated capacities (as in Table 5.5) were simulated for the whole cooling season. Each capacity was also simulated under all four scenarios (as in Table 5.4). For each capacity under each scenario three types of simulation were carried out. Table 5.6 shows the difference between the three types of simulation. With four capacities, four scenarios, and three simulation types, there are a total of 48 simulation runs.


Table 5.6 Three Types of Simulation

Type Simulation Time Time-step Internal gains

Type 1 Whole cooling season(1-May through 30-Sep)

60 min Varying over time of day with reasonable schedules

Type 2 Whole cooling season(1-May through 30-Sep)

60 min Constant at maximum value

Type 3 One day (11-Aug),the measurement day

1 min Constant at maximum value

5.3.4.1 Energy Penalty of Oversizing

Table 5.7 shows the result of the simulations. Not all the combinations of capacities and scenarios are relevant in the calculation of the penalty. The energy penalty associated to oversizing should be calculated based on the most reasonable scenario for each of the capacities.

Table 5.7 Cooling Energy Consumption for 'Building A' under Various Scenarios

Cooling(kWh)

Fan(kWh)

Total(kWh)

Scenario 1

Capacity 1 1,939 3,081 5,019Capacity 2 1,231 1,581 2,811Capacity 3 1,314 1,722 3,036Capacity 4 1,456 1,722 3,178

Scenario 2

Capacity 1 1,114 3,081 4,194Capacity 2 764 1,581 2,344Capacity 3 803 1,722 2,525Capacity 4 872 1,722 2,594

Scenario 3

Capacity 1 1,194 3,081 4,275Capacity 2 822 1,581 2,403Capacity 3 864 1,722 2,586Capacity 4 942 1,722 2,664

Scenario 4

Capacity 1 1,403 3,081 4,483Capacity 2 944 1,581 2,525Capacity 3 1,000 1,722 2,722Capacity 4 1,094 1,722 2,817


Capacity 1 of Scenario 4 (as-designed capacity under calibrated internal gains with full capacity) is taken as the basis of the comparison. Scenario 1, which is the as-designed condition, was not taken as the basis of the comparison because it represents the maximum condition possible and will not occur all the time. It is true that Capacity 1 of Scenario 4 will not occur all the time either, but for the penalty calculation this combination will result in a conservative value of the penalty.

The other capacities use its corresponding scenario (i.e. the scenario that was used to size the capacity) for the simulation. Table 5.8 shows the summary of the penalties.

Depending on which capacity is used to replace the installed RTU, the energy savings range from 37% to 48%. Note that these numbers are significantly higher than the savings estimated from the measurement data (see Section 4.3.3 on page 49), and even higher than the theoretical estimate (see Section 2.8 on page 22).

Table 5.8 Energy Penalty Estimate for Building A

in kWhCapacity 1 Capacity 2 Capacity 3 Capacity 4

Scenario 4 Scenario 2 Scenario 3 Scenario 4

Cooling 1,403 764 864 1,094Fan 3,081 1,581 1,722 1,722Total 4,483 2,344 2,586 2,817Penalty - 48% 42% 37%

5.3.4.2 Peak Demand Penalty of Oversizing

Table 5.9 shows the maximum power demand for the compressor and the fan. As discussed earlier, not all capacities under all scenarios are relevant for peak demand penalty estimation. Table 5.10 shows the estimate of the peak demand penalty for Building A.

The peak demand savings is between 48% to 62%, which translates to approximately 2.4 kW to 3.1 kW.


Table 5.9 Maximum Compressor and Fan Power for Building A under Various Scenarios

Compressor(kW)

Fan(kW)

Total(kW)

Scenario 1

Capacity 1 4,284 839 5,123Capacity 2 1,635 430 2,065Capacity 3 1,811 469 2,280Capacity 4 2,153 546 2,699

Scenario 2


Scenario 3


Scenario 4


Table 5.10 Peak Demand Penalty Estimate for Building A

Compressor(W)

Fan(W)

Total(W)

Penalty

Capacity 1 Scenario 4 4,227 839 5,066 -Capacity 2 Scenario 2 1,481 430 1,911 62%Capacity 3 Scenario 3 1,687 469 2,156 57%Capacity 4 Scenario 4 2,095 546 2,641 48%

5.3.4.3 Comfort Penalty of Undersizing

Figures 5.14 to 5.17 show the comparison of the indoor air temperature and the compressor power under different scenarios throughout a peak day. These are the results of Type 3 simulations (see Table 5.6). The figures show (1) how compressors with different capacities cycle under different scenarios and (2) whether the indoor air temperature rises above the set-point temperature.


For Scenario 1 (Figure 5.14), it is obvious that only Capacity 1 can maintain the set-point temperature throughout the day. The other capacities will result in indoor air temperatures above the set-point temperature.

For Scenario 2 (Figure 5.15), it is also obvious that all capacities can maintain the set-point temperature in the room because the calibrated internal load is low. The compressor for Capacity 1 cycles throughout the day, while the other compressors do not.

Figure 5.16 shows the result for Scenario 3. The indoor air temperature is maintained close to the set-point temperature except for Capacity 2 and Capacity 3 which produce indoor temperatures above the set-point. For Capacity 2, the indoor air temperature exceeded the set point at around 3pm, while for Capacity 3 it maintained temperature until approximately 4pm.

Figure 5.17 shows the result for Scenario 4. Since the occupant load is higher than in Scenario 3, only Capacity 1 can maintain the indoor air temperature near the set-point temperature. The other capacities cannot maintain the set-point temperature. For Capacity 2, the indoor air temperature started to exceed the set point around noon, while for Capacity 3 it exceed it at about 1pm. For Capacity 4, the temperature started to rise at about 3pm.

Figure 5.14 Simulation Results for Scenario 1



Figures 5.18 and 5.19 show the results of Type 2 simulation, which covers the whole cooling season (see Table 5.6 for the types of simulation). The indoor air temperature was plotted against the time of day.

Figure 5.18 shows that under Scenario 3 (with 10 people in the conference room), Capacity 2 (which is the smallest) can maintain the set-point temperature until to 2pm, and it will only get uncomfortable sometime after 5pm. To put it in another way, if the owner is willing to schedule any large meetings before the 2pm on the hottest days, Capacity 2 (which is only about 35% of the as-designed capacity) can handle all the cooling load for the zone.

Figure 5.18 shows the maximum indoor air temperature, but it does not have any information on how many hours the maximum temperature occurs. Table 5.11 shows the number of hours for the whole cooling season (from the beginning of May to the end of September). Capacity 2 has 62 hours above 76 °F, which is only 1.66% of the total hours in the cooling season (3260 hours). If we limit to the working hours, Capacity 2 has only 9 hours above 76 °F, which is only 0.32% of the working hours in the cooling season (2640 hours).


Figure 5.18 Maximum Indoor Air Temperature as a Function of Time of Day (Scenario 3)

What if we have 20 persons in the conference room (Scenario 4)? Figure 5.19 shows the maximum indoor air temperature for Scenario 4. Capacity 2 obviously cannot handle the load past 11am. Capacity 4 can maintain the set-point temperature until 2pm, and can keep the room below 78 °F until 6pm.

Table 5.12 shows the unmet hours of Scenario 4. Capacity 2 has 350 unmet hours, which is more than 9% of the total hours. It terms of working hours Capacity 2 has about 149 hours above 76 °F, which is almost four weeks. This is certainly not acceptable.

Table 5.11 Unmet Hours for Building A (Scenario 3)

All hours Capacity 1 Capacity 2 Capacity 3 Capacity 4

Hours above 76 °F

0 61 18 00.00% 1.66% 0.50% 0.00%

Hours above 78 °F

0 5 0 00.00% 0.14% 0.00% 0.00%

Office Hours Capacity 1 Capacity 2 Capacity 3 Capacity 4

Hours above 76 °F

0 9 1 00.00% 0.32% 0.03% 0.00%

Hours above 78 °F

0 0 0 00.00% 0.00% 0.00% 0.00%


Figure 5.19 Maximum Indoor Air Temperature as a Function of Time of Day (Scenario 4)

Table 5.12 Unmet Hours for Building A (Scenario 4)

All hours Capacity 1 Capacity 2 Capacity 3 Capacity 4

Hours above 76 °F

0 350 212 710.00% 9.54% 5.77% 1.93%

Hours above 78 °F

0 147 82 140.00% 4.01% 2.22% 0.37%

Office Hours Capacity 1 Capacity 2 Capacity 3 Capacity 4

Hours above 76 °F

0 149 78 180.00% 5.63% 2.94% 0.68%

Hours above 78 °F

0 49 21 20.00% 1.85% 0.80% 0.06%

Capacity 4, however, has more acceptable unmet hours. With only 71 total hours (and 18 working hours) above 76 °F. Capacity 4 is a the best alternative to the as-designed capacity and is less than half of the as-designed capacity. It certainly offers savings with acceptable risk regarding human comfort.

Figures 5.20 and 5.21 show the maximum indoor temperature as a function of outdoor temperature, for Scenario 3 and 4 respectively. Figure 5.20 shows that for Scenario 3 (with 10 people in the conference room), Capacity 2 can maintain the set-point temperature until the outdoor temperature is about 85 °F, and will only become uncomfortably hot when the outside air temperature is above 90 °F.


Figure 5.20 Maximum Indoor Air Temperature as a Function of Outside Air Temperature (Scenario 3)

Figure 5.21 Maximum Indoor Air Temperature as a Function of Outside Air Temperature (Scenario 4)

Figure 5.21 shows that for Scenario 4 (with 20 people in the conference room), Capacity 4 can maintain the indoor air temperature at the set-point temperature until the outdoor air temperature is about 85 °F, and will only be uncomfortably hot when the outdoor air temperature is above 92 °F.


5.4 Putting it all togetherThere are a number of lessons learned from the simulation work:

1. Energy simulation can be used as a sizing tool. Designers can present the client with a more complete picture of performance based on simulation results.

2. Sizing calculations should be presented not only based on the maximum internal gains scenario, but also based on the likelihood or frequency that such scenarios will ever happen.

3. Alternative scenario(s) should be presented along with the maximum internal gains scenario, and the comparison of the various penalties (energy, peak demand and comfort) should be made.

4. Instead of bearing all the risk and liability alone (and because of that, size the system to the maximum capacity), the designer can ask the owner to bear the consequences of choosing a smaller capacity. The owner will have first cost and energy cost savings with a certain calculated risk.


References

Armstrong, P.R., Sullivan, G.P. & Parker, G.B., 2006. Field Demonstration of a High-Efficiency Packaged Rooftop Air Conditioning Unit at Fort Gordon, Augusta, GA, Richland, WA: Pacific Northwest National Laboratory. Available at: http://www.osti.gov/energycitations/servlets/purl/894472-eyCW5b/ [Accessed August 24, 2009].

Bell, A.A., 2008. HVAC: Equations, Data, and Rules of Thumb 2nd ed., McGraw-Hill.

Cowan, A., 2004. Review of Recent Commercial Roof Top Unit Field Studies in the Pacific Northwest and California, White Salmon, WA, USA: New Buildings Institute. Available at: http://www.peci.org/ComRetail/docs/NWPCC_SmallHVAC_Report_R3_.pdf.

Felts, D.R., 1998. Rooftop Unit Performance Analysis Tool - A Case Study, California, USA: Pacific Gas and Electric Company.

Felts, D.R. & Bailey, P., 2000. The State of Affairs - Packaged Cooling Equipment in California. Available at: http://www.eceee.org/conference_proceedings/ACEEE_buildings/2000/Panel_3/p3_11/.

Guthrie, P., 2003. The architect's portable handbook : first step rules of thumb for building design 3rd ed., McGraw-Hill.

Haines, R.W. & Wilson, C.L., 2003. HVAC Systems Design Handbook 4th ed., McGraw-Hill.


Henderson, H., Parker, D. & Huang, J., 2000. Improving DOE-2’s RESYS routine: User Defined Functions to Provide More Accurate Part Load Energy Use and Humidity Predictions, Berkeley, CA, USA: Lawrence Berkeley National Laboratory.

Henderson, H., Raustad, R. & Rengarajan, K., 1991. Measuring Thermostat and Air Conditioner Performance in Florida Homes, Florida, USA: Florida Solar Energy Center. Available at: http://www.fsec.ucf.edu/en/publications/pdf/FSEC-RR-24-91.pdf.

Jacobs, P., 2003a. Small HVAC Field and Survey Information, Califonia Energy Comission. Available at: http://www.energy.ca.gov/2003publications/CEC-500-2003-082/CEC-500-2003-082-A-23.PDF.

Jacobs, P., 2003b. Small HVAC System Design Guide, Califonia Energy Comission. Available at: http://www.newbuildings.org/downloads/FinalAttachments/A-12_Sm_HVAC_Guide_4.7.5.pdf.

Jacobs, P. & Conlon, T., 2002. State-of-the-Art Review Whole Building, Building Envelope, and HVAC Component and System Simulation and Design Tools - Part 1: Whole-Building and Building Envelope Simulation Design Tools. Available at: http://tc47.ashraetcs.org/pdf/Presentations/Jacobs_Cincinnati.pdf.

Jacobs, P. & Henderson, H., 2002. State-of-the-Art Review Whole Building, Building Envelope, and HVAC Component and System Simulation and Design Tools, Arlington, VA, USA: Air-Conditioning and Refrigeration Technology Institute.

Komor, P., 1997. Space Cooling Demands for Office Plug Loads. ASHRAE Journal, 39(12), 41-44.

McLain, H. & Goldberg, D., 1984. Benefits of replacing residential central air conditioning systems. In Washington DC, USA: American Council for an Energy-Efficient Economy, pp. E226 - E227.

NEEA, 2004. Assessment of the Commercial Building Stock in the Pacific Northwest, Portland, OR, USA: Northwest Energy Efficiency Alliance. Available at: http://www.nwalliance.org/research/reports/125.pdf.


NEEA, 2005. Light Commercial HVAC, Portland, OR, USA: Northwest Energy Efficiency Alliance. Available at: http://www.nwalliance.org/research/reports/143.pdf.

Neme, C., Proctor, J. & Nadel, S., 1999. Energy savings potential from addressing residential air conditioner and heat pump installation problems, Washington DC, USA: American Council for an Energy-Efficient Economy.

Parken, W.H. et al., 1985. Field performance of three residential heat pumps in the cooling mode, Gaithersberg, MD, USA: National Bureau of Standards.

PGE, 1997. Commercial Building Survey Report, California, USA: Pacific Gas and Electric Company.


A Survey and interview result

A.1 Survey resultTable A.1 below shows the response for the survey.


Table A.1 Survey Response

No Building Name

Description Year Total SF (office only)

Total SF (excluding office)

Total Tons

(office only)

Total Tons

(excluding office)

Heating source

Heating capacity (Btu/hr)

Glazing Area

Firm A Building 1 One story office building 2000 13700 0 41 0 GAS 664,000 HIGH

Firm A Building 2 One story office building 2001 4400 0 16 0 GAS 240,000 MEDIUM

Firm A Building 3 One story office/exam building 1995 9300 8400 26 22 GAS 918,000 MEDIUM


Firm A Building 5 Polygon Two storey office building 2000 28900 0 69 0 GAS 1,603,000 MEDIUM

Firm A Building 6 Polygon One story office/garage building 2005 6300 1600 21 0 GAS 440,000 MEDIUM

Firm A Building 7 Two storey office building 2006 2800 0 8 0 GAS 183,000 LOW

Firm A Building 8 One story office/garage building (with mezzanine) 2005 10500 7832 37 0 GAS 632,000 LOW


Firm B Building 10 Office 2001 28800 0 76.5

Firm B Building 11 Office 1999 11125 32

Firm B Building 12 Office 2006 4875 27915 10 50

Firm C Building 13 Office 1999 10878 36

Firm C Building 14 club house LOST 1999

Firm C Building 15 High school 1999 91054 65


Firm C Building 17 Medical office 2004 16527 37

Firm C Building 18 Industrial office 2004 46515 59.5

Firm C Building 19 Medical office 2004 33460 45


Firm C Building 21 Bank 2004 3459 11

Firm C Building 22 Office 2001 6194 18

Firm D Building 23 Office

Firm E Attempted to fill in the survey, but has not building that fits into the category

Firm F No response

A.2 Interview result

Table A.2 Interview Result with Firm A for Building A

Assumptions Remarks BUILDING ABuilding Geometry

Did you use printed drawing? Yes

Did you use a CAD file? No

What were geometrical simplifications used in sizing calculations?

Average roof heights was used.

Building Material

Who specified the materials? Architect

Was it assumed to be code compliance? If yes, what code?

Yes, IECC 2003

How was this information input into the load calculation? U.A.deltaT?

-

Was thermal mass somehow included in the load calculation?

Yes

Glazing What is the glazing specs? VLT? SHGC? -

What is the WWR? <40%

Was there any shading? Was it included in the calculation?

Yes

Load Calculation Method

What method was used? CLTD? RTS? Simulation Software?

CLTD(ELITE software)

Zoning What are the main considerations? One big unit or several small units?

Several small Units

Is there sub-zoning? What system was used? CVVT?

No

Outdoor Design condition

What specific conditions were used for peak sizing? Air temp and RH?

Air temperature

Was an hourly weather file used? From what data source?

No

Lighting load Was there a lighting designer on the project?Yes

What was the lighting power density used? -

Was another assumption used? From where?No

Occupancy load Who determine the occupancy? Did the architect? Did someone else describe the occupancy? If not, how did you determine occupancy for sizing calculations?

Based on number of workstation

Was other assumption were used? How were they generated?

No


Assumptions Remarks BUILDING AEquipment load What assumptions were used? How were

they determined?150 watts per workstation

Any other loads?

Specify what load, what number was used, and how they were determined.

No

Outside air flow What number was used and how was it determined?

20 cfm/person

Design airflow What is the design airflow based on for the sizing calculation?

Proportioned by sensible load

Safety factor Was a safety factor applied in this project? If so, what %?

No

Was the safety factor applied at the end of the calculation?

-

Was the safety factor applied as part of the assumptions?

Yes

Design “dynamic”

Was there a set budget for mechanical systems ($/sf)?

No

Was the project open bidding? Pre-existing relationship? With owner? With architect?

Design build

What is the duration of the project? Any time constraint for HVAC?

4 weeks

Did the owner dictate a certain indoor condition?

No

Any mandate to accommodate future growth? How

No

Was the installed capacity “upsized” to the next available unit?

Yes

Did the HVAC contractor changed the capacity?

No

Did the owner has special requirement that changed the design?

No

Construction, operation and maintenance “dynamic”

Was there a commissioning? Startup procedure? Who is the commissioning agent?

Start-up by Mechanical contractorAir balance only

If yes, is there any concern during commissioning?

-

Is there any report/complaint from the tenants/users?

-


Table A.3 Interview Result with Firm B for Building C and D

Assumptions Remarks Building C (2000) Building D (2006)Building Geometry

Did you use printed drawing? Yes Yes

Did you use a CAD file? Yes Yes


No No

Building Material

Who specified the materials? Architect Architect


No energy code IECC 2003


U.A.deltaT U.A.deltaT


No No

Glazing What is the glazing specs? VLT? SHGC? U-0.48, SC-0.63 U-0.3, SC-0.44

What is the WWR?


No Yes, No it was not used in the calculation



CLTD CLTD


Small Units Small Units


No No



Outsise drybulb temperature-96˚F

Outsise drybulb temperature-97˚F


No No

Lighting load Was there a lighting designer on the project?

Yes Yes

What was the lighting power density used? 1.5 W/sqft as per ASHRAE 90.1-1999

1.3 for the selected zone as per residential IECC 2003

Was another assumption used? From where?

No -

Occupancy load

Who determine the occupancy? Did the architect? Did someone else describe the occupancy? If not, how did you determine occupancy for sizing calculations?

Furniture plan or Max occupancy per IMC.

Furniture count or ASHRAE 62.1Owner/Architect provided the occupancy schedules.


No NA

Equipment load

What assumptions were used? How were they determined?

1.25 W/Sqft, Old in house standard

-


Assumptions Remarks Building C (2000) Building D (2006)Any other loads?


NA NA

Outside air flow

What number was used and how was it determined?

IMC, 20cfm/person 20 cfm/person , 1330 Total (140 for the selected zone- RTU 4)

Design airflow What is the design airflow based on for the sizing calculation?

20˚ ΔT sensible heat gain

20˚ ΔT sensible heat gain

Safety factor Was a safety factor applied in this project? If so, what %?

Not in percentages, some conservative assumptions

No


NA NA


Some conservative assumptions

NA



- -


Selective bidding, proir relation with Architect,

Open bidding, Prior relation with Architect and owner.


- 5 months designing, 8 months for construction, No special HVAC constraints.


No No


No No


Yes Yes


No No


Raised floor,but no mechanical impact

No



No Yes there was commissioning, startup contractor <firm name deleted>


- -


No -


Table A.4 Interview Result with Firm C for Building A and B

Assumptions Remarks Building A (2000) Building B (2006)

Building Geometry

Did you use printed drawing? No No

Did you use a CAD file? Yes Yes


No, Simple rectangles

None, Rectangle

Building Material

Who specified the materials? Architects Architect


Yes, Either IBC or UBC

International Building Code 2006


U Value/R Value R Value


No No

Glazing What is the glazing specs? VLT? SHGC? U-.59, SC-0.81 U-0.35, SC-0.72

What is the WWR?


Yes, there is shading; No, it was not included in the calculations

No



CLTD CLTD


One big zone One big zone based on use


No No



Air temperature and time of the day

Air temperature and Time of day


No No

Lighting load Was there a lighting designer on the project?

Yes Yes

What was the lighting power density used? Yes 1 w/sqft

Was another assumption used? From 2w/sqft as per No



where? ASHRAE

Occupancy load

Who determine the occupancy? Did the architect? Did someone else describe the occupancy? If not, how did you determine occupancy for sizing calculations?

Architect , Code and ASHRAE

ArchitectMechanical code based on sqft of zone


None None

Equipment load

What assumptions were used? How were they determined?

Computer Load-ASHRAE

None

Any other loads?


No Computer/Equipment loads information from the owner

Outside air flow

What number was used and how was it determined?

Code, per person 20cfm/person/percode/ASHRAE

Design airflowSafety factor

What is the design airflow based on for the sizing calculation?

1600 cfm 1600 cfm

Was a safety factor applied in this project? If so, what %?

Yes, 10% in load calculation only

Yes, 10%


Yes, Safety Factor was taken out when equipment was selected and installed

Yes


Yes No



- Yes, $ 15/sqft


Open bidding for MEP

Open bidding for MEP


2months, No 3 months , no constraints


No Yes, Standard office conditions


No Yes, shell and core space built




Yes


No No


No No



Startup procedure by Mechanical contractor

Yes, <firm name deleted>


- No


- No


B Measurement results

B.1 Introduction

B.2 The outdoor temperature comparisonThis appendix shows the outdoor temperature during the measurement periods. The measured outdoor temperatures are compared with the TMY weather data and the measured data at the Boise airport for the same day.

The Boise Airport weather data is downloaded from WunderGround website.

Figure B.1 Outdoor temperature comparison – Building A


Figure B.2 Outdoor temperature comparison – Building B

Figure B.3 Outdoor temperature comparison – Building B


Figure B.4 Outdoor temperature comparison – Building C

Figure B.5 Outdoor temperature comparison – Building D


Figure B.6 Outdoor temperature comparison – Building D

Figure B.7 Outdoor temperature comparison – Building E-1


Figure B.8 Outdoor temperature comparison – Building E-2

Figure B.9 Outdoor temperature comparison – Building F


Figure B.10 Outdoor temperature comparison – Building G

Figure B.11 Outdoor temperature comparison – Building H


B.3 Measurement results

B.3.1 Building A

B.3.1.1 Building and RTU Description: Building A

Building description

The building is a single story narrow office building. The total floor space area is 6,194 ft2. The year of design is 2001. It was not designed to meet any energy code.

Zone description

The measured space is a combination of 3 rooms and is served by a single HVAC system. The rooms are a conference room, a single office and a storage. The bulk of the cooling load comes from the conference room.

The zone is located in the northwest corner of the building, with its north and west walls exposed to external environment. All the windows have a fixed external shading.

The total floor space area served by this RTU is 915 ft2. Table B.1 shows the description of the RTU for Building A.

Table B.1 RTU Description for Building A

RTU Building AManufacturer Carrier 48TF-005Tons 4Compressor/control 1/single stageEER 11CFM 1600


B.3.1.2 Measurement result: Building A

Figure B.12 Measurement Result – Building A (11-Aug)

B.3.1.3 Data analysis: Building A


Figure B.13 Nmax calculation for Building A

Table B.2 Summary of Measurement Result for Building A

unit τ=30 τ=60

Number of cycles # 15 15Cycling rate (Ave) cycle/hour 1.27 1.27Cycling rate (Max) cycle/hour 2.66 2.66RTF (ratio) 0.15 0.15PLR (ratio) 0.224 0.207EERavg (Btu/hr)/W 10.23 9.46Energy Penalty % 7.54 16.27Oversized % 346.15 383.26Peak-load penalty W 3386 3461Peak-load penalty kW/ton 0.846 0.865


B.3.2 Building B

B.3.2.1 Building and RTU Description: Building B


The building is a irregular polygonal two story clinical building. Top floor of a two story building is dedicated to clinical services, while the basement is dedicated to clinical wards. The total floor space area is 16,527 ft2. The year of design is 2004. It was not designed to meet any energy code.

Zone description

The measured space is a rectangular double height structure, located at the second floor level of the building. The nurse station is an internally dominated zone, with walls and floor being adiabatic. The roof is the only heat transfer surface in the building. The total floor space area served by this RTU is 1368 ft2. Table B shows the description of the RTU.

Table B.3 RTU Description for Building B

RTU Building BManufacturer Carrier 48HJ-005

High EfficiencyTons 4Compressor/control 1/single stageEER 13CFM 1600

B.3.2.2 Measurement result: Building B


Figure B.14 Measurement Result – Building B (14-Jul)



B.3.2.3 Data analysis: Building B

Figure B.18 Nmax calculation for Building B

Table B.4 Summary of Measurement Result for Building B

unit τ=30 τ=60

Number of cycles # 32 32Cycling rate (Ave) cycle/hour 1.63 1.63Cycling rate (Max) cycle/hour 2.66 2.66RTF (ratio) 0.55 0.55PLR (ratio) 0.754 0.746EERavg (Btu/hr)/W 12.29 11.97Energy Penalty % 2.55 5.24Oversized (ave) % 99.42 105.48Oversized (peak) % 32.58 34.09Peak-load penalty W 936 969Peak-load penalty kW/ton 0.234 0.242


B.3.3 Building C

B.3.3.1 Building and RTU Description: Building C


The building is a polygon shaped single story office building. It has two wings on either side, with a double height lobby space. The total floor space area is 13,700 ft2. The year of design is 2000. It was not designed to meet any energy code.

Zone description

The measured space is a section of the right wing, with its north western wall exposed to external environment. The rectangular space is a part of an open office plan served by a separate HVAC system. The total floor served by this RTU is 1242 ft2. Table C shows the description of the RTU.

Table B.5 Measured RTU for Building C

RTU Building CManufacturer Bryant 580 DPV

048 074 Tons 4Compressor/control 1/single stageEER 10.9CFM 1600

B.3.3.2 Measurement result: Building C

Figures B.19 to B.23 show the measurement results use for the analysis. There are a total of five days of data.

The data shows that the RTU is constantly cycling. The compressor cycled 161 times over the period of 32 hours of data analyzed.

The fan mode of operation is not continuous, but cycle together with the compressor. The RTU has no economizer mode.


Figure B.19 Measurement Result – Building C (19-Aug)



B.3.3.3 Data analysis: Building C

Figure B.24 Nmax calculation for Building C

Table B.6 Summary of Measurement Result for Building B

unit τ=30 τ=60



B.3.4 Building D

B.3.4.1 Building and RTU Description: Building D


The building is a single story office building with a garage. The building has enhanced daylighting features. It has a central double height hallway, which leads to either side of the building. The total floor space area is 6,300 ft2. The year of design is 2005.

Zone description

The measured zone is an irregular polygonal office space, having 2 huge skylights on a sloped roof surface. It is located in the southeastern corner of the building with its south and eastern walls exposed to external environment. The total floor served by this RTU is 588 ft2. Table B.7 shows the description of the RTU.

Table B.7 Measured RTU for Building D

RTU Building DManufacturer Trane T/YSC036-

E3 Tons 3Compressor/control 1/single stageEER 10.9CFM 1200

B.3.4.2 Measurement result: Building D


Figure B.25 Measurement Result – Building D (20-Aug)



B.3.4.3 Data analysis: Building D

Figure B.33 Nmax calculation for Building D

Table B.8 Summary of Measurement Result for Building D

unit τ=30 τ=60



B.3.5 Building E – RTU7

B.3.5.1 Building and RTU Description: Building E


The building is a single story office building. The total floor space area is 28,800 ft2. The year of design is 2001.

Zone description

The measured zone served by RTU 7, is located in the northeastern corner of the building, with its north and eastern side exposed to external environment. It has a large open square shaped area, which includes entrance lobby and restrooms in the northeast corner. The total floor served by this RTU is 3060 ft2. Table B.7 shows the description of RTU7.

Table B.9 Measured RTU for Building E

RTU7 Building EManufacturer YORK

D7CG036N040 Tons 3Compressor/control 1/single stageEER 9CFM 1200

B.3.5.2 Measurement result: Building E


Figure B.34 Measurement Result – Building E-7 (30-Jul)

Figure B.35 Measurement Result – Building E-7 (31-Jul)


Figure B.36 Measurement Result – Building E-7 (01-Aug)



B.3.5.3 Data analysis: Building E

Figure B.38 Nmax calculation for Building E-7

Table B.10 Summary of Measurement Result for Building E – RTU7

unit τ=30 τ=60



B.3.6 Building E – RTU6

B.3.6.1 Building and RTU Description: Building E – RTU6


The building is a single story office building. The total floor space area is 28,800 ft2. The year of design is 2001.

Zone description

The measured zone served by RTU 2 is located in the middle of other zones, with north wall exposed to external environment. It serves the narrow enclosed office spaces. The total floor served by this RTU is 932 ft2. TableB.11 shows the description of RTU6.

Table B.11 Measured RTU for Building E – RTU6

RTU 6 Building EManufacturer YORK 259335-

YTG-B-0606Tons 10Compressor/control 2/dual stageEER 10.9CFM 1600

B.3.6.2 Measurement result: Building E – RTU6



B.3.6.3 Data analysis: Building E – RTU6


Figure B.42 Nmax calculation for Building E-6

Table B.12 Summary of Measurement Result for Building E – RTU6

unit τ=30 τ=60



B.3.7 Building F

B.3.7.1 Building and RTU Description: Building F


The building is a two story showroom building. The second floor is dedicated to office space which is overlooking the showroom below. The total floor space area is 4,875 ft2. The year of design is 2006.

Zone description

The measured zone is an irregular shaped office space, located in the north eastern corner of the second floor, with its north and eastern walls exposed to external environment. The zone is part of an open office plan overlooking the space below. The total floor served by this RTU is 1762 ft2. Table B.13 shows the description of the RTU.

Table B.13 Measured RTU for Building F

RTU Building FManufacturer Bryant 580F073150Tons 6Compressor/control 1/single stageEER 10.1CFM 2400

B.3.7.2 Measurement result: Building F


Figure B.43 Measurement Result – Building F (11-Aug)

B.3.7.3 Data analysis: Building F

Figure B.44 Nmax calculation for Building F


Table B.14 Summary of Measurement Result for Building F

unit τ=30 τ=60

Number of cycles # 3 3

Cycling rate (Ave) cycle/hour 0.32 0.32

Cycling rate (Max) cycle/hour 1.13 1.13

RTF (ratio) 0.9 0.9

PLR (ratio) 0.997 0.996

EERavg (Btu/hr)/W 6.21 6.19

Energy Penalty % 0.29 0.59

Oversized (ave) % 57.75 59.85

Oversized (peak) % 0.33 0.44

Peak-load penalty W 10 14

Peak-load penalty kW/ton 0.003 0.005


B.3.8 Building G

B.3.8.1 Building and RTU Description: Building G


The building is a rectangular two story narrow office building. It has enclosed office spaces on the periphery and open office spaces in the center. The total floor space area is 11,125 ft2. The year of design is 1999.

Zone description

The measured space is a combination of open and enclosed office space, located in the north eastern corner of the second floor, with its north and eastern walls exposed to external environment. The total floor served by this RTU is 1,586 ft2. Table B.15 shows the description of the RTU.

Table B.15 Measured RTU for Building G

RTU 6 Building GManufacturer Carrier 48TJE007-

521Tons 6Compressor/control 1/single stageEER 9CFM 2400


B.3.8.2 Measurement result: Building G

Figure B.45 Measurement Result – Building G (24-Jul)

Figure B.46 Measurement Result – Building F (27-Jul)


B.3.8.3 Data analysis: Building G

Note: The Nmax calculation was not attempted because the compressor was always ON (and never cycled OFF) during the day after it was ON at the beginning of the day.

Table B.16 Summary of Measurement Result for Building G

unit τ=30 τ=60

Number of cycles # 3 3Cycling rate (Ave) cycle/hour 0.12 0.12Cycling rate (Max) cycle/hour 0 0RTF (ratio) 1 1PLR (ratio) 0.999 0.998EERavg (Btu/hr)/W 6.22 6.21Energy Penalty % 0.1 0.2Oversized (ave) % 0.12 0.23Oversized (peak) % 0.08 0.17Peak-load penalty W 3 5Peak-load penalty kW/ton 0.001 0.002


B.3.9 Building H

B.3.9.1 Building and RTU Description: Building H


Two stories office building.

Zone description

The measured space is an office space, located in western part of second floor. The total floor served by this RTU is 2400 ft2. Table B.17 shows the description of RTU6.

Table B.17 Measured RTU for Building H

RTU 6 Building HManufacturer Trane

YCD210C3LCCATons 17.5Compressor/control 2/dual stageEER 11CFM 7000

B.3.9.2 Measurement result: Building H


Figure B.47 Measurement Result – Building H (26-Aug)



B.3.9.3 Data analysis: Building H

Figure B.50 Nmax calculation for Building H – Compressor 2 Stage 1 (26-Aug)

Table B.18 Summary of Measurement Result for Building H – Compressor 2 Stage 1 (26-Aug)

unit τ=30 τ=60





unit τ=30 τ=60





unit τ=30 τ=60

Number of cycles # 6 6Cycling rate (Ave) cycle/hour 1.88 1.88Cycling rate (Max) cycle/hour 2.51 2.51RTF (ratio) 0.27 0.27PLR (ratio) 0.386 0.364EERavg (Btu/hr)/W 10.37 9.73Energy Penalty % 6.12 13.04Oversized (ave) % 309.28 337.99Oversized (peak) % 158.82 175Peak-load penalty W 11715 12149Peak-load penalty kW/ton 0.669 0.694


rightsizing of rooftop hvac system - · pdf fileright sizing of rooftop hvac systems,...

Documents