building serv. eng. res. technol. operational energy use andorca.cf.ac.uk/82172/1/building serv eng...

Special symposium issue

Operational energy use andpower demands in EuropeanHVAC components

IP Knight

Abstract

This paper is presented in the context of a gradual move from energy management at a Building level to

energy management at Component level. This move requires specific energy consumption and power

demand benchmarks to allow measured component energy use and power demands to be put into

context. The paper presents a methodology for collecting and collating the data needed for producing

these benchmarks from measured operational data. This methodology was tested in the European

iSERVcmb project. Example benchmarks from iSERVcmb are provided along with links to the full published

data. The paper concludes that large scale measurement and comparison of detailed operational data are

possible and useful in support of achieving operationally low energy buildings.

Practical application: The explosion of low-cost operational data availability at the level of individual

components can be used to improve understanding of their operational energy efficiency in the context of

the overall efficiency of a building. The paper shows this data can:

. Be collected from all types of buildings across all EU Member States

. Produce benchmark ranges of operational energy consumption and power demands at the level of individualcomponents serving given activities

Applying this benchmarked data in the design and operation of buildings will:

. Improve their operational energy efficiency,

. Enable legislation using this data to show compliance

Keywords

Energy, power, benchmarks, HVAC, components, measurements, Europe, standards

Introduction

This paper examines how detailed energy con-sumption data can be used to help producebenchmarks of operational energy use and

Welsh School of Architecture, Cardiff University, UK

Corresponding author:

IP Knight, Cardiff University, Bute Building, King Edward VII

Avenue, Cardiff, CF10 3NB, UK.

Email: [email protected]

Building Serv. Eng. Res. Technol.

0(0) 1–15

! The Chartered Institution of Building

Services Engineers 2015

DOI: 10.1177/0143624415613954

bse.sagepub.com

at Cardiff University on November 24, 2015bse.sagepub.comDownloaded from

http://bse.sagepub.com/

power demands in building HVAC components.It is part of a wider body of work startedin the IEE HARMONAC Project (www.harmonac.info) and continued through into theiSERVcmb project (www.iservcmb.info). Theindividual HVAC component operational per-formance data derived from the measured datafor iSERVcmb are available to download fromthe iSERVcmb website.

The paper presents the methodology used forcollecting and collating detailed, physically-based data about HVAC components and thebuildings in which they operate. The use of acommon framework is a pre-requisite to enablecomparisons to be drawn between the consump-tion and the power demands of individual com-ponents in different buildings.

The use of Operational Data to BenchmarkBuildings is now gaining commercial credence,with schemes such as the Australian NABERSRating System1 using operational data at theircore, and the mass collection of building oper-ational data underpinning the database2 operatedby the US Office of Energy Efficiency &Renewable Energy. CIBSE also provide a guideto evaluating the operational energy performanceof buildings3 and there is a growing body of workaround the world that is now using these newoperational data streams to provide a betterunderstanding of the use of energy in buildingsand systems. From a bottom-up viewpoint,manufacturers are also attempting to leveragethis information in the form of add-on commer-cial products with companies now starting tomarket the abilities of their products to providereal-time insights into their performance.However, this information often exists with nocontext of what is being serviced by the compo-nent, and therefore whether the data provided areindicating good or poor behaviour from anenergy viewpoint.

At present, iSERVcmb is believed to be theonly approach currently benchmarking oper-ational performance at the level of individualcomponents and activities, right up to wholebuildings and Estates.

This look at detailed operational energy useand power demand data is timely becauselow cost energy use data at the level of individ-ual components and processes is now becomingavailable from all aspects of the built environ-ment. This new data availability heralds aninevitable move from theoretical assessment ofbuilding energy performance to measuredassessment, with all the benefits that the cer-tainty of measurement should bring.

The aims of iSERVcmb in the context of thispaper were to establish firstly that it was prac-tical to collect and collate data at this level ofdetail, secondly to understand which parts of theprocedure were simple and which problematicfrom a practice viewpoint and, finally, to estab-lish a first set of published data on the measuredenergy and power demands of the HVAC com-ponents in the participating buildings

As this is the first time that a comparison ofEuropean-wide continuous sub-hourly datacollection on HVAC component energy use ser-vicing given end use activities has beenundertaken, the data produced should be con-sidered only as a first indication of the energyconsumption and power demands beingachieved by these components in practice inoperational buildings.

Whilst not the subject of this paper directly, itshould be noted that a driving motivation forthe iSERVcmb project was the observationfrom HARMONAC that providing detailedinformation on the operational energy use andpower demands of individual HVAC compo-nents often led to rapid improvements in theirperformance, as the operators could see clearlywhat needed to be addressed. Data obtained atthis level of detail can therefore be seen to beapplicable to both energy management practiceand legislative procedures via benchmarking.

There are many other operational and man-agement benefits of having this data, includingbusiness continuity, cost-optimality calculations,etc., but the rest of this paper focusses on theprocess of obtaining the energy and powerdemand data.

2 Journal of Building Services Engineering Research & Technology 0(0)


www.harmonac.info

www.harmonac.info

www.iservcmb.info


Background

The Energy Performance of Buildings Directive(EPBD)4 legislation to formally provide moreaccountability for energy use in EuropeanBuildings was introduced over 11 years ago in2002. This had at its heart measures and require-ments which were suitable for the technologyand knowledge of building energy use at thetime. It also reflected the skills available inthe marketplace, with an eye on improvingrequirements over time.

Since 2002, however, we have had a revolutionin computing, metering and communication. It isnow commonplace for us to be able to use mobiledevices to control and access information fromalmost anywhere in the world. Despite their rela-tively slow pace of change compared with othermarkets, Buildings Services have not beenimmune to this change and we can now usemobile phones to turn our domestic heating onor off, or to interrogate individual items ofBuilding Services plant.

It therefore seems an appropriate time toreappraise the information we have availableto us to understand and control the energy useof our buildings, and in particular the energy useof Heating Ventilation and Air Conditioningsystem components, to see if there is a moreeffective way of assessing building energy usewhich allows better understanding of opportu-nities for reducing this consumption.

The current mainstream approach to under-standing energy end use in buildings has been tomodel the building and then use either top-downor bottom-up approaches to estimate where theenergy is being used e.g. CIBSE TM22.5 Thishas led to the emergence of the ‘performancegap’ term where the energy consumed by thebuilding as a whole does not match that pre-dicted by these models and techniques.

There is a wide body of research dealing withthe need to use measured data to calibratemodels to reduce this gap, with some examplesgiven in the references.3,6–9 At present, theirmain conclusions are that it is possible to tune

some building models to accurately predict whatthe measurements are providing – meaningthat for those particular buildings the elementscreating the energy demand are reasonably wellunderstood. However, most buildings do not yethave models and are unlikely to have access tothe resources and expertise to undertake such anexercise.

Other research examining the development ofbenchmarking methodologies10–13 shows thatalmost all benchmarking approaches are under-pinned by modelling. There are at present nobuilding benchmarking and performance assess-ment approaches implemented which useoperational data to provide both the bench-marks and assessment. This paper forms partof a procedure which proposes to introducesuch an assessment approach.

The approach used in this paper arises fromresearch at both the UK and European level. TheEU IEE AUDITAC project14 aimed to show thelikely impact on energy use in AC systems ofInspection as required by EPBD Article 9.4

AUDITAC concluded that the EPBD wouldnot have the impact on energy use in AC systemshoped for, mainly due to a lack of skilledInspectors to implement Article 9 across theAC systems falling within its scope but alsodue to a severe lack of independent data onthe actual energy use of AC systems in oper-ational buildings. There was also a concern(borne out by the subsequent development ofthe Inspection market into primarily a compli-ance market) about the lack of feedback on theimpact of Inspection as there was no mechanismfor assessing energy savings achieved as a resultof Inspection.

The IEE HARMONAC project15 was pro-posed and coordinated by the author between2007 and 2010. It obtained information fromacross Europe on the energy savings identifiedby Inspections compared with the energy sav-ings identified from detailed energy metering of42 AC systems across Europe. One of the mainconclusions was that a good Inspection wouldon average identify only �37% of the energy

Knight 3



savings that sub-hourly energy metering sug-gested existed. However, perhaps more import-antly, it was clear from HARMONAC thatowners would act on the detailed meter data toimplement many of the savings identified, as thebenefits could be clearly calculated and hencethe investment risk was reduced.

iSERVcmb16 built on the findings fromHARMONAC. It proposed that empirical datafrom automatic metering and physical observa-tions should be obtained at the level of individualHVAC components. This data, in conjunctionwith knowledge of the activities served by thesecomponents and other parameters, such as thefloor area being serviced, allowed benchmarkranges of achieved performance to be produced,enabling comparisons to then be made acrossbuildings and systems.

An important part of the procedure is thatiSERVcmb obtains the area and activity beingserviced by an HVAC component through con-necting each HVAC component via an HVACsystem to the specific areas they service withinbuildings. This means that the energy and thepower demand benchmarks derived are basedon a more focussed understanding of the actualload being imposed on a component, rather thanwhole building areas which are more commonlyused in benchmarking at present. A consequenceof this is that iSERVcmb can benchmark HVACsystem performance at the level of individualspaces and activities – a feature that is of valuein determining appropriate benchmarks forHVAC systems in buildings which cross trad-itional sectoral definitions. This data also havethe potential to be of value to legislation.

The debate about how we might reappraise ourmethod of assessing energy use in buildings andsystems is one that the HARMONAC andiSERVcmb projects have presented to Europe asa whole17 and will not be covered in depth in thispaper. However, the project also demonstratedthat the actual energy use in operational buildingscan be reduced significantly and cost-effectivelythrough the use of detailed operational energyuse data at the level of individual HVACcomponents.

The value of detailed operational data inachieving operational energy efficiency hasalready been recognised by the EuropeanCommission following the finding from theIEE HARMONAC project that detailed moni-toring can help identify and reduce excessiveenergy consumption in Air-Conditioning sys-tems. As a result, the recast of the EPBD in201018 incorporates the requirement forMember States to ‘encourage the introductionof intelligent metering systems’ in Article 8 aswell as specifically noting in Articles 14 and 15(relating to the Inspection of Heating and Air-Conditioning Systems) that ‘Member Statesmay reduce the frequency of such inspectionsor lighten them as appropriate, where anelectronic monitoring and control system is inplace’.

The rest of this paper deals only with theissues surrounding the derivation of compara-tive energy use and power demands in HVACcomponents and systems.

Data collection methodology

Data collection – spaces, activities,meters and HVAC

The iSERVcmb project revealed that the great-est uncertainty in the use of operational data iswhen the data themselves are clearly ‘wrong’ forsome reason. This uncertainty most often comesfrom the following:

. Poor data collection techniques and proto-cols, including frequent meter errors andmissing data (which is particularly import-ant for pulse meters which simply recordconsumption per interval, not overallconsumption)

. Poorly installed meters, leading to largerinaccuracies than expected

. Poor understanding of the installed meteringsystems and what they are connected to

. Lack of certainty over where the HVAC com-ponents/systems the meters are connected toactually serve




The key to achieving meaningful results, aswith any comparative process, is to minimisethe uncertainty in the data that underpins thebenchmark ranges we intend to create. A stand-ard methodology for collecting this data wastrialled during iSERVcmb. An Excel spread-sheet (available for free from www.iservcmb.infoor www.k2nenergy.com) was produced as aframework for describing a buildings’ HVACcomponents, the spaces and activities theyserve, and their meters. Figure 1 shows howthe iSERVcmb spreadsheet is structured to cap-ture these details and, as importantly, connectthe various elements to each other. The figurealso shows many of the fields to be completed.

Figure 2 shows in graphical form how theasset data in the spreadsheet is inter-connected.

The feedback from the project for thisapproach from end users is promising. The con-cept is easy to understand as it relies on physicaldata which can be described, measured and ver-ified on site – which means the data have valuebeyond the task of helping establish bench-marks. Essentially, the spreadsheet acts as anenhanced asset register for many of the physicalassets of the building that influence its energyconsumption. The main omission, for practicalreasons, were most of the physical descriptions

of the building and its fabric, meaning that atthis stage the framework could not be used tomodel heating and cooling demands for a build-ing in much detail. However, given that we wererecording the actual energy use of the buildingwhich incorporated all the interactions betweenoccupancy, climate and fabric then this was con-sidered an acceptable compromise at this stage.

The main hurdles to overcome for the wide-spread use of the process are:

. It became clear from undertaking theHARMONAC and iSERVcmb studies thatfew buildings in Europe have records of thisbasic information, let alone understand howtheir services work in detail. This means thereis a time investment needed to provide theinitial details required. There is a case toargue that this information should beknown by any professional before anyenergy efficiency retrofit measures areattempted, but practical and economic rea-sons have meant that these have not beenvalued in operational buildings until recently.

. With any measurement and descriptive exer-cise there is also the scope for uncertainty.This includes the choice of activity to describethe primary use of a space and the more

Figure 1. Empty iSERVcmb spreadsheet showing some of the data fields to be collected.

Knight 5


www.iservcmb.info


obvious potential for errors in the measure-ment of area and energy use already noted.

. Achieving robust operational data collectionand transmission on a continuous basis

Data collection – energy consumptionand sensor data

Having collected and collated each building’sphysical elements, iSERVcmb then collectedthe sub-hourly data for its ELECTRICALenergy meters, and any provided sensors, atleast once a month. The iSERVcmb project col-lected information from 330 buildings in 20 EUMember States. The data covered 2 831 HVACsystems, 7 685 HVAC components, 2 230m, 11173 Spaces, 72 different Activity Types and 1551 638 m2 of floor area. The sub-hourly datacollected for these systems covers from a fewmonths to over 5 years, including some historicdata. Most of this information is at 15 or 30mintime intervals.

Figure 3 shows an example of this data in theform of a 15min carpet plot, taken fromOctober 2012. It shows the measured consump-tion through the electricity meter serving aPortuguese HVAC system consisting ofChillers, Pumps and AHUs. This carpet plotseparates the data into 10 energy use bands,enabling periods of high and low energy use tobe quickly identified by day and time over a spe-cific month.

We can also view this information in the formof scatter graphs showing time of day powerdemand or energy use over an extended period.Figure 4 shows the range of average recordedpower demand per m2 over a year for a completeHVAC system serving the activity of ‘GenericCheck-In’ for each hour of the day for weekdaysin Portugal.

The maximum, minimum and average powerdemands for the system can be clearly seen byhour of the day for the weekdays across a wholeyear. From knowledge, taken from the spread-sheet, of the components, their nominal power

1. The meters, spaces and HVAC systems are described first in the spreadsheet2. Then connect the meters to the HVAC components and spaces they serve3. Then connect the HVAC components to the systems they serve4. Finally, connect the HVAC systems to the spaces they serve

Incomingmeter

Meter 1

Chillers

Small power and lighting

HVAC system 1

HVAC system 2

NB: AN HVAC system is acollection of components

CHWpumps

HW pumps

AHU 1

Room 1

Room 2

Room 3

Room 4

Room 5

AHU 2

Boilers

Meter 2

Meter 3

Meter 4

Meter 5

Figure 2. Connecting the building assets.




ratings and other components on the same meterserving the same end uses, along with analysingthe profiles from various parts of the year, wecan apportion the power demand between thevarious HVAC components that comprise thesystem being metered.

The same information can be used for oper-ational, diagnostic and benchmarking purposes.

Data reliability and accuracy

The production of benchmarks of any kindrequires confidence in the data from which

they are derived – in this instance that includesmeasurements of area, activity, etc., as well asthe recorded consumed energy. For any initialattempt at deriving this information in the realworld then, as noted earlier, there will always beuncertainties, especially when applied across dif-fering cultures in the EU, as with the iSERVcmbdataset.

iSERVcmb used the RICS definition of GrossInternal Area19 ‘the area of a building measuredto the internal face of the perimeter walls at eachfloor level’ as its measure of area – which meansthat individual spaces should be measured to the

C_ANAM_CHILLER SUL, UTAS 5 E 6 BOMBAS

Date range for plot: 01 Oct 12 to 31 Oct 12

23:00

22:00

Bands of 10%

based on the max value of 63.0 kw

h in one time interval

21:00

20:00

19:00

18:00

17:00

16:00

15:00

14:00

13:00

12:00

11:00

10:00

09:00

08:00

07:00

06:00

05:00

04:00

03:00

02:00

01:00

00:00 01 Oct 12

02 Oct 12

03 Oct 12

04 Oct 12

05 Oct 12

06 Oct 12

07 Oct 12

08 Oct 12

09 Oct 12

10 Oct 12

12 Oct 12

13 Oct 12

14 Oct 12

15 Oct 12

16 Oct 12

17 Oct 12

18 Oct 12

19 Oct 12

20 Oct 12

11 Oct 12

22 Oct 12

23 Oct 12

24 Oct 12

25 Oct 12

26 Oct 12

27 Oct 12

28 Oct 12

29 Oct 12

30 Oct 12

31 Oct 12

21 Oct 12

No Data0.0 - 6.3

6.3 - 12.6

12.6 - 18.9

18.9 - 25.2

25.2 - 31.5

31.5 - 37.8

37.8 - 44.1

44.1 - 50.4

50.4 - 56.7

56.7 - 63.0

Figure 3. A Carpet Plot showing the 15-min energy use of a Portuguese HVAC system consisting of Chillers, Pumps

and AHUs in October 2012.

Knight 7



centreline of any internal walls or partitions.However, floor area within buildings submittedto the database may not always adhere to thisstandard. The paper by Warren20 on the world-wide variation in measurement standards offloor area across commercial buildings indicatesthat the largest likely floor area measurementvariation due to different measurement practiceswill be around 5% (4.8% in the paper byWarren) across all areas supplied to the projectdatabase. As iSERVcmb refers to spaces by indi-vidual sizes, it is likely the room areas suppliedhave been taken to the inner faces of the roompartitions rather than the partition centre lines, soit is also more likely that the areas supplied willbe up to 5% too small rather than too large. Theerror ranges calculated for iSERVcmb take thisinto consideration by assuming the actual area isin the range of 99 to 104% of the area provided.

For electricity meters, the project assumesthey have been installed correctly when the read-ings are not clearly erroneous (in which casethese data are queried and checked by the rele-vant iSERVcmb Partner for the EU Member

State) so a general error of �2% has beenallowed for each reading, based on UKOFGEM guidance21 indicating a range of+2.5% to �3.5% to be acceptable in suchmeters. Other significant sources of error includemultiplication errors and corrupted datastreams, though these are relatively simpler toidentify and either correct or discard.

However, the greatest error likely to be pre-sent with the data is associated with uncertaintyover exactly what energy end uses each meterserves. A common finding across Europe, fromboth the HARMONAC and iSERVcmb pro-jects, has been that, along with a lack of detailabout their services, very few buildings have up-to-date records of exactly what is served fromtheir metered distribution boards. When this iscombined with the uncertainty over which areassome of the HVAC systems service, it is clearthat deriving quantitative values for ranges ofnormalised ‘in use’ energy consumption andpower demands for the 2 831 EU HVAC sys-tems in iSERVcmb requires both subjective andobjective approaches at this initial stage.

System power : Weekdays : Portugal, ANAM, P1_PULICO_PARTIDAS50

40

30

20

W/m

2

10

00 1 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 21

12345

22 23Timeslot

10

Figure 4. A Scattergraph plot of the average weekday power demand per m2 by time of day for an HVAC system

serving the ‘Generic Check-In’ activity in Portugal.




For iSERVcmb, the approach used to recordthe in-use energy consumption and powerdemands of HVAC systems and components isempirically-based. At this stage of their develop-ment, benchmarks derived from this data rely onthe subjective judgement and experience of theauthor to interpret the reliability of the objectivedata being collected and produced. This inter-pretation is undertaken with reference to thelarger dataset to ensure outliers are spottedand discussed with the original provider of thedata before deciding whether to include or dis-card the data.

The author has been continuously involved inmonitoring, analysing and predicting the energyuse of operational buildings and HVAC systemssince 1987 and has published numerous outputsranging from UK Energy Consumption Guides,through Building and HVAC system energy usepublications, to the final reports fromiSERVcmb.22 This experience is important toprovide confidence in the initial operationalenergy use and power demand benchmarkranges published by iSERVcmb, examples ofwhich are presented here.

To summarise, iSERVcmb produced a signifi-cant amount of information of the type pre-sented in Figures 3 and 4 for individualoperational HVAC components. These weremetered either individually, or as part of asystem, in buildings located across 20 EU MS.The reliability of this data for use in the bench-marks has been manually examined and queriedwhere it varied significantly from other data forthe same activities and components.

It is important that the benchmarks presentedfrom this data are understood to be an initialview of what is happening in practice, basedon pre-installed metering systems and buildingowners understanding of their Estate assets. Noclaim is made for their absolute accuracy as thiswas not possible to do over such a scale andvariety of cultures within the project fundingand time constraints. What is presented there-fore is what the participating buildings and com-ponents would report as their energy use andassets (including floor areas) if queried.

Operational data benchmarks

A benchmark, for the purposes of this paper, isdefined as a figure or range of figures encom-passing the measured performance of oper-ational components in buildings. They areNOT considered to be STANDARDS to bemet, but the approach could form the basis ofoperational data standards if needed.

Benchmark types

From the iSERVcmb data, it is clear that thereare categories of electrical energy use and powerdemand that are comparable across the EU.There are also other categories, usually to dowith geographical location and climate, whichare better dealt with through self-reference orreference only to local consumption figures.iSERVcmb had insufficient data to enable thislatter analysis, so the range of achieved con-sumption and power demand figures presentedcovers all systems in the EU countries – thoughsome of these figures are also presented at coun-try level on the iSERVcmb website to enablethese figures to be assessed further if needed.

We will further normalise some of the datafor weather data in a later paper, but felt it wasmore important at this stage to produce a set ofactual operational energy use figures with aslittle normalisation as possible to enable rela-tively ‘raw’ operational energy consumptiondata to be available for the debate on whatrole building services should play in achievinglow or zero energy buildings in the near future.The only normalisation used in this paper is byfloor area in m2.

If benchmarks are to have an impact in help-ing reduce the operational energy use of realbuildings and systems then, from iSERVcmband HARMONAC, they need to provide infor-mation in a way that allows clear connections toactions that can be taken to improve energy useor power demand.

We can produce two main types of bench-mark from sub-hourly Operational Data –energy consumption (kWh) and power demands(W). Each can be normalised by a variety of

Knight 9



parameters including floor area served – themost commonly used parameter at present.

The operational energy consumption bench-

mark is used for Operational Buildings in exist-ing procedures, including Energy PerformanceCertificates and Display Energy Certificates,usually normalised by total floor area andassessed over a full year i.e. kWh/m2.a. It hasthe advantage of being easy to calculate for abuilding through using its billed annual energyuse. At this level of detail, it has the disadvan-tages, amongst others, of:

. Insufficient detail to know WHERE actionshould be taken to improve.

. Building descriptions are too vague to allowmeaningful comparison

. Floor area is not a good parameter for bench-marking some important activities, notablyServer rooms, LAN rooms and high densityIT rooms.

. Ignoring the value of the more detailed sub-hourly data often available for the mainmeter

. Requiring a full years’ data before a view onthe operational performance can be derived.This makes it of limited use to commissioningof new build or evaluation of the impact ofchanges.

The main failing of this energy consumptionbenchmark however, is that reductions orincreases in the kWh/m2.a figure for a buildingare very difficult to attribute to their causes. Thismakes it difficult to claim success from a particu-lar energy conservation strategy or to under-stand what has caused the increase i.e. it tellsyou WHAT has happened, but not WHY orWHERE. This leads to increased risk inmaking investment decisions and poor recogni-tion of the value of energy efficiency. This prob-lem has led to the emergence of protocols suchas IPMVP23 to try and remedy this deficiency.

Conversely, an operational power demand

benchmark which looks at peak and averagepower demands from individual components orsystems, can quickly identify where components

or systems are consuming more than would beexpected, and can do so during the importantcommissioning and design phases. This bench-mark is more suited to achieving action in redu-cing the overall energy use of buildings butneeds to be used in conjunction with the oper-ational energy consumption benchmark to pro-vide an overall view of the energy efficiency of aspecific building.

For both benchmark types, the approachused enables benchmark ranges to be producedfrom measured data at the resolution level ofindividual components serving individual activ-ity types. This enables unique, tailored bench-mark ranges for any building, space or system.iSERVcmb showed that this level of detailprompted action to be taken where the bench-marks suggested savings were possible.

Having ranges of achieved performance asthe benchmarks also allows use of the measuredenergy data to predict potential ranges of energyand cost savings for each situation, enablingcost-optimality to be quickly addressed. Thisability to predict savings is key to much existingand future legislation.

Operational energy use and power demandbenchmark figures. The initial benchmarktypes produced from the measured data, foreach HVAC component serving a specific enduse activity, are:

. ‘Annual energy use’ in kWh/m2

. ‘Power in use’ in W/m2

Future benchmark sub-types possible oncemore data sources are available could include:

. ‘Monthly energy use’ in kWh/m2 for eachmonth of the year

. ‘Power in use’ in W/m2 by time of day andmonth of year

These are likely to need to be location andclimate specific.

The data analysis undertaken to produce thefirst benchmarks compared the normalised




0.9

0.8

0.7w

/m2

0.6

0.5

0.4

0.3

0.2

0.1

0.0

8.0

7.0

w/m

2w

/m2

6.0

5.0

4.0

3.0

2.0

1.0

0.0

0.0

5.0

10.0

15.0

25.0

20.0

Min w/m2 (meas)

Average w/m2 (meas)

Min w/m2/a (bench)

Max w/m2 (meas)

Max w/m2/a (bench)

50% w/m2/a (bench)

Number of component plus sub-component samples - all activities

Minimum measured and benchmark power demandsAir handling units: Supply with heating and cooling variants

Average measured and benchmark power demandsAir handling units: Supply with heating and cooling variants

Maximum measured and benchmark power demandsAir handling units: Supply with heating and cooling variants

Number of component plus sub-component samples - all activities

0 5 10 15 20 25 30 35

0 5 10 15 20 25 30 35

0 5 10 15 20 25 30 35Number of component plus sub-component samples - all activities

Figure 5. Example of measured data vs predicted power demand/m2 benchmarks for a Supply AHU component

serving various end use activities.

Knight 11



measured data with initial predictions preparedby the author of HVAC component power andbenchmark ranges when servicing specifiedactivities. This was done to provide a ‘sanity

check’ to the figures produced from the mea-sured data. The 73,000+ initial separate bench-mark predictions were based primarily onexisting confidential commercial datasets

Air handling units - average annual kWh/m2 by activity type served

0Retail warehouse sales area-chilled

Workshop

Lecture theatreDiagnostic imaging

Laundry

Laboratory

Laboratory - sterile

Lounges

Catering: Bars

Open plan office areaMeeting room

Catering: Vending machinesCellular office area - multiple occupation

Dept store sales area - general

Library - reading roomRecreational: Sports ground changing rooms

Recreational: Changing facilities with showers

Recreational: Fitness studioIT: LAN rooms

Bedroom

ToiletGeneric checkin areas

Teaching areasRetail warehouse sales area - general

Circulation area (corridors and stairways)Light plant room

Cellular office area

Waiting roomsIT: Server room

Operation theatreSmall shop unit sales area - chilled

Consulting/treatment roomLibrary - stacks and storeroom

Warehouse storageExhibition rooms, museum

BathroomPhysiotherapy studioLibrary - open stacks

Unoccupied spaceLifts

ReceptionEscalators

Nursery

Hotel roomRecreational: Recreational pool

Multi-storey car parks (office and private use)

Small shop unit sales area - general

Generic ward

Catering: Snack bar with chilled cabinets

Catering: Kitchenette (small appliances, fridge and sink)

Catering: Eating/drinking areaStorage area/cupboard

Stage (theatres and event buildings)

Assembly areas / hallsDept store sales area - chilled

IT: High density IT suite

Recreational: Fitness suite/gymSpectator area (theatres and even buildings)

Catering: Limited hot food preparation area

Catering: Full kitchen preparing hot mealsLaboratory with fume cupboards

282.2278.6

176.1167.8

159.1102.8

87.583.2

7368.9

58.456.956.256

48.641.939.839.538.5

35.434.7

2625.8

20.720.620.519.619.118.918.717.917.717.317.116.414.914.512.21211.61110.910.29.6

7.86.56.15.85.65.44.94.94.73.92.72.62.52.4

0.2

50 100 150 200 300250

Figure 6. Example of normalised average annual energy consumption benchmark by activity and component type.




available to the author, along with reference toother publications dealing with operationalenergy use benchmarks, such as VDI380724

and CIBSE Guide F.25,26

An example of how operational data andthese predicted benchmarks compared for a lim-ited sample drawn from the ‘Air HandlingUnits: Supply with Heating and Cooling vari-ants’ subset of data is shown in Figure 5. The

graphs show the minimum, average and max-imum power demand/m2 values measured andpredicted are generally similar across a numberof end use activities taken from a range of build-ings across the EU from warm to cold countries.

At present all benchmarks produced byiSERVcmb are based on the /m2 parameter forall combinations. These will evolve as the pro-cedure is taken forward. For example, data

Air handling units - average W/m2 by activity type served

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 50.044.1

37.133.5

20.518.5

17.616.8

16.113.5

13.111.911.6

8.68.2

7.86.86.76.6

6.15.85.85.85.85.7

5.45.14.94.74.44.34.34.2

3.83.6

3.33.03.0

2.52.42.42.42.22.22.12.01.71.61.61.61.51.4

1.11.00.80.80.80.60.50.40.30.20.0

45.040.0

Retail warehouse sales area - generalRetail warehouse sales area - electrical

Workshop

Lecture theatre

Diagnostic imaging

Laundry

Laboratory

Laboratory - sterile

Lounges

Catering: Bars

Open plan office area

Meeting room

Catering: Vending machines

Cellular office area - multiple occupation

Dept store sales area - general

Library - reading room

Recreational: Sports ground changing roomsRecreational: Changing facilities with showers

Recreational: Fitness studio

IT: LAN rooms

Bedroom

Toilet

Generic checkin areas

Teaching areas

Retail warehouse sales area - chilled

Circulation area (corridors and stairways)Light plant room

Cellular office area

Waiting rooms

IT: Server room

Operation theatre

Small shop unit sales area - chilled

Consulting/treatment room

Library - stacks and storeroom

Warehouse storage

Exhibition rooms, museum

Bathroom

Physiotherapy studio

Library - open stacks

Industrial process area

LiftsReception

Unoccupied space

Heavy plant room

Post mortern facility

Nursery

Recreational: Recreational pool

Multi-storey car parks (office and private use)

Small shop unit sales area - general

Small shop unit sales area - electrical

Generic ward

Catering: Snack bar with chilled cabinets

Catering: Kitchenette (small appliances, fridge and sink)

Catering : Eating/drinking area

Storage area/cupboard

Stage (theatres and event buildings)

Assembly areas / halls

Dept store sales area - chilled

IT: High density IT suite

Recreational: Fitness suite/gym

Spectator area (theatres and even buildings)

Catering: Limited hot food preparation area

Catering: Full kitchen preparing hot meals

Laboratory with fume cupboards

Figure 7. Example of normalised average power demand benchmark by activity and component type.

Knight 13



collected by iSERVcmb indicate that IT LANrooms and Servers, along with their associatedservices, should be separately metered to enabletheir consumption to be removed from that ofthe building as a whole. This is because volumeor space occupied is a very poor guide to con-sumption efficiency for this activity.

Other activities and processes show similarcharacteristics regarding poor correlation withfloor area, and these will be discussed furtherin other papers as the data are analysed inmore detail. It is important to note that theiSERVcmb procedure is using the operationaldata to show what is happening in practicerather than assuming any form of relationshipwith any parameter. This means that the oper-ational data benchmark ranges produced canshow their derivation very clearly. Improvingthe accuracy of these benchmarks will be atrade-off between which normalisation param-eter is the most accurate and which is the mostpractical to apply in operational buildings.

Figures 6 and 7 are examples from theiSERVcmb project27 of the average annualenergy use and power demand figures foundfor operational AHUs serving various end useactivities around Europe.

From the figures, it can be seen that the mea-sured energy use and power demands generallyconfirm what might be expected for each activityand component pair/m2, providing confidencethat this data provide a reasonable indicationof performance being achieved in use.

Conclusions

This paper presents a first look at deriving EUMember States operational energy consumptionand power demands at the level of individualcomponents and activities in operational build-ings. The key hurdles to overcome in implement-ing such an approach are:

. The initial description of each building in theiSERVcmb template

. Producing standards for products and pro-cesses to meet in providing the data

. Obtaining the support of the main actors inthis area to implement the approach

. Producing legislation that supports theapproach as an alternative or complementto existing Member State legislation

The operational data collected show that theprocedure used in iSERVcmb can be success-fully used in all EU MS to provide informationat this level of detail. This is a key finding as theEU strives to provide practical guidance toachieving operationally low energy buildings inthe transition to a lower energy intensity future.Without information at this level of detail thenachieving and maintaining low energy use andreduced power demands in operational buildingswill be difficult to manage.

The findings are also of great value to build-ing energy modellers as they provide potentialboundary parameters that can be used in themodelling process to improve the accuracywith which models can be used to predict con-sumption and power demands in operationalbuildings.

Further work is already underway to helpaddress many of the hurdles identified, withpractical experience in operating the approachacross a large estate being a key component ofthis work. It is hoped to recruit more Estates tothe approach in the coming months.

Acknowledgments

The author wishes to acknowledge the contributions ofall the Partners and contributors to the iSERVcmb

project in producing the benchmarks presented inthis paper, in particular K2n Ltd for their assistancewith generating the benchmark data. The sole respon-

sibility for the content of this paper lies with theauthor. It does not necessarily reflect the opinion ofthe European Union. Neither the EACI nor theEuropean Commission are responsible for any use

that may be made of the information contained herein.

Declaration of conflicting interests

The author(s) declared no potential conflicts of inter-est with respect to the research, authorship, and/or

publication of this article.




Funding

The author wishes to thank the European IEE pro-gramme for funding the iSERVcmb project.

References

1. NSW Office of Environment and Heritage Australia.

NABERS – national Australian built environment rating

scheme. Sydney: NSW Office of Environment and

Heritage, Australia.

2. United States Office of Energy Efficiency and

Renewable Energy. Building Performance Database.

2015.

3. CIBSE. TM54: Evaluating operational energy perform-

ance of buildings at the design stage. Chartered

Institution of Building Services Engineers, 2013.

4. European Parliament. Directive 2002/91/EC of the

European Parliament and of the Council of 16

December 2002 : on the energy performance of buildings.

S.l: EC, 2003, p. 7 p.

5. CIBSE. TM22: Energy assessment and reporting method,

2nd ed. London: Chartered Institution of Building

Services Engineers, 2006.

6. CIBSE. CIBSE Guide A: Environmental design, 7th ed.

London: Chartered Institution of Building Services

Engineers, 2006.

7. de Wilde P. The gap between predicted and measured

energy performance of buildings: a framework for inves-

tigation. Autom Construc 2014; 41: 40–49.

8. Knight I, Stravoravdis S and Lasvaux S. Predicting

operational energy consumption profiles – findings

from detailed surveys and modelling in a UK educa-

tional building compared to measured consumption.

Int J Ventil 2008; 7: 49–57.

9. Norford LK, Socolow RH, Hsieh ES and Spadaro GV.

Two-to-one discrepancy between measured and pre-

dicted performance of a ‘low-energy’ office building:

insights from a reconciliation based on the DOE-2

model. Energy Build 1994; 21: 121–131.

10. Chung W. Review of building energy-use performance

benchmarking methodologies. Appl Energy 2011; 88:

1470–1479.

11. Hernandez P, Burke K and Owen Lewis J. Development

of energy performance benchmarks and building energy

ratings for non-domestic buildings: an example for Irish

primary schools. Energy Build 2008; 40: 249–254.

12. Holley AM and King A. Worldwide KPIs and bench-

marking. Bracknell: BSRIA, 2006.

13. Perez-Lombard L, Ortiz J, Gonzalez R and Maestre I. A

review of benchmarking, rating and labelling concepts

within the framework of building energy certification

schemes. Energy Build 2009; 41: 272–278.

14. Adnot J. AUDITAC – field benchmarking and Market

development for audit methods in air conditioning.

Final Report to European Commission. European

Commission. Grant Agreement No/ EIE/04/104/

S07.38632. February 2007.

15. Knight I. HARMONAC – harmonizing air conditioning

inspection and audit procedures in the tertiary building

sector. Final Report for the Energy Consumption in

European Air Conditioning Systems and the Air

Conditioning System Inspection Process. Intelligent

Energy Europe Project No: EISAV/EIE/07/132/2007.

European Commission, October 2010.

16. Knight I. Final Report: the inspection of building services

through continuous monitoring and benchmarking – the

iSERVcmb project. Intelligent Energy Europe Project

No. IEE-10-272. European Commission July 2014.

17. Knight I. Automatic Monitoring and Legislation to

meet EPBD requirements. In: Concerted Action III

Project, Vienna, December 2011, Paper prepared for

the European Commission’s Concerted Action III

Project.

18. European Commission. Directive 2010/31/EU of the

European Parliament and of the Council on the energy

performance of buildings (recast), http://eur-lex.euro-

pa.eu/LexUriServ/LexUriServ.do?uri¼CELEX:

32010L0031:EN:NOT2010 (accessed 19 May 2010).

19. Royal Institute of Chartered Surveyors. RICS code of

measuring practice, 6th ed. London: Royal Institute of

Chartered Surveyors, 2007.

20. Warren CMJ. Benchmarking corporate real estate: funda-

mentals of measurement. Pacific Rim Property Research

Journal: Taylor and Francis, 2002, pp. 277–285.

21. OFGEM. OFGEM factsheet – meter accuracy and bill-

ing disputes. OFGEM, 2013.

22. Knight I. Personal webpage – research projects and

publications.

23. Efficiency Valuation Organisation. International per-

formance measurement and verification protocol: con-

cepts and options for determining energy and water

savings Volume 1. EVO 10000 – 1:2012, www.evo-worl-

d.org January (2012).

24. Verein Deutscher Ingenieure. VDI 3807 Part 4 –

Characteristic values of energy and water consumption

of buildings. Characteristic values for electrical energy.

Berlin: Beuth Verlag, 2008.

25. CIBSE. CIBSE guide F – energy efficiency in buildings,

3rd ed. London: The Chartered Institution of Building

Services Engineers, 2012.

26. CIBSE. TM46: energy benchmarks. London: Chartered

Institution of Building Services Engineers, 2008.

27. Knight I. iSERVcmb measured energy consumption

data by HVAC component and activity: European

union as a whole. Intelligent Energy Europe Project

No. IEE-10-272, European Commission, July 2014.

Knight 15


http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32010L0031:EN:NOT2010




www.evo-world.org

www.evo-world.org


building serv. eng. res. technol. operational energy use andorca.cf.ac.uk/82172/1/building serv eng...

Documents