Dt

YC

a

ARRAA

KPFTIAs

1

bteesatc

BbziBcm

MT

h0

Energy and Buildings 84 (2014) 575–585

Contents lists available at ScienceDirect

Energy and Buildings

j ourna l ho me page: www.elsev ier .com/ locate /enbui ld

esign and operation methodology for active building-integratedhermal energy storage systems

uxiang Chen ∗, Khaled E. Galal, Andreas K. Athienitisoncordia University, 1455 De Maisonneuve Blvd. W., Montreal, Quebec, Canada H3G 1M8

r t i c l e i n f o

rticle history:eceived 31 May 2014eceived in revised form 7 August 2014ccepted 10 August 2014vailable online 20 August 2014

eywords:redictive control

a b s t r a c t

A methodology is presented for integrating the design and operation of active building-integrated thermalenergy storage (BITES) systems to enhance their thermal and energy performance. A bounding-conditionbased design approach is proposed in conjunction with predictive control strategies. The predictive con-trol uses frequency domain models and room air temperature set-point profile as input. The set-pointprofiles and BITES design are improved in a holistic manner according to the thermal dynamic responseof active BITES systems and their thermal zones. The dynamic response is obtained from the transferfunctions of frequency domain models. The methodology is demonstrated on ventilated systems. The

requency domain modelransfer functionntegrated design and operationctive building-integrated thermal energytorage

results show that the methodology can significantly improve the design and operation of active BITESsystems, and hence improve their thermal and energy performance. The dynamic response of differentsizes of systems is presented to provide useful information for design selection. It is shown that concretethickness of 0.2–0.3 m is a good value to initiate design. Other important application considerations arealso discussed.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Building-integrated thermal energy storage (BITES) systems useuilding fabric (e.g. masonry block walls and concrete slabs) ashermal storage mass. They are considered as active BITES if theymbody internal charging system, such as hydronic, air-based orlectric systems. They are sometimes referred to as fabric thermaltorage [1], fabric energy storage [2], or thermo-active (or thermallyctivated) building systems (TABS) [3]. Active charging enhanceshe engagement of core mass for thermal energy storage by utilizingore area for heat transfer.

Primary space conditioning can be supplied through activeITES systems. The active charging systems may heat/cool theuilding fabric (wall, floor or ceiling), which in turn heat/cool theirones through radiation and convection, like large radiant heat-ng/cooling panels [4] integrated with storage mass. Using active

ITES systems with proper control can provide low energy spaceonditioning with relatively flat profile of power demand, whileaintaining or improving thermal comfort. The demand of space

∗ Corresponding author. Dept. BCEE, Rm: EV-6.139, Concordia University, 1455 deaisonneuve Blvd. West, Montréal, Québec, Canada H3G 1M8.

el.: +1 514 848 2424x7244; fax: +1 514 848 7965.E-mail address: yuxia ch@hotmail.com (Y. Chen).

ttp://dx.doi.org/10.1016/j.enbuild.2014.08.013378-7788/© 2014 Elsevier B.V. All rights reserved.

conditioning and the supply of ambient renewable energy and off-peak power can also be well matched [5,6].

Fig. 1 shows two configurations of such active BITES sys-tems. They will be used for demonstration in this paper. In thewithout-airflow-to-zone configuration (Fig. 1a, e.g. hydronic floorheating/cooling systems), the only function of heat transfer fluids(e.g. air or water passing through the BITES) is to heat or cool theBITES mass. In the with-airflow-to-zone configuration (Fig. 1b, onlyfor air-based systems), the airflow enters the room and mixes withroom air after passing through the BITES. It can provide ventilationand space conditioning to the zone, besides exchanging heat withthe BITES mass.

There are two main criteria for the thermal functions of suchactive BITES systems. First, they can provide sufficient space ther-mal conditioning to their thermal zones. Second, they can storeideal (sometimes large) amount of thermal energy for appropriateoperating temperature ranges and time. The acceptable thermalcomfort range offers flexibility but also imposes limits on theiroperations.

Since an active BITES systems has a considerable amount ofthermal storage mass, its thermal response is slow. Its time con-

stant is of the order of hours. To achieve their functionality,predictive control has to be implemented. Different predictivecontrol strategies for BITES systems have been proposed in theliterature [4–12].

576 Y. Chen et al. / Energy and Buildings 84 (2014) 575–585

Nomenclature

Symbols∼ oscillatory value/response∼= approximately equal to� difference or potential← order of layers in an assembly. 1 ← N means the

assembly contains layers from 1 to N, and the exci-tations are on surface l of layer N

Area surface areaArg{} argument (phase angle) of the complex numberB total number of bounding surfaces, or BITESc specific heat capacity (J/kg/K)CR combined convection and radiationd mathematical symbol for differentiale TES capacity per unit area (J/m2) or exponential baseE energy (J)f fluids (air for with-airflow-to-zone configuration,

and air or water for without-airflow-to-zone con-figuration)

h convective heat transfer coefficient (W/m2/K) orharmonic index

M matrix of transfer functionsP heat power (W) or period in secondsQ volumetric flow rate (m2/sec)rm room or room airsc source or source layerslr solart time or duration (second/sec unless specified)ttl totalT temperature (◦C)�T temperature difference or potential (◦C)Th thickness or equivalent thickness (m)u heat transfer coefficient per unit area (W/m2/K)Y self-/transfer-admittance, transfer function in fre-

quency domain

Greek� density (kg/m3)�c volumetric heat capacity (J/m3/K)ω angular frequency (rad/s), ωf = 2�/P and ωh = h·ωf� phase angle of complex number

AcronymsACH air changes per hour (air flow rate in terms of how

many times of room volume in 1 h)AHU air handling unitBITES building-integrated thermal energy storageCHTC convective heat transfer coefficientDFS discrete Fourier seriesTES thermal energy storage

Variables0 ← rma21 the element at the second row and the first column

of the admittance matrix 0←rmadm

M of the assemblybetween node “0” and room air nodes

sch CHTC between the path inner surface and the air-flow

toph combined heat transfer coefficient on BITES top sur-face

0 ← rm p heat exchange between the room air and the otherside of the BITES (node “0” in this case)

Bp total thermal output of BITES to the room (i.e.CR.Bp + A.Bp)

CR.Bp combined convective and radiative thermal outputof BITES

csRatio area ratio of internal heat transfer surface to room-side surface

0 ← rmt12 the element at the first row and the second columnof the transmission matrix 0←rm

trs M of the assemblybetween node “0” and room air nodes

Fig. 1. Conceptual schematics of active BITES systems and their thermal couplingwith the interior space: (a) without-airflow-to-zone configuration, and (b) with-airflow-to-zone configuration (“Indoor mass” includes room air, wallboards and

A.Bp advective thermal output of BITES

furniture; CNV: convection; ADV: advection; LR: long-wave radiation; HRV: heatrecovery ventilator).

Some guidelines for design and operation are provided in liter-ature. Ma and Wang [13,14] and Athienitis and Santamouris [15]investigated the dynamic response of common building fabric com-ponents and provided some sizing guidelines for passive storage.Athienitis and Chen [16] studied the thermal performance and con-trol strategies of an electric radiant heating floor with thermalstorage. Howard and Fraker [17] reviewed the design principalsof ventilated BITES systems that use concrete masonry units. Sim-plified mathematical models and graphical methods are developedin the literature for sizing the active charging system of hydronicBITES [4,18,19]. Similar approach is provided for ventilated BITESsystem by Fort [20].

The design and operation of buildings are interrelated, and the

operation should be taken into account in the design as a primaryconsideration. Frequency domain thermal modeling of active BITESsystems provides a promising approach to integrate design and pre-dictive operation. It offers a convenient means for analyzing the

d Buildings 84 (2014) 575–585 577

dbiin

afitmdqeemnqA

coati

•

•

•

c

2

icctid(ocdlTidoo

2

mpclco

Y. Chen et al. / Energy an

ynamic response (e.g. time lag and magnitude) of multi-layereduilding assemblies (e.g. wall/roof) [21,22]. Through the compar-

son of dynamic response, design of active BITES systems can bemproved on a relative basis. A thermal-admittance-based tech-ique has been widely used in the UK [23,24].

For predictive control, frequency domain transfer functions ofctive BITES systems can be used to analytically calculate the pro-le of the required active charging rates based on a room airemperature set-point profile and the corresponding required ther-

al output [25–28]. Set-point profiles can be enhanced based onynamic response. Z-transfer functions can be derived from fre-uency domain transfer functions. It provides a computationallyfficient way for model-based predictive control [21,29,30]. Chent al. [31] presented a design of model-based predictive controlethodology using frequency domain functions. That study is a

ecessary reference of this one. Some brief information about fre-uency domain modeling and transfer functions can be found inppendix A.

In this paper, optimizing design in conjunction with predictiveontrol strategies is proposed to improve the energy performancef active BITES systems. Frequency domain transfer functions ofctive BITES systems are used for the integrated design and opera-ion. An integrated design and operation methodology is presentedn this paper. It includes the following key steps:

Analysis of dynamic response using frequency domain transferfunctions.Design of predictive control strategies with frequency domainmodels.Design optimization based on operation under bounding designconditions.

The application of the methodology is demonstrated and keyonsiderations for potential applications are discussed.

. Methodology for integrated design and operation

The design procedure that takes operation into account isllustrated in Fig. 2. Bounding conditions (e.g. design weatheronditions, internal heat gain) define the most demanding spaceonditioning load profiles and consequently the required capaci-ies of a building’s space conditioning system (active BITES systemsn this case). Based on operation objectives (e.g. thermal comfort,emand reduction) and bounding conditions, operation strategiese.g. BITES precooling, utilization of ambient renewable energy orff-peak utility energy) should be defined. Preliminary BITES designan be obtained according to bounding loads and guidelines to berawn in this study. Predictive control strategies can then be estab-

ished heuristically based on the dynamic response of active BITES.hey can be further optimized numerically. An iterative fine-tunings needed to further integrate the design and operation. After theesign is finished, the predictive control strategies will be used forngoing operations. In the following subsections, the key elementsf the methodology will be discussed.

.1. Key design parameters of active BITES

The thermal properties of an active BITES system have to accom-odate its functions and operations. There are five critical design

arameters, namely thermal energy storage (TES) capacity, active

harge capacity, thermal output capacity, and thermal output timeags and magnitudes. These five parameters are closely related. Theapacities can be readily calculated based on the construction andperations of an active BITES system. Meanwhile, time lags and

Fig. 2. Active BITES systems design procedure.

magnitudes are dynamic response parameters determined from afrequency domain analysis.

2.1.1. Thermal storage and charge/discharge capacitiesThermal capacities (storage and charge/discharge) are defined

under thermal comfort compliant operating temperatures. The TEScapacity is the maximum possible amount of thermal energy thatcan be stored in a BITES system. It is related to the volume, spe-cific heat and allowable operating temperature range. The activecharge capacity is the maximum heat exchange rate between anactive BITES system and the heat transfer fluid of its active chargingsystem.

The thermal output capacity is the maximum heat exchange ratebetween a BITES system and its thermal zone. In an active BITES sys-tem without airflow to zone (Fig. 1a), its thermal output capacityis the total radiative and convective heat exchange between theexposed BITES surface and the rest of the room. For a with-airflow-to-zone configuration (Fig. 1b), the thermal output capacity equalsto the sum of heat exchange on the exposed surface and the advec-tion component due to the airflow to the thermal zone. The activecharging rate and the advective thermal output rate are commonlyaffected by the airflow rate. The calculation of the required thermalcapacities is demonstrated in the next section.

2.1.2. Dynamic responseA time lag indicates the time delay between an excitation and

its corresponding output response, while a magnitude is the ratioof the response amplitude to the excitation amplitude. Time lagsand magnitudes can be obtained from frequency domain transfer

functions (e.g. thermal admittances or impedances). As an example,Fig. 3 shows the heat flux (response wave) of the interior surfacea BITES system responded to its exterior temperature (excitationwave). Y is a admittance transfer function in the frequency domain.

578 Y. Chen et al. / Energy and Build

�tot

T

woiωi

obdaa

Bfiprsrsu

atsus

2

eionoBpTnatotp

and its zone.

Fig. 3. Conceptual schematic of dynamic response [15].

is the phase angle of Y. As shown, the phase angle and magni-ude of this assembly define the output wave. The time lag can bebtained from the phase angle of Y (Eq. (1)). The magnitude equalshe absolute value of Y (i.e. |Y|).

ime lag = Arg{Y}/ω = �/ω (1)

here function Arg{Y} calculates the argument (i.e. phase angle)f a complex number Y (i.e. transfer function or admittancen frequency domain). ω is the angular frequency of interest.

= 2�/86400 s for a frequency of one cycle per day (the mostmportant frequency for building dynamics).

The dynamic response to excitations of one cycle per day (i.e.ne harmonic) is of major interest for design and operation. This isecause the dominant harmonic of most loads is 1-cycle-per-dayue to the characteristic of weather conditions (e.g. solar radiationnd exterior air temperature) and the high thermal inertia of BITESssemblies [32].

Generally, a larger time lag and smaller magnitude (e.g. thickerITES) is preferable in practice for two main considerations. Therst one is to enable a longer charging period during an off-peakeriod (e.g. nighttime for commercial buildings) and hence toeduce the peak charging demand during an on-peak period. Theecond consideration is to stagger in time the heat inputs to theoom. More thermal energy from individual sources (e.g. directolar gain and heat released from the BITES system) can then betilized.

By comparing the dynamic response of different configurations, suitable configuration with desired thermal energy retainingime (e.g. time lag) and releasing intensity (e.g. magnitude) can beelected. Furthermore, proper active charging rates can be sched-led based on the time lag of a BITES system, as shown later in nextection.

.2. Bounding conditions for design

Bounding conditions (e.g. weather and internal heat gain forxtreme conditions) define the most demanding space condition-ng load profiles and consequently the required thermal capacitiesf an active BITES system. The maximum space conditioning loadeeds a matching thermal output capacity with low temperatureperation. Active charge capacity should allow a full charge of theITES within a limited time with desirable energy sources androvide required space conditioning during the charging period.he TES capacity should be close to the amount of thermal energyeeded for the most demanding discharging period. Furthermore,ctive BITES systems with suitable dynamic response better matchhe energy supply (e.g. renewable and off-peak) and demand (e.g.

n-peak space heating). Therefore, the bounding conditions needo be used for design. An example using this design approach isresented in the next section.

ings 84 (2014) 575–585

For a heating-dominated climate, a sunny cold day followed byan overcast mildly cold day would be an ideal bounding weathercondition. The active BITES system will need to store as muchenergy as possible during the sunny period and then release theheat to its thermal zone to sustain its comfort level until the nextfavorable charging time. A period consisting of several consecutivecold and overcast days may cause a larger space conditioning load.However, this period is not suitable for the TES design if storage ofsolar energy is considered. This is similar in a cooling-dominatedclimate, where a relatively cool night followed by a hot day shouldbe used as an ideal bounding weather condition to utilize cool out-door air for night precooling of BITES.

2.3. Predictive control strategies

To reduce peak demand and energy cost, ideal operations shouldallow active BITES systems to store desirable amount of off-peakand/or renewable energy during favorable periods (e.g. sunnyperiod in a winter day). Yet, thermal comfort compliant room tem-perature profiles bound the operations. However, if the room airtemperature set-point profile is set according to the time lags ofthe active BITES system and its zone, the BITES can be charged atpredicted favorable periods. Take BITES pre-heating for example,the peak room air temperature can be set at a known time lag afterthe peak sol-air temperature [33] is reached. Then the BITES heatingwill take place at peak sol-air temperature.

The concept of the predictive control strategies proposed in thisstudy is as follows. The room temperature set-point profile is corre-lated with the time lags of the active BITES system and its thermalzone and the exterior sol-air temperature based on weather fore-cast. The time lags of active BITES systems can be obtained from thetransfer functions of their frequency domain models. These mod-els will also be used to calculate the active charging rates based onknown room air temperature and the required thermal output ofan active BITES system. The establishment of the set-point profileis based on a heuristic approach. Future research can involve opti-mization techniques to improve the set-point profile. The design ofpredictive control strategies will be demonstrated in the followingsection.

3. Methodology demonstration

In this section, the presented methodology will be demon-strated. The analysis of the dynamic response of an active BITESsystem with airflow to zone (Fig. 1b) will be presented first. Thedynamic responses of other configurations (without-airflow-to-zone configuration or passive BITES) can be obtained in a similarway. Using the dynamic response to establish a suitable roomair temperature set-point profile for predictive control will thenbe presented, followed by design under bounding conditions. Theapplication approach of the methodology will be summarized atthe end.

3.1. Dynamic response analysis

The original cross section of a ventilated BITES system is trans-formed to fit one-dimensional heat transfer modeling (Fig. 4 a)required for frequency domain modeling. Chen et al. [27,28] havepresented relevant modeling techniques that were verified withfull-scale experiment. Their work is adopted in this study. Otherkinds of construction assemblies can be represented in a similarway. Fig. 4b shows the thermal network of the active BITES system

The total thermal output Bp from the BITES to its zone consistsof an advective part A.Bp due to the airflow released directly to theroom and a combined convective and radiative part CR.Bp from the

Y. Chen et al. / Energy and Buildings 84 (2014) 575–585 579

work

ecmflfro

B

wots

aaitd(si

s

wsmw

tebstrpt

1tcihaTb

Fig. 4. (a) Transformed cross section and (b) thermal net

xposed surface of the BITES (Fig. 4b). For without-airflow-to-zoneonfiguration, A.Bp is zero. Bp is normally resulting from three ther-al potentials—temperature difference across the assembly, heat

ux absorbed on the exposed surface, and the charging heat fluxrom the source layer (Fig. 4 and Eq. (2)). The respective dynamicesponse of Bp to each term on the right hand side of Eq. (2) can bebtained by analyzing their corresponding transfer functions.

pi = A.Bpi + CR.Bpi = 0 rmpi + slr rmpi + sc rmpi (2)

here 0 rmp is the heat exchange between the room air and thether side of the BITES (node “0” in this case). slr rmp is the contribu-ion from the solar radiation absorbed by the exposed top surface.c rmp is due to the heat flux from the active charging system.

The dynamic response of room air temperature related to thective charging is most important for design and operation since thective charge is the main heat source for primary space condition-ng in this case. The oscillatory part of sc rmp, which is resulted fromhe dynamic response, can be calculated using Eq. (3). Hence, theynamic response can be found by analyzing the transfer function0←sct12/0←rmt12

)in Eq. (3). It is analogous to the current divi-

ion method [34]. See references Chen et al. [27,28,31] for detailednformation. Some brief information can be found in Appendix A.

c rmpi = scpi ·0←sct120←rmt12

(3)

here scp is active charging heat flux. 0← sct12 is from transmis-ion matrix trsM of assembly 0 ← sc, 0 ← rmt12 is from transmissionatrix trsM of assembly 0 ← rm. They account for the advection forith-airflow-to-zone configuration [31].

The analysis for different equivalent concrete thicknesses (thehickness in the transformed cross section) is plotted in Fig. 5. Theffects of the room thermal capacitance (air, furniture and walloards) and the location of the source layer are included. Fig. 5ahows that the closer the source layer is to the bottom, the longerhe time lag and slightly smaller the magnitude, regardless of theoom thermal capacitance (data for thermally light zones are notlotted). The change of magnitude is more sensitive to changes ofhick concrete and high air flow rate.

Fig. 5b shows the dynamic response for a source layer located/5 of the concrete layer thickness above the bottom surface ofhe concrete layer. Maximum time lags for different zone thermalapacitance levels are shown. The difference in magnitude dimin-shes as the room thermal capacitance level increases. For thermally

eavy zones, maximum time lag and minimum magnitude canlmost be reached with an equivalent concrete thickness of 0.4 m.his means in a periodic charge, 1 W/m2 of charging heat flux wille released after 10.5 h with an intensity of about 0.05 W/m2. For

of a zone with an active BITES slab with airflow to zone.

an extremely light zone, the maximum time lag is slightly morethan 6 h, with a corresponding magnitude of about 0.07 W/m2.

The air flow rate for Fig. 5 is set to two air changes per hour (ACH)of the room air volume (i.e. 2 ACH for every unit floor area). A largerair flow rate will shorten the time lags. With a flow rate of 3 ACH,the largest thermal lag (about 9 h) was found for a concrete thick-ness between 0.3 and 0.4 m (data are not plotted here). The thermalproperties of the material of a BITES assembly also have significantinfluence on its dynamic response. Considering the practical appli-cation of this study mainly concerns concrete building fabric, thechoice of storage mass only considers normal-weight concrete. Thethermo-physical properties of the concrete used in this study are840 J/kg/K for specific heat, 2200 kg/m3 for density, and 1.7 W/m/Kfor thermal conductivity.

The bottom insulation value is 0.5 W/m2/K, and the combinedconvective and radiative heat transfer coefficient (i.e. film coeffi-cient) between the top surface and the rest of the room is assumedto be 9 W/m2/K. More accurate relationships including consid-erations of the direction and flow rate of the outlet air, outletslocations, and heat flow direction on the room-side surface can beused in the future. For a without-airflow-to-zone configuration, thetop surface temperature can be calculated and hence the convectionheat transfer coefficient.

The dynamic response of a concrete slab assembly betweenfloors also has been investigated (data are not plotted here) bychanging the insulation (Fig. 4a) value to 9 W/m2/K—the typicalcombined convective and radiative heat transfer coefficient on asurface. The results show that this change does not significantlyaffect the dynamic response for concrete thicknesses larger than0.2 m. Since a thickness of 0.2 m or larger is generally used for thepurpose of providing enough TES and structural capacities, and thedynamic response will deviate from the theoretical values in realpractice, Fig. 5 alone will be sufficient for design and operationpurposes for similar BITES assemblies.

For dynamic response related to 0 rmp and slr rmp from Eq. (2),the maximum time lags and minimal magnitudes are reached with0.15–0.2 m of equivalent concrete thickness [15]. Therefore overall,an equivalent thickness of 0.2–0.3 m is a good value with which toinitiate design (largest time lags and peak admittances are near thatthickness range).

3.2. Predictive control setup and simulation results

The establishment of the room air temperature set-point profilefor predictive control during a space heating design period (Fig. 6)is used here for demonstration. The resulting thermal behavior ofthe BITES and its thermal zone under this set-point profile is used

580 Y. Chen et al. / Energy and Buildings 84 (2014) 575–585

(a) location of source for thermally heavy zones (b) buildings with different level s of effective

thermal storage mass ; source layer at “1/5” level

Mag

nitu

de

Tim

e la

g (h

r)

Concrete th ickness (m)0 0.1 0.2 0.3 0.4 0.5

12

10

8

6

4

2

0

0

0.1

0.2

0.3

0.4

0.5

0.6Time lag (1/2)Time lag (1/5)Magnitude (1/2)Magnitude (1/5)

Mag

nitu

de

Tim

e la

g (h

r)

Concret e thickness (m)0 0.1 0.2 0.3 0.4 0.5

12

10

8

6

4

2

0

0

0.1

0.2

0.3

0.4

0.5

0.6Time lag (light)Time lag (heavy)Magnitude (l ight)Magnitude (heav y)

Fig. 5. Dynamic response of room air temperature due to 1-cycle-per-day active charge (negative values in the left y-axis means time lag; “concrete thickness” in x-axis isthe thickness of the transformed concrete layer, the equivalent thickness; “light”: thermally light zones, room thermal capacitance is insignificant; “heavy”: thermally heavyzones, effective thermal capacitance (including wall boards) is about 70 times that of enclosed air; “1/2”: source layer (Fig. 4a) is in the middle of the concrete; “1/5”: sourcel

ti

rtptpvdt2fsB

fie

Fd

ayer is 1/5 of the concrete thickness above the bottom surface of the concrete.

o improve the initial design of the active BITES system, as shownn a later subsection.

At first, 22.5 ◦C is used as the average set point with a diurnalange of 5 ◦C. In other words, the room temperature will be allowedo float between 20 and 25 ◦C. One peak and one low temperatureivot points will be identified for each day. The room air tempera-ure set-point profile will be generated by connecting the adjacentivots with straight lines (e.g. ramp segments). The temperaturealues (y-axis) of the peak and low pivots are a function of out-oor peak and low sol-air temperature, respectively. For example,he peak room air temperature for the first day (sunny and cold) is5 ◦C. See Appendix B for information of the function. The time dif-erence (x-axis) between the peak room temperature and the peakol-air temperature equals the time lag of the thermal zone (i.e.

ITES and the room).

There are three reasons for adopting this set-point profile:rst, to reduce space conditioning load by reducing the differ-nce between room and exterior sol-air temperature (for example,

0 4 8 12 16 20 24 28 32 36 40 44 4810−

6−

2−

2

6

10

14

18

22

26

0

100

200

300

400

500

600

700

800

900

Exterior temperatureSet room air temperatureEquator-oriented facade global solar radiation

Time (hr)

Tem

pera

ture

(C

)

Rad

iatio

n (W

/m^2

)

TL TL LTLT

ig. 6. Weather conditions and room air temperature set-point profile during a two-ay space heating period (Round dots: pivots, TL: time lag).

higher set point during hot sunny days and lower set point duringcold and cloudy winter days); secondly, to take more advantage ofthe TES capacity of BITES by allowing the BITES temperature to fluc-tuate within comfort range; thirdly to pre-condition the BITES andits associated thermal zone during favorable periods. For example,allowing the temperatures of the room and its BITES to rise dur-ing a sunny daytime in winter periods will allow more storage ofsolar thermal energy from active and/or passive solar heating, sothat auxiliary heating requirement is reduced for the following coldnight.

In this case, the zone is thermally medium heavy. The advec-tion rate through the BITES is 2 ACH of the room. The sourcelayer is located at the middle of the concrete slab, and the equiv-alent thickness of the slab is 0.3 m. Therefore, based on Fig. 5,the time lag for active charge is about 8 h. Hence, the peak roomtemperature is set to occur 8 h after the time when the peaksol-air temperature occurs (Fig. 6). By setting the wet-point pro-file in this way, the charging of the BITES will start at the mostfavorable time (e.g. sunny time for slab heating). The dynamicresponse of the active BITES system allows the heat to be releasedslowly after the passive solar heating period. Maximum amountof solar thermal energy is stored in the room and its BITES with-out space overheating. The time lag can be fine-tuned duringcommissioning since the actual room capacitance is not preciselyknown.

In general, for a space heating scenario, the global solar radia-tion on equator-facing fac ades or solar thermal collector surfacesis recommended for the calculation of the sol-air temperature. Fora space cooling scenario, the horizontal global solar radiation canbe used, although for a building with large glazed facades (e.g.facing west), the facades may be used for an average sol-air tem-perature. To enable off-peak period charging even when ambientrenewable energy is unavailable, a weighting factor (higher valuefor space heating, and lower value for cooling) can be assigned tothe exterior temperature of off-peak period in the calculation ofthe sol-air temperature. Similarly, other weighting factors (e.g. fornatural ventilation, photovoltaic) can also be applied. In this study,

weighting factors have not been applied. However, since favorableperiods generally coincide with off-peak periods (e.g. sunny day-time during winter, and nighttime during summer), the adoptedapproach also enables utilization of off-peak purchased energy to

Y. Chen et al. / Energy and Buildings 84 (2014) 575–585 581

(a)

(b)

0 4 8 12 16 20 24 28 32 36 40 44 4819

20

21

22

23

24

25

Time (hr)

Tem

pera

ture

(C)

BITES r oom -sid e surface temperature Simulated ro om

air temper ature

BITES source layer temper atu re

0 4 8 12 16 20 24 28 32 36 40 44 4830

20

10

0

10

20

30

40

50

Time (hr)

Hea

t flu

x (W

/m^2

floo

r are

a)

Spa ce condit ioningload

Simu lated thermal output from BI TES to room air

Smoothed heat inj ection at AHU

F ITES tc

atp

imctdtswsisthbg(tc

3

vrsts

ig. 7. Thermal energy injection rates at the AHU and corresponding room air and Bonfiguration).

certain extent. A systematic approach for the optimization ofhe room temperature set-point profile that accounts for differentarameters is worth further study.

The profiles of room air temperature set-point and correspond-ng space conditioning load are then used in the frequency domain

odel of the ventilated BITES slab to calculate the desired activeharging rates. With the charging rate, the thermal behavior ofhe zone in question is simulated with its lumped-parameter finiteifference model (Fig. 7). See Chen et al. [31] for more informa-ion. Even though the temperature set-point profile shown in Fig. 6eems preliminary, the final room air temperature is satisfactory,ith the calculation approach adopted from Chen et al. [31]. As

hown in Fig. 7, the resulting room air temperature rises signif-cantly during the sunny period of the first day thus allowingignificant solar heat gain to be stored in the BITES and the rest ofhe room. The room air temperature profile also reduces the spaceeating load after sunset by reducing the temperature differenceetween exterior and interior (allowing the room temperature dropradually). Peak power demands take place during the sunny periodpossibly also the off-peak period). Regarding thermal comfort, theemperature profile satisfies temperature limits and avoids rapidhanges.

.3. Design under bounding conditions

The previous subsection presents the thermal behavior of theentilated BITES system under a set of bounding conditions. The

esults will be used to improve the initial design of the BITESystem in this subsection. To generalize the approach, calcula-ions are conducted on unit room-side surface area of the BITESystem.

emperatures: (a) temperature profiles; (b) heat flux profiles (with-airflow-to-zone

3.3.1. TES capacityThe required TES capacity, TES.requirese (unit: kWh/m2 or J/m2

of BITES surface) can be approximately calculated by summing theenergy flows (i.e. thermal output and thermal energy charged) dur-ing the period unfavorable for charging (Eq. (4)). For the spaceheating bounding conditions in this case, the energy flows betweenthe peak sol-air temperatures in the two days of the design periodcan be used for Eq. (4). Hence, the unfavorable period equals 24 h.This choice of time period is also applicable to the space coolingscenario.

TES.requirede ∼=∑i=I

1(Bpi + scpi) �t (4)

where I is the total time steps, �t, in the chosen period.For the profiles shown in Fig. 7b, the total thermal output from

the BITES to the room (∑

Bpi · �t in Eq. (4)) between hour 12 and36 is about 0.24 kWh/m2 (0.87 × 106 J/m2), and the thermal energycharged (

∑scpi · �t) is 0.14 kWh/m2 (0.51 × 106 J/m2). Also note

that the slab was not charged to an ideal temperature (i.e. 25 ◦C asthe set upper limit of the room air temperature) during the sunnyperiod. This is because there was excess solar gain, and hence somespace cooling is required from the slab. Ideally in practice, usefulsolar heat gain and free cooling should be controlled at a suitablelevel so that no excess heat or cool is admitted to the thermal zone.Under that condition, the space conditioning load (e.g. thermal out-put from BITES) will be zero (or set to zero). Hence, the BITES willbe fully charged, and less auxiliary thermal energy will be required.

The TES capacity for unit square meter of exposed surface area(unit: J/m2) can be calculated as follow:

TESe = BTh · B�T · B�c (5)

5 d Build

wBo

c(a(aattiitTt

3

dcwoa

B

wbWpaohc

Basrs2f

stp89rbval

mflaiAmatsB

82 Y. Chen et al. / Energy an

here BTh is the equivalent thickness of the thermal storage mass,�c is the volumetric heat capacity of the mass, and B�T is theperation temperature range of the BITES.

Let us take 0.38 kWh/m2 (0.24 + 0.14) as the required TESapacity, and B�c of normal weight concrete of 0.51 kWh/m3/K1.85 × 106 J/m3/K). The temperature drop (i.e. operation temper-ture range) of the BITES after the sunny day is about 1.75 ◦CFig. 7a, source layer is about 0.5 ◦C warmer). For design purposesnd based on previous studies [35], it is practical and sufficient tossume 0.5 ◦C, or no significant temperature difference betweenhe exposed surface and the source layer of the BITES. Let us takehe operation temperature range (B�T) of 2.25 ◦C (extra 0.5 ◦C withdeal operation). Therefore, the required concrete thickness accord-ng to Eq. (5) is about 0.33 m, by setting TESe = TES.requirede. This ishicker than the initial slab thickness used (0.3 m). The requiredESe can be achieved either increase the concrete thickness and/orhe area of the BITES.

.3.2. Thermal output capacityIn satisfying the required thermal output under bounding con-

itions, the design parameters related to thermal output capacityan be determined. By assuming uniform temperature distributionithin the slab and the temperature of the airflow equals to that

f the concrete, the thermal output capacity per unit BITES surfacerea B.maxp (unit: W/m2) can be calculated as follow:

. maxp ≈ rm B. max�T(

toph + f.ttlQC

BArea

)(6)

here rm B.max�T is the maximum temperature differenceetween the room air and the BITES temperature. toph (unit:/m2/K) is the combined convective and radiative thermal out-

ut from the exposed surface. BArea is the top exposed surfacerea. f.ttlQc = f.ttlQ · f�c with f.ttlQ being the total volumetric flow ratef the heat transfer fluid (air in this case), and f�c the volumetriceat capacity of the fluid. f.ttlQ is zero for a without-airflow-to-zoneonfiguration.

The allowable maximum and minimum BITES temperature,.maxT and B.minT, and maximum and minimum room air temper-ture, rm.maxT and rm minT, should comply with thermal comforttandards and are subjective to occupants. To accommodate a largerange of occupant preferences, less extreme values can be cho-en. For a space heating scenario, rm B.max�T = B.maxT − rm.minT ≈8 − 20 = 8 ◦C, and rm B.max�T = (rm.minT − B.maxT) ≈ 25 − 19 = 16 ◦Cor a space cooling scenario.

Once the allowable maximum f.ttlQc is determined, the exposedurface area of the BITES can be obtained. toph is tempera-ure dependent, but approximate values can be used for designurposes. For a without-airflow-to-zone configuration, 6 and

W/m2/K can be used for space cooling and heating, respectively. W/m2/K can be used for with-airflow-to-zone configurationegardless of cooling or heating, since room air is being stirredy BITES outlet airflow. Hence, the combined radiative and con-ective thermal output capacity is about 64–72 W/m2 for heatingnd 36–45 W/m2 for cooling, comparable to reported values in theiterature [36,37].

For the profiles shown in Fig. 7, the maximum required ther-al output (i.e. B.maxp in Eq. (6)) is about 20 W/m2 (630 W for a

oor area of 31.5 m2). It can be satisfied alone by the convectivend radiative thermal output from the exposed surface. If the max-mum flow rate is 2 ACH, f.ttlQc/BArea will be about 2.2 W/m2/K.

relatively high temperature operation is needed for creating theaximum allowable temperature difference between the room air

nd average BITES temperature, (i.e. rm B.max�T). To perform lowemperature space conditioning, suitably large BITES surface areahould be used. Furthermore, a large floor area will be needed if theITES surface is partially covered.

ings 84 (2014) 575–585

3.3.3. Active charge capacityFor unit square meter of room-side surface area, the active

charge capacity scu (unit: W/m2/K) can be calculated with Eq. (7).After knowing the required TES capacity under bounding condi-tions, the needed charge capacity can be calculated with Eq. (8).

scu = sch · coreArea/BArea = sch · csRatio (7)

where sch is the CHTC between the path inner surface and the air-flow (e.g. source layer). csRatio is the area ratio of internal heattransfer surface to room-side surface. It is created in this study tocalculate the required active charge capacity based on unit room-side surface area, since the calculation of other thermal capacitiesare also based on unit room-side surface area.

TES.reqiurede = sc f�T · scu · charget = sc f�T · sch · csRatio · charget (8)

where sc f�T is the average temperature difference between sourcelayer and heat transfer fluid. charget is the charging time (e.g. sunnyhour during a typical winter day). It can be assumed to be half ofthe annual shortest daytime in this space heating case.

Since the outlet air is close to the source layer temperature(around 23 ◦C during charging period (Fig. 7a)), and inlet air is lessthan 35 ◦C for low temperature operation or using solar-heated airdirectly from solar thermal collectors [38], sc f�T can be conser-vatively assumed to be about 6 ◦C. Once flow rate and hence schis known, csRatio and hence the air channel (i.e. core) surface canbe obtained. Note that the charging capacity also has to satisfy thethermal output capacity requirement.

For the profiles shown in Figs. 6 and 7, the charging periodis about 4 h (daytime is about 9 h). The TES.Requirede is about0.38 kWh/m2 as calculated above. The sch is about 5.1 W/m2/K for 2ACH flow rate (0.62 m/sec air velocity in the air channel) for the airchannel cross section used in this case. The resulted csRatio is about3.1. That means the required air channel surface is 3.1 m2 per squaremeter of BITES surface area. Comparison with the values (less than2.5) of hollow core slabs [39] indicates that 3.1 is too high and hencenot practical. Further design fine-tuning is needed. For a given TEScapacity, increasing the BITES surface area can reduce the requiredstorage capacity per unit surface area. Hence, a smaller csRatio willbe required.

3.4. Application approach

The BITES room-side area BArea and the thermal storage massthickness play important roles. The capacities of TES, active chargeand thermal output are commonly affected by the room-side areaof the BITES. The thickness significantly influences the dynamicresponse and the TES capacity. Hence, designers can initiate thedesign with an initial BITES effective thickness (0.2–0.3 m forconcrete as suggested previously) and one interior surface area(normally the floor or the ceiling). One entire interior surface canbe used for practical construction and uniform room air tem-perature considerations. Normally, larger thermal output meansmore thermal energy needs to be stored. The area requirementfor capacities of TES and active charge is usually higher than thatfor thermal output as shown in the previous calculation, exceptin the case where buildings have high space conditioning spikesoccasionally (e.g. conference room). In this case, the exposed sur-face area should be first determined through the required thermaloutput capacity. Then the thickness can be determined by therequired TES capacity. Sometimes, the requirement of the surfacearea has to consider the possibility that the room side surface of

BITES maybe partially covered (e.g. carpet, acoustic panels). Thelargest impact due to covering is on the thermal output capacityand the dynamic response, especially for a without-airflow-to-zoneconfiguration.

d Build

taatraipt

(

(

(

(

(

(

(

(

(

(((

(

sa

attsb

4

tkt

Y. Chen et al. / Energy an

If the initial surface area is not sufficient for the required peakhermal output or charging capacity (assuming maximum allow-ble flow rate have been reached), additional surface area can bedded to increase the total exposed surface area. In the meantime,hickness can be adjusted while maintaining the desired dynamicesponse. This will slightly change the suitable room air temper-ture set-point profile and hence the space conditioning load. Anterative process is necessary. However, since the temperature set-oint profile will not change significantly, one iteration is expectedo be sufficient.

The general design procedure can be summarized as followsFig. 2):

1) Identify bounding conditions (e.g. design weather conditions,internal heat gain).

2) Decide on operation strategies that enhance energy per-formance under the bounding conditions (e.g. allow roomtemperature to swing within comfort zone; pre-conditionactive system with ambient renewable energy or off-peak util-ity energy).

3) Design active BITES (e.g. storage mass thickness) to providesuitable dynamic response (Fig. 5) for the decided operationstrategies.

4) Select control approach. In this study, active BITES is operatedbased on room air temperature set-point profile (with linearsegments connected by pivot points), which is heuristicallyestablished based on the information from steps (2) and (3).

5) Estimate the corresponding space conditioning load profileusing numerical models of the thermal zone, excluding theBITES system.

6) Select one or multiple available interior surfaces as the BITESroom-side surface according to the maximum thermal outputcapacities of the selected active BITES systems and the peakspace conditioning load under bounding conditions. Attentionis needed for potentially covered area.

7) Calculate the required active charging rates for the active BITESsystem, and the thermal output of the system (e.g. Fig. 7).

8) Then:(i) Check the inlet fluid flow rate and temperature, if they are

within design ranges.ii) Check the sufficiency of the TES capacity.

iii) Check the maximum thermal output capacity.iv) Calculate the required minimum area ratio of internal heat

transfer to room-side surface (i.e. air channel surface area forventilated systems and pipe perimetric area for hydronic sys-tems).

9) Adjust the active BITES design if needed and repeat the steps(3) to (8), which is the “Design optimization” in Fig. 2.

Steps (4) and (5) are for the control approach presented in thistudy. This design procedure can be applied with other controlpproaches.

The design of the building should consider the design and oper-tion of BITES. For example, if the space conditioning load exceedshe thermal output capacity of the selected active BITES system,he envelope of the building (e.g. insulation level or shading device)hould be improved or other types of active BITES systems shoulde chosen.

. BITES application considerations

Supplying primary space conditioning through active BITES sys-ems is not suitable for two kinds of load conditions. The firstind is with large space conditioning loads, and the second one ishermally lightweight buildings with rapidly varying thermal load.

ings 84 (2014) 575–585 583

Nevertheless, active BITES systems can still be used to assist spaceconditioning in these two situations. Reported studies showed thathydronic BITES floors have a cooling capacity of 40–60 W/m2 and aheating capacity of 30–40 W/m2, under operations compliant withthermal comfort standards [40,41]. Active BITES with airflow tozone can handle larger ranges of space conditioning load and fluc-tuation due to their advective thermal output (e.g. the supply airfrom the BITES in Fig. 1). The second limiting condition is mainlydue to the large thermal inertia of the BITES. The system’s tem-perature cannot change rapidly enough to accommodate a suddenlarge change of space conditioning loads. To eliminate these twokinds of load conditions, the building’s heat gain and loss need to becontrolled with a properly designed building envelope. For exam-ple, proper shading design is important for passive solar buildings.Internal heat gain needs to be controlled as well (e.g. small lightingpower intensity). The effective thermal storage mass level shouldbe medium to high in order to reduce the temperature fluctuationin case of large load fluctuations.

Thermal output rate per unit degree temperature can be seenas the thermal coupling between BITES system and the rest of theroom. Sufficient thermal coupling between the BITES systems andtheir thermal zones should be provided in order to enable effec-tive heat exchange [42], especially in passive solar design [38]. Onthe room side surface, the heat exchanges may be weakened bycoverings such as furniture, carpet or wooden flooring. However,the advection from with-airflow-to-zone configuration (Fig. 1b)can compensate this weakness. Furthermore, local re-circulationof room air through BITES systems, which acts as a short cut toavoid sending air back to the plant unnecessarily, can be adoptedto enhance the thermal coupling.

Increasing the thermal coupling can also reduce the temper-ature difference between the BITES and its thermal zone for agiven heat exchange rate. This reduction has significant benefits.The operating temperature of the BITES systems can be lower forspace heating and higher for cooling. This will reduce the operationenergy consumption and the initial costs of the mechanical equip-ment and service systems (by downsizing them). It also creates awide set of solutions in the choice of energy sources (e.g. renewableand recovered waste heat).

To accomplish desirable operations, a matching design (e.g. suit-able thermal capacities) is needed. On the other hand, design hasconstrains, and operations have to adapt. For example, the max-imum amount of thermal storage mass may be limited by theallowable structural load.

5. Conclusions

A methodology for the integrated design and operation of activebuilding-integrated thermal energy storage (BITES) systems waspresented in this study. The methodology is demonstrated on ven-tilated BITES systems, and its application approach is provided.Considerations for the application of active BITES are also discussed.

The methodology uses frequency domain models for design andpredictive control. Dynamic response of BITES systems is obtainedfrom the transfer functions of their models. It is used to improvethe BITES design on a relative basis and to establish room air tem-perature set-point profiles. The dynamic response of different sizesof ventilated systems is presented to provide useful design infor-mation for designers. It shows that concrete thickness of 0.2–0.3 mis a good value to initiate design. Using time lags obtained from rel-evant transfer functions to improve the set-point profiles is robust

and it allows BITES to be charged at desirable period.

Bounding conditions (based on extreme load/weather condi-tions) are proposed for design in this methodology. With initialdesign and predictive control, the thermal behaviors of the BITES

5 d Build

sdntr

A

SNwFs

A

A

af

[aTfm0

d10p[wps

1t

ws

84 Y. Chen et al. / Energy an

ystems under bounding conditions can be used to fine tune theesign. The equivalent thickness of the BITES storage mass shouldot be more than 0.3 m for ventilated systems; otherwise, prac-ical active charge capacity will not be sufficient for fully chargeequirements.

cknowledgments

The work is funded by Postgraduate Scholarship from Naturalciences and Engineering Research Council (NSERC) of Canada, theSERC Smart Net-zero Energy Buildings Strategic Research Net-ork (SNEBRN), and Graduate Student Support Program from the

aculty of Engineering and Computer Science of Concordia Univer-ity, Montreal, Canada.

ppendix A. Appendix

.1. Frequency domain modeling

An active BITES system (Fig. 4) can be represented with anssembly consisting of N layers of material (Fig. A1). Using discreterequency response modeling, the oscillatory responses of heat flux10pi,h and temperature 1

0Ti,h at surface 0 of layer 1 due to excitationsN1 pi,h and Ti,h on surface l of layer N can be calculated with Eq. (A1)26,43]. The right-hand side subscript i means the ith time interval,nd h is the harmonic index (h = 1 for 1 cycle per day excitations).he mean responses can be obtained in a similar way with the trans-er functions matrix 1←N

trs Mh replaced with a thermal resistanceatrix. Then the total responses in frequency domain at surface

will be 10Ti = 1

0T +∑H

h=1

(10Ti,h

), and 1

0pi=10p+

∑Hh=1( 1

0pi,h). Time

omain values can then be obtained through 10Ti = Re

{10Ti

}and

pi = Re{

10pi

}, where Re{}takes the real part value from the com-

lex number.

10Ti,h

10pi,h

]= 1←N

trs Mh ·[

Nl

Ti,h

Nl

pi,h

](A1)

here 1←Ntrs Mh is the overall transmission matrix, equal to the

roduct of the individual transmission matrix in the order corre-ponding to their locations in the assembly (Eq. (A-2)).

←Nrs Mh = 1

trsMh · 2trsMh · 3trsMh. . . · NtrsMh (A2)

here left-hand-side superscript “1←N” indicates this transmis-ion matrix is of the assembly containing layers from 1 to N. As

Fig. A1. Schematic of an N-layer assembly, and its excitations.

[

[

[

ings 84 (2014) 575–585

indicated by Eq. (A2), the temperature and heat flux on surface l oflayer N have to be the excitations for this formulation. n

trsMh is thetransmission matrix of layer n (Eq. (A3)). In this study, surfaces 0and 1 are the opposite outer surfaces of any layer, such as the roomair and the wall surface in an air film layer.

ntrsMh =

[cosh(nl · �h) sinh(nl · �h)

nk · �h sinh(nl · �h) cosh(nl · �h)

]=

[t11h t12h

t21h t22h

](A3)

where nl is the thickness of layer n, and nk is the thermal conductiv-ity (W/m/K).

√jωf h/n˛, and n

= nk/n� · nc is the thermal diffusivity(m2/sec) of the material of layer n.

For a layer that can be considered as purely resistive/conductive(e.g. insulation, air film), the transmission matrix becomes trsMh =[

1 r0 1

]. r is the thermal resistance of the corresponding layer. For

an exterior air film, r = 1/cnvh, with cnvh being the exterior CHTC.

A.2. Formula for establishing room air temperature set-pointprofile

The temperature value of the pivots is calculated based on theexterior sol-air temperature, but limited within the throttling range(Eq. (A4)).

pivotTi = spT + sp�T · Limit[−1, saTi − 1c

2c, 1

](A4)

where spT is the room air temperature set point, and sp�T is halfthe throttling range. saT is the maximum or minimum exterior sol-air temperatures of one day. Function Limit[a,x,b] takes a valuefrom x, but limits it between a and b, inclusively. Coefficient 1c is23.3 ◦C, and 2c is 11.1 ◦C in this case. Their values are obtained fromsimultaneously solving

sp�T = (sa. maxT − 1c) /2c and − sp�T = (sa. minT − 1c) /2c

where sa.maxT of 56.7 ◦C and sa.minT of −10 ◦C are the maximum andminimum sol-air temperatures in this study, respectively.

References

[1] G.D. Braham, Mechanical ventilation and fabric thermal storage, Indoor andBuilt Environment 9 (2) (2000) 102–110.

[2] R. Winwood, R. Benstead, R. Edwards, Advanced fabric energy storage I: review,Building Services Engineering Research and Technology 18 (1997) 1–6.

[3] B. Lehmann, V. Dorer, M. Gwerder, F. Renggli, J. Todtli, Thermally activatedbuilding systems (TABS): energy efficiency as a function of control strategy,hydronic circuit topology and (cold) generation system, Applied Energy 88 (1)(2011) 180–191.

[4] ASHRAE, Panel heating and cooling, in: ASHRAE Handbook—HVAC Systems andEquipment, American Society of Heating Refrigerating and Air-ConditioningEngineers (ASHRAE), Atlanta, GA, US, 2008.

[5] J.E. Braun, Load control using building thermal mass, Journal of Solar EnergyEngineering—Transactions of the ASME 125 (3) (2003) 292–301.

[6] G.P. Henze, C. Felsmann, D.E. Kalz, S. Herkel, Primary energy and comfort per-formance of ventilation assisted thermo-active building systems in continentalclimates, Energy and Buildings 40 (2) (2008) 99–111.

[7] P. Simmonds, Control strategies for combined heating and cooling radiant sys-tems, ASHRAE Transactions 100 (1) (1994) 1031–1039.

[8] C.H. Turner, N.K. Tovey, Case study on the energy performance of the Zuc-kerman Institute for Connective Environmental Research (ZICER) building,ASHRAE Transactions 112 (2) (2006) 320–329.

[9] M. Kummert, P. Andre, J. Nicolas, Optimal heating control in a passive solarcommercial building, Solar Energy 69 (1–6) (2000) 103–116.

10] M. LeBreux, M. Lacroix, G. Lachiver, Fuzzy and feedforward control of anhybrid thermal energy storage system, Energy and Buildings 38 (10) (2006)1149–1155.

11] M. LeBreux, M. Lacroix, G. Lachiver, Control of a hybrid solar/electric thermalenergy storage system, International Journal of Thermal Sciences 48 (3) (2009)645–654.

12] J.A. Candanedo, A.K. Athienitis, Investigation of anticipatory control strategiesin a net-zero energy solar house, ASHRAE Transactions 116 (1) (2010) 246–259.

d Build

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

Y. Chen et al. / Energy an

13] P. Ma, L.-S. Wang, Effective heat capacity of exterior Planar Thermal Mass(ePTM) subject to periodic heating and cooling, Energy and Buildings 47 (2012)394–401.

14] P. Ma, L.-S. Wang, Effective heat capacity of interior planar thermal mass(iPTM) subject to periodic heating and cooling, Energy and Buildings 47 (2012)44–52.

15] A.K. Athienitis, M. Santamouris, Thermal Analysis and Design of Passive SolarBuildings, James & James, London, 2002.

16] A.K. Athienitis, T.Y. Chen, Experimental and theoretical investigation of floorheating with thermal storage, ASHRAE Transactions 99 (1) (1993) 1049–1057.

17] B.D. Howard, H. Fraker, Thermal energy storage in building interiors, in: B.Anderson (Ed.), Solar Building Architecture, Vol. 9, MIT press, Cambridge, MA,USA, 1990.

18] M.K. Guyot, S.J. Ormiston, H.M. Soliman, Development of a design procedure forradiant heating in slab-on-grade floors in a cold climate, ASHRAE Transactions117 (2) (2011) 639–651.

19] M. Koschenz, V. Dorer, Interaction of an air system with concrete core condi-tioning, Energy and Buildings 30 (2) (1999) 139–145.

20] K. Fort, Hypocaust and murocaust storage, in: O.C. Mørck, R. Hastings (Eds.),Solar Air Systems: a Design Handbook, James & James, London, 2000.

21] A.K. Athienitis, M. Stylianou, J. Shou, A methodology for building thermaldynamics studies and control applications, ASHRAE Transactions 96 (2) (1990)839–848.

22] A.K. Athienitis, Building Thermal Analysis, MathSoft Inc., Boston, MA, USA,1994.

23] M.G. Davies, Thermal response of an enclosure to periodic excitation: theCIBSE approach, Industrial and Engineering Chemistry Research 33 (5) (1994)217–235.

24] CIBSE, CIBSE Guide, Volume A, Design Data, The Chartered Institution of Build-ing Services Engineers (CIBSE), London, 1988.

25] L.A. Pipes, Matrix analysis of heat transfer problems, Journal of the FranklinInstitute 263 (3) (1957) 195–206.

26] H.S. Carslaw, J.C. Jaeger (Eds.), Conduction of heat in solids, 2nd ed., ClarendonPress, Oxford, 1959.

27] Y. Chen, A.K. Athienitis, K.E. Galal, Frequency domain and finite difference mod-eling of ventilated concrete slabs and comparison with field measurements:Part 1, modeling methodology, International Journal of Heat and Mass Transfer

66 (2013) 948–956.

28] Y. Chen, A.K. Athienitis, K.E. Galal, Frequency domain and finite difference mod-eling of ventilated concrete slabs and comparison with field measurements:Part 2. Application, International Journal of Heat and Mass Transfer 66 (2013)957–966.

[

[

ings 84 (2014) 575–585 585

29] D.G. Stephenson, G.P. Mitalas, Calculation of heat conduction transfer functionsfor multi-layer slabs, ASHRAE Transactions 77 (2) (1971) 117–126.

30] J.A. Candanedo, A. Allard, A.K. Athienitis, Predictive control of radiant floorheating and transmitted irradiance in a room with high solar gains, ASHRAETransactions 117 (2) (2011) 652–665.

31] Y. Chen, A. Athienitis, K. Galal, A charging control strategy for active building-integrated thermal energy storage systems using frequency domain modeling,Energy and Buildings (2014), in press.

32] D.C. Hittle, C.O. Pedersen, Calculating building heating loads using the fre-quency response of multi-layered slabs, ASHRAE Transactions 87 (2) (1981)545–568.

33] ASHRAE, Non-residential cooling and heating load calculations, in: ASHRAEHandbook—Fundamentals, American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE), Atlanta, GA, USA, 2009.

34] J.O. Bird, Electrical circuit theory and technology, 3rd ed., Newnes, Boston, US,2007.

35] Y. Chen, K. Galal, A.K. Athienitis, Modeling, design and thermal performanceof a BIPV/T system thermally coupled with a ventilated concrete slab in a lowenergy solar house: Part 2, ventilated concrete slab, Solar Energy 84 (11) (2010)1908–1919.

36] G. Hauser, C. Kempkes, B. Olesen, Computer simulation of the performanceof a hydronichydronic heating/cooling system with embedded pipes, ASHRAETransactions 106 (1) (2000) 702–712.

37] P. Simmonds, S. Holst, S. Reuss, W. Gaw, Using radiant cooled floors to conditionlarge spaces and maintain comfort conditions, ASHRAE Transactions 106 (1)(2000) 695–701.

38] Y. Chen, A.K. Athienitis, K. Galal, Modeling, design and thermal performance of aBIPV/T system thermally coupled with a ventilated concrete slab in a low energysolar house: Part 1 BIPV/T system and house energy concept, Solar Energy 84(11) (2010) 1892–1907.

39] CPCI, Design Manual for Precast and Prestressed Concrete, Canadian PrecastPrestressed Concrete Institute (CPCI), Ottawa, ON, Canada, 2010.

40] G. Hauser, C. Kempkes, B. Olesen, Computer simulation of the performance of ahydronic heating/cooling system with embedded pipes, ASHRAE Transactions106 (1) (2000) 702–712.

41] B. Lehmann, V. Dorer, M. Koschenz, Application range of thermally activatedbuilding systems tabs, Energy and Buildings 39 (5) (2007) 593–598.

42] M.R. Shaw, K.W. Treadaway, S.T.P. Willis, Effective use of building mass, Renew-able Energy 5 (2) (1994) 1028–1038.

43] K.-I. Kimura, Unsteady state heat conduction through walls and slabs, in: K.-I.Kimura (Ed.), Scientific basis for air-conditioning, Applied Science PublishersLtd., London, UK, 1977.