Академический Документы
Профессиональный Документы
Культура Документы
com/robuart
Building Energy
Management Systems
The Age of Intelligent and Adaptive Buildings
B
uilding automation systems (BAS), or building control systems
(BCS), typically consist of building energy management systems
MILOS MANIC,
(BEMSs), physical security and access control, fire/life safety, and
DUMIDU WIJAYASEKARA,
other systems (elevators, public announcements, and closed-circuit
KASUN AMARASINGHE,
television). BEMSs control heating, ventilation, and air conditioning
and JUAN J. RODRIGUEZ-ANDINA
(HVAC) and lighting systems in buildings; more specifically, they con-
trol HVAC’s primary components such as air handling units (AHUs), chillers, and
heating elements. BEMSs are essential components of modern buildings, tasked
with seemingly contradicting requirements—minimizing energy consumption
while maintaining occupants’ comfort [1]. In the United States, about 40% of total
energy consumption and 70% of electricity consumption are spent on buildings
Buildings
Smartphones
Connected
Vehicles
Computers Wearable
Tablets Devices
every year. These numbers are com- This article provides an overview of are required in BEMSs to address the
parable to global statistics that about issues related to modern BEMSs with a three primary issues highlighted above
30% of total energy consumption and multitude of (often conflicting) require- because of the highly complex and
60% of electricity consumption are ments. Because of massive and often in- changing nature of buildings, such as a
spent on buildings. Buildings are an complete data sets, control, sensing, and large number of heterogeneous sensors
integral part of global cyberphysical the evolving nature of these complex and controls, constant changes inside
systems (smart cities) and evolve and systems, computational intelligence (CI) and outside the building (occupancy
interact with their surroundings (Fig- techniques present a natural solution to patterns, aging of materials and equip-
ure 1) [2]. As buildings undergo years optimal energy efficiency, energy secu- ment, floor plan changes, etc.), and
of exploitation, their thermal char- rity, and occupant comfort in buildings. the need to address occupant comfort
acteristics deteriorate, indoor spaces The article further presents an overall while maximizing energy efficiency.
(especially in commercial buildings) architecture where CI can be used in
get rearranged, and usage patterns BEMSs and concludes with a case study Building Energy Consumption
change. In time, their inner (and outer) of the practical applications of using CI Relative to Other Industry Sectors
microclimates adjust to changes in sur- techniques in the BEMS domain [6]. According to the U.S. Energy Information
rounding buildings, overshadowing The primary areas of interest in Administration (Figure 2), 41% of total
patterns, and city climates, not to men- BEMSs can be categorized into three U.S. energy consumption is consumed
tion building retrofitting [3], [4]. Thus, areas: 1) energy efficiency, 2) integra- in residential and commercial buildings
even in cases of “ideally” designed tion of BEMSs with utilities and smart [7], with commercial and residential
BEMS/HVAC systems, because of ever- grid technologies, and 3) resilience buildings consuming 72% of all electric-
changing and uncertain indoor and and security. These are problematic ity [8]. The International Energy Agency
outdoor environments, their perfor- and difficult to effectively address with (IEA) estimates that, throughout the
mance frequently falls short of expec- typical BEMSs that do not have the world, buildings represent 32% of total
tations. Unfortunately, the complexity necessary data processing, evaluat- final energy consumption (energy that is
of BEMSs, large amounts of constantly ing, and control methodologies. Some supplied to the consumer for all final en-
changing data, and evolving interrela- of the aspects that are lacking in typi- ergy uses, such as heating, cooling, and
tions among sensor feeds make iden- cal BEMSs include (but are not limited lighting), and around 40% of primary
tifying these suboptimal behaviors to) adaptability, predictive modeling, (crude) energy consumption in most IEA
difficult [1], [5]. Therefore, traditional multisensor fusion, dynamic optimiza- countries [9]. The United Nations Envi-
data-mining algorithms and data-anal- tion, state-awareness, providing action- ronment Program estimates that residen-
ysis tools are often inadequate. able information, etc. These aspects tial and commercial buildings consume
Buildings
15%
Industry Buildings
24% (Domestic and
Industrial Transportation
Tertiary)
Process 8%
Buildings 37%
48% 21%
Transportation Industry
28% Transportation 77%
33%
FIGURE 2 – The energy consumption by sector for (a) the United States, (b) the European Union, and (c) China [14]–[16].
approximately 60% of the world’s elec- coordinating integration and optimization consumers. Electric vehicles will charge
tricity, in addition to using 40% of global of homes and commercial buildings with through plug-in connections managed by
energy, 25% of global water, and 40% of the nation’s grid [21]. Pacific Northwest home and BAS. Buildings’ utility-scale re-
global resources. Because of the high National Laboratory, with support from newable generation systems will require
energy consumption, buildings are also the U.S. DOE, developed VOLTTRON—an responsive loads to match the fluctua-
one of the major contributors to green- open-source common platform offering tions caused by varying wind and solar
house gas production [10], [11]–[12], in-depth understanding of complex sys- conditions [20]. And finally, consumers
but also offer the greatest cost potential tems that integrate new challenges, such will access their own energy consump-
for achieving significant greenhouse gas as renewable energy generation, energy tion data to make informed decisions
emission reductions, with numbers pro- storage, and electric vehicles [22]. How- about energy habits.
jected to increase [11], [13]. ever, integration of highly variable fac- Thus, the integration of building sys-
Thus, energy efficiency in buildings tors, such as renewable energy, demands tems with the grid is a critical part of the
is an important issue on a global scale control methodologies that are adapt- stability and success of the smart grid
[1]. However, by implementing very low- able and dynamic [20]. The Engineering [20]. NIST efforts resulted in tools such as
cost building management strategies Laboratory of the National Institute of simulation and testing in the Virtual Cy-
alone, it has been shown that the energy Standards and Technology (NIST) has bernetic Building Testbed and Net Zero
efficiency of modern HVAC systems can been investing in building integration Energy Residential Test Facility [23].
be improved by more than 5% [17], [18]. with the smart grid since 2011 [23]. NIST
Furthermore, it has been shown that recognized the need for new standards New Aspects—Resilient
the energy efficiency of HVAC systems enabling homes and buildings to interact and Secure BEMS
can be improved by up to 40% by close with the grid, with buildings becoming Buildings are inherent components of
monitoring and advanced control [19]. both energy renewable generators and global cyberphysical systems and are
User
Existing Sensors
Lighting
Temperature
Carbon • Threshold-Based Alarms
Dioxide Database • Large Amount of Data
Outside Air • Data Rich−Information Poor
User Input
HVAC • Low Understandability
• No Actionable Information
AHU Sensors • No Predictions
Real-Time Data
Legacy Control
• Static/Model Based
• Nonpredictive
Control Signals • No Uncertainty Modeling
In contrast, various CI-based algo- to be capable of outperforming tradi- converging on near- optimal results
rithms such as artificial neural networks tional control methods in a variety of when the search space is too large to
(ANNs), fuzzy logic (FL) modeling, and industrial applications, including, but be searched exhaustively.
evolutionary algorithms (EAs) [42], [43] not limited to, building control [6], As depicted in Figure 5, CI-based algo-
enable the implementation of advanced [42], [44]–[48]. ANNs enable dynamic, rithms can be used not only for control,
control architectures, data-mining tech- predictive, and holistic modeling of but also for providing the user with under-
niques, and optimization capabilities the system by learning the underlying standable and actionable information [1].
that can lead to better situational aware- interdependencies of the system and ANN-based system modeling, advanced
ness and more efficient, dynamic, and generalizing overall system behavior. clustering-based system modeling, or
adaptive control, as well as grid resil- These inherent generalization capabil- FL-based expert rules can be used for
ience and building data security. ities enable ANNs to accurately handle anomaly detection of the entire system
Figure 5 shows a possible frame- previously unseen and unexpected as well as the subsystems of the overall
work of CI algorithms used in a BEMS. behavior. Furthermore, ANNs have the BEMS [1], [5], [50]. These anomaly detec-
CI-based techniques have the unique capability to adaptively change to new tion techniques identify and make use of
capability of handling large quantities requirements via online learning tech- the underlying interdependencies of the
of heterogeneous data from multiple niques [44], [45], [48]. While ANNs are system for determining potential subop-
sources and extracting the generalized capable of extracting and modeling the timal anomalous behavior and, therefore,
behavior of the system. Furthermore, general behavior of the system, FL mod- are more expressive and useful than tra-
CI-based techniques are inherently eling can be used to model and quantify ditional threshold-based alarms [1], [5].
adaptive and optimizable while being uncertainties that inherently appear in Furthermore, linguistic summarization
able to model uncertainties inherent in data to ensure adaptable control even can be used to provide clear, concise,
real-world measurements. Similarly, this in the presence of noisy, unreliable data and understandable information about
data-driven approach can be comple- [1], [49]. FL also assists in easily in- suboptimal building behavior and its po-
mented by expert-driven CI-based meth- corporating expert domain knowledge tential root-cause to the user [1], [50].
odologies to enable the system to be into the control system by means of ANN, FL systems, and other CI-
more dynamic and accurate. human interpretable linguistic rules. based predictive algorithms have
CI-based control (depicted in To achieve near optimal control, EAs been successfully applied in various
Figure 5) can benefit from various, may be used for dynamically optimiz- control systems, including BEMSs, for
proven CI-based techniques such as ing both ANN and FL techniques as sensor data prediction [3], [4], [51],
ANN, FL, evolutionary optimization, well as classical control methods [6], [52]. Sensor value prediction enables
etc. Such methods have been shown [47]. EAs provide the capability of the system to be proactive rather than
Multitude of
Real-Time Data Output Devices
AHU Sensors
CI
User Input
CI-Based Control
Wind Solar PEVs Storage
• Predictive, Holistic Modeling
Control Signals • Dynamic
• Data Driven and Rule Driven
• Real-Time Optimization
• Uncertainty Modeling
PEVs - Plug-in-Electric Vehicles
Fuzzifier
X
Maximum Operators
Minimum Operators
Defuzzifier
Out
Fuzzifier
Y
Fuzzy
+1 +1 +1 Rules
“selection of the fittest” in terms of bio- use of ANNs to control BEMSs [66], [67], Dealing with uncertain information
logical reproduction. showing advantages of self-learning with has been evidenced in fuzzy and fuzzy-
Other techniques that are related to fast convergence time and fast learning agent-based control. Yordanova et al.
and/or derived from these methods, such in the presence of time delays and model [71] used a simple two-variable fuzzy
as various clustering and classification uncertainties and predictive control for data acquisition control [71]. Hurtado
techniques, are also significant areas of thermal comfort and energy savings in et al. [72] present fuzzy multiagent con-
CI [66]. These methods rely on the under- public buildings. Ferreira et al. [66] dem- trol of a BEMS in a smart grid framework
lying dependencies of multidimensional onstrated ANN deployment for balancing that shows reduced energy demand and
data to generate generalized clusters and desired thermal comfort level and energy points to even bigger savings in larger
classification boundaries. A multitude of savings at the University of Algarve with buildings due to thermal inertia.
CI techniques that are capable of han- energy savings of more than 50%. Martirano et al. [73] demonstrated a
dling large sets of highly multidimension- Energy consumption has been the fuzzy building automation control system
al data is available for these tasks. focus of research interests as well. For with several fuzzy rules for energy and
example, Li et al. [51] have used classi- comfort balancing on three case studies
Concrete Applications fication techniques for daily electricity of a smart office equipped with an auto-
CI techniques are predominantly da- consumption in buildings in Birming- mated/dynamic shading, lighting, and
ta-driven techniques. While theory ham, England, demonstrating 99% cor- HVAC control. Keshtkar et al. [74] present
teaches us to combine physics-based rect prediction. Yuce and Rezgui [68] an FL rule-based algorithm using outdoor
and data-driven techniques, in prac- have used an ANN-GA approach for se- temperature, load demand, electricity
tice, data-driven techniques prove to mantic rule generation for better energy prices, and occupant presence as inputs
be easier, more accurate, and more ca- performance prediction, demonstrating for residential building load management.
pable of keeping up with the evolution a 25% energy reduction while satisfying Genetic and evolutionary approach-
of buildings as floor plan and usage occupants’ comfort. es have been evidenced in energy effi-
pattern change, thermal features dete- Zhang and Chen [46] demonstrated ciency predictions. Wang and Wang [75]
riorate, and HVAC systems age and get a particle swarm optimization (PSO)-ra- have demonstrated 15% electricity sav-
replaced. However, in cases where suf- dial basis function (RBF) ANN solution ings using intelligent control (fuzzy-PSO
ficient data are not available, the ther- for building energy consumption, while approach) of ventilation for maintaining
mal energy storage (TES) model can be Quintero and Mares [69] presented indoor carbon dioxide in comfort zones
created as a physics-based model. an ANN-FL approach demonstrating with reduced energy consumption. Such
Smart buildings offer the possibility 25% energy savings while maintaining techniques have been used to develop
to buffer excess energy from fluctuating customers’ comfort. RBF ANNs have thermal models [76], energy assessment
renewable sources in thermal and elec- been used for electric load forecasting [77], and planning of daily consump-
trical storage units increasing autonomy for large office building load forecast- tions and occupant satisfaction [78].
from utilities and resilience to brown- ing [70], capturing 97% of variability Combined CI techniques attracted
outs. From shifting energy-intensive pro- in hourly electric load of test buildings special attention of researchers in build-
cesses, saving energy in another form, (based on weather and electric power ing energy systems. Techniques such
or producing regenerative energy them- consumption alone). Tran and Tan [48] as fuzzy c-means clustering, support
selves, buildings have become an inte- used feedforward ANN for improving vector machines, and GAs have been
gral part of an intelligent cyberphysical building illumination energy efficiency, used for energy consumption behav-
systems to produce, store, and consume demonstrating 95% accuracy with more iors profiling [79], [80]. Adaptive HVAC
energy. Researchers have examined the than 28% energy savings. control was discussed by Bruckner et al.
TES BEMS
Return Water from HVAC
Chiller
Utility
BEMS/TES
ANN Controller ULP (ANN) Framework
BPR
Predictor (ANN)
Control Valves that Control
the Flow Rate of Water
TES Model (Physics)
Data Stream
Experimental Results Current outside air The current outside air temperature is taken as an indication of the prevailing
temperature weather conditions.
The data set for the case study was gener-
ated using the U.S. DOE’s free, open-source Current averaged room This is the averaged room temperature of the whole building across all floors.
temperature
EnergyPlus simulation software [86]. The
data set was a time period of five months,
where one month was used for training
and the remaining four months were used to evaluate the effects of capacity and Therefore, with the experimental
for testing the ANN architecture. other aspects of storage devices. The setup for TES and chillers, six test
Figures 9 and 10 show the mean ab- parameters of the test tanks were driven cases can be specified:
solute percent errors (MAPE) obtained by best industry practices (Table 2). ■■ test case one: small tank and low-
in the BPR and ULP, respectively. It can Each tank had a different size and maxi- powered chiller
be observed that as the prediction mum flow rate. Table 2 provides details ■■ test case two: small tank and high-
time increases, the errors and stan- of the different tanks used for testing. powered chiller
dard deviations increase. However, the Furthermore, two sizes of chillers were ■■ test case three: medium tank and
prediction errors are below 1% for the considered for each tank, a high-power low-powered chiller
BPR and 5% for the ULP. chiller (with the capacity to recool the ■■ test case four: medium tank and
A simulated building was used to tank to a desired temperature at night) high-powered chiller
carry out the experimentation due to and a low-power chiller (which does ■■ test case five: large tank and low-
privacy issues, but the lessons learned not have the capacity to cool down the powered chiller
from this study would remain the same TES). The last two rows in Table 2 list the ■■ test case six: large tank and high-
regardless whether real building data specifications of the considered chillers. powered chiller.
were used. The simulated build- For each test case, three dif-
ing consisted of ten floors, with ferent TES usage controllers
five occupant zones on each were tested. These controllers
floor. For the sake of simplic- TABLE 2 – Values of the tested tanks. will determine the amount of
ity, the assumption of identical Tank Variable Unit Small Medium Large power that will be sent from
Tank Tank Tank
floors with identical cooling de- the TES to the building in the
mand was made. TES height m 5 10 15 next time step. The three con-
To calculate the costs for the TES radius m 1 7 15 trollers tested were a classical
peak and off-peak hours, real- Tank wall thickness m 0.3 0.3 0.3 nonpredictive PD controller, an
world pricing information was Thermal conductance W/ 0.1 0.1 0.1 ANN 1 Hour (1H) control look-
used. The prices were obtained mK ing 1 h (time step) ahead, and
from freely available data from Gallons in tank gal 4,000 400,000 2,800,000 an ANN 6 Hours (6H) control
a power company in Richmond, TES temprature. upper o
C 17 30 30 looking 6 h ahead.
Virginia [87]. The pricing sched- limit The PD controller (Fig-
ule used for calculations was TES temprature. lower o
C 2 2 2 ure 11) represents a classical
US$0.09001 per kWh during peak limit controller with no predictive
hours (from 7:00 a.m. to 8:00 Maximum flow rate kg/s 0.5 5 capability. The PD control-
p.m.) and US$0.02405 per kWh Efficiency ratio of TES None 0.85 0.85 0.85 ler determines the amount of
during off-peak hours (from 8:00 Efficiency ratio of chiller None 0.9 0.9 0.9
power to be used by the TES
p.m. to 7:00 a.m.) [87]. for the next time step, based on
Chiller watts low W 4,000 50,000 75,000
Three TES tanks of different current and previous power re-
Chiller watts high W 40,00 100,000 150,000
sizes were used in this case study quirements of the building.
Conclusion
FIGURE 9 – The MAPE for the predictions of BPR. This article analyzes the main issues
associated with the development of
efficient BEMSs. It was shown that CI
Absoulte Percentage Errors in Prediction (ULP) techniques are particularly well suited
Percentage Error in Prediction (%)
TABLE 3 – Experimental results for the small tank (test cases one and two).
Tank Chiller Wattage Control Cost without Cost with TES Savings (US$) Savings (%) Wasted Peak Usage (%)
Size TES (US$) (US$) Cost (US$)
Low (test case one) PD 8,370.508 8,290.878 79.630 0.951 134.737 78.144
ANN 1H 8,370.508 8,188.689 181.818 2.172 0 78.301
ANN 6H 8,370.508 8,187.480 183.027 2.187 0 78.079
Small
High (test case two) PD 8,370.508 7,455.746 914.761 10.928 3.703 68.415
ANN 1H 8,370.508 7,432.460 938.048 11.207 3.100 67.687
ANN 6H 8,370.508 7,428.485 942.023 11.254 3.326 68.219
TABLE 4. Experimental results for the medium tank (test cases THREE AND FOUR).
Tank Size Chiller Wattage Control Cost without Cost with TES Savings (US$) Savings (%) Wasted Peak Usage
TES (US$) (US$) Cost (US$) (%)
Low (test case three) PD 8,370.508 6,532.240 1,838.267 21.961 0 39.316
ANN 1H 8,370.508 6,422.214 1,948.293 23.276 0 38.180
ANN 6H 8,370.508 6,431.000 1,939.508 23.171 8.627 35.980
Medium
High (test case four) PD 8,370.508 4,761.879 3,608.628 43.111 546.578 14.304
ANN 1H 8,370.508 4,421.861 3,948.647 47.173 144.123 13.897
ANN 6H 8,370.508 4,217.703 4,152.804 49.612 57.504 12.282
TABLE 5 – Experimental results for the large tank (test cases five and six).
Tank Size Chiller Control Cost without Cost with Savings (US$) Savings (%) Wasted Cost (US$) Peak Usage (%)
Wattage TES (US$) TES (US$)
Low (test case five) PD 8,370.508 7,591.848 778.660 9.302 0 33.547
ANN 1H 8,370.508 7,682.410 688.098 8.221 28.370 31.323
ANN 6H 8,370.508 7,570.861 799.647 9.553 0 30.910
Large
High (test case six) PD 8,370.508 5,459.915 2,910.592 34.772 494.534 9.795
ANN 1H 8,370.508 5,459.547 2,910.960 34.776 253.989 10.733
ANN 6H 8,370.508 5,257.788 3,112.719 37.187 136.023 9.635