You are on page 1of 8

Proceedings of the 2007 American Control Conference FrA05.

2
Marriott Marquis Hotel at Times Square
New York City, USA, July 11-13, 2007

Intelligent Building Systems – Past, Present, and Future


James E. Braun

Abstract—This paper provides a vision for intelligent facilities. Only about 10% of the 4,650,000 commercial
HVAC&R equipment that after many years of unfulfilled buildings in the U.S. have an EMCS (EIA, 1999). Figure 2
promise seems closer to reality. The paper describes some shows that the probability of having an EMCS increases
enabling technologies and incentives that are driving the dramatically with building size. Overall, EMCS serve about
development of embedded intelligence for HVAC&R
33% of the approximately 67 billion ft2 of commercial floor
equipment and then provides examples of some specific
developments of intelligent servicing and control algorithms. space. However, this statistic exaggerates the impacts that
EMCS have had on buildings. Surveys of building
I. INTRODUCTION operators suggest that only a fraction of possible EMCS
functionality is utilized leading to higher operating costs
I magine appliances and HVAC&R equipment that self-
diagnose problems and automatically provide information
to service providers so that parts can be ordered and service
than would have been expected for use of these systems
(Gordon and Haasl, 1996; Lowry, 2002).
scheduled. Imagine various building systems that can
automatically coordinate themselves to optimize energy
efficiency and limit energy demand in response to utility
rates and incentives, weather forecasts, and warnings/alerts.
Imagine devices that provide adaptable interfaces for elderly
with limited hearing and also make information available
throughout the building and remotely accessible from a
variety of interfaces, such as a personal computer, PDA
(personal data assistant), television, or display on any other
appliance (Figure 1).

Fig. 2. EMCS Installations According to Building Size


(from EIA 1999)
The good news is that the number of EMCS installations
increased by 80% between 1995 and 1999 (EIA, 1995 and
1999), with much of the gains occurring in small to medium
sized commercial buildings (10,000 to 200,000 square feet).
The controls companies appear to be providing more and
better products geared towards smaller commercial
buildings at lower costs (BCS, 2002). Costs are coming
down for a variety of reasons, including greater competition
resulting from open protocols and the evolution of
Fig. 1. Selected Components of Intelligent Buildings distributed systems, lower cost computing, and easier-to-use
The notion of intelligent buildings received a lot of press software configuration tools. In addition, facility
in the 1980’s after the advent of computerized control for management and energy service companies are expanding
buildings and yet reality has fallen well short of this vision. and utilizing EMCS to manage and quantify building
In fact, energy management and control systems (EMCS) for performance.
buildings have been underutilized except for very large As a result of these changes, the cost-to-benefit ratio has
improved significantly for application of computerized
Manuscript received September 22, 2006. control systems for buildings. Furthermore, additional
James E. Braun is a professor at the School of Mechanical Engineering, developments are underway and anticipated that will
Purdue University, West Lafayette, IN 47909 USA (phone: 765-494-9157; ultimately lead to even more dramatic reductions in the cost-
fax: 765-494-0787; e-mail: jbraun@ purdue.edu)
. to-benefit ratio and widespread application of computerized

1-4244-0989-6/07/$25.00 ©2007 IEEE. 4374


FrA05.2

system control, including the residential market.


The key to realizing the potential for computerized
control in buildings has been to reduce installed costs. Early
EMCS were very centralized and had exceedingly high
wiring and software programming costs. In order to reduce
these costs, there has been a transition from a “centralized”
to “distributed” control architecture over the past twenty
years. The next great opportunity to reduce installation
Fig. 3. Integration of Sensing, Communication, and Computation
costs and realize the vision of intelligent buildings is to
embed intelligence directly in manufactured building For instance, a compressor might have an integrated chip
equipment and components with the ability to communicate that contains a performance map and has sensor inputs for
information with standard protocols that can be used for suction and discharge conditions that allow use of the map
intelligent system operations (control, maintenance, and to determine refrigerant mass flow rate and compressor
service). It does not seem far fetched to envision smart power consumption. These “virtual sensors” would provide
equipment and appliances having integrated chips that are the same information provided by expensive mass flow and
embedded with documentation, models, and algorithms that power meters at a fraction of the cost and could be used for
could enable coordination with other devices, condition a variety of purposes, including continuous performance
monitoring, and automated diagnostics. Embedded monitoring of equipment EER, inputs to a diagnostic
intelligence could ultimately be integrated within most system, inputs to a building-wide demand-limiting
components found in buildings, including individual lights, algorithm, etc. Similarly, a condenser could have a chip
sensors, compressors, valves, heat exchangers, pumps, with a model used along with measurements of state
refrigerators, and dishwashers. Service providers could variables and a virtual refrigerant mass flow measurement to
provide free diagnoses and cost estimates for appliance provide a virtual sensor for condenser air flow. Condenser
repairs remotely. Appliances could be designed to be grid air flow could then be used as a diagnostic indicator for
friendly so that they respond to utility control or price fouling or fan problems. Embedded intelligence will allow a
signals and coordinate themselves to limit electrical demand proliferation of low-cost information for monitoring,
during periods of critical demand. Ultimately, the use of diagnostics, and improved control.
embedded intelligence and high-level protocols could enable The application of MEMS technology to HVAC&R
“plug-and-play” system controls and monitoring that would sensors could also have a dramatic effect on the costs of
dramatically reduce the costs associated with configuring the embedding sensors and intelligence in HVAC&R equipment
software necessary to reap the great cost savings potential (Yashar and Domanski, 2004). MEMS technologies involve
for computerized operations of buildings. manufacture of many solid-state sensors on a single
This paper describes some of the enabling technologies semiconductor wafer, which reduces cost. MEMS sensors
and incentives (see Roth et al., 2005) that are driving the are low cost when manufactured in large volumes, small and
development of embedded intelligence for building systems lightweight, and can have good accuracy and resistance to
and then provides examples of some specific developments long term drift (Figure 4).
of intelligent servicing and control algorithms.

II. ENABLING TECHNOLOGIES AND INCENTIVES


The cost-to-benefit ratio for implementation of embedded
intelligence within building equipment has been decreasing
in the past decade for a variety of reasons. Costs are
decreasing because of reductions in hardware, installation,
and infrastructure costs. The benefits are increasing as
utilities offer new incentives to reduce peak demand and
energy usage and as the overall cost of energy increases.
A. Lower Cost Sensing and Processing Fig. 4. Spider Mite on a MEMS Device (Sandia National Labs)
Costs, size, and power requirements for processors, They are widely used in the automotive industry due to
sensors, and networking have dropped dramatically over the the large volumes employed. However, for adaptation to
last decade. Small-scale, low-cost devices are now available HVAC&R applications, MEMS sensors need to provide data
that integrate a CPU with a sensor and network output in formats commonly used by HVAC&R controllers
communication (Figure 3). This is pushing processing (e.g., 4-20ma). In order to realize price benefits from
closer to the sources of data and opens up many possibilities economies of scale, the HVAC&R sensors market would
for embedded intelligence. need to have greater standardization in sensor components

4375
FrA05.2

and sufficient buyer demand to drive up production ultimately be widely applied for monitoring and control in
volumes. This increased demand would be a logical buildings because of reduced installation costs. In
outgrowth of a general trend towards HVAC&R equipment particular, wireless solutions make the most sense for retrofit
manufacturers’ embedding intelligence in their products. applications where adding wiring is much more labor
intensive and expensive than for new structures. Recently,
B. Lower Installation Costs
the building controls companies have been offering wireless
The costs of field installation for monitoring and control temperature sensors products.
systems are a major part of total system costs and are due to One of most compelling applications for wireless
wiring of individual components, sensor installation, and temperature sensors is for zone temperature thermostats.
software implementation. Although control and monitoring Wireless technology should allow larger flexibility in the
companies have expended considerable resources to reduce placement of thermostats leading to improved comfort
these costs, several developments are taking place that will conditions. Ultimately, individual occupants could have
provide significant additional cost savings. their own zone temperature sensors that communicate with a
In order to reduce wiring costs, control and monitoring control system that attempts to optimize thermal conditions
systems have migrated towards highly distributed for occupants. Furthermore, battery powered sensors can
architectures that rely on low-level controllers that are adjust transmission rates according to the rate at which the
networked together. This architecture tends to minimize measurements are changing to conserve battery life.
lengths of individual wires between processors and For small-scale retrofits, wireless can be implemented
individual sensors and actuators (Figure 5). This trend will with simple point-to-point wireless schemes (Kele, 2004) for
continue as processors get integrated at even lower levels wire replacement. These technologies have existed for
within sensors and actuators and directly connected to a commercial buildings for a few two years (Turpin, 2004).
local network. Fully distributed networked solutions Wireless systems compatible with BACNet are also
significantly reduce lengths of wiring compared to more available (Wang and Nova, 2004, Kiyon, 2004). For larger
centralized processing solutions. This will most likely scale applications, wireless mesh networks that are
involve a hierarchical structure for networking with lower interconnected, self-enabling and self-healing (Kintner-
cost networking employed at lower levels (e.g., within Meyer and Conant, 2004) could be employed. In mesh
individual air conditioner units) where information networks, individual nodes (Figure 6) can serve as
transmission requirements are less than at higher levels (e.g., transmitters, receivers, and routers that relay messages
between all building controllers and monitoring devices). (CABA, 2004; Turpin, 2004). Mesh networks are beginning
to be deployed for a variety of applications, including for
agricultural monitoring using moisture and temperature
sensors (Delin, 2004; Electronic News, 2004) and for
industrial settings (Marshall, 2003). Controls companies
and wireless mesh network providers are entering into
business relationships for commercial buildings (Turpin,
2004) and companies have begun to offer building-wide
wireless solutions.

Fig. 5. Transition from Centralized to Distributed Network


Architectures

Two other approaches for reducing the costs associated


with wiring have received considerable attention in recent
years: power-line carrier (PLC) and wireless. PLC uses
existing power lines within a building to transmit digital
information. Digital signals are encoded and transmitted at
a much higher frequency than the electric power signal.
Fig. 6. Wireless Sensor Node (www.sensorsmag.com, Riedel
Although this would appear to be a very attractive approach (March 2004))
from a cost perspective, electrical noise from other devices
connected to the electrical distribution network can cause The costs of many wireless solutions are coming down
significant problems (Knisley, 2003) and to date, PLC has and recent studies have demonstrated that they are currently
primarily been used for lighting control in residential cost effective in many situations (Kintner-Meyer and
markets. It is not expected that PLC will play a major role Brambley, 2002; Katipamula and Brambley, 2004)
in building control and monitoring applications (BCS, depending on the number of sensors and wireless
2002). components employed. The economics of wireless systems
There are a variety of wireless approaches that could improve as the number of sensors increases due to a tradeoff

4376
FrA05.2

between higher fixed costs for wireless systems and higher years.
installation costs for wired systems. Wireless system costs A communication protocol is a common language that
are expected to decrease at a significant rate over the next allows devices to share essential information for effective
few years (Kintner-Meyer and Brambley, 2002; Kintner- control and performance monitoring. In the 1990’s, several
Meyer and Conant, 2004) meaning that wireless solutions open communications protocols came to market for building
will undoubtedly have a major impact on building controls, with BACNet and LonTalk receiving the most
applications in the near future. attention in the U.S. LonWorks products appear to have
In addition to wiring, installation of sensors in the field is gained greater market share for equipment control and
time consuming and prone to errors. These difficulties and stand-alone controllers than BACNet (DeNamur, 2002).
costs would be reduced if more sensors were embedded in However, BACNet may represent the best option for
equipment at the factory and could communicate through system-wide monitoring and control (Kranz and Gisler,
standard protocols with other devices. The incentives for 2002) because it offers greater top-level functionality and
factory-installed sensors will improve as low-cost MEMS interoperability with enterprise networks (i.e., with Ethernet
sensors are developed for building applications and the cost and IP).
effectiveness of embedding sensors and intelligence in The use of existing enterprise networks could
equipment is improved. significantly reduce the installed cost of system-wide
Systems that provide facility wide monitoring and control monitoring and controls and allow remote access and control
require a significant amount of software implementation and from any device with Internet access (PC, hand-held device,
configuration. Control companies have invested heavily in cell phone) and appropriate access permission. Ultimately,
developing software configuration tools that reduce these web-based browsers could allow occupants to take control
programming costs. The tools have graphical interfaces of local building environments. To a limited extent, this has
with standardized control and monitoring solutions for already started to occur (Tom, 2005). The seamless sharing
different types of equipment. However, this process is still and transfer of information via IP (or other protocols) can
time consuming, costly, and prone to errors. Also, re- also substantially increase the efficiency of building
programming of the system is required whenever changes to management. For example, a study of U.S. capital facilities
the HVAC&R or other building systems occur. As a result, estimates that seamless interoperability reduces operations
computerized control and monitoring systems are typically and maintenance costs by $0.23/ft2.
underutilized except for larger commercial applications Communications protocols have been modified to
where a support staff takes advantage of continuous system incorporate specifications to communicate over IP networks.
monitoring and can manage software reprogramming and For example, ASHRAE has published an addition to the
upgrades that are necessary. Additional cost reductions and BACNet standard known as “BACNet/IP” that specifies
flexibility should be possible by embedding intelligence how BACNet devices can directly communicate over IP-
within equipment leading to wider application of system- based networks. Major controls manufacturers now offer
wide control and monitoring. Ultimately, the availability of “BACNetTM/IP” workstations and controllers. Ultimately,
embedded intelligence will facilitate the development of building controls may communicate entirely in IP-based
“plug-and-play” controllers that automatically configure protocols, bypassing the need for other protocols. Web-
themselves to provide coordination between different based building control systems that use eXtensible Markup
devices for operation cost minimization and provide Language (XML), a data format for sharing device-specific
performance and diagnostic information for system-wide information over enterprise networks, are under
monitoring and diagnostics. development and at least two manufacturers have launched
native XML building control products.
C. Communications Infrastructure
Another development that could reduce software
The number of providers for control systems used in configuration costs and facilitate commissioning would be
building applications has increased dramatically from the the extension of Universal Plug ‘n Play (UPnP) to
1980’s when three companies dominated the marketplace for commercial building control applications and devices. UPnP
buildings. This trend has been facilitated by the move is an open technology that allows IP-based communication
towards distributed control systems that utilize an array of between UPnP-enabled devices on a network that supports
small-scale, application-specific controllers and the IP. It incorporates a discovery function that largely
development of standard communication protocols, both of automates the incorporation of new UPnP devices into the
which lowered the barriers to entry in the marketplace. The network. When a new device is added to the network, it
increased competition for hardware and services has “advertises” its presence to the other UPnP-enabled devices.
undoubtedly led to lower costs and more installations. The devices exchange information about their functionality
There are many exciting developments in communication (in XML format), which then enables devices to share
protocols for building applications that could have even information or allow one device to control another device.
greater effects on costs and product offerings in the next few Power over Ethernet (PoE) is another promising

4377
FrA05.2

application of enterprise network technology to buildings and improved occupant comfort and satisfaction. Utility
that could significantly impact costs. PoE can provide up to costs are a relatively small percentage of facility operating
13W of power over Ethernet cabling to devices up to 100m costs and have been relatively stable over the past 20 years.
from an Ethernet hub or port (Thomas, 2005). PoE However, energy costs have begun to rise recently in
consolidates separate communications and power wiring response to dramatically increased world-wide demand.
into one connection, reducing the cost of installing a wired Higher energy costs provide a greater incentive for
device. In buildings, for example, it could potentially power intelligent features that lead to more efficient operations.
and provide information to and from sensors and some In many locations, increased electrical demand has put a
controllers. strain on utilities to meet peak requirements and in response,
Widespread use of the internet has created a greater they are experimenting with new utility rate structures that
expectation among customers for accessibility to rely on electronic notification in real time to influence
information for use in decision making. It is likely that the customers to level their electrical demand during times of
marketplace will reward hardware and software providers in peak usage. For instance, some utilities in the U.S. are
the HVAC&R arena who respond to these expectations. offering real-time electric rates that more-or-less reflect the
true cost of providing electricity at any time. The rates can
D. Consolidation within the Facility Operations Industry
increase dramatically during periods of peak demand when
The structure of the industry that provides controls and less efficient generating capacity is employed and when it is
services for the operation and maintenance of buildings has necessary purchase additional generating capacity from
changed dramatically in the past 20 years. In the 1980’s, other utilities. The real-time rates are transmitted
only a handful of companies provided computerized control electronically some hours in advance. Some utilities in
systems for buildings. The control systems tended to be California are also experimenting with critical peak pricing
centralized with a single vendor providing all of the (CPP) rate structures. With CPP, utilities can impose very
necessary hardware and software. As control systems have high (super-peak) rates during designated times of the day
become more distributed and compatible with standard (e.g., 12 – 6 pm) on a limited number of CPP days (e.g., 6)
communication protocols, lower-level control hardware with throughout each year. The utilities provide an electronic
integrated applications software have become more or less signal to the customer in advance of the CPP event. With
commodities that can be provided by any number of either real-time or CPP rates, intelligent controls have the
competing companies based upon first cost and specific potential to automatically respond to these dynamic price
features. In this environment, the large control companies signals and reduce overall costs.
have significantly expanded and broadened their service There also is a trend in the U.S. of utilities offering
offerings to provide complete facility management and incentive programs to commercial customers to provide
maintenance. In fact, sales of controls hardware are only a “tune-ups” for packaged HVAC&R equipment. Programs
relatively small part of the income of the large control have been developed on both the east and west coasts that
companies that serve the buildings’ market. Other rely on diagnostic tools for diagnosing problems and
companies have entered and/or expanded in this growing evaluating performance before and after service. This is part
market segment to provide different aspects of facilities of a growing concern about the degradation in performance
management. The cost-to-benefit ratio for intelligent of packaged equipment that occurs in the field due to
operations improves as the industry moves towards large inadequate service. There are indications that the U.S.
organizations managing the operations and maintenance of government may eventually mandate minimum performance
many buildings. In particular, the cost of developing and requirements for in-situ performance of HVAC&R
managing the necessary software tools can be spread out equipment. This type of requirement would provide a
over a larger revenue base. tremendous boon for embedded tools that provide
In addition, large companies that provide HVAC&R continuous feedback on equipment performance and
equipment and service have purchased controls companies developing faults.
and recently a large controls and facility management
company purchased an HVAC&R equipment manufacturer. III. AUTOMATED DIAGNOSTICS
Companies that provide HVAC&R equipment, controls,
One of the most significant near-term opportunities for
operations, and service can embed equipment-specific
embedded intelligence in HVAC&R equipment is automated
intelligence within their products at a lower cost which helps
diagnostics and service scheduling. This technology
them in selling packages to customers for facility operations,
involves the use of sensors and information technology to
maintenance, and service.
monitor the health of equipment, diagnose problems, and
E. Greater Incentives for Intelligent Features recommend service. Diagnostic systems have been applied
The use of embedded intelligence within HVAC&R and for many years to critical systems, such as nuclear power
other equipment can lead to reduced utility costs, reduced plants and air craft. However, their development and
costs for providing operations, maintenance, and service, application for HVAC&R has been much slower because the

4378
FrA05.2

industry is extremely cost sensitive and the benefits are more Packaged air conditioning equipment accounts for
difficult to quantify and don’t generally include human approximately 60% of the installed cooling capacity within
safety. As a result, research and development efforts for the U.S and yet they generally are not well maintained with
HVAC&R applications have focused on low-cost solutions. current practices (Breuker and Braun, 1998; Jacobs, 2003;
Embedded diagnostics for packaged air conditioning Rossi, 2004; Li and Braun, 2007a). In a best case scenario,
equipment is an application of tremendous potential that is a customer enters into a maintenance contract that involves
near commercialization. Figure 7 depicts a vision for regular inspections and preventative maintenance. With
embedded diagnostic systems that is currently under skilled technicians, this can result in equipment that
development for rooftop air conditioners. Low cost sensors, performs efficiently with minimal down time. However, the
data monitoring, and equipment safety algorithms are contracts are relatively expensive and the customer has no
embedded within individual equipment. In the system under direct feedback on the benefits of service that is performed.
development, individual devices on each rooftop unit An embedded system would eliminate the need for
communicate through a wireless network on the roof to a expensive on-site inspections and provide continual
local server that provides the primary diagnostic capabilities feedback to customers on the condition of their equipment
and can communicate with a centralized server through an and the cost effectiveness of performing service. This
internet or cell phone connection. The central server would significantly reduce the costs of providing service
provides web-based reporting and is accessible by contracts and make the service contractor more accountable
authorized personnel across the Internet. This architecture to the customer, fostering a better relationship between the
tends to minimize costs for hardware and installation at the two. In the current environment, service tends to be
lowest level associated with individual pieces of equipment. performed in response to comfort complaints when
The use of low-cost wireless communications on the roof equipment is either not operating or when the cooling
significantly reduces installation costs. In this initial capacity has degraded to a point where space conditions
implementation, the higher cost elements, such as the local cannot be maintained. This leads to poor equipment energy
Shopping Mall efficiency and reduced equipment life.
Transmit data
Most of the diagnostic method development for packaged
Micro-
a/c a/c
air conditioners has been driven by the goal of keeping
controller
a/c Store C Store D
installed costs low and has focused on faults that degrade
B

Transducers a/c
A
system cooling capacity and efficiency and impact the life of
Thermo / Mechanical
Building 1 equipment (e.g., heat exchanger or filter fouling, high or low
Equipment
air flow, evaporator frosting, liquid line restrictions,
A/C Unit Transmit data
Building 2 refrigerant leakage, loss of compressor performance,
Benefits
Central presence of non-condensable gases, etc.). Most methods use
• improved comfort
• lower energy costs
Processing
Facility Building 3 relatively low-cost temperature and pressure measurements
• lower maintenance costs and use differences between measurements and expectations
of thermodynamic states at steady state conditions for
Fig. 7. Vision for a fully automated diagnostic system detection and diagnoses of faults. The expected values for
server and external communications, are shared leading to thermodynamic states under normal operation depend upon
lower costs per unit than if their functions were integrated. ambient and room conditions and the diagnostic methods
For sites with a small number of units, the local server and generally incorporate simple models for normal states that
wireless network could be eliminated and data and are specific to the particular unit (see Braun, 2003a).
diagnoses could be accessed using a hand-tool interface to Diagnostic methods that are based on changes in
individual units carried by a service technician. thermodynamic states do not handle multiple faults
The primary benefits of automated fault detection and occurring simultaneously because all the state variables
diagnostic systems for HVAC&R equipment are reduced change in response to the development of any fault. In
downtime, service costs, and utility costs. The use of an addition, these methods require measurements over a wide
embedded diagnostic system will allow a relatively small range of conditions to train reference models for each of the
support staff in a central location to monitor and manage thermodynamic state variables. In order to handle multi-
service for a large number of HVAC&R units and buildings simultaneous faults and keep costs low, decoupling features
in a more cost effective manner than is currently possible. can be incorporated that are uniquely dependent on
Although computer control and monitoring systems are individual faults and that use ‘virtual sensors’ (see Li and
employed in large commercial buildings, they do not yet Braun, 2007b,c,d). Virtual sensors estimate system
include many diagnostic capabilities. Furthermore, small to parameters using mathematical models in conjunction with
medium size commercial buildings do not typically employ physical sensors. Therefore, they rely on low cost
monitoring systems and yet they may incorporate a number measurements and models to estimate quantities necessary in
of packaged air conditioners. determining the decoupling features. For instance, a virtual

4379
FrA05.2

sensor was developed for refrigerant charge that can be used operating costs that would be enabled through availability of
to diagnose low or high refrigerant charge and only requires embedded equipment intelligence.
measurements of four temperatures. There has been relatively little development of intelligent
Diagnostic systems should incorporate indices that control strategies for small commercial buildings that utilize
measure equipment performance degradations in order to packaged air conditioning equipment because these
assess fault severity and evaluate whether service should be applications generally do not have computerized control
performed. Several investigators (see Braun, 2003a) have systems. However, there are significant opportunities for
used coefficient of performance (COP) as an economic cost savings in small buildings that could be realized at low
index and cooling capacity degradation as a comfort index installed costs through intelligent HVAC&R equipment.
for evaluating fault severity. More recently, Li and Braun One of the most exciting opportunities is the development of
(2007e) developed an economic performance degradation a smart thermostat. A smart thermostat would be a
index (EPDI) that is determined from low-cost replacement for existing thermostats used to control
measurements and includes the effects of energy efficiency individual packaged air conditioners. The thermostat would
and equipment life. attempt to minimize operating costs for air conditioning by
adjusting zone temperature setpoints over time within the
IV. INTELLIGENT CONTROLS comfort range to respond to time-varying electrical energy
There has been a significant amount of research and and demand rates and ambient conditions. It would need to
development related to intelligent control strategies that have access to utility rates and equipment performance
attempt to minimize utility costs for large commercial characteristics. The building load dynamics would need to
HVAC&R systems. Chapter 41 of the ASHRAE Handbook be learned through adaptation of a simple dynamic model
of HVAC Applications (ASHRAE, 2003) presents a number that could relate thermostat calls for cooling with occupancy
of different strategies for systems used in commercial schedules and weather conditions. A smart thermostat might
buildings along with results that demonstrate the cost also incorporate an occupancy sensor. In addition to being
savings potential. For the most part, these strategies are useful as an input to an adaptive model for building load
geared towards large commercial systems that incorporate dynamics, the smart thermostat could communicate with
chilled water distribution. Most of the strategies require intelligent lighting controllers to reduce local lighting levels
performance information for more than one piece of when appropriate. Although there is a significant
equipment in order to consider tradeoffs in energy usage. development effort required for smart thermostats, the
For instance, the strategy in the ASHRAE Handbook for economic benefits are substantial (see Braun , 2003b).
control of cooling towers requires information on tower fan
power requirements and the sensitivity of chiller power to V. CONCLUSIONS AND OUTLOOK
changes in condenser water temperature. Similarly, a The incentives and opportunities for intelligent HVAC&R
strategy for optimal sequencing of multiple chillers requires equipment have never been greater. However, the industry
part-load information for the individual chillers. In order to has been relatively slow to adopt new technologies and the
implement these strategies, a control engineer needs to pace is unlikely to increase substantially unless driven by
collect equipment performance information and configure customer expectations and/or significant cost reduction
controllers where the algorithms have been implemented. If opportunities. Customer expectations for intelligent features
the configuration of the plant changed, then the controllers are certainly increasing due to explosion of smart consumer
would need to be reconfigured. Unfortunately, these electronics. However, there is a need to develop more
requirements have been a barrier to widespread application documentation of the benefits and cost savings potential for
of these control strategies. intelligent HVAC&R equipment. This information is
The availability of embedded intelligence in individual needed by developers in the design of cost effective systems
HVAC&R equipment could have a tremendous impact on and also by marketers who will need to promote this
the cost of implementing intelligent control strategies technology.
leading to much wider application. For instance, a cooling The most likely first implementation of embedded
tower fan controller could have access to embedded intelligence for HVAC&R equipment and appliances will be
performance information from each tower fan and each for documentation. A chip integrated in the factory could
chiller across the local network. Similarly, a chiller contain detailed information necessary for installation,
sequencing controller could obtain part-load information commissioning, operation, maintenance, warranty, and
from individual chillers. The use of embedded intelligence repair and be accessible from a local computer or handheld
would substantially reduce the human effort required to device through a USB and/or wireless connection. For
configure or reconfigure these controllers and also lead to medium to larger commercial buildings, this information
better performance. For large commercial cooling systems, could be accessible across a local network provided for
the existence of intelligent control algorithms means that system control and monitoring. Embedded documentation
there are many near-term opportunities for reducing could dramatically reduce the chances of losing this

4380
FrA05.2

information and improve the prospects for proper [10] EIA, 1999, “Commercial Buildings Energy Consumption Survey
(CBECS),” U.S. Department of Energy, Energy Information
installation and operation. Administration, http://www.eia.doe.gov/emeu/cbecs/contents.html .
Performance monitoring and diagnostics will begin to be [11] ElectronicNews, 2004, “MEMS Come to OZ Wine Industry,” Online
integrated within HVAC&R equipment used in commercial Staff of Electronics News, June (Electronics News is the Australian
sister publication of Electronic News).
buildings within the next five years. In the near term for
[12] Gordon, L.M. and T. Haasl, 1996, “Operation and Maintenance in
small commercial buildings, this capability will allow a Office Buildings: Defining Baseline,” Proc. ACEEE Summer Study on
service technician or owner to monitor and manage service Energy Efficiency in Buildings. Pacific Grove, CA, August.
using a handheld or local computer with a USB and/or [13] Jacobs, P., V. Smith, C. Higgins, and M. Brost, 2003, “Small
Commercial Rooftops: Field Problems, Solutions and the Role of
wireless connection. As customer confidence and Manufacturers,” Proc. of the 11h National Conference on Building
expectations for better economic performance grow, there Commissioning, 20-22 May.
could be a dramatic expansion of contracts with service [14] Katipamula, S. and M.R. Brambley, 2004, “Wireless Condition
Monitoring and Maintenance for Rooftop Packaged Heating,
providers who manage maintenance and repair using remote Ventilation and Air-Conditioning,” Proc. ACEEE Summer Study on
monitoring connected across a network to intelligent Energy Efficiency in Buildings, Pacific Grove, CA, August.
HVAC&R equipment. For commercial HVAC&R [15] Kele, 2004, Kele Frontier Product Line, www.kele.com, July.
[16] Kintner-Meyer, M., and M.R. Brambley, 2002, “Pros & Cons of
applications, diagnostic capabilities should ultimately be Wireless,” ASHRAE Journal, November.
embedded in nearly all controllers that are packaged with the [17] Kintner-Meyer, M. and R. Conant, 2004, “Opportunities of Wireless
devices that they control. This includes chillers, rooftop Sensors and Controls for Building Operation,” Proc. ACEEE Summer
units, air handling units, VAV boxes, etc. Study on Energy Efficiency in Buildings, Pacific Grove, CA, August.
[18] Kiyon, 2004, Kiyon, Online Product Information,
The outlook for widespread application of intelligent www.kiyon.com/Bacnet_eval_kits.html, July.
control of HVAC&R equipment for minimum utility costs is [19] Knisley, J., 2003, “Bright Ideas in Lighting Control,” Electrical
not as optimistic. In order to properly consider tradeoffs in Construction & Maintenance, April.
[20] Kranz, H.R. and O. Gisler, 2002, “Standardization & IT Technology
performance for intelligent control, it is necessary to have Will shape the BACS Industry,” AutomatedBuildings.com, January.
specific knowledge of equipment performance. Although it [21] Li, H. and J.E. Braun, 2007a, “An Economic Evaluation of the
is clearly possible to integrate the necessary information Benefits Associated with Application of Automated Fault Detection
and Diagnosis in Rooftop Air Conditioners,” To appear in ASHRAE
within chips on individual equipment, equipment Transactions.
manufacturers are unlikely to do so without a mandate from [22] Li, H. and J.E Braun, 2007b, “A Methodology for Diagnosing
the government or a recognized standard. A standard could Multiple-Simultaneous Faults in Vapor Compression Air
Conditioners,” To appear in International Journal of Heating,
provide an opportunity for smaller players to gain a
Ventilating, Air-Conditioning and Refrigeration Research.
competitive advantage which could ultimately compel the [23] Li, H. and J.E. Braun, 2007c, “Decoupling Features and Virtual
larger players into compliance. This is essentially what Sensors for Diagnosis of Faults in Vapor Compression Air
happened in the development of and compliance with the Conditioners, To appear in International Journal of Refrigeration.
[24] Li, H. and J.E. Braun, 2007d, “Evaluation of a Decoupling-Based
BACNet communications protocol. Perhaps the BACNet Fault Detection and Diagnostic Technique – Part I: Field Emulation
protocol could eventually be amended to include equipment Evaluation and Part II: Field Evaluation and Application, To appear in
performance characteristics necessary for implementation of Journal of Harbin Institute of Technology.
[25] Li, H. and J.E. Braun, 2006e, “An Overall Performance Index for
intelligent control strategies. Characterizing the Economic Impact of Faults in Direct Expansion
Cooling Equipment,” International Journal of Refrigeration, Vol.30,
REFERENCES No. 2, Pages 299-310,.
[26] Lowry, G., 2002, “Factors Affecting the Success of Building
[1] ASHRAE, 2003, ASHRAE Handbook of HVAC Applications, Management System Installations,” Building Serv. Eng. Res.
“Supervisory Control Strategies and Optimization,” Chapter 41, pp. Technology, vol. 23, no. 1, pp. 57-66.
41.1 - 41.39, ASHRAE. [27] Marshall, R., 2003, “Practically Wireless,” Chemical Engineering,
[2] BCS Partners, 2002, “The Building Control Systems Market (2001- December.
2006)”, Report by BCS Partners, July. [28] Roth, K.W, D. Westphalen, M.Y. Feng, P. Llana, L. Quatararo, 2005,
[3] Braun, J.E., 2003a, “Automated Fault Detection and Diagnostics for “Energy Impact of Commercial Building Controls and Performance
Vapor Compression Equipment,” ASME Journal of Solar Energy Diagnostics,” TIAX Reference No. D0180, November.
Engineering, Vol. 125, pp. 266-274. [29] Rossi, T.M., 2004, “Unitary Air Conditioner Field Performance,”
[4] Braun, J.E., 2003b, “Load Control using Building Thermal Mass,” Proceedings of the 2004 International Refrigeration and Air
ASME Journal of Solar Energy Engineering, Vol. 125, pp. 292-301. Conditioning Conference at Purdue, Purdue University, W. Laf., IN.
[5] Breuker, M.S. and J.E. Braun, J.E., 1998, “Common Faults and Their [30] Thomas, G., 2005, "Industrial Ethernet is Now a True Fieldbus,"
Impacts for Rooftop Air Conditioners,” International Journal of automatedbuidlings.com, May.
Heating, Ventilating, Air-Conditioning and Refrigerating Research, [31] Tom, S., 2005, “XML in the Real World with Real-World Examples,”
Vol. 4, No. 3, pp. 303-318. Presentation summarized in CABA Event Report: “XML Symposium
[6] CABA, 2004, “Reliable Wireless Networks for Industrial Systems” 2005,” ER 2005-03.
May, Report #IS 2004-21. [32] Turpin, J., 2004, “The Wireless Technology Revolution: What It
[7] Delin, K., 2004, “The Sensor Web: A Distributed, Wireless Means For BAS,” Engineered Systems Magazine, February.
Monitoring System,” Sensors Online, April. [33] Wang, W. and M. Nova, “Wireless Bridges the Technology Gap in
[8] DeNamur, M., 2002, “BAS Trends,” HPAC Engineering, June, pp. Building Automation,” Industrial Ethernet Book, Issue 17, November.
19-22, [34] Yashar Y. and P.A. Domanski, 2004, “MEMS Sensors for
[9] EIA, 1995, “Commercial Buildings Energy Consumption Survey HVAC&R,” ASHRAE Journal, May, pp. 69-74.
(CBECS),” U.S. Department of Energy, Energy Information
Administration, http://www.eia.doe.gov/emeu/cbecs/contents.html .

4381

You might also like