Proactive Management for Smart Buildings
Resource Management Strategy
Milan Prodanović, Maria Belén Téllez Molina, Alessandro Gallo and José González Aguilar
Institute IMDEA Energía, Avenida Ramón de la Sagra 3, 28933 Móstoles, Madrid, Spain
Keywords: Energy Efficiency, Demand Management, Proactive Control, Smart Buildings.
Abstract: Self-sustainability and near-zero energy consumption have already become common requirements for
construction of new buildings. Apart from the existing traditional ways for improving building energy
efficiency like introducing novel architectural solutions or new building materials, the role of energy
management has been increasingly seen as pivotal. Nowadays buildings rely on their management systems
to provide optimised energy supply solutions and meet their energy demand targets from on-site generation
and storage installations. In this paper the latest developments in proactive approach in management of
building resources will be presented and critically compared to the existing electrical energy efficiency
methods based on reactive type of control. Simulation and cost/benefit results are used to demonstrate the
performance improvements achieved by deploying the proactive management system.
1 INTRODUCTION
The term “Smart Buildings” normally refers to
various control techniques used for integration of
renewable energy sources to buildings, energy
efficiency improvement, reduction of greenhouse
emissions and application of demand side
management. The results of various studies on
incorporation of renewable energy technologies into
buildings reveal in many cases lack of a holistic
approach to investigation and optimisation of small
energy supply systems. The main drawbacks are
identified as an absence of analysis of end-use
energy demand, lack of pre-definition of energy
system structure in initial phases of building design
and finally not taking advantage of more sensitive,
proactive energy management schemes.
Once the energy supply mix has been
determined, further advances can be achieved
through resource management (Téllez, 2011),
(Prodanovic, 2012). What most of the existing
building energy management systems (BEMS)
(ABB, 2010), (Vikon, 2009) have in common is
their reactive nature of control as they mostly
attempt to react on an unexpected event in order to
accomplish actions to rectify the situation. Although
these technologies may lead to construction of more
“ecologically friendly” buildings, novel BEMS
should not be limited to energy conservation
measures only, but also to look for any proactive
ways to improve energy performance and potentially
achieve self-sustainability. Yet, recent advances in
energy storage technologies, demand side
management techniques, on-site CHPs, intelligent
appliances along with development of sensor
networks and information technology suggest that
near real-time optimal dispatch of all installed
resources is possible (Doukas, 2007) and
(Gupta,2010).
New hierarchical methods for proactive
management of multiple energy sources to meet total
energy demand of a small, local scale energy system
are being under investigation (Brooks, 2010). In
order to propose and develop management systems
capable of taking strategic decisions, all the relevant
control aspects need to be considered such as
prediction, scheduling and real-time control for
generation and demand resources installed
(including demand side management and demand
dispatch). Energy storage devices and energy
exchange with electricity grid play an important role
in these new schemes as they add more options in
management of installed energy sources by
accommodating for their different dynamic
properties and by compensating for any supply or
demand uncertainties. To further improve the real-
time energy balance, shorter intervals for the overall
optimisation so that the system is in position to
165
Prodanovic M., Belén Téllez Molina M., Gallo A. and Gonzalez Aguilar J..
Proactive Management for Smart Buildings - Resource Management Strategy.
DOI: 10.5220/0004408801650170
In Proceedings of the 2nd International Conference on Smart Grids and Green IT Systems (SMARTGREENS-2013), pages 165-170
ISBN: 978-989-8565-55-6
Copyright
c
2013 SCITEPRESS (Science and Technology Publications, Lda.)
proactively manage any various dynamic
uncertainties of renewable energy sources (RES, like
solar radiation, wind speed), demand profiles, prices,
efficiency of energy conversion subsystems etc.
Figure 1: Building management system.
In this paper the results of the ongoing research
aimed to achieve additional cost savings and
performance improvements in the building resource
management will be presented. The flexibility of
both demand and generation will be discussed and a
hierarchical control structure for prediction based
proactive management will be discussed. The
simulation results are used to demonstrate the
system performance and assess the cost benefits of
the proposed scheme.
2 ENERGY DEMAND
In order to create a representative energy demand
profile for a building that can be used for demand
prediction and energy management, it is necessary to
consider all the specific functional details and
services available. User profiling is one of the most
important aspects that can be linked directly with the
demand. The building occupancy and external
ambient data are some of the crucial aspects.
Figure 2 shows a daily profile of the hotel
occupancy indicating when the clients are inside the
building and when they are active. This information
directly relates to the way the services are used.
In Figure 3 an aggregated daily demand profile is
depicted for a hotel. The main contributors to the
energy demand can be easily identified: laundry,
restaurant food preparation, HVAC etc.
What is relevant for the management system is to
determine what loads may be shifted in time or
trimmed in power. In this way some additional load
flexibility can be obtained and used for demand
prediction and system optimisation.
Figure 2: Guest behaviour daily profile.
Figure 3: Aggregated profile of daily demand in summer.
3 ENERGY RESOURCES
In the past most of the buildings depended solely on
the electricity grid and on-site diesel generation to
provide for their energy needs. In recent years there
is a significant effort to introduce concepts of energy
self-sufficiency and near zero consumption in the
building sector. To achieve that, more conventional
and RES have been integrated to buildings.
The creation of the generation mix including the
energy storage devices is a tecno-economic exercise
and several references (Téllez, 2011),
(Todorovic, 2011) deal with this issue. Just as an
example, different energy system configurations
have been compared in Figure 4 for their cost
savings, self-consumption and self-sufficiency when
a special “super off-peak” tariff is used (favouring
nocturnal energy use and having three pricing
periods).
Once the battery system is installed there are
many more options for proactive energy
management and cost savings. Just charging the
batteries during the low pricing periods can account
to about 10% of savings. Yet, the batteries provide
more flexibility and can improve energy capture
from intermittent resources so that the locally
generated energy is neither exported (at lower
SMARTGREENS2013-2ndInternationalConferenceonSmartGridsandGreenITSystems
166
prices) nor left unused.
What is relevant from the management point of
view is to determine the main control parameters of
the installed resources. In this way it is important to
identify what resources are dispatchable, what
resources are introducing the uncertainty to the
system, dynamic properties etc.
Figure 4: Cost savings, self-consumption and self-
sufficiency for different generation configurations.
4 PROACTIVE MANAGEMENT
In contrast to the most of energy efficiency methods
used in the building sector to achieve energy and
operational cost savings (based on predefined
resource scheduling and reactive actions), the
proactive management uses historical and real-time
data to predict the demand and then optimise in
advance the commitment of all installed system
resources.
4.1 Control Steps
In order to integrate a mix of diverse technologies,
the dynamic properties of the installed equipment
and control processes need to be taken into account.
It is, therefore, important to adopt a hierarchical
management structure that is the most suitable for
the respective response times required. Figure 5
depicts the necessary levels for successful
application of the proactive management.
Figure 5: Hierarchical structure of management system.
In the first step (Prediction), the short-term weather
forecast, user behaviour predictions, external factors,
historical data and system ratings are all used to
obtain one-hour estimates of the RES supply and the
consumption patterns for the following day. This
prediction step can be updated up to one-hour ahead
to improve the accuracy of control.
An optimal plan for each hour is calculated in the
next step (Scheduling) where the set-points of
controllable sources are adjusted to meet the
expected demand according to the RES profiles,
cost, DSM and storage strategies aforementioned.
The role of the next steps (RT control) is to
actively balance supply and demand by managing
disturbances introduced to the system by weather
conditions, prices, human behaviour etc. The
demand management actions applied continuously
(demand dispatch, DD) here include only an
adjustment of the controllable loads. DD selects
loads with scheduling and/or intensity freedom so
that their modification will not affect the user
comfort. For example, refrigerators, dishwashers,
etc. could advance start times of their cycles or
comfort temperatures could be slightly increased in
air conditioning devices. Another option to provide
additional balancing capacity is to reschedule the
energy storage devices. Finally, the fine adjustment
step includes any necessary minute based balancing
requirements of the system.
By applying the four optimisation steps and
taking into account the dynamic properties, the
prime exploitation of the installed resources can be
achieved.
4.2 Real-time Operation
The RT control is continuously making decisions for
all the dispatchable energy resources in order to
reduce the instantaneous mismatch between the
supply and consumption. Energy imports (exports)
from (to) the grid are considered in order to capture
any change originated from the applied access
tariffs.
The main entities in the control process are
shown in Figure 6 and can be classified into one of
these four different categories: on-site generation,
load, storage or the grid connection. The sources are
divided in dispatchable and non-dispatchable ones.
Majority of the RES are considered as non-
dispatchable because of their unpredictable
behaviour. RES mostly present in commercial
buildings are small wind turbines, photovoltaic and
solar thermal systems.
On contrary, dispatchable sources can be
adjusted to any control demand. Along with gas
turbines, reciprocating engines, etc. small scale
Combined Heat and Power (CHP) plants are
frequently present to avoid any heat waste during the
ProactiveManagementforSmartBuildings-ResourceManagementStrategy
167
electricity production.
Conversely, sinks of energy comprise of
controllable and non-controllable loads. Controllable
load is any schedulable, curtailable, deferrable and
adjustable load under some restrictions, hence it is
suitable for application of DD. Lighting, HVAC
systems, and some kitchen and cleaning appliances
generally belong to this group. Yet, entertainment
and essential personal appliances fit in the
uncontrolled group.
Storage units and grid connection can be used
either as sources or sinks according to system needs.
Various tariffs can be applied when accessing the
electricity grid for purchase or sale. Regarding
storage systems, batteries and thermal tanks are
frequently found in commercial buildings.
Finally, the core of the RT controller consists of
a sensor network providing continuous system
measurements, data base for storing and using the
historical data and an intelligent decision maker to
coordinate the operations.
Two most common situations the controller is
subjected to are:
- Energy shortage, caused mainly by reduction of
RES availability due to unpredicted weather
conditions or sudden increases of demand. Typical
control actions include: curtailment or postponement
of controllable loads, utilization of energy storage,
using additional on-site (dispatchable) generation
and importing (buying) energy from the grid (in case
it is more economical and faster than the
aforementioned alternatives).
- Energy surplus, when more favourable weather
conditions than predicted arise or when there is a
demand reduction by, for example, a lower
occupancy level than expected. Typical solution is to
store the energy excess for future use (if possible),
advancing cycles of some controllable device or
exporting energy to the grid.
Figure 6: Information exchange in real-time.
In case a shortage or a surplus of energy remains
for considerable time, a permanent change of
weather or behaviour conditions should be
considered.
4.3 Optimisation Problem
The dispatch decisions are continuously made by the
controller. This complex task takes into
consideration the stochastic nature of the parameters
involved, dependence on the past and current
decisions, future (scheduled) actions and multi-
objective options of the optimisation problem.
Favouring local renewable generation and occupant
comfort maximization are also valid optimisation
targets.
The decisions are made for a particular interval,
but the optimisation problem also affects the
successive intervals as they are not independent.
In the proactive control, the system is configured
to be sensitive to any variation that may affect the
demand or generation. Figure 7 shows the block
diagram of how the optimised set-point levels for
each unit are generated. This generation includes
both, present time and future intervals depending on
the predicted future scenarios.
Figure 7: Generation of control actions.
4.4 Test Case
An example of the daily demand coverage of a hotel
during summer months is used to show the features
of proactive management. It is assumed the “super
off-peak” tariff is used, favouring the night time
consumption from the grid. The daily scheduler
optimized the generation, storage and demand
resources based on the most recent available data at
midnight. The washing services were scheduled to
coincide with the maximum PV generation and the
lighting during these hours is permitted to operate at
the maximum power. The batteries are charged from
the PV generation during the daytime and they are
used during the evening hours when the food
preparation increased the total demand. Figure 8a
shows the demand coverage in more details.
The solar irradiance forecast was then updated
with more accurate data at 5.00h indicating lower
SMARTGREENS2013-2ndInternationalConferenceonSmartGridsandGreenITSystems
168
than expected insolation between 12.00h and 16.00h.
As a result of the updated optimization, the batteries
are charged during the night time and discharged
during that interval, some additional import from the
grid is necessary and the washing was scheduled to
finish before then. Also, the levels of lighting were
reduced. The updated demand coverage is shown in
Figure 8b.
Finally, at 9.00h the demand prediction for the
evening hours was once again revised, as the
information about the cancellation of the dinner
party booked at the restaurant was received. The
reduction of the demand shown in Figure 8c allowed
more accumulated battery energy to be used to cover
the evening demand and reduced the purchase of
energy from the grid.
a) Optimisation at = 00:00 h.
b) Optimisation at = 05:00 h.
c) Optimisation at t = 09:00 h.
Figure 8: Updates of demand coverage according to
revised scheduling.
Figure 9 shows the proactive management and
the applied control updates in more details. In these
diagrams on the left side the total generation is split
between non-dispatchable generation Eg,nd, energy
storage discharge Eb,dschg, electricity purchase
Egrid,in and then power adjustment El,pc. On the
right side electrical non-controllable load El,nc,
energy storage charge Eb,chg, electricity export
Egrid,out and time-shiftable load El,tc are shown.
The main difference between the plots was
introduced by lower than expected non-dispatchable
generation Eg,nd. This as an effect rescheduled the
times of the charging and discharging the batteries
Eb,chg and Eb,dschg, the grid purchases and sales
Egrid,in and Egrid,out. Also, some changes to these
profiles were caused by the demand reduction in the
evening hours.
a) Optimisation at = 00:00 h.
b) Optimisation at = 05:00 h.
c) Optimisation at = 09:00 h.
Figure 9: Daily evolution of main control parameters.
ProactiveManagementforSmartBuildings-ResourceManagementStrategy
169
5 BENEFITS ANALYSIS
To demonstrate the advantages of proactive control
in terms of minimising the running costs, the savings
have been calculated and compared for the three
different tariffs applied. As the base case for this
comparison it is assumed the fixed daily schedule
without any changes applied by the proactive
management. In this base case the battery charge and
discharge cycles are fixed and any demand variation
is covered by importing/exporting the power from/to
the grid. In Figure 10 the increase in running costs as
a result of not applying the hourly adjustment is
shown. For the single tariff case the cost saving is
around 5% while for the two-tariff case and the
“super off-peak” tariff the potential for the savings
increase to about 9%.
Figure 10: Increase in operational costs when proactive
control updates are not applied.
The proactive management clearly demonstrates its
capacity to achieve cost savings with or without the
tariff discrimination system by matching the demand
peaks with the available local generation. In case of
the differential pricing system, the additional saving
benefits are gained by the demand shifting
techniques and the battery management.
6 CONCLUSIONS
This paper has discussed the operational and cost
related aspects of the proactive resource
management system for Smart buildings. It has been
demonstrated that the proposed solution is sensitive
to any unforeseen disturbance in demand,
generation, access tariffs, ambient conditions,
building occupancy, building services etc. and by
using optimised dispatch significantly reduces
operational costs.
The proposed methodology introduces a
hierarchical approach to the multi-energy system
design, long term resource scheduling and proactive
RT response. Energy surplus and energy shortage
scenarios have been both considered and with
optional capacity restrictions in the system grid
connection.
Simulation results and cost/benefit analysis have
been used to demonstrate some of the features of the
proposed proactive control.
ACKNOWLEDGEMENTS
The authors would like to thank Sacyr Vallehermoso
and Ministry of Science and Innovation,
Government of Spain, for their help and funding
research project THOFU (Cenit).
REFERENCES
Téllez M. B; Gafurov, T.; Prodanović M., 2011: Proactive
control for energy systems in smart buildings, IEEE
PES Innovative Smart Grid Technologies 2011 Europe
Prodanović, M., Gafurov, T., Tellez, M. B., 2012. A
demand based approach to optimisation of energy
supply mix for smart buildings, IEEE PES Innovative
Smart Grid Technologies, 2012
ABB, 2010., Technology of choice smart home and
intelligent building control ABB i-bus® KNX.
[Online]. Available: http://knx-gebaeudesysteme.de/
sto_g/English/GENERAL_DOCUMENTATION/Tech
nology_of_choice_2CDC500064M0201.pdf
Vikon by Tridium 2009. Creating better buildings.
[Online]. Available: http://www.vykon.com/galleries/
brochures/VYKONOverviewBroch.pdf
Doukas, H., Patlitzianas, K. D., Iatropoulos, K., and
Psarras, J., 2007, Intelligent building energy
management system using rule sets, Building and
Environment
Gupta, A., Saini R. P.,and Sharma, M. P., 2010, Modelling
of hybrid energy system -Part I: Problem formulation
and model development, Renewable Energy, 36: 459-
465, June 2010.
Brooks, A., Lu, E., Reicher, D., Spirakis, C., and Weihl.
B., 2010, Demand dispatch: Using real-time control of
demand to help balance generation and load, Power
and Energy Magazine, IEEE, 8(3):20–29, June 2010.
Todorović, M. S., Djurić, O. E., Matinović, I., Ličina, D.,
2011, Renewable Energy Sources and Energy
Efficiency for Building's Greening: From traditional
village houses via high-rise residential building's BPS
and RES powered co- and tri-generation towards net
ZEBuildings and Cities, in Proc. 2011 IEEE
International Symposium on Exploitation of
Renewable Energy Sources, pp. 29-37.
SMARTGREENS2013-2ndInternationalConferenceonSmartGridsandGreenITSystems
170
