Energy and Buildings 90 (2015) 65–75
Contents lists available at ScienceDirect
Energy and Buildings
journal homepage: www.elsevier.com/locate/enbuild
Optimal electrical and thermal energy management of a residential
energy hub, integrating demand response and energy storage system
Faeze Brahman, Masoud Honarmand, Shahram Jadid ∗
Center of Excellence for Power System Automation and Operation, Department of Electrical Engineering, Iran University of Science and Technology,
P.O. Box 1684613114, Tehran, Iran
a r t i c l e
i n f o
Article history:
Received 31 August 2014
Received in revised form 3 November 2014
Accepted 21 December 2014
Available online 31 December 2014
Keywords:
Combined cooling heating and power
(CCHP)
Energy hub
Demand response (DR)
Plugged in hybrid electric vehicle
Thermal energy storage
a b s t r a c t
Energy crisis along with environmental concerns are some principal motivations for introducing “energy
hubs” by integrating energy production, conversion and storage technologies such as combined cooling,
heating and power systems (CCHPs), renewable energy resources (RESs), batteries and thermal energy
storages (TESs). In this paper, a residential energy hub model is proposed which receives electricity, natural gas and solar radiation at its input port to supply required electrical, heating and cooling demands
at the output port. Augmenting the operational flexibility of the proposed hub in supplying the required
demands, an inclusive demand response (DR) program including load shifting, load curtailing and flexible
thermal load modeling is employed. A thermal and electrical energy management is developed to optimally schedule major household appliances, production and storage components (i.e. CCHP unit, PHEV
and TES). For this purpose, an optimization problem has been formulated and solved for three different
case studies with objective function of minimizing total energy cost while considering customer preferences in terms of desired hot water and air temperature. Additionally, a multi-objective optimization
is conducted to consider consumer’s contribution to CO2 , NOx and SOx emissions. The results indicate
the impact of incorporating DR program, smart PHEV management and TES on energy cost reduction of
proposed energy hub model.
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Recently, environmental issues such as increasing level of pollutant emissions along with the rapid growing of energy demand
and the surge in fuel cost has drawn particular attention to distributed energy resources (DERs). DERs contribute to system by
cutting the expenses and being more environmental friendly in
comparison with centralized power systems. In this regard, combined heat and power (CHP) units are one of the most beneficial
technologies and their performance will precisely be clarified in
this paper. The main advantage of a CHP unit is its potential of
generating both power and heat simultaneously. Accordingly, the
whole system efficiency is improved by using the wasted heat to
satisfy the heating demand. Moreover, combined cooling, heating
and power units (CCHPs) are becoming widely desirable and even
more economical due to the fact that they can meet the cooling
demand along with electrical and heating one. As stated in [1], efficiency of CCHP system is up to 60–80%, which is considerably higher
∗ Corresponding author. Tel.: +98 21 77491223; fax: +98 21 77491242.
E-mail addresses: faeze brahman@elec.iust.ac.ir (F. Brahman),
honarmand@elec.iust.ac.ir (M. Honarmand), jadid@iust.ac.ir (S. Jadid).
http://dx.doi.org/10.1016/j.enbuild.2014.12.039
0378-7788/© 2014 Elsevier B.V. All rights reserved.
than those of traditional energy supply systems. Although, CCHPs
can reach their highest capabilities in residential and commercial
sectors, their collaboration with the upstream grid, by providing
reserve and peak shaving services, are beyond doubt [2].
Several studies have proposed optimal operation of CCHPs with
regard to electrical, heating and cooling demand as well as economic and environmental consideration [3–6]. The authors in [7],
presented mathematical models of a CCHP components, then a
multi-objective optimization has been formulated so as to minimize energy cost and greenhouse gas emissions of a commercial
Micro-Grid (MG). In this respect, “Energy Hub” concept has been
firstly initiated at ETH Zurich [8]. As defined in [9], an Energy
Hub is an integrated system with multiple energy carriers at its
input where energy production, conversion and storage technologies such as CCHP, renewable energy resources and batteries are
deployed in order to supply certain required services such as electricity, heating and cooling at its output.
The application of CCHP systems accompanied by photovoltaic
(PV) systems can enhance the system overall performance particularly in residential energy hubs. Albeit the efficacy of PV and other
DERs has been proved, their unsteady nature poses big challenges
to the operation of power systems. Several candidate solutions have
been proposed to address this issue. Energy storage systems (ESSs)
66
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
Sets
t
Ti
i
A
time interval t ∈ {1, . . ., 24}
desired time window of appliance i Ti = [˛i , ˇi ]
index of appliances
set of shiftable appliances A = {wm,dw,dry,pp,ir}
Subscripts
must run hours of appliance
MRH
MUT
minimum up time of appliance
crit
critical
Constants
electrical and thermal efficiency of CCHP unit
e ,th
ˇ
converting factor of 1 kWh to m3 natural gas.
min , P max minimum and maximum electrical output of
Pcchp
cchp
CCHP
min , H max minimum and maximum thermal output of
Hcchp
cchp
CCHP
rre
electrical ramp rate of CCHP (kW/h)
PV array area (m2 )
S
I
solar radiation (kW/m)
t
outdoor illumination at time t
ILout
t
ILreq
required illumination at time t
temperature of entering cold water (◦ C)
Tcw
V
water storage volume (L)
Cair
heat capacity of air (kWh/◦ C)
t
Vcold
volume of entering cold water (L)
specific heat of water (kWh/◦ C)
Cwater
R
thermal resistance of the house shell (◦ C/kW)
des
Tws
desired temperature of water storage (◦ C)
min , T max minimum and maximum water storage temperTws
ws
ature deviation
des
Tin
desired indoor temperature (◦ C)
min , T max minimum and maximum indoor temperature
Tin
in
deviation (◦ C)
t
hourly outdoor temperature (◦ C)
Tout
in , dr injecting and drawing heat efficiencies
PHEV’s battery capacity (kWh)
Cap
Pli
rated power of lighting (kW)
N
number of shiftable loads
t
TOU
time-of-use price tariff at time t
natural gas price
NG
2 ,NOx ,SOx emission factors of network electricity
CO
net
2 ,NOx ,SOx emission factors of natural gas
CO
f
Variables
t
Pcchp
CCHP electrical output at time t
t
Hcchp
recovered heat from PGU at time t
t
Fcchp
fuel consumption of CCHP at time t (m3 )
t
ILin
t
Tws
t
Tin
t
Pnet
t
Hws
t
Hair
indoor illumination level at time t
water storage temperature at time t (◦ C)
indoor temperature at time t (◦ C)
exchange power with network at time t (kW)
exchange heat with water storage at time t
required thermal energy to set the home temperature at time t (kWh)
injected/drawn heat from TES at time t(kWh)
TES energy content at time t(kWh)
PHEV’s battery charging (kW)
t , Ht
Hin
dr
t
Qtes
t
Pch,phe
v
t
Pdch,phe
v PHEV’s battery discharging (kW)
t
DEtotal
total electrical demand at time t (kW)
t
Dcrit
critical electricity demand at time t (kW)
t , Ht
EEC
electricity and heat supplied to electric and absorpAC
tion chiller at time t
Binary variables
t
scchp
CCHP on/off status at time t
sit , uti , dit on/off, start up and shut down status of appliance i
utin , utdr int injecting/drawing state of TES at time t
chtphev , dchtphev PHEV’s battery charging/discharging state
are being installed to reduce the mismatch between energy supply
and demand. Recent developments in energy storage technologies
have introduced Plug-in Hybrid Electric Vehicle (PHEV) as a decent
solution [10,11]. Smart building and PHEV are two promising technologies. The integration of these two emerging technologies holds
great promises in improving the power supply reliability and the
flexibility of building energy and comfort management [12]. The
owner of a PHEV usually use the vehicle for a couple of hours during a day and the vehicle is available in the home garage for the rest
of the day. Hence, PHEV’s battery can be considered as ESS [13–15].
In a building integrating DERs, PHEVs can contribute to the system
by charging or discharging that are called grid-to-vehicle (G2V) and
vehicle-to-grid (V2G), respectively [16].
According to [17] an indisputable fact about micro-CHP-based
building is to accurately coordinate its thermal and electrical loads.
Therefore, an energy management system (EMS) is established
as a promising mean to optimally coordinate all generation, consumption and energy storage resources. Various studies have been
conducted to develop the EMS model and to demonstrate its advantages by serving both economic and technical facets. Ref. [17] has
proposed a smart scheduling of a micro-CHP based MG and a temperature dependent thermal load modeling. Several uncertainties
associated with temperature, electrical and thermal load are considered in the proposed model. A smart EMS has been presented
in [18] incorporating power forecasting module, ESS, and optimization module to achieve a great coordination between power
production of DG units and ESS.
The urge to manage the unpredictability of RESs caused a
great interest in the deployment of Demand Response (DR) programs [19]. As defined in [20], “DR is when consumers voluntarily
change their energy consumption pattern from their nominal one,
in response to price signal variations or when motivated by incentive payments offered by utilities in order to maintain the system
reliability during on-peak periods”. It can be concluded that DR
programs not only benefit participated consumers with incentive
payments as well as bill savings, but also the utilities by enhanced
system reliability, modified load shape and better market performance [21]. The impact of Demand Side Management (DSM)
programs on systems incorporating renewable energy resources
has been addressed in several studies [22,23]. In [23] a Demand
Response Provider (DRP) aggregates customer load reduction offers
and a stochastic method has been used to capture wind and solar
power uncertainties as well as demand forecast errors. The authors
in [24] proposed a stochastic model to schedule both energy and
reserve by generating units and responsive loads. The applied
DR programs helped to cover uncertainties associated with wind
power forecasting. Economic value of an extended energy hub with
heat DSM capability is determined in [25] based on Monte Carlo
simulation.
In this regard, this paper aims at presenting an optimal energy
management strategies for a residential energy hub in order to well
coordinate CCHP unit, PV panels, PHEV and Thermal Energy Storage
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
(TES) while satisfying electrical and thermal demand. Two potential
solutions are considered in order to mitigate the negative impacts
of PV generation intermittency and to reduce energy cost of the
proposed hub. Load shifting and load curtailing are two specific
forms of DR programs in which load is shifted to low price periods or
curtailed according to customer preferences, and PHEV is an active
load which can be smartly scheduled based on TOU price tariffs.
For this purpose, an optimization algorithm has been applied on a
residential energy hub with the objective of minimizing its energy
cost as well as considering consumer’s contribution to CO2 , NOx and
SOx gas emissions. Briefly, the novelty of this work is highlighted
as following:
• A comprehensive EMS model to schedule CCHP, PV, PHEV and TES
in a residential energy hub.
• Implementing a thorough DR scheme, including load shifting,
load curtailing and flexible thermal load modeling.
• Considering two types of ESS, which are PHEV and TES.
The rest of the paper is organized as follows: mathematical
model of the residential energy hub is developed in Section 2. The
optimization problem is described in Section 3. Case studies and
simulation results are discussed in Section 4. Finally, main conclusions are provided in Section 5.
2. Residential energy hub model
The energy hub studied in this paper, consists of some forms
of generation and storage devices like CCHP, PV panels, PHEV and
TES, as shown in Fig. 1. A thermal energy storage ensures a more
efficient usage of the collected solar energy and CCHP [26]. Enhancing the reliability of the CCHP system, it is also connected to the
upstream grid for selling/purchasing energy when there is energy
excess/shortage, respectively. It is expected that each consumer
will manage not only the loads, but also small generation units,
storage systems, and electric vehicles. Each consumer can participate in different demand response events promoted by system
operators or aggregation entities [27]. Three types of demand are
considered at the hub’s output: Electrical demand associated with
lighting and other home appliances, heating demand in terms of
desired hot water temperature and cooling demand to have the
desired air temperature. It is also assumed that the proposed energy
hub is equipped with smart meter, which provides all the required
information such as energy consumptions and weather conditions.
Mathematical models of energy hub components, energy balances
and efficiency constraints will be proposed in an optimization
framework. The ultimate goal is to investigate the impact of incorporating both DR programs, PHEV and TES on cost reduction and
energy saving strategies.
67
hub is equipped with TES to achieve the most cost efficient operation. It has been shown that CCHP with an absorption chiller cannot
save as much energy in comparison with separate production system. This is due to the fact that the Coefficient Of Performance (COP)
of absorption chiller is lower than that of electric chiller [28].
To improve the energy efficiency of the system, the cooling
unit is considered in two parts; Absorption Chiller (AC) and Electric Chiller (EC). AC uses the waste heat and EC has a high COP;
therefore, the combination of the two contributes to energy efficiency improvement. The operational constraints of CCHP unit are
provided in the following. The relation between Pcchp and Hcchp is
modeled as follows:
t
t
Hcchp
(1a)
= Pcchp
· th
e
The fuel consumption of CCHP in m3 is given as:
t
Fcchp
=
t
Pcchp
e
·ˇ
(1b)
where ˇ is converting factor of 1 kWh to m3 natural gas. Other
operational constraints of CCHP are expressed as:
t
min
t
t
max
scchp
· Pcchp
≤ Pcchp
≤ scchp
· Pcchp
(1c)
t
min
t
t
max
scchp
· Hcchp
≤ Hcchp
≤ scchp
· Hcchp
(1d)
t
t−1 Pcchp − Pcchp
≤ rre
t
t−1 Hcchp − Hcchp
≤ rre · th
(1e)
(1f)
e
2.1.2. PV panels
Recently photovoltaic systems have become a promising energy
source for residential end-users due to their clean nature. Meanwhile, their output power strongly relies on solar irradiance
availability and tends to fluctuate dramatically depending on the
time of the day. Such intermittency can be partially covered by
performing an accurate forecasting. In this study, a deterministic
model is used and it is assumed that a forecasting module provides the output power of PV, using a time-series model to forecast
solar radiation based on historical data [29] collected over a certain period. More details on time-series technique are available in
[30]. Fig. 2 shows the historical and forecasted solar radiation using
time-series model. Subsequently, maximum output power of PV is
calculated by Eq. (2) [31]:
t
Ppt v = pv · S · I · (1 − 0.005 · (Tout
− 25))
(2)
where pv is the conversion efficiency of the solar cell array (%); S
t
is the array area (m2 ); I is the solar radiation (kW/m2 ); and Tout
is
the outdoor temperature (◦ C).
2.1. Local generators
2.2. Responsive loads
2.1.1. Combined cooling, heating and power (CCHP) system
A comprehensive model of CCHP consists of four main parts: a
power generation unit (PGU), a heat recovery unit (HRU), heating
and cooling units. As indicated in Fig. 1, fuel is fed into the PGU
to generate electricity, then the HRU uses the waste heat to meet
the heating and cooling demand. Auxiliary boiler is used whenever
the recovered heat is insufficient for the demands. The scope of
this paper is limited to a single home, where peak thermal demand
can be covered by CCHP thermal output and no auxiliary boiler
is needed. However, in the case of multiple homes/buildings, an
auxiliary boiler can be used along with CCHP unit to account for
the thermal demand. In order to enable CCHP following a hybrid
electric-thermal load [4], flexible thermal load is presented and the
Power consumption of residential sectors, represent about 30%
of total power consumption, among which 40% is associate with
washing and cooling appliances. These appliances show great
potential of performing load shifting [19]. In this regard, Responsive
loads are classified into: shiftable loads that are washer, dishwasher, dryer, iron and pool pump, curtailable load that is the
lighting system and flexible heating and cooling demand.
Having known the accurate anticipated load along with price
tariffs, provided by the embedded smart meter, customers can
effectively participate in DR programs by a well-timed respond
to energy and price signal changes. In this study, customers can
participate in DR programs by means of:
68
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
Fig. 1. The proposed residential energy hub model.
Fig. 2. Historical and forecasted solar radiation using time series model [15].
2.2.1. Shiftable loads
Five appliances with lower priority are considered for participating in load shifting program. These appliances can be readily
reschedule based on price signal variations, without causing significant discomfort for customers. This study considers a time window
for the operation of each shiftable appliance, which is defined
according to the customer preferences. The operational constraints
of these appliances are expressed as:
uti − dit = sit − sit−1 ,
uti + dit ≤ 1,
sit
= 0,
∀t ∈ [˛i , ˇi ], ∀i ∈ A, uti , dit , sit ∈ {0, 1}
∀t ∈ [˛i , ˇi ], ∀i ∈ A
(3a)
(3b)
t∈Ti
uki ≤ sit ,
∀t ∈ [˛i , ˇi ], ∀i ∈ A
(3e)
k=t−MUTi +1
In this model, Eq. (3c) is considered to ensure that appliances
only operate within their preferred time window. Eqs. (3d) and
(3e) represent total must run hours of each appliance in a day and
minimum required time to finish a task, respectively.
Unmistakably, there should be a logical sequence between the
operations of some devices, as there is between washer and dryer
operation. The dryer should start to run right after the washer has
completed its task or with a maximum allowed time of 1 h gap:
t−MRHwm
∀t ∈ T − [˛i , ˇi ], ∀i ∈ A
(3c)
sit = MRHi ,
t
sdr
≤
k
swm
(3f)
k=t−1
t∈
/Ti
t
∀t ∈ [˛i , ˇi ], ∀i ∈ A
(3d)
2.2.2. Curtailable load
The lighting system of the house is considered as a curtailable
load, due to the fact that it can be dimmed to a predefined level
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
during high electricity price periods. According to mathematical
model for the indoor lighting, presented in [32], the lighting load
of a house is modeled using a new index, named illumination level.
In this paper, the illumination level index is used with little modification focused on its price elasticity. According to Eq. (4), required
illumination can be satisfied through the house lighting system and
outdoor illumination, considering customers tendency to reduce
their lighting demand by up to 20% during peak hours:
t
t
t
ILin
+ ILout
≥ (1 − 0.2t ) · ILreq
(4)
where t , 0 ≤ t ≤ 1, is a linear function of electricity price, being
equal to 1 during peak hours and 0 during off-peak hours. Note that,
t is set to 0 for all hours, when load curtailing is not considered.
The required and outside illumination level are assumed to be in
per unit and normalized data can be found in [33].
2.2.3. Flexible thermal loads
In this paper, both heating and cooling demands are modeled
within the context of desired hot water and air temperature based
on the model described in [17], with little modification to supply
cooling demand. The dynamic of water draw-off flow is not taken
into consideration, since the focus of this paper is on thermal and
electrical energy management of system components. The water
storage is assumed to be always full and the consumed hot water
is substituted with the same volume of cold water in each time
interval. The water temperature is calculated in Eq. (5a):
t+1
Tws
=
[Vtcold
t )+V
· (Tcw − Tws
V
t ]
· Tws
+
t
Hws
V · Cw
(5a)
Eq. (5a) relates the water storage temperature to equilibrium
temperature of entering cold water and remaining hot water, and
to the injected thermal power by CCHP in order to heat the water
to a desired level. According to second law of thermodynamics, the
natural flow of energy is from hot to cold area, that is, for a summer
day, the transferred heat through the building materials in each
time interval is:
Qt =
1 t
t
Tout − Tin
R
(5b)
In the above equation, R is the thermal resistance of the house
shell. The required thermal energy to set the home temperature at
t is calculated from:
time t, Hair
Cair
t
dTin
dt
t
= Q t − Hair
(5c)
Where Cair is the air specific heat (kWh/◦ C). By substituting Eq.
(5b) in Eq. (5c):
t
dTin
−1
1
1
t
t
t
=
· Tin
+
· −Hair
+ · Tout
R.Cair
Cair
R
dt
t
t
t
= Tin
· e−1/R·Cair + (−R · Hair
+ Tout
) · (1 − e−1/R·Cair )
(5e)
min
t
max
Tws
≤ Tws
≤ Tws
(5f)
min
Tin
(5g)
≤
≤
max
Tin
1
max
· Hin
· utin
in
(6a)
t
max
0 ≤ Hdr
≤ Hdr
· utdr · dr
(6b)
t
0 ≤ Hin
≤
utin
+ utdr
≤ 1;
t+1
t
Qtes
= Qtes
+
∀t
utin , utdr
t
Hin
· in −
∈ {0, 1}
t
Hdr
(6c)
(6d)
dr
tes
t
tes
Qmin
≤ Qtes
≤ Qmax
(6e)
2.2.5. Active load (PHEV)
PHEV can play a role as an active load. This kind of load can be
charged in off-peak hours with low electricity prices and sell the
stored energy to the grid during peak hours. The following constraints should be considered in order to optimally manage the
charging and discharging procedures. Eqs. (7a)–(7c) present PHEV’s
battery energy balance and charger constraint, respectively. The
stored energy in the battery is considered jointly with the energy
remaining from the previous period and the charging or discharging
in the period t.
t+1
t
t
Ephe
v = Ephev + G2V · Pch,phev · t −
1
t
· Pdch,phe
v · t
V 2G
(7a)
t
max
t
Pch,phe
v ≤ Pcharger · chphev
(7b)
t
max
t
Pdch,phe
v ≤ Pcharger · dchphev
(7c)
where G2V and V2G are the PHEV’s battery charging and dismax
is the maximum
charging efficiencies, respectively and Pcharger
charging/discharging power of the charger.
The upper and lower limitations of SOC and battery charging/discharging constraints are expressed by:
t
SOCmin ≤ SOCphe
v ≤ SOCmax
(7d)
t
t
Pch,phe
v · G2V · t ≤ Cap − Ephev
(7e)
t
Pdch,phe
v·
1
t
· t ≤ Ephe
v
V 2G
(7f)
Eqs. (7e) and (7f) ensure that the energy charged to the battery
and the energy discharged from it are lower than the empty capacity and the energy content of battery, respectively. Simultaneous
charging/discharging is restricted using Eq. (7g):
chtphev + dchtphev ≤ 1;
∀tchtphev , dchtphev ∈ {0,1}
(7g)
3. Smart decision maker
In case of flexible thermal load modeling, a maximum temperature deviation from desired set points, that costumers are willing
to accept, is considered rather than applying a fixed temperature.
Therefore, the following two constraints should be considered:
t
Tin
Simultaneous injecting and drawing thermal energy is restrained
by Eq. (6c). TES energy content and its minimum and maximum
limitations are expressed by (6d) and (6e), respectively.
(5d)
As it can be seen, Eq. (5d) is in form of ẋ = −ax + bu and its
discrete model with time interval of one hour (t) will result in:
t+1
Tin
69
2.2.4. Thermal energy storage (TES)
In order to model TES, the following sets of constraints are used,
in which Eqs. (6a) and (6b) ensure that injecting and drawing heat
from TES in each time interval are less than a maximum value.
Smart home automation system (SHAS) and smart home EMS
have been recently developed in many studies [34–36]. This system
helps consumers to track their energy consumption patterns as well
as their local energy production. Fundamentally, SHAS is composed
of a Smart Decision Making core (SDM), a set of smart appliances, a
smart metering system and a communication link in order to facilitate two-way communication between system components. As
depicted in Fig. 3, smart meter provides energy consumption patterns and energy production profiles of local generators and sends
them to a data base manager. The forecasting module receives
required data for forecasting PV production, temperature and load
and sends its output back to the data base manager. The collected
data in data base along with the energy price signals and appliance
parameters are analyzed by SDM to generate the optimal dispatch
of each system components.
70
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
Fig. 3. The proposed smart decision maker.
3.1. Demand–supply balance
t represents the purchased power
Note that a positive value of Pnet
from the network and a negative value represents the sold power
to the network.
The smart decision maker is the main part of the energy
hub management system. It helps minimizing the energy cost by
optimally regulating household appliances, generation units and
charging/discharging of electric vehicle. In addition to mathematical model of system components, energy and power balances have
to be included in the optimization.
3.1.2. Thermal balances
The heat balance after HRU is expressed as:
3.1.1. Electrical power balances
(1) No DR:
The energy balance equation for cooling demand is given as
follows:
t
Pnet
=
−
t
Pch,phe
v
t
DEtotal
· G2V −
+ (ILtin
t
Pdch,phe
v
V 2G
t
+ Pcchp
t
Pch,phe
v · G2V −
t
Pdch,phe
v
V 2G
N
t
= Dcrit
+ (ILtin · Pli ) + (
i=1
t
Hdr
dr
t
t
= Hws
+ HAC
(8c)
(8d)
3.2. Objective functions
(8a)
(2) Considering DR:
t
Pnet
−
t
Hin
· in −
t
t
t
(HAC
· COPAC ) + (EEC
· COPEC ) = Hair
+ Ppt v
t
· Pli ) + PEC
t
Hcchp
+
t
+ Pcchp
+ Ppt v
3.2.1. Energy cost
The end-user’s total energy cost is composed of the electrical
power which is imported from the network, revenue from selling
power to the network as well as natural gas consumption of CCHP
over the scheduling horizon (24 h).
OF cos t =
24
t
t
t
[Pnet
· TOU
+ Fcchp
· NG ]
(9a)
t=1
t
sit · Pi ) + PEC
(8b)
3.2.2. Multi-objective optimization
Almost 30% of total green house gas (GHG) emissions is expected
from power generation sectors [37]. Fossil-fuel based generators,
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
mainly gas- and coal-fired units, are mostly responsible for CO2 ,
NOx and SOx emissions. Power generators have different operational cost and emission rates based on their type [24]; for this
reason, emission is considered as a subsidiary objective function
and a multi-objective (MO) optimization is carried out to find the
best solution between these two conflicting objective functions.
The subsidiary objective and the function for MO optimization are
presented as follows:
OF Emission =
24
t
NOx
SOx
2
[Pnet
· (CO
net + net + net )
t=1
t
+ Fcchp
2 + NOx + SOx )]
· (CO
f
f
f
(9b)
min(OF Cost , OF Emission )
(9c)
Emission objective function consists of CO2 , NOx and SOx emissions from network electricity and natural gas consumption.
Generally, MO optimization is used to find an optimal solution
between two or more objectives with opposing behavior; that is,
an increase in one objective, results in a decrease in another one.
Several MO optimization methods are introduced such as normal
boundary intersection (NBI) [38], goal attainment [39], imperialist
competitive algorithm [40] and epsilon constraint method [24]. In
this paper, the epsilon constraint method is used to minimize both
energy cost and emission in Eq. (9c). Subsequently, the best solution
is determined on Pareto front set by using fuzzy method explained
in [41].
4. Simulation and results
The proposed model is conducted on a residential energy hub,
over a daily time horizon with time interval of one hour and is
illustrated in Fig. 1. Fig. 4 shows total electrical demand and the
contribution from each appliance which are obtained by means
of embedded smart meter in the house. It is worth noting that
consumption patterns of appliances in Fig. 4 is resulted from customer comfort maximization, in which all the house appliances are
switched on/off according to customer preferences. A CCHP with
electrical output of 3 kW and a PHEV with the capacity of 6.68 kWh
are considered. The arrival and departure times of PHEV is assumed
to be 8:00 a.m. and 4:00 p.m., respectively. The batteries used in
PHEVs are supposed to operate over a large SOC window. After an
“overnight charge” of the batteries from the utility grid to a certain high SOC, the vehicle will operate in a charge-depleting or an
electric-only mode until a low state of charge is reached, which
is when the internal combustion engine will provide the required
propulsion. As a result, the arriving SOC after a daily trip is assumed
to be greater or equal to the minimum SOC.
In this paper, five shiftable loads are considered and their operational parameters are listed in Table 1. Other technical parameters
of system components and TOU price tariffs are given in [17],
Tables 2 and 3, respectively. The proposed model is solved using
Mixed Integer Linear Programming (MILP) by CPLEX solver under
GAMS on a Pentium IV, 2.6 GHz processor with 4 GB of RAM.
Table 1
Appliance operational parameters.
Shiftable loads
Rated power
(kW)
MRHi
(hour)
MUTi
(hour)
Time window
Washer
Dish washer
Dryer
Pool pump
Iron
0.5
0.7
1.1
0.7
1.3
2
2
1
10
1
2
2
1
2
1
15–23
20–24
15–23
9–24
18–24
71
Table 2
Parameters assumption.
Parameter
Value
Unit
ˇ
pv
S
des
max
min
Tin
, Tin
, Tin
des
max
min
Tws
, Tws
, Tws
max
max
, Hdr
Hin
max
min
Qtes
, Qtes
in , dr
COPAC , COPEC
Pli
NOx
SOx
2
CO
net , net , net
2 , NOx , SOx
CO
f
f
f
0.0925
15.7
20
22,21,25
70,60,80
1.5
2.25,7.5
98,98
0.7,3
0.15
968,0.5,2.1
220, 0.019, 2.62 × 10–4
(m3 /h)
(%)
(m2 )
(◦ C)
(◦ C)
(kWh)
(kWh)
(%)
–
(kW)
(g/kWh)
(g/m3 )
Table 3
TOU pricing for summer weekdays [42].
Time
TOU
Price (cents/kWh)
8:00 a.m. to 11:00 a.m.
11:00 a.m. to 5:00 p.m.
6:00 p.m. to 10:00 p.m.
10:00 p.m. to 7:00 a.m.
Mid-peak
On-peak
Mid-peak
Off-peak
11.4
12.7
11.4
7.8
4.1. Case studies
In order to evaluate the proposed model, three case studies are
considered according to Table 4. The simulation results are then
compared and discussed. To put it more precisely, in the first case
study no DR programs are considered. This is the current status
of power system, where many infrastructures such as AMI system and V2G technology have not been established properly. PHEV
owners prefer to connect their vehicle to the grid for charging rather
than participation in V2G services. This is mainly due to battery
degradation concerns. In case 2, the house is equipped with smart
meter to collect customers’ real-time data. Therefore, the basic
requirement for conducting DR program by determined responsive loads are enabled. It is likely that V2G-enabled PHEVs become
more profitable in the near future due to recent battery technology
development. In case 3, a controllable bi-directional energy flow
between the vehicle and utility grid is considered using V2G technology along with DR program and TES to better manage thermal
energy.
4.2. Energy cost optimization in three case studies
In order to evaluate the effect of incorporating DR programs
including load shifting, load curtailing and flexible thermal load
modeling on CCHP operation, the electrical output of CCHP in cases
1 and 2 is shown in Fig. 5. It is seen that in case 1, CCHP mostly
follows the thermal demand as it should maintain the water and
air temperatures at a fix level. While, after applying a wider range
of temperature for both water and air, CCHP is able to generate
more electricity especially in high price hours (i.e. 12th–16th). In
this case, the excess energy is sold to the network and consequently
more profit is gained.
Fig. 6 illustrates the electrical demand profile of aggregate
shiftable, curtailable and critical loads in cases 1 and 2 before and
after conducting load management. It can be said that the load profile is modified in such a way that results in much more smooth
load profile by shaving the peaks. It is observed that after applying DR program, the electricity peak load has been reduced from
4.593 kWh to 3.449 kWh, i.e. 25% reduction. Moreover, during the
high price hours (i.e. 13th–17th) shiftable loads are rescheduled to
operate at late night when the market prices are the least.
72
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
Fig. 4. Total and segregated electrical demands.
Table 4
Summary of case studies.
Case studies
Case 1
Case 2
Case 3
Dumb charging
√
√
DR program
Load shifting
Load curtailing
Flexible thermal load
√
√
√
√
√
√
Fig. 5. Electrical output of CCHP in cases 1 and 2.
Fig. 6. Total electrical demand before and after DR implement.
TES
Smart PHEV management
√
√
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
73
Table 5
Summary results of three case studies.
Case studies
Case 1
Case 2
Case 3
Energy cost (¢/day)
197.019
157.310
112.597
CCHP production (kW)
11.922
14.575
17.697
Network
Revenue (¢/day)
Purchased power (kW)
Sold power (kW)
24.272
25.514
23.989
9.132
12.723
16.982
Fig. 7. Optimal dispatch of electric resources for case 3.
Fig. 8. Optimal dispatch of thermal resources for case 3.
Fig. 9. Indoor and water temperature for case 3.
102.709
160.637
210.000
74
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
Table 5 summarizes the case studies performance in different
criteria. Conducting DR programs, using V2G option and TES technology leads to more efficient operation of CCHP which is directly
reflected in higher production rate in case 3 in comparison with
cases 1 and 2. Furthermore, less purchased power and more sold
power is observed in case 3. Consequently, daily saving ratio of
case 3 in comparison with cases 1 and 2 are calculated more than
40% and 28%; respectively. This can bring significant benefit in a
monthly time period.
Figs. 7 and 8 demonstrate the optimal dispatch of electric and
thermal resources in case 3. As expected, CCHP has the most contribution to the system due to PHEV’s smart management and TES
presence. During the 3th–7th hours the CCHP is shut down and
its output is covered by purchasing electricity from the network
due to low price market. While, in high price hours (i.e. 12th–17th)
considerable amount of power is sold back to the network. In the
afternoon, when solar radiation is at its highest level, PV panels are mainly responsible for satisfying the electrical demand
and CCHP unit has less contribution, except for those hours with
thermal demand. When PHEV is available in the garage, charging/discharging decision of PHEV’s battery is made with respect
to TOU price tariffs to achieve the most economic operation.
By installing TES and applying flexible thermal load modeling in case 3 the surplus recovered heat from PGU can be either
stored in TES during low thermal demand periods or used for
pre-heating/cooling of the water/air as depicted in Figs. 8 and 9.
Moreover, at the 10th, 15th, 16th and 18th hour; the thermal
demand is satisfied by drawing heat from TES, since CCHP is shut
down.
According to Fig. 9, both indoor and water temperature try to
track the desired user defined set points for most of the hours,
except for high price and peak-load hours while CCHP is almost
operating at its maximum capacity. In those hours, wider temperature deviation can be observed. Increased water temperature,
decreased air temperature and injecting heat to TES are observed
during high electrical demand or when the house owner has the
opportunity of gaining profit by selling electricity to the upstream
grid.
The PHEV’s battery should be fully charged at the departure
time. Therefore, it should be immediately plugged in when arriving
home in the evening in order to be ready for the next day’s drive. It
can be seen in Fig. 10 that the battery is charged during mid-night
hours when electricity price and network load is the least. Since the
PHEV is mostly in travel and uses up its energy out of the house, it
is more often in charging mode than in discharging one.
4.3. Multi-objective optimization for case 3
Table 6 presents energy cost and MO optimization performances with respect to their power exchange with the network. The
amount of power purchased from the network in MO optimization is 15.962 kW, which is remarkably less than that of energy
cost optimization (i.e. 23.982 kW). This difference is attributed to
higher emission rates of grid electricity compared with natural gas
emission rate. Moreover, this will lead to selling less electricity to
the network, particularly in high price hours, which is reflected in
higher energy cost (Fig. 11).
Table 6
Power exchange with network for two optimizations.
Optimization
Purchased power
from network (kW)
Sold power to
network (kW)
Energy cost
Multi-objective
23.982
15.962
16.982
9.049
Fig. 10. PHEV’s battery state of charge.
Fig. 11. Pareto front set of energy cost and emission.
Fig. 11 illustrates the Pareto front set of cost and emission minimization using epsilon constraint method. The lowest-right point
in Pareto curve is the optimal solution of energy cost objective function which has the least energy cost. As shown in Fig. 11, the best
solution is almost obtained at the knee point of the curve with the
energy cost and emission of 122.4 ¢ and 16.7 kg, respectively.
5. Conclusion
Integrating renewable energy resources in the power networks
poses new challenges associated with their volatile nature; accordingly, various forms of energy carriers (i.e. electricity, natural gas,
etc.) are considered within the energy hub, offering a certain degree
of freedom in satisfying the loads. In this paper, an optimal thermal
and electrical energy management has been developed for a typical residential energy hub. Mathematical models of hub’s energy
generator, i.e., the CCHP unit, household shiftable appliances such
as washer, dryer, dishwasher, iron, pool pump, and the lighting
system along with PHEV as an active load and TES are presented.
Based on the proposed architecture, an optimization problem has
been formulated and solved for three different case studies with the
objective function of minimizing total energy cost while considering customer preferences in terms of desired operation time for
appliances and water and indoor temperature. Simulation results
evident that case 1 has the highest energy cost and the least CCHP
F. Brahman et al. / Energy and Buildings 90 (2015) 65–75
contribution due to its inflexible thermal load. Whereas, applying
an inclusive DR program in case 2 prevents drastic peak demand
during on-peak hours. Finally, case 3 results in the most efficient
operation by integrating both V2G capability and TES as a mean
to hinder energy spillage. It also enables more CCHP production
which leads to cost savings of up to 40% and 28% compared with
cases 1 and 2, respectively. Moreover, a multi-objective optimization has been conducted to consider customer’s contribution to
CO2 , NOx and SOx emissions along with energy cost which leads
to less power exchange with the network due to high emission rate
of grid electricity.
References
[1] D.W. Wu, R.Z. Wang, Combined cooling, heating and power: a review, Prog.
Energy Combust. Sci. 32 (2006) 459–495.
[2] W. Gu, Z. Wu, R. Bo, W. Liu, G. Zhou, W. Chen, et al., Modeling, planning and optimal energy management of combined cooling, heating and power microgrid:
a review, Int. J. Electr. Power Energy Syst. 54 (2014) 26–37.
[3] F. Fang, Q.H. Wang, Y. Shi, A novel optimal operational strategy for the CCHP
system based on two operating modes, Power Syst. IEEE Trans. 27 (2012)
1032–1041.
[4] P.J. Mago, L.M. Chamra, Analysis and optimization of CCHP systems based on
energy, economical, and environmental considerations, Energy Build. 41 (2009)
1099–1106.
[5] P.J. Mago, N. Fumo, L.M. Chamra, Performance analysis of CCHP and CHP systems operating following the thermal and electric load, Int. J. Energy Res. 33
(2009) 852–864.
[6] H. Ren, W. Gao, Economic and environmental evaluation of micro CHP systems
with different operating modes for residential buildings in Japan, Energy Build.
42 (2010) 853–861.
[7] M.C. Bozchalui, R. Sharma, Optimal operation of commercial building microgrids using multi-objective optimization to achieve emissions and efficiency
targets, in: Power Energy Society General Meeting 2012 IEEE, 2012, pp. 1–8.
[8] M. Geidl, G. Koeppel, P. Favre-Perrod, B. Klockl, G. Andersson, K. Frohlich, Energy
hubs for the future, Power Energy Mag. IEEE 5 (2007) 24–30.
[9] M. Geidl, Integrated Modeling and Optimization of Multi-carrier Energy Systems, Power Systems Laboratory, ETH Zurich, 2007.
[10] S.W. Hadley, A.A. Tsvetkova, Potential impacts of plug-in hybrid electric vehicles on regional power generation, Electr. J. 22 (2009) 56–68.
[11] A. El-Zonkoly, Intelligent energy management of optimally located renewable energy systems incorporating PHEV, Energy Convers. Manage. 84 (2014)
427–435.
[12] Z. Wang, L. Wang, A.I. Dounis, R. Yang, Integration of plug-in hybrid electric
vehicles into energy and comfort management for smart building, Energy Build.
47 (2012) 260–266.
[13] M. Honarmand, A. Zakariazadeh, S. Jadid, Optimal scheduling of electric vehicles
in an intelligent parking lot considering vehicle-to-grid concept and battery
condition, Energy 65 (2014) 572–579.
[14] M. Honarmand, A. Zakariazadeh, S. Jadid, Self-scheduling of electric vehicles in
an intelligent parking lot using stochastic optimization, J. Franklin Inst. (2014),
http://dx.doi.org/10.1016/j.jfranklin.2014.01.019.
[15] M. Honarmand, A. Zakariazadeh, S. Jadid, Integrated scheduling of renewable
generation and electric vehicles parking lot in a smart microgrid, Energy Convers. Manage. 86 (2014) 745–755.
[16] C. Pang, P. Dutta, M. Kezunovic, BEVs/PHEVs as dispersed energy storage for
V2B uses in the smart grid, Smart Grid. IEEE Trans. 3 (2012) 473–482.
[17] M. Tasdighi, H. Ghasemi, A. Rahimi-Kian, Residential microgrid scheduling
based on smart meters data and temperature dependent thermal load modeling, Smart Grid. IEEE Trans. 5 (2014) 349–357.
[18] C. Chen, S. Duan, T. Cai, B. Liu, G. Hu, Smart energy management system for optimal microgrid economic operation, Renew. Power Gener. IET 5 (2011) 258–267,
http://dx.doi.org/10.1049/iet-rpg.2010.0052.
75
[19] M.C. Vlot, J.D. Knigge, J.G. Slootweg, Economical regulation power through
load shifting with smart energy appliances, Smart Grid. IEEE Trans. 4 (2013)
1705–1712.
[20] U.S. Department of Energy, Benefits of Demand Response in Electricity Markets and Recommendations for Achieving Them, Report to
U.S. Congress, February 2006, Available from: http://www.oe.energy.gov/
DocumentsandMedia/congress 1252d.pdf [online].
[21] P. Siano, Demand response and smart grids – a survey, Renew. Sustain. Energy
Rev. 30 (2014) 461–478.
[22] A. Pina, C. Silva, P. Ferrão, The impact of demand side management strategies
in the penetration of renewable electricity, Energy 41 (2012) 128–137.
[23] A. Zakariazadeh, S. Jadid, P. Siano, Stochastic multi-objective operational planning of smart distribution systems considering demand response programs,
Electr. Power Syst. Res. 111 (2014) 156–168.
[24] H. Falsafi, A. Zakariazadeh, S. Jadid, The role of demand response in single and
multi-objective wind-thermal generation scheduling: a stochastic programming, Energy 64 (2014) 853–867.
[25] F. Kienzle, P. Ahčin, G. Andersson, Valuing investments in multi-energy conversion, storage, and demand-side management systems under uncertainty,
Sustain. Energy IEEE Trans. 2 (2011) 194–202.
[26] V. Badescu, Model of a solar-assisted heat-pump system for space heating integrating a thermal energy storage unit, Energy Build. 34 (2002)
715–726.
[27] F. Fernandes, H. Morais, Z. Vale, C. Ramos, Dynamic load management in a
smart home to participate in demand response events, Energy Build. 82 (2014)
592–606.
[28] J. Wang, Z. John Zhai, Y. Jing, C. Zhang, Particle swarm optimization for redundant building cooling heating and power system, Appl. Energy 87 (2010)
3668–3679.
[29] Iran’s Meteorological Organization, Historical wind speed and solar radiation
data, 2014, http://www.weather.ir (accessed 10.08.14).
[30] J.W. Taylor, P.E. McSharry, R. Buizza, Wind power density forecasting using
ensemble predictions and time series models, Energy Convers. IEEE Trans. 24
(2009) 775–782.
[31] A. Yona, T. Senjyu, T. Funabashi, Application of recurrent neural network to
short-term-ahead generating power forecasting for photovoltaic system, in:
Power Engineering Society General Meeting 2007, IEEE, 2007, pp. 1–6.
[32] M.C. Bozchalui, S.A. Hashmi, H. Hassen, C.A. Canizares, K. Bhattacharya, Optimal
operation of residential energy hubs in smart grids, Smart Grid. IEEE Trans. 3
(2012) 1755–1766.
[33] UW Weather Station, Data Archives – Incoming Shortwave Radiations, 2009,
Available from: http://weather.uwaterloo.ca/data.html#select day [online].
[34] D.-M. Han, J.-H. Lim, Smart home energy management system using
IEEE 802.15.4 and zigbee, Consum. Electron. IEEE Trans. 56 (2010)
1403–1410.
[35] J. Han, H. Lee, K.-R. Park, Remote-controllable and energy-saving room architecture based on ZigBee communication, in: Consum. Electron. 2009. ICCE ‘09,
Dig. Tech. Pap. Int. Conf., 2009, pp. 1–2.
[36] C.-H. Lien, Y.-W. Bai, M.-B. Lin, Remote-controllable power outlet system for
home power management, Consum. Electron. IEEE Trans. 53 (2007) 1634–1641.
[37] United States Environmental protection of Agency, Available from:
http://www.epa.gov/climatechange/ghgemissions/sources/electricity.html
[online].
[38] A. Zangeneh, S. Jadid, A. Rahimi-Kian, Uncertainty based distributed generation
expansion planning in electricity markets, Electr. Eng. 91 (2010) 369–382.
[39] F. Gembicki, Y.Y. Haimes, Approach to performance and sensitivity multiobjective optimization: the goal attainment method, Automat. Control. IEEE Trans.
20 (1975) 769–771.
[40] O. Abednia, N. Amjady, K. Kiani, H. Shayanfar, A. Ghasemi, A multiobjective environmental and economic dispatch using imperialist competitive algorithm, Int.
J. Tech. Phys. Probl. Eng. 4 (2012) 63–70.
[41] G.D. Yalcin, N. Erginel, Determining weights in multi-objective linear programming under fuzziness, in: Proceedigs of WCE, 2011, p. 2.
[42] Electricity prices in Ontario. Available from: http://oeb.gov.on.ca/OEB/For+
Consumers/Underestanding+Your+Bill+Rates+and+Prices/Electricity+in+
ontario [online].
ID
262605
Title
Optimalelectricalandthermalenergymanagementofaresidentialenergyhub,integratingdemandresponse
andenergystoragesystem
http://fulltext.study/journal/191
http://FullText.Study
Pages
11