Journal of Energy Storage 100 (2024) 113526

Contents lists available at ScienceDirect

Journal of Energy Storage

journal homepage: www.elsevier.com/locate/est

Research Papers

Economy and energy flexibility optimization of the photovoltaic heat pump

system with thermal energy storage
Benjia Li a , Zhongbing Liu a,d,* , Yue Zheng b, Huahui Xie c , Ling Zhang a,d
a
College of Civil Engineering, Hunan University, Changsha 410082, China
b
ClusterTech Limited, No. 8 Science Park West Avenue, Hong Kong Science Park, N.T., Hong Kong
c
Department of Civil and Architecture Engineering, Wuyi University, Jiangmen, Guangdong Province 529020, China
d
Key Laboratory of Building Safety and Energy Efficiency of the Ministry of Education, Hunan University, Changsha 410082, China

A R T I C L E I N F O A B S T R A C T

Keywords: Photovoltaic systems have been widely applied in building energy systems. However, with the sharp increase in
Thermal energy storage demand of electricity in buildings during heating and cooling seasons, the uncertainty of photovoltaic generation
PV self-consumption further exacerbates the peak-valley difference in electricity use. To improve the economy and energy flexibility
Total annual cost
of buildings in hot summer and cold winter zones of China, a rule-based operation strategy was proposed for the
Optimal size
Rule-based operation strategy
photovoltaic heat pump with thermal energy storage system to optimize the size of the thermal energy storage
system and system operation, with aims of minimizing the total annual cost of the system and maximizing the
self-consumption rate of the photovoltaic generation. The effects of grid export power limits, grid import power
limits, feed-in tariffs, and PV generation on the system operation optimization results were also investigated. The
results indicated that by integrating the thermal energy storage system into the photovoltaic heat pump system,
the self-consumption rate of the photovoltaic generation was reduced by 2.39 %, the total annual cost of the
system was decreased by 6.61 %, and the payback period of the thermal energy storage system was 1.31 years.
The optimum size of the thermal energy storage system and the self-consumption rate of the photovoltaic
generation decreased with the increasing grid export power limits and grid import power limits. In contrast,
higher grid export power limits, grid import power limits, feed-in tariffs, and PV generation all resulted in a
lower total annual cost. This study provided a systematic design and operation optimization method for building
energy systems.

randomness of PV generation, there will be a mismatch between energy

supply and demand during the operation of the PV-HP system [12]. The
1. Introduction addition of energy storage systems [13], including electrical energy
storage [1,14,15] and thermal energy storage systems [16], to the PV-
Building energy consumption accounts for about 33 % of global HP system can effectively improve the building energy flexibility and
energy consumption [1], and building carbon emissions exceed 28 % of realize the real-time balance between PV generation and building
global carbon emissions [2]. To cut down building energy consumption electricity consumption [17]. Nevertheless, owing to the low lifespan,
and carbon emissions, photovoltaic (PV) generation has been widely limited cycles, and expensive initial investment costs, applying electrical
used as clean energy in the building energy system (BES) [3,4]. How­ energy storage systems on a large scale in buildings is difficult [18,19].
ever, it is not easy to match PV generation with the building electricity Therefore, there are more researches for the photovoltaic heat pump
load owing to the uncertainty of PV generation, which is challenging the with thermal energy storage (PV-HP-TES) system in recent years. With
stability of the power grid [5,6]. long lifetime and low initial investment cost, the PV-HP-TES system is a
Photovoltaic heat pump (PV-HP) system integrates a variety of en­ more economical and sustainable energy system [20,21].
ergy technologies [7], which can meet demands of space heating (SH), Currently, there are several design optimization methods for the PV-
space cooling (SC), and electricity, and has the advantages of high en­ HP-TES system, including optimization methods based on parameter
ergy utilization efficiency [8,9], good economic benefits [10] and variation and simulation, the mathematical planning, and rule-based
environment-friendly [11]. However, due to the volatility and

* Corresponding author at: College of Civil Engineering, Hunan University, Changsha 410082, China.
E-mail address: [email protected] (Z. Liu).

https://doi.org/10.1016/j.est.2024.113526
Received 26 February 2024; Received in revised form 21 July 2024; Accepted 24 August 2024
Available online 31 August 2024
2352-152X/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Nomenclature IL The Module photocurrent (A)

Io The diode reverse saturation current (A)
Abbreviations γ The empirical PV curve-fitting parameter
BES Building energy system Tc The module temperature (K)
HSCW Hot summer and cold winter Rs The module series resistance (Ω)
TES Thermal energy storage q The electron charge constant (C)
TOU Time-of-use k The Boltzmann constant (J/K)
PV-HP Photovoltaic heat pump system
PV-HP-TES Photovoltaic heat pump with thermal energy storage Variables
TC Total annual cost VTES The size of TES (m3)
SCR Self-consumption rate
max
VTES The maximum volume of TES (m3)
GEP Grid export power PPV− g (t) Power from the PV system to the power grid (kW)
GIP Grid import power PPV− HP (t) Power from the PV system to the ASHP (kW)
PBP Payback period PPV− E (t) Power from the PV system to meet the building electricity
SH Space heating load (kW)
SC Space cooling PPV− L (t) Power from the PV system to drive the ASHP to meet the
MILP Mixed-integer linear programming SH and SC loads (kW)
ASHP Air-source heat pump PPV− TL (t) Power from the PV system to meet the total building
RT Retail tariff electricity load (kW)
FIT Feed-in tariff PPV− TES (t) Power from the PV system to drive the ASHP to produce
energy stored in the TES (kW)
Parameters PPV dump (t) The surplus PV power generation which is dumped (kW)
COP(t) The coefficient of performance of the ASHP Pg− L (t) Power from the power grid to drive the ASHP to meet the
Q(t) The heating/cooling capacity from the ASHP (kW) SH and SC loads (kW)
P(t) The input electricity power (kW) Pg− E (t) Power from the power grid to meet the building electricity
Tout (t) The ambient temperature (◦ C) load (kW)
Ptotal,load (t) The total building electricity load (concerning lighting, Pg− TES (t) Power from the power grid to drive the ASHP to produce
equipment, and heating/cooling loads.) (kW) energy stored in the TES (kW)
Pelec,load (t) The building electricity load (concerning lighting, Pg− HP (t) Power from the power grid to drive the ASHP (kW)
equipment loads.) (kW) QHP− TES (t) Energy stored in the TES (kW)
load(t) The SH and SC loads (kW) QHP− L (t) Energy from the ASHP to meet the SH and SC loads (kW)
PPV (t) The PV generation (kW) QTES− L (t) Energy from the TES to meet the SH and SC loads (kW)
Qmax The maximum stored energy in TES (kW) QTES (t) The net energy storage of the TES system (kW)
ηTES The storage efficiency of TES Qch (t) The energy that is charged into the TES (kW)
ηTES,c The charge efficiency of TES Qdisch (t) The energy that is discharged from the TES (kW)
ηTES,d The discharge efficiency of TES CTES The annual investment cost (CNY)
Pim The grid power import limit (kW) CO The annual operation cost (CNY)
max
Pex The grid power export limit (kW) Cm The annual maintenance cost (CNY)
max
μTES The price of unit volume of TES (m3/CNY) Cre The annual replacement cost (CNY)
i The annual interest rate Cbuy (t) The annual cost of purchasing electricity from the power
n The life cycle of the system grid (CNY)
Δt Time interval Csell (t) The annual revenue of selling electricity to the power grid
φbuy The retail electricity price (CNY) (CNY)
φsell The feed-in tariffs (CNY) Qmax
TES,c The maximum energy charged into the TES system, kW
I The current (A) Qmax
TES,d The maximum energy discharged from the TES system, kW
V The voltage (V)

control [22]. The parameter variation and simulation-based optimiza­ The optimization method based on the mathematical planning model
tion method determines the optimal size of components of the PV-HP- is widely applied, and it is primarily used to establish an optimization
TES system by varying the system parameters and re-running the model for the PV-HP-TES system and solve the optimization problems by
simulation [23]. For example, Li et al. [24] focused on the PV-HP-TES the optimization software [25–27]. For example, Fischer et al. [22]
system for residential buildings in Australia. They analyzed the elec­ developed a mixed-integer linear programming (MILP) model for the
tricity consumption and self-consumption rate (SCR) of the PV genera­ PV-HP-TES system for a residential building in Germany. And they
tion of the system under different sizes of the water tank by using optimized the capacity of the water tank in the system with the objective
TRNSYS software, and then determined the optimal water tank capacity of minimizing the total cost in the whole life of the system and analyzed
of the PV-HP-TES system. Heinz et al. [21] simulated the operation of the effects of PV power generation, heating demand, and variable tariffs
the PV-HP-TES system in the heating season with different control on the optimal size of the water tank. The results indicated that the
strategies and water tank sizes using TRNSYS for renovated residential heating demand of the building had the most significant influence on the
buildings, and they finally determined the optimal size of the water tank optimal size of the water tank. And Beck et al. [26] also presented a
with the minimum payback period. Their research showed that the PV- MILP model for the PV-HP-TES system in a residential building to
HP-TES system with optimal size of the water tank reduced grid elec­ optimize the optimal capacity of the thermal energy storage (TES) sys­
tricity consumption by 29 % and net electricity costs by 35 % compared tem and system operation with the objective of minimizing the total
to a conventional heating system. annual cost (TC) of the system, and the optimization problem was solved

2
B. Li et al. Journal of Energy Storage 100 (2024) 113526

by CPLEX software. Its results showed that the optimal capacity of the respectively, which provided theoretical guidance for system design and
TES system depended on the building heating demand. Savolainen et al. operation optimization.
[28] proposed a MILP model to optimize the optimal size of the water The paper is structured into six sections. The system model is
tank and PV system and system operation of the PV-HP-TES system for described in Section 2. The energy management strategy and optimi­
residential and office buildings, respectively. The SH and electricity load zation method are illustrated in Section 3. The study cases in this paper
of buildings was taken into account, and the optimization objective was are presented in Section 4. The results analysis and discussion are pro­
minimizing the TC of the system. The results showed that the optimized vided in Section 5, and conclusions are summarized in Section 6.
configuration of the PV system and the water tank saved the energy cost
of the building significantly, and the system cost-effectiveness of the 2. System model
office building was higher than that of the residential building. How­
ever, the optimization method based on the mathematical planning The PV-HP-TES system of an office building studied in this paper is
model has a long solution time and is not easy to be combined with the described in Fig. 1, which mainly consisted of PV arrays, the power grid,
system operation and control, which makes it difficult to be popularized the TES system, and the air-source heat pump (ASHP). The building load
and applied in practical engineering [20]. models were established in TRNSYS software. And the building load
The rule-based control optimization method is mainly used to includes the electricity load of lighting and equipment and the thermal
formulate optimization objectives and corresponding system operation load of SH and SC. The demand for SH and SC was met by an ASHP. The
strategies for the PV-HP-TES system and utilize heuristic algorithms to system was connected to the power grid to ensure that the building
solve optimization problems [29] This optimization method is solved electricity demand was covered and to avoid the shortage of the PV
quickly and can be combined with the real-time operation control of the generation. The TES system was used to store energy in the valley tariff
system [30]. Considering the SH load, SC load, and electricity load de­ period. The energy was released from the TES system to meet the de­
mands of the residential building, Efkarpidis et al. [31] studied the mand of SH and SC at the peak tariff period, which would make more PV
optimal capacity of the water tank and the system operational perfor­ generation be fed into the power grid to obtain profits at peak time. In
mance of the PV-HP-TES system under the PV self-consumption maxi­ this section, simplified models were chosen for the ASHP, TES system,
mization strategy and time-of-use (TOU) tariff arbitrage strategy, and PV generation, respectively.
respectively. The results of their study showed that the utilization of the
water tank for energy storage could improve the system performance, 2.1. ASHP model
and different operation strategies greatly affected the optimal capacity
of the water tank in the system. Salpakari et al. [32] proposed a rule- The ASHP can convert electricity power from the power grid and PV
based control strategy and carried out operation optimization of the system into thermal energy that can satisfy the demand for SH and SC.
PV-HP-TES system with the objective of maximizing the SCR of the PV The thermal energy generated by the ASHP could be directly transferred
generation, considering the SH and electricity load demands of a resi­ to the building. Alternatively, it was primarily stored in the TES system
dential building in Finland. Their results showed that this control during the valley tariff period, and then the energy in the TES system
strategy reduced the PV generation fed into the grid by 8–88 %, reducing was released to cover the SH and SC loads in peak hours. The main
the impact of PV generation on the power grid. parameters of the ASHP selected in this paper are shown in Table 1. The
In summary, the system design and operational optimization of the ASHP model is simply described by Eq. (1), where COP(t) is the coeffi­
PV-HP-TES system are comprehensive processes involving complex cient of performance of the ASHP, which is closely related to the ambient
power interactions among the PV systems, power grid, and the TES temperature [33]. The relationship between the coefficient of perfor­
system. The economy of the system not only depend on electricity retail mance of the ASHP and the ambient temperature is described by Eqs.
tariff and feed-in tariff (FIT), but also on the characteristics of the energy (2)–(3) for the heating and cooling operating conditions, respectively.
use of the buildings and the corresponding system operation control
COP(t) = Q(t)/P(t) (1)
strategies. The current research on the PV-HP-TES system mainly fo­
cuses on residential buildings, and many studies only consider the
heating conditions of the building. While less research has been carried COPh (t) = 0.0001Tout (t)2 + 0.0547Tout (t) + 2.7967 (2)
out on the PV-HP-TES system for office buildings and took the SH, SC,
and electricity loads into account. Due to the highly matching between COPc (t) = 0.0006Tout (t)2 − 0.1049Tout (t) + 6.1580 (3)
the characteristics of building energy use and PV generation in office
where Q(t) is the heating or cooling generation from the ASHP, kW; P(t)
buildings, the PV-HP-TES system has a greater potential for energy
is the input power, kW; Tout (t) is the ambient temperature, ◦ C.
saving and flexibility. Especially in the HSCW regions of China, the peak-
valley difference in building electricity consumption is further exacer­
2.2. TES system model
bated by the surge in building electricity demand during the summer air
conditioning and winter heating periods, and the instability of PV
The energy generated by the ASHP was stored in the TES system,
generation.
which was transformed from electricity power that was from the power
Therefore, to further improve the economy and energy flexibility of
grid at the valley tariff or the excess PV generation in day time. Then, the
the PV-HP-TES system in office buildings in the HSCW regions of China,
TES system was discharged at peak tariff to satisfy the demand for SH
an economic and energy flexibility optimization research of the PV-HP-
and SC. The hybrid tank model was chosen to model the TES system. The
TES system was carried out in this paper, which comprehensively took
stored energy of the TES system is related to the energy charged and
the SH, SC, and electricity load demands into account. The main con­
discharged at the current moment and the stored energy at the previous
tributions of this study are as follows:
moment, and the energy of the TES system will be dissipated with time
(1) A rule-based optimization strategy was proposed for the optimal
[28]. The energy storage balance of the TES system is described by Eq.
size of the TES system and the operation optimization of the PV-HP-TES
(4).
system, with aims of minimizing the TC of the system and maximizing
the SCR of the PV generation, improving the system economic efficiency 1
QTES (t) = ηTES • QTES (t − 1) + ηTES,c • QTES,c (t) − • QTES,d (t) (4)
and energy flexibility. ηTES,d
(2) The effects of grid export power (GEP) limits, grid import power
(GIP) limits, FITs, and PV generation on the TC of the system, the SCR of where QTES (t) is the stored energy in the TES system at t moment, kW.
the PV generation, and optimal size of the TES system were studied, ηTES is the efficiency of stored energy in the TES system, which is 0.98.

3
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 1. Scheme of the PV-HP-TES system in the study case.

storage systems need to be optimized in this part using the right oper­
Table 1
ational strategies to improve the energy efficiency and system stability
The main parameters of the ASHP.
of the BES. A rule-based optimization strategy for the optimum size of
Main parameters Unit Value the TES system and operation of the PV-HP-TES system under the TOU
Rated heating capacity kW 138 tariffs was proposed.
Rated cooling capacity kW 130
Rated electric power (h/c) kW 38.1/40.3
Rated COP (h/c) – 3.62/3.23 3.1. Energy management strategy

A suitable operation strategy can efficiently improve performance of

QTES (t − 1) is the net energy stored in the TES system at t-1 moment, kW.
the system. The detailed process of the optimization methodology is
ηTES,c and ηTES,d are the efficiency of charging and discharging, respec­
shown in Fig. 2. As presented in Fig. 2, the PV generation, the total load,
tively, their values are set as 0.95.
the SH/SC load, the electricity load, the COP of the ASHP, and the TOU
tariffs were the inputs of this optimization model. And the optimization
2.3. PV generation model variable was the size of the TES system. On the basis of TOU tariffs, there
were three operation strategies formulated according to the operation
There were 238 PV modules in the roof. The module short circuit states of not working, charging, and discharging of the TES system,
current and module open circuit voltage are 13.89 A and 37.89 V, respectively. Further, comprehensively considering the local consump­
respectively. Module current and module voltage and at max power tion of the PV generation, load transfer, and tariff arbitrage, a rule-based
point and reference conditions are 13.29 A and 31.61 V, respectively. control optimization method was proposed with the optimization ob­
The maximum power of the module is 420 W, and all the modules were jectives of minimizing the TC of the system and maximizing the SCR of
arranged on the south facing 28◦ tilted surface. The four-parameter the PV generation. And the bi-objective optimization problem was
equivalent circuit model was used as the PV generation model [34], as solved by the NSGA-II algorithm.
shown in Eq. (5). The proposed energy management strategy of the PV-HP-TES system
{ [ ] } in this paper is presented in Fig. 3. The strategy optimized the power
q interactions among the PV system, the building, the TES system, and the
I = IL − Io exp (V + IRs ) − 1 (5)
γkTc power grid, detailed descriptions of the energy management strategy are
as follows:
where I is the current, A; V is the voltage, V; IL is the module photo­
current, A; Io is the diode reverse saturation current, A. γ is the empirical (1) TES not working: as shown in Fig. 3(a), at the first shoulder tariff
PV curve-fitting parameter. Tc is the module temperature, K. Rs is the period, the TES system was set not to be discharged or charged
module series resistance, Ω. The module temperature Tc and four pa­ because of the lower electricity retail price (RT). The electricity
rameters IL ,Io , γ,and Rs of the PV module vary with ambient temperature load of equipment and lighting was firstly covered by the PV
and solar irradiation, and the relationship of each parameter between generation and then by the power grid. The SH and SC loads was
the actual operation and the standard test conditions can be found in met by the heating and cooling energy generated from the ASHP
[35]. q is the electronic charge constant, its value is 1.60217646 × that was driven by the electricity power from the power grid. The
10− 19 , C; k is the Boltzmann constant, which value is 1.3806503 × PV generation would be sold into the power grid and then be
10− 23 , J/K. transferred into thermal energy stored in the TES after covering
the loads of SH, SC, equipment, and lighting.
3. Energy management strategy and optimization methodology (2) TES discharge: as presented in Fig. 3(b), at the peak and the
second shoulder tariff period, the TES system was both set to be
To solve the intermittency and instability of the PV generation, en­ discharged. The reason why the operation strategies were iden­
ergy storage systems are applied to the PV systems. By regulating the tical at the peak and the second shoulder tariff period was that the
charging and discharging process in the energy storage systems, users energy stored in the TES system must be fully released during an
can design an appropriate operation strategy to manage the system operating cycle. For SH and SC loads, they were firstly satisfied by
autonomously. Therefore, the size and system operation of energy the TES system discharging and then supplemented by electricity

4
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 2. The proposed optimization framework.

power of the PV generation and the power grid. For the electricity system, so the annual replacement cost of the TES system was without
load of the equipment and lighting, it was firstly satisfied by the consideration. In order to convert the cost of initial investment to the
PV generation, then by the electricity power from the power grid. cost of annual investment, the capital recovery factor (CRF) [36,37]
After covering the loads of SH, SC, equipment, and lighting, the defined by Eq. (8) was introduced for discounting.
PV generation would be sold into the power grid firstly and then
Minimize TC = CTES + CO + Cm + Cre (6)
the surplus would be transferred into thermal energy stored in the
TES.
CTES = VTES × μTES × CRF (7)
(3) TES charge: During the valley tariff period, as shown in Fig. 3(c),
thanks to the lowest retail tariff, the TES was set to be charged. i(1 + i)n
The electricity power was taken from the power grid to drive the CRF = (8)
(1 + i)n − 1
ASHP to generate heating/cooling energy stored in the TES sys­
tem until the water tank was fully charged. It is beneficial to 8
∑ 760
( )
transfer SH and SC loads from peak hours to valley hours. CO = Cbuy (t) − Csell (t) (9)
Further, there was no load of SH, SC, equipment, and lighting at t=1

23:00–7:00. Therefore, the PV generation would firstly charge ( )

the TES if the water tank was not fully charged, and then the Cbuy (t) = Pg− TES (t) + Pg− E (t) + Pg− L (t) × Δt × φbuy (10)
surplus PV generation would be sold into the power grid.
Csell (t) = PPV− g (t) × Δt × φsell (11)
3.2. Objective functions
where CTES is the annual investment cost of the TES system, calculated
by Eq. (7), CNY. μTES is the price of the unit volume of the TES system,
Concerning the economy and energy flexibility optimization of the
which is 320 CNY/m3 . CRF is the capital recovery factor calculated by
PV-HP-TES system, there were two objectives set in this paper respec­
Eq. (8), and i is the annual interest rate, which is 5 %. n is the life cycle of
tively. Detailed descriptions are as follows:
the system, which is 20 years. CO is the cost of system annual operation,
CNY. Cbuy (t) is the cost of buying electricity power from the power grid
(1) The TC of the system
at moment t and is calculated by Eq. (10), CNY. Δt is the calculation step,
To improve the system economy, minimizing the TC of the system which is 1 h. φbuy is the electricity retail tariff (RT), CNY. Csell (t) is the
was set as one of the objective functions of this study. The TC includes revenue of selling electricity power to the power grid at time t, calcu­
the costs of annual investment, operation, maintenance, and replace­ lated by Eq. (11), CNY. φsell is the feed-in tariff (FIT), CNY.
ment, which is calculated as shown in Eq. (6). In this article, it was
supposed that the energy storage efficiency of the TES system remained (2) The SCR of the PV generation
constant during its service life term, so the annual maintenance cost was
not considered in this study. Furthermore, it was also hypothesized that PV generation can be directly used to satisfy the building electricity
the service life of the TES system was equivalent to the life cycle of the load or drive the operation of the ASHP to generate heating/cooling

5
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 3. Energy management strategy at different tariff periods (a) The first shoulder period (b)The peak and second shoulder period (c)The valley period.

energy stored in the TES system, which in turn improves the mismatches where PPV− E (t), PPV− L (t), and PPV− TES (t) are the power interactions be­
of PV generation and the building electricity load, enhances the self- tween the PV system with the building electricity load, the SH and SC
regulation capability of the system, and decreases the impacts of PV loads, and the TES system, respectively, kW. PPV (t) is the PV generation
generation on the power grid [8,38]. To reflect the on-site consumption from the PV system, kW.
capacity of the PV generation and improve the energy flexibility of the
building, the SCR of the PV generation was maximized in this study. It is
described as the ratio of the directly used PV generation to the total 3.3. Optimization constraints
power generation of the PV system, as shown in Eq. (12).
The optimization constraints were set considering the system power
8760
∑
(PPV− E (t) + PPV− L (t) + PPV− TES (t) )
balance, TES operation, and grid power interaction. The detailed de­
Maximize SCR = i=1
(12) scriptions are as follows:
8760
∑
PPV (t)
i=1 (1) System power balance constraints

6
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 3. (continued).

It is a vital requirement to make sure the balance between the energy (3) Grid power interaction constraints
use and supply of the BES. Therefore, system power balance constraints
were set from the energy demand and supply side, as shown in Eqs. (13)– The export power and import power limits of the power grid should
(14). Ptotal,load (t) is the total electricity of the building, kW. PPV− E (t) satisfy Eqs. (18)–(19). The export power limit Pim max is the maximum
and PPV− L (t) are the power from the PV system to cover the loads of electricity power that can be used to store energy in the TES system
equipment and lighting, SH and SC, respectively, kW. Pg− E (t) and Pg− L (t) during the valley tariff period, kW. The import power limit Pexmax is the
are the power from the power grid to meet the loads of equipment and maximum PV generation that can be fed into the power grid, kW.
lighting, SH and SC, respectively, kW. Pg− TES (t) is the power from the Pg− TES (t) is the power from the power grid to the TES system, kW.
power grid to drive the ASHP to produce energy stored in the TES, kW. PPV− g (t) is the power from the PV system to the power grid, kW.
Qdisch (t) is the energy discharged from the TES system to meet the SH
and SC loads, kW. Qch (t) is the energy charged into the TES, kW. COP(t) Pg− TES (t) ≤ Pim
max (18)
is the coefficient of performance of the ASHP.
PPV− g (t) ≤ Pex
max (19)
Qch (t)
TES charge : Ptotal,load (t) +
COP(t)
= PPV− E (t) + PPV− L (t) + Pg− E (t) + Pg− L (t) + Pg− (13) 3.4. Optimization algorithm
TES (t)

Qdisch (t) NSGA-II algorithm is a multi-objective optimization algorithm

TES discharge : Ptotal,load (t) − developed based on genetic algorithm, with strong search abilities and
COP(t)
the capability of obtaining uniformly distributed Pareto front. The
= PPV− E (t) + PPV− L (t) + Pg− E (t) + Pg− L (t) (14)
NSGA-II algorithm proceeds as follows: I. Selecting parents for the next
generation using the selection function of the current population (initial
(2) TES operation constraints sample); II. Generating children from selected parents through crossover
and mutation; III. Scoring the children by calculating their objective
The energy charged into and discharged from the TES system was function values and viability; IV. Merging the current population and the
with constraints of Eqs. (15)–(17). QTES,c (t) and QTES,d (t) are the energy children. V. Pruning the extended population to determine the popula­
charged into and discharged from the TES system, kW. Qmax max tion size. The program stopped when the number of maximum genera­
TES,c and QTES,d
are the maximum energy charged into and discharged from the TES tions is exceeded or the time limit is exceeded. During the program
max
system, kW. VTES is the maximum volume of the water tank, which was operation, the Pareto front distribution is updated once as the algorithm
evolves for one generation, continuously approximating the uniform
the volume of water tank to meet maximum load at peak tariffs period,
distribution and the Pareto optimum solutions, and finally obtaining the
m3 .
Pareto optimal solution set. And then decision-makers need to select the
0 ≤ Qch (t) ≤ Qmax
TES,c (15) most appropriate solution from them according to their demand [39].
Owing to the excellent optimization searching ability, the NSGA-II
0 ≤ Qdisch (t) ≤ Qmax
TES,d (16) algorithm was applied as the primary optimization tool to solve the
proposed bi-objective optimization problem, and the proposed optimi­
0 ≤ VTES ≤ VTES
max
(17) zation framework was shown in Fig. 2 in Section 3.1. The gamultiobj
function was used for solving the bi-objective optimization problem in
this paper. And the parameter settings of the solver were showed in

7
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Table 2 hours are identified as valley hours, which promote the consumption of
The parameter settings of the solver. renewable energy and lead users to adjust their loads. This tariff struc­
Parameters Unit Value ture can motivate users to take an active role in adjusting loads and
reducing the pressure on the power grid during the peak period [40].
Pareto fraction Dimensionless 0.6
Population size 400 There were two cases set up in this paper to compare and analyze the
Generations 200 properties of the proposed operation strategy, where the PV-HP system
Stopping algebra 200 was set as Case 1 and the reference case. Under the TOU tariff structure
Deviation of fitness function values 10− 5 in Fig. 6, the electricity load of the building was prioritized to be covered
by the PV generation. Furthermore, if the PV generation could not
Table 2. The detailed setting of optimization constraints was presented completely cover the electricity load, the electricity power was taken
in Section 3.3. from the power grid to satisfy the remaining electricity demand. If the
PV generation could totally satisfy the electricity load of the building,
3.5. Evaluation indicator the excess PV generation would be delivered into the grid. The PV-HP-
TES system proposed in this paper was set as Case 2, and its system
The economic efficiency of TES system in building is the focus of structure is shown in Fig. 1. Based on Case 1, a TES system was added to
users. In this paper, the payback period (PBP) was used to evaluate the the PV-HP system. The operational strategy for Case 2 was the energy
economics of the TES system. The PBP is the number of years required to management strategy presented in Section 3 of this paper. The elec­
compensate for the initial investment cost, defined as the ratio of the tricity power at the valley tariff was utilized to produce energy that was
initial investment cost to the reduced cost of the system after the stored in the TES system at night. Then, the energy in the TES system
configuration of water tank [6], which was defined as shown in Eq. (20). was released at the peak tariff to meet the SH and SC loads, which could
reduce the usage of electricity power from the power grid and sell more
CTES PV generation to the power grid at peak tariff.
PBP = (20)
Creduced
5. Results analysis and discussions
where, CTES is the initial investment of the water tank, CNY; Creduced is
the reduced cost, CNY. In this part, the performance of the PV-HP-TES system was evalu­
ated, and the effects of the GEP limits, GIP limits, and FITs on the system
4. Study cases operation optimization results were also analyzed.

A three-story office building in Changsha was taken as the research

5.1. Optimization results
object. The total area of the building was 1439.1 m2, and the main pa­
rameters of this building envelope are presented in Table 3. The SH and
By solving the bi-objective optimization problem of Case 2, the op­
SC loads were simulated using the multi-zone model in TRNSYS soft­
timum size of the TES system was selected based on the Pareto front
ware, and the meteorological parameters used in the simulation were
acquired after the bi-objective optimization. The performance of the PV-
obtained from the data of a typical meteorological year in Meteonorm
HP-TES system was evaluated by comparing the system operation, the
software. The SH and SC loads were met by the ASHP. The electricity
TC of the system, and the SCR of the PV generation in two cases. The
loads of lighting and equipment were calculated using the unit-area
power interactions of the TES system, the power grid, and the PV system
method, and the time schedules of lighting and equipment were
were also analyzed by analyzing the typical operation data from August
referred to the design standard of public buildings in China [6]. The
1st to August 3rd.
hourly electricity, SH, and SC loads of the building for the whole year is
shown in Fig. 4.
5.1.1. Analysis of bi-objective optimization results
The PV arrays were installed on the roof of the building, and the area
This study carried out a bi-objective optimization of the size of the
of the PV arrays was 476 m2. The hourly PV generation of the PV arrays
TES system by the NSGA-II algorithm to realize the minimization of TC
for the whole year was calculated by the PV generation model in Section
of the system and maximization of the SCR of the PV generation. The
2.3, as shown in Fig. 5. In Fig. 5, the maximum of the hourly PV gen­
Pareto front of the optimization results is presented in Fig. 7 and the
eration is close to 100 kW. As the solar radiation intensity in the summer
detailed data of the optimization results for the two cases are shown in
is higher, there is more PV generation compared to the winter situation.
Table 4. The TC and the SCR of the system corresponding to the opti­
It was assumed that the annual PV generation remained unchanged in
mized size of the TES system were normalized, and the Euclidean dis­
the life cycle of the PV-HP-TES system.
tance method was used to select the optimum size of the TES system.
In this paper, the RT and the FIT were TOU tariff and the structures of
Therefore, the point P was the optimal optimization result.
the TOU tariff are shown in Fig. 6. The TOU tariff structure is based on
As shown in Fig. 7 and Table 4, at point P, the TC of the system was
the fluctuation of electricity consumption load of users, the electricity
54,858.7 CNY, the SCR was 78.30 %, the size of the TES system was 15.9
will be divided into several periods of 24 h in a day. When the system
m3, and the PBP of the TES system was 1.31 years. Compared to Case 1,
supply and demand are nervous, and the marginal cost of power supply
Case 2 achieved building heating and cooling loads transfer at peak tariff
is high, the hours are identified as peak hours, which guide users to
period by storing energy in the TES system at the valley tariff moment
reduce consumption in peak hours. And when the system supply and
and releasing energy at peak tariff moment, contributing to a 6.61 %
demand are relaxed, and the marginal cost of power supply is low, the
reduction of the TC of the system. While more PV generation was
delivered to the grid at peak tariff period to gain revenue, which led to a
Table 3 decrease of the SCR by 2.39 %.
Information about the building envelope.
Construction type Thickness (mm) U-value (W/m2 • K− 1 ) 5.1.2. Operational analysis of typical days
External walls 240 0.767 Operational data from August 1st to August 3rd were taken for
Roof 260 0.596 operation analysis of the two study cases on typical days in the summer.
Internal walls 240 0.950 The hourly building electricity load of Case 1 and Case 2 are shown in
Floor/ceilings 260 0.896 Fig. 8 and Fig. 9, respectively. The hourly building electricity load
External windows 24 2.82
concludes the loads of SC, equipment and lighting. As shown in Fig. 8

8
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 4. Hourly electricity, SH, and SC loads of building.

Fig. 5. Hourly power generation of the PV system.

and Fig. 9, due to fewer occupants and lower lighting and equipment The hourly power flows of the TES system, the power grid, and the
usage in the office building, the hourly building electricity load of Case 1 PV system of the PV-HP-TES system in Case 2 for August 1st to August
and Case 2 at 13:00–14:00 was lower. The building hourly electricity 3rd is shown in Fig. 11. In Fig. 11(a), the power flows of the TES system
load in Case 1 was met by the electricity power from the power grid and are shown. The power grid was the primary source of energy stored in
the PV generation, and it could be fully met by the PV generation at the the TES system. The ASHP generated cooling energy by utilizing the
peak tariff period on August 2nd when the PV generation was enough. electricity power from the power grid at the valley tariff period, and the
Compared to Case 1, the TES system was charged at the valley tariff cooling energy was released from the TES system to meet the SC load at
period and discharged during the peak tariff period in Case 2 to meet the the peak tariff period. However, there was a certain amount of energy
SC load, which significantly reduced the amount of electricity power loss from the TES system in the whole process. Furthermore, owing to
from the power grid to drive the ASHP. the priority to gain revenue by feed-in of PV generation, there was no PV
The hourly building electricity load curve between the two cases can generation to drive the ASHP producing energy stored in the TES system.
be clearly compared in Fig. 10. Compared to Case 1, due to configuration The power flows of the power grid are presented in Fig. 11(b). It was
of the TES system, the building electricity load of Case 2 was lower at the obvious that the system operation cost mainly came from storing energy
peak tariff period. Further, since the optimum size of the TES system was in the TES system at the valley tariff period and satisfying building
small, the energy stored in the TES system at valley tariff period could electricity and cooling loads during the day. Since the SC load was firstly
not fully satisfy the cooling load of the building at the whole peak tariff satisfied by the TES system discharging and the building electricity load
period in Case 2. Therefore, the energy of the TES system was fully was firstly met by the PV generation, there would be a situation where
discharged in the early hours at the peak tariff period. the building electricity load was covered without taking electricity

9
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 6. The structure of the TOU tariff.

Fig. 7. The Pareto front of the optimization results.

5.2. Effects of GEP limits on the system operation optimization results

Table 4
The optimization results of the study case.
The optimization results of the system operation under various GEP
Study Total annual Self-consumption rate The size of the TES limits are presented in Fig. 12. As shown in Fig. 12, when the GEP limit
case cost (%) system (m3)
was increased from 5 kW to 10 kW, the optimum size of the TES system
(CNY)
was increased from 10.1m3 to 14.7m3, the TC of the system was reduced
Case1 58,741.7 80.69 –
by 1.7 %, and the SCR of the PV generation was decreased by 0.7 %. The
Case2 54,858.7 78.30 15.9
reason for this was that when the GEP limit was relatively low, to make
full use of the TES system at valley tariff period and reduce the cost of
power from the power grid at peak tariff period, taking the power in­ initial investment of the TES system, the optimal capacity of the TES
teractions of the power grid on August 2nd for example. system at this time was negligible. The smaller optimal volume of the
The power interactions of the PV system are shown in Fig. 11(c). It TES system meant that less energy was stored in the TES system at valley
was clearly shown that the PV generation was used to cover the building tariff period, less energy would be released from the TES system at peak
electricity load in priority, followed by driving the ASHP to operate the tariff period, and the PV generation delivered into the power grid was
cooling energy to satisfy the building cooling load, which realized the reduced, which in turn increased the TC of the system and the SCR of the
maximum SCR of the PV generation. When the hourly PV generation was PV generation. When the GEP limit was more significant than 40 kW, it
large on August 2nd, the PV generation would be sold into the power had almost no effect on the optimal volume of the TES system, the TC of
grid for profits after meeting the hourly building electricity load, ulti­ the system, and the SCR of the PV generation.
mately realizing the aim of minimizing the TC of the system.

10
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 8. The building electricity load on typical days of Case 1.

system. The PV generation exceeding the GIP limit would be trans­

formed into thermal energy stored in the TES system by driving the
ASHP operation, and the optimal volume of the TES system was
increased, which contributed to the higher SCR of the PV generation.

5.4. Effects of FITs on the system operation optimization results

The effects of different FITs on the system operation optimization

results are shown in Fig. 14. This paper compared the effects of different
FITs on the TC of the system, the SCR of the PV generation, and the
optimal size of the TES system, respectively. In Fig. 14(a), it was evident
that as the FIT increased, the TC of the system decreased, and there was a
lower increase in the optimal size of the TES system. When the online
price of electricity increased from FITs to RT, the TC of the system was
Fig. 9. The building electricity load on typical days of Case 2. decreased by 7.90 % and the optimum size of the TES system was
increased by 0.2 m3. The reason for this was the parallel increased in the
FITs for each period, which in turn increased the revenues obtained from
the PV generation fed into the power grid and then resulted in a decrease
in the TC of the system. The increasing optimal size of the TES system
could not result in a higher operating profit, leading to a lower increase
in the optimal size of the TES system. In Fig. 14(b), the SCR of the PV
generation remained almost constant with the increasing FIT. When the
price of electricity power sold into the power grid was increased from
FITs to RT, the SCR of the PV generation was reduced by only 0.03 %.
This was due to the fact that the PV generation was primarily used to
satisfy the electricity load of the building. And another reason was that
the SH and SC loads as well as the optimal size of the TES system
remained almost constant. Therefore, at the peak period, the PV gen­
eration delivered into the power grid was practically unchanged, mak­
ing the SCR of the PV generation invariable.
Fig. 10. The building electricity load curves for Case 1 and Case 2.
5.5. Effects of PV generation on the system operation optimization results
5.3. Effects of GIP limits on the system operation optimization results
The PV panel area of 60 %, 70 %, 80 %, 90 %, and 100 % of the roof
The impacts of different GIP limits on the system operation optimi­ area were set to investigate the effects of PV generation on the optimi­
zation results are shown in Fig. 13. The TC of the system and the SCR of zation results of the system operation. As clearly shown in Fig. 15, with a
the PV generation increased as the GIP limit decreased, as shown in gradual increase in PV generation, the TC of the system, the SCR of the
Fig. 13. When the GIP limit was reduced from 40 kW to 20 kW, the TC of PV generation, and the optimal size of the TES system gradually
the system was increased by 6.4 %, the SCR was increased by 0.81 %, decrease. When the PV panel area was increased from 60 % roof to 100
and the optimum size of the TES system was added by 0.5m3. This was % roof, the CT of the system was reduced from 86,560.4 CNY to
because the reduction in the GIP limit reduced the PV generation 54,858.7 CNY, the SCR of the PV generation was decreased from 92.4 %
delivered into the grid, which in turn reduced the revenue gained from to 78.3 %, and the optimal capacity of the TES system was reduced from
the PV generation fed into the power grid and increased the TC of the 32.9 m3 to 15.9 m3. This was due to the fact that the more PV generation,

11
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 11. The hourly power flows of the component systems (a) the TES system (b) the power grid (c) the PV system.

12
B. Li et al. Journal of Energy Storage 100 (2024) 113526

Fig. 12. The effects of different GEP limits on system operation optimization (a) TC (b) SCR.

Fig. 13. The effects of different GIP limits on system operation optimization (a) TC (b) SCR.

Fig. 14. The effects of different FITS on system operation optimization (a) TC (b) SCR.

Fig. 15. The effects of PV generation on system operation optimization (a) TC (b) SCR.

13
B. Li et al. Journal of Energy Storage 100 (2024) 113526

the more PV generation the building could utilize to meet the load de­ CRediT authorship contribution statement
mand during operation, which in turn reduced the electricity power
purchased from the power grid, thus reducing the TC of the system. And Benjia Li: Writing – review & editing, Writing – original draft,
as PV generation increased, the system could sell more PV generation Validation, Software, Resources, Methodology, Investigation, Formal
into the power grid at the peak tariff period, allowing the SCR of the PV analysis, Data curation, Conceptualization. Zhongbing Liu: Writing –
generation to decrease. The TES system shifted the SH and SC loads at review & editing, Validation, Supervision, Methodology, Investigation,
the peak tariff period by storing energy at the valley tariff period and Funding acquisition, Formal analysis, Conceptualization. Yue Zheng:
releasing energy at the peak tariff period. When PV generation Writing – review & editing, Software. Huahui Xie: Writing – review &
decreased, the total amount of electricity power purchased from the editing. Ling Zhang: Writing – review & editing.
power grid increased and the PV generation sold into the power grid
decreased, which increased the TC of the system. Therefore, in order to
Declaration of competing interest
minimize the TC of the system, the optimal size of the TES system was
increased to shift more SH and SC loads at the period of peak tariff.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
6. Conclusions
the work reported in this paper.
With the objectives to minimize the TC of the system and maximize
Data availability
the SCR of the PV generation, a rule-based optimization strategy for the
optimum size of the TES system and the system operation under TOU
The authors do not have permission to share data.
tariffs was proposed. And a bi-objective optimization of the PV-HP-TES
system for an office building in HSCW regions of China was carried out.
Acknowledgement
The effects of the GEP limits, GIP limits, FITs, and PV generation on the
operation optimization results of the PV-HP-TES system were investi­
This work was supported by the National Key Research & Develop­
gated in detail. The main conclusions are as follows:
ment Program of China [grant number 2022YFC3801503] and the Na­
tional Natural Science Foundation of China [grant number 52378097].
(1) The PV-HP-TES system could effectively minimize the TC of the
system and maximize the SCR of the PV generation. The ideal
optimization result was selected using the Euclidean distance References
method based on the Pareto front obtained after bi-objective
[1] B. Li, Z. Liu, Y. Wu, P. Wang, R. Liu, L. Zhang, Review on photovoltaic with battery
optimization. The optimum size of the TES system was 15.9 m3, energy storage system for power supply to buildings: challenges and opportunities,
and the PBP of the TES system was only 1.31 years. Compared to J. Energy Stor. 61 (2023) 106763.
Case 1, the SCR of the PV generation in the PV-HP-TES system [2] K. Ahmed Ali, M.I. Ahmad, Y. Yusup, Issues, impacts, and mitigations of carbon
dioxide emissions in the building sector, Sustainability 12 (18) (2020) 7427.
was decreased by 2.39 %, and the TC of the system was reduced [3] M. Hu, F. Xiao, S. Wang, Neighborhood-level coordination and negotiation
by 6.61 %. techniques for managing demand-side flexibility in residential microgrids, Renew.
(2) The GEP limits had a critical influence on the TC of the system Sust. Energ. Rev. 135 (2021) 110248.
[4] T. Han, Y. Zheng, G. Gong, Exergy analysis of building thermal load and related
and the optimum size of the TES system. When the GEP limit was energy flows in buildings, Indoor Built Environ. 26 (9) (2015) 1257–1273.
increased from 5 kW to 10 kW, the optimum size of the TES [5] X. Chen, Z. Liu, P. Wang, B. Li, R. Liu, L. Zhang, et al., Multi-objective optimization
system was raised from 10.1m3 to 14.7m3, the TC of the system of battery capacity of grid-connected PV-BESS system in hybrid building energy
sharing community considering time-of-use tariff, Appl. Energy 350 (2023)
was reduced by 1.70 %, and the SCR of the PV generation was
121727.
decreased by 0.70 %. In contrast, when the GEP limit was greater [6] J. Liu, Z. Liu, Y. Wu, X. Chen, H. Xiao, L. Zhang, Impact of climate on photovoltaic
than 40 kW, there would be no significant difference in the op­ battery energy storage system optimization, Renew. Energy 191 (2022) 625–638.
[7] A. Baniasadi, D. Habibi, W. Al-Saedi, M.A.S. Masoum, PV self-consumption
timum size of the TES system, the TC of the system, or the SCR of
enhancement with optimal residential thermal energy management, in: 2019 9th
PV generation due to the increasing GEP limit. International Conference on Power and Energy Systems (ICPES), Perth, WA,
(3) As the GIP limit decreased, the TC of the system, the SCR of the Australia, 2019, pp. 1–6.
PV generation and the optimum size of the TES system increased. [8] R. Luthander, J. Widén, D. Nilsson, J. Palm, Photovoltaic self-consumption in
buildings: a review, Appl. Energy 142 (2015) 80–94.
When the GIP limit was reduced from 40 kW to 20 kW, the TC of [9] R. Thygesen, B. Karlsson, Simulation and analysis of a solar assisted heat pump
the system was increased by 6.4 %, the SCR of the PV generation system with two different storage types for high levels of PV electricity self-
was added by 0.81 %, and the optimum size of the TES system consumption, Sol. Energy 103 (2014) 19–27.
[10] A. Arteconi, E. Ciarrocchi, Q. Pan, F. Carducci, G. Comodi, F. Polonara, et al.,
had an increase of 0.5 m3. Thermal energy storage coupled with PV panels for demand side management of
(4) With an increase in the FITs, the TC of the system decreased, the industrial building cooling loads, Appl. Energy 185 (2017) 1984–1993.
SCR of the PV generation remained almost the same, and there [11] L. Liu, W. Liu, J. Yao, T. Jia, Y. Zhao, Y. Dai, Life cycle energy, economic, and
environmental analysis for the direct-expansion photovoltaic-thermal heat pump
was a lower increase in the optimum size of the TES system. When system in China, J. Clean. Prod. 434 (2024) 139730.
the price of electricity power sold to the power grid was increased [12] F. Reda, S. Paiho, R. Pasonen, M. Helm, F. Menhart, R. Schex, et al., Comparison of
from FITs to RT, the TC of the system was reduced by 7.90 %, the solar assisted heat pump solutions for office building applications in northern
climate, Renew. Energy 147 (2020) 1392–1417.
SCR of the PV generation was decreased by 0.03 %, and the op­
[13] A. Baniasadi, D. Habibi, W. Al-Saedi, M.A.S. Masoum, C.K. Das, N. Mousavi,
timum size of the TES system was added by 0.2 m3. Optimal sizing design and operation of electrical and thermal energy storage
(5) PV generation had a significant impact on the optimization of PV- systems in smart buildings, J. Energy Stor. 28 (2020) 101186.
[14] C. Lange, A. Rueß, A. Nuß, R. Öchsner, M. März, Dimensioning battery energy
HP-TES system operation. With a reduction in the PV generation,
storage systems for peak shaving based on a real-time control algorithm, Appl.
the electricity power purchased from the power grid increased, Energy 280 (2020) 115993.
the PV generation sold into the power grid decreased, and the TC [15] A.J. Pimm, T.T. Cockerill, P.G. Taylor, The potential for peak shaving on low
of the system and the SCR of the PV generation increased. In order voltage distribution networks using electricity storage, J. Energy Stor. 16 (2018)
231–242.
to minimize the TC of the system, the optimal size of the TES [16] J. Liu, X. Yang, Z. Liu, J. Zou, Y. Wu, L. Zhang, et al., Investigation and evaluation
system was increased to transfer more SH and SC loads of the of building energy flexibility with energy storage system in hot summer and cold
building at the peak tariff period. winter zones, J. Energy Stor. 46 (2022) 103877.
[17] Y. Wu, Z. Liu, B. Li, J. Liu, L. Zhang, Energy management strategy and optimal
battery capacity for flexible PV-battery system under time-of-use tariff, Renew.
Energy 200 (2022) 558–570.

14
B. Li et al. Journal of Energy Storage 100 (2024) 113526

[18] M. Shabani, J. Mahmoudimehr, Techno-economic role of PV tracking technology in [30] M. Pinamonti, A. Prada, P. Baggio, Rule-based control strategy to increase
a hybrid PV-hydroelectric standalone power system, Appl. Energy 212 (2018) photovoltaic self-consumption of a modulating heat pump using water storages and
84–108. building mass activation, Energies 13 (23) (2020) 6282.
[19] N. Mousavi, G. Kothapalli, D. Habibi, M. Khiadani, C.K. Das, An improved [31] N.A. Efkarpidis, S.A. Vomva, G.C. Christoforidis, G.K. Papagiannis, Optimal day-to-
mathematical model for a pumped hydro storage system considering electrical, day scheduling of multiple energy assets in residential buildings equipped with
mechanical, and hydraulic losses, Appl. Energy 247 (2019) 228–236. variable-speed heat pumps, Appl. Energy 312 (2022) 118702.
[20] M.H. Abokersh, M. Vallès, K. Saikia, L.F. Cabeza, D. Boer, Techno-economic [32] J. Salpakari, P. Lund, Optimal and rule-based control strategies for energy
analysis of control strategies for heat pumps integrated into solar district heating flexibility in buildings with PV, Appl. Energy 161 (2016) 425–436.
systems, J. Energy Stor. 42 (2021) 103011. [33] J. Tarragona, C. Fernández, L.F. Cabeza, A. de Gracia, Economic evaluation of a
[21] A. Heinz, R. Rieberer, Energetic and economic analysis of a PV-assisted air-to-water hybrid heating system in different climate zones based on model predictive control,
heat pump system for renovated residential buildings with high-temperature heat Energy Convers. Manag. 221 (2020) 113205.
emission system, Appl. Energy 293 (2021) 116953. [34] J.A. Gow, C.D. Manning, Development of a model for photovoltaic arrays suitable
[22] D. Fischer, K.B. Lindberg, H. Madani, C. Wittwer, Impact of PV and variable prices for use in simulation studies of solar energy conversion systems, in: 1996 Sixth
on optimal system sizing for heat pumps and thermal storage, Energ. Buildings 128 International Conference on Power Electronics and Variable Speed Drives (Conf.
(2016) 723–733. Publ. No. 429), Nottingham, UK, 1996, pp. 69–74.
[23] E. Ampatzi, I. Knight, R. Wiltshire, The potential contribution of solar thermal [35] Q. Kou, S.A. Klein, W.A. Beckman, A method for estimating the long-term
collection and storage systems to meeting the energy requirements of north performance of direct-coupled PV pumping systems, Sol. Energy 64 (1) (1998)
European housing, Sol. Energy 91 (2013) 402–421. 33–40.
[24] Y. Li, A. Mojiri, G. Rosengarten, C. Stanley, Residential demand-side management [36] A. Maleki, A. Askarzadeh, Optimal sizing of a PV/wind/diesel system with battery
using integrated solar-powered heat pump and thermal storage, Energ. Buildings storage for electrification to an off-grid remote region: a case study of Rafsanjan,
250 (2021) 111234. Iran. Sustain. Energy Technol. Assess. 7 (2014) 147–153.
[25] M. Tostado-Véliz, D. Icaza-Alvarez, F. Jurado, A novel methodology for optimal [37] E. Shayan, V. Zare, I. Mirzaee, On the use of different gasification agents in a
sizing photovoltaic-battery systems in smart homes considering grid outages and biomass fueled SOFC by integrated gasifier: a comparative exergo-economic
demand response, Renew. Energy 170 (2021) 884–896. evaluation and optimization, Energy 171 (2019) 1126–1138.
[26] T. Beck, H. Kondziella, G. Huard, T. Bruckner, Optimal operation, configuration [38] J.C. Hernández, F. Sanchez-Sutil, F.J. Muñoz-Rodríguez, Design criteria for the
and sizing of generation and storage technologies for residential heat pump optimal sizing of a hybrid energy storage system in PV household-prosumers to
systems in the spotlight of self-consumption of photovoltaic electricity, Appl. maximize self-consumption and self-sufficiency, Energy 186 (2019) 115827.
Energy 188 (2017) 604–619. [39] N. Delgarm, B. Sajadi, S. Delgarm, F. Kowsary, A novel approach for the
[27] T. Beck, H. Kondziella, G. Huard, T. Bruckner, Assessing the influence of the simulation-based optimization of the buildings energy consumption using NSGA-II:
temporal resolution of electrical load and PV generation profiles on self- case study in Iran, Energ. Buildings 127 (2016) 552–560.
consumption and sizing of PV-battery systems, Appl. Energy 173 (2016) 331–342. [40] S. Amiri-Pebdani, M. Alinaghian, S. Safarzadeh, Time-of-use pricing in an energy
[28] R. Savolainen, R. Lahdelma, Optimization of renewable energy for buildings with sustainable supply chain with government interventions: a game theory approach,
energy storages and 15-minute power balance, Energy 243 (2022) 123046. Energy 255 (2022) 124380.
[29] L. Maier, M. Schönegge, S. Henn, D. Hering, D. Müller, Assessing mixed-integer-
based heat pump modeling approaches for model predictive control applications in
buildings, Appl. Energy 326 (2022) 119894.

15