Energy & Buildings 252 (2021) 111385

Contents lists available at ScienceDirect

Energy & Buildings

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

Optimisation of multi-residential building retrofit, cost-optimal

and net-zero emission targets
Maria Panagiotidou, Lu Aye ⇑, Behzad Rismanchi
Renewable Energy and Energy Efficiency Group, Department of Infrastructure Engineering, Faculty of Engineering and Information Technology, The University of Melbourne,
Victoria 3010, Australia

a r t i c l e i n f o a b s t r a c t

Article history: The European Union’s (EU) building stock is characterised by low energy efficiency and slow growth
Received 9 July 2021 rates. To achieve EU’s greenhouse gas (GHG) emission targets, doubling building retrofit rates is one of
Revised 13 August 2021 the focuses of the European Green Deal. In this article, a real-world retrofit study was conducted, testing
Accepted 20 August 2021
the limits towards carbon neutrality. A multi-objective optimisation process was developed, aiming to
Available online 25 August 2021
minimise the operating GHG emissions and the life-cycle cost. The process was applied to a typical
multi-residential building and was tested in the four Greek climate zones. It was found that the cost-
Keywords:
optimal retrofit agrees with the observed market trends (envelope insulation, double-glazed windows,
Medium-rise building
Residential building
air-to-air heat pumps (HP) and solar thermal collectors), leading to more than 60% reduction in GHG
Multi-objective optimisation emissions. A maximum of 87% to 96% reduction was achieved by applying thicker envelope insulation,
Greenhouse gas emissions low-carbon (biomass boiler) or high-efficiency (gas-condensing boiler, air-to-air or air-to-water HPs)
Life-cycle cost heating and cooling systems, photovoltaic-thermal and facade-integrated photovoltaic systems. A net-
Net-zero carbon zero GHG emission retrofit could not be achieved within the building premises without considering
the future decarbonisation of the electricity grid and the installation of efficient electricity-driven
systems.
Crown Copyright Ó 2021 Published by Elsevier B.V. All rights reserved.

1. Introduction ‘Energy Performance of Buildings Directive’ (EPBD) [8] that

requires all new buildings and buildings that undergo a major ret-
Almost 75% of the existing European buildings are not energy rofit to have a nearly-zero energy demand, known as nearly-zero
efficient [1]. The primary reason is their high energy consumption energy buildings (near-ZEB). The pathway to energy mitigation
(40%) and related greenhouse gas (GHG) emissions (36%) [2]. Being presupposes the improvement of buildings’ energy efficiency and
the 75% of the total built area in the European Union (EU), the res- the on-site renewable energy generation. However, multi-
idential sector dominates the building stock [3], holding a large residential buildings located in highly urbanised areas struggle to
share of the total final energy consumption (26% of the total in meet their energy demand through the installation of renewable
2018) [4]. Its annual growth rate across Europe is about 1%, while energy systems, such as solar, due to their limited suitable land
the average annual refurbishment rate is between 0.5% and 2.5% and envelope space. The significance of addressing that problem
[5]. At the same time, about 75% of the projected buildings in also arises from the fact that, as of 2018, 46% of the EU-27 popula-
2050 have already been built [6]. Recently, the European Commis- tion lived in multi-residential buildings, while for the Mediter-
sion set radical targets for the reduction of GHG emissions by 2030 ranean countries (Italy, Greece and Spain) that value ranges from
(50–55% compared with 1990 levels) and 2050 (climate neutral- 52% to 65% [9].
ity). To achieve those targets, the building stock is expected to play Retrofit is a complex process that involves the analysis of mul-
a critical role, doubling its retrofit rates [7]. Therefore, there is a tiple interactions among building components and the building
clear opportunity to retrofit the existing buildings, leading to sig- environment. Considering the ‘whole building’, retrofit measures
nificant energy savings and GHG emissions reduction. can be classified into two groups, namely the energy saving mea-
Retrofit has been seen as a way to improve the energy efficiency sures (ESM) targeting the building envelope and the energy supply
of the building sector, especially after being introduced by the systems (ESS) that include heating, ventilation and air-
conditioning (HVAC) systems, domestic hot water (DHW) systems
⇑ Corresponding author. and renewable energy supply systems (RESS), such as solar ther-
E-mail address: [email protected] (L. Aye). mal and photovoltaic. Retrofit also involves the reconciliation of

https://doi.org/10.1016/j.enbuild.2021.111385
0378-7788/Crown Copyright Ó 2021 Published by Elsevier B.V. All rights reserved.
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Nomenclature

Symbols EU-15 15 member countries of the European Union (prior to

C cost (€ m
2) the accession of ten candidate countries on 1 May 2004)
EF greenhouse gas emission factor (kg CO2-e kWh
1) EU-27 27 member countries of the European Union
EC energy consumption (kWh m
2) GA Genetic Algorithm
EM annual greenhouse gas emissions (kg CO2-e m
2 a
1) GAHP Gas Absorption Heat Pump
LCC life-cycle cost (€ m
2) GHG Greenhouse Gas
r real discount rate (–) GUE Gas Utilisation Efficiency
t project lifespan (a) HP Heat Pump
HPWH Heat Pump Water Heating
Subscripts HVAC Heating, Ventilation and Air-Conditioning
E operating LCC Life-Cycle Cost
I initial LHV Lower Heating Value
M maintenance LowE Low Emissivity
j index for the fuel type MOGA Multi-Objective Genetic Algorithms
n index for the year NCC National Construction Code
near-ZEB nearly Zero Energy Building
Abbreviations net-ZEB net Zero Energy Building
PLF Part-Load Factor
a-a air-to-air
a-w air-to-water PLR Part-Load Ratio
Cfa Temperate Oceanic Köppen climate zone PV Photovoltaic
COP Coefficient of Performance PV/T Photovoltaic-Thermal
RESS Renewable Energy Supply System
Csa Warm Mediterranean Köppen climate zone
DAKOTA Design Analysis Kit for Optimization and Terascale SHGC Solar Heat Gain Coefficient
Applications TABULA Typology Approach for Building Stock Energy Assess-
Dfa Hot summer continental Köppen climate zone ment
TCG Technical Chamber of Greece
DHW Domestic Hot Water
EPBD Energy Performance of Buildings Directive TESS Thermal Energy System Specialists
EPC Energy Performance Certificates TRNSYS TRaNsient SYstems Simulation (energy simulation soft-
ware package)
EPS Expanded Polystyrene
ESM Energy Saving Measures uPVC unplasticised Poly Vinyl Chloride
ESS Energy Supply Systems U-value Thermal transmittance value
EU European Union WIPV Wall-Integrated Photovoltaic

environmental, energy-related, financial, legal and social factors. In formance of the building envelope and systems. The most widely
real-world applications, market-availability, system sizing, initial used building energy performance simulation tools include TRaN-
retrofit cost, fuel cost, fuel emission factors and fuel availability, sient SYstems Simulation (TRNSYS), EnergyPlus and DOE-2
as well as other aspects, might have a considerable impact on [10,12,16].
the retrofit feasibility and outcome. Most developed methodologies for building design or retrofit
At the same time, the process of identifying the optimal retrofit energy performance optimisation introduce a single-stage simula-
solutions, among a variety of alternative measures, can be signifi- tion process. However, real-world applications would require, at
cantly demanding in terms of computing time and resources. least, a three-step approach: i) the calculation of building hourly
Multi-objective optimisation is the most commonly employed loads and peak loads, ii) system capacity sizing and selection based
method [10] as it enables the identification of trade-offs between on commercial availability, iii) calculation of the systems’ part-load
the competing objective functions. The essential concept of performance and expected energy consumption. Multi-stage
multi-objective optimisation is Pareto optimality [11], which is sequential optimisation is also linked to processes that prioritise
the set of feasible solutions for which no improvement of all objec- the optimisation of a single sub-system at every step, such as
tive functions is possible simultaneously. This is factual for real- ESMs. This approach is linked to the priority that is given to build-
world problems, where a large number of non-dominated solu- ing energy efficiency measures by the EU policies. However, it does
tions are likely to exist. not allow proper consideration of the mutual relationships
Optimisation algorithms can identify the optimal solutions between the variables affecting the energy demand and the others
without exhaustive computing time. Recent reviews of research impacting the energy supply, thus limiting the exploration of the
studies on simulation-based optimisation methods applied in entire building retrofit space [17].
building performance analysis showed that genetic algorithms Morelli et al. [18] were the first to identify the increased diffi-
(GA) were employed by more than 40% of the studies [12,13]. culty that multi-residential buildings face towards a near-ZEB ret-
The main reason is their ability to tackle non-linear, discontinuous rofit. They pointed out the need to consider not only ESM but also
problems with many local minima [14], such as building retrofit ESS, especially RESS. Ferreira et al. [19] performed a comparative
optimisation, since they are less susceptible to the shape of the study between cost-optimal and net-ZEB retrofit solutions for a
non-dominated front [15]. Optimisation tools are coupled with multi-residential building, targeting the minimisation of non-
building performance simulation software in an iterative process. renewable primary energy and life-cycle cost (LCC), considering
Dynamic simulation programs are widely used to analyse the per- both ESM and ESS. The study is more theoretical than real-world

2
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

as it assumes a single-value coefficient of performance for the feasibility of a net-zero emission retrofit. An application-oriented
building systems, as well as unlimited rooftop space where photo- framework was developed considering a wide and comprehensive
voltaic (PV) systems can be installed to supply the remaining set of design variables and targeting the minimisation of the oper-
building energy needs. ating GHG emissions and LCC, in order to provide stakeholders
According to the authors’ previous review [20], as well as other with optimal sets of commercially available solutions. Focusing
investigations [21] and recent reviews of building retrofit studies on multi-residential buildings, a challenging building typology
[22], there is a lack of retrofit methods that address the ‘whole- for the minimisation of GHG emissions, the feasibility of the goal
building’. Most identified studies that consider a large number of is tested.
interventions were conducted either for the retrofit of simplified The contributions of this study reside in the following:
hypothetical buildings [23,24] or non-residential buildings
[21,25], the design of new buildings [17,26,27] and the perfor-  The development of a three-step optimisation framework to
mance assessment of a developed computational method [28]. consider the impact of system sizing and capacity range of the
available systems. It is analogous to a reverse Sankey diagram,
1.1. Greek context starting with the calculation of building loads and ending with
the amount of energy consumption and the equivalent GHG
In Greece, the residential sector accounts for about 30% of the emissions. It employs empirical part-load performance curves
total final energy consumption [29], 65% of which is dedicated to for building energy supply technologies, considering the market
heating and cooling and 5.7% to DHW [30]. The analysis of energy available products and improving the simulation accuracy.
performance certificates (EPC) of residential buildings [22] indi-  The simultaneous consideration of ESMs, as well as energy sup-
cated that, for multi-residential buildings being constructed prior ply and demand-side technologies, in an integrated ‘whole-
to 1980 the primary energy consumption for heating is about building’ approach, exploring the entire design space within
150 kWh m
2 a
1 in Greek climate zone A and about 370 kWh the same iteration, without prioritising ESMs over the others
m
2 a
1 in Greek climate zone D. The numbers for space cooling and thus leading to better results in terms of the minimisation
are about 50 kWh m
2 a
1 for climate zone A and 20 kWh m
2 of LCC.
a
1 for climate zone D, while for DHW about 50 kWh m
2 a
1 for  The assessment of near and net-zero emissions target for multi-
all climate zones. Balaras et al. [31] compared the measured and residential buildings, considering the constraints (fuel availabil-
calculated primary energy for heating, using data from the EPCs ity, market availability of products and systems, building envel-
that were calculated using the tool developed by EU members to ope space availability for the installation of RESS and more) and
support the EPBD. Results indicated that multi-residential build- the opportunities (decarbonisation of the electricity grid) dic-
ings consume 43% less energy than the amount calculated. tated by the building environment.
The most popular retrofit interventions in multi-residential
buildings are the replacement of single-glazed windows with The following sections describe the structure of the framework,
double-glazed and the central diesel oil conventional boiler with upon which the case study is demonstrated.
a natural gas central or local boiler, saving 15% and 21% of primary
energy respectively [31]. Kolaitis et al. [32] compared the impact of 2. Methods
an external to an internal insulation system on the heating and
cooling load, for an apartment located in a mid-floor of a multi- The developed optimisation framework quantitatively analyses
residential building, in two Köppen climate zones; warm Mediter- the cost-optimal solutions and the solutions that minimise the
ranean (Csa) and temperate Oceanic (Cfa). Both configurations sig- annual GHG emissions, with reference to the retrofit of multi-
nificantly reduced the total energy needs, however, external residential buildings. It combines energy simulation and multi-
insulation outperformed internal by 8% for both climate zones. objective optimisation. Scenario analysis is also employed for a
Regarding the identification of the gap between cost-optimal and range of electricity GHG emission factors, to assess the impact of
near-ZEB retrofit solutions, Pallis et al. [33] studied the comparative future decarbonisation of the electricity grid on the minimisation
impact of a large number of retrofit interventions on the minimisa- of annual GHG emissions. The following sections present the
tion of the annual primary energy use and the LCC, considering framework structure and the employed tools, after providing the
single-family and multi-residential buildings in climate zones B and objective functions and design variables considered.
C that were constructed before 1980 and between 1980 and 2000.
For multi-residential buildings of climate zone B, results indicated 2.1. Objective functions and design variables
that gas condensing boiler heating solutions dominate the cost-
optimal area, while heat pumps (HP) dominate the near-ZEB area. A bi-objective optimisation approach was selected for the per-
Regarding the fulfilment of the EU GHG mitigation targets, Das- formance assessment of the retrofit combinations. Considering
calaki et al. [34] modelled the energy savings of the Greek building more than two objective functions would increase the complexity
stock, resulting from a number of retrofit scenarios, based on the in the outcome interpretation [35]. For the calculation of the
Greek ‘Typology Approach for Building Stock Energy Assessment’ annual GHG emissions, the energy consumption for heating, cool-
(TABULA). There were two scenarios considered: one for ESMs (en- ing and DHW systems was estimated by using the GHG emissions
velope scenario) and one for heating and DHW systems (system factor of each fuel type, as shown in Eq. (1).
scenario). Results indicated that both envelope and system scenar-
ios fell short to meet the GHG emissions mitigation targets, even X
J
EM ¼ ðEC j EF j Þ ð1Þ
when considering unrealistically high retrofit rates. Only when j¼1
the envelope and system scenarios were combined the EU targets
could be achieved. For the calculation of LCC, the discounted cash flow of each year
was determined at the starting point of the project, according to
1.2. Objectives and contribution of the work Eq. (2):
Xt
C En þ C M n þ C Rn
The present study aims a real-world investigation regarding the LCC ¼ C I þ ð2Þ
ð1 þ rÞn
multi-objective optimisation of multi-residential buildings and the n¼1

3
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

The proposed lifespan for the calculation of the global cost of The advantage of the three-step optimisation process is the
residential buildings is 30 years (EU regulation 244/2012). How- impact consideration of system sizing and the capacities of com-
ever, provided the fact that the method is developed for an existing mercially available products, as well as the part-load performance
building, the considered lifespan is 20 years and is within the lifes- of the simulated systems. For that purpose, TRSNSYS capabilities
pan of most building systems (according to EN 15459:2014 and for the calculation of energy consumption, based on part-load per-
manufacturers’ supplied information). The candidate ESM and formance curves developed using empirical or manufacturer’s sup-
ESS are chosen from the list of Table 1. The considered measures plied data, were employed, improving simulation accuracy.
are those that affect the energy consumption and related emissions
for heating, cooling and DHW. The energy consumptions for light-
3. Case study application
ing, equipment and plug loads were not considered. Finally, price
offers from local contractors and construction material retailers
This section presents the retrofit study of a multi-residential
are used for the accurate estimation of the current market prices.
building, located in urban areas of the four Greek climate zones,
as well the building’s financial, legal and social environment. The
preparation process and the specifics of the optimisation are
2.2. The optimisation framework and tools described in Section 3.1 and 3.2, while Section 3.3 provides infor-
mation of the model verification and validation.
The developed framework of Fig. 1 consists of two stages, the According to the last national census data (2011), almost 50% of
preparation process and the main optimisation process. The prepa- the Greek dwellings were in multi-residential buildings, located in
ration process starts with the pre-screening phase, where the urban areas, while 50% of the multi-residential buildings were built
building expert assesses the competency and applicability of the before 1980 [39] and the introduction of the thermal insulation
design variables and decides on their value range, considering legislation. Looking at the multi-residential buildings constructed
the legislative, climatic, technological and financial constraints prior to 1980, approximately 90% do not have wall insulation, more
applied. An additional phase of the preparation process is the cal- than 60% have single-glazed windows and almost 60% have a cen-
culation of the energy consumed or produced by the systems tral heating system.
whose energy performance can be calculated regardless of the Those figures justify the selection of the case study building,
building envelope performance, such as the water heating systems which is a multi-residential building, constructed prior to 1980
or RESS. Results are saved in ‘csv’ format files in order to be (year 1961) and located in Athens, Greece (climate zone B, see
accessed later by the main optimisation process, reducing compu- Table 2). It was selected from the EU-funded programme TABULA
tation time and complexity. [40]. It is a six-storey building, which has two shops on the ground
The main optimisation process contains the optimisation rou- floor and 27 apartments (ground floor: 3 apartments, 1st to 4th
tine that calls the simulation sub-processes, Model 1 and Model floor: 5 apartments, 5th floor: 4 apartments). In addition, the build-
2, iteratively. Each model is self-contained. The first one estimates ing has a basement area and a staircase exit to the roof. In total, the
the building heating and cooling energy demand and peak loads. residential conditioned area is 1,360 m2. Fig. 2 shows the building
Based on those outputs, a developed python code performs system in ‘Google Street View’ and a typical floorplan. In order to identify
sizing for Model 2, which in turn calculates systems’ energy con- the way that various parameters of the building environment
sumption and the annual building’s GHG emissions and LCC. might affect the method application and obtained results, four
Design Analysis Kit for Optimization and Terascale Applications locations were considered, as listed in Table 2.
(DAKOTA) developed by the Department of Energy Sandia National
Laboratories [36] handles the optimisation process. TRNSYS,
energy simulation software package developed by the Solar Energy 3.1. Pre-screening of the design variables
Laboratory, the University of Wisconsin-Madison, College of Engi-
neering [37] is the environment for the transient simulation of The pre-screening of the design variables was performed con-
the building and systems’ performance. DAKOTA sees TRNSYS as sidering the legal, climatic, technological and financial environ-
a black box; the developed Python code interfaces them. The ment of the case study building. For this purpose, a building
Multi-Objective Genetic Algorithm (MOGA) [38] is employed by audit was conducted to assess the general condition of the build-
DAKOTA to reduce the required computation time. Within the ing, verify TABULA parameter values and results and thus identify
optimisation process, three primary functions are performed: i) the variations between the archetype and reality. The following
the assignment of input parameters, ii) the execution of the sections present the design variables of the optimisation process
‘whole-building’ simulation, and iii) the assessment of simulation in two categories: i) ESMs, and ii) measures considering the ESS,
results and the generation of new input parameters, as required. including heating and cooling, DHW and RESS.
The procedure goes on until convergence is achieved or a ‘stopping
criterion’ is satisfied. Along with the iterations’ results summary, 3.1.1. Energy saving measures
Pareto front is obtained. Information about the building was extracted from TABULA [40]
and the national construction codes (NCC) [44]. Table 3 provides
the details of the construction types of the building envelope, while
Table 1
Table 4 the information regarding the building thermal gains and
‘Whole-building’ energy retrofit measures.
losses. According to the NCC of the construction year (prior to
Category Retrofit measure 1980), there were no insulation requirements for the building
Energy saving Insulation of external walls envelope. Currently, the Greek regulation for the energy perfor-
Insulation of roof mance of buildings requires buildings that undergo a major retrofit
Insulation of floor/basement ceiling to meet the maximum allowed thermal conductivity requirement,
Window replacement
Installation of shadings
to the extent that it is technically and financially feasible. Thus, the
Energy supply Replacement of HVAC systems design variables regarding the thermal conductivity of the building
Installation of HVAC control systems envelope were constrained by those regulations. The upper con-
Replacement of DHW systems straint and step of each variable were defined by the market avail-
Installation of RESS (PV panels, ground source HPs, etc.)
ability of the products. Expanded Polystyrene (EPS) panels were
4
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 1. Schematic view of the developed optimisation framework.

Table 2 Fig. 2a). The initial cost for window replacement is provided in
The selected cities and the Greek and Köppen climate zone classification [42,43]. Table A.2 of the Appendix.
City Greek climate zone Köppen climate zone
Heraklion A Csa 3.1.2. Measures considering the ESS
Athens B Csa ESS are distinguished to HVAC systems, DHW systems and
Thessaloniki C Csa/Cfa RESS. Regarding the existing systems, the case study building has
Florina D Cfa/Dfa a double-pipe high temperature central hydronic heating system,
Csa: Mediterranean hot summer climates, Cfa: Humid subtropical climates, Dfa: Hot coupled with a natural gas non-condensing boiler and radiators.
summer continental climates During the heating season, heating is supplied to the conditioned
zones for a few hours during midday and evening. This is the tra-
ditional way central heating systems were designed to operate in
selected as the insulation material due to their low thermal con- Greece. There is no central cooling system, however, most of the
ductivity and low environmental impact [45]. Their mechanical apartments have air-to-air (a-a) HPs installed in key apartment
(compressive strength) and physical (moisture resistance) proper- areas, such as the living room and/or the master bedroom. DHW
ties allow their use for the insulation of all external and internal is supplied by electric water heaters and storage tanks located in
building surfaces. Table 5 provides the range of the envelope insu- each apartment. No RESS were installed at the building. It should
lation design variables. The step of the variable is 10 mm, equiva- be mentioned that mechanical ventilation wasn’t considered a
lent to the market-available insulation products. For simplicity competent retrofit measure as the case study building lacks ade-
purposes, envelope insulation alternatives were grouped into four quate ceiling or underfloor space for the required ductwork.
major categories: non-insulated envelope elements, EPS-thin (30 –
70 mm), EPS-medium (80–100 mm) and EPS-thick (110 – 150 mm)
3.1.2.1. HVAC systems. For modelling purpose, several assumptions
insulation. The initial cost for the building envelope insulation is
were made. The 27 apartments were grouped per floor level in 7
provided in Table A.1 of the Appendix.
conditioned zones. Common building areas, such as the basement,
The windows of the case study building have a higher thermal
the staircase and corridor and the staircase rooftop exit were mod-
conductivity value than the NCC requirements for buildings that
elled as non-conditioned zones. Apartment autonomy was intro-
undergo major retrofit, leading to significant heat losses during
duced through the installation of water valves in every radiator,
the heating season and gains during the cooling season. The market
controlled by thermostats in each apartment. During the heating
available window frame materials are mainly aluminium and
season (from November 1 to April 15 for climate zones A and B,
unplasticised poly vinyl chloride (uPVC). The prior has inferior
and from October 15 to April 30 for climate zones C and D) [44],
thermal properties and a higher cost than the later; thus, it wasn’t
the heating system was designed to operate in two modes. The
considered. Two window replacement options were selected, using
daytime set-point temperature was 20 °C [44]. The set-back tem-
uPVC as frame material; the first one is double-grazed while the
perature for night-time was 18 °C.
second is triple-grazed (see Table 6). Finally, no additional ESM,
The site visit revealed that the existing boiler is 45% oversized
such as shading systems, were considered as most glazed areas
(200 kW), compared to the peak load (137 kW) estimated using
are shadowed by the building balconies and awnings (see
TRNSYS for a continuous operation. This is justified by the current
5
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 2. The case study building, a. ‘Google Street View’ [41], b. plan of a typical floor (red: conditioned zone, blue: non-conditioned zone). (For interpretation of the references
to colour in this figure legend, the reader is referred to the web version of this article.)

Table 3 perature of the selected alternative low-temperature systems, the

Construction types and their thermal conductivity (U-value) (W m
2 K
1) of the case assumed water distribution temperature for all hydronic heating
study building, TABULA (2017).
systems was 55 °C.
Elements Construction type U- During the cooling season (from May 15 to the September 15 for
value climate zones A and B, and from June 1 to August 31 for climate
Wall 1 Double brickwork - plastered on both sides 2.20 zones C and D) [44] the set-point temperature for cooling was
Wall 2 Load bearing structure – reinforced concrete - plastered 3.40 26 °C. For the alternative hydronic cooling system, the assumed
on both sides
water supply temperature was 7 °C.
Roof Flat roof 3.05
Floor Slab on grade 3.10 System selection was constrained by the climate conditions,
Window Single glazed, wooden or synthetic frame 4.70 market availability, fuel availability and space limitations. Build-
ings located in South European countries have moderate heating
demand while cooling demand is significant [47]. For that reason,
operation mode, which is a couple of hours in the morning and heating and cooling were studied together, grouping the alterna-
evening. In this study, system sizing followed the method reported tive systems into one design variable. The reference building that
by Lhendup [46] (80% of peak load), also considering the market the retrofit cases were compared to was named ‘base case’. For
available system capacities. Provided the oversized components locations without natural gas (climate zone A and D), the ‘base
(distribution pipes and radiators) of the existing high temperature case’ heating system was a central diesel oil conventional boiler
hydronic system and considering the maximum water supply tem- and the alternatives a diesel oil condensing boiler or a biomass boi-

Table 4
Case study building thermal losses and gains.

Parameter Value Unit References

2
1
Thermal bridge 0.15 Wm K TABULA [40]
Infiltration 0.40 hr
1
Ventilation 0.75 m3 hr
1 m
2 Technical Chamber of Greece (2010b)
Occupants’ thermal power 4.00 W m
2
Lighting capacity 6.40 W m
2
Lighting in-parallel use factor 0.50 –
Equipment capacity 4.00 W m
2
Equipment operation factor 0.75 –
Equipment in-parallel factor 0.50 –

6
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Table 5
ESMs design variables and values for commercially available envelope insulation.

Building envelope element EPS 80 insulation panel thickness (mm) alternatives

Greek climate zone
A B C D
External walls – outdoors boundary condition 1) no insulation 1) no insulation 1) no insulation 1) no insulation
2) 50–70 2) 60–70 2) 70 2) 80 – 100
3) 80–100 3) 80–100 3) 80–100 3) 110–150
4) 110–150 4) 110–150 4) 110–150 –
Horizontal roof – outdoors boundary condition 1) no insulation 1) no insulation 1) no insulation 1) no insulation
2) 60–70 2) 70 2) 80 – 100 2) 90–100
3) 80–100 3) 80–100 3) 110–150 3) 110–150
4) 110–150 4) 110–150 – –
Basement ceiling – non-conditioned space boundary conditions 1) no insulation 1) no insulation 1) no insulation 1) no insulation
2) 30–70 2) 30–70 2) 40–70 2) 50–70
3) 80–100 3) 80–100 3) 80–100 3) 80–100
4) 110–150 4) 110–150 4) 110–150 4) 110–150

Table 6 operation they were calculated based on DOE-2 boiler curve [50].
ESMs design variables and U-values (W m
2 K
1) for window replacement. The indicative commercial heating and cooling products and their
Window type Frame Glazing SHGC available capacity range, a breakdown of their initial cost and the
1) No replacement 2.2 5.6 0.8
maintenance cost can be respectively found in Table A.3,
2) uPVC frame, argon filled double glazing 1.1 1.1 0.6 Table A.4 and Table A.5 of the Appendix.
(4  15  5 mm)
3) uPVC frame, lowE argon filled triple glazing 1.0 0.6 0.5
3.1.2.2. Domestic hot water systems. The most commonly used
(4LowE  15  5  15  5 mm)
domestic water heater is the electric. Most of the times, it has a
SHGC: Solar Heat Gain Coefficient. small volume and the user has to turn it on a few minutes before
using it. At the same time, the total installed solar thermal collector
capacity in Greece is 2,301 MWth, placing the country at the second
ler. For locations with natural gas (climate zone B and C), a central rank among the EU-15 [51]. However, rooftop space limitations
natural gas conventional boiler was the ‘base case’, while a natural apply in multi-residential buildings. For the case study building,
gas condensing boiler or a gas absorption heat pump (GAHP) were it was estimated that the available rooftop space is not enough
the alternative central systems. A-a HPs were considered as the to accommodate solar thermal panels that, combined with a con-
‘base case’ cooling system for all locations. Autonomous (apart- ventional auxiliary system, can cover more than 20% of the build-
ment) systems, that replace the central boiler, were also consid- ing’s space heating requirement. However, the estimated hot water
ered for all locations; either new high-efficient a-a HPs (two production can cover 85% of the apartment-based DHW heating
units per apartment) or an air-to-water (a-w) HP (3 fan coil units system needs [52]. Thus, the installation of DHW and RESS systems
and new water distribution system per apartment) both for heat- were studied in parallel.
ing and cooling. The conducted building audit assessed the existing Most of the existing systems were evaluated by the conducted
systems as being close to the end of their life, thus the ‘no replace- building audit as being close to the end of their life, thus the ‘no
ment’ option was not considered. Table 7 lists the existing and replacement’ option was not considered. The ‘base case’ water
selected alternative system combinations. heating system is electric, while the heat pump water heating
For the calculation of systems’ energy consumption, part-load (HPWH) system was proposed as an alternative. It should also be
performance curves were developed (see Fig. 3) from empirical mentioned that the hot water produced by the solar thermal pan-
or manufacturer’s supplied data. HPs and high efficiency boilers els (a retrofit alternative under the design variable of RESS pre-
have modulating capabilities that significantly limit the on/off sented in the following section) will be supplied to the electric
operation energy losses. For HPs we employed the method devel- water heating system. For the purpose of this study, the setpoint
oped by Schibuola, Scarpa & Tambani [48], that using HP products’ water temperature of 60 °C and a 200-litre cylindrical storage tank
EN 14825:2016 test reports (efficiencies for a range of part-load were assumed for all water heating systems. A detailed study of
conditions for four ambient temperatures) they produced part- the DHW systems was conducted by the authors [52], containing
load ratio (PLR) – part-load factor (PLF) curves. For boilers, the technical and cost information.
part-load performance curves during capacity modulation were
developed and expanded for a range of return water temperatures, 3.1.2.3. Renewable energy supply systems. Solar systems are the
using the method developed by Cockroft et al. [49]. During on/off most widely installed RESS in the urban environment. However,

Table 7
Energy supply design variable for heating and cooling systems.

Locations with natural gas Locations without natural gas

Heating Cooling Heating Cooling
1) Natural gas condensing boiler a-a HP (autonomous - two units per Diesel oil condensing boiler a-a HP (autonomous - two units per
(central) apartment) (central) apartment)
2) GAHP (central-single or multiple Biomass pellet boiler (central)
units)
3) a-w HP (autonomous - one unit per apartment)
4) a-a HP (autonomous - two units per apartment)

7
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 3. Thermal efficiency (LHV)/COP curves of selected heating and cooling systems, a. boilers, b. HPs. [Note: The boiler efficiencies were calculated for 55 °C/40 °C water
supply/return temperature, while the HP efficiencies for 10 °C (heating mode) and 30 °C (cooling mode) ambient temperature.

multi-residential buildings have limited suitable areas for the supplied to the DHW tank (with electric boosting) of each apart-
installation of RESS, due to the high site coverage and the extended ment. It should be mentioned that in case the HPWH system was
envelope areas shadowed by the nearby buildings. Market limita- selected for the DHW system design variable, the solar thermal
tions make PV, solar thermal and photovoltaic-thermal (PV/T) pan- panels and the PV/T panels were excluded from the RESS alterna-
els the most competent alternatives for the building’s suitable roof tives. A detailed study regarding the performance of the selected
area, while for facade-integrated systems, only the installation of roof-applied RESS was conducted by the authors [52], containing
wall-integrated PV (WIPV) panels was considered (see Table 8). technical and cost information.
A preliminary study was conducted to identify the suitable
areas for installing solar systems at the case study building. The 3.2. Optimisation settings
annual cumulative solar radiation (kWh m
2 a
1) was calculated
in Grasshopper-Rhino environment, using the Ladybug plugin that For the calculation of the annual GHG emissions, the energy
interface with Daysim/Radiance software [53]. The rooftop and consumption was multiplied by the GHG emission factor of each
facade surfaces that exceed the minimum annual solar radiation electricity/fuel type, provided in Table 9. For the calculation of
threshold (1,000 kWh m
2 a
1 and 800 kWh m
2 a
1 equivalently), the annual energy cost, the energy consumption was multiplied
as proposed by Compagnon [54], were considered to be suitable for by the energy price of Table 10.
PV installation, in order to respond to the present technical limita- A discount rate between 2% and 4% was advised to be used for
tions and economic factors. Fig. 4 shows the suitable PV areas for energy efficiency investments made by building occupants, as it
climate zone B: 181 m2 for rooftop and only 73 m2 of facade area. reflects the actual building owners’ benefits, over the entire life-
The maximum panel area that can be installed on the unshaded time [56]. According to Buildings Performance Institute Europe,
rooftop area of the rooftop (panels are facing north and have a the applicable discount rate for space heating and hot water of a
25° tilt angle) is 81 m2, while for the facade (parallel to the wall), household is 3.1% to 3.7% [57,58]. This investigation used the real
considering accessibility, is 65 m2 for climate zones B, C and D discount rate of 4% assuming higher uncertainty.
and 47 m2 for climate zone A. The same module efficiency was The steps of a GA application are initialisation, selection of indi-
assumed for both wall-integrated and roof-applied PV systems viduals, crossover, mutation and termination. The GA control
(16.5%), considering 0.7% annual degradation. The produced elec- parameters can significantly affect its performance in terms of
trical energy was equally divided into the 27 apartments through speed and reliability. For example, a high mutation rate value
virtual net-metering [55]. might lead to the loss of good solutions, while a high crossover rate
The same panel area was also assumed for rooftop solar thermal to an early convergence. The population size should also be about
and PV/T installations. For those cases, a 3 m2 panel area was allo- 2% – 6% of the number of genes. For this investigation, MOGA con-
cated to each household and the produced hot water was directly trol parameters were set as shown in Table 11, based on previous
studies [59] and tests carried out by the authors in order to obtain
Table 8 the best trade-off between the computational time and reliability
Renewable energy supply systems. of results.
Roof-installed Wall-integrated PV
1) No RESS 1) No WIPV panels 3.3. Model verifications and validations
2) PV panels 2) WIPV panels
3) Solar thermal panels – There are six categories of computer models applied in this
4) PV/T panels –
work:
8
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 4. The identification of suitable rooftop and facade areas for the installation of PV systems in Athens (climate zone B), a. annual solar radiation, b. suitable rooftop areas, c.
suitable facade areas (yellow: suitable surfaces, purple: nearby buildings and shading elements). (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)

Table 9 with the manufacturers’ and test reports’ data (the coefficient of
GHG emission factors (kg CO2-e kWh
1) for Greece. determination, R2 ranges from 0.958 to 0.997)[67].
Fuel type Factor Reference The annual heating load calculated in TABULA was 70.5 kWh
Electricity 0.810 [60]
m
2 a
1. The modelled annual heating load per floor area of the
Natural gas 0.240 [60] building was 97.6 kWh m
2 a
1. For DHW, the calculated TABULA
Diesel oil 0.306 [60] input electrical energy was 18.3 kWh m
2 a
1, while the modelled
Biomass pellet 0.063 [61] was 21.2 kWh m
2 a
1. The modelled annual cooling load per floor
area was 76 kWh m
2 a
1. No data for the cooling load was avail-
able in TABULA. The difference between TABULA and simulated
Table 10 results was mainly attributed to the different calculation methods
Energy prices (€ kWh
1) in Greece. employed and assumptions made; TABULA uses a static calculation
Fuel type Price Reference
method, while TRNSYS uses a dynamic approach.
Validation of results via comparison with electricity/fuel bills or
Electricity 0.18001 [62]
building monitoring was not feasible as the heating and cooling of
Natural gas 0.0509 [63]
Diesel oil 0.1000 [64] the building or individual apartments is not based on thermostat
Biomass pellet 0.05002 [65] settings but on fixed heating hours. In fact, the conducted field sur-
1 vey by the authors confirmed poor indoor thermal comfort condi-
Based on the average electricity consumption of 2,000 kWh per four months.
2
The pellet has a calorific value of 5 kWh kg
1 and costs 0.250 € kg
1. tions, resulting from reduced heating and cooling operating hours
and lower thermostat settings that deviate from the calculation
assumptions. In addition, some apartments have improved their
Table 11 energy performance individually, via either wall insulation, win-
MOGA control parameters. dow replacement or radiators’ replacement; thus, building moni-
Parameter Value
toring would have been a tremendous task.

Population size 14
Crossover rate 0.8 4. Results
Mutation rate 0.08
Maximum number of evaluations 1500 The obtained results (see Figs. 5–8) are presented and discussed
considering three major parameters: the climate zone, the avail-
ability of natural gas infrastructure and the future decarbonisation
1. Models from Daysim/Radiance software tool; of the electricity grid. It should be mentioned that natural gas
2. Models from Ladybug plug-in for Grasshopper-Rhino; infrastructure network is not available nationwide. Among the
3. Optimisation models from DAKOTA; considered locations, it is available in Athens and Thessaloniki,
4. TRNSYS models (or Types) from Standard Component Library; located in climate zones B and C, respectively.
5. TRNSYS models from Thermal Energy System Specialists (TESS) To visualise the solution space, non-dominated sets of retrofit
Component Library Package; measures are presented using black colour, while dominated using
6. TRNSYS models from User-Written Component Libraires. grey. To illustrate the optimal solutions, six additional graphs are
provided; namely for wall, roof and basement ceiling insulation,
In general, validations of Category 1–5 models were reported in window replacement, heating and cooling systems and RESS.
the literature. For example, TRNSYS Type 56, which was used for DHW optimal solutions are not illustrated as only the electric
the modelling of the case study building, was validated for Rome heater/booster was part of the Pareto front, for all climate zones.
(Mediterranean climate) and its high accuracy (the coefficient of
determination, R2 ranges from 0.854 to 0.990) was reported by 4.1. Locations without natural gas
Mazzeo et al. [66]. In addition, The TRNSYS Types developed for
the performance assessment of the heating and cooling systems Two locations are under this category; Heraklion (climate zone
were validated through the comparison of the simulated results A) and Florina (climate zone D). Comparing the results of those
9
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

a
500
LCC (€ m-2)

400

300

200

100 base case

dominated solutions
optimal solutions
0
0 10 20 30 40 50 60 70
GHG (kgCO2-e m a-1)
-2

b c d
400
EPS-thin no insulation
LCC (€ m-2)

no insulation
EPS-thin EPS-medium EPS-thin
EPS-medium EPS-thick EPS-medium
EPS-thick EPS-thick

300

200
e f
400
LCC (€ m-2)

no replacement Biomass boiler & a-a HPs solar thermal

double glazing A-a HPs solar thermal & WIPV
tripe glazing A-w HP PV/T
PV/T & WIPV

300

200
0 10 20 30 0 10 20 30 0 10 20 30
GHG (kgCO2-e m-2 a-1) GHG (kgCO2-e m-2 a-1) GHG (kgCO2-e m-2 a-1)

Fig 5. Retrofit optimisation results for Greek climate zone A (Heraklion), a. dominated and non-dominated Pareto front results, b. wall insulation, c. roof insulation, d.
basement ceiling insulation, e. window replacement, f. heating and cooling systems, g. RESS optimal retrofit solutions.

locations (see Fig. 5 and Fig. 6) we can see the impact of climate on (minimum of 5.3 kg CO2-e m
2 a
1) for Florina. The target of net-
the optimal solution sets and the objective functions. High GHG zero GHG emissions cannot be achieved for the current emission
emissions reduction can be achieved for both locations, up to factor of the electricity grid, under the assumptions of this
95% (minimum of 3.1 kg CO2-e m
2 a
1) for Heraklion and 94.7% investigation.

10
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

a
500
LCC (€ m-2)

400

300

200

100 base case

dominated solutions
optimal solutions
0
0 10 20 30 40 50 60 70 80 90 100
GHG (kgCO2-e m-2 a-1)
b c d
400
no insulation EPS-medium EPS-medium
LCC (€ m-2)

EPS-medium EPS-thick EPS-thick

EPS-thick

300

200
e f g
400
LCC (€ m-2)

300
solar thermal
no replacement Biomass boiler solar thermal & WIPV
double glazing & a-a HPs PV/T
tripe glazing A-a HPs PV/T & WIPV
200
0 20 40 0 20 40 0 20 40
GHG (kgCO2-e m-2 a-1) GHG (kgCO2-e m-2 a-1) GHG (kgCO2-e m-2 a-1)

Fig 6. Retrofit optimisation results for Greek climate zone D (Florina), a. dominated and non-dominated Pareto front results, b. wall insulation, c) roof insulation, d. basement
ceiling insulation, e. window replacement, f. heating and cooling systems, g. RESS optimal retrofit solutions.

11
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 7. Retrofit optimisation results for Greek climate zone B (Athens), a. dominated and non-dominated Pareto front results, b. wall insulation, c. roof insulation, d. basement
ceiling insulation, e. window replacement, f. heating and cooling systems, g. RESS optimal retrofit solutions.

The cost-optimal solutions for both locations require the insula- ‘base case’. The equivalent number for Florina is 59.6%. Regarding
tion of roof and basement ceiling, a-a HPs for heating and cooling the LCC of the cost-optimal solution, it can be reduced by 32.5%
and solar thermal panels for DHW. No wall insulation and window in Heraklion and 44.7% in Florina. When the target of GHG emis-
replacement are included. The cost-optimal solution set for Herak- sions mitigation is greater, additional ESMs are needed, such as
lion releases 61.9% less annual GHG emissions compared to the thick wall insulation and triple-glazed window replacement, while

12
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 8. Retrofit optimisation results for Greek climate zone C (Thessaloniki), a. dominated and non-dominated Pareto front results, b. wall insulation, c. roof insulation,
d. basement ceiling insulation, e. window replacement, f. heating and cooling systems, and g. RESS optimal retrofit solutions.

the biomass boiler for space heating and the PV/T and/or WIPV 4.2. Locations with natural gas
become part of the optimal solutions, for both locations. In Herak-
lion, the a-w HP for heating and cooling is also among the optimal The studied locations under this category are Athens (climate
retrofit sets that minimise GHG emissions, however, the LCC zone B) and Thessaloniki (climate zone C) and their optimisation
increases significantly, due to the high replacement cost of the results are presented in Fig. 7 and Fig. 8 equivalently. Overall, great
existing water distribution system with new pipes and fan coil GHG emissions reduction can be achieved for both locations, up to
units. 96.1% (minimum of 2.8 kg CO2-e m
2 a
1) for Athens and 87.5%
13
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Fig 9. Scenario analysis for a range of electricity GHG emission factors (kgCO2-e kWh
1), a. Heraklion, b. Athens, c. Thessaloniki and d. Florina.

(minimum of 10.9 kg CO2-e m
2 a
1) for Thessaloniki. Similarly to 4.3. The future decarbonisation of the electricity grid
locations without natural gas infrastructure, the target of net-zero
GHG emissions cannot be achieved for the current emission factor Building retrofit, being an environmental impact study, is sensi-
of the electricity grid. tive to fuel and electricity GHG emission factors. While fuel GHG
The cost-optimal retrofit set consists of roof and basement ceil- emission factors do not change drastically over time, there is an
ing insulation, a-a HPs for heating and cooling and solar thermal effort towards the decarbonisation of the electricity grid. In order
panels for DHW. The operating GHG emissions are 27.5 kg CO2-e to study the potential of net-zero GHG emissions building retrofit,
m
2 a
1 (62% reduction compared to ‘base case’) in Athens and a scenario analysis was conducted for a range of electricity GHG
34.5 kg CO2-e m
2 a
1 (60.5% reduction compared to ‘base case’) emission factors, derived from the EU 2050 targets for carbon neu-
in Thessaloniki. The LCC of the cost-optimal solution is reduced trality and the considered retrofit lifespan. Assuming a linear
by 28.8% in Athens and by 30.5% in Thessaloniki. When lower decarbonisation process, the electricity GHG emission factor will
annual GHG emissions are targeted, retrofit solutions are a combi- not exceed 0.54 kgCO2-e kWh
1 by 2030 and 0.27 kgCO2-e kWh
1
nation of thick wall, roof and basement ceiling insulation, installa- by 2040. The scenario analysis results are graphically presented in
tion of triple-gazed windows, replacement of the conventional gas Fig. 9, while the retrofit interventions that minimise the annual
boiler with a condensing for heating, a-a HPs for cooling, PV/T pan- GHG emissions are listed in Table 12.
els and WIPV for DHW and electricity production. Similarly to Her- Overall, it can be noted that buildings located in climate zones A
aklion, the retrofit solutions that minimise the annual GHG and B can get very close to achieve a net-zero GHG emissions retro-
emissions of the building in Athens require the installation of a- fit (minimum GHG emissions equal to 1 and 0.9 kgCO2-e m
2 a
1
w HPs; once more, the LCC is significantly increased as a result. equivalently), while for buildings located in climate zones C and
On the other hand, in Thessaloniki, the GAHP is the heating alter- D the target can be approached by 2040 (minimum GHG emissions
native that minimises GHG emissions; compared with the natural equal to 3.8 and 4.1 kgCO2-e m
2 a
1 equivalently). Regarding the
gas condensing boiler, the GHG emissions additional savings are retrofit interventions that minimise GHG emissions, increased
negligible, while the LCC increases dramatically due to the high ini- envelope insulation and efficient electricity-driven heating and
tial cost of the units. cooling solutions, such as a-a and a-w HPs, as well as PV/T and

14
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Table 12
Optimal set of solutions that minimise the GHG emissions considering the gradual decarbonisation of the electricity grid.

GHG emission Retrofit Locations without natural gas Locations with natural gas
factor intervention
Greek climate zone A Greek climate zone B Greek climate zone C Greek climate zone D
(kgCO2-e kWh
1)
(Heraklion) (Florina) (Athens) (Thessaloniki)
0.81 Envelope EPS-thick EPS-thick EPS-thick EPS-thick
insulation
Window uPVC triple-glazed uPVC triple-glazed uPVC triple-glazed uPVC triple-glazed
replacement
Heating/cooling Biomass boiler/a-a HPs or A- Biomass boiler/a-a HPs NG cond. boiler/a-a HPs or A- NG cond. boiler/a-a HPs or GAHP/
w HPs w HPs a-a HPs
DHW Electric booster Electric booster Electric booster Electric booster
RESS PV/T & WIPV PV/T & WIPV PV/T & WIPV PV/T & WIPV
0.54 Envelope EPS-thick EPS-thick EPS-thick EPS-thick
insulation
Window uPVC triple-glazed uPVC triple-glazed uPVC triple-glazed uPVC triple-glazed
replacement
Heating & cooling Biomass boiler/a-a HPs or A- Biomass boiler/a-a HPs A-w HPs A-w HPs
w HPs
DHW Electric booster Electric booster Electric booster Electric booster
RESS PV/T & WIPV PV/T & WIPV PV/T & WIPV PV/T & WIPV
0.27 Envelope EPS 150 mm EPS 150 mm EPS 150 mm EPS 150 mm
insulation
Window uPVC triple-glazed uPVC triple-glazed uPVC triple-glazed uPVC triple-glazed
replacement
Heating & cooling Biomass boiler/a-a HPs or A- A-a HPs A-w HPs A-w HPs
w HPs
DHW Electric booster Electric booster Electric booster Electric booster
RESS PV/T & WIPV PV/T & WIPV PV/T & WIPV PV/T & WIPV

WIPV are required. Table 12 shows the retrofit sets of solutions quate description or visualisation of results for the rest of the
that minimise the GHG emissions for each scenario, highlighting design variables, thus no further comparison can be made. It
the switch from gas-driven to electricity driven systems. should be mentioned that Pallis et al. [33] employed a monthly
quasi-steady state method and a brute-force process, while the
present study a three-step dynamic simulation and evolutionary
5. Discussion optimisation process, performing system sizing that considers
the market-available range of system capacities. The present study
The identified cost-optimal retrofit solutions are in line with the went beyond the findings of Pallis et al. [33], exploring a broader
observed market trends: the insulation of the building envelope, solution space, considering advanced technologies, such as GAHPs
the replacement of single with double-glazed windows, the instal- PV/T and WIPV, and a langer variation of parameters that affect the
lation of a-a HPs for heating and cooling and solar thermal panels building retrofit, such as the climate (all Greek climate zones) and
for DHW [31]. Other commonly applied retrofit interventions, such fuel availability (areas with and without natural gas infrastruc-
as the installation of condensing gas boilers (for locations with nat- ture). Here, we also aimed for carbon neutrality, exploring the
ural gas) and biomass boilers (for locations without natural gas) impact of electricity grid decarbonisation on the retrofit solution
can also be found in the middle of the obtained Pareto front. Mov- space and results.
ing towards carbon neutrality, there is a noticeable market pene- Targeting a near-ZEB retrofit, Ferreira et al. [19], compared cost-
tration potential for energy-efficient systems, such as a-w HPs efficient with net-zero-energy retrofit solutions of a multi-
and GAHPs, as well as RESS, such as PV/T panels and WIPV; how- residential building in Porto, Portugal, considering heating, cooling
ever, their initial cost remains high. and DHW; thus it is relevant to the present study. Their study held
It is also noted that the obtained results rely heavily on a large as the cost-optimal solution the envelope insulation with low-
number of parameters, such as the building type and construction thickness EPS panels, the installation of PVC framed double-
period, the climate, the selected retrofit measures and technolo- glazed windows, the replacement of the existing electric systems
gies, the fuel and electricity availability, the price, the discount rate for heating and DHW with a-a HPs for heating and cooling and a
and GHG emission factors. Consequently, the comparison of results natural gas boiler for DHW. When cooling was not considered,
with state-of-the-art studies is challenging, however, a high-level the cost-optimal heating system was a natural gas boiler. Those
comparison is worthwhile. results are aligned with the findings of the present study. In order
A seminal study [33] for the retrofit of reference single-family to achieve a net-ZEB retrofit, the study considered unlimited roof-
and multi-residential buildings, being constructed before 1980 top space where PV systems are installed to supply the remaining
and between 1980 and 2000, in Greek climate zones B and C, cal- building energy needs; thus, those results are not comparable with
culated the gap between the cost-optimal and near-ZEB retrofit. the present study.
Comparing their results with those of the present study for climate Regarding the ultimate goal of carbon neutrality, most reviewed
zone C, there are apparent similarities; the gas condensing boiler studies did not achieve a net-zero-energy or emissions retrofit
for heating and a-a HPs for cooling dominate the cost-optimal solu- under the determined building system boundaries and the consid-
tion space (a-a HPs for heating were not considered by Pallis et al. ered interventions [23,35,68,69]. Salata et al. [70] minimised the
[33]), while a-w HPs for heating and cooling populate the near-ZEB energy use for heating, cooling and DHW from 59.62 kWh m
2
solution space. The financial gap between cost-optimal and near- a
1 to 0.33 kWh m
2 a
1, by replacing the conventional boiler with
ZEB was calculated by Pallis et al. [33] equal to 140 € m
2, similar an a-w HP and installing a large area of PV panels on the building
to the results of the present study (see Fig. 7f). There is not ade- rooftop. The selected interventions are in agreement with the
15
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

findings of the present study. However, the ratio of PV panel area due to the limited available rooftop space for the installation of
to conditioned floor area of their case study building was 50% more RESS. However, considering the future decarbonisation of the elec-
compared to this study, fact that explains their high energy tricity grid, the annual operating GHG emissions can approach net-
reduction. zero-carbon in 2040, while efficient electricity-driven systems are
installed, such as air-to-air and air-to-water HPs, and renewable
energy supply system, such as PV/T and WIPV.
6. Conclusions Although the results obtained from the optimisation are specific
to the selected case study, the findings of the study can be gener-
This article reports on the identification of cost-optimal solu- alised to similar buildings and construction types in urban areas of
tions and solutions that minimise the annual greenhouse gas the Mediterranean climate.
(GHG) emissions, targeting a net-zero emission outcome, for the
retrofit of multi-residential buildings. The developed simulation- CRediT authorship contribution statement
based multi-objective optimisation process evaluated the perfor-
mance of a large number of retrofit interventions, addressing the Maria Panagiotidou: Conceptualization, Methodology, Soft-
whole building. Being a real-world investigation, it considered ware, Validation, Formal analysis, Investigation, Data curation,
the impact of system sizing and the capacity range of market- Writing - original draft, Writing - review & editing, Visualization.
available systems, adopting a three-step dynamic modelling pro- Lu Aye: Conceptualization, Validation, Resources, Writing - review
cess, based on hourly components’ part-load performances, that & editing, Supervision, Project administration, Funding acquisition.
improves the accuracy of the outputs. The software tools that facil- Behzad Rismanchi: Writing - review & editing, Supervision.
itated the process are DAKOTA and TRNSYS, and the Multi-
Objective Genetic Algorithm (MOGA) which avoids exhaustive
Declaration of Competing Interest
computation time and resources.
A case study application illustrated the functionality of the
The authors declare that they have no known competing finan-
developed process. The selected case study building is a typical
cial interests or personal relationships that could have appeared
multi-residential building, constructed prior to 1980, in Greek cli-
to influence the work reported in this paper.
mate zone B. To identify how various parameters of the existing
building might affect the application and the obtained results,
the building retrofit was optimised for the four Greek climate Acknowledgements
zones.
It was found that the cost-optimal retrofit solutions are in line This research has been conducted within the Department of
with the observed trends: insulation of the building envelope, Infrastructure Engineering at the University of Melbourne, Aus-
installation of air-to-air HPs for heating and cooling and solar ther- tralia. Part of this research is included in the PhD thesis of Maria
mal collectors for domestic hot water. While reducing the life-cycle Panagiotidou, whom has been supported by the Melbourne
cost they also lead to GHG emissions reduction of about 60% com- Research Scholarship award.
pared to the base case, for all climate zones. If the retrofit target is
the minimisation of GHG emissions, additional interventions are Appendix
required, such as complete insulation of the building envelope,
installation of triple-glazed windows, installation of a central bio- The appendix provides data regarding the costing of the build-
mass boiler in locations without natural gas or a gas condensing ing envelope insulation, window replacement and alternative
boiler in locations with natural gas (heating) and air-to-air HPs space heating and cooling systems. It also includes information
(cooling), or air-to-water HPs (heating and cooling) for climate about the capacity range of commercially available heating and
zones A and B, photovoltaic-thermal (PV/T) panels and wall- cooling systems. Costing of the alternative hot water heating sys-
integrated photovoltaics (WIPV). This way, the building carbon tems and renewable energy supply systems are provided in a
footprint can be significantly decreased by more than 87% and up detailed study conducted by the authors [52]. Prices were obtained
to 96%, depending on the climate zone. via personal communication with local contractors, and construc-
It was revealed that a net-zero-carbon retrofit could not be tion material retailers. The multi-residential building has 27 apart-
achieved for any locations, within the building premises, mainly ments and a conditioned area of 1,360.4 m2.

16
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Table A.1
Initial cost per unit of conditioned area (€ m
2) for the envelope insulation thickness in mm (EPS80).

Element Thickness (mm) Cost Notes

External walls 50 25.98 The cost is based on an insulated area of approximately 950 m2.
60 26.84 It includes the cost of materials and installation.
70 27.71
80 28.58
90 29.44
100 30.31
110 31.17
120 32.04
130 32.90
140 33.77
150 34.64
Horizontal roof 60 9.89 The cost is based on an insulated area of approximately 350 m2.
70 10.21 It includes the cost of materials and installation.
80 10.50
90 10.85
100 11.17
110 11.48
120 11.8
130 12.12
140 12.44
150 12.76
Basement ceiling 30 3.83 The cost is based on an insulated area of approximately 150 m2.
40 3.97 It includes the cost of materials and installation.
50 4.10
60 4.24
70 4.37
80 4.51
90 4.65
100 4.78
110 4.92
120 5.06
130 5.19
140 5.33
150 5.46
–
Scaffold cost 2.35 The cost is added only once in case of external
wall insulation and/or window replacement.

Table A.2
Initial cost per unit of conditioned area (€ m
2) for window replacement.

Window type Cost Notes

uPVC frame, argon filled double glazing 16.42 The cost is based on a total window area of 225 m2. It includes the cost
(4  15  5 mm) of materials and installation.
uPVC frame, lowE argon filled triple glazing 25.43
(4LowE  15  5  15  5 mm)
Scaffold cost 2.35 The cost added only once in case of external wall insulation and/or
window replacement.

17
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

Table A.3
Heating and cooling commercial products and their available capacities (kW).

System Nominal capacity1 Reference2

Heating Cooling
Natural gas condensing boiler 147.4 – [71]
114.6
94.0
84.5
62.3
43.6
Diesel oil condensing boiler 157.3 – [72]
120.3
93.4
85.0
67.0
50.0
Biomass boiler 174.0 – [73]
140.0
112.0
93.0
71.0
52.0
40.0
29.0
GAHP 38.3 – [74]
a-w HP 9.9 11.0 [75]
8.2 9.1
6.0 7.2
4.1 5.1
3.3 4.0
a-a HP 5.4 5.0 [76,77]
3.6 3.4
3.2 2.5
1
Nominal conditions vary.
2
Commercial product that their capacity ranges were used as a reference for the modelled systems.

Table A.4
Initial cost per unit of conditioned area (€ m
2) breakdown of the alternative systems.

Item/service cost Diesel oil cond. Biomass Natural gas cond. GAHP a-w a-a
boiler boiler boiler HP HP
Installation of thermal autonomy 4.47 4.47 4.47 4.47 – –
Pipe insulation of non-conditioned zones 0.37 0.37 0.37 0.37 – –
Dismantling of the existing central heating 0.19 0.19 0.19 0.19 – –
system
Chimney replacement 1.20 – 1.20 – – –
Secondary circuit water pump 0.60 0.60 0.60 0.60 – –
Piping installation of the autonomous system – – – – 11.91 –
Fan coils – – – – 11.91 –
New system installation 1.25 1.05 1.25 3.16 8.93 3.18
New system1 11.83 2.73 3.31 26.91 56.57 19.26
Total initial cost 19.91 9.41 11.39 35.70 89.32 22.44
1
A-a and a-w HPs are autonomous. In case of a-a HPs, two units per apartment were assumed, while in case of a-w HPs, one unit.

18
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

[20] M. Panagiotidou, B. Rishmanchi, L. Aye, Low energy building retrofit: A review

Table A.5
of objectives and solutions, in: ZEMCH 2018 Int. Conf. Conf. Proc., Melbourne,
Cost of maintenance per unit of conditioned area (€ m
2). 2018: pp. 59–74.
[21] F. Ascione, N. Bianco, G.M. Mauro, G.P. Vanoli, A new comprehensive
framework for the multi-objective optimization of building energy design:
System Cost Harlequin, Appl. Energy 241 (2019) 331–361, https://doi.org/10.1016/j.
apenergy.2019.03.028.
Natural gas cond. boiler 0.11 [22] Z. Ma, P. Cooper, D. Daly, L. Ledo, Existing building retrofits: methodology and
Biomass boiler 0.11 state-of-the-art, Energy Build. 55 (2012) 889–902, https://doi.org/10.1016/j.
enbuild.2012.08.018.
GAHP 0.11 [23] C. Diakaki, E. Grigoroudis, D. Kolokotsa, Towards a multi-objective
a-w HP 1.59 optimization approach for improving energy efficiency in buildings, Energy
a-a HP 1.19 Build. 40 (9) (2008) 1747–1754, https://doi.org/10.1016/j.
enbuild.2008.03.002.
[24] P. Penna, A. Prada, F. Cappelletti, A. Gasparella, Multi-objectives optimization
of Energy Efficiency Measures in existing buildings, Energy Build. 95 (2015)
57–69, https://doi.org/10.1016/j.enbuild.2014.11.003.
[25] F. Ascione, N. Bianco, R.F. De Masi, G.M. Mauro, G.P. Vanoli, Energy retrofit of
References educational buildings: transient energy simulations, model calibration and
multi-objective optimization towards nearly zero-energy performance, Energy
[1] J. Rosenow, R. Cowart, E. Bayer, M. Fabbri, Assessing the European Union’s Build. 144 (2017) 303–319, https://doi.org/10.1016/j.enbuild.2017.03.056.
Energy Efficiency Policy: Will the Winter Package deliver on ‘‘Efficiency First”?, [26] M. Hamdy, A. Hasan, K. Siren, A multi-stage optimization method for cost-
Energy Res Soc. Sci. 26 (2017) 72–79. optimal and nearly-zero-energy building solutions in line with the EPBD-
[2] UN Environment and IEA, Towards a zero-emission, efficient, and resilient recast 2010, Energy Build. 56 (2013) 189–203, https://doi.org/10.1016/j.
buildings and construction sector. Global status report 2017, 2017. enbuild.2012.08.023.
www.globalabc.org (accessed March 1, 2019). [27] F. Rosso, V. Ciancio, J. Dell’Olmo, F. Salata, Multi-objective optimization of
[3] L. Gynther, B. Lapillone, K. Pollier, Energy efficiency trends and policies in the building retrofit in the Mediterranean climate by means of genetic algorithm
household and tertiary sectors, An analysis based on the ODYSSEE and MURE application, Energy Build. 216 (2020) 109945, https://doi.org/10.1016/j.
databases, 2015. http://www.odyssee-mure.eu/publications/br/energy- enbuild.2020.109945.
efficiency-trends-policies-buildings.pdf. [28] M. Hamdy, A.-T. Nguyen, J.L. Hensen, A performance comparison of multi-
[4] Eurostat, Energy data 2020 edition, 2020. https://ec.europa.eu/ objective optimization algorithms for solving nearly-zero-energy-building
eurostat/documents/3217494/11099022/KS-HB-20-001-EN-N.pdf/bf891880- design problems, Energy Build. 121 (2016) 57–71, https://doi.org/10.1016/j.
1e3e-b4ba-0061-19810ebf2c64%0Ahttps://ec.europa.eu/eurostat/fr/web/ enbuild.2016.03.035.
products-statistical-books/-/KS-HB-20-001?inheritRedirect=true&redirect=% [29] European Commission, EU energy in figures. Statistical pocketbook., 2016.
2Feurostat%2Ffr%2Fpublicati. doi:10.2832/91509.
[5] BPIE, Europe’s Buildings Under the Microscope: A country-by-country review [30] Hellenic Statistical Authority, Press Release - Survey on Energy Consumption
of the energy performance of buildings, Buildings Performance Institute in Households, 2013. https://www.statistics.gr/en/statistics?p_p_
Europe, 2011. id=documents_WAR_publicationsportlet_INSTANCE_qDQ8fBKKo4lN&p_p_
[6] IEA, Technology Roadmap - Energy Efficient Building Envelopes, Paris, 2013. lifecycle=2&p_p_state=normal&p_p_mode=view&p_
https://www.iea.org/reports/technology-roadmap-energy-efficient-building- p_cacheability=cacheLevelPage&p_p_col_id=column-2&p_p_col_count=4&p_
envelopes. p_col_pos=1&_documents_WAR_publicat.
[7] European Commission, The European Green Deal, Eur. Comm. 53 (2019) 24. [31] C.A. Balaras, E.G. Dascalaki, K.G. Droutsa, S. Kontoyiannidis, Empirical
https://ec.europa.eu/info/sites/info/files/european-green-deal- assessment of calculated and actual heating energy use in Hellenic
communication_en.pdf. residential buildings, Appl. Energy. 164 (2016) 115–132, https://doi.org/
[8] EU, Directive (EU) 2018/844 of the European Parliament and of the Counsil of 10.1016/j.apenergy.2015.11.027.
30 May 2018 amending Directive 2010/31/EU on the energy performance of [32] D.I. Kolaitis, E. Malliotakis, D.A. Kontogeorgos, I. Mandilaras, D.I. Katsourinis,
buildings and Directive 2012/27/EU on energy efficiency, Off. J. Eur. Union. M.A. Founti, Comparative assessment of internal and external thermal
(2018) 75–91. insulation systems for energy efficient retrofitting of residential buildings,
[9] Eurostat, Distribution of the population by type of dwelling, 2018, (2018). Energy Build. 64 (2013) 123–131, https://doi.org/10.1016/j.
https://ec.europa.eu/eurostat/statistics-explained/index.php?title=File: enbuild.2013.04.004.
Distribution_of_the_population_by_type_of_dwelling,_2018_(%25)_LCIE20. [33] P. Pallis, N. Gkonis, E. Varvagiannis, K. Braimakis, S. Karellas, M. Katsaros, P.
png (accessed March 5, 2021). Vourliotis, Cost effectiveness assessment and beyond: a study on energy
[10] L. De Boeck, S. Verbeke, A. Audenaert, L. De Mesmaeker, Improving the energy efficiency interventions in Greek residential building stock, Energy Build. 182
performance of residential buildings: a literature review, Renew. Sustain. (2019) 1–18, https://doi.org/10.1016/j.enbuild.2018.10.024.
Energy Rev. 52 (2015) 960–975, https://doi.org/10.1016/j.rser.2015.07.037. [34] E.G. Dascalaki, C.A. Balaras, S. Kontoyiannidis, K.G. Droutsa, Modeling energy
[11] V. Pareto, Cours d’économie politique, F. Rouge, Lausanne, Switzerland, 1896. refurbishment scenarios for the Hellenic residential building stock towards the
[12] A.-T. Nguyen, S. Reiter, P. Rigo, A review on simulation-based optimization 2020 & 2030 targets, Energy Build. 132 (2020) 74–90, https://doi.org/10.1016/
methods applied to building performance analysis, Appl. Energy 113 (2014) j.enbuild.2016.06.003.
1043–1058, https://doi.org/10.1016/j.apenergy.2013.08.061. [35] F. Ascione, N. Bianco, C. De Stasio, G.M. Mauro, G.P. Vanoli, A new methodology
[13] R. Evins, A review of computational optimisation methods applied to for cost-optimal analysis by means of the multi-objective optimization of
sustainable building design, Renew. Sustain. Energy Rev. 22 (2013) 230–245, building energy performance, Energy Build. 88 (2015) 78–90, https://doi.org/
https://doi.org/10.1016/j.rser.2013.02.004. 10.1016/j.enbuild.2014.11.058.
[14] V. Machairas, A. Tsangrassoulis, K. Axarli, Algorithms for optimization of [36] B.M. Adams, L.E. Bauman, W.J. Bohnhoff, K.R. Dalbey, J.P. Eddy, M.S. Ebeida, M.
building design: a review, Renew. Sustain. Energy Rev. 31 (2014) 101–112, S. Eldred, P.D. Hough, K.T. Hu, J.D. Jakeman, L.P. Swiler, J.A. Stephens, D.M. Vigil,
https://doi.org/10.1016/j.rser.2013.11.036. T.M. Wildey, Dakota, A Multilevel Parallel Object-Oriented Framework for
[15] M. Wetter, J. Wright, Comparison of a generalized pattern search and a genetic Design Optimization, Parameter Estimation, Uncertainty Quantification, and
algorithm optimisation method, in: Eighth Int. IBPSA Conf., Eindhoven, Sensitivity Analysis: Version 6.0 User’s Manual, (2014). http://prod.sandia.gov/
Netherlands, 2003. http://www.inive.org/members_area/medias/pdf/Inive/ techlib/access-control.cgi/2014/144633.pdf (accessed December 9, 2017).
IBPSA/UFSC89.pdf (accessed December 7, 2017). [37] Solar Energy Laboratory, TRNSYS 18: a TRaNsient SYstem Simulation program,
[16] S. Attia, M. Hamdy, W. O’Brien, S. Carlucci, Assessing gaps and needs for (2017). http://sel.me.wisc.edu/trnsys (accessed March 5, 2018).
integrating building performance optimization tools in net zero energy [38] T. Murata, H. Ishibuchi, MOGA: Multi-objective genetic algorithms, in: Proc.
buildings design, Energy Build. 60 (2013) 110–124, https://doi.org/10.1016/j. 1995 IEEE Int. Conf. Evol. Comput., Perth, Australia, 1995: pp. 289–294.
enbuild.2013.01.016. [39] HSA, 2011 Population and Housing Census: Characteristics and amenities of
[17] M. Ferrara, A. Rolfo, F. Prunotto, E. Fabrizio, EDeSSOpt – energy demand and dwellings , 2014.
supply simultaneous optimization for cost-optimized design: application to a [40] TABULA, IEE Project TABULA, (2017). https://webtool.building-typology.eu/
multi-family building, Appl. Energy 236 (2019) 1231–1248, https://doi.org/ #bm (accessed June 19, 2017).
10.1016/j.apenergy.2018.12.043. [41] Google Maps, (2018). https://www.google.com.au/maps/@-37.7996526,144.
[18] M. Morelli, L. Rønby, S.E. Mikkelsen, M.G. Minzari, T. Kildemoes, H.M. 9577026,15z (accessed September 25, 2018).
Tommerup, Energy retrofitting of a typical old Danish multi-family building [42] Technical Chamber of Greece, Greek Regulation for the Energy Efficiency of
to a ‘‘nearly-zero” energy building based on experiences from a test Buildings- T.O.T.E.E. 20701-3/2010 - Guidelines on the Greek climatic data,
apartment, Energy Build. 54 (2012) 395–406, https://doi.org/10.1016/j. (2010).
enbuild.2012.07.046. [43] H.E. Beck, N.E. Zimmermann, T.R. Mcvicar, N. Vergopolan, A. Berg, E.F. Wood,
[19] M. Ferreira, M. Almeida, A. Rodrigues, S.M. Silva, Comparing cost-optimal and Data Descriptor: present and future Köppen-Geiger climate classification maps
net-zero energy targets in building retrofit, Build. Res. Inf. 44 (2) (2016) 188– at 1 -km resolution, Sientific Data 5 (2018) 1–12, https://doi.org/10.1038/
201, https://doi.org/10.1080/09613218.2014.975412. sdata.2018.214.

19
M. Panagiotidou, L. Aye and B. Rismanchi Energy & Buildings 252 (2021) 111385

[44] Technical Chamber of Greece, Greek Regulation for the Energy Efficiency of [59] T. Li, G. Shao, W. Zuo, S. Huang, Genetic algorithm for building optimization -
Buildings - T.O.T.E.E. 20701-1/2010 - Guidelines on the evaluation of the State-of-the-art survey, in: ACM Int. Conf. Proceeding Ser., 2017: pp. 205–210.
energy performance of buildings, (2010). doi:10.1145/3055635.3056591.
[45] D. Densley Tingley, A. Hathway, B. Davison, An environmental impact [60] B. Koffi, A. Cerutti, M. Duerr, A. Iancu, A. Kona, G. Janssens-Maenhout,
comparison of external wall insulation types, Build. Environ. 85 (2015) 182– Covenant of Mayors for Climate and Energy: Default emission factors for local
189, https://doi.org/10.1016/j.buildenv.2014.11.021. emission inventories, Publ. Off. Eur. Union. (2017), https://doi.org/10.2760/
[46] T. Lhendup, Development of an Inter-Seasonal Thermal Storage System, The 290197.
University of Melbourne, 2013. [61] P. Bertoldi, D. Bornas Cayuela, S. Monni, R. Piers De Raveschoot, How to
[47] U. Persson, S. Werner, Quantifying the Heating and Cooling Demand in develop a Sustainable Energy Action Plan (SEAP) - Guidebook., (2010).
Europe., 2015. https://heatroadmap.eu/wp-content/uploads/2018/09/ doi:10.2790/20638.
STRATEGO-WP2-Background-Report-4-Heat-Cold-Demands.pdf (accessed [62] Public Power Corporation, Residential non-Off-Peak Tariff (Residential Tariff) -
November 15, 2019). C1 | PPC S.A., (2018). https://www.dei.gr/en/oikiakoi-pelates/timologia/
[48] L. Schibuola, M. Scarpa, C. Tambani, Modelling of HVAC system components oikiako-timologio-xwris-xronoxrewsi-g1 (accessed January 23, 2019).
for building dynamic simulation Air-to-water on-off control., in: Proc. BS2013 [63] Eurostat, Natural gas price statistics - Statistics Explained, (2019). https://ec.
13th Conf. Int. Build. Perform. Simul. Assoc., Chambéry, France, 2013: pp. europa.eu/eurostat/statistics-explained/index.php?title=Natural_gas_price_
1103–1109. statistics#Natural_gas_prices_for_household_consumers (accessed October
[49] J. Cockroft, A. Samuel, P. Tuohy, Development of a methodology for the 28, 2019).
evaluation of domestic heating controls - Phase 2 of a DEFRA Market [64] POPEK, Fuel oil prices (GR), (2019). https://www.popek.gr/index.php/el/times-
Transformation Programme project, carried out under contract to BRE kafsimon/home (accessed October 28, 2019).
Environment, 2007. http://www.esru.strath.ac.uk/Documents/07/BRE_E302_ [65] Hellas pellet, Pellet price list, (2019). https://www.hellaspellet.gr/so-
040707.pdf. jaύrilo-pekkes/oi-silέ1-la1/ (accessed October 2, 2019).
[50] H.I. Henderson, D. Parker, Y.J. Huang, Improving DOE-2’s RESYS routine: User [66] D. Mazzeo, N. Matera, C. Cornaro, G. Oliveti, P. Romagnoni, L. De Santoli,
defined functions to provide more accurate part load energy use and humidity Energy & Buildings EnergyPlus , IDA ICE and TRNSYS predictive simulation
predictions, Proc. ACEEE Summer Study Energy Effic. Build. 1 (2000) 1113– accuracy for building thermal behaviour evaluation by using an experimental
1124. campaign in solar test boxes with and without a PCM module, 212 (2020).
[51] A.N. Celik, T. Muneer, P. Clarke, A review of installed solar photovoltaic and doi:10.1016/j.enbuild.2020.109812.
thermal collector capacities in relation to solar potential for the EU-15, Renew. [67] M. Panagiotidou, L. Aye, B. Rismanchi, Alternative heating and cooling systems
Energy. 34 (3) (2009) 849–856, https://doi.org/10.1016/j.renene.2008.05.025. for the retrofit of medium-rise residential buildings in Greece, Energy Technol.
[52] M. Panagiotidou, L. Aye, B. Rismanchi, Solar driven water heating systems for (2021), https://doi.org/10.1002/ente.202100377, In press.
medium-rise residential buildings in urban mediterranean areas, Renew. [68] Jarek Kurnitski, Arto Saari, Targo Kalamees, Mika Vuolle, Jouko Niemelä, Teet
Energy. 147 (2020) 556–569, https://doi.org/10.1016/j.renene.2019.09.020. Tark, Cost optimal and nearly zero (nZEB) energy performance calculations for
[53] C.F. Reinhart, O. Walkenhorst, Validation of dynamic RADIANCE-based residential buildings with REHVA definition for nZEB national implementation,
daylight simulations for a test office with external blinds, Energy Build. 33 Energy Build. 43 (11) (2011) 3279–3288, https://doi.org/10.1016/j.
(7) (2001) 683–697, https://doi.org/10.1016/S0378-7788(01)00058-5. enbuild.2011.08.033.
[54] R. Compagnon, Solar and daylight availability in the urban fabric, Energy Build. [69] G. Verbeeck, H. Hens, Energy savings in retrofitted dwellings: economically
36 (4) (2004) 321–328, https://doi.org/10.1016/j.enbuild.2004.01.009. viable?, Energy Build 37 (7) (2005) 747–754, https://doi.org/10.1016/j.
[55] National Printing House of Hellenic Republic, Installation of RES modules by enbuild.2004.10.003.
energy producers under net-metering system or virtual net-metering system [70] Ferdinando Salata, Iacopo Golasi, Umberto Domestico, Matteo Banditelli,
pursuant to the article 14Α sot Ν. 3468/2006., Gov. Gaz. Hell. Repub. (2017) Gianluigi Lo Basso, Benedetto Nastasi, Andrea de Lieto Vollaro, Heading
14641–14660. http://www.deddie.gr/Documents2/Fotovoltaika/FV virtual towards the nZEB through CHP+HP systems. A comparison between retrofit
metering/YA_ΑPΕΗK_Α_U1_oij175067-UΕΚ 1547Β_5.5.2017.pdf (accessed solutions able to increase the energy performance for the heating and
November 26, 2018). domestic hot water production in residential buildings, Energy Convers.
[56] European Commission, Guidelines accompanying Commission Delegated Manag. 138 (2017) 61–76.
Regulation (EU) No 244/2012 of 16 January 2012 supplementing Directive [71] Ariston, Genus Premium EVO: Installation Guide (GR), (2019).
2010/31/EU of the European Parliament and of the Council, Off. J. Eur. Union. [72] De Dietrich, GTU C 220 C 330 Floor standing oil-fired condensing boiler,
(2012) 1–28. https://publications.europa.eu/en/publication-detail/-/ (2019). http://www.dedietrich-heating.com/content/download/161690/
publication/b6d03fa4-b936-43f1-9120-10b027c04e3a/language-en. 2148087/version/1/file/GTUC_220_330_EN.pdf.
[57] S. Braungardt, W. Eichhammer, R. Elsland, T. Fleiter, M. Klobasa, M. Krail, B. [73] Mavil SA, Mavil Primus - Pellet/Wood boilers, (2015). http://mavil-sa.gr/wp-
Pfluger, M. Reuter, B. Schlomann, F. Sensfuss, S. Tariq, Study evaluating the content/uploads/2012/09/mavil-primus.pdf.
current energy efficiency policy framework in the EU and providing [74] Robur S.p.A., GAHP A: Installation , use and maintenance manual, (2019) 1–25.
orientation on policy options for realising the cost-effective energy- https://www.robur.com/downloads/1208/184/D-LBR548_Z_19MCMSDC024_
efficiency/saving potential until 2020 and beyond - Report on behalf of DG GAHP_A_AEE_DDC_EN.pdf.
ENER., 2014. http://www.isi.fraunhofer.de/isi-wAssets/docs/x/en/projects/ [75] Ariston, Nimbus Flex , Plus , Compact , Pocket M Net: Energy labels and
FinalReport_EvaluationEnergyEfficiency_2020_2030.pdf (accessed August 9, product fiche, (2019). https://www.ariston.com/ita/it/file/
2018). 8840/download?token=USlHBuHa.
[58] J. Steinbach, D. Staniaszek, Discount rates in energy system analysis, Build. [76] NREL, Laboratory Test Report for Fujitsu 12RLS and Mitsubishi FE12NA Mini-
Perform. Institue Eur. (2015). http://bpie.eu/uploads/lib/document/ Split Heat Pumps, (2011).
attachment/142/Discount_rates_in_energy_system-discussion_paper_2015_ [77] Mitsubishi , MSZ-HR R32 Classic Inverter Heat Pump , Mitsubishi ( 2020 ).
ISI_BPIE.pdf. http://library.mitsubishielectric.co.uk/pdf/book/MSZ-HR_R32_2020_PI_Sheet.

20