You are on page 1of 15

Energy & Buildings 249 (2021) 111253

Contents lists available at ScienceDirect

Energy & Buildings


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

In search of optimal consumption: A review of causes and solutions to


the Energy Performance Gap in residential buildings
Stefano Cozza a,⇑, Jonathan Chambers a, Arianna Brambilla b, Martin K. Patel a
a
Energy Efficiency Group, Institute for Environmental Sciences (ISE) – University of Geneva, Switzerland
b
School of Architecture, Design and Planning – The University of Sydney, Australia

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

Article history: The assessment of building performance through energy certificates is important for tracking and
Received 1 February 2021 improving the energy efficiency of the building stock. The reliability of these assessments is critical for
Revised 20 April 2021 achieving future energy targets. However, there is evidence of a significant Energy Performance Gap
Accepted 3 July 2021
(EPG) in buildings, defined as the difference between measured and calculated energy consumption.
Available online 7 July 2021
This work performs a systematic review of EPG causes and reduction strategies in the context of heating
of residential buildings. It introduces the concept of ‘‘optimal” consumption, in contrast to ‘‘theoretical”
Keywords:
(i.e. calculated with standards) and ‘‘actual” (i.e. measured) consumption, which enables a more rigorous
Energy certificates
Governance improvement
classification of causes and potential solutions to the EPG. This review found that inaccuracies in mod-
Measured consumption elling of building characteristics and occupant behaviour has been most studied by researchers. It found
Theoretical consumption that many EPG reduction strategies have been proposed, which can be categorized into two groups. The
Environmental target first aims to improve the energy consumption calculation by correcting the standard assumptions and/or
considering new approaches to create energy certificates. The second group focuses on improving the
actual performance of the building’s energy systems, through better monitoring, maintenance, and gen-
eral usage of the building. A range of practical strategies were identified, which are relevant to a range of
stakeholder groups. At the same time, this work also highlights that understanding of the relative impor-
tance of EPG causes and the potential impact of the corresponding solutions is incomplete.
Ó 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2. Aim and scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1. Systematic review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2. The concept of optimal consumption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3. Causes of the EPG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.1. Theoretical consumption deviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.1.1. Inaccuracy of inputs and assumptions for building modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.1.2. Inaccuracy of climate data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.1.3. Inaccuracy of occupant behaviour modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2. Actual consumption deviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2.1. Malfunctioning equipment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2.2. Measurement systems limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2.3. Execution of the work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2.4. Non-optimal use of the building by the occupant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.3. Assessing the causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

⇑ Corresponding author.
E-mail address: stefano.cozza@unige.ch (S. Cozza).

https://doi.org/10.1016/j.enbuild.2021.111253
0378-7788/Ó 2021 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

4. Solutions to address the EPG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8


4.1. Improvements to theoretical calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1.1. Approaches to energy labelling and calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1.2. Accuracy of standard values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1.3. Additional considerations for retrofitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.2. Improving actual consumption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.2.1. Technical factors to improve actual consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.2.2. Monitoring practices and continuous iteration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.2.3. Collaboration between stakeholders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.3. Solutions at different building life stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
5. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Declaration of Competing Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1. Introduction 1.2. Aim and scope

1.1. Background Energy performance certificates certify the design more than
the actual building, relying on standards and assumptions to com-
Improving building performance is key to tackling the chal- pare buildings in different local contexts [19]. However, energy
lenges of climate change, considering that buildings account for performance certificates are frequently used in many applications
40% of the global energy consumption [1]. The urgent need for in an incorrect manner (e.g. to predict the savings of an energy ret-
energy efficiency improvement calls for the implementation of rofit [20,21]), under the assumption that they are a reliable indica-
suitable policy instruments, including performance certificates tor of the buildings’ measured energy consumption. This has
[2]. The European Union’s Energy Performance of Buildings Direc- resulted in observations of differences between the theoretical
tive (EPBD) [3] establishes general energy performance certificates energy consumption according to certificates and standards on
calculation requirements, while implementation details differ the one hand and the actual energy consumption on the other,
between countries. Performance certificates were introduced to i.e. the EPG. This work reviews research identifying the causes of
raise awareness of building performance among owner and tenants the EPG and synthesises research findings that give solutions to
and to motivate performance improvements, as well as to ensure reduce the EPG.
that the performance of the building stock meets national require- To enable a rigorous classification of causes and potential solu-
ments and complies with broader energy policy [4]. tions to the EPG we introduce the concept of ‘‘optimal” consump-
Energy and climate policy targets for the building sector are fre- tion, in contrast to ‘‘theoretical” (i.e. calculated with standards) and
quently based on the evaluation and improvement of building ‘‘actual” (i.e. measured) consumption. This definition is developed
energy performance through certification schemes [5–7]. It is in Section 2.2 and is essential for identifying which aspect of the
therefore critical to determine to what extent these can be consid- energy performance and certification process should be targeted
ered reliable tools for fulfilling this purpose [8]. The Energy Perfor- by the various solutions.
mance Gap (EPG), generally defined as the difference between We apply systematic review methods, considering only the
expected energy consumption calculated by a building perfor- research on space heating of residential buildings in Europe. We
mance assessment and the actual consumption, is a significant focus on a) studies that identify causes of the EPG and b) studies
challenge to the achievement of energy efficiency targets [9–11]. on practical solutions to reduce the EPG. These works are cate-
Although the concept of the EPG is broadly accepted, its interpre- gorised following the optimal consumption concept definition.
tation and implementation differs considerably between authors This classification approach allows us to synthesize the common
[12]. While there is a range of literature on the EPG in buildings, findings between research coming from different European coun-
there is only limited consistency between different works in terms tries into a single conceptual framework. From this review, we pre-
of definitions and methodologies [13,14]. The lack of consistency in sent concrete strategies which may be applied for different
definitions reflects the challenge to find suitable solutions to close stakeholders and at different stages in building construction, reno-
the EPG. Some solutions are in fact addressed to different stake- vation, and use. These are highly relevant to construction practi-
holders, in different contexts, for different building types, and for tioners and policy makers.
the different design phases of a building. Therefore, not all of them
can be readily implemented in the context of energy performance 2. Method
certificates in Europe. Indeed, previous extensive review in Europe
addressed only non-residential buildings [15], while a review 2.1. Systematic review
which also included residential buildings considered all regions
of the world and did not target the distinctive and specific nature We apply a systematic review method to investigate causes of
of the EPG identified by the use of EU energy certificates/standards and solutions to the EPG, composed of three steps as follows [22].
[16]. This review addresses this gap in the existing research by (1) Collection of relevant publications: We considered publica-
proposing solutions for closing the EPG in the context of energy tions including journal articles, conference papers, reports, build-
certification practice over the buildings’ life. Finally, although the ing standards, and national guidelines. An initial list of four
EPG has been investigated in numerous publications, little work review publications on the EPG was identified [15–17,23] as a basis
to date has aimed to combine the findings and to propose a consis- to generate keywords for electronic searches. The search has been
tent strategy, in spite of the growing interest in the subject [17,18]. directed using the keywords ‘‘energy performance gap”, ‘‘measured
consumption”, ‘‘building actual consumption” that are included in

2
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

Fig. 1. PRISMA flow diagram of review process indicating the number of publications included in the analysis at each stage following filtering based on the exclusion criteria,
from the initial identification through the screening and eligibility steps.

use considered (i.e. space heating, domestic hot water, . . .); EPG
size; EPG causes; and approaches for reducing the EPG.
(3) Content synthesis: After analysing relevant publications, the
underlying causes and solutions of the EPG were identified and
grouped. Future research and policy directions were identified to
make the energy performance certificates a more robust and com-
prehensive method in building energy assessment.
To conclude, Fig. 1 describes how the records identified through
the searches were processed for this review. A total of 160 publica-
tions were included in the analysis, divided in: 108 Journal articles;
20 Conference papers; 20 Reports; 9 Standards and directives; and
3 PhD theses.
Fig. 2. Overview of the optimal consumption. The EPG is the sum of the theoretical
and actual deviations relative to the optimal consumption. The first is defined as the
difference between the optimal and the theoretical consumption, the second is the 2.2. The concept of optimal consumption
difference between the optimal and actual consumption.
Building energy performance is influenced by several intercor-
related factors, making it particularly challenging to identify and
the article title, abstract, and keywords, by year and country. Fur-
quantify each individual contribution to the EPG. In this review
thermore, publications within the reference lists with titles fitting
we introduce a new concept to classify the causes and solutions
the search criteria were also reviewed. The keyword search was
of the EPG: the optimal consumption (see Fig. 2). The optimal
conducted using Google Scholar and Scopus databases. All the jour-
consumption is the energy consumption when the building is per-
nal or conference papers were required to meet the following cri-
forming in an ideal way with respect to its intended design and
teria: peer-reviewed publications; focused on Europe; relevant to
construction quality, while satisfying the reasonable needs of its
the topic; focus on recent research findings. Publications were fil-
occupants. These conditions are usually different from both the
tered according to the inclusion criteria outlined above on the basis
theoretical ones defined by the standards and the actual ones.
of title and abstract.
We introduce this concept because there is a lack of clarity in the
(2) Content analysis: The publications selected were reviewed
literature about whether the causes for the EPG should be ascribed
with regards to nine aspects: building type; number of case stud-
to problems in the certificate calculations or problems with con-
ies; origin of the dataset; method applied to calculate the theoret-
struction or use of the building (often, albeit incorrectly, described
ical consumption; data used for the actual consumption; energy
as ‘occupant behaviour’). While it is trivial to note that the reality
can be a mixture of all of those, there is a need to introduce the
3
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

Fig. 3. Classification of the Energy Performance Gap’s principal causes, categorised into theoretical and actual deviation.

optimal consumption concept to allow a clear organisation of find- 3. Causes of the EPG
ings. Following this approach, the EPG can then be broken down
into I) the difference between theoretical consumption and opti- By introducing the concept of optimal consumption, we propose
mal consumption and II) the difference between actual consump- a new, clearer classification of causes of the EPG, as shown in the
tion and optimal consumption. The first difference is associated Fig. 3. In this section we separately discuss the causes found for
with the limitations of using a theoretical model to describe the the deviation between I) the optimal consumption and theoretical
real conditions of use of a building, while the second difference consumption and II) the optimal consumption and actual
is associated with malfunctioning and unrecommended use of consumption.
the building. The sum of the two, as shown in Fig. 2, gives the
EPG - the difference between actual and theoretical consumption. 3.1. Theoretical consumption deviation
A common approach to reduce the EPG is to perform field mea-
surements to obtain values for model parameters based on actual Building performance certificates rely on simplified building
operating conditions, in order to improve the prediction of actual energy models for determining the theoretical energy use [30].
energy use ; thereby reducing the level of uncertainty in the results These models are based on standards assuming uniform conditions
[12]. This kind of reasoning, however, is based on current stan- and allowing the comparison of buildings which differ significantly
dards, both normative (e.g. ventilation rate) and in terms of expec- [31]. However, these models are subject to simplification of reality,
tations of the actual use of a building (e.g. comfort level), which are overlooking the complexity underlying building energy use [32].
very difficult to modify. In order not only to close the EPG, but also This section focuses on reasons why a simplified building energy
to significantly reduce the consumption, it is important to consider model (as used to determine theoretical consumption according
the level of service that buildings should provide [24]. This is the to the building energy certificate) may not describe the energy
role of the optimal parameters: to challenge both theoretical and consumption of a building as accurately as it should.
actual values, in order to define the optimal consumption for each
building whereby efficiency and comfort needs are reached bearing
3.1.1. Inaccuracy of inputs and assumptions for building modelling
in mind realistic building usage.
As basis for certificates, usually simplified ‘‘white-box” models
The below is an example that can further clarify the concept of
on the building’s thermal balance are used [33]. White-box models
optimal consumption. Consider the three types of consumption
require detailed physical input data to predict the energy con-
(i.e. theoretical, optimal, and actual) as a function of indoor tem-
sumption of a building based on standard thermodynamic equa-
perature. The calculation of the theoretical energy consumption
tions, and their accuracy is determined by the correctness of the
assumes an indoor temperature of 20 °C according to Swiss
physical model and the quality of the inputs [34]. There is a general
national standards [25], while Flourentzou et al. [26] report an
consensus in the literature as to which of the building model
observed value of around 23.5 °C, resulting in a correspondingly
parameters are most commonly based on assumptions that have
higher actual energy consumption. Since an indoor temperature
a significant impact on energy consumption [35]. These are: the
of 20 °C is typically considered as uncomfortably cold in central
mean indoor temperature [36,37], the building elements’ assumed
Europe [23,27], Flourentzou et al. [26] propose 21.5 °C as the ‘‘op-
U-values [38,39], and the ventilation rate [40,41]. Indoor tempera-
timal” indoor temperature for which they report no significant
ture has been shown to have a large impact on energy consump-
complaints about cold/discomfort by the occupants, while the
tion estimates and therefore on the EPG [36]. Real indoor
energy consumption is not increased excessively. We refer to the
temperature are typically higher than required by the various stan-
corresponding energy consumption as optimal (Fig. 2). Another
dards [42,43], e.g. with values of 23 °C measured for Germany [44]
example is the efficiency of the heating system, such as the coeffi-
or the 22 °C in Denmark [45]. One study estimated that an increase
cient of performance (COP) of an air source heat pump. In this case
in room temperature by 1 °C resulted in approximately 5% higher
the theoretical energy consumption may assume a COP of 4.5 [28],
energy consumption [46]. A comprehensive review on the relation-
while an ‘‘optimal” value may amount to 3.0, based on empirical
ship between indoor temperature and building energy consump-
evidence of the real-world performance of a well-maintained heat
tion can be found in [47].
pump [26]. In contrast, the actual COP may only reach a value of
In new constructions, U-values are calculated in the design
2.0 due to poor installation or maintenance of the heat pump
phase and therefore variations during the building process will
[29]. The difference in energy consumption under theoretical as
add uncertainty to the actual U-values [48]. In existing buildings
opposed to actual conditions represents the EPG.
on the other hand, typical values for the period of construction
4
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

and the type of building are used. In both cases (also in carefully what was planned in the design of the building [72], and usually
performed U-value calculations in which these are corrected for it varies over time [73]. In this section we refer only to differences
thermal bridges) the uncertainty of the theoretically calculated in occupant behaviour between standard conditions, as repre-
U-value and the corresponding actual U-value is very high [49]. sented by the model, and optimal conditions in the sense of accu-
Research in UK statistically analysed (along with a review of ther- rately modelled conditions [74]. A major challenge is the
mal imaging data) this difference revealing widespread deviations representation of thermal comfort with current models (e.g. pre-
from the design intent across the majority of dwellings [50]. dicted mean vote and adaptive models), which still have many lim-
Another study showed a 2.5% difference in the heating consump- itations and are largely inadequate for predicting individual
tion for each building element (e.g. window) for which the U- responses [75]. Other examples are the number of people per
value was poorly assessed [51]. metre square or the daily presence in the building [76]. Sec-
Defining a standard value for the ventilation of the building has tion 3.2.4 deals with the non-optimal use of the building by the
also been found to be very complex [40]. The value for natural ven- occupants (e.g. overheating of rooms). In their review, Zou et al.
tilation is constantly changing due to the weather, with a large [16] listed the numerous factors that can affect the occupant beha-
impact on the heating consumption [51]. In one study it was viour, including occupants’ characteristics (e.g. lifestyle, cultural
shown that an increase rise in ventilation rate by 0.2 m3/(m2h) background, geographical origin, and gender [77]), the interaction
leads to an increase of the heating demand by about 5 kWh/ between occupants, their cost-consciousness, and occupants’ com-
(m2y) [52]. A large part of the prediction error is due to the free- fort. Other authors pointed out the importance of the building
dom of choice between default and measured values as input for characteristics (e.g. windows orientation) on the use made of it
the calculation, especially for the air tightness of the building by its occupants [78], occupants’ knowledge [79], and occupants’
envelope [53]. experience with the building technical system [80].
It has also been shown that current static calculations on the
energy consumption of buildings with high thermal mass lead to 3.2. Actual consumption deviation
different results than those measured [54], especially during inter-
mittent occupancy [55]. In standard calculations, the importance of The optimal energy performance of a building relies on certain
the dynamic thermal properties is neglected or underestimated design assumptions, such as set-point temperature, control sched-
[56], leading to up to 26% of the variation in theoretical predictions ules, and performance of the technical systems. In reality, many of
[57]. Inaccuracies in the geometrical representation of the building these assumptions do not apply due to malfunctioning of the
in the model (i.e. dimensions of its internal and external spaces), building systems, non-optimal use by occupants, or other reasons,
have been indicated as another important cause of the EPG contributing to the EPG. In Section 3.1 we discussed why optimal
[36,58]. Research found a discrepancy in the buildings’ surface area values are different from theoretical values used in standards,
overestimating the theoretical area of 8–10%, which impacts the now we will investigate what generates the deviation of actual val-
resulting energy calculation [59]. To conclude, because there is ues from optimal ones (Fig. 3).
not a clear trend found for the errors, the issues with these param-
eters are not easy to resolve. Strong variations are indeed noted 3.2.1. Malfunctioning equipment
between countries and across different buildings types. Malfunctioning equipment is a common issue, where the differ-
ence between the theoretical and actual consumption is no longer
3.1.2. Inaccuracy of climate data caused by problems in modelling or data collection, but by a real
Building energy models have been shown to be especially sen- technical problem in the building [16,81]. Heating systems in par-
sitive to the weather data used [60]. Several studies showed the ticular are frequently identified as the cause of an EPG [44,82]. The
substantial performance difference depending on the weather data examples of technical problems in literature are numerous and
(up to 40% in [61]), and the need to use climate adjusted data for an diverse, highlighting the critical importance of monitoring and
energy certificate [62–64]. The main challenges in the generation maintenance [83]. Reimann et al. [82] demonstrated that issues
of weather data include how to deal with uncertainty, urban heat with these systems can have a much larger influence on energy
islands, climate change, and extreme events [65]. An energy label- use than user behaviour. Building HVAC systems, especially for
ling simulation of a residential building in seven different cities innovative systems (e.g. smart building controls), often suffer from
was performed with the Norwegian certification tool and the stan- control systems malfunctions and lack of fine-tuning of control
dard weather file, and gave a D-label as result [66], while when strategies [84]. Fine-tuning may be omitted by the project devel-
using typical reference years from the local site the energy label oper but even if performed, the results may be disappointing
for three cities was upgraded to a C. because proper execution is time-intensive and implementation
Current standard weather files for building models are not sui- according to state of the art may conflict with requests made by
ted to the assessment of the potential impacts of a changing cli- occupants [85]. The absence of a facility manager (typically only
mate [67,68] which impacts the EPG [61]. Kočí et al. [63] present in larger buildings) has also been found to be an important
investigated the effect of warming trends on the calculated energy cause of the EPG [16]. Poor building operation practise has been
demand and found the 2013–2017 warming trend reduced average found to increase energy use by 50–80% relative to baseline, while
heating demand by 4% relative to the Test Reference Year temper- good practice may save 15–30%. However, this is contingent on
atures used for Czech Republic. As a consequence, the current Test managers having access to information and tools to apply optimi-
Reference Year for Prague, (and likely for other European cities sation – without these, expectations for optimisation are less likely
[69]), no longer reflects the actual weather conditions. In summary, to be met [86]. Before concluding, it is important to mention that in
energy certificates currently have a systematic problem in terms of the event of a malfunctioning equipment, the impact on the energy
not using the best locally adjusted climate data that represents the consumption, and therefore on the EPG, could also be positive (i.e.
local conditions, including the effects of climate change. a lower actual consumption due to a partial failure of the heating
system) [37].
3.1.3. Inaccuracy of occupant behaviour modelling
Even when a building is correctly designed and used, occupant 3.2.2. Measurement systems limitations
behaviour characterisation is a major challenge in building mod- Monitoring systems, while critical to identifying performance
elling [70,71]. The actual use by the inhabitants is different from problems, can themselves be the source of inaccuracies [87]. Incor-
5
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

the design phase, resulting in a deviations from the design specifi-


cations, especially with regard to insulation and air-tightness [89].
Construction quality depends on several factors [16], such as lack
of expertise on-site that may, for example, cause unexpected ther-
mal bridges [92]. Bell et al. [93] showed that an architect or project
developer with limited experience often fails to specify all the
details of the project, leaving important decisions to the builder
which results in a gap between designed and as-built. Van Dronke-
laar et al. [15] listed the most common issues related to on-site
workmanship for non-residential buildings which can easily be
extended for residential buildings: these include the imperfect
eaves to wall junction insulation, the improper installation of drai-
nage, air ducts and electrical pipe work, and the incorrect position-
ing of windows and doors, that reduce the actual performance of
the thermal envelope.
An important factor is the general lack of attention in the pro-
Fig. 4. Number of references analysed in this study, categorised by their main topic curement of materials and construction contractors (and sub-
(s). contractors) – to the extent that this is often not even considered
as a distinct phase of the design and building process [94]. This
is important as all the practical implementation of designs and
rect installation or malfunction of monitoring systems leads to
technical solutions ultimately rests with the construction contrac-
errors in the interpretation of the building energy performance
tors. Finally, it is important for the original design to both meet the
[88]. Metered energy use obtained from measurement systems
energy efficiency targets, and for it to be actually possible to build.
needs to be validated to ensure accuracy of the data [15]. The most
If insufficient attention is paid to the executability of the construc-
common sources of errors identified are mislabelling (i.e. sensors
tion, it becomes likely that flaws in the final building will arise and
not measuring what it was thought they were measuring), incor-
cause performance problems [92].
rect installation [89], and lack of calibration of sensors [90]. For
certain types of measurement, for example measuring building ele-
ment U-values, the need for a qualified expert using sufficiently 3.2.4. Non-optimal use of the building by the occupant
capable measurement devices has been highlighted [91]. The way occupants operate the building in regard to opening
windows, handling controls or shading, lighting, and indoor set
point temperature, significantly affects the EPG [95,96]. Contrary
3.2.3. Execution of the work to Section 3.1.3, this section is not concerned with inaccurate mod-
The building energy performance is greatly influenced by the elling of the occupant behaviour but instead discusses non-optimal
quality of construction, which is difficult to account for during use of the building by the occupants. A clear example is the contra-

Table 1
Causes of the EPG. Cause ID is used to link causes with solutions given in Table 2. *The number reported in each box represents the number of times that cause was indicated as
the cause of the EPG.

Deviation (N*) Macro-causes (N*) Causes description (N*) Cause


ID
Deviation of theoretical consumption (272) Inaccuracy of inputs and (149) Inappropriate initial standard assumptions (61) 1
relative to optimal consumption assumptions for building modelling Inappropriate modelling of building element (42) 2
(façade, roof)
Inaccurate representation of the exact geometry (13) 3
of the building
Incompleteness and inaccuracy of available (21) 4
project information (for retrofit)
Uncertainty in prediction of the actual use of the (12) 5
building
Inaccuracy of climate data (30) Uncertainty in definition of the actual external (24) 6
environmental conditions
Uncertainty in definition of future weather (6) 7
scenarios
Inaccuracy of occupant behaviour (93) Uniform assumptions for occupant behaviour in (47) 8
modelling the standards
Variability in comfort needs of the occupants (46) 9
across the population
Deviation of actual consumption relative (159) Malfunctioning equipment (56) Partial failure or malfunctioning of energy (19) 10
to optimal consumption systems
Lack of fine-tuning during the operation stage (26) 11
Controls strategies/software do not work as (11) 12
predicted
Measurement system limitations (8) Poor quality of the measurement equipment (8) 13
Execution of the work (34) Poor quality of the building envelope (12) 14
Poor workmanship (15) 15
Lack of attention to executability of construction (7) 16
Non-optimal use of the building by (61) Occupants’ experience with the installed (26) 17
the occupants technology
Poor communication between the stakeholders (35) 18
(little information to the occupant)

6
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

Fig. 5. Treemap visualization of the causes of EPG. Each cause is represented using rectangles, the dimensions of which are calculated using the number of times that cause
was indicated as the cause of the EPG.

Fig. 6. Classification of the Energy Performance Gap’s principal solutions identified in the literature. They allow to reduce the deviation between the theoretical or/and actual
consumption on the one hand and the optimal consumption on the other.

dictory use of the thermostatic valves set at their maximum, while lated [100]. The problem identified through this review is that
simultaneously keeping the windows open in the tilted position all most reviewed studies focus on one or two causes separately, but
day long [78], mainly to counteract the indoor overheating [26], never in combination. Furthermore, studies dealing with theoreti-
and causing an increase of actual consumption by up to a factor cal deviation only focus on calculation problems, and rarely worry
of two [97]. However, occupant behaviour and its impact on energy about the actual technical problems of the buildings, and vice
consumption is notoriously hard to predict [70]. In one study, man- versa. This is even more evident from Fig. 4, in which all the studies
ual operation of shading systems by users had the effect of reduc- analysed in this review are reported: 60.5% of the studies deal with
ing solar gains and increasing heating consumption [98], while in theoretical deviation, 25.5% treat only the technical problems of
another, manual shading operations instead increased solar heat the buildings, and only 14% deal with both causes of the EPG, those
gains and reduced heating consumption [99]. related to theoretical deviation and those related to actual
deviation.
Finally, making use of the insights of the literature review pre-
3.3. Assessing the causes sented above, we prepared a full list of all the causes found, which
have been classified in Table 1 according to the macro-causes
The combination of these causes can significantly influence defined previously. For each study, we counted the covered causes,
building energy performance. Moreover, it was found that the and reported them in Table 1, whereby the number reported rep-
causes of EPG are not independent, but are often strongly interre-
7
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

resents the times when that cause was indicated as the cause of the neighbouring properties have cavity walls, then the software could
EPG. ask for confirmation that the wall type is correct. A similar
As can be observed from Table 1, all these causes have been approach has also been applied to detect outliers in energy use pre-
identified with a different frequency in the literature and most dictions [105]. Finally, the work of Semple and Jenkins [106] high-
importantly, they contribute differently to both energy consump- lighted that there is still much to be done to integrate actual data
tion and EPG. These different contributions can be visualized in within the assessment methods used to generate performance cer-
Fig. 5, in which the dimension of each box represents the number tificates. They found that only 21% of the homes in the six largest
of times that cause was indicated as the cause of the EPG. European countries can use actual consumption to generate a
In conclusion, this literature review shows that it is challenging certificate.
to arrive at a ranking of the causes based on their influence on the This review highlights that improving the procedures for
EPG. Indeed, each of these macro-causes may be the main reason obtaining and validating certificates is necessary. It also highlights
for the EPG. However, while one of these occur, it is also possible how diverse the actors involved in gathering data reported in the
that two or more of these factors are jointly present. Literature certificates are, and how different the expectations are for each
does not offer quantitative evidence whether single-factor or mul- of them [107]. Furthermore, it would be useful to add a value for
tifactor causes are more frequent and if so, which ones. target (optimal) consumption for the building in operation, sepa-
rate from the value calculation used for labelling and compliance.
It has been suggested [108] to use a two sets of inputs for the cer-
4. Solutions to address the EPG
tification, one standardised for labelling and a second focused on
the potential for energy saving measures.
Several strategies can be adopted to reduce the EPG by optimiz-
ing the actual consumption or increasing the accuracy of the theo-
4.1.2. Accuracy of standard values
retical consumption assessment. We grouped those strategies into
It has been argued that national building performance certifi-
suggestions to I) improve the calculation of theoretical energy use
cates require increased data quality [39]. The variability of user
to better represent optimal consumption (Section 4.1), and II)
behaviour and the challenges to define it in building energy mod-
increase the actual performance of buildings to be closer to their
elling is a widely acknowledged issue [53]. To address this prob-
optimal consumption (Section 4.2). This grouping is illustrated in
lem, the use of large-scale occupant behaviour datasets, gathered
Fig. 6. Finally, we determined the phase of building projects where
through different data collection methods (e.g. questionnaires,
the different strategies should be applied and linked those with the
wearable devices), has been recommended to bring the theoretical
causes categorised previously with respect to the optimal con-
values closer to the optimal values [71].
sumption (Section 4.3).
Currently, information on the building characteristics is col-
lected from drawings and in-situ observations, and then calculated
4.1. Improvements to theoretical calculations according to appropriate standards [109]. However so far only
indirect approaches have been proposed to improve these calcula-
This section addresses solutions that aim to correct the theoret- tions, mainly through proposing to update standard values (e.g.
ical energy consumption calculation, as well as considering new standard U-value) [110], or by applying adjustment factors
approaches to facilitate the application of energy certificates. for groups of buildings (e.g. to the indoor temperature profiles)
[31]. It has been demonstrated that through the substitution of
4.1.1. Approaches to energy labelling and calculations the standard values in the model with values based on in-situ mea-
Even though the energy performance certificate is regarded as a surements of air tightness and U-value, it was possible to reduce
key instrument in promoting energy efficiency in the building sec- the EPG to 2.5% [35]. However, to improve the accuracy of the U-
tor, so far its impact is quite limited [4]. These schemes need to be values, a comprehensive fabric test including air permeability, U-
improved not only concerning theoretical calculations, but also values (heat flux) measurements, and thermal imaging surveys
more generally in their implementation modalities. Taranu and have been suggested [50]. It has also been proposed to replace
Verbeeck [101] highlight that the framing of the key messages the use of static U-value in certificates with an ‘‘effective U-
and of the information about the technical aspects of the certificate value” to quantify the dynamic performance of walls [55].
is an important aspect in the certificate implementation. They note Various authors have proposed increasing the indoor tempera-
that some certificates provide a confusing mix of information and ture heating set-point to 21.0 °C or 21.5 °C [26,52], instead of the
recommend that each EU state should define a priori what the pur- current 20 °C set in the national standard [42,111,112]. Further-
pose of the certificate is. For example, if the purpose of the certifi- more, while heating setpoint temperatures can vary greatly
cate is to encourage the owner to seek advice on energy efficiency between dwellings, research suggested that this variation is not
improvements, it is important that it clearly reports the current random [43]. Systematic variations were found according to occu-
and optimal potential energy performance of the building [102]. pancy profiles, household characteristics, motivation, and beha-
Several studies have been conducted with the aim of enhancing viour; this could be accounted for by modifying standard values
the quality of energy performance certificates. An algorithm was accordingly [113]. Therefore, the optimal values should also be dif-
developed for Spain’s certificates with the aim of improving the ferent depending on the building type and the respective occu-
baseline certificates according to building type and climate zones, pants. For semi-detached houses a higher set-point value (23 °C),
and it proved to reduce the EPG [103]. Another suggestion to while for semi-detached bungalows a lower one (20 °C), are sug-
improve the quality of the certificate is ‘‘smart auditing” [104]. gested compared to other dwelling types [43]. In newer houses,
To-date, energy certificates have been randomly audited, but a tar- constructed after 2007, an even lower indoor temperature set-
geted procedure could audit those that meet certain criteria (e.g. point (19 °C) should be assumed due to well-insulated building
having no gas heating systems declared, but a gas supply present) envelopes [51]. Using the household characteristics instead, in
or have been identified as having a discrepancy. UK for single parent a setpoint temperature of 18 °C was reported,
An alternative to smart auditing would be to allow the energy while for dwelling occupied by inhabitants rating their general
assessors to correct potential errors while the certificate is pre- health as ‘‘very bad”, a setpoint temperature of 20 °C was reported
pared, based on machine learning approach. For example, if a cer- [43]. In this sense, personal comfort models have also been pro-
tificate that is produced describes a solid wall property, but posed, which can be crucial for ‘‘generating accurate predictions of
8
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

individuals’ comfort requirements and closing the loop between occu- costs [123]. New methods for rapid and less intrusive performance
pants and heating systems” [75]. However, these models are not cur- assessment may help in this respect [124].
rently compatible with the approach to thermal comfort foreseen Finally, the role of economic subsidies that can be obtained
in energy standards/certificates. Certificates should be further through an energy certificate is very important. Given the high
developed to cater for individual difference in thermal comfort. capital cost of energy retrofitting it has been argued that, in order
Several studies suggested changing the ventilation rate to to be more effective, subsidies should be paid before the start of
reduce the EPG [52,114]. Conservative values were indeed used the work of energy retrofitting, and not after its completion, as is
in certificates to encourage a better thermal performance of the the case today [125]. This would avoid that the owner has to obtain
envelope (walls, floor, roof) [53]. However, most of the recent an interim loan from the bank which can act as further barrier
new buildings have become much more air-tight, calling for an [126]. At the same time, compliance with minimum energy perfor-
update of these values. Some authors have therefore proposed mance targets could be better guaranteed if part of the subsidy is
revised ventilation rates of 1 m3/(m2h) instead of 0.3 m3/(m2h) only paid once proper execution and functioning has been proven.
for forced ventilation [98], and 1.1 m3/(m2h) during use and To conclude, simplification and acceleration of legal procedures for
0.5 m3/(m2h) out of use instead of the default of 0.7 m3/(m2h) for energy retrofits may be required [127].
hybrid ventilation [37] in order to reduce the EPG. These values
are considered optimal because they take into account the air flow 4.2. Improving actual consumption
resulting from window opening (by counting the opened windows
several times during the day), which is not sufficiently considered This section focuses on solutions that aim to improve the actual
in the standards [26]. performance of the building’s energy systems, enhancing the
To address the lack of precision in the standard weather data design and construction phases, as well as driving better monitor-
used in the certificates, new tools that support the quantification ing, maintenance, and general use of the building.
of local and micro-climate conditions have been created [115].
Tools have also been introduced to generate weather files that take 4.2.1. Technical factors to improve actual consumption
climate change into account [63]. Jentsch et al. [116] developed a Most researchers and practitioners agree that adjusting build-
tool to integrate the future UK climate change scenarios into the ing technical systems (e.g. building controls, hydraulic balancing,
widely used Typical Meteorological Year file formats, using the etc.) is among the most effective ways to bring consumption closer
‘morphing’ methodology [117]. These were all found to reduce to optimum and reduce the EPG [128,129]. Continuous post-
the EPG. occupancy data collection and performance monitoring of the
While the various solutions to improving the accuracy of stan- building can ensure that optimal conditions are met under actual
dard values have been found to reduce the EPG, the synthesis of operating conditions [130] identifying deviations caused by the
these findings indicates that small changes to any of these values lack of fine-tuning during operation [131]. To solve this problem,
could significantly change the resulting energy consumption esti- sensors [132], advanced electricity meters [133], building monitor-
mate. This allows to close the EPG but without necessarily being ing system [134], Wi-Fi [135], and wireless camera networks [136]
a better representation of reality. Great care must be taken to avoid can be used for data collection. It is then essential that facility man-
over-fitting the theoretical model by performing arbitrary changes agers are given access to detailed data on energy consumption in
to key parameters without significant experimental evidence for buildings [135]. At the same time, the cost of monitoring needs
these changes. to be kept in mind. It will be necessary to develop and implement
cost-effective solutions which are likely to differ depending on the
4.1.3. Additional considerations for retrofitting type and the size of the building [137].
The EPG gives rise to uncertainty also when property owners The optimization of the HVAC systems tailored to the occu-
take decisions about energy retrofit [42]. European studies using pants’ needs is essential to reduce the EPG [85,138]. The function-
energy performance certificates found that the ratio of actual to ing and comfort provision of ventilation systems has been
expected energy savings after retrofit ranged from 40% to 60% highlighted as needing more attention during the design phase
[21,118]. Considering that standard inputs for calculations can be as well as improved monitoring [139]. In the context of building
up to 50% higher than the optimal inputs, unrealistic expectations retrofit, field measurements indicated that the ventilation effi-
of energy savings and profitability can be created if recommenda- ciency was significantly lower than assumed in the design phase
tions are based only on standardised calculations [119]. To reduce and that simple adjustment of the settings improved the overall
this problem, the study of Cozza et al. [20] indicated that a more performance and reduced the EPG [40]. In general, it has been
realistic assessment (3.6% difference) of real energy savings can observed that actively tuning and maintaining the building’s
be achieved by comparing the actual current consumption with energy systems, e.g. by hydraulic balancing of the heating system,
the expected theoretical consumption defined by the energy cer- adjusting heating profiles and temperatures, or correctly setting up
tificate after retrofit. Moreover, the energy certificate has to-date heat recovery system, can help to attain the optimal consumption
a limited influence on homeowners’ energy retrofit practices. Own- and can drastically reduce the EPG [133].
ers in general regard both comfort and energy saving as important
reasons when renovating [120]. Further developments to the per- 4.2.2. Monitoring practices and continuous iteration
formance certificate concept, such as Building Renovation Pass- The engineering and architectural knowledge exists to allow the
ports outlining a long-term building renovation history and construction of energy-efficient buildings and to conduct effective
future roadmap, may also help [121]. deep energy retrofitting. However, the legislative framework is
Another important aspect with respect to the certificate for ret- essential to enforce good practice in construction, commissioning
rofit is the mode of verifying retrofits. For existing buildings, an and operation. The EPBD has improved building energy efficiency
energy performance certificate is typically only required if a build- ratings in the EU [140] but it does not mandate monitoring actual
ing or a large building element is subject to major renovation [3]. consumption of the building. The 2010 EPBD update added the
To-date, these policies do not require proof of performance after possibility of inspection and monitoring of building HVAC systems
commissioning [122], i.e. existing policy does not currently take [141], but did not make this mandatory [142]. The 2018 EPBD [3]
the EPG into account. A part of the challenge is to achieve energy finally required Member States to ensure that residential buildings
policy goals without imposing excessive monitoring labour and are equipped with electronic monitoring. Nevertheless, there
9
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

remains significant opportunities to integrate measurement and out by Way and Bordass [149], ‘‘most designers and contractors have
verification in the EPBD requirements for building performance traditionally shown little interest in learning from how their buildings
monitoring [143]. A major limitation of the certification schemes actually perform in use; and most owners have certainly not wanted to
is the absence of performance monitoring once the certificate has pay them to do so”. It is precisely on this aspect that specific regu-
been issued [144]. Many authors indicate that continuous monitor- lations are needed. A solution is the implementation of a role (per-
ing should lead to optimal consumption and reduce the EPG son) responsible for implementation of the energy project [150], to
[57,92]. Monitoring should therefore be required for certification ensure that the energy objectives are met [23]. The core part of this
of buildings [44,145]. A positive experience has been made in this approach is the support to the inhabitants during the operating
respect by the Passive House institute, which requires monitoring phase, accompanied by an optimization phase after renovation,
of consumption, as well as occupant satisfaction and thermal com- which further reduces heating consumption and therefore the
fort, even after awarding their energy standard, with the aim of EPG [151].
reducing the EPG [146]. While the measures discussed above concern primarily multi-
It has been suggested that building designers or construction family buildings (and especially larger ones), the installation of
contractors should be responsible for verifying energy perfor- innovative technologies such as heat pumps in single-family build-
mance of completed buildings to identify over-consumption ings is generally less demanding. Accompanying measures may
resulting from poor construction [147]. This makes it possible to nevertheless be necessary in order to ensure optimal performance
hold contractors responsible for resolving problems in the project and to avoid a bad reputation of innovative technological solutions.
execution. As simple solutions, requirements for certification could As adequate approach, quality labels for recommendable installers
be the documentation of the progress of construction with pho- may be introduced or lists of accredited installers may be pub-
tographs as well as airtightness tests during the construction per- lished, helping homeowners to make a choice [120]. Good experi-
iod, as recommended by the checklist for planning and ence has been made with this approach in Switzerland [152].
implementation of Passive Houses [148]. However, as pointed
4.2.3. Collaboration between stakeholders
Several studies highlight the importance of collaboration and
Table 2
communication between stakeholders to reduce the EPG
Overview of the recommendations to reduce the EPG according to the stage of the [153,154]. However, this is complex due to the different objectives
project. Cause ID refers to the cause or set causes listed in Table 1. between stakeholders as well as their different knowledge and
Design phase Recommendation to reduce the EPG Cause
experience with technologies, and especially due to the different
ID definition of optimal consumption among the stakeholders [16].
It is important to increase the transparency in all the phases of
Initial Well defined expectations and resources available 1-5
assessment in order to set realistic objectives design and construction, from the administration to the architect,
Better defined the external environmental 2-3-6- from the energy utility to the inhabitant [155]. The objective is
conditions (weather data) and initial state of the 7 to share the knowledge gained in various projects with a larger
building
audience, informing all building occupants about responsible use
Take in account the experience learned from 1-8-9
previous similar projects
of the building systems [156]. Schröder et al. [112] clearly stated
Plan the maintenance of the building systems 10-11 that ‘‘the influence of individual user behaviour on the integral energy
Take in account the feedback and comments of 8-9- consumption will further rise, because the energy release into the liv-
the future occupants 17 ing environment gets more and more disconnected from the human
Design Unambiguous building plan/retrofit plan that can 1-4
perception”. This means that it becomes increasingly difficult for
be realistically implemented
Promote simple and reliable solutions, taking into 5-10- users to understand the energy impact of their own actions (e.g.
account possible future change of use of the 16 raise the thermostat by one degree). To address this, some work
building has applied Behaviour Change Programs using workshops, coach-
Clarify the responsibility and the role of each 18 ing, and feedback programs to promote energy saving behaviours
stakeholder
Employ only experienced and qualified personnel 14-
and reduce energy consumption [157].
15-16 To facilitate the collaboration, researchers have tested the
Define a strategy for monitoring 11 application of Building Information Modelling and cloud sharing
Construction Regularly control the quality of execution and 1-13- platforms [158,159]. However, these technologies are still at an
eventually update the theoretical values with the 14
early stage and need all stakeholders to be trained to cope with
real ones
Set up a system of sensors for measurements and 10-11 these new tools before they can be implemented in the energy cer-
alarms to detect malfunctions tification scheme [89]. It is indeed important to improve the skills
Encourage collaboration/communication among 15-18 of all the actors involved in the design process [160], especially by
all stakeholders promoting more integrated and cross-disciplinary teamwork, [161]
Commissioning Adjust the systems depending on the real usage of 11-12
the building
in order to move towards a common understanding of the optimal
Convene all stakeholders when the building is 11-18 consumption.
commissioned and implement periodic
inspections to be performed in the first year of 4.3. Solutions at different building life stages
operation
Train building managers on different aspects of 11-12
energy management The energy certificate represents a snapshot of building perfor-
Ensure that follow-up is organized and 11- mance at a given time, while the causes of the EPG stem from dif-
communicated in writing 15-18 ferent stages of its design, construction and operation. It is
Operation Monitor the performance and eventually optimize 7-11 therefore important that the solutions presented in Sections 4.1
it
Inform and raise awareness among inhabitants 9-17-
and 4.2 also be applied at their most appropriate phase of the pro-
and assess their level of satisfaction/comfort 18 ject. The final goal of these solutions is to bring the consumption of
Give feedback to the designers and all the other 18 a building as closely as possible to optimal consumption. These
stakeholders involved, disseminate the good solutions have been summarised in the form of a series of recom-
practice
mendations in Table 2. These aim to reduce the EPG by avoiding
10
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

breaks in the chain of responsibility throughout the stages of the The objective is to make theoretical values more precise and accu-
buildings’ life. The recommendations are grouped according to rate, and therefore close the EPG by bringing theoretical consump-
the different phases of the project following the classification of tion closer to the actual one. However, current certification
Khoury et al. [23] and matched with the specific cause(s) identified methods should not be expected to predict actual consumptions
in Table 1 using the Cause ID. Since some causes may be tackled as they are primarily designed to enable comparisons between
differently, more than one solution may refer to the same cause. buildings. The problem is that certificates are often used to predict
actual energy consumption and energy savings from building ret-
rofits. This is due to a lack of valid alternatives, which highlights
5. Conclusions the pressing need for quick and accessible performance informa-
tion. To this end, improved standard values could be used to more
This study reviewed the causes of the Energy Performance Gap accurately model energy consumption, although - as models are
between calculated energy consumption in residential perfor- sensitive to parameter changes - this could lead to problems of
mance certificates and actual consumption, as well as strategies over-fitting. Alternatively, methods using actual energy data could
proposed for reducing the gap. We introduce the concept of opti- be adopted.
mal consumption to classify the causes and mitigation strategies The second group focuses on improving the actual performance
for EPG. Optimal consumption is the real energy consumption of of the building’s energy systems, through better monitoring, main-
a building performing in an ideal way, in which the comfort of tenance, and general usage of the building. In this case the objec-
the inhabitants is guaranteed, their use of the building is as tive is to close the EPG by focusing only on actual consumption
intended, and the technical systems function as planned. The EPG to bring it closer to theoretical consumption.
can then be broken down into i) the difference between theoretical Our work highlights that neither approach should necessarily
consumption and optimal consumption and ii) the difference aim to eliminate the gap between current performance certificate
between actual consumption and optimal consumption. The first calculations and measurements, since the true optimal operation
difference is associated with the limitations of using a theoretical point of the building may correspond to neither of these values.
model based using standardised methods and inputs to describe A key first step is understanding to what extent observed discrep-
the real conditions of use of a building, while the second difference ancies stem from the calculation side compared to the implemen-
is associated with malfunctioning and unrecommended use of the tation. This is why we propose the concept of optimal consumption
building. as something distinct from theoretical and actual consumption
Introducing the optimal consumption concept allows to clarify because it brings greater clarity to the study of the EPG. However,
how the various causes and solutions fit into the building lifecycle further research is required to develop this concept of optimal val-
and what stakeholders might be involved. It highlights that a ues, in order to properly represent occupant and buildings diversi-
building’s theoretical consumption provided by a certificate will ties. This implies that the attributes of individual occupants and
systematically be different from the actual consumption. It is their buildings, such as the set-point temperature and the effi-
therefore necessary to define an optimal consumption as a target, ciency of the systems, should be used to better describe, and
towards which improvements to both the building certification achieve, optimal consumption. This review also identifies the need
and building operation can strive. Furthermore, optimal values, for studies on the very definition of optimal consumption for dif-
as opposed to the theoretical values that are used for today’s per- ferent conditions (e.g. climate zone, urban area, technologies), so
formance calculations, offer the basis for a debate around what is that it can be shared, regulated, and included within the existing
optimal performance of a building and what services it should pro- certification process.
vide. As highlighted by Shove [24] ‘‘the conclusion that technologies An overview of possible solutions has been presented based on
and practices are interwoven suggests that there might be ways of the existing body of knowledge. Significant work remains to refine
crafting buildings that do not meet present needs, and that do not deli- these solutions and ensure their effective implementation. It is
ver equivalent levels of service, but that do enable and sustain much notably essential to ensure better alignment between the various
lower-carbon ways of life”, and the concept of optimal consumption construction phases, in combination with complementary projects
can be seen as a step in this direction. and activities, leading to more specific recommendations.
This work further highlighted that EPG causes are generally The EPG remains a complex and incompletely understood topic.
considered in isolation. Most studies focus on one or two causes The related research questions are complex and require continuous
separately, but never in combination. Only 14% of the studies deal efforts across the multiple domains of research. The causes classi-
with both deviations (theoretical and actual, Fig. 2) of the EPG with fied can all influence the building’s energy performance, but
a more holistic approach. We find that the majority of studies tar- researchers have not been able to reliably quantify each impact
geted inaccuracy of building physical modelling (35%) and occu- independently due to intercorrelation and case-specific variability.
pant behaviour modelling (22%) to explain the EPG. Large-scale quantitative multi-parameter analyses remain scarce,
Malfunctioning equipment (13%) and execution of the work (8%) and with significant limitations, rendering it difficult to make
have been mentioned less often as causes of the EPG, although they definitive judgements about the relative importance of the causes
are directly related to the actual consumption of a building. This of the EPG and its solutions. Among the strategies to better under-
work therefore highlights the need to consider the diversity of pos- stand causes and solutions, we recommend in-depth monitoring
sible causes when developing strategies to reduce the EPG. Those and analysis of sufficiently large samples of buildings that are
mentioned most often in literature are not necessarily the ones prominent in the building stock, thereby allowing to draw conclu-
with the biggest impact on the EPG. Notably, previous works have sions for a significant share of today’s buildings.
often defined the role of the occupant as the main, if not sole, cause
of the EPG, but our findings indicate that this is not sufficiently
substantiated by evidence, in line with other recent works [162].
This review found that many strategies have been proposed to Declaration of Competing Interest
reduce the EPG, and that these can be categorized into two groups.
The first group, which appears most often, aims to improve the The authors declare that they have no known competing finan-
energy consumption calculation by correcting the standard values cial interests or personal relationships that could have appeared
and/or considering new approaches to create energy certificates. to influence the work reported in this paper.
11
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

Acknowledgement [23] J. Khoury, P. Hollmuller, B. Lachal, S. Schneider, U. Lehmann, COMPARE


RENOVE: du catalogue de solutions à la performance réelle des rénovations
énergétiques, Office fédéral de l’énergie (OFEN) (2018).
This research was financially supported by the Swiss Innovation [24] E. Shove, What is wrong with energy efficiency?, Build. Res. Inf. 46 (2018)
Agency Innosuisse and is part of the Swiss Competence Center for 779–789, https://doi.org/10.1080/09613218.2017.1361746.
[25] SIA, SIA 380 - Basi per il calcolo energetico di edifici, Swiss Society of
Energy Research on Future Energy Efficient Buildings & Districts
Engineers and Architects (SIA), 2015.
(SCCER FEEB&D). It builds on work done in the framework of the [26] F. Flourentzou, A.Y. Ivanov, P. Samuel, Understand, simulate, anticipate and
University of Sydney – University of Geneva Partnership Collabora- correct performance gap in NZEB refurbishment of residential buildings, J.
Phys. Conf. Ser. (2019), https://doi.org/10.1088/1742-6596/1343/1/012177.
tion Award.
[27] T. Frank, Climate change impacts on building heating and cooling energy
demand in Switzerland, Energy Build. 37 (2005) 1175–1185, https://doi.org/
10.1016/j.enbuild.2005.06.019.
[28] SIA, SIA 384/3 - Impianti di riscaldamento negli edifici – Fabbisogno di
References energia, Swiss Society of Engineers and Architects (SIA), 2013.
[29] M. Dominguez, O.M.D., C. Fraga, S. Callegari, P. Hollmuller, Air-to-water heat
pumps as a substitution of oil-boiler in a non-retrofitted multi-family
[1] International Energy Agency, World Energy Outlook 2018, 2018.
building of the 70’s. In-situ monitoring, actual energy balance and
[2] S.V. Russell-Smith, M.D. Lepech, R. Fruchter, Y.B. Meyer, Sustainable target
performance, in: 13th IEA Heat Pump Conf., 2020.
value design: Integrating life cycle assessment and target value design to
[30] P. de Wilde, Ten questions concerning building performance analysis, Build.
improve building energy and environmental performance, J. Clean. Prod.
Environ. (2019), https://doi.org/10.1016/j.buildenv.2019.02.019.
(2015), https://doi.org/10.1016/j.jclepro.2014.03.025.
[31] M. Mangold, M. Österbring, H. Wallbaum, Handling data uncertainties when
[3] EU Parliament, Directive 2018/844/EU Energy performance of buildings, in:
using Swedish energy performance certificate data to describe energy usage
Off. J. Eur. Communities, 2018: pp. 75–91.
in the building stock, Energy Build. 102 (2015) 328–336, https://doi.org/
[4] G. Sutherland, P.G. Audi, A. Lacourt, A. Deliyannis, D. Sotiropoulos, K.
10.1016/j.enbuild.2015.05.045.
Poseidon, P. Mcelmuray, M. Skrivanou, S. Koutsou, M.F. Davis, A. Fytrou, M.
[32] C. Gorse, M. Brooke-Peat, J. Parker, F. Thomas, Building Simulation and
Tsagkla, Implementing the Energy Performance of Building Directive (EPBD),
Models: Closing the Performance Gap, in: Build. Sustain. Futur. Des. Built
Intelligent Energy Europe Programme, 2015. www.epbd-ca.eu (accessed
Environ., 2015: pp. 1–351. doi:10.1007/978-3-319-19348-9.
November 14, 2017).
[33] J. Langevin, J.L. Reyna, S. Ebrahimigharehbaghi, N. Sandberg, P. Fennell, C.
[5] R. Galvin, M. Sunikka-Blank, Quantification of (p)rebound effects in retrofit
Nägeli, J. Laverge, M. Delghust, Mata, M. Van Hove, J. Webster, F. Federico, M.
policies – why does it matter?, Energy. (2016), https://doi.org/10.1016/j.
Jakob, C. Camarasa, Developing a common approach for classifying building
energy.2015.12.034.
stock energy models, Renew. Sustain. Energy Rev. 133 (2020). doi:10.1016/j.
[6] M. Brøgger, P. Bacher, H. Madsen, K.B. Wittchen, Estimating the influence of
rser.2020.110276.
rebound effects on the energy-saving potential in building stocks, Energy
[34] Y. Wei, X. Zhang, Y. Shi, L. Xia, S. Pan, J. Wu, M. Han, X. Zhao, A review of data-
Build. 181 (2018) 62–74, https://doi.org/10.1016/j.enbuild.2018.10.006.
driven approaches for prediction and classification of building energy
[7] IPCC, IPCC, 2018: Summary for Policymakers, 2018.
consumption, Renew. Sustain. Energy Rev. 82 (2018) 1027–1047, https://
[8] O. Pasichnyi, J. Wallin, F. Levihn, H. Shahrokni, O. Kordas, Energy performance
doi.org/10.1016/j.rser.2017.09.108.
certificates — new opportunities for data-enabled urban energy policy
[35] A. Marshall, R. Fitton, W. Swan, D. Farmer, D. Johnston, M. Benjaber, Y. Ji,
instruments?, Energy Policy. 127 (2019) 486–499, https://doi.org/10.1016/j.
Domestic building fabric performance: closing the gap between the in situ
enpol.2018.11.051.
measured and modelled performance, Energy Build. 150 (2017) 307–317,
[9] K. Gram-Hanssen, S. Georg, Energy performance gaps: promises, people,
https://doi.org/10.1016/j.enbuild.2017.06.028.
practices, Build. Res Inf. (2017).
[36] M. Hughes, J. Palmer, V. Cheng, D. Shipworth, Global sensitivity analysis of
[10] M. Ménard, Building energy Performance Gap-possible causes and the role of
England’s housing energy model, J. Build. Perform. Simul. 8 (2015) 283–294,
users, operators and design standards, (2016).
https://doi.org/10.1080/19401493.2014.925505.
[11] M. Sunikka-Blank, R. Galvin, Introducing the prebound effect: the gap
[37] V. Fournier, Comparaison des coûts face à l’impact environnemental entre
between performance and actual energy consumption, Build. Res. Inf. 40
travaux de rénovation sur l’enveloppe et changement de vecteur énergétique
(2012) 260–273, https://doi.org/10.1080/09613218.2012.690952.
de plusieurs projets immobiliers à Genève, EPFL (2018).
[12] M.S. Geraldi, E. Ghisi, Building-level and Stock-level in contrast: a literature
[38] A. Merzkirch, T. Hoos, S. Maas, F. Scholzen, D. Waldmann, Wie genau sind
review of the energy performance of buildings during the operational stage,
unsere Energiepässe?, Bauphysik (2014) 40–43, https://doi.org/10.1002/
Energy Build. (2020), https://doi.org/10.1016/J.ENBUILD.2020.109810.
bapi.201410007.
[13] A.J. Summerfield, T. Oreszczyn, J. Palmer, I.G. Hamilton, F.G.N. Li, J. Crawley, R.
[39] C. Loucari, J. Taylor, R. Raslan, E. Oikonomou, A. Mavrogianni, Retrofit
J. Lowe, What do empirical findings reveal about modelled energy demand
solutions for solid wall dwellings in England: the impact of uncertainty upon
and energy ratings? Comparisons of gas consumption across the English
the energy performance gap, Build. Serv. Eng. Res. Technol. 37 (2016) 614–
residential sector, Energy Policy 129 (2019) 997–1007, https://doi.org/
634, https://doi.org/10.1177/0143624416647758.
10.1016/j.enpol.2019.02.033.
[40] L. La Fleur, B. Moshfegh, P. Rohdin, Measured and predicted energy use and
[14] P. Palma, J.P. Gouveia, S.G. Simoes, Mapping the energy performance gap of
indoor climate before and after a major renovation of an apartment building
dwelling stock at high-resolution scale: implications for thermal comfort in
in Sweden, Energy Build. 146 (2017) 98–110, https://doi.org/10.1016/j.
Portuguese households, Energy Build. (2019), https://doi.org/10.1016/j.
enbuild.2017.04.042.
enbuild.2019.03.002.
[41] J. Ferreira, M. Pinheiro, In search of better energy performance in the
[15] C. Van Dronkelaar, M. Dowson, C. Spataru, D. Mumovic, A review of the
Portuguese buildings — The case of the Portuguese regulation, Energy Policy.
energy performance gap and its underlying causes in non-domestic buildings,
39 (2011) 7666–7683, https://doi.org/10.1016/j.enpol.2011.08.062.
Front. Mech. Eng. 1 (2016) 1–14, https://doi.org/10.3389/fmech.2015.00017.
[42] H. Hens, W. Parijs, M. Deurinck, Energy consumption for heating and rebound
[16] P.X.W. Zou, X. Xu, J. Sanjayan, J. Wang, Review of 10 years research on
effects, Energy Build. 42 (2010) 105–110, https://doi.org/10.1016/j.
building energy performance gap: life-cycle and stakeholder perspectives,
enbuild.2009.07.017.
Energy Build. 178 (2018) 165–181, https://doi.org/10.1016/j.
[43] R.V. Jones, A. Fuertes, C. Boomsma, S. Pahl, Space heating preferences in UK
enbuild.2018.08.040.
social housing: a socio-technical household survey combined with building
[17] B. Frei, C. Sagerschnig, D. Gyalistras, ParkGap – Performance Gap Gebäude,
audits, Energy Build. 127 (2015) 382–398, https://doi.org/10.1016/j.
2018.
enbuild.2016.06.006.
[18] I. Mojic, M. Luzzatto, M. Haller, M. Lehmann, M. Benz, S. van Velsen,
[44] D. Calì, T. Osterhage, R. Streblow, D. Müller, Energy performance gap in
ImmoGap: Einfluss der Kombination aus Nutzerverhalten und
refurbished German dwellings: Lesson learned from a field test, Energy Build.
Gebäudetechnik auf den Performance Gap bei Mehrfamilienhäuser, 2018.
127 (2016) 1146–1158, https://doi.org/10.1016/j.enbuild.2016.05.020.
[19] S. Petersen, C.A. Hviid, The European energy performance of buildings
[45] K.E. Thomsen, J. Rose, O. Mørck, S.Ø. Jensen, I. Østergaard, H.N. Knudsen, N.C.
directive: Comparison of Calculated and Actual Energy Use in a Danish Office
Bergsøe, Energy consumption and indoor climate in a residential building
Building, in: IBPSA-England; Build. Simul. Optim. Conf., 2012: pp. 43–48.
before and after comprehensive energy retrofitting, Energy Build. 123 (2016)
http://www.ibpsa-england.org/resources/files/bso-2012/1B3.pdf (accessed
8–16, https://doi.org/10.1016/j.enbuild.2016.04.049.
October 31, 2017).
[46] J.W. Moon, S.H. Han, Thermostat strategies impact on energy consumption in
[20] S. Cozza, J. Chambers, C. Deb, J.L. Scartezzini, A. Schlüter, M.K. Patel, Do energy
residential buildings, Energy Build. 43 (2011) 338–346, https://doi.org/
performance certificates allow reliable predictions of actual energy
10.1016/j.enbuild.2010.09.024.
consumption and savings?, Learning from the Swiss national database,
[47] H.X. Zhao, F. Magoulès, A review on the prediction of building energy
Energy Build. 224 (2020), https://doi.org/10.1016/j.enbuild.2020.110235
consumption, Renew. Sustain. Energy Rev. 16 (2012) 3586–3592, https://doi.
110235.
org/10.1016/j.rser.2012.02.049.
[21] F. Filippidou, N. Nieboer, H. Visscher, Effectiveness of energy renovations: a
[48] D. Majcen, L. Itard, H. Visscher, Actual and theoretical gas consumption in
reassessment based on actual consumption savings, Energy Effic. 12 (2018)
Dutch dwellings: what causes the differences?, Energy Policy. 61 (2013) 460–
19–35, https://doi.org/10.1007/s12053-018-9634-8.
471, https://doi.org/10.1016/j.enpol.2013.06.018.
[22] M.J. Grant, A. Booth, A typology of reviews: An analysis of 14 review types
[49] C. Hoffmann, A. Geissler, The prebound-effect in detail: real indoor
and associated methodologies, Health Info. Libr. J. 26 (2009) 91–108, https://
temperatures in basements and measured versus calculated U-values,
doi.org/10.1111/j.1471-1842.2009.00848.x.

12
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

Energy Procedia 122 (2017) 32–37, https://doi.org/10.1016/J. gap, Procedia Eng. 118 (2015) 573–580, https://doi.org/10.1016/j.
EGYPRO.2017.07.301. proeng.2015.08.487.
[50] R. Gupta, A. Kotopouleas, Magnitude and extent of building fabric thermal [74] D. Yan, W. O’Brien, T. Hong, X. Feng, H. Burak Gunay, F. Tahmasebi, A.
performance gap in UK low energy housing, Appl. Energy. 222 (2018) 673– Mahdavi, Occupant behavior modeling for building performance simulation:
686, https://doi.org/10.1016/j.apenergy.2018.03.096. Current state and future challenges, Energy Build. 107 (2015) 264–278.
[51] J. Kragh, J. Rose, H.N. Knudsen, O.M. Jensen, Possible explanations for the gap doi:10.1016/j.enbuild.2015.08.032.
between calculated and measured energy consumption of new houses, [75] J. Kim, S. Schiavon, G. Brager, Personal comfort models – a new paradigm in
Energy Procedia 132 (2017) 69–74, https://doi.org/10.1016/ thermal comfort for occupant-centric environmental control, Build. Environ.
j.egypro.2017.09.638. 132 (2018) 114–124, https://doi.org/10.1016/j.buildenv.2018.01.023.
[52] J. Khoury, Z. Alameddine, P. Hollmuller, Understanding and bridging the [76] F. Flourentzou, Possible strategies and obstacles in the pathway towards
energy performance gap in building retrofit, Energy Procedia 122 (2017) 217– energy transition of residential building stocks in Switzerland, IOP Conf. Ser.
222, https://doi.org/10.1016/j.egypro.2017.07.348. Earth Environ. Sci. 323 (2019), https://doi.org/10.1088/1755-1315/323/1/
[53] M. Delghust, W. Roelens, T. Tanghe, Y. De Weerdt, A. Janssens, Regulatory 012171 012171.
energy calculations versus real energy use in high-performance houses, Build. [77] J. Wang, J. Zhu, Z. Ding, P.X.W. Zou, J. Li, Typical energy-related behaviors and
Res. Inf. 43 (2015) 675–690, https://doi.org/10.1080/ gender difference for cooling energy consumption, J. Clean. Prod. (2019),
09613218.2015.1033874. https://doi.org/10.1016/j.jclepro.2019.117846.
[54] E. Rodrigues, M.S. Fernandes, A.R. Gaspar, Á. Gomes, J.J. Costa, Thermal [78] D. Calì, R.K. Andersen, D. Müller, B.W. Olesen, Analysis of occupants’ behavior
transmittance effect on energy consumption of Mediterranean buildings with related to the use of windows in German households, Build. Environ. (2016),
different thermal mass, Appl. Energy. 252 (2019), https://doi.org/10.1016/j. https://doi.org/10.1016/j.buildenv.2016.03.024.
apenergy.2019.113437 113437. [79] T. Dwyer, Knowledge is Power: Benchmarking and prediction of building
[55] A. Reilly, O. Kinnane, The impact of thermal mass on building energy energy consumption, Build. Serv. Eng. Res. Technol. 34 (2013) 5–7, https://
consumption, Appl. Energy. 198 (2017) 108–121, https://doi.org/10.1016/j. doi.org/10.1177/0143624412471130.
apenergy.2017.04.024. [80] U.I. Dar, L. Georges, I. Sartori, V. Novakovic, Influence of occupant’s behavior
[56] N. Aste, F. Leonforte, M. Manfren, M. Mazzon, Thermal inertia and energy on heating needs and energy system performance: a case of well-insulated
efficiency – parametric simulation assessment on a calibrated case study, detached houses in cold climates, Build. Simul. 8 (2015) 499–513, https://doi.
Appl. Energy. 145 (2015) 111–123, https://doi.org/10.1016/j. org/10.1007/s12273-015-0230-y.
apenergy.2015.01.084. [81] R. Galvin, Making the ‘rebound effect’ more useful for performance evaluation
[57] E. Mantesi, C.J. Hopfe, M.J. Cook, J. Glass, P. Strachan, The modelling gap: of thermal retrofits of existing homes: defining the ‘energy savings deficit’
quantifying the discrepancy in the representation of thermal mass in building and the ‘energy performance gap’, Energy Build. 69 (2013) 515–524, https://
simulation, Build. Environ. 131 (2017) 74–98, https://doi.org/10.1016/j. doi.org/10.1016/j.enbuild.2013.11.004.
buildenv.2017.12.017. [82] W. Reimann, E. Buhlmann, M. Lehmann, Erfolgskontrolle
[58] R. De Lieto Vollaro, C. Guattari, L. Evangelisti, G. Battista, E. Carnielo, P. Gori, Gebäudeenergiestandards 2014-2015, Bundesamt für Energie (BFE), 2016.
Building energy performance analysis: a case study, Energy Build. 87 (2015) [83] O. Guerra-Santin, C.A. Tweed, In-use monitoring of buildings: an overview of
87–94, https://doi.org/10.1016/j.enbuild.2014.10.080. data collection methods, Energy Build. 93 (2015) 189–207, https://doi.org/
[59] T. Schluck, K.N. Streicher, S. Mennel, Statistical modelling of the energy 10.1016/j.enbuild.2015.02.042.
reference area based on the Swiss building stock, J. Phys. Conf. Ser. 86 (2019) [84] C. Hopfe, C. Struck, G.U. Harputlugil, J. Hensen, Computational Simulation
35–42, https://doi.org/10.1002/cjce.20015. Tools For Building Services Design – Professional’s Practice and Wishes, in:
[60] A. GhaffarianHoseini, T. Zhang, O. Nwadigo, A. GhaffarianHoseini, N. 17th Int. Air-Conditioning Vent. Conf., 2006. http://www.bwk.tue.nl/bps/
Naismith, J. Tookey, K. Raahemifar, Application of nD BIM Integrated hensen/publications/06_acv_ccg.pdf.
Knowledge-based Building Management System (BIM-IKBMS) for [85] C. Ahern, B. Norton, Energy savings across EU domestic building stock by
inspecting post-construction energy efficiency, Renew. Sustain. Energy Rev. optimizing hydraulic distribution in domestic space heating systems, Energy
72 (2017) 935–949, https://doi.org/10.1016/j.rser.2016.12.061. Build. (2015), https://doi.org/10.1016/j.enbuild.2015.01.014.
[61] S. Erba, F. Causone, R. Armani, The effect of weather datasets on building [86] M. Moezzi, C. Hammer, J. Goins, A. Meier, Behavioral strategies to bridge the
energy simulation outputs, Energy Procedia 134 (2017) 545–554, https://doi. gap between potential and actual savings in buildings, Natl. Res. Cent.
org/10.1016/j.egypro.2017.09.561. Canada. (2014), https://doi.org/10.1289/ehp.1002503.
[62] G. Branco, B. Lachal, P. Gallinelli, W. Weber, Predicted versus observed heat [87] M.W. Ahmad, M. Mourshed, D. Mundow, M. Sisinni, Y. Rezgui, Building
consumption of a low energy multifamily complex in Switzerland based on energy metering and environmental monitoring – a state-of-the-art review
long-term experimental data, Energy Build. 36 (2004) 543–555, https://doi. and directions for future research, Energy Build. 120 (2016) 85–102, https://
org/10.1016/j.enbuild.2004.01.028. doi.org/10.1016/j.enbuild.2016.03.059.
[63] J. Kočí, V. Kočí, J. Maděra, R. Černý, Effect of applied weather data sets in [88] J. Wingfield, M. Bell, D. Miles-Shenton, J. Seavers, Elm Tree Mews Field Trial –
simulation of building energy demands: comparison of design years with Evaluation and Monitory of Dwellings Performance, Final Technical Report,
recent weather data, Renew. Sustain. Energy Rev. 100 (2019) 22–32, https:// 2011. http://www.leedsbeckett.ac.uk/as/cebe/projects/elmtree/
doi.org/10.1016/j.rser.2018.10.022. elmtree_finalreport.pdf (accessed October 6, 2017).
[64] L. Thaler, D. Kellenberger, Addressing gaps: user on behavior and sufficiency [89] D.J. McElroy, J. Rosenow, Policy implications for the performance gap of low-
in the planning and operation phase of a 2000-Watt Site, Energy Procedia 122 carbon building technologies, Build. Res. Inf. 47 (2019) 611–623, https://doi.
(2017) 961–966, https://doi.org/10.1016/j.egypro.2017.07.440. org/10.1080/09613218.2018.1469285.
[65] M. Herrera, S. Natarajan, D.A. Coley, T. Kershaw, A.P.R. González, M. Eames, D. [90] T. Maile, V. Bazjanac, M. Fischer, A method to compare simulated and
Fosas, M. Wood, A review of current and future weather data for building measured data to assess building energy performance, Build. Environ. 56
simulation, Build. Serv. Eng. Res. Technol. (2017), https://doi.org/10.1177/ (2012) 241–251, https://doi.org/10.1016/j.buildenv.2012.03.012.
0143624417705937. [91] S. Wyss, W. Hässig, UFELD: Feldmessungen von U-Werten zur Überprüfung
[66] A. Gonzalez Caceres, D.G. Zenginis, T.A. Vik, The impact of the weather data der im Gebäudeenergieausweis (GEAK) hinterlegten U-Werte, Bundesamt für
file on the energy performance certificate, the case of Norway, Proc. 59th Energie (BFE), 2016.
Conf. Imulation Model. (SIMS 59), 26-28 Sept. 2018, Oslo Metrop. Univ. Norw. [92] P. de Wilde, The gap between predicted and measured energy performance of
153 (2018) 342–349. doi:10.3384/ecp18153342. buildings: a framework for investigation, Autom. Constr. 41 (2014) 40–49,
[67] M.F. Jentsch, P.A.B. James, L. Bourikas, A.B.S. Bahaj, Transforming existing https://doi.org/10.1016/j.autcon.2014.02.009.
weather data for worldwide locations to enable energy and building [93] M. Bell, J. Wingfi, D. Miles-Shenton, J. Seavers, Low carbon housing Lessons
performance simulation under future climates, Renew. Energy. 55 (2013) from Elm Tree Mews, (2010) 118. doi:978-1-85935-766-8.
514–524, https://doi.org/10.1016/j.renene.2012.12.049. [94] C. Turner, M. Frankel, Energy performance of LEED for new construction
[68] L. Kallab, R. Chbeir, P. Bourreau, P. Brassier, M. Mrissa, HIT2GAP: towards a buildings, New Build. Inst. (2008) 1–46.
better building energy management, Energy Procedia 122 (2017) 895–900, [95] J. Love, Understanding the Interactions Between Occupants, Heating Systems,
https://doi.org/10.1016/j.egypro.2017.07.399. and Building Fabric in the Context of Social Housing, University College
[69] E. Rodrigues, M.S. Fernandes, Overheating risk in Mediterranean residential London, 2014.
buildings: comparison of current and future climate scenarios, Appl. Energy. [96] E. Cayre, B. Allibe, M.-H. Laurent, D. Osso, There are people in the house! How
259 (2020), https://doi.org/10.1016/j.apenergy.2019.114110 114110. the results of purely technical analysis of residential energy consumption are
[70] Y. Zhang, X. Bai, F.P. Mills, J.C.V. Pezzey, Rethinking the role of occupant misleading for energy policies, in: ECEEE 2011 Summer Study - Energy Effic.
behavior in building energy performance: a review, Energy Build. 172 (2018) First Found. a Low-Carbon Soc., 2011.
279–294, https://doi.org/10.1016/j.enbuild.2018.05.017. [97] D. Marchio, A. Rabl, Energy-efficient gas-heated housing in France: predicted
[71] S. Hu, D. Yan, E. Azar, F. Guo, A systematic review of occupant behavior in and observed performance, Energy Build. 17 (1991) 131–139, https://doi.org/
building energy policy, Build. Environ. 175 (2020), https://doi.org/10.1016/j. 10.1016/0378-7788(91)90005-N.
buildenv.2020.106807 106807. [98] U. Lehmann, J. Khoury, M.K. Patel, Actual energy performance of student
[72] P. Van den Brom, A. Meijer, H. Visscher, Performance gaps in energy housing: case study, benchmarking and performance gap analysis, Energy
consumption: household groups and building characteristics, Build. Res. Inf. Procedia 122 (2017) 163–168, https://doi.org/10.1016/J.
46 (2017) 54–70, https://doi.org/10.1080/09613218.2017.1312897. EGYPRO.2017.07.339.
[73] S. Niu, W. Pan, Y. Zhao, A virtual reality supported approach to occupancy [99] B. Risholt, T. Berker, Success for energy efficient renovation of dwellings—
engagement in building energy design for closing the energy performance Learning from private homeowners, Energy Policy. 61 (2013) 1022–1030,
https://doi.org/10.1016/j.enpol.2013.06.011.

13
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

[100] W. Tian, Y. Heo, P. de Wilde, Z. Li, D. Yan, C.S. Park, X. Feng, G. Augenbroe, A [123] D. Jonlin, Bridging the Energy Performance Gap: Real-World Tools, ACEEE
review of uncertainty analysis in building energy assessment, Renew. Summer Study Energy Effic. Build. (2014) 168–179. http://aceee.org/files/
Sustain. Energy Rev. 93 (2018) 285–301, https://doi.org/10.1016/j. proceedings/2014/data/papers/3-300.pdf.
rser.2018.05.029. [124] J.D. Chambers, T. Oreszczyn, Deconstruct: a scalable method of as-built heat
[101] V. Taranu, G. Verbeeck, A closer look into the European Energy Performance power loss coefficient inference for UK dwellings using smart meter data,
Certificates under the lenses of behavioural insights — a comparative Energy Build. (2019), https://doi.org/10.1016/j.enbuild.2018.11.016.
analysis, Energy Effic. (2018) 1745–1761. [125] H. Hufen, H. De Bruijn, Getting the incentives right. Energy performance
[102] J.O. Olaussen, A. Oust, J.T. Solstad, Energy performance certificates – contracts as a tool for property management by local government, J. Clean.
informing the informed or the indifferent?, Energy Policy. 111 (2017) 246– Prod. (2016), https://doi.org/10.1016/j.jclepro.2015.10.036.
254, https://doi.org/10.1016/j.enpol.2017.09.029. [126] S. Pätäri, K. Sinkkonen, Energy service companies and energy performance
[103] J. Las-Heras-Casas, L.M. López-Ochoa, L.M. López-González, J.P. Paredes- contracting: is there a need to renew the business model? insights from a
Sánchez, A tool for verifying energy performance certificates and improving Delphi study, J. Clean. Prod. (2014), https://doi.org/10.1016/j.
the knowledge of the residential sector : a case study of the Autonomous jclepro.2013.10.017.
Community of Aragón (Spain), Sustain. Cities Soc. 41 (2018) 62–72, https:// [127] Conferenza Cantonale dei Direttori dell’Energia CDE, Il Programma Edifici nel
doi.org/10.1016/j.scs.2018.05.016. 2016 - Rapporto annuale, (2016).
[104] A. Hardy, D. Glew, An analysis of errors in the Energy Performance certificate [128] E. Cuerda, O. Guerra-Santin, J.J. Sendra, F.J. Neila, Understanding the
database, Energy Policy. 129 (2019) 1168–1178, https://doi.org/10.1016/j. performance gap in energy retrofitting: measured input data for adjusting
enpol.2019.03.022. building simulation models, Energy Build. (2019), https://doi.org/10.1016/j.
[105] H. Do, K.S. Cetin, Evaluation of the causes and impact of outliers on enbuild.2019.109688.
residential building energy use prediction using inverse modeling, Build. [129] S. Naylor, M. Gillott, T. Lau, A review of occupant-centric building control
Environ. (2018), https://doi.org/10.1016/j.buildenv.2018.04.039. strategies to reduce building energy use, Renew. Sustain. Energy Rev. 96
[106] S. Semple, D. Jenkins, Variation of energy performance certificate (2018) 1–10, https://doi.org/10.1016/j.rser.2018.07.019.
assessments in the European Union, Energy Policy. 137 (2020), https://doi. [130] S. Niu, W. Pan, Y. Zhao, A virtual reality integrated design approach to
org/10.1016/j.enpol.2019.111127 111127. improving occupancy information integrity for closing the building energy
[107] L. Pérez-Lombard, J. Ortiz, R. González, I.R. Maestre, A review of performance gap, Sustain. Cities Soc. 27 (2016) 275–286, https://doi.org/
benchmarking, rating and labelling concepts within the framework of 10.1016/j.scs.2016.03.010.
building energy certification schemes, Energy Build. 41 (2009) 272–278, [131] P. Torcellini, M. Deru, B. Griffith, N. Long, S. Pless, R. Judkoff, D.B. Crawley,
https://doi.org/10.1016/j.enbuild.2008.10.004. Lessons learned from field evaluation of six high-performance buildings,
[108] A. Gonzalez-Caceres, T.A. Vik, Improving the Energy Performance Certificate ACEEE Summer Study Energy Effic Build. (2004) 325–337.
recommendations accuracy for residential building through simple [132] M. Frei, J. Hofer, A. Schlüter, Z. Nagy, An easily-deployable wireless sensor
measurements of key inputs, IOP Conf. Ser. Mater. Sci. Eng. (2019), https:// network for building energy performance assessment, Energy Procedia 122
doi.org/10.1088/1757-899X/609/3/032065. (2017) 523–528, https://doi.org/10.1016/j.egypro.2017.07.309.
[109] I. Ballarini, V. Corrado, Application of energy rating methods to the existing [133] C. Wang, D. Yan, Y. Jiang, A novel approach for building occupancy
building stock: analysis of some residential buildings in Turin, Energy Build. simulation, Build. Simul. 4 (2011) 149–167, https://doi.org/10.1007/
41 (2009) 790–800, https://doi.org/10.1016/j.enbuild.2009.02.009. s12273-011-0044-5.
[110] A. Stone, D. Shipworth, P. Biddulph, T. Oreszczyn, Key factors determining the [134] C. Carpino, E. Loukou, P. Heiselberg, N. Arcuri, Energy performance gap of a
energy rating of existing English houses, Build. Res. Inf. 42 (2014) 725–738, nearly Zero Energy Building (nZEB) in Denmark: the influence of occupancy
https://doi.org/10.1080/09613218.2014.905383. modelling, Build. Res. Inf. (2020) 1–23, https://doi.org/10.1080/
[111] SIA, SIA 380/1 - Besoins de chaleur pour le chauffage, Swiss Society of 09613218.2019.1707639.
Engineers and Architects (SIA), 2016. [135] X. Shi, B. Si, J. Zhao, Z. Tian, C. Wang, X. Jin, X. Zhou, Magnitude, causes, and
[112] F.P. Schröder, O. Papert, T. Boegelein, H. Navarro, B. Mundry, Reale Trends des solutions of the performance gap of buildings: a review, Sustainability. 11
spezifischen Energieverbrauchs und repräsentativer (2019) 1–21, https://doi.org/10.3390/su11030937.
Wohnraumtemperierung bei steigendem Modernisierungsgrad im [136] H.S. Alavi, H. Verma, J. Mlynář, D. Lalanne, The hide and seek of workspace:
Wohnungsbestand, in: Bauphysik, 2014: pp. 309–324. doi:10.1002/ towards human-centric sustainable architecture, Conf. Hum. Factors Comput.
bapi.201410045. Syst. - Proc. (2018-April (2018)) 1–12, https://doi.org/10.1145/
[113] H. Elsharkawy, S. Zahiri, The significance of occupancy profiles in 3173574.3173649.
determining post retrofit indoor thermal comfort, overheating risk and [137] C. Fan, F. Xiao, Z. Li, J. Wang, Unsupervised data analytics in mining big
building energy performance, Build. Environ. 172 (2020), https://doi.org/ building operational data for energy efficiency enhancement: a review,
10.1016/j.buildenv.2020.106676 106676. Energy Build. 159 (2018) 296–308, https://doi.org/10.1016/j.
[114] C. Deb, M. Frei, A. Schlueter, Identifying temporal properties of building enbuild.2017.11.008.
components and indoor environment for building performance assessment, [138] D. Johnston, D. Farmer, M. Brooke-Peat, D. Miles-Shenton, Bridging the
Build. Environ. (2020), https://doi.org/10.1016/j.buildenv.2019.106506. domestic building fabric performance gap, Build. Res. Inf. 44 (2016) 147–159,
[115] C. Bianchi, A.D. Smith, Localized Actual Meteorological Year File Creator https://doi.org/10.1080/09613218.2014.979093.
(LAF): A tool for using locally observed weather data in building energy [139] A. Dodoo, L. Gustavsson, U.Y.A. Tettey, Final energy savings and cost-
simulations, SoftwareX. 10 (2019), https://doi.org/10.1016/ effectiveness of deep energy renovation of a multi-storey residential
j.softx.2019.100299 100299. building, Energy 135 (2017) 563–576, https://doi.org/10.1016/j.
[116] M.F. Jentsch, A.B.S. Bahaj, P.A.B. James, Climate change future proofing of energy.2017.06.123.
buildings-Generation and assessment of building simulation weather files, [140] EU Parliament, Directive 2002/91/EC of the European parliament of the
Energy Build. 40 (2008) 2148–2168, https://doi.org/10.1016/j. council of 16 December 2002 on the energy performance of buildings, in: Off.
enbuild.2008.06.005. J. Eur. Communities L1/65-71, 2003.
[117] J. Amoako-Attah, A.B. Jahromi, The impact of different weather files on [141] EU Parliament, Directive 2010/31/EU of the European parliament and of the
London detached residential building performance-deterministic, council of 19 May 2010 on the energy performance of buildings, in: Off. J. Eur.
uncertainty, and sensitivity analysis on CIBSE TM48 and CIBSE TM49 future Communities L153/13-35, 2010.
weather variables using CIBSE TM52 as overheating criteria, Sustainability. 8 [142] K.G. Droutsa, S. Kontoyiannidis, E.G. Dascalaki, C.A. Balaras, Mapping the
(2016) 1–18, https://doi.org/10.3390/su8111194. energy performance of hellenic residential buildings from EPC (energy
[118] P. van den Brom, A. Meijer, H. Visscher, Actual energy saving effects of performance certificate) data, Energy. 98 (2016) 284–295, https://doi.org/
thermal renovations in dwellings—longitudinal data analysis including 10.1016/j.energy.2015.12.137.
building and occupant characteristics, Energy Build. 182 (2019) 251–263, [143] R. Fazeli, B. Davidsdottir, Energy performance of dwelling stock in Iceland:
https://doi.org/10.1016/j.enbuild.2018.10.025. system dynamics approach, J. Clean. Prod. (2017), https://doi.org/10.1016/j.
[119] M. Laurent B. Allibe C. Tigchelaar T. Oreszczyn I. Hamilton R. Galvin Back to jclepro.2017.05.009.
reality : How domestic energy efficiency policies in four European countries [144] Y. Li, S. Kubicki, A. Guerriero, Y. Rezgui, Review of building energy
can be improved by using empirical data instead of normative calculation performance certification schemes towards future improvement, Renew.
Eceee Summer Study Proceedings. 2013 2057 2070 http://discovery.ucl.ac. Sustain. Energy Rev. (2019), https://doi.org/10.1016/j.rser.2019.109244.
uk/1403583/. [145] C. Struck, M. Benz, V. Dorer, B. Frei, M. Hall, M. Ménard, S. Moosberger, K.
[120] T.H. Christensen, K. Gram-hanssen, M. De Best-, A. Adjei, Energy retrofits of Orehouring, C. Sagerschnig, Performance Gap in der Schweiz - Brisanz,
Danish homes: is the Energy Performance Certificate useful?, Build. Res. Inf. Ursachen und Einflüsse auf die Differenz von geplantem Energiebedarf und
3218 (2014), https://doi.org/10.1080/09613218.2014.908265. dem gemessenen Verbauch von Gebäuden, Brenet, 18. Status-Seminar,
[121] M.M. Sesana, G. Salvalai, A review on Building Renovation Passport: Zürich. (2014) 1–10.
potentialities and barriers on current initiatives, Energy Build. (2018), [146] Passivhaus Institut, Criteria for the Passive House, EnerPHit and PHI Low
https://doi.org/10.1016/j.enbuild.2018.05.027. Energy Building Standard, 2016. https://passiv.de/downloads/
[122] L.M. López-González, L.M. López-Ochoa, J. Las-Heras-Casas, C. García-Lozano, 03_building_criteria_en.pdf
Energy performance certificates as tools for energy planning in the [147] I. Allard, T. Olofsson, G. Nair, Energy evaluation of residential buildings:
residential sector. The case of La Rioja (Spain), J. Clean. Prod. (2016), performance gap analysis incorporating uncertainties in the evaluation
https://doi.org/10.1016/j.jclepro.2016.08.007. methods, Build. Simul. 11 (2018) 725–737, https://doi.org/10.1007/s12273-
018-0439-7.

14
S. Cozza, J. Chambers, A. Brambilla et al. Energy & Buildings 249 (2021) 111253

[148] Passivhaus Institut, Passive House checklist, (2021). https://passipedia.org/ [157] M. Anda, J. Temmen, Smart metering for residential energy efficiency: the use
basics/passive_house_checklist. of community based social marketing for behavioural change and smart grid
[149] M. Way, B. Bordass, Making feedback and post-occupancy evaluation routine introduction, Renew. Energy. 67 (2014) 119–127, https://doi.org/10.1016/j.
2: Soft Landings – involving design and building teams in improving renene.2013.11.020.
performance, Build. Res. Inf. 33 (2005) 353–360, https://doi.org/10.1080/ [158] S.-A. Kim, Y. Choe, M. Jang, W. Seol, Design process visualization system
09613210500162008. intergrating BIM data and performance-oriented design information, in: Proc.
[150] Geneve – ReC, GEnergie 2050: Des mesures concrètes pour accélérer la 28th Int. Symp. Autom. Robot. Constr., 2011: pp. 728–733.
transition énergétique du canton, 2019. [159] P.G. Tuohy, G.B. Murphy, Closing the gap in building performance: Learning
[151] Z. Min, P. Morgenstern, L. Marjanovic-Halburd, Facilities management added from BIM benchmark industries, Archit. Sci. Rev. 58 (2015) 47–56, https://
value in closing the energy performance gap, Int. J. Sustain. Built Environ. doi.org/10.1080/00038628.2014.975780.
(2016), https://doi.org/10.1016/j.ijsbe.2016.06.004. [160] Ö. Göçer, Y. Hua, K. Göçer, Completing the missing link in building design
[152] A. Freyre, Energy programs for buildings; analysis in the context of the Swiss process: enhancing post-occupancy evaluation method for effective feedback
energy transition, PhD Thesis - Univ. Geneva. (2019). for building performance, Build. Environ. (2015), https://doi.org/10.1016/j.
[153] D. Hsu, How much information disclosure of building energy performance is buildenv.2015.02.011.
necessary?, Energy Policy. (2014), https://doi.org/10.1016/j. [161] C. Brunsgaard, P. Dvořáková, A. Wyckmans, W. Stutterecker, M. Laskari, M.
enpol.2013.08.094. Almeida, K. Kabele, Z. Magyar, P. Bartkiewicz, P. Op ’t Veld, Integrated energy
[154] M. Brøgger, K.B. Wittchen, Energy performance certificate classifications design – education and training in cross-disciplinary teams implementing
across shifting frameworks, Procedia Eng. 161 (2016) 845–849, https://doi. energy performance of buildings directive (EPBD), Build. Environ. 72 (2014)
org/10.1016/j.proeng.2016.08.727. 1–14, https://doi.org/10.1016/j.buildenv.2013.10.011.
[155] X. Xu, C.Z. Li, J. Wang, W. Huang, Collaboration between designers and [162] A. Mahdavi, C. Berger, H. Amin, E. Ampatzi, R.K. Andersen, E. Azar, V.M.
contractors to improve building energy performance, J. Clean. Prod. (2019), Barthelmes, M. Favero, J. Hahn, D. Khovalyg, H.N. Knudsen, A.L. Navarro, A.
https://doi.org/10.1016/j.jclepro.2019.02.036. Roetzel, F.C. Sangogboye, M. Schweiker, M. Taheri, The role of occupants in
[156] H. Rau, P. Moran, R. Manton, J. Goggins, Changing energy cultures? household buildings’ energy performance gap: myth or reality?, Sustainability. (2021)
energy use before and after a building energy efficiency retrofit, Sustain. 1–44.
Cities Soc. (2020), https://doi.org/10.1016/j.scs.2019.101983.

15

You might also like