Energy & Buildings 288 (2023) 113019

Contents lists available at ScienceDirect

Energy & Buildings

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

A detailed view of the Adaptive-Comfort-Control-Implementation Script

(ACCIS): The capabilities of the automation system for adaptive setpoint
temperatures in building energy models
Daniel Sánchez-García a,⇑, Jorge Martínez-Crespo a, Ulpiano Ruiz-Rivas Hernando b, Carmen Alonso c
a
Department of Electrical Engineering, University Carlos III of Madrid, Leganés, Spain
b
Department of Thermal and Fluid Engineering, University Carlos III of Madrid, Leganés, Spain
c
Department of construction, Eduardo Torroja Institute for Construction Science, CSIC, Madrid, Spain

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

Article history: Recent research has looked into the potential for energy savings from adopting setpoint temperatures
Received 20 October 2022 based on adaptive comfort. The recently suggested Adaptive-Comfort-Control-Implementation Script
Revised 30 January 2023 (ACCIS), a computational method that expands the use of air conditioning to adaptive comfort, can be
Accepted 23 March 2023
used to accomplish this. The user’s setup and an Input Data File (IDF) are used by ACCIS to convert
Available online 27 March 2023
PMV-based EnergyPlus models into adaptive setpoint models. The Adaptive Comfort Control
Implemented Model (‘‘accim”) Python module has been used to nest ACCIS, which was originally an
Keywords:
Energy Management System (EMS) script. This research focuses on the creation of ACCIS and offers a
Adaptive thermal comfort
Energy efficiency
more in-depth understanding of how the computational method functions based on the connections
Heating and cooling setpoint temperatures between the objects that compose it. Since all weather file quantity constraints have been overcome
Computational approach and studies can now be generated based on millions of simulations, ACCIS opens up new opportunities
Building energy modelling for performing energy simulations on a worldwide scale rather than only on a national or continental
Building performance simulation level. This potential is increased considering that, until recently only international models ASHRAE
Standard 55 and European standard EN16798-1 were considered in ACCIS, but now a recent update
allows to select local comfort models developed for specific areas. As a result, this potential is firstly
demonstrated by means of the study of the 3 arguments that have the most important impact on the set-
point temperatures: the thermal comfort model, the acceptability levels and the behaviour of setpoint
temperatures when adaptive comfort models are not applicable; and secondly by means of an actual
example of use of accim.
Ó 2023 The Author(s). 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/).

1. Introduction door environment. Air conditioning systems are occasionally used

to maintain a comfortable interior temperature. But utilising these
One of the most significant challenges facing humanity today is systems, which frequently have extremely constrained setpoint
climate change. As a result of human behaviour that disregards the temperatures that are fairly comfortable for the human body,
environment, climate change’s severity increases every year [1]. may lessen people’s physiological, psychological, and behavioural
Using the preindustrial mean surface temperature as a baseline, ability to adapt to changes in ambient temperature [4], in addition
temperature rises might range from 2.5 °C in 2050 to 4.5 °C in to requiring high energy rates. Recent studies have concentrated
2100, 7.5 °C in 2200, and 8 °C in 2300 [2]. To prevent resource on finding ways to mitigate this issue by utilising adaptive comfort
depletion and pollutant gas emissions, the building sector in Eur- settings for air conditioning, or, in other words, adopting adaptive
ope is required to reduce greenhouse gas emissions by 90% by setpoint temperatures (AST). According to the standards ASHRAE
2050 [3]. 55 and EN16798-1, adaptive comfort models are only appropriate
In light of climate change, it is essential to remember that the for naturally ventilated spaces. Recent studies, however, have
primary function of buildings is to provide protection from the out- shown that regardless of the technology used to create the interior
environment, humans seem to be able to adapt to a considerably
wider range of indoor temperatures [5]. It is also conceivable to
⇑ Corresponding author. create statistically significant adaptive comfort models for cooling
E-mail address: dsgarcia@ing.uc3m.es (D. Sánchez-García). operations in air-conditioned office buildings, according to

https://doi.org/10.1016/j.enbuild.2023.113019
0378-7788/Ó 2023 The Author(s). 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/).
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Nomenclature

ACCIM Adaptive-Comfort-Control-Implemented Model GUI Graphical User Interface

ACCIS Adaptive-Comfort-Control-Implementation Script HVAC Heating, Ventilation and Air-Conditioning
ACST Adaptive Cooling Setpoint Temperature IDF Input Data File
AHST Adaptive Heating Setpoint Temperature PMOT Prevailing Mean Outdoor Temperature
AST Adaptive Setpoint Temperatures PMV Predicted Mean Vote
BES Building Energy Simulation RMOT Running Mean Outdoor Temperature
EMS EnergyManagementSystem VST Ventilation Setpoint Temperature
EPW EnergyPlus Weather
ERL EnergyPlus Runtime Language

scientific data [6]. The values of these AST, which were determined of this, particular instances may differ for the various criteria taken
using adaptive thermal comfort techniques, correspond to upper into consideration.
and lower comfort thresholds. It is therefore expected that energy Recently, a number of national standards have been created,
consumption will be reduced due to the use of less-restricted set- such as GB/T50785 [15] from China and ISSO 74 [16,17] from the
point temperatures while maintaining adequate thermal human Netherlands. For the Chinese standard, two distinct models are
sensation because the use of AST shows that the operating temper- established based on the cold, warm, and moderate climatic zones.
ature fluctuates within the adaptive thermal comfort zone: (i) R.P. Additionally, each model creates two distinct groups, and the lim-
Kramer et al. [7] investigated energy conservation at an Amster- itations deviate from international norms. Additionally, China built
dam museum by using a variety of setpoint temperatures and particular models for particular locations [18]. In case of the Dutch
adaptive thermal comfort levels. A 74% energy savings was standard, the second revision from 2014 [19] focuses on creating
achieved by allocating the lower comfort limit to the heating set- interior settings that fall into one of four categories of acceptability
point temperature and only using the Heating, Ventilation and utilising both international databases and local research with var-
Air-Conditioning (HVAC) systems when the museum was open; ious upper and lower bounds.
(ii) Sánchez-Guevara Sánchez et al. used monthly setpoint temper- Carrying out detailed research for various building uses or to
atures in three Spanish residential buildings based on the comfort comprehend actual living conditions is not new, however it has
limitations of the adaptive model from ASHRAE Standard 55-2013 become a global trend since many adaptive comfort field studies
[8]. Energy consumption was reduced by 20% and 80%, respec- have arised in the last years. For instance, these studies have been
tively, in heating and cooling [9]. carried out for countries such as Pakistan [20], Iran [21], China [22–
Adaptive comfort models have been used up till now to suggest 25], Tunisia [26], Japan [27–29], India [30–32], Australia [33,34],
the integration of consumers’ climatic adaptability into the low Qatar [35], Brazil [36], Colombia [37], Mexico [38,39], Chile [40]
energy consumption. Standards like EN 16798-1:2019 [10] and and Romania [41], building uses such as hospitals [42], shelters
ASHRAE 55-2020 [11], which take into account that users can [43], dormitories [44], prefab construction site offices [45] or work-
interact with the environment, incorporate such models. The Smart shops [46], or social categories as age and gender, e.g. children in
Control and Thermal Comfort SCATs (conducted in Europe) and RP- primary school [47], females [48] and older residents [49].
884 international initiatives served as the foundation for both stan- DesignBuilder [50], which makes use of the simulation engine
dards (carried out in various world locations). The outcomes of EnergyPlus [51], was largely used to create exploratory investiga-
these experiments demonstrated the connection between the tions using building energy simulations (BES). Previously, adaptive
operative temperature and the outdoor temperature in terms of setpoint temperatures were applied to BES using 2 methods: first,
user comfort. very simplistic ones like changing setpoint temperatures monthly
Field studies revealed that occupants’ temperature reactions and then adding the outcomes of the 12 simulations; second, more
vary from those of occupants in buildings with centralised HVAC sophisticated ones like using Schedule:Compact objects, Energy-
systems primarily because of variances in the thermal experience, Plus objects that can precisely define setpoint temperatures and
the availability of controls, and changes in occupant expectations operation schedules. However, it included a laborious, challenging,
[12]. and time-consuming manual approach that was prone to mistakes
In recent years, distinct adaptive thermal comfort models have which took the following steps: (i) setpoint temperatures were
been created to (i) escape the restrictions of the global models, (ii) previously calculated using an Excel spreadsheet and then (ii)
comprehend the uniqueness of each country, and (iii) to address pasted into Schedule:Compact objects which should be created in
individual living or building usage conditions: advance, then (iii) the appropriate EPW file was selected, and lastly
The comfort standard is based on a limited amount of data col- (iv) the simulation of the setpoint temperature and climate zone
lecting for outdoor temperatures more than 25 °C, according to the took place. Additional file management tasks were also required.
EN 16798-1:2019 standard. This is due to the fact that, among all However, this process needed to be repeated for each combination
the case studies of the countries participating in the development of AST and EPW, which could be numbered over hundreds or
of the standard, just two buildings in Greece have data under such thousands.
circumstances, and the majority of the sample is composed of To address this flaw, a computational method called Adaptive-
colder-climate countries (e.g., United Kingdom and France). In Comfort-Control-Implemented Model (ACCIS) was developed
other words, applying this model for warm climates limits the [52]. It was created as part of the conceptual framework called
results and reduces the model’s applicability, especially when tak- Adaptive-Comfort-Control-Implemented Model (ACCIM), which
ing into account the studies on the effects of climate change aimed to change the paradigm, from using fixed PMV-based to
[13,14]. Even though the cases studies database takes into account adaptive setpoint temperatures. Additionally, earlier versions of
examples from all across the world, many of the ASHRAE 55-2020 ACCIS were used, enabling studies based on numerous simulations
cases include mixed-mode and centralised HVAC facilities. Because in different locations and AST, such as 780 [53] or 48,786 [54]

2
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

unique location-AST combinations. The only accessible software or Table 2

procedure at this time is the one described in this article. Other Available comfort models in ACCIS and accim 0.6.2.

methods provide comparable results, but they require physical ComfStand No. ComfStand Name Area Reference
labour, which is time-consuming, exhausting, and prone to mis- 0 CS_ESP CTE Spain [58]
take. Some of the key research articles for this computational 1 CS_INT EN16798 Europe [10]
approach are compiled in Table 1. 2 CS_INT ASHRAE55 Worldwide [11]
The advantages of accim compared to previous methods for 3 CS_JPN Rijal Japan [29]
4 CS_CHN GBT50785 Cold China [15]
applying adaptive setpoint temperatures are: 5 CS_CHN GBT50785 HotMild China
6 CS_CHN Yang China [25]
– The high customization potential: the multiple arguments allow 7 CS_IND IMAC C NV India [31]
the user to transform static setpoints into adaptive based on 11 8 CS_IND IMAC C MM India
9 CS_IND IMAC R 7DRM India [32]
parameters;
10 CS_IND IMAC R 30DRM India
– The implementation of setpoints ‘‘on the go”: with previous 11 CS_IND Dhaka India [30]
methods, the daily setpoints needed to be calculated in 12 CS_ROU Udrea Romania [41]
advance; however, accim allows to implement the adaptive set- 13 CS_AUS Williamson Australia [34]
points as the simulation runs, since the EnergyManagementSys- 14 CS_AUS DeDear Australia [33]
15 CS_BRA Rupp NV Brazil [36]
tem objects computes them in each simulation timestep.
16 CS_BRA Rupp AC Brazil
17 CS_MEX Oropeza Arid Mexico [39]
Until recently, only international models (namely ASHRAE 55 18 CS_MEX Oropeza DryTropic Mexico
and EN16798-1) and the Spanish Building Code could be selected 19 CS_MEX Oropeza Temperate Mexico
20 CS_MEX Oropeza HumTropic Mexico
in ACCIS. However, ACCIS has recently been updated, and in accim
21 CS_CHL Perez-Fargallo Chile [40]
0.6.2, multiple local comfort models developed for specific areas
can be used, which has been collected in Table 2. In this table,
the ComfStand Name has been coded following the pattern
by engineers, architects, and researchers to model how buildings
‘‘CS_Country Code Model name or author other specifications”,
consume energy and water. This software runs on a console and
where CS stands for ComfStand, the argument used to select the
processes input data from the user before writing it to text files.
comfort standard. This pattern has been kept for consistency with
It was first developed in FORTRAN, but as of version 8.2.0, it has
the naming code for the output IDF files generated by accim.
been converted to C++. Due of the complexity of this software
The EnergyManagementSystem (EMS) script ACCIS and the
and its programming language, graphical user interfaces like
Python module accim have already been investigated in another
DesignBuilder or OpenStudio [33] have been developed to assist
study [52]. However, that research only offered a broad description
non-programmers.
for each, not an explanation of the specifics. This research offers a
EnergyPlus is a powerful programme, but with limited modifi-
thorough understanding of ACCIS by considering the usage of var-
cation possibilities. However, there is a built-in element called
ious items in the EnergyPlus and EnergyManagementSystem envi-
EnergyManagementSystem (EMS) that makes up for this short-
ronments, as well as the functioning and interactions of various
coming. This is a powerful tool for developing custom modelling
variables. The novelty also resides in the local comfort models that
and control procedures as well as for giving supervisory control
now can be used to generate building energy models with setpoint
over particular EnergyPlus modelling domains. However, because
temperatures. The aforementioned points are covered in Section 2,
it requires the development of computer programmes in the Ener-
along with a brief overview of the software that ACCIS depends on,
gyPlus Runtime Language (ERL), which is required to describe the
a brief description of the many objects used in the EnergyManage-
control algorithms, it is challenging to use.
mentSystem environment, an explanation of ACCIS’s limits, and a
clarification of how ACCIS functions, split down into 4 stages. The
results are presented in Section 3, in which the main three argu- 2.2. EnergyManagementSystem objects
ments are studied, and a demonstration of capabilities is provided.
The debate is presented in Section 4, and the conclusions in It is necessary to explain some EnergyManagementSystem
Section 5. items in order to properly comprehend how ACCIS functions:
EnergyManagementSystem:Sensor (hereinafter, sensors) objects are
used to gather information from model components that could
2. Methodology vary over the course of a single run period for use in control calcu-
lations. EnergyManagementSystem:Actuator (hereinafter, actuators)
2.1. Methods: EnergyPlus and EnergyManagementSystem objects are used to actuate some EnergyPlus functionalities. They
typically replace existing functionalities rather than introducing a
The US Department of Energy created EnergyPlus, the ACCIS new set of EMS-aware controls and component models.
simulation engine, as an open-source BES application. It is used EnergyManagementSystem:Program objects hold the overridden

Table 1
Relevant research articles related to ACCIS.

Authors Citation Setpoints Energy Saving (%) Level of automation

R.P. Kramer et al. [7] Adaptive setpoints (based on Adaptive Thermal 74 Not specified
Guideline)
Sánchez-Guevara [9] Adaptive setpoints (ASHRAE Standard 55-2013) 80 (Cooling); 20 Low (based on several monthly simulations and
Sánchez et al. (Heating) following merge)
Sánchez-García et al. [55] Adaptive setpoints (EN 15251:2007) between 23 and 46 Medium (based on Schedule:Compact objects)
Sánchez-García et al. [56] Adaptive setpoints (EN 15251:2007) between 31 and 70 Medium (based on Schedule:Compact objects)
Sánchez-García et al. [57] Adaptive setpoints (EN 15251:2007) between 40 and 62 Medium (based on Schedule:Compact objects)
Sánchez-García et al. [52] Adaptive setpoints (EN 16798-1:2021) 82.8 High (by means of accim)

3
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

behavior that is applied to the simulation run using a variety of 2.4. ACCIS operation
statements, including set-statements to assign variables, if-blocks
to alter behavior based on conditions, or while-blocks to iterate The ACCIS EMS code, added to the IDF file when accim 0.6.2 is
as long as a condition is satisfied. The time for when the ERL pro- used, is included in Appendix A.1 of this document. There are four
grams are executed, at some particular EMS Calling Points, is main stages in ACCIS: (i) Stage 1, where some user-defined data is
described using EnergyManagementSystem:ProgramCallingManager stored at some global variables, and sensors, actuators, and global
(hence, program calling manager) objects. These objects are used variables are added; (ii) Stage 2, where some programs are run to
to describe the timing for when the ERL programs are run. In order apply the user-defined input data to set up the customized out-
to declare variables with a user-defined name and a global scope, puts; (iii) Stage 3, where data to be reported is computed and actu-
EnergyManagementSystem:GlobalVariable (hereafter, global vari- ators override the values of the components; and (iv) Stage 4,
able) objects must be used. EnergyManagementSystem:OutputVari where the output variables are added to show which variables
able (hereafter, output variable) objects are used to create a custom need to be reported. The operation is described in detail below,
output variable that is mapped to an EMS variable. The normal and it is advised to read the script in Appendix A.1, in which line
EnergyPlus output mechanisms, like the Output:Variable object, numbers referred to the code are indicated, and flowcharts (Figs. 1
can then be used to report the custom output variable to the out- and 2) as this section is read. The highlighted syntax has been
put file. maintained, while comments and clarifications in the code are
shown by a ‘‘!” symbol in grey.

2.4.1. Stage 1
2.3. Limitations (i) First, the SetInputData Program object and the Pro-
gramCallingManager are added (lines 5 to 33). As implied by the
This section presents the limitations for accim 0.6.2. The build- name, its purpose is to store the input data using set-statements
ing for which the script has been written only contains one zone so that the simulation run can be customized later. ComfStand,
(designated Block1:Zone1) and one window (designated Block1_- CAT, ComfMod, HVACmode, VentCtrl, VSToffset, MinOToffset, Max-
Zone1_Wall_Win) in order to simplify the script. The EMS sensors, WindSpeed, ACSTtol, and AHSTtol are the global variables where
actuators, program calling managers, programs, output variables, the input data is saved. The documentation page [19] provides a
and global variables connected to each particular zone or window detailed explanation of these many characteristics. Some input
must therefore be added for all buildings with numerous zones data variables, which are used later in conditional statements,
and/or windows. The name of the room or window appears in regard the input values as categorical data. These variables are:
the names of each of these things. Additionally, the HVAC system
interior units need to be watched over in order to use mixed mode. - ComfStand is used to control the comfort standard. It can range
To do this, a VRF system must be individually designed for each from 0 to 21, based on Table 2;
zone, with the Outdoor units bearing names in the format ‘‘VRF - CAT is used to control the occupant expectations. It depends on
Outdoor Unit_BlockX_ZoneY”. the ComfStand chosen, but 1, 2, or 3 can be stored for those
In addition to the aforementioned, ACCIS needs the following developed similarly to EN16798-1, otherwise 80 or 90 can be
elements to function properly: stored for those developed similarly to ASHRAE 55;
- ComfMod is used to control if limits are PMV-based (in that case,
– For EnergyPlus to calculate the prevailing mean outdoor tem- 0 is stored), or adaptive (in which 1, 2 or 3 can be stored, which
perature (PMOT) and running mean outdoor temperature controls the static behaviour of the setpoints when the adaptive
(RMOT), People objects must have the values ‘‘AdaptiveASH5500 model is not applicable).
and ”AdaptiveCEN1525100 in the ‘‘Thermal Comfort Model 1 - HVACmode is used to control the operating mode of the build-
Type and 2 Type” fields. ing. Natural ventilation is not allowed and the full air-
– ‘‘ThermostatSetpoint:DualSetpoint” objects must be used to conditioning setting is activated if 0 is stored; free-running
assign Schedule:Compact objects with the name pattern mode is applied, and natural ventilation is allowed if 1 is stored;
‘‘FORSCRIPT_AHST_BLOCKX_ZONEY” to the associated mixed-mode is applied if 2 is stored, therefore natural ventila-
‘‘BLOCKX_ZONEY” in the IDF file. tion is prioritised, otherwise windows are closed and HVAC sys-
– All zones must have access to natural ventilation, windows tem starts to work.
must constantly be open, and the ventilation technique must - VentCtrl is used to control the ventilation mode. The ventilation
be ‘‘Calculated” for mixed-mode to function. If the weather is setpoint temperature is set to the comfort temperature if 0 is
not favourable, ACCIS will close the windows. stored. If 1, the top comfort limit is assigned. Ventilation beha-
– It is recommended to name the zones as short as possible, viours for values 2 and 3 are are similar respectively to 0 and 1,
because the lines of any EMS program cannot exceed 100 however in this case doors and windows are opened based on
characters. the customised venting opening factor.
– The language of the Graphical User Interface (GUI) must be
English. The remaining variables in the supplied data are all regarded as
numbers. Since adding this script to the IDF file was the only way
These prerequisites must be satisfied in order to apply setpoint to utilize it prior to the creation of accim, the user had to modify
temperatures using either copying and pasting ACCIS in the IDF or these settings to customize the simulation.
in the DesignBuilder Scripting module. Therefore, using the Python Similar to SetInputData, the SetVOFinputData Program and Pro-
package accim is advised since the limitations above except the gramCallingManager objects are added (lines 35 to 49), although in
language of the GUI are not applicable in that case. It automatically this case, the global variables where input data is saved are Max-
verifies that the IDF file complies with the limits indicated earlier TempDiffVOF, MinTempDiffVOF, and MultiplierVOF. These arguments
and, if not, automatically modifies the IDF to ensure that the oper- are used to modulate the gradual opening of the windows and
ation is correct. The open-source code of accim is available at the doors depending on the outdoor temperature. When MaxTempDiff-
web repository [59], and a user manual is available in the docu- VOF (the maximum temperature difference for ventilation), win-
mentation [60]. dows and doors remain open the fractional value entered in
4
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Fig. 1. Stages 1 and 2 of ACCIS.

MultiplierVOF. On the other hand, when the temperature difference (iii) Then, a number of sensors are attached (lines 129 to 199),
is smaller than MinTempDiffVOF (the minimum temperature differ- which will be utilized to gather data from certain components
ence for ventilation) but not a negative value, windows and doors (Table 3). In light of this, these will make it possible to determine
are fully opened. Therefore, these programs were initially placed at the values for the RMOT, PMOT, and for each thermal zone and
the beginning of the script to distinguish them from the non- window, the zone operational temperature, the cooling rate, the
editable code. heating rate, the outdoor air wind speed, and the outdoor air
(ii) To declare all of the many variables utilized by ACCIS, a glo- dry-bulb temperature at each timestep. In the case of the zones,
bal variable object is created (line 54 to 124). While some of these these last five components are related to the zones themselves or
connect to a certain zone, others do not. The Appendix A.1 contains the VRF indoor units situated in those zones, and in the case of
further information for each global variable. the windows, these final five components refer to the zones where

5
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Fig. 2. Stages 3 and 4 of ACCIS.

6
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Table 3 sensors, and actuators have been added. The program calling man-
List of sensors. agers, who will choose when to call the programs, are also
Name Output: Output:Variable or included.
Variable or Output:Meter Name The programs that are used to set the adaptive setpoint temper-
Output:Meter atures are incorporated first. Based on the input data supplied at
Index Key
Name
the SetInputData program, these will calculate and save the rele-
vant values to various global variables.
RMOT People Block1: Zone Thermal Comfort
Zone1 CEN 15251 Adaptive
The SetComfTemp program (lines 230 to 295) is used to store at
Model Running Average the global variable ComfTemp the comfort temperature at each
Outdoor Air timestep. Depending on the value given to ComfStand in the SetIn-
Temperature putData program, the comfort temperature for the selected comfort
PMOT People Block1: Zone Thermal Comfort
model or standard will be stored. All the comfort models or stan-
Zone1 ASHRAE 55 Adaptive
Model Running Average dards are gathered in Table 2, as well as their associated ComfStand
Outdoor Air value. Mostly, the PMOT sensor is utilized to determine the comfort
Temperature temperature, however there are some cases in which the RMOT
Block1_Zone1_OpT Block1:Zone1 Zone Operative sensor is used (i.e., when EN16798-1 is used, or when the regres-
Temperature
sion equation has been developed for 30-days-long weighted mean
Block1_Zone1_CoolCoil Block1:Zone1 Cooling Coil Total
VRF Indoor Cooling Rate outdoor temperatures).
Unit DX The SetAppLimits program (lines 297 to 435) is used to store the
Cooling Coil values of the applicability limitations in the global variables cre-
Block1_Zone1_HeatCoil Block1:Zone1 Heating Coil Heating
ated for that purpose (AHSTaul, AHSTall, ACSTaul and ACSTall). The
VRF Indoor Rate
Unit DX applicability thresholds will be established for the associated com-
Heating Coil fort model based on the value supplied to ComfStand as well. The
Block1_Zone1_WindSpeed Block1:Zone1 Zone Outdoor Air Wind applicability limitations are set to 50 if the recorded value does not
Speed fall under any of the categories mentioned above (which means the
Block1_Zone1_OutT Block1:Zone1 Zone Outdoor Air
model will not be applicable).
Drybulb Temperature
Block1_Zone1_Win_OpT Block1:Zone1 Zone Operative The ApplyCAT program (lines 437 to 598) is used to record the
Temperature values of the offset for the heating and cooling setpoint tempera-
Block1_Zone1_Win_CoolCoil Block1:Zone1 Cooling Coil Total tures from comfort temperature in the global variables created
VRF Indoor Cooling Rate
specifically for this purpose (AHSToffset and ACSToffset). These also
Unit DX
Cooling Coil
rely on the settings that were previously saved in CAT and
Block1_Zone1_Win_HeatCoil Block1:Zone1 Heating Coil Heating ComfStand.
VRF Indoor Rate The setpoint temperatures for the global variables AHST and
Unit DX ACST are calculated and stored using the SetAST program (lines
Heating Coil
600 to 2743). ComfStand and ComfMod values that were previously
Block1_Zone1_Win_WindSpeed Block1:Zone1 Zone Outdoor Air Wind
Speed stored, together with those of other sensors and global variables,
Block1_Zone1_Win_OutT Block1:Zone1 Zone Outdoor Air are used to select the desired setpoint temperature. This program
Drybulb Temperature is the longest and one of the most significant since it has 101 if-
OutT Environment Site Outdoor Air Drybulb
blocks to pass through all ComfStand, CAT and ComfMod values,
Temperature
and uses the set-statement patterns to compute and save the val-
ues for the setpoints (Equations (1) and (2)).
the windows are placed or the VRF indoor units of those zones.
According to the EnergyPlus Input Output Reference, the RMOT is set ACST ¼ RMOT  m þ n þ ACSToffset þ ACSTtol ð1Þ
the weighted mean outdoor temperature of the previous 7 days,
while in case of the PMOT, the number of days prior to the day
in question is 30. set AHST ¼ RMOT  m þ n þ AHSToffset þ AHSTtol ð2Þ
(vi) Afterwards, some actuators are also introduced (lines 203 to In both equations, the directive ‘‘set” instructs ERL to save the
226, Table 4). With these, it will be possible to change the open or value; the global variables for the adaptive cooling and heating set-
closed state of the current window (with Block1_- point temperatures are ACST and AHST, respectively; ‘‘RMOT” is the
Zone1_Win_VentOpenFact) as well as the values of the schedules sensor for the running mean outdoor temperature, which can also
allocated to the heating and cooling setpoint temperatures for alternatively be ‘‘PMOT”; ‘‘m” and ‘‘n” are coefficients of the com-
the existing zone (ACST_Act_Block1_Zone1 and AHST_Act_Block1_- fort model regression equation (for example, 0.33 + 18.8 in case
Zone1, respectively). of EN 16798–1 adaptive model); ‘‘ACSToffset” and ‘‘AHSToffset” are
global variables for the offset from comfort temperature previously
2.4.2. Stage 2 stored in ApplyCAT; and ‘‘ACSTtol” and ‘‘AHSTtol” are global vari-
The following objects are the programs, which will ascertain the ables added to allow for certain tolerances, stored in SetInputData.
link between the previously added sensors and actuators and These tolerances, whose typical values might be ±0.1 °C, are
define the various global variables stated, once the global variables, intended to be utilized to guarantee that temperature in all

Table 4
List of actuators.

Name Actuated Component Unique Name Actuated Component Type Actuated Component Control Type
ACST_Act_Block1_Zone1 ACST_Sch_Block1_Zone1 Schedule:Compact Schedule Value
AHST_Act_Block1_Zone1 AHST_Sch_Block1_Zone1 Schedule:Compact Schedule Value
Block1_Zone1_Win_VentOpenFact Block1:Zone1_Win AirFlow Network Window/Door Opening Venting Opening Factor

7
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

timesteps is within the comfort zone because the simulated oper- Zone1 based on the value of the global variable HVACmode, stored
ating temperature can occasionally be a few decimals of degrees at SetInputData. The actuators ACST_Act_Block1_Zone1 and
higher than the setpoint temperatures. AHST_Act_Block1_Zone1 are given the values of ACST and AHST,
The setpoint temperatures without tolerances (AHSTnoTol and respectively, if HVACmode is 0; 100 and 100, respectively, are
ACSTnoTol), which are the same value as the adaptive comfort lim- assigned to the cooling and heating setpoint actuators if HVACmode
its, are computed and stored using the SetASTnoTol program (lines is 1; and if HVACmode is 2, and the previously evaluated environ-
2745 to 2753). AHSTnoTol must store a positive value, whereas mental conditions are not acceptable, ACST and AHST are also given
ACSTnoTol must store a negative value. to the cooling and heating actuators. Additionally, the global vari-
To apply the mixed-mode and the ventilation setpoint temper- able VentHours_Block1_Zone1 stores the period of time during
ature at global variables MinOutTemp and VST, the SetVST program which ventilation is permitted. Since the scope of the local variable
(lines 2755 to 2795) is used to compute and save the minimum VentHours_Block1_Zone1 is exclusive to this application, it has not
outside temperature. When the adaptive model is applicable, the been defined. All sensors are connected to thermal zones since this
value for MinOutTemp is determined by subtracting the global vari- software operates at the zone level (all sensors include the zone
able MinOToffset (stored at SetInputData program) from the AHST, name as a prefix or suffix).
while the value for VST is determined by adding the global variable The structure and goals of SetWindowOperation_Block1_Zone1_
VSToffset (also stored at SetInputData program); otherwise, it is Win (lines 3035 to 3145) are very similar to those of the preceding
stored as the average of the two setpoint temperatures. programme ApplyAST_Block1_Zone1, with the exception that the
Finally, the program SetMyVOF is added. Since it is used to grad- overridden component is the state of the windows, which is deter-
ually modulate the fractional opening of windows and doors as a mined by the actuator Block1_Zone1_Win_VentOpenFact (1 is open,
function of the difference between outdoor and indoor tempera- 0 is closed). Since the software in this instance operates at the win-
tures, firstly computes the slope of the function based on the global dow level, all sensors are connected to the window and use the
variables MultiplierVOF, MinTempDiffVOF and MaxTempDiffVOF. window name as a prefix or suffix.
After calculating the actual temperature difference, it is input in
an if-block to decide to what extent windows and doors should 2.4.4. Stage 4
be opened, and this fractional value is stored in the global variable The output variables for the stated global variables are then
Block1_Zone1_Win_MyVOF. included (lines 3149 to 3330). These objects are used to give Ener-
gyPlus data on the global variables linked to EMS that must be
2.4.3. Stage 3 given at the simulation results. The complete list of output variable
At this point, the setpoint temperatures have already been objects is shown in Table 5 with the global variable’s name in the
determined using all programs. Therefore, to execute computa- field labelled ‘‘EMS Variable Name” and the name and reporting
tions or modify the setpoint temperatures schedule object and units in the fields labelled ‘‘Name” and ‘‘Units”.
the window’s open or closed state, all of the following programs The distinction between ‘‘Output:Variable” and ‘‘-
are utilized. EnergyManagementSystem:OutputVariables” objects must be made.
When the related zone’s operative temperature (i) is within The usage of ‘‘Output:Variable” objects extends beyond EMS, and
thermal comfort limits, when it is not due to (ii) heat or (iii) cold they specify the variables that must be reported.
but is still applicable, and when it is not applicable due to (vi) heat
or (v) cold, the CountHours_Block1_Zone1 program (lines 2826 to
2905) computes these times and stores them in the global vari- 3. Results
ables (i) ComfHours_Block1_Zone1, (ii) DiscomfAppHotHours_Bloc
k1_Zone1, (iii) DiscomfAppColdHours_Block1_Zone1, (vi) 3.1. Customization potential of the main arguments
DiscomfNonAppHotHours_Block1_Zone1, and (v)
DiscomfNonAppColdHours_Block1_Zone1, added for that reason. ACCIS has a great potential for the development of parametric
All other variables at this time are either sensors or previously studies based on adaptive thermal comfort: firstly, it allows the
saved global variables. simulation at any location, since adaptive setpoint temperatures
Similar to the previous programme, CountHoursNoApp_Block1_- are calculated ‘on the go’; secondly, the wide range of arguments
Zone1 (lines 2907 to 2922) only checks to see if the operating tem- it admits provide an extensive scope of different settings, cus-
perature is within comfort limits (stored at the global variables tomised by the user. The latter is demonstrated in this section, con-
ACSTnoTol and AHSTnoTol), and if it is, it stores the time at the glo- sidering the version of accim is 0.6.2, by means of the study of the
bal variable ComfHoursNoApp_Block1_Zone1. Otherwise, it ignores main 3 parameters that have an influence on the adaptive setpoint
the applicability limits. temperatures: ComfStand, CAT and ComfMod.
Five if-blocks with many nested if-blocks make up
ApplyAST_Block1_Zone1 (lines 2924 to 3033). The first and second 3.1.1. ComfStand: Thermal comfort models involved
if-blocks are used to gather conditions into a single variable, Firstly, the argument ComfStand allows to define the thermal
respectively related to the non-usage of heating and cooling sys- comfort model or standard to work with. Available models are
tems and some environmental conditions, since nested if-blocks those shown in Table 2. In order to show the fitting of the oper-
are allowed only up to 5 levels in EMS. The third determines if ative temperature to the heating and cooling setpoint tempera-
the environment is suitable for mixed-mode ventilation, and if it tures, a number of building performance simulations has been
is, 1 is saved as a boolean at the local variable Ventilates_HVAC carried out, which are shown in Figs. 3 and 4. These figures
mode2_Block1_Zone1 for usage at the fifth if-block. The fourth if- show all models and standards from Table 2. Specifically, Fig. 3
block, which is dependent on the global variable VentCtrl saved shows the models with PMOT, while Fig. 4 depicts the models
at SetInputData, is also used to determine if the surrounding cir- with RMOT. These simulations have been performed using suit-
cumstances are suitable for ventilation, but in this case in free- able EPW files for each model: for Australia, Sydney; for Brazil,
running mode. If the circumstances are suitable, 1 is saved as a Florianópolis; for China, Turpan and Hong Kong; for Chile, Con-
boolean in a new local variable called Ventilates_HVACmode1_Bloc cepcion; for Spain, Madrid; for India, Ahmedabad; for interna-
k1_Zone1. The fifth if-block modifies the values of the Schedule: tional models (EN16798-1 and ASHRAE 55), London; for Japan,
Compact objects ACST_Sch_Block1_Zone1 and AHST_Sch_Block1_- Tokyo; for Mexico, Chihuahua, Acapulco, Veracruz and Mexico
8
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Table 5
List of EMS output variables.

Name EMS Variable Name Units

Comfort Temperature ComfTemp C
Adaptive Cooling Setpoint Temperature ACST C
Adaptive Heating Setpoint Temperature AHST C
Adaptive Cooling Setpoint Temperature_No Tolerance ACSTnoTol C
Adaptive Heating Setpoint Temperature_No Tolerance AHSTnoTol C
Ventilation Setpoint Temperature VST C
Minimum Outdoor Temperature for mixed mode ventilation MinOutTemp C
Comfortable Hours_No Applicability_Block1_Zone1 (summed) ComfHoursNoApp_Block1_Zone1 h
Comfortable Hours_Block1_Zone1 (summed) ComfHours_Block1_Zone1 h
Discomfortable Applicable Hot Hours_Block1_Zone1 (summed) DiscomfAppHotHours_Block1_Zone1 h
Discomfortable Applicable Cold Hours_Block1_Zone1 (summed) DiscomfAppColdHours_Block1_Zone1 h
Discomfortable Non Applicable Hot Hours_Block1_Zone1 (summed) DiscomfNonAppHotHours_Block1_Zone1 h
Discomfortable Non Applicable Cold Hours_Block1_Zone1 (summed) DiscomfNonAppColdHours_Block1_Zone1 h
Ventilation Hours_Block1_Zone1 (summed) VentHours_Block1_Zone1 h

City respectively for climates arid, dry tropical, humid tropical 3.1.3. ComfMod: Control of the setpoint temperature behavior
and temperate; and for Romania, Bucharest. The building energy Lastly, the argument ComfMod allows to control the behavior of
model used for these has been the ASHRAE 90.1–2016 DOE Com- the setpoint temperature. For almost all models, including the
mercial Reference Building Prototype Small Office, since a typical CS_IND IMAC C NV (Fig. 6, left), the control of the behavior is as
building would be different for all countries and comfort models. follows:
Therefore, the purpose is the demonstration of the potential for
parametric studies rather than the presentation of the simulation - if ComfMod is 0, setpoint temperatures take the values of the
results. In all of these, different slopes of the heating and cooling PMV-based regulation or model relevant for each specific area;
setpoint temperatures can be observed, which shed light on the - if ComfMod is 1, 2 or 3, the adaptive model selected in ComfS-
different capabilities of adaptation to the indoor environment of tand is used when PMOT or RMOT falls within applicability lim-
the people from these different areas. The only models which are its; otherwise, when it falls outside:
not adaptive are CS_BRA Rupp AC, a model developed by Rupp o if ComfMod is 1, the previously explained PMV-based regu-
et al [36] for air-condition spaces, and CS_ESP CTE, which is lation or model relevant for that specific area is used;
the Spanish Building Code. Except for the latter, all simulations o if ComfMod is 2, PMV-based setpoint temperatures based on
have been performed using the 80% acceptability level (in case worldwide EN-ISO 7730 [61];
of the CS_INT EN16798, the equivalent category is II), and the o If ComfMod is 3, the heating and cooling setpoint tempera-
ComfMod argument has taken the value 3, therefore horizontally tures are horizontally extended.
extending the heating and cooling setpoint temperature when
the PMOT or RMOT exceeded the applicability limits. The main These exceptions are CS_ESP CTE, CS_INT EN16798 and CS_INT
advantage of this parameter is that it allows to select the com- ASHRAE55, and the reason of these different controls is that this
fort standard or model among a wide range of possibilities, con- tool was initially designed to be used mainly in the Spanish energy
sidering not only international models, but also local models, framework. However, these are likely to change in the future for
and therefore the new possibilities it opens, since it allows to coherence with the other models. In case of CS_ESP CTE, there
bring a totally novel field of research. are no different controls based on ComfMod. In case of the interna-
tional standards, the differences with the previously explained
behavior occur when the PMOT or RMOT falls outside applicability
limits. For clarity purposes, all behaviors for all models are shown
3.1.2. CAT: Occupant expectations or acceptability levels in Appendix A.2. The main advantage of this parameter is that it
Secondly, the argument CAT allows the user to define the allows to select the desired behavior of the setpoint temperatures:
acceptability levels. Fig. 5 shows the different acceptability levels whether these should be static and take relevant values for that
available, which depends on the ComfStand selected, since CAT country based on national building energy regulations, or whether
has been developed by taking the thresholds and range values con- these should be adaptive, and in that case, the static setpoint tem-
sidered in each original model, referenced in Table 2. This figure peratures that to be considered if the adaptive model falls out of
shows on the left the CS_IND IMAC C NV model (which stands the applicability range.
for India Model for Adaptive Comfort, specifically spaces with com-
mercial use and natural ventilation), with 3 acceptability levels
defined in the model: 80%, 85% and 90%, which are translated 3.2. Demonstration of capabilities
respectively into an offset of ±4.1 °C, ±3.3 °C and ±2.4 °C from neu-
tral or comfort temperature; and, on the right, the CS_INT In this section, an example of the use of accim is shown with the
EN16798, also with 3 acceptability levels (although defined as cat- purpose of demonstrating the capabilities of ACCIS and the Python
egories or occupant expectation in the standard): 3, 2 and 1 (or III, library accim, as well as allowing the reader to replicate this sec-
II and I), which are projected respectively as an offset of 5°C and tion with its own IDF and EPW files. This example shows the rec-
4 °C, 4°C and 3 °C, and 3°C and 2 °C from neutral temperatures, ommended method for applying adaptive setpoint temperatures,
for the heating and cooling setpoint temperatures. The available which is the use of the Python library accim rather than the man-
acceptability levels for each ComfStand can be consulted in Appen- ual addition of the EMS script ACCIS to the IDF file.
dix A.2. The main advantage of this parameter is that it allows to Considering Python 3.9 and EnergyPlus (any version from 9.0
easily select the acceptability levels to be applied to the setpoint to 22.2) must be previously installed, the first step is installing
temperatures. accim, if it is not already installed. To do this, the user needs
9
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Fig. 3. Thermal comfort standards based on the PMOT as a function of the ComfStand.

to open a CMD terminal, and enter ‘‘pip install accim”. development environment (IDE), importing the library and call-
Afterwards, the second step is running Python or an integrated ing the function. There are multiple procedures to do this, but

10
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Fig. 4. Thermal comfort standards based on the RMOT as a function of the ComfStand.

Fig. 5. Thermal comfort acceptability levels for CS_IND IMAC C NV and CS_INT EN16798 as a function of the argument CAT.

one of the simplest is opening a CMD terminal pointing at the

path where the IDF or IDFs to be transformed are located and
then running Python by entering ‘‘python”. In this case, the
Small Office model as well as two copies of it with some modi-
fications has been used, which have been renamed to Model_A,
Model_B and Model_C. Then, the library should be imported
and the function should be called, which can be done mainly
by means of two different methods:

- The shorter and easier method, designed mainly to be used by

users without programming experience: with this method, the
user will be asked to enter all the required information on the
CMD terminal. The code to be entered is:

- The longer method, which consists of the definition of all the In this case, the longer method has been used. The user have
required arguments when calling the function. The code to be asked accim to generate output IDF files for the IDFs Model_A,
entered, for the version of accim 0.6.2 is: Model_B and Model_C considering:

11
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

Fig. 6. Behaviour of the heating and cooling setpoint temperatures as a function of the argument ComfMod.

- With regard to the comfort standard (ComfStand), the EN mentioned in this study, are beyond the scope of ACCIS, since these
16798-1, ASHRAE 55, the IMAC-C, and the local adaptive model are related to Python algorithms and EnergyPlus objects rather
developed by Ricardo Forgiarini Rupp et al. for Florianopolis than merely EnergyManagementSystem objects. Nevertheless, all
(Brazil) [36] (respectively values 1, 2, 9 and 15); arguments are carefully explained in the documentation [60], open
- With regard to the acceptability levels (CAT), all main expecta- source and available for the user.
tions, which are categories I, II and III (values 1, 2 and 3) for As a result, the number of input IDFs was 3, and the size of each
EN16798-1, and 80 and 90% acceptability levels for ASHRAE one 217 KBs, while the number of output IDFs has been 108 (full
55 and IMAC-C and Brazilian model (values 80 and 90); list in Appendix A.3) and the size of each one 546 KBs. The process
- With regard to the behaviour of setpoints (ComfMod), all avail- has taken 18.32 s, although it depends on the number and size of
able types (values 0 to 3); the input IDFs and the performance of the computer.
- With regard to the building operation mode (HVACmode), full
air conditioning and naturally ventilated (respectively values
4. Discussion
0 and 1).
The fact that adaptive setpoint temperatures are not first com-
The output IDFs are named based on the data the user have
puted for a particular environment and then placed in the relevant
entered for the arguments joined with the characted ‘‘[‘‘, so that
Schedule:Compact object is one of the most significant benefits of
it can be separated in categories in the data analytics stage, there-
ACCIS. Since EnergyPlus calculates and stores the running and pre-
fore following the pattern:
vailing mean outdoor temperatures values at the RMOT and PMOT
Name[ComfStand[CAT[ComfMod[HVACmode[VentCtrl[VSToff
sensors as the simulation runs, then use these values to calculate
set[MinOToffset[MaxWindSpeed[ASTtol[NameSuffix.idf
ACST and AHST, and finally override the Schedule:Compact values
Thus, an example is:
associated with the ThermostatSetpoint:DualSetpoint objects with
Model_A[CS_BRA Rupp NV[CA_80[CM_0[HM_0[VC_X[VO_X[M
the stored values for ACST and AHST, this code is applicable to sim-
T_X[MW_X[AT_0.1[NS_X.idf
ulations with any climate file. As a result, the availability of these
Where Model_A is the name; CS refers to ComfStand; CA refers
sensors is necessary for ACCIS to function properly.
to CAT; CM refers to ComfMod; HM refers to HVACmode; VC refers
As a consequence, based on the user-defined parameters ComfS-
to VentCtrl; VO refers to VSToffset; MT refers to MinOToffset; MW
tand, CAT, and ComfMod, all potential setpoint temperatures are
refers to MaxWindSpeed; AT refers to ASTtol; and NS refers to the
provided in Appendix A.2. It is important to remark the high cus-
name suffix. Some of the arguments, which have not been
tomisation possibilities that ACCIS provides, since there are 165
12
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

available ComfStand-CAT-ComfMod combinations, and the remain- Finally, the arguments ComfStand, CAT and ComfMod are stud-
ing 7 arguments have not been studied in this research. All of these ied, and the high customisation potential is presented, and the
setpoint temperatures combine to create what can be referred to as capabilities of accim are demonstrated, providing 108 IDFs with
‘‘hybrid” comfort models, which employ adaptive setpoint temper- adaptive setpoints in less than 20 s. The extensive simulation runs
atures when the adaptive model is applicable and static setpoint it is possible to carry out with ACCIS and the potential energy sav-
temperatures when it is not (except when ComfMod is 0, which is ings open up new opportunities for doing global energy simula-
entirely static). tions as opposed to just national or regional ones. As a result,
Because of ACCIS’s potential for energy savings and the lengthy this tool broadens the scope of several studies on adaptive comfort,
simulation runs it permits, it is now possible to conduct energy such as those looking at energy poverty or climate change
simulations on a global scale as opposed to only on a national or mitigation.
continental size. This instrument therefore creates new possibili-
ties for enlarging the scope of numerous adaptive comfort research
CRediT authorship contribution statement
inquiries, such studies of energy poverty or mitigation strategies
for climate change. To the best of the authors’ knowledge, the pro-
Daniel Sánchez-García: Conceptualization, Methodology, Soft-
posed computational strategy is the first to do this, and it also
ware, Formal analysis, Investigation, Writing – original draft, Writ-
enables faster completion of additional typical adaptive-setpoint
ing – review & editing. Jorge Martínez-Crespo: Conceptualization,
simulations than any other technique. Additionally, this approach
Investigation, Formal analysis. Ulpiano Ruiz-Rivas Hernando:
has been automated and designed to be simple to use. In contrast,
Visualization, Investigation, Validation. Carmen Alonso: Visualiza-
other systems entail executing file management tasks for dozens or
tion, Investigation, Validation.
hundreds of EPW files and accurately simulating IDF files with
adaptive setpoints with the appropriate EPW file.
Data availability

5. Conclusions The authors do not have permission to share data.

Adaptive setpoint temperatures might be a potential mean of Declaration of Competing Interest

enhancing a building’s energy efficiency. Its potential for large
energy savings has been widely shown. Nevertheless, the method The authors declare that they have no known competing finan-
used to carry out most of the studies related to adaptive setpoint cial interests or personal relationships that could have appeared
temperatures was time-consuming and error-prone, given that to influence the work reported in this paper.
most of the processes were carried out manually. This problem
has recently been addressed using a computational technique Acknowledgments
known as ACCIS. There is currently no alternative method that
can automate the process of applying adaptive setpoint tempera- The authors acknowledge funding from the European Commis-
tures. However, ACCIS had certain restrictions since a number of sion through Urban Innovative Actions of the EPIU Getafe Project
conditions had to be met in order for the procedure to be success- under Grant UIA04-212.
ful. In order to automate all of these activities, ACCIS has been
nested in a Python module named accim. As a consequence, ACCIS
Appendix A. Supplementary data
converts PMV-based building energy models into adaptive setpoint
building energy models in accordance with an IDF and the setup
Supplementary data to this article can be found online at
provided by the user. The main advantages of accim are:
https://doi.org/10.1016/j.enbuild.2023.113019.
- The strong customizability potential: The user may change sta-
tic setpoints into adaptive ones depending on 11 factors by Appendix B. Latest version of accim
using multiple arguments;
- The ability to implement setpoints ‘‘on the fly”: Previously, the Please note accim 0.6.2 has been explained in this research arti-
daily setpoints had to be computed in preparation. With accim, cle. For the latest version of accim, visit the following websites:
however, this is no longer necessary because the EnergyMan- - Web repository: https://github.com/dsanchez-garcia/accim
agementSystem objects compute the adaptive setpoints at the - Python Package Index (PyPI): https://pypi.org/project/accim/
start of each simulation timestep. - Documentation: https://accim.readthedocs.io/en/latest/

Following some earlier considerations, this research offers a References

thorough understanding of ACCIS by deconstructing the code into
4 stages: [1] World Wildlife Fund, Living Planet Report 2014: Species and spaces, people
and places, WWF International, Gland, Switzerland, 2014. 10.1007/s13398-
014-0173-7.2.
- Stage 1: the user adds the required EMS items and defines the [2] M. Meinshausen, S.J. Smith, K. Calvin, J.S. Daniel, M.L.T. Kainuma, J. Lamarque,
input data for customising the adaptive setpoint simulation K. Matsumoto, S.A. Montzka, S.C.B. Raper, K. Riahi, A. Thomson, G.J.M. Velders,
D.P.P. van Vuuren, The RCP greenhouse gas concentrations and their
parameters. extensions from 1765 to 2300, Clim. Change. 109 (2011) 213–241, https://
- Stage 2: the user-defined input data is applied by EMS pro- doi.org/10.1007/s10584-011-0156-z.
grammes, and the global variables are computed and stored [3] European Commission, A Roadmap for Moving to a Competitive Low Carbon
Economy in 2050, (2011) 1–15. https://eur-lex.europa.eu/legal-content/EN/
to build the framework of the customised adaptive setpoint
ALL/?uri=CELEX:52011DC0112 (accessed July 18, 2021).
simulation. [4] M.A. Humphreys, Thermal comfort temperatures world-wide - the current
- Stage 3, the global variables already computed in Stage 2 are position, Renew. Energy. 8 (1996) 139–144, https://doi.org/10.1016/0960-
used to override the setpoint schedule and the open or closed 1481(96)88833-1.
[5] T. Parkinson, R. de Dear, G. Brager, Nudging the adaptive thermal comfort
state of the windows. model, Energy Build. 206 (2020), https://doi.org/10.1016/j.
- Stage 4: the variables that will be reported are specified. enbuild.2019.109559.

13
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

[6] G.Y. Yun, J.H. Lee, K. Steemers, Extending the applicability of the adaptive thermal comfort model based on field studies in five climate zones across
comfort model to the control of air-conditioning systems, Build. Environ. 105 India, Build. Environ. 219 (2022), https://doi.org/10.1016/j.
(2016) 13–23, https://doi.org/10.1016/j.buildenv.2016.05.027. buildenv.2022.109187.
[7] R.P. Kramer, M.P.E. Maas, M.H.J. Martens, A.W.M. van Schijndel, H.L. Schellen, [33] R. de Dear, J. Kim, T. Parkinson, Residential adaptive comfort in a humid
Energy conservation in museums using different setpoint strategies: A case subtropical climate—Sydney Australia, Energy Build. 158 (2018) 1296–1305,
study for a state-of-the-art museum using building simulations, Appl. Energy https://doi.org/10.1016/j.enbuild.2017.11.028.
158 (2015) 446–458, https://doi.org/10.1016/j.apenergy.2015.08.044. [34] T. Williamson, L. Daniel, A new adaptive thermal comfort model for homes in
[8] ANSI/ASHRAE, ASHRAE Standard 55-2013 Thermal Environmental Conditions temperate climates of Australia, Energy Build. 210 (2020), https://doi.org/
for Human Occupancy, GA, United States, 2013. 10.1016/j.enbuild.2019.109728.
[9] C. Sánchez-Guevara Sánchez, A. Mavrogianni, F.J. Neila González, On the [35] M. Indraganti, D. Boussaa, An adaptive relationship of thermal comfort for the
minimal thermal habitability conditions in low income dwellings in Spain for a Gulf Cooperation Council (GCC) Countries: The case of offices in Qatar, Energy
new definition of fuel poverty, Build. Environ. 114 (2017) 344–356, https://doi. Build. 159 (2018) 201–212, https://doi.org/10.1016/J.ENBUILD.2017.10.087.
org/10.1016/j.buildenv.2016.12.029. [36] R.F. Rupp, R. de Dear, E. Ghisi, Field study of mixed-mode office buildings in
[10] European Committee for Standardization, EN 16798-1:2019 Energy Southern Brazil using an adaptive thermal comfort framework, Energy Build.
Performance of Buildings. Ventilation for Buildings. Indoor Environmental 158 (2018) 1475–1486, https://doi.org/10.1016/j.enbuild.2017.11.047.
Input Parameters for Design and Assessment of Energy Performance of [37] A. García, F. Olivieri, E. Larrumbide, P. Ávila, Thermal comfort assessment in
Buildings Addressing Indoor Air Quality, Thermal Environment, Lighting and naturally ventilated offices located in a cold tropical climate, Bogotá, Build.
Acoustics, (2019). https://en.tienda.aenor.com/norma-bsi-bs-en-16798-1- Environ. 158 (2019) 237–247, https://doi.org/10.1016/J.
2019-000000000030297474 (accessed August 6, 2021). BUILDENV.2019.05.013.
[11] ANSI/ASHRAE, ASHRAE Standard 55-2020 Thermal Environmental Conditions [38] L.A. López-Pérez, J.J. Flores-Prieto, C. Ríos-Rojas, Adaptive thermal comfort
for Human Occupancy, Atlanta, GA, United States, 2020. model for educational buildings in a hot-humid climate, Build. Environ. 150
[12] R. de Dear, G.S. Brager, Thermal comfort in naturally ventilated buildings: (2019) 181–194, https://doi.org/10.1016/j.buildenv.2018.12.011.
revision to ASHRAE standards 55, J. Energy Build. 34 (2002) 549–561. [39] I. Oropeza-Perez, A.H. Petzold-Rodriguez, C. Bonilla-Lopez, Adaptive thermal
[13] D. Sánchez-García, C. Rubio-Bellido, J. Pulido-Arcas, F. Guevara-García, J. comfort in the main Mexican climate conditions with and without passive
Canivell, Adaptive comfort models applied to existing dwellings in cooling, Energy Build. 145 (2017) 251–258, https://doi.org/10.1016/j.
mediterranean climate considering global warming, Sustainability. 10 (2018) enbuild.2017.04.031.
3507, https://doi.org/10.3390/su10103507. [40] A. Pérez-Fargallo, J.A. Pulido-Arcas, C. Rubio-Bellido, M. Trebilcock, B. Piderit, S.
[14] R. Barbosa, R. Vicente, R. Santos, Climate change and thermal comfort in Attia, Development of a new adaptive comfort model for low income housing
Southern Europe housing: A case study from Lisbon, Build. Environ. 92 (2015) in the central-south of Chile, Energy Build. 178 (2018) 94–106, https://doi.org/
440–451, https://doi.org/10.1016/j.buildenv.2015.05.019. 10.1016/j.enbuild.2018.08.030.
[15] Ministry of Housing and Urban-Rural Development (China), (GB/T 50785- [41] I. Udrea, C. Croitoru, I. Nastase, R. Crutescu, V. Badescu, First adaptive thermal
2012) Evaluation standard for indoor thermal environment in civil buildings, comfort equation for naturally ventilated buildings in Bucharest, Romania, Int.
Standardization Administration of China Beijing, China, 2012. J. Vent. 17 (2018) 149–165, https://doi.org/10.1080/14733315.2017.1356057.
[16] Instituut Voor Studie En Stimulering Van Onderzoek, ISSO-publicatie 74 [42] Y.H. Yau, B.T. Chew, Adaptive thermal comfort model for air-conditioned
Thermische behaaglijkheid, Rotterdam, Netherlands, 2004. hospitals in Malaysia, (n.d.). 10.1177/0143624412474829.
[17] Instituut Voor Studie En Stimulering Van Onderzoek, ISSO-publicatie 74 [43] R. Thapa, H.B. Rijal, M. Shukuya, Field study on acceptable indoor temperature
Thermische behaaglijkheid, 2014. in temporary shelters built in Nepal after massive earthquake 2015, Build.
[18] L. Yang, R. Fu, W. He, Q. He, Y. Liu, Adaptive thermal comfort and climate Environ. 135 (2018) 330–343. 10.1016/J.BUILDENV.2018.03.001.
responsive building design strategies in dry–hot and dry–cold areas: Case [44] Z. Wu, N. Li, P. Wargocki, J. Peng, J. Li, H. Cui, Adaptive thermal comfort in
study in Turpan, China, Energy Build. 209 (2020), https://doi.org/10.1016/j. naturally ventilated dormitory buildings in Changsha, China, Energy Build. 186
enbuild.2019.109678. (2019) 56–70, https://doi.org/10.1016/J.ENBUILD.2019.01.029.
[19] A.C. Boerstra, J. van Hoof, A.M. van Weele, A new hybrid thermal comfort [45] C. Fu, Z. zheng, C.M. Mak, Z. Fang, M.O. Oladokun, Y. Zhang, T. Tang, Thermal
guideline for the Netherlands: background and development, Archit. Sci. Rev. comfort study in prefab construction site office in subtropical China, Energy
58 (2015) 24–34, https://doi.org/10.1080/00038628.2014.971702. Build. 217 (2020) 109958. 10.1016/J.ENBUILD.2020.109958.
[20] F. Nicol, S. Roaf, Pioneering new indoor temperature standards: the Pakistan [46] S. Kumar, M.K. Singh, A. Mathur, M. Košir, Occupant’s thermal comfort
project, Energy Build. 23 (1996) 169–174, https://doi.org/10.1016/0378-7788 expectations in naturally ventilated engineering workshop building: A case
(95)00941-8. study at high metabolic rates, Energy Build. 217 (2020), https://doi.org/
[21] S. Heidari, S. Sharples, A comparative analysis of short-term and long-term 10.1016/J.ENBUILD.2020.109970.
thermal comfort surveys in Iran, Energy Build. 34 (2002) 607–614, https://doi. [47] S. Haddad, P. Osmond, S. King, Application of adaptive thermal comfort
org/10.1016/S0378-7788(02)00011-7. methods for Iranian schoolchildren, (2016). 10.1080/09613218.2016.1259290.
[22] K.W.H. Mui, W.T.D. Chan, Adaptive comfort temperature model of air- [48] S. Thapa, Thermal comfort in high altitude Himalayan residential houses in
conditioned building in Hong Kong, Build. Environ. 38 (2003) 837–852, Darjeeling, India – An adaptive approach, Indoor Built Environ. 29 (2020) 84–
https://doi.org/10.1016/S0360-1323(03)00020-9. 100, https://doi.org/10.1177/1420326X19853877.
[23] Z. Wang, L. Zhang, J. Zhao, Y. He, Thermal comfort for naturally ventilated [49] Y. Jiao, H. Yu, Y. Yu, Z. Wang, Q. Wei, Adaptive thermal comfort models for
residential buildings in Harbin, Energy Build. 42 (2010) 2406–2415, https:// homes for older people in Shanghai, China, Energy Build. 215 (2020), https://
doi.org/10.1016/J.ENBUILD.2010.08.010. doi.org/10.1016/J.ENBUILD.2020.109918.
[24] L. Yang, W. Zheng, Y. Mao, J.C. Lam, Y. Zhai, Thermal adaptive models in built [50] DesignBuilder Software Ltd, DesignBuilder, (2021). https://designbuilder.co.uk
environment and its energy implications in Eastern China, Energy Procedia. 75 (accessed July 18, 2021).
(2015) 1413–1418, https://doi.org/10.1016/J.EGYPRO.2015.07.237. [51] D.B. Crawley, L.K. Lawrie, F.C. Winkelmann, W.F. Buhl, Y.J. Huang, C.O.
[25] L. Yang, R. Fu, W. He, Q. He, Y. Liu, Adaptive thermal comfort and climate Pedersen, R.K. Strand, R.J. Liesen, D.E. Fisher, M.J. Witte, J. Glazer,
responsive building design strategies in dry–hot and dry–cold areas: Case EnergyPlus: Creating a new-generation building energy simulation program,
study in Turpan, China, Energy Build. 209 (2020), https://doi.org/10.1016/j. Energy Build. 33 (2001) 319–331, https://doi.org/10.1016/S0378-7788(00)
enbuild.2019.109678. 00114-6.
[26] C. Bouden, N. Ghrab, An adaptive thermal comfort model for the Tunisian [52] D. Sánchez-García, D. Bienvenido-Huertas, C. Rubio-Bellido, Computational
context: a field study results, Energy Build. 37 (2005) 952–963, https://doi.org/ approach to extend the air-conditioning usage to adaptive comfort: Adaptive-
10.1016/J.ENBUILD.2004.12.003. Comfort-Control-Implementation Script, Autom. Constr. 131 (2021), https://
[27] H.B. Rijal, M. Honjo, R. Kobayashi, T. Nakaya, Investigation of comfort doi.org/10.1016/j.autcon.2021.103900.
temperature, adaptive model and the window-opening behaviour in [53] D. Bienvenido-Huertas, D. Sánchez-García, C. Rubio-Bellido, M.J. Oliveira,
Japanese houses, Archit. Sci. Rev. 56 (2013) 54–69, https://doi.org/10.1080/ Influence of adaptive energy saving techniques on office buildings located in
00038628.2012.744295. cities of the Iberian Peninsula, Sustain. Cities Soc. 53 (2020), https://doi.org/
[28] H.B. Rijal, M.A. Humphreys, J.F. Nicol, Towards an adaptive model for thermal 10.1016/j.scs.2019.101944.
comfort in Japanese offices, Build. Res. Inf. 45 (2017) 717–729, https://doi.org/ [54] D. Bienvenido-Huertas, D. Sánchez-García, C. Rubio-Bellido, J.A. Pulido-Arcas,
10.1080/09613218.2017.1288450. Analysing the inequitable energy framework for the implementation of nearly
[29] H.B. Rijal, M.A. Humphreys, J.F. Nicol, Adaptive model and the adaptive zero energy buildings (nZEB) in Spain, J. Build. Eng. 35 (2021), https://doi.org/
mechanisms for thermal comfort in Japanese dwellings, Energy Build. 202 10.1016/j.jobe.2020.102011.
(2019), https://doi.org/10.1016/j.enbuild.2019.109371. [55] D. Sánchez-García, D. Bienvenido-Huertas, M. Tristancho-Carvajal, C. Rubio-
[30] S. Dhaka, J. Mathur, G. Brager, A. Honnekeri, Assessment of thermal Bellido, Adaptive comfort control implemented model (ACCIM) for energy
environmental conditions and quantification of thermal adaptation in consumption predictions in dwellings under current and future climate
naturally ventilated buildings in composite climate of India, Build. Environ. conditions: A case study located in Spain, Energies. 12 (8) (2019), https://
86 (2015) 17–28, https://doi.org/10.1016/J.BUILDENV.2014.11.024. doi.org/10.3390/en12081498.
[31] S. Manu, Y. Shukla, R. Rawal, L.E. Thomas, R. de Dear, Field studies of thermal [56] D. Sánchez-García, C. Rubio-Bellido, M. Tristancho, M. Marrero, A comparative
comfort across multiple climate zones for the subcontinent: India Model for study on energy demand through the adaptive thermal comfort approach
Adaptive Comfort (IMAC), Build. Environ. 98 (2016) 55–70, https://doi.org/ considering climate change in office buildings of Spain, Build. Simul. 13 (2020)
10.1016/j.buildenv.2015.12.019. 51–63, https://doi.org/10.1007/s12273-019-0560-2.
[32] R. Rawal, Y. Shukla, V. Vardhan, S. Asrani, M. Schweiker, R. de Dear, V. Garg, J. [57] D. Sánchez-García, C. Rubio-Bellido, J.J.M. del Río, A. Pérez-Fargallo, Towards
Mathur, S. Prakash, S. Diddi, S.V. Ranjan, A.N. Siddiqui, G. Somani, Adaptive the quantification of energy demand and consumption through the adaptive

14
D. Sánchez-García, J. Martínez-Crespo, Ulpiano Ruiz-Rivas Hernando et al. Energy & Buildings 288 (2023) 113019

comfort approach in mixed mode office buildings considering climate change, [60] D. Sánchez-García, accim’s documentation, (2021). https://accim.readthedocs.
Energy Build. 187 (2019) 173–185, https://doi.org/10.1016/j. io/en/latest/index.html (accessed April 16, 2021).
enbuild.2019.02.002. [61] ISO, ISO-7730: Ergonomics of the thermal environment Analytical
[58] The Government of Spain, Royal Decree 314/2006. Approving the Spanish determination and interpretation of thermal comfort using calculation of the
Technical Building Code, Madrid, Spain, 2006. PMV and PPD indices and local thermal comfort criteria, Management 3 (2005)
[59] D. Sánchez-García, accim web repository, (2021). https://github.com/ 605–615, https://doi.org/10.1016/j.soildyn.2004.11.005.
dsanchez-garcia/accim (accessed July 19, 2021).

15