BUILDING PERFORMANCE SIMULATION IN DESIGN EDUCATION: DESIGN-INTEGRATED VERSUS ADDITIVE USE

Abstract

In the light of an extensive literature research and the Building Performance Simulation (BPS) in teaching survey, conducted among German universities, it is deduced that using BPS as a part of design process rather than an evaluation tool for existing designs is a prominent way to teach BPS, particularly in architectural education. New methods are needed to adopt BPS not only as a performance evaluator, but also as a design stimulator during design exploration. This paper shares teaching experiences gained through architectural design courses by comparing the additive and the design-integrated uses of BPS.

Full text

BUILDING PERFORMANCE SIMULATION IN DESIGN EDUCATION:
DESIGN-INTEGRATED VERSUS ADDITIVE USE
Isil Kalpkirmaz Rizaoglu1, Karsten Voss1
1 University of Wuppertal, Germany, E-Mail: kalp@uni-wuppertal.de
Abstract
In the light of an extensive literature research and the
Building Performance Simulation (BPS) in teaching
survey, conducted among German universities, it is
deduced that using BPS as a part of design process
rather than an evaluation tool for existing designs is a
prominent way to teach BPS, particularly in
architectural education. New methods are needed to
adopt BPS not only as a performance evaluator, but
also as a design stimulator during design exploration.
This paper shares teaching experiences gained through
architectural design courses by comparing the additive
and the design-integrated uses of BPS.
Introduction
Utilizing design and BPS tools in tandem can be an
important pillar for supporting efficiency, variety and
flexibility of the design process. Most architects meet
BPS during their education, and the experiences
gained throughout play a significant role in terms of
adoption of BPS tools in their future works. However,
BPS is often an additive in design education rather
than an inherent part, therefore new methods are
needed to provide design-integrated use experiences.
In an earlier paper (Kalpkirmaz-Rizaoglu, et al., 2020)
authors shared an extensive literature review
regarding pros and cons of early integration of BPS in
architectural design, as well as the results of the BPS
in Teaching survey, which aimed to find out how BPS
is taught especially in high education in Germany. It
is conducted by the participation of 18 lecturers from
13 different universities, who have been using BPS
tools for teaching in architecture and civil engineering
education. Most mentioned reason for choosing a
specific BPS tool in teaching was ease-of-use.
Another highlight from the survey was that CAD tools
rated as most used design and documentation tools by
72% of the respondents, while Building Information
Modeling (BIM) was the least preferred with 8%.
Another important finding of the survey is that 76% of
the courses are independent courses and 12% are part
of a design studio, while the remaining 12% are a
separate course, but support the design course.
The literature review over the topic corresponds with
the survey results and emphasize the ease-of-use.
While some claims that BPS is not very common in
undergraduate level courses due to the students’ lack
of knowledge on building physics (Soebarto, 2005;
Fernandez‐Antolin, et al., 2020) Some others claim
that available BPS tools are too advanced for
beginners (Delbin, et al., 2006). One author
(Augenbroe, 2011) says that: “In many cases
simulation only needs to be adequate for comparative
analysis of design variants.”. Others (Hensen, et al.,
2011) signify that although design parameters have to
be considered in an integrated manner, different
design stages focus on different parameters. A gradual
interaction with BPS promotes confidence among
students and allows them to move independently to
more specific tools at later stages (Fernandez‐Antolin,
et al., 2020). In early design, BPS is used for detecting
and ranking the key parameters and gaining an insight
for a possible future performance of a building (Hopfe,
et al., 2005). Further, it is important to gain an
awareness regarding consequences of decisions and a
critical understanding of limitations inherent in BPS
tools (Beausoleil-Morrison, et al., 2016).
A need of a stronger link between design and BPS
tools was another most mentioned issue in the
literature. Regarding the early design phase, it is found
out that 3D CAD modeling environments are more
prevalent than BIM environments (Soebarto, et al.,
2015). Due to required high level of detail and
complexity of BIM, architects are more likely start
their design works in CAD. Also, according to the
result of one international survey, CAD tools were the
most used design and documentation tools by
architects with 80%, while BIM tools with less than
30%. (Hopfe, et al., 2017). Also, it is stated (Attia, et
al., 2011; Fernandez‐Antolin, et al., 2020) that
architecture students generally tend to prefer the BPS
tools providing visual representation instead of
numeric representation.
Regarding the interdisciplinary approach, another
study (Fernandez‐Antolin, et al., 2020) states that:
“Energy simulation and architectural design have
traditionally been considered as separate elements in
the design process.”. The aforementioned
international survey reveals that the percentage of the
mentioned courses that BPS is taught as a part of
design studio was only 8% (Hopfe, et al., 2017).
In response, this paper presents methods, which are
developed and tested through design courses in an
architecture master program, for integrating BPS in
architectural design workflows.

Experiences gained through Architecture
Master Courses
Studio - Sustainable Building and Building
Performance (S.NB) is a 2-semester-long master
design course as part 1 and part 2 with a total of 12
credits. And the part 2 is directly addressing BPS.
In this paper, the experiences gained at part 2 through
two separate winter semesters are shared. Not to be
repetitive, S.NB-2 in winter semester (WS) 2020/2021
is called Studio A, and S.NB-2 in WS 2021/2022 is
called Studio B.
Introduction to courses
Structure of the master program supports students with
two preceding courses as a base for the studio.
Although these are not prerequisite for the studio,
students are advised and guided in that order. One of
them (NB.1) is about improving the energy
performance to an annually net energy positive
building, while maintaining architectural quality, via
well-known residential case-studies from the
classical-modern style. The other one (NB.2) is about
simplified indoor climate simulations and real site
measurements by comparison of confidence intervals.
Therefore, it can be said that students come to the
studio with a certain level of BPS knowledge and a
skill of critical thinking. Further, BPS can accompany
the integrated design course (E5) and the master thesis
(MA) based on the content and student’s preference.
The courses in master program that BPS is addressed
can be seen in Figure 1.
Figure 1: The courses that BPS is addressed.
In Studio A, students investigated performance of
their designs, which they created in S.NB-1, by using
BPS tools integrated to design tools. Differently, in
Studio B, students started from scratch for a new
design. In Studio B, the same design and BPS tools
were used, as well as other design and simulation
technics; i.e. parametrization, optimization.
Main performance tasks of the both courses were the
same: climate and site integration, solar energy
utilization, daylight availability and achieving indoor
thermal comfort while minimizing active heating and
cooling.
A design upgrade for an overall performance was
requested as a final work of Studio A, and only a
design proposal for Studio B. Final submission and the
study structure, i.e. individual and group works, were
decided considering the semester duration, the work
load and the level of collaborations between students.
While Studio A phases were structured based on
performance tasks, Studio B phases by the methods
applied. Moreover, in Studio B, students are grouped
for locations with different climate patterns and started
their investigations considering both present and
future climate scenarios of assigned locations for the
same design task.
Methods
Speaking of the teaching methods, student-centered
approach was adopted. Both students and the lecturers
played an active role throughout workshops and
supervision. Students’ learning has been continuously
measured via comprehension questions and
assignments. Booklets and exhibition posters were
aimed as final outputs to highlight the key topics and
have a clear picture of intended learning outcomes. To
balance the theory and the hands-on-sessions, lectures
were held only to give theoretical information, when
needed. Instead of long lectures, case-studies were
used as warm-up sessions before the main tasks. Case-
studies served to refresh the students’ knowledge and
introduce them new BPS tool and simplified
approaches of the course. Also, discussions on default
settings of simulations, as well as simplified methods
were made throughout these sessions. It is thought to
be useful for students to be cognizant of the pros and
cons, and be able to decide when and how to use that
default and /or simplified approaches.
Based-on the experiences gained in Studio A, more
time was invested in workshops in Studio B, in order
to promote learning-by-doing (Fig. 2).
Figure 2: Time percentage spent for activities of
each studio throughout a 14 week-long semester
In the context of this paper, while additive-use refers
to using BPS only for performance evaluation for
advanced or completed designs, mostly via BPS tools
not integrated to architectural design ecosystem,
design-integrated use refers to adopting BPS in design
process as earliest as possible, not only for
performance evaluation, but also as informer and
stimulator, and utilizing BPS tools that is integrated
with design tools. In this regard, for the students
evaluated already existing designs, although they used
a BPS tool, which is available in their design
ecosystem, Studio A can be considered as semi-
additive. On the other hand, the Studio B students
started using BPS tool during their early form
investigations, therefore Studio B was design-
integrated. Examples of solar energy utilization
investigations by using solar radiation analysis from
Studio A and B can be seen in Figure 3. While in

Studio A, only slight revisions were possible for the
fractions of existing building; in Studio B, massing
studies were accompanied by solar radiation analysis.
(a)
(b)
Figure 3: Solar radiation examples from the Studios.
(a) existing roof design upgrade from Studio A and
(b) massing studies from Studio B.
While the Studio A students were asked to consider all
geometric and non-geometric properties of the
existing designs for first BPSs, the Studio B students
started their investigations by considering only
geometric properties, and other required simulation
inputs were provided by custom templates for a start.
It is observed that the Studio A students had
difficulties to comprehend BPS inputs for they had to
cope with all inputs at once. The amount of simulation
inputs was likely to be overwhelming. For early design
phase focuses on investigation of possible form
alternatives, rather than the optical and thermo-
physical properties of a design, including BPS in this
phase seems to be advantageous in attracting the
attention of students and enabling an easy get-in BPS.
In both studios, step-by-step approach, i.e. one
parameter at-a-time, was adopted. This was useful for
students to see the individual effect of each parameter
on a performance task. Later, combinations of selected
parameters were also tested. It was possible to
encourage the Studio B students to try more extreme
variations, especially regarding geometric parameters,
to enable them to see the effect of a specific parameter
clearly. Investigating the limit values of geometric
properties was also likely to bring new design ideas in
process. This was not the case in Studio A, due to the
already decided forms; only small revisions were
possible. Moreover, the students were likely to be
reluctant to apply even these small revisions, for they
were already highly coalesced with their designs.
Tools in architects’ ecosystem
ClimateStudio (CS) [V.1.6.8] (Solemma, 2019),
which is a BPS plugin with educational free license for
Rhinoceros3D CAD software [V.6] (McNeel, 2010),
was used in both studios. CS was preferred, for it is
integrated in the design tool, as well as built on the
validated simulation engines EnergyPlus and
Radiance. CS uses Rhinoceros3D’s interface and also
provides additional visually rich graphical user
interface (GUI) for simulation inputs and results. CS
is also available in Grasshopper (CS for Rhino), which
is a graphical algorithm editor in Rhinoceros, and
enables coupling number of tools; in the context of the
studios, for modeling, parametrization, simulation and
optimization. Model-based optimization tool
Opossum (Wortmann, 2017) [V.2.2.4], which is a –
free plugin for Grasshopper, was used for multi
objective optimization (MOO). This machine
learning-related optimization strategy was selected,
because it is appropriate for time-intensive
performance simulations. While parametric modeling
allows fast and flexible generation of design
alternatives, model-based optimization supports
finding well-performing variants, based on defined
performance goals. In Studio A, only CS for Rhino
was used; in Studio B, both CS for Rhino and CS for
GH were used. For the Studio B students developed
their models parametrically from scratch, it was
possible to extend the investigations from Rhino to
GH environment. Moreover, working in GH
environment enabled using the optimization solver in
workflows of Studio B (see Fig. 4).
Figure 4: Tools and Workflows of the Studios
Previous knowledge of the tools was not a prerequisite
for the studios. Most of the students were already
familiar with Rhino, but not with the GH, CS and
Opossum. Therefore, required training was provided
via case-studies and workshops.
Simplifications of BPS for teaching
In early design investigations, less complex and less
time-intensive design integrated methods and tools of
BPS have a significant potential to be adopted by
larger number of architects, because early design
seeks detection and quick evaluation of possible
design alternatives in relatively short time and with
relatively less input (Kalpkirmaz-Rizaoglu, et al.,
2020). In addition, considering beginners, adopting
simplified methods integrated to design workflows
seems to be promising way to enable easy and
attractive start. Besides developing knowledge of
building physics and BPS skills, one of the important

aims of teaching BPS might be to show students how
important their role is as future architects for the
indoor comfort and the sustainability of the built
environment, as well as how important a leverage can
be using BPS tools.
Simplifications applied in the studios can be grouped
as technical and theoretical. Theory related
simplification only applied in Studio B. Regarding the
technical simplifications, in both studios, custom
templates were prepared. After first run of
simulations, students were assisted to figure out the
most significant parameters in the context of their
designs, and detailed inputs were introduced in the
later steps. Also, in Studio B, a range of types for each
construction element were prepared, and students
were enabled to test different types by simply
selecting them via dropdown menus. For example,
instead of asking students to find the most appropriate
specific heat capacity of an exterior wall for a design
condition, a range of walls, i.e. light, medium and
heavy, are provided to simplify the early
investigations. Custom schedules for occupancy,
lighting and equipment and ventilation by residential
and non-residential options were also provided.
During parametric design and simulation, students
started with example workflows, which are tailored
for the studio by lecturers. In the further steps, students
were asked to adapt these workflows to their works.
Active solar energy utilization
First simplification regarding the theory applied for
active solar energy utilization. Differently from Studio
A, solar radiation analyses are integrated to massing
studies in studio B. Students searched the most
promising forms in terms of maximizing suitable
surfaces for active solar energy utilization.
Additionally, instead of directly getting in detailed
photovoltaic (PV) calculations, students were asked to
run annual solar radiation simulation and shadow
analysis for detecting the building surfaces with high
potential for PV installation with provided
specifications; e.g. panel dimensions, panel and
inverter efficiency. Later, students were asked to
calculate a solar production to energy load by using
an energy use intensity of a case study. This task
aimed to give an insight about the potential of
renewables by simply comparing potential of energy
generation and load.
Daylight availability
The massing studies were also coupled by daylight
availability analyses to draw the students’ attention to
the possible relations between the form and daylight
performance. Metrics, which are spatial daylight
autonomy (sDA) and useful daylight illuminance
(UDI), selected for daylight analysis. This preference
was quite useful to show the students high level of
daylight availability does not necessarily leads to high
level of visual comfort.
Thermal comfort
Another simplification was applied for thermal
comfort. The inspiration was gained from Building
2226 - low-tech of Baumslager & Eberle in Lustenau,
Austria - which was the case study of the first phase
of Studio B. It is noticed that the thermal comfort
concept of the building, which is defined in a range of
temperatures was easy to comprehend for students.
Therefore, similar approach was adopted for the
Studio. Percentage of heating, cooling and neutral
hours were used as an indicator of thermal comfort
performance. Hours equal to and between 20°C and 26
°C of operative temperature (Top) were considered as
“neutral hours”, which refers to the capacity of a
building to run without active heating and cooling.
Neutral hours is more of a simplified approach to give
students an insight about thermal comfort rather than
a definitive method of a building's cooling and heating
demand.
Integration of intelligent methods
In Studio B, parametrization and optimization
methods were adopted after a certain level of
knowledge was assured to make the investigation
more attractive and less time consuming, as well as the
investigation space larger. In the second phase of
Studio B students were introduced parametric
modeling and simulation. In this phase, performance
investigations were made manually. Combinations of
geometric and non-geometric parameters are tested
separately for each performance task to detect the key
parameters and their optimum variations. An example
parametric investigation for daylight availability can
be seen below (Fig. 5).
Figure 5: Parametric investigations in Studio B.
In the phase 3, students were requested to come up
with specific problems for their designs. They defined
the parameters to further investigate the designs for
the selected performance goals, and they were ready
to get in optimization (Fig. 6). After an introduction
lecture and a workshop about the optimization theory
and the tool, students were supported to create their
own parametric geometries and optimization
workflows connected to simulation workflows.

(a)
(b)
Figure 6: Example optimization sketches in Studio B:
Façade (a) and PV angle (b) optimizations.
Evaluation surveys – capturing students’ views
Anonymous studio evaluation surveys were
conducted to capture the students’ views about the
studio experiences by the end of each course. Each
survey had 8 respondents. A rating range is given
between 0 (low) and 4 (high) points, and the results
are presented as weighted arithmetic means.
It is asked that how useful did they find the studio
activities. While the mean in Studio A was 2,3; in
Studio B it was 3,25. Detailed ratings for studio
activities can be seen in Figure 7.
Figure 7: Rating for the studio activities.
Both studios selected thermal comfort as the most
difficult topic. The most prominent reason given by
students was that although they could directly use their
3D model for other simulations, for thermal comfort
they had to create a thermal zone model, and this was
difficult. When students are requested to compare
their expectation to the simulation results, vast
majority pointed out that the whole investigation
showed them something new, but not much different
than their expectations. Possible reasons for this result
might be that students had already a good level of
knowledge or / and results were likely predictable for
the design cases.
Several questions were asked about the BPS tool CS
(Fig. 8). Both studios students found the tool easy-to-
get-in (2,8) and the simulation time acceptable (2,6).
While the Studio A students were likely to be neutral
about the ease-of-use, Studio B students agreed that it
was easy to use. The difference can be interpreted
based on the amount of inputs to be defined at once in
each studio. It is asked if they have found the
integration of BPS and design tools attractive. The
Studio A students almost fully agreed (3,5) and the
Studio B students almost agreed (2,88). The lower
rating of Studio B might be resulted from extending
the CS workflows from Rhino to GH, which has less
user-friendly GUI, as well as lack of previous
knowledge of GH.
Figure 8: Feedback for the BPS tool.
The students were asked the level of improvement
regarding their skills and self-confidence for using a
BPS Tool by the completion of the studio. While the
Studio A students answered the question between
medium and high (2,5), Studio B students indicated a
level as very high (3,88).
Only half of the Studio A students and all studio B
students, except for one, indicated that they plan to use
BPS in their future studies. Reasons for not
considering the use of BPS were feeling mentally
limited during design, overwhelming amount of
simulation inputs, time intensive simulation process
and not having concrete plans for the master thesis yet.
Main findings
Adoption of BPS tools within design ecosystem at the
earliest stage possible, structuring a course by
gradually increasing level, starting by more geometry
related inputs, appropriate simplifications for a design
phase, hands-on-session, intensive supervision and
utilization of intelligent technics were found as the
highlights of the research in terms of integrating BPS
in a design process. When the learning objectives of
the courses were reviewed at the end, it is seen that
outcomes highly achieved the goals of the studios. On
the other hand, especially in Studio B, although
students were quite satisfied with the course in
general, when they were asked their opinions about
possible revisions, the common request was
decreasing the workload. This might be due to the high
number of performance tasks, and the tools they had
to learn and apply in one semester.

One limitation of applying BPS in early design was
the uncertainties regarding the simulation inputs. Yet,
for it was an early phase, custom templates and pre-
defined workflows were helpful as plausible solutions.
After students were introduced to BPS with simplified
workflows, they were more likely to move towards
detailed investigations in further steps. Theoretical
simplifications were helpful for students to gain
insight into performance at the early steps, as long as
they were accompanied by critical thinking. For
instance, regarding the thermal comfort simplification,
it was important to consider the risk of the
accumulation of cooling or heating hours just behind
the boundaries of limit temperatures that define the
neutral hours; and to check the operative temperature
curve graphics.
Including the parametric design and simulation
techniques increased the flexibility, and encouraged
the students to test more design variables, for it was
much easier to change the design geometry and other
simulation inputs. But for some, geometry related
predefined parametric workflows appeared as a
limitation for design aesthetics. By the integration of
optimization solver, the performance investigation
process was extremely sped up. On the other hand, it
is seen that before using optimization techniques,
students need to go through manual investigations to
recognize key parameters and optimization goals.
Acknowledging the necessity of simplifications for
early learners and/or early design phase, it is important
to guide students to understand pros and cons of these
simplifications through critical thinking, for them to
be able to decide when to use these simplifications and
when to turn to more advanced methods and tools.
Conclusion
This paper presented an example for integrating BPS
within the design course by sharing the pros and cons
of the developed simplification methods. Outcome of
the study shows that interdisciplinary and simplified
methods supported by intelligent design and
simulation techniques within architects’ design
ecosystem have a high potential to achieve integration
of BPS in architectural education.
A future international survey with broader population
is planned, as well as collaborations with other chairs
in order to structure new courses, in which design,
BPS and computational techniques are smoothly
integrated, providing a comfortable and adequate
learning process; e.g. two-semester-long courses.
Acknowledgement
This research was partially supported by German
Academic Exchange Service (DAAD) through
teaching assistant fund within the STIBET doctoral
program in 2021.
References
Attia, S., De Herde, A. 2011. Early design simulation
tools for net zero energy buildings: A comparison
of ten tools, 12th IBPSA Conf. Proc. Build. Simu.
Augenbroe, G. 2011. The role of the simulation in
performance-based building / Building simulation
for design and operation, ed. Hensen, J.,
Lamberts, R. London, Spon.
Beausoleil-Morrison, I., Hopfe C. 2016 Developing
and testing a new course for teaching the
fundamentals of building performance
simulation, eSIM 2016, Hamilton.
Delbin S. [et al.] 2006. Implementing building energy
simulation into the design process: A teaching
experience in Brazil, 23rd Int. PLEA Conf. Pro.
Geneva.
Fernandez‐Antolin, M.M. [et al.] 2020. The
Relationship between the use of building
performance simulation tools by recent graduate
architects and the deficiencies in architectural
education, Energies, Vol. 13. - p. 1134.
Hensen J., Lamberts, R. 2011. Building performance
simulation for design and operation. London,
Spon.
Hopfe, C. [et al.] 2005. Exploration of using building
performance simulation tools for conceptual
building design, IBPSA-NVL Conf. Pro. Delft.
Hopfe Christina J. [et al.] 2017.Understanding the
differences of integrating building performance
simulation in the architectural education system,
15th IBPSA Conf. Pro., p. 1249
Kalpkirmaz-Rizaoglu, I., Voss, K. 2020. Building
performance simulation to stimulate architectural
early design, 35th PLEA Conf. Pro.Coruña, p.
1525.
McNeel R., [et al.] 2010. Rhino3D Version 6.0.
Seattle : Robert McNeel & amp Associates, WA.
Soebarto, V. [et al.] 2015. Capturing the views of
architects about building performance simulation
to be used during design processes, 14th Int.
IBPSA Conf. Pro.
Soebarto, V. 2005. Teaching an energy simulation
program in an architecture school: Lessons
learned, 9th Int. IBPSA Conf. Pro. Montréal p.
1147.
Solemma. 2019. ClimateStudio.
[https://www.solemma.com/climatestudio]
Wortmann, T. 2017. Opossum: Introducing and
evaluating a model-based optimization tool for
Grasshopper.