2020 - 30 ITcon Khosakitchalert
2020 - 30 ITcon Khosakitchalert
2020 - 30 ITcon Khosakitchalert
net/publication/347510625
CITATIONS READS
4 36,993
3 authors:
Tomohiro Fukuda
Osaka University
218 PUBLICATIONS 1,542 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Chavanont Khosakitchalert on 20 December 2020.
SUMMARY: Quantity takeoff based on building information modeling (BIM) is more reliable, accurate, and rapid
than the traditional quantity takeoff approach. However, the quality of BIM models affects the quality of BIM-
based quantity takeoff. Our research focuses on drywalls, which consist of wall framings and wall panels. If BIM
models from the design phases do not contain wall framing models, contractors or sub-contractors cannot perform
quantity takeoff for purchasing materials. Developing wall framing models under a tight schedule in the
construction phase is time-consuming, cost-intensive, and error-prone. The increased geometries in a BIM model
also slow down the software performance. Therefore, in this research, an automatic method is proposed for
calculating quantities of wall framings from drywalls in a BIM model. Building elements that overlap with the
drywalls are subtracted from the drywall surfaces before calculation. The quantities of wall framings are then
embedded into the properties of drywall in the BIM model and hence they can be extracted directly from the BIM
model. A prototype system is developed and the proposed method is validated in an actual construction project.
The results of the case study showed that the prototype system took 282 s to deliver accurate quantities of wall
framings with deviations of 0.11 to 0.30% when compared to a baseline, and the file size of the BIM model after
applying the proposed method was increased very slightly from 47.0 MB to 47.1 MB. This research contributes to
developing an approach for quantity takeoff of wall framings that are not present in a BIM model. Accurate
quantities of wall framings can be obtained while the time and cost of developing wall framings for quantity takeoff
can be saved. The proposed method does not increase the geometries in the BIM model; therefore, the file size of
the model does not increase greatly, which stabilizes the software performance.
KEYWORDS: Building information modeling (BIM), Quantity takeoff, BIM-based quantity takeoff,
Quantification, Light-gauge steel framing, Wall framing
REFERENCE: Chavanont Khosakitchalert, Nobuyoshi Yabuki, Tomohiro Fukuda (2020). Development of BIM-
based quantity takeoff for light-gauge steel wall framing systems. Journal of Information Technology in
Construction (ITcon), Vol. 25, pg. 522-544, DOI: 10.36680/j.itcon.2020.030
COPYRIGHT: © 2020 The author(s). This is an open access article distributed under the terms of the Creative
Commons Attribution 4.0 International (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted
use, distribution, and reproduction in any medium, provided the original work is properly cited.
2. LITERATURE REVIEW
2.1 BIM-based quantity takeoff
BIM is the use of a digital representation of a built asset as a shared knowledge resource to form a reliable basis
for decisions during the design, construction, and operation process (ISO, 2018; National Institute of Building
Sciences, 2007). Because all building elements in BIM are object-based models, BIM can be used to extract
quantities and other information directly from the building model elements (Sacks et al., 2018); thus, BIM-based
quantity takeoff is more reliable, accurate, and rapid than the traditional quantity takeoff (Bečvarovská and
Matějka, 2014; Nadeem et al., 2015; Sacks et al., 2018; Sattineni and Bradford, 2011).
In traditional quantity takeoff, the lack of details in construction drawings lowers the accuracy of quantity takeoff
(Nani and Adjei-Kumi, 2007; Smith, 2014). Similarly, the lack of details in BIM models affects the quality of
BIM-based quantity takeoff (Khosakitchalert et al., 2019c; S. Kim et al., 2019; Smith, 2014). Moreover, checking
the quality of complex BIM models is challenging for construction practitioners (Smith, 2014, 2016). Therefore,
the level of development (LOD) and modeling guidelines should be determined before developing BIM models
for quantity takeoff (Firat et al., 2010; Smith, 2016; Wood et al., 2014). The LOD specifies the degree of geometry
and attached information of each building element in a BIM model (BIMForum, 2019), which affects the accuracy
of quantity takeoff directly (Peansupap and Thuanthongdee, 2016).
The effects of the modeling method and the LOD on BIM-based quantity takeoff have been studied. Monteiro and
Martins (2012) surveyed the strengths and weaknesses of different modeling methods for walls and floors. Zima
(2017) found that separating each wall layer according to the actual construction provided the best quantity takeoff
accuracy. Yun and Kim (2013) studied various modeling methods for reinforced concrete structure and proposed
a suitable modeling method for BIM-based quantity takeoff. S. Kim et al. (2019) examined the deviation of
material quantities of building interior components for different modeling methods. Peansupap and Thuanthongdee
(2016) investigated the LOD for a precast concrete floor to use for cost estimations during bidding and material
purchasing. They found that the BIM model of the precast concrete floor for material purchasing needed additional
details, which are wire mesh and rebar models.
In general, modeling problems in BIM-based quantity takeoff have been tackled by methods to automate the
modeling process or by algorithms to calculate the material quantity from the existing BIM models. S. A. Kim et
al. (2009) developed an automatic system that generates interior wall finishes for accurate BIM-based quantity
takeoff. Kannan and Santhi (2013) created parametric BIM components for concrete formwork and used them in
a high-rise BIM model. Liu et al. (2015, 2018) reported an automatic system to generate the boarding layout of
drywalls in a BIM model with the lowest material waste. Lim et al. (2016) proposed an automatic algorithm to
calculate quantities of rebar from 3D structural models. Cho and Chun (2015) developed a method that integrates
quantity takeoff logics and data mining techniques to estimate quantities of reinforced concrete structures based
on BIM models. Rajabi et al. (2015) created a system to estimate quantities of lighting and heating appliances from
BIM models that do not contain those appliances. Khosakitchalert et al. (2018, 2019c) reported a method that uses
the capability of BIM-based clash detection to subtract the overlapping quantities and add missing quantities of
walls and floors. Khosakitchalert et al. (2019a) proposed an algorithm to calculate quantities of concrete formwork
from BIM models with no concrete formwork models.
2.2 Light-gauge steel framing system
At present, a light-gauge steel framing system is used as an alternative to a wood framing system in drywall
construction (Packer, 2016). The system is similar to the wood framing system in principle although the wood
Fig. 2: Two possible details at intersections and corners of a light-gauge steel framing system.
Fig. 5: Flowchart for process 1: generate drywall surfaces from a BIM model.
Fig. 6: (a) Building model elements are imported from a BIM model to the system. (b) Wall elements that are not
drywalls are discarded. (c) Wall surfaces created by the system. (d) Overlapping building elements are
subtracted from the wall surfaces.
4.2 Process 2: Generate lines that represent U-tracks
Fig. 7 shows the flowchart for process 2: generate lines that represent U-tracks. The process is broken down into
processes for the top, bottom, nogging, header, and sill trimmer tracks. For the top tracks and the bottom tracks,
the perimeter lines of the subtracted wall surfaces from process 1 are extracted. The algorithm selects the lines at
the top edge of the wall surfaces to obtain the lines that represent the top tracks (see Fig. 8a). Another algorithm
Fig. 8: Lines that represent (a) top tracks, (b) bottom tracks, (c) nogging tracks, (d) header tracks at doors, and
(e) header tracks and sill trimmer tracks at windows.
For the nogging tracks, first, the number of nogging tracks on a wall is calculated by dividing the wall heights
from process 1 by the input spacing value of U-tracks and rounding up the results to a whole number. The number
of nogging tracks and the spacing value of U-tracks are used to create a sequence of points on the Z-axis of the
subtracted wall surfaces. Second, the wall location lines are duplicated to the points except for the first point,
which is the location of the bottom track. The lines are intersected with the subtracted wall surfaces. The parts of
the lines that are outside the subtracted wall surfaces, such as the regions of doors and windows, are eliminated to
obtain the lines that represent nogging tracks (see Fig. 8c).
Fig. 10: Lines that represent (a) common studs and (b) jack studs.
For the jack studs, first, the locations of the jack studs are calculated. The perimeter lines of the door surfaces and
the window surfaces from process 2 are extracted. The algorithm selects the lines at the top edge of the door
surfaces. Another algorithm selects the lines at the top edge and bottom edge of the window surfaces. The lengths
are calculated from the selected lines and divided by the input spacing value of the jack studs, and the results are
rounded up to a whole number to give the number of jack studs on each door and window. The number and spacing
value of the jack studs are used to create a sequence of points on the lines. Second, the vertical lines that represent
the jack studs are created. For the jack studs between the top tracks and the header tracks, the vertical lines are
created from the points at the top edge of the doors and the windows to the top of the walls. For the jack studs
between the sill trimmer tracks and the bottom tracks, the vertical lines are created from the points on the bottom
edge of the windows to the base of the walls. Finally, the vertical lines are intersected with the subtracted wall
surfaces from process 1 and the parts of lines that are outside the subtracted wall surfaces are eliminated to obtain
the lines that represent jack studs (see Fig. 10b).
The final products of process 3 are the lines that represent common studs and jack studs, which are used in process
6.
4.4 Process 4: Generate lines that represent extra C-studs at intersections and corners
Fig. 11 shows the flowchart for process 4: generate lines that represent extra C-studs at intersections and corners.
Initially, construction detail A or B is chosen by a user. If construction detail A is chosen, the system generates
one line that represents extra C-studs at corners and T-intersections and two lines that represent extra C-studs at
cross intersections. To do this, an algorithm finds intersection points on the wall location lines from process 1. The
duplicate points that are in the same locations are removed. Each point is then checked for an intersection with the
lines that represent common studs from process 3. The lines that intersect the points are collected into a list as lines
that represent extra studs at corners, T-intersections, and cross intersections (see Fig. 12a). Next, the algorithm
finds the second extra studs at cross intersections. The cross intersection points are identified by finding the
intersection points on the wall location lines that have four duplicates at the same position. The cross intersection
points are then checked for intersections with the lines that represent common studs. The lines that intersect the
points are collected into another list as the lines that represent the second extra stud at cross intersections (see Fig.
12b).
If construction detail B is chosen, the system generates one line that represents extra C-studs at cross intersections.
To do this, the algorithm finds intersection points on the wall location lines from process 1. The points of cross
intersections are identified by finding the intersection points on the wall location lines that have four duplicates at
the same position. The cross intersection points are then checked for intersections with the lines that represent
Fig. 11: Flowchart for process 4: generate lines that represent extra C-studs at intersections and corners.
Fig. 12: Lines that represent (a) extra studs at corners, T-intersections, and cross intersections for construction
detail A, (b) second extra studs at cross intersections for construction detail A, and (c) extra studs at cross
intersections for construction detail B.
The final products of process 4 are the lines that represent extra C-studs at intersections and corners, which are
used in process 6.
4.5 Process 5: Generate lines that represent extra C-studs at both sides of openings
Fig. 13 shows the flowchart for process 5: generate lines that represent extra C-studs at both sides of openings.
Fig. 13: Flowchart for process 5: generate lines that represent extra C-studs at both sides of openings.
Fig. 14: Lines that represent extra studs at both sides of openings.
4.6 Process 6: Calculate quantities of wall framings and embed them into the drywall
properties in the BIM model
Fig. 15 shows the flowchart for process 6: calculate quantities of wall framings and embed them into the properties
of drywalls in the BIM model.
Fig. 15: Flowchart for process 6: calculate quantities of wall framings and embed them into the properties of
drywalls in the BIM model.
Fig. 16: Properties of a wall in which the quantities of C-studs and U-tracks are embedded.
5. VALIDATION
The BWFQT method was validated in an actual construction project. Quantities of wall framings from the BWFQT
prototype system were compared with the quantities of wall framings from the complete detailed BIM model and
two traditional quantity takeoff methods. Furthermore, the execution time of the BWFQT prototype system was
compared with the modeling time of the BIM model and the manual calculation times of the traditional methods.
Autodesk Revit 2018.2, Dynamo 1.3.3.4111, and a computer with an Intel Core i7-8770 3.2 GHz CPU, 32 GB
RAM, and NVIDIA GeForce GTX 1060 6 GB graphics card were used to perform the prototype system.
5.1 Overview of the case study
The case study was an interior project on the fifth and sixth floors of Chulapat 14 building located at Chulalongkorn
University, Bangkok, Thailand (see Fig. 17a). The fifth and sixth floors were chosen because they contain drywalls
meeting at corners, T-intersections, and cross intersections. Furthermore, these floors contain drywalls intersecting
with doors, windows, structural columns, beams, and floors. The gross internal area of these floors is about 2,560
m2. There are 28 rooms on the fifth floor and 25 rooms on the sixth floor separated by drywalls.
The typical height of the drywall is 3.65 m and the height of the drywall under the beams is 3.45 m. The spacing
of C-studs, including common studs and jack studs, is a maximum of 600 mm center-to-center. The spacing of U-
tracks is a maximum of 600 mm center-to-center. The construction detail of the intersections and corners is
construction detail A (see Fig. 2). One extra C-stud is needed at a corner and a T-intersection and two extra C-
studs are needed at a cross intersection. The C-studs at both sides of openings are two C-studs combined into boxed
studs. Fig. 17b shows the wall framings in the construction site.
The incomplete and complete detailed BIM models were both used in the validation. Both BIM models were
created by a graduate student with a background in architecture who has been using Autodesk Revit for seven
Fig. 17: (a) Chulapat 14 building. (b) Interiors of the fifth and the sixth floors during construction.
Fig. 19: Blue lines that represent wall framings in the incomplete detailed BIM model (drywalls were hidden
from the viewport).
Fig. 21: Complete detailed BIM model (drywalls were hidden from the viewport).
The BWFQT method provided similar quantities of wall framings compared with the complete detailed BIM
model, which was the baseline (see Table 1). The quantities of C-studs from the BWFQT method had deviations
of 0.12% on the fifth floor, 0.11% on the sixth floor, and 0.11% in total. The quantities of U-tracks from the
BWFQT method had deviations of 0.16% on the fifth floor, 0.44% on the sixth floor, and 0.30% in total. The
deviation of C-studs and U-tracks may arise from the thickness of the wall framings in the complete detailed BIM
model making the lengths of wall framings shorter when they connect to each other.
The quantities from traditional methods 1 and 2 were less accurate than those from the BWFQT method and had
deviations ranging from −3.84% to 12.05%. A negative deviation indicates that the quantity is less than the
baseline.
The quantities of C-studs from traditional method 1 had deviations of 4.83% on the fifth floor, 1.83% on the sixth
floor, and 3.35% in total. The quantities of U-tracks from traditional method 1 had deviations of 11.56% on the
fifth floor, 12.05% on the sixth floor, and 11.80% in total. The excess C-studs and U-tracks were calculated because
traditional method 1 does not consider wall openings, such as doors and windows, and the parts of walls that
overlap with structural elements.
The quantities of C-studs from traditional method 2 had deviations of 4.08% on the fifth floor, 3.10% on the sixth
floor, and 3.60% in total. The quantities of U-tracks from traditional method 1 had deviations of −3.84% on the
fifth floor, −0.79% on the sixth floor, and −2.34% in total. The excess C-studs and the insufficient U-tracks were
calculated because the method calculates the lengths of wall framings based on the average wall framings per
square meter, which may not perfectly fit walls that have doors and windows.
Table 2: Comparison of the execution time of the BWFQT prototype system, the modeling time of the complete
detailed BIM model, and the manual calculation times of traditional methods 1 and 2.
The BWFQT Prototype Complete Detailed BIM Traditional Manual 1 Traditional Manual 2
System Model
Time Time Time Time Time Time Time Time
(seconds) (seconds) (seconds) (seconds)
4 min 42 s 282 25 hr 40 min 92,400 27 min 1,620 53 min 3,180
The BWFQT prototype system took the least time (4 min 42 s [282 s]) to find the quantities of wall framings (see
Table 2). Modeling the complete detailed BIM model took the most time (25 h 40 min [92,400 s]). The calculation
time of traditional method 1 was 27 min (1,620 s). In this method, most of the time was spent measuring the wall
lengths and counting the numbers of corners, intersections, doors, and windows from the 2D construction
drawings. The calculation time of traditional method 2 was 53 min (3,180 s). In this method, most of the time was
spent measuring the wall areas from the 2D construction and subtracting the areas of doors, windows, columns,
and beams that overlap with the walls.
The BWFQT prototype system was 328 times faster than the modeling time of the complete detailed BIM model,
6 times faster than traditional method 1, and 11 times faster than traditional method 2.
In the case study, the BWFQT prototype system calculated the quantities of wall framings and embedded them
into the properties of drywalls in the BIM model. The quantities of wall framings obtained from the BWFQT
method were as accurate as those extracted from the complete detailed BIM model, and the execution time of the
BWFQT method was much faster than the manual modeling time of the complete detailed BIM model. The
BWFQT method can reduce the working time and cost of creating wall framings in a BIM model for quantity
takeoff. The quantities of wall framings are embedded into wall elements; therefore, this information can be
extracted directly from the BIM model.
However, if wall framing models are needed for other tasks, such as visualizing a construction process or
generating shop drawing documents, modeling for wall framings is still required. The BWFQT method could be
developed by the converting lines that represent the wall framings generated from the system to the wall framing
components in a BIM model. Nevertheless, the increased wall framings in a BIM model result in a larger file size
that would consume more memory resources and slow down the software performance. According to the case
study, the file size of the complete detailed BIM model that contains wall framings is 66% larger than the file size
of the incomplete detailed BIM model that does not contains wall framings, while the file size of the incomplete
detailed BIM model after applying the BWFQT prototype system is still similar to the original model (see Fig.
23). The impact of difference in file sizes between the BIM model that contains wall framings and does not contain
wall framings would be bigger in a large-scale building that has a huge number of drywalls. Therefore, the BWFQT
method is a better approach for BIM-based quantity takeoff, and if the wall framing models are needed for
producing shop drawing documents, they should only be created in a selection of representative areas.
Fig. 23: The file sizes of (a) the complete detailed BIM model, (b) the incomplete detailed BIM model, and (c)
the incomplete detailed BIM model after applying the BWFQT prototype system.
The BWFQT method also delivers more accurate quantities of wall framings than the traditional methods because
it calculates the lengths of wall framings based on the lines that represent the wall framings in the correct positions
in the BIM model. Traditional method 1 calculates rough quantities of wall framings based on the lengths of walls
and spacing values of wall framing without considering wall openings. Traditional method 2 calculates
approximate quantities of wall framings based on the wall surface area and the wall framings per area. Both
methods also need to add the extra studs at corners, T-intersections, and cross intersections, and the studs at both
sides of openings to the length of C-studs. In this case study, the total additional C-stud was 1,788.50 m. If this
quantity were not added to the length of C-studs, the deviation of the quantity of C-studs would be −29.82% instead
of 3.35% in traditional method 1 and −29.56% instead of 3.60% in traditional method 2. Therefore, the deviations
could vary if quantity surveyors with different levels of experience perform the quantity takeoff.
The BWFQT method is also faster than traditional methods 1 and 2. Most of the time is spent on measuring wall
length and net wall area in traditional methods 1 and 2, respectively. The traditional methods could be sped up by
using semi-automatic BIM-based methods to extract the length and the area of walls from a BIM model. However,
the walls in the BIM model must be checked and the overlapping regions corrected manually, whereas the BWFQT
method eliminates the overlapping regions automatically.
In future research, more case studies should be used to validate the BWFQT method. Furthermore, because the
accuracy and speed of the traditional methods could vary depending on the experience and expertise of the quantity
surveyors, different quantity surveyors should be invited to validate the BWFQT method.
7. CONCLUSION
General contractors or sub-contractors require the quantities of wall framings during the construction phase.
Traditionally, the quantities of wall framings are estimated using manual calculations. However, BIM can improve
the accuracy and speed of quantity takeoff by allowing quantities of building elements to be extracted directly
from a BIM model. However, wall framings are detailed elements that are not usually present in BIM models.
Creating wall framing models only for quantity takeoff is a time-consuming, costly task. Our approach provides
an alternative method of calculating quantities of wall framings from a BIM model that does not contain wall
framing models.
The BWFQT method calculates quantities of wall framings from drywalls in a BIM model. The algorithms
automatically calculate the locations and the lengths of wall framings from the drywall surfaces and the input
spacing values of wall framings. The results are embedded into the properties of drywalls in the BIM model, and
the quantities of wall framings can then be obtained from the model directly.
The BWFQT prototype system was developed using the Dynamo extension in Autodesk Revit. Validation was
conducted using an actual construction project. The BWFQT method provided accurate quantities of wall framings
compared with the complete detailed BIM model in which the wall framings were modeled manually. The BWFQT
method was more accurate than the traditional quantity takeoff methods, and the execution time for the prototype
system was faster than the modeling time of the complete detailed BIM model and the calculation times of the
traditional methods.
The BWFQT method contributes to developing an approach for quantity takeoff of wall framings that are not
present in a BIM model. Using this method, construction practitioners can obtain accurate quantities of wall
framings from a BIM model that does not contain wall framing models. Deviations due to manual calculations can
be reduced and the time and cost of creating wall framings in a BIM model for quantity takeoff can be saved. The
accurate quantities of wall framings are beneficial for purchasing materials and reducing the costs associated with
excess materials. The BWFQT method does not increase the geometries in the BIM model; therefore, the file size
of the model does not increase, which stabilizes performance, and the model is easy to be edited. Furthermore, the
study highlights the issues when performing quantity takeoff from an incomplete detailed BIM model. The case
study in this research also provides an understanding of deviations of material quantities of wall framings from the
traditional quantity takeoff methods, which is beneficial for academic use. Finally, the logic of BWFQT method
can be applied to other research that deals with other types of framing systems and can be used to develop a wall
framing calculation system in any BIM software product.
Future research could focus on two areas. First, the accuracy and speed of the BWFQT method could be validated
by further case studies and various quantity surveyors. Second, the BWFQT method could be extended and
improved. For example, the BWFQT method could be used as a base to develop a system that can generate wall
framing components in a BIM model automatically, which is beneficial for producing shop drawings and as-built
drawings. Moreover, the limitations of the BWFQT method should be addressed. The method should be extended
to calculate quantities of wall framings in slanted walls and curved walls and to calculate wall framings for
different drywall types simultaneously. In addition, other wall openings shapes than rectangular should be
supported in future systems.