Contents lists available at ScienceDirect

Automation in Construction

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

Process, productivity, and economic analyses of BIM–based multi-tradeprefabrication—A case study

Sejun Janga,c, Ghang Leeb,⁎

a Department of Architecture & Architectural Engineering, Yonsei University, Republic of KoreabDepartment of Architecture & Architectural Engineering, Yonsei University, Seoul 120-749, Republic of Koreac Smart Construction Team of Hyundai Engineering and Construction, Seoul, Republic of Korea

A R T I C L E I N F O

Keywords:PrefabricationMulti-trade prefabricationCorridor MEP rackProductivityBuilding information modeling (BIM)

A B S T R A C T

Previous studies yielded contradicting results regarding the benefits of multi-trade prefabrication (MTP) basedon building information modeling (BIM). This study investigates the causes of the contradicting results byanalyzing the process, productivity, and economic benefits of BIM-based MTP through a case study. The processanalysis results indicate that coordinating mechanical, electrical, and plumbing (MEP) systems took longer inMTP than in the conventional method because of the newly added offsite coordination activities. Nevertheless,the overall project duration was reduced because of the parallel execution of MTP and concrete work. Theproductivity analysis shows that the newly added MTP activities also increased person-hours. However, as theassembly process was repeated, the required person-hours decreased by 40% from that of the initial stage be-cause of the learning effect. The case study revealed that the management of coordination activities and theselection of projects were critical for the successful implementation of BIM-based MTP.

1. Introduction

Prefabrication, which was first introduced in the manufacturingindustry (e.g., the automobile industry), is being increasingly applied inconstruction projects [4]. Compared with conventional onsite fabrica-tion, prefabrication is reported to shorten the overall project schedule,improve product quality, increase onsite safety, and reduce the need forskilled onsite workers, waste, and carbon emissions [23,33,46,49].Conversely, prefabrication also has the potential to increase the overallproject duration and cost due to high production, transportation, andinstallation time and costs [2] and difficulties in implementation, par-ticularly when prefabrication details are not reflected in the design orworkers are unfamiliar with the prefabrication processes [27].

Advances in design and construction technologies, such as buildinginformation modeling (BIM), design for manufacture and assembly(DfMA), and laser scanning, can help overcome the aforementioneddrawbacks [4,18]. For example, BIM can be used to efficiently detaileach building trade [25] and detect and correct the interferences be-tween trades to obtain precise and fully coordinated geometric models.Such BIM environments improve the applicability of prefabrication toconstruction projects [48]. Furthermore, these technological develop-ments support the expansion of prefabrication from single-trade pre-fabrication (e.g., curtain walls, precast concrete, and prefabricated

pipes) to multi-trade prefabrication (MTP), wherein various trades mustbe coordinated in the same space [15].

Yet, some studies have expressed concerns regarding the practic-ability of MTP in the construction industry [8,24,27] and claimed thatprefabrication is not economically effective despite the shortened pro-ject durations [2,13,45]. The present study investigates the causes thatresulted in the contradicting views of previous studies on the benefits ofMTP in BIM projects. Several previous studies discussed the potentialeconomic benefits of prefabrication but only on a high level (i.e., finalreduction values or perceived savings) [7,17,27,28]. This study con-ducts detailed process, productivity, and economic analyses of a caseproject of multi-trade corridor racks in an exhibition complex.

The rest of this paper is organized as follows: Section 2 reviews thecurrent literature on prefabrication in the construction industry andhighlights relevant issues; Section 3 describes the research metho-dology; Section 4 introduces the case project; Sections 5 to 7 report theresults of the process, productivity, and economic analyses, respec-tively; and Section 8 summarizes the study's findings and limitations,presents the industry implications of the results, and discusses the scopefor future work.

https://doi.org/10.1016/j.autcon.2017.12.035Received 22 March 2017; Received in revised form 1 December 2017; Accepted 27 December 2017

⁎ Corresponding author.E-mail address: glee@yonsei.ac.kr (G. Lee).

Automation in Construction 89 (2018) 86–98

0926-5805/ © 2018 Elsevier B.V. All rights reserved.

T

2. Literature review

Prefabrication was initially applied in the construction industry toreduce cycle time and costs and improve construction quality [46] andrecently to reduce waste and carbon emissions [23,49]. Prefabricationtechnologies are applied in various contexts to meet different objec-tives. For example, as housing demand exceeds supply in the UK, pre-fabrication has been increasingly used since the 1990s to improveconstruction speed [33]. The prefabrication technology was also ap-plied in the Graves Avenue Bridge renewal project in Florida, in the USto minimize bridge downtime [9]. Meanwhile, the Morgantown CoalUnloader Pier construction project in Maryland (US) adopted pre-fabrication to reduce worker exposure to cold temperatures outside andimprove productivity, which decreased the project duration [37]. Si-milarly, a plant construction project deployed prefabrication during theengineering and fabrication phases to improve productivity throughdesign standardization, modularization, and recycling [31].

Prefabrication technologies for various trades have gradually ad-vanced. With the development of prefabricated housing frames in the1620s [3], the prefabrication for housing buildings is among the earliestand most explored technologies [6]. The main prefabricated elementsin these early structures, which were not as extensively prefabricated asthe current structures, were timber frames and complex joints [38]. Thecosts of timber and plywood increased, and with the increasing ac-ceptance of concrete as a building material, precast concrete work wereattempted in the 1850s [12]. With various reinforcement technologies,precast concrete is currently used to manufacture linear and curvedmembers [41].

Conventional prefabrication technologies focus on single-tradeprefabrication, such as precast concrete products, facade panels, andwindow frames. However, the demand for MTP, wherein various tradesmust be coordinated in the same space, is rapidly growing [15,17,27].Bathroom pods (Fig. 1), which are commonly known as unit bathrooms,or UBRs in Asia, are among the oldest and most well-known examples ofMTP that include MEP, tiling, interior finishing, and door installation.These bathroom pods were introduced in the 1960s when Korea and

Japan were striving for economic revitalization. They have been widelyimplemented in apartment units and hotels since the 1980s [11].

Modern buildings tend to have heavy MEP systems [27], for whichcontractors tend to use prefabrication [40,47]. MTP has recently beenapplied to MEP corridor racks with complex MEP components(Fig. 2)—the coordination, manufacturing, and assembly of which aremuch more challenging than those of relatively simpler MTP compo-nents (e.g., bathroom pods) because of the higher MEP density.

BIM-based DfMA, coordination, and laser scanning technologieshelp overcome some of the complex coordination, manufacturing, andassembly issues. For example, the BIM-enhanced prefabrication ofprecast concrete and structural elements through the automation ofdetailing and the utilization of computer numerical control machines[5,14,39] has facilitated the prefabrication of complex structures, in-cluding irregular facade panels and MEP systems [22,26]. A BIM-baseddesign process and a model view definition for prefabricated buildingswere also specified [36,42]. In addition, the use of the Internet ofThings in prefabrication is being explored to improve data exchangeand communication during prefabrication and assembly [43,52]. As aresult, the adoption of prefabrication has been recently accelerated[1,4,18,50].

Despite many successful cases of BIM-based MTP [1,4,18,50], sev-eral studies have expressed concerns about the effectiveness of pre-fabrication [8,21,24]. A survey conducted by Goodier and Gibb [8]identified the cost increase as the main barrier to the adoption of pre-fabrication in the UK. Lu [24] reports that the top three challenges inprefabrication in the US are the inability to make changes onsite,transportation constraints, and limited design options. Lawson et al.[21] argued that MTP increases onsite risks if the geometric differencesbetween the prefabricated components and the structural frames of thebuilding are not controlled.

Several industry reports have analyzed the economic impact of MTPand have also revealed contradicting results. Among the industry re-ports, the most documented are the Miami Valley Hospital project bySkanska US, wherein MTP was applied to corridor racks and bathroompods [7,20], and the Saint Joseph Heritage project by Mortenson

Fig. 1. Bathroom pod (UBR) (photographed by SejunJang).

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

87

Construction. Skanska US reported that employing MTP saved 4–8% ofthe total construction costs and 20% of the labor costs and increasedproductivity by 30%, thereby shortening the construction schedule byeight weeks. In contrast, Mortenson Construction reported a 15% re-duction in the construction schedule of the Saint Joseph Heritage pro-ject and a 6% increase in the construction cost [2,29,30]. Nevertheless,neither study provided an in-depth analysis of the causes for the con-tradicting results. The present study investigated the BIM-based MTP indetail to identify the causes for the contradicting results from the pro-cess, productivity, and cost perspectives.

3. Research method

This section describes the research method. This study aims toanalyze the process, productivity, and economic impacts of MTP. Tothis end, a set of criteria is defined to identify a suitable project. Thesecriteria are described in Section 4. This study includes three steps(Fig. 3), and the research method is as follows.

The first step is to analyze the process changes compared withconventional offsite construction in the design, coordination, manu-facturing, and assembly phases caused by the MTP application.Identifying and understanding the changes that must be made to theprocess to which the technology is applied, determining the time re-quired to conduct the process, and reviewing whether any new tech-nical requirements apply to the process are all critical functions thatmust be considered before any new technology is implemented.

The second step is to analyze the changes in productivity.Contractors primarily adopt MTP to increase productivity[10,27,32,34]. Productivity is defined as the amount of installation andlabor input in person-days. This study focuses on the labor input be-cause the core comparative analysis herein is conducted for the samebuilding. The unit amount of installation is similar for both the MTPand conventional methods. Moreover, this study analyzes the pro-ductivity gain caused by the learning effect of repetitive work.

The third step is to analyze the economic benefits of MTP comparedwith the conventional method. The baseline for the comparison is set

Fig. 2. MTP corridor rack (photographed by SejunJang).

Fig. 3. Research method.

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

88

based on the productivity data measured in the second step and thosereported in the literature.

4. Case selection

The selection criteria for the case project were defined as uni-versally as possible to maximize the generality of the case study: thebuilding should not be too tall or too short, should not be too large ortoo small, and should not be too complex or too easy to build. High-rise,small, free-form, and commodity buildings were excluded from thecandidate list. The specific selection criteria were as follows:

1. The project deploys MTP.2. The project involves various components and multiple MEP trades:

– The project uses general corridor MEP racks that can be used invarious types of buildings (e.g., offices, housing, hospitals, anddata centers).

– MEP trades involve various systems, such as fire protection pipes;heating, ventilating, and air conditioning (HVAC) ducts and pipes;sanitary pipes; and electrical and intelligent building system (IBS)trays.

3. The project and site are of typical task difficulty:– The ceiling height must not be too high or too low (ideally 4–5 m),and the working space must not be confined because the KoreaEstimation Standard [19] assigns a relatively higher baseline forthe MEP labor input when the ceiling height exceeds 5 m.

– Lifting tasks must be moderate (5–10 stories), and the construc-tion site must not be underground.

– The prefabrication space must be of a regular shape and must notbe sloped.

4. The project working environment is typical:– The site should not be in a military or restricted area because theKorea Estimation Standard [19] assigns an increased estimationbaseline for the labor input for work in military or restricted areas.

– The roads around the building should be easily accessible for in-stallation and material procurement. The site should not be in amountainous area.

An automotive exhibition complex located in the Gyeong-gi pro-vince of the Republic of Korea met the selection criteria and, hence, waschosen as the case project (Fig. 4).

Apartment projects were also considered as candidates for the casestudy because of their repetitive patterns. However, they were un-suitable for this case study because the MEP designs were too simple forMTP usage. Another candidate was a hospital project. Hospital projectswere also determined to be unsuitable because the MEP design inhospital projects frequently changes even during the construction

phase. Offsite construction methods, including MTP, are weak at copingwith such last-minute design changes.

The selected building consisted of four underground floors and nineaboveground floors organized into various spaces (e.g., a complex of-fice, an exhibition space, and an automobile repair space), with a totalfloor area of 62,755 m2 and a total construction period of 37 months.The ground floors were structured with steel, while the undergroundfloors were structured with steel-reinforced concrete. MTP moduleswere used to fabricate the corridor of the complex office area located onfloors five to eight. A total of 15 modules were used, as follows: fifthfloor—four modules, sixth floor—four modules, seventh floor—fourmodules, and eighth floor—three modules. Each floor had a ceilingheight of 4650 mm, which satisfied the general construction conditionand precluded the usage of the working-height labor adjustment factor[19]. The complex office corridor required HVAC ducts, sprinkler pipes,sanitary pipes, HVAC pipes, and electrical and IBS trays, thereby alsorequiring MTP involving various trades (Fig. 5).

5. Process analysis

We analyzed the process changes and identified the technical re-quirements necessary for implementing the MTP in this section. Theanalysis of the process changes for design and coordination was con-cerned with the additional structural supporting system for pre-fabrication and the increased number of coordination steps. Meanwhile,the analysis of the process changes for manufacturing and assembly wasconcerned with the comparative process analysis of conventional andMTP construction. The technical requirements for construction usingMTP, which many workers tend to be unaware of, were identified.

5.1. Analysis of the process changes for design and coordination

A new MEP support system was necessary for implementing MTP. Inthe case of the conventional prefabrication, distinct hanging systemswere designed for each individual trade according to the project spe-cification. In MTP, complex elements of various trades should be sup-ported within a single integrated structural framing. Therefore, si-multaneously designing a new supporting structure that can enduredead and operational loads of multiple trade elements was essential.Accordingly, a structural analysis for the new structural framing wasconducted in the design phase, and the members of the structuralframing were determined for each MTP module.

The structural framing of the MTP module can be described as fol-lows: the MTP module was 640 mm× 1590 mm× 5000 mm in sizewith an inter-module separation of 1000 mm. The module height(640 mm) was set considering the plenum space, while the width(1590 mm) was determined considering the working space and the

Fig. 4. Case project and MTP zone (complex officearea).

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

89

corridor width (1800 mm). Meanwhile, the length of the MTP module(5000 mm) was determined considering the length of the standard pipeunits in Korea (6000 mm) and the workability for connecting the MTPmodule. The inter-module spacing (1000 mm) was set considering theworkability for joint connection during the onsite assembly. The designprocess of the structural framing accounts for the requirements of thevarious contractors responsible for the MEP prefabrication and onsiteassembly. Fig. 6 shows the design of the MTP module used in the caseproject.

MTP entailed more complicated steps than conventional design andcoordination. The conventional process involved BIM modeling, co-ordination, and shop drawing before the actual construction. The de-sign and coordination process for MTP involved two additional steps,namely designing the structural framing for supporting multiple tradesand the re-coordination process following the structural framing. Thebasic (first) coordination step was similar to that in the conventionalmethod. In the first coordination step, the elements of each trade arerelocated, and the interference between trades is eliminated. In the re-coordination step, fine modifications are made in addition to elim-inating the interference with the structural framing to connect eachelement to the structural framing. Conventional construction involvedonly shop drawing, where prefabrication involved detailed spooldrawings (see Fig. 8 for an example). Fig. 7 highlights the differences inthe design and coordination processes between the conventionalmethod and MTP. MTP required more preparation time before the

actual construction phase than the conventional method because of theadditional steps. The additional time required in our case was fourweeks.

5.2. Analysis of the process changes for manufacturing and assembly

A manufacturing factory can be placed within a construction site ifthe site is spacious enough; however, in our case, most of the space atthe construction site was utilized for construction activities. Therefore,the factory was established offsite. The field workers of the same con-tractor working at the site were allocated to the factory. In other words,no additional (special) contractor was hired to manufacture the MTPmodules, ensuring that the productivity would not be affected by theinvolvement of a new subcontractor. A change in subcontractor mayaffect the productivity analysis because of the rare understanding of thedesign and MEP detail. The MTP module was assembled in the fol-lowing order: structural framing, duct, pipe, and tray. This sequencewas set after reviewing the inputs from the representatives of eachtrade.

Compared with the conventional method, utilizing the MTP mod-ules had the four following advantages. First, factory manufacturingminimized the time spent in non-value-adding activities that did notcontribute to the final output. For example, the time lost to workers'movement was reduced. Workers in Korea typically conduct a safetymeeting near the field office and then move to the work place, which is

Fig. 5. MTP zone (corridor MEP racks for plenumspace).

Fig. 6. MTP module design (corridor MEP racks).

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

90

distant from the meeting place. This movement time can be minimizedif factory prefabrication is adopted. Factory prefabrication also elim-inates the need for moving tools from one place to another because thetools are already present and organized at the factory.

Second, workability increased because the working height was

lower in factory prefabrication than in the conventional method,wherein the workers needed to climb ladders or operate lifting equip-ment for corridor MEP work at heights. Factory manufacturing mini-mized the operation of the lifting equipment and the use of ladders.Moreover, the eye-level working environment increased worker

Fig. 7. Design and coordination process of the conventional method and MTP.

Fig. 8. Spool drawing of an MTP module.

Fig. 9. Conventional construction and factory manufacturing (photographed by Sejun Jang).

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

91

comfort.Third, MTP minimized the number of sub-workers for the MEP

work. In the conventional method, MEP work is sometimes undertakenby a team rather than a single worker. In the case of installation work

using the scissor lift equipment, the main worker requires a sub-workerto deliver the construction materials or tools to the main worker. Thesafety standard in Korea regulates that MEP work using lifting equip-ment must be performed together by the main worker and the

Fig. 10. Comparative analysis of the manufacturing and assembly process in the conventional method and MTP.

Fig. 11. Laser-scanned corridor for sleeve positioning (left) and a detected difference (right).

Fig. 12. Locating the MTP module (left) and BIM-based site logistics (right).

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

92

supplementary worker. Hence, less skilled workers are deployed at thebottom of the main MEP work. However, lifting equipment is not re-quired when performing factory manufacturing. Accordingly, a sub-worker is not required for MEP installation.

Fourth, factory manufacturing minimized the spatial interferencescaused by workers of different trades working and moving in the samespace in the conventional method (Fig. 9). Due to these spatial inter-ferences, the work often needed to be temporarily stopped and thenrestarted. However, the sequential process employed in factory

manufacturing prevented the interference between tasks (Fig. 9).Hence, productivity was expected to be higher in MTP than in the

conventional method.The multi-trade construction process was expected to differ from the

conventional method in the following ways: the MEP installation workin the conventional method was executed after the pouring and curingof the concrete, sequentially followed by the trades. The total time re-quired after the concrete work was [a + b + c + d], where [a, b, c, d]is the installation time for each trade [A, B, C, D] (Fig. 10, upper panel).

Fig. 13. Connecting modules to each other on the slab and lifting them as one piece.

Fig. 14. Analysis of the labor input between the conventionalmethod and MTP.

Fig. 15. Labor input decrease through continuous MEP fabrica-tion work compared to conventional construction baseline.

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

93

The various trade activities in the conventional method may overlapbecause the relationship between the activities was of the start-to-finishtype, thereby leading to inefficiency because of worker overcrowdingand shared material stocking.

Meanwhile, in MTP, MEP trade tasks can be executed either inparallel with or before the concrete pouring and curing. Almost all theMEP installations in the conventional method were conducted on a slab,while a large portion of the MEP installations in MTP can be executedaway from the site (i.e., the concrete work and MEP manufacturingshare a finish-to-finish relationship). The time for the MEP work [a + b+ c + d] may overlap with the concrete work, which means that thesame finish time can be set for both the MEP and concrete work(Fig. 10, lower panel). The MTP modules can be assembled and con-nected after the concrete curing. Therefore, the total factory pre-fabrication duration was shorter than that the duration in the conven-tional process.

In the conventional method, fine adjustments to the MEP elementswere made on the site when the sleeve position in the shop drawingswas different from the site conditions. In contrast, with an MTP module,such differences entail high risks and may necessitate the modificationof the entire MTP module. Therefore, in the coordination phase, itsconnection with the structural framing must be considered in additionto the routing of each MEP trade. Moreover, each MEP element must beconnected to the structural framing in the prefabrication phase.Therefore, an unexpected onsite error could necessitate multiple mod-ifications (e.g., in MEP routing, structural framing, or connection de-tail). Thus, as an additional method for overcoming this increased risk,laser scanning must be performed after concrete work to obtain anaccurate site condition. The obtained point cloud data can be over-lapped with the BIM models for modifying the MTP modules if neces-sary (see Fig. 11 for an example involving a cable tray). Thus, MTPrequired a higher level of technological capabilities than the conven-tional method. MTP clearly entails a high risk of rework.

In addition, a BIM-based site logistics was provided to simulate thelifting and moving of the MTP modules (Fig. 12). Each MEP element inthe conventional method was relatively small and light and, thus, couldbe easily moved by the workers. On the contrary, the size and theweight of the construction unit in MTP were much higher than in theconventional method because various trade elements were combinedand integrated with the structural framing, thereby increasing the dif-ficulty of site logistics. Therefore, BIM-based site logistics should beperformed before moving and lifting the MTP module to place themodule at the accurate location at the right time.

In conclusion, the use of MTP modules increased the constructionrisks, thereby necessitating more advanced techniques (e.g., laserscanning and BIM-based site logistics) than the conventional method.

MTP modules on the same floor were connected to each other on theslab and lifted as one piece to the ceiling (Fig. 13). The MEP modules

were assembled on the slab because it was more efficient than con-necting them at the ceiling-height level using a ladder or lift. A singleworker can connect four modules on the same floor in a single workingday.

The installation of each MEP trade in the conventional methodproceeded sequentially after the concrete work. Therefore, the timerequired for each MEP installation contributed separately to the totalMEP installation duration. In contrast, MTP shortened the schedule byapproximately five weeks because only the connection and assembly ofthe MTP module needed to be executed onsite. In other words, the MTPmodule can be manufactured while the concrete work is being con-ducted. The MEP work durations in MTP and the conventional methodwere calculated as follows.

Total duration for MEP after concrete work:

= + + + + + =Duration of the conventional method a b c d e f 36 days(1)

= + =Duration of the MTP f g 3 days, (2)

where

a: duration of duct work (6 days)b: duration of HVAC work (6 days)c: duration of sanitary pipe work (6 days)d: duration of electrical and IBS pipe work (6 days)e: duration of fire protection pipe work (8 days)f: duration of inspection (1 day)g: duration of connecting and installing prefabricated modules(2 days)

6. Productivity analysis

One of the major objectives of adopting the prefabrication tech-nology is to ensure economy through productivity improvements[10,27,31,34]. In this study, a productivity analysis was conducted bymeasuring the labor input (in hours) for manufacturing and assembly.Approximately five weeks were spent in measuring the labor input in-volved in manufacturing, transporting, lifting, and assembling the MTPmodules for four floors.

The measured total labor input in MTP was compared with thebaseline of the Korea Estimation Standard [19]. The Korean govern-ment baseline standard formulated the labor input (in person-days)considering the material quantity, duct work (m2), pipe work (m), andtypes and number of connections. The baseline of the labor inputwithout adopting an adjustment factor accounting for the increasedestimated labor input for more strenuous working conditions (i.e.,conventional method) was compared with the measured labor input(MTP).

For transporting and lifting work, the Korean government does not

Fig. 16. Labor input decrease through continuous structuralframing fabrication work.

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

94

provide an accurate baseline or guide for the estimation of labor input.Therefore, the transporting and lifting work was excluded from theproductivity analysis.

6.1. Analysis of the labor inputs (conventional vs. MTP)

MTP on a corridor MEP rack has not yet been implemented in Korea.Therefore, the workers participating in the case project had no ex-perience as regards MTP. The workers in the manufacturing factoryinitially seemed unfamiliar with the changed environment and process.The supervisor had to instruct the workers individually in detail at theearly stage. However, the workers acclimatized to the changes and latermanufactured the MTP module without direct supervision.

The total labor input (manufacturing work in factory + assemblywork onsite) in MTP was 129.7 person-days, whereas the baseline in theKorean government standards was 114.3 person-days, meaning thatcompared with the conventional method, an additional 13.5% labor isrequired when applying MTP.

The sum of the labor input in the MTP factory, except for thestructural framing and the site assembly, was 77.3 person-days (67.7%in the conventional method). When considering manufacturing only,the labor input of almost all the trades, except that of fire protection,decreased compared with their counterparts in the conventionalmethod. The largest decrease in labor was in the HVAC duct trade(63.8%). The labor input for the HVAC pipe decreased by 27.4%, for thesanitary pipe by 41.8%, and for the electrical and IBS tray by 8.1%. Thelabor for fire protection increased by 3.2%. Fire protection entailedhigher levels of connection work (welding type) with the main pipe andthe branch pipe. The modification of the connection joints onsite wascumbersome because of the nature of MTP. Arranging a workbench forwelding was also labor-intensive. Therefore, the fire-protection laborslightly increased.

The labor input of the structural framing and the onsite assemblyaccounted for 40.4% of the total labor input in MTP. These were the keyactivities that decreased the productivity in MTP. Fig. 14 illustrates thedifferences in the labor input between the conventional method andMTP in our case study.

The result of our productivity analysis differed considerably fromthat of Skanska, which reported a 300% increase in productivity in MTP[7] but was similar to that of Mortenson Construction, who reported a15.2% productivity decrease in applying MTP to the corridor MEP rackwork [2]. These differences likely emerged because of the differences infabrication, MTP module design, factory environments, and the skilllevel of workers. Effective comparisons may be impossible.

6.2. Analysis of the labor inputs of MTP

The learning effect was identified by the labor input analysis. Themanufacturing work was sequentially performed from the fifth to theeighth floors. We separately calculated the labor input for each floor(Fig. 15) and found that the productivity progressively increased.Compared with the baseline, a 36.7% labor input was required in theHVAC duct work on the fifth floor. This value slightly decreased withrepetitive work. In the electrical and IBS tray work on the initial floor,99.0% labor input was required compared with the baseline, whichthen decreased to 70.3% in the final floor. Fire-protection work re-quired 135.0% labor input compared with the baseline for the initialfloor, which decreased to 69.5% in the final floor. Thus, the pro-ductivity improved because of the repetitive nature of the work and wasidentified as one of the major factors affecting the prefabrication pro-ductivity.

The productivity showed an increasing pattern with the repeatedmanufacturing work, which was also identified in the structuralframing work. The structural framing of the MTP module is yet to be-come a common construction method. Thus, the government baselinewas not yet available in the Korea Estimation Standard [19]. The

productivity of the structural framing work was measured in person-days per 1 m of the framing members. The results of the productivityanalysis showed that the labor input decreased from 0.133 to 0.029person-days with repeated work, which was similar to the case of theMEP trade (Fig. 16).

The variation in the labor input after repeated work can be pre-dicted using the learning curve model. The oldest and most commonlyused learning curve model is the Wright Model [16]:

= −Y A X ,x 1n (3)

where Yx is the production time (or labor input) of the Xth product; A1

is the working time of the first unit (or labor input); and X is the re-petition cycle number. The learning rate L is 2−n. The productivitychanges discussed in Section 7.2 were calculated using the WrightModel.

7. Economic analysis

An economic analysis was conducted in this section using the pro-ductivity analysis results from the case study. The productivity analysisshowed that the MTP application increased the labor input by 13.5%compared with that in the conventional method. However, MTP couldbe economic in certain situations because the baseline labor inputchanges according to the working condition. In addition, the learningeffect increased the productivity. Finally, the relatively lower wagecosts for the factory work reduced the project costs the project sche-dule.

7.1. Setup of a baseline for the labor input

The labor input of the case study was calculated using the Koreangovernment standard. The MTP application in a typical, simple, smallproject could be disadvantageous in terms of productivity. However,the compared baseline [19] varied with the working conditions, such asclimate, ceiling height, and building height. In addition, the US esti-mation standard for labor input has an adjustment factor to increase thebaseline according to the climate, working height, building story, andbuilding size [35]. In addition, a consistent productivity could be ex-pected in MTP by ensuring a uniform working environment, which isnot the case in the conventional method [31]. For these reasons, thegovernment has presented some labor input adjustment factors for useby contractors to arrive at a rational bid price.

The following example is an adjustment factor of the labor inputestimation by the Korean government standard and a contractor (ENRInternational, which is ranked among the top 15). The labor input ad-justment factor (%) adds a premium to the baseline labor input ac-cording to the status of the project and the site conditions. For example,when the expected labor is 100 person-days by the government stan-dard, the total labor input is changed to 110 person-days if the workingsituation indicates a 10% labor input adjustment factor.

1. Standard of the Labor Input Adjustment Factor According to theBuilding Specification(A) Increase of the Labor Input Estimation According to the Building

Story—(Korea Estimation Standard)– 2–5 stories: 1%– 6–10 stories: 3%– 11–15 stories: 4%– 16–20 stories: 5%– 21–25 stories: 6%– Increase 1% per 5 stories after 26 stories

(B) Increase of the Labor Input Estimation According to the BuildingType—(A Contractor)– Hospital and Health Care Project: 15%– Biology Medicine Drug Plant: 20%– Semiconductor Plant: 20%

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

95

– Complex Building: 10%2. Standard of the Labor Input Adjustment Factor According to the

Working Space Specification(A) Increase of the Expected Labor According to the Ceiling

Height—(Korea Estimation Standard)– 0–5 m: 0%– 5–10 m: 20%– 10–15 m: 30%– 15–20 m: 40%– 20–30 m: 50%– 30–40 m: 60%– Increase 10% per 10 m after 40 m

(B) Increase of the Labor Input Estimation to Underground Work —(Korea Estimation Standard)– First underground floor: 1%– Second–fifth underground floors: 2%

(C) Increase of the Labor Input Estimation to MEP Room—(KoreaEstimation Standard)– MEP room or similar: 50%

3. Standard of the Labor Input Adjustment Factor According to the SiteSurroundings—(Korea Estimation Standard)(A) Site location is in a military operation area: 20%(B) Site location is not on a modernized road: 50%

Using the adjustment factor of the labor input, the MTP applicationcould be more economical than the conventional method in terms ofproductivity (see Table 1 for an example). This study compared thetotal labor of MTP to the baseline of the government estimation stan-dard (100%) without the adjustment factor. However, the adjustmentfactor would increase the estimation of the labor input by 3%, 15%, and20% if the working conditions are assumed to be eight hospital floorsand 8 m ceiling height. Thus, the baseline (100%) would become 138%.In the productivity results of the case project, the labor input increasedto 113.5% compared with the conventional method (100%), but theMTP was more efficient than the conventional method (Table 1). Thus,whether to apply MTP according to the type of working situation of aproject can be determined. The MTP application could be more eco-nomical in terms of productivity if the project is complex and theworking conditions are difficult.

7.2. Economic impact of the learning effect

The productivity for MTP modules was changed according to therepeated work by floor number. Compared with the baseline of theKorea Estimation Standard [19], the labor input (manufacturing andassembly) on the fifth floor was 137.0%, the labor input on the sixthfloor was 121.6%, that on the seventh floor was 98.6%, and that on thefinal floor was 95.6%. That is, the productivity of MTP had alreadyexceeded that of the traditional method by the seventh floor due to thelearning effect [51].

The accumulated productivity, however, did not reach a breakevenpoint on the eighth floor. To estimate the breakeven point for the ac-cumulated productivity, the learning curve was calculated based on the

Wright Model (Eq. (3)) using the data points up to the eighth floor andthe goodness measures calculated by a curve-fitting tool. The learningcurve was y= 1.389175 × x−0.2703122. Although the productivity ofMTP had already exceeded that of the traditional method by the se-venth floor (Fig. 17), mathematically the lost productivity at the be-ginning of MTP was expected to be recovered after the twelfth floor.This means that when untrained workers work on an MTP project, thesize of the building should be large enough to recover productivity lossthrough the learning effect.

The initial productivity of MTP was inefficient compared with theconventional method because of the workers' unfamiliarity with MTP.However, as the number of projects that deploy MPT increases, the skilllevel of workers and the productivity of MTP will increase.

7.3. Analysis of the total cost impact

This section analyzes the total economic impact of MTP. Manystudies have reported that MTP yields positive qualitative effects, suchas improvement in quality, reductions in accidents, and reductions inenvironmental problems [2,7,27–29]. This study, however, excludesqualitative factors from the cost analysis because they cannot be ob-jectively calculated.

As discussed in the Process Analysis section, the application of MTPrequired more time and effort for designing and detailing before theproduction phase than the conventional method, which conforms to thefindings of a previous study [44]. The positive effects to direct cost wereincurred as the person-hours, the labor cost, and the material unit pricedecreased. First, the person-hours decreased because of the factorymanufacturing and the repeated working condition. Second, the laborcost decreased because the factory workers had relatively lower wages(i.e., approximately 80% of that of the field workers) [7]. Third, thematerial unit price could be reduced through bulk purchasing [40]. Thenegative cost effects of MTP included the costs for additional structuralframing members, which was the key for maintaining the stability ofthe delivery, lifting, and assembly of MTP modules and for structurallysupporting the combined MEP system. Therefore, the material costs andquantity in MTP increased [29].

The MTP application increased the transportation and (indirect)factory operation costs. Each piece of MEP material was moved to themanufacturing factory, and the manufactured MTP module had to bedelivered to the construction site. Thus, MTP increased the transpor-tation costs. These costs varied depending on the location and thedistance of the manufacturing factory and could be minimized by es-tablishing an onsite prefabrication yard. In addition, the factory op-eration cost was added to the overall MTP cost. The manufacturingfactory issue is known to be one of the main reasons why the con-tractors were reluctant to adopt MTP [27].

Table 1Illustrative implementation of the adjustment factor to the estimation baseline of thelabor input.

Baseline Adjustment factor Changed baseline

100% a b c 100% × (a + b+ c)

Example 8-Storybuilding

Hospital Ceilingheight: 8 m

100% 3% 15% 20% 100% × (3%+ 15%+ 20%) = 138%

Fig. 17. The productivity improvement from the learning effect.

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

96

Many studies have reported that one of the biggest changes in MTPapplication is the reduction in schedule [2,7,27,29]. Schedule reductioncan provide additional cost-savings in regard to the total constructioncosts, and buildings can be handed over to clients more quickly. Thus,the contractor can obtain additional incentives from the client throughMTP. Reducing the cost for the supervisor and the site manager alsobecame possible because of the schedule reduction [7,30]. Further-more, the cost of safety management can be reduced by reducing theworking time onsite [2].

These economic changes can be summarized using the followingequation, where the denominator is the total cost in the conventionalmethod, and the numerator is the total cost in MTP. The MTP appli-cation would be economical if the cost–benefit ratio is greater than one.

=

∑ + ∑

+ ∑

∑ + ∑

+ ∑

=− + − + +

+ + + +

Cost‐Benefit Ratio( Material Cost Savings) ( Labor Cost Savings)

(Additional Benefits)(Structural Framing Cost) (Additional Design Cost)

(Additional Operation Cost)(MC MC ) (LC LC ) (IC SC)

(MC LC ) (DC ) (TC RC)o p o p

F F P (4)

where

MCO = onsite material cost (purchasing by each project).MCP = expected offsite material cost (purchasing by bulk)LCO = onsite labor costLCP = expected offsite labor costIC = cost incentive from client by schedule reductionSC = additional cost benefits from minimized supervisionMCF = additional material cost for structural framingLCF = additional labor cost for structural framingDCP = design and detailing cost for offsite constructionTC = additional transportation costRC = additional rent and operation cost for factoryEconomical when Cost Benefit Ratio ≥ 1Not Economical when Cost Benefit Ratio < 1

The total cost impact varies depending on the various factors of theproject properties. For example, even if the expected labor input rangesbetween 100% and 120% of the baseline, MTP could be more efficientthan the conventional method (when the other conditions remain thesame) if the wage of a worker can be lowered to 80% of that with theconventional method [7]. In addition, even if productivity and othercost factors are predicted, compared with the conventional method, alower material cost caused by bulk purchasing can lead to a positiveeconomic impact. Securing a positive economic impact through aschedule shortening incentive is also possible, which is one of thebiggest reasons for adopting MTP.

8. Conclusions

This study examined the effectiveness of BIM-based MTP throughthe process, productivity, and economic analyses of a case project in-volving a heavily equipped multi-trade corridor MEP rack installationin a nine-story 62,755 m2 multi-purpose exhibition complex building.Unlike the general expectations based on some of the previous casestudies, the analyses herein showed that BIM-based MTP and pro-ductivity were lower compared to the traditional corridor rack in-stallation method even if the productivity was improved through thelearning effect. The major analysis results and findings are as follows:

1) Process analysis: MTP took longer to design and coordinate than theconventional onsite installation method. In the case study, MTP tookapproximately four weeks to design and coordinate, whereas

manufacturing and onsite installation took approximately fiveweeks because a structural framing used to hold the MEP elementswas required to deliver the assembled corridor MEP rack offsite tothe construction site and so were the additional coordination pro-cesses between structural framing and MEP elements and betweenstructural framing and site conditions. In addition, a spool drawinggeneration process for factory assembly was required for a detailedfactory production. Another factor was the unfamiliarity with newtechnologies for 3D coordination, such as 3D laser scanning andBIM-based site logistics. Nevertheless, the overall project durationwas shorter than that in the conventional process because of theparallel execution of the concrete and MTP work. As describedbelow, the case study also demonstrated the possibility of shorteningthese processes as the workers became used to the processes and thetechnologies.

2) Productivity analysis: MTP had a lower productivity than the onsiteinstallation. MTP required 113.5% labor input compared to theconventional onsite installation. When only factory work was con-sidered, the labor input of each trade was significantly reduced to67.7% compared to the conventional method. However, the addi-tional two processes (i.e., structural framing and onsite assembly),and coordination between them took 40.4% of the labor input of theoverall MTP work, which was the main reason for the lower pro-ductivity in MTP. Nevertheless, this inefficiency was mitigatedthrough the learning effect.

3) Economic analysis: The economic analysis showed that a projectshould be of a certain size to be able to gain economic benefits. Inthe future, the productivity and economic benefits are also likely toincrease with the increase in the number of MTP-trained workers. Inaddition, the productivity and economic benefits can be improved ifa building is designed considering MTP or work environments, andthe site logistics for MTP.

The findings are expected to contribute to academia and industry inthe following aspects: First, previous MTP case studies reported only thecontradicting results of MTP productivity without reasoning or a de-tailed analysis about the outcomes [7,30], whereas the present studyanalyzed and identified the factors that affected the productivity andthe economic benefits of BIM-based MTP. Second, the factors identifiedthrough this study can be used as a basis for improving BIM-based MTPin the future. The identified factors included increased coordinationtime onsite and offsite, spool drawing generation time for factory as-sembly, use of unfamiliar technology (e.g., 3D laser scanning and BIM-based site logistics), and workers' unfamiliarity with MTP. Fortunately,all these are types of factors that can be addressed through training andprocess improvement.

This study inherited all the limitations of a case study, including asmall sample size and weak generalizability and reliability of the data,as well as the advantages of a case study, including an in-depth analysisof a case with rich and qualitative information. This case study wasconducted for over 37 months, and the data were analyzed and col-lected through daily meetings, emails, etc. The specific numbers re-garding time, productivity, and economic benefits may vary by project,but the influential factors and their potential impact on the durationand productivity of MTP will still be valid. Studies to improve the BIM-based MTP process are expected to follow in the near future.

Acknowledgments

This research was supported by a grant (17RTRP-B104237-03) fromthe Architecture & Urban Development Research Program funded bythe Ministry of Land, Infrastructure and Transport of the KoreanGovernment and by the Human Resources Program in EnergyTechnology of the Korea Institute of Energy Technology Evaluation and

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

97

Planning (KETEP), granted financial resource from the Ministry ofTrade, Industry and Energy, Republic of Korea (No. 20174010201320).

References

[1] F.H. Abanda, J.H.M. Tah, F.K.T. Cheung, BIM in off-site manufacturing for build-ings, J. Build. Eng. 14 (2017) 89–102, http://dx.doi.org/10.1016/j.jobe.2017.10.002.

[2] E.I. Antillón, M.R. Morris, W. Gregor, A value-based cost-benefit analysis of pre-fabrication processes in the healthcare sector: a case study, Proc. 22nd AnnualConference of the International Group for Lean Construction, June 25–27, 2014, pp.995–1006 Oslo, Norway https://iglcstorage.blob.core.windows.net/papers/attachment-198ef25e-8612-4637-a5d6-ef885d897844.pdf.

[3] A. Arieff, B. Burkhart, Prefab, Gibbs Smith, Layton, Utah, USA, 2002 (ISBN:9781586851323).

[4] L.E. Bygballe, M. Ingemansson, The logic of innovation in construction, Ind. Mark.Manag. 43 (2014) 512–524, http://dx.doi.org/10.1016/j.indmarman.2013.12.019.

[5] C.M. Clevenger, R. Khan, Impact of BIM-enabled design-to-fabrication on buildingdelivery, Pract. Period. Struct. Des. Constr. 19 (2014) 122–128, http://dx.doi.org/10.1061/(ASCE)SC.1943-5576.0000176.

[6] C. Davies, The Prefabricated Home, Reaktion Books, Islington, UK, 2005 (ISBN:9781861892430).

[7] T. Fishking, Integrating BIM and Prefabrication in Healthcare Facility Design, NBBJ,Columbus, Ohio, USA, 2011 Available http://www.ahcaseminar.com/Speaker%20pdfs/FacilityEngineeringSessions/Fishking_Integrating%20BIM%20and%20Prefabrication%20Presentation.pdf , Accessed date: 22 July 2017.

[8] C.I. Goodier, A.G.F. Gibb, Barriers and opportunities for offsite in the UK, Proc. 11thJoint CIB International Symposium Combining Forces, June 13–16, 2005, pp.148–158 Helsinki, Finland https://dspace.lboro.ac.uk/2134/5451.

[9] R. Hällmark, H. White, P. Collin, Prefabricated bridge construction across Europeand America, Pract. Period. Struct. Des. Constr. 17 (2012) 82–92, http://dx.doi.org/10.1061/(ASCE)SC.1943-5576.0000116.

[10] S. Hoover, P. Trombitas, E. Cowle, FMI/BIM Forum Prefabrication Survey, FMICorporation, Raleigh, North Carolina, USA, 2017, p. 2017. Available https://www.fminet.com/wp-content/uploads/2017/02/PrefabricationSurvey_2017.pdf ,Accessed date: 3 September 2017.

[11] Huamanchina, Development of SMC Sanitaryware, [Online], Available http://www.huamanchina.com/en/pro.asp?id=1, (2011) , Accessed date: 26 July 2016.

[12] M. Hurd, PCI J.: Precast Concrete Homes for Safety, Strength and Durability,Available https://www.pci.org/PCI_Docs/Project_Resources/housing/jl-94-march-april-4.pdf, (1994) , Accessed date: 17 June 2017.

[13] L. Jaillon, C.S. Poon, Sustainable construction aspects of using prefabrication indense urban environment: a Hong Kong case study, Constr. Manag. Econ. 26 (2008)953–966, http://dx.doi.org/10.1080/01446190802259043.

[14] Y.S. Jeong, C.M. Eastman, R. Sacks, I. Kaner, Benchmark tests for BIM data ex-changes of precast concrete, Autom. Constr. 18 (2009) 469–484, http://dx.doi.org/10.1016/j.autcon.2008.11.001.

[15] S.A. Jones, Global Industry Trends with Building Information Modeling (BIM),Construction Innovation 2013 Forum, Melbourne, Australia, Available https://www.bimmepaus.com.au/wp-content/uploads/2016/05/mcgraw%20hill%20global%20trends%20with%20bim_.pdf, (2013) , Accessed date: 13 June 2017.

[16] F. Jordan Srour, D. Kiomjian, I. Srour, Learning curves in construction: a criticalreview and new model, J. Constr. Eng. Manag. (2015) 06015004, , http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0001096.

[17] KHS&S, Prefabrication: How KHS&S is Changing the Future of Construction,Anaheim, California, USA, Available http://besbuilt.com/uploads/khs_and_n/Prefabrication_Credential.pdf, (2013) , Accessed date: 2 January 2017.

[18] M.-K. Kim, Q. Wang, J.-W. Park, J.C.P. Cheng, H. Sohn, C.-C. Chang, Automateddimensional quality assurance of full-scale precast concrete elements using laserscanning and BIM, Autom. Constr. 72 (2016) 102–114, http://dx.doi.org/10.1016/j.autcon.2016.08.035.

[19] Korea Estimation Standard, Korea Labor Estimation Standard, Degunsa, Seoul,South Korea, (2015) (ISBN: 8989502551).

[20] K. Ku, P. Broadstone, A. Colonna, Towards on-site fabrication: a case study onmulti-trade prefabrication, Proc. 2012 ACSA Fall Conference, September 27–29,2012, pp. 52–57 Philadelphia, Pennsylvania, USA http://apps.acsa-arch.org/resources/proceedings/uploads/streamfile.aspx?path=ACSA.FALL.12&name=ACSA.FALL.12.10.pdf.

[21] M. Lawson, R. Ogden, C. Goodier, Design in Modular Construction, Taylor &Francis, Abingdon, UK, 2014 (ISBN: 9780415554503).

[22] G. Lee, S. Kim, Case study of mass customization of double-curved metal Façadepanels using a new hybrid sheet metal processing technique, J. Constr. Eng. Manag.138 (2012) 1322–1330, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000551.

[23] Z. Li, G.Q. Shen, M. Alshawi, Measuring the impact of prefabrication on construc-tion waste reduction: an empirical study in China, Resour. Conserv. Recycl. 91(2014) 27–39, http://dx.doi.org/10.1016/j.resconrec.2014.07.013.

[24] N. Lu, The current use of offsite construction techniques in the United States con-struction industry, Proc. Construction Research Congress 2009: Building aSustainable Future, April 5–7, 2009, pp. 946–955 Seattle, Washington, USA https://doi.org/10.1061/41020(339)96.

[25] N. Lu, T. Korman, Implementation of Building Information Modeling (BIM) inmodular construction: benefits and challenges, Proc. Construction ResearchCongress 2010: Innovation for Reshaping Construction Practice, May 8–10, 2010,pp. 1136–1145 Banff, Alberta, Canada https://doi.org/10.1061/41109(373)114.

[26] G.P. Luth, VDC and the engineering continuum, J. Constr. Eng. Manag. 137 (2011)

906–915, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000359.[27] McGraw Hill, SmartMarket Report: Prefabrication and Modularization: Increasing

Productivity in the Construction Industry, New york city, New york, USA, Availablehttps://www.nist.gov/sites/default/files/documents/el/economics/Prefabrication-Modularization-in-the-Construction-Industry-SMR-2011R.pdf, (2013) , Accesseddate: 7 January 2017.

[28] Mortenson Construction, Building Knowledge in Design and Construction: LeanConcepts Across Delivery Methods: Mortenson's Lean Journey, Minneapolis,Minnesota, USA, Available https://www.leanconstruction.org/media/docs/chapterpdf/chicago/6-LCI-Chicago-Presentation-Reduced.pdf, (2014) , Accesseddate: 30 October 2017.

[29] Mortenson Construction, Building What's Next: Prefabrication Faster, Better,Smarter: Reduce the Overall Cost and Time of Project Delivery While Increasing theQuality and Scope, Minneapolis, Minnesota, USA, Available https://www.mortenson.com/~/media/files/pdfs/prefabrication-building-whats-next-cost-time-quality-scope.ashx, (2014) , Accessed date: 3 October 2017.

[30] Mortenson Construction, Prefabrication Research Study Results: Prefabrication:Benefits & Drivers for Successful Implementation, Minneapolis, Minnesota, USA,Available http://www.mortenson.com/prefab-study, (2014) , Accessed date: 19October 2017.

[31] J. O'Connor, W. O'Brien, J. Choi, Standardization strategy for modular industrialplants, J. Constr. Eng. Manag. 141 (2015) 04015026, , http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0001001.

[32] J.T. O'Connor, W.J. O'Brien, J.O. Choi, Critical success factors and enablers foroptimum and maximum industrial modularization, J. Constr. Eng. Manag. 140(2014) 04014012, , http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000842.

[33] W. Pan, A.G.F. Gibb, A.R.J. Dainty, Perspectives of UK housebuilders on the use ofoffsite modern methods of construction, Constr. Manag. Econ. 25 (2007) 183–194,http://dx.doi.org/10.1080/01446190600827058.

[34] E.A. Poirier, S. Staub-French, D. Forgues, Measuring the impact of BIM on laborproductivity in a small specialty contracting enterprise through action-research,Autom. Constr. 58 (2015) 74–84, http://dx.doi.org/10.1016/j.autcon.2015.07.002.

[35] R. Pray, National Construction Estimator, Craftsman Book Company, Carlsbad,California, USA, 2015 (ISBN: 9781572183162).

[36] I.J. Ramaji, A.M. Memari, Extending the current model view definition standards tosupport multi-storey modular building projects, Architectural Engineering andDesign Management, 2017, pp. 1–19, , http://dx.doi.org/10.1080/17452007.2017.1386083.

[37] C. Rosner, T. Shafer, R. Byres, Rapid pier delivery using precast concrete compo-nents, Proc. Ports 2010: Building on the Past, Respecting the Future, April 25–28,2010, pp. 1098–1107 Jacksonville, Florida, USA https://doi.org/10.1061/41098(368)113.

[38] Ryan E. Smith, History of prefabrication: a cultural survey, Proc. ThirdInternational Congress on Construction History, May 20–24, 2009, pp. 1355–1364Cottbus, Germany http://www.bma.arch.unige.it/pdf/construction_history_2009/vol3/smith-ryan_vw_paper_layouted.pdf.

[39] R. Sacks, C. Eastman, G. Lee, D. Orndorff, A target benchmark of the impact ofthree-dimensional parametric modeling in precast construction, PCI J. 50 (2005)126–139, http://dx.doi.org/10.15554/pcij.07012005.126.139.

[40] H. Said, Prefabrication best practices and improvement opportunities for electricalconstruction, J. Constr. Eng. Manag. 141 (2015) 04015045, , http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0001018.

[41] K. Saindon, M. McMullen, Curved precast prestressed concrete bridge girder sys-tems, Proc. Structures Congress 2010, May 12–15, 2010, pp. 51–58 Orlando,Florida, USA https://doi.org/10.1061/41130(369)6.

[42] S. Sharma, A. Sawhney, M. Arif, Parametric modelling for designing offsite con-struction, Procedia Eng. 196 (2017) 1114–1121, http://dx.doi.org/10.1016/j.proeng.2017.08.069.

[43] M.M. Singh, A. Sawhney, A. Borrmann, Integrating rules of modular coordination toimprove model authoring in BIM, Int. J. Confl. Manag. (2017) 1–17, http://dx.doi.org/10.1080/15623599.2017.1358077.

[44] R.E. Smith, J. Timberlake, Prefab Architecture: A Guide to Modular Design andConstruction, Wiley, Hoboken, New jersey, USA, 2011 (ISBN: 9780470880463).

[45] V.W.Y. Tam, C.M. Tam, S.X. Zeng, W.C.Y. Ng, Towards adoption of prefabricationin construction, Build. Environ. 42 (2007) 3642–3654, http://dx.doi.org/10.1016/j.buildenv.2006.10.003.

[46] C. Tatum, Management challenges of integrating construction methods and designapproaches, J. Manag. Eng. 5 (1989) 139–154, http://dx.doi.org/10.1061/(ASCE)9742-597X(1989)5:2(139.

[47] H.P. Tserng, Y.L. Yin, E.J. Jaselskis, W.-C. Hung, Y.-C. Lin, Modularization andassembly algorithm for efficient MEP construction, Autom. Constr. 20 (2011)837–863, http://dx.doi.org/10.1016/j.autcon.2011.03.002.

[48] A. Watson, Digital buildings – challenges and opportunities, Adv. Eng. Inform. 25(2011) 573–581, http://dx.doi.org/10.1016/j.aei.2011.07.003.

[49] P. Wu, S. Low, Lean management and low carbon emissions in precast concretefactories in Singapore, J. Archit. Eng. 18 (2012) 176–186, http://dx.doi.org/10.1061/(ASCE)AE.1943-5568.0000053.

[50] J. Zhang, Y. Long, S. Lv, Y. Xiang, BIM-enabled modular and industrialized con-struction in China, Procedia Eng. 145 (2016) 1456–1461, http://dx.doi.org/10.1016/j.proeng.2016.04.183.

[51] L. Zhang, X. Zou, Z. Kan, Improved strategy for resource allocation in repetitiveprojects considering the learning effect, J. Constr. Eng. Manag. 140 (2014)04014053, , http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000896.

[52] R.Y. Zhong, Y. Peng, F. Xue, J. Fang, W. Zou, H. Luo, S. Thomas Ng, W. Lu,G.Q.P. Shen, G.Q. Huang, Prefabricated construction enabled by the internet-of-things, Autom. Constr. 76 (2017) 59–70, http://dx.doi.org/10.1016/j.autcon.2017.01.006.

S. Jang, G. Lee Automation in Construction 89 (2018) 86–98

98