Post simulation visualization model for effectivescheduling of modular building construction1

Mansooreh Moghadam, Mohamed Al-Hussein, Saad Al-Jibouri, and Avi Telyas

Abstract: The factory-based modular construction process has proven to increase the speed of construction, and improvequality and safety, while providing value to the customer and a rapid return on investment to the builder and owner. How-ever, onsite module assembly creates new schedule demands, as activities are scheduled on a minute-by-minute basis; there-fore simulation of the process becomes essential at early stages of a project. Although simulation proves to be an effectivetool for project engineers to assess complex construction operations, it remains a symbolic base model with no visual linkto the actual physical shape and look of the project’s activities. This paper presents the application of integrated simulationand post simulation visualization as a tool to assist the modular construction industry in scheduling onsite installation ofprefabricated modules. The proposed methodology uses simulation model output as an ASCII file in a binary format andimports this ASCII file to 3D Studio Max to perform the animation. The output from the high level simulation model istransformed into frames/second in 3D Studio Max. The proposed methodology was tested on the planned construction of a34-storey building in Brooklyn, New York, USA. Simulation visualization of the process proved to be effective in communi-cating the value and simplicity of a minute-by-minute schedule. Based on the output information, the most efficient solu-tions were generated. The use of post simulation visualization was effective in analyzing the construction methods of thecase study which consisted of 950 structural steel modules. Issues related to construction activities’ productivity weresynchronized to achieve onsite installation of the project in only 56 working days.

Key words: simulation, visualization, modular building, schedule.

Résumé : Il a été démontré que le procédé de construction modulaire en usine augmente la vitesse de construction et amé-liore la qualité et la sécurité tout en fournissant une valeur au client ainsi qu’un rapide retour sur l’investissement pour leconstructeur et le propriétaire. Toutefois, l’assemblage des modules sur place engendre de nouvelles demandes sur l’échéan-cier puisque certaines activités sont programmées à la minute; la simulation d’un tel procédé devient donc essentielle auxpremières étapes d’un projet. Bien que la simulation s’avère un outil efficace pour les ingénieurs de projet pour évaluer lesopérations complexes de construction, elle demeure un modèle de base symbolique sans lien visuel avec la forme physiqueréelle et l’apparence des activités d’un projet. Le présent article montre l’application de la simulation intégrée et de la visua-lisation post-simulation comme outil pour aider l’industrie de la construction modulaire à programmer sur le site l’installa-tion de modules préfabriqués. La méthodologie proposée utilise les résultats d’un modèle de simulation comme un dossierASCII en format binaire et importe ce dossier ASCII dans 3D Studio Max pour réaliser l’animation. Les résultats de ce mo-dèle de simulation avancé est transformé en images par seconde grâce à 3D Studio Max. La méthodologie proposée a étémise à l’épreuve sur la construction planifiée d’un bâtiment de 34 étages â Brooklyn, New York, É.-U. La visualisation dela simulation du procédé s’est avérée efficace pour communiquer la valeur et la simplicité d’un échéancier minute par mi-nute. Les solutions les plus efficaces ont été générées en se basant sur l’information obtenue des résultats. L’utilisation de lavisualisation post-simulation était efficace pour analyser les méthodes de construction de l’étude de cas, lequel comportait950 modules structuraux en acier. Les enjeux reliés à la productivité des activités de construction ont été synchronisés afinde compléter l’installation sur le site du projet en seulement 56 jours de travail.

Mots‐clés : simulation, visualisation, bâtiment modulaire, échéancier.

[Traduit par la Rédaction]

1. Introduction

Modular buildings are built through a more efficient and

cost-effective engineering method that can deliver market re-quirements for increased construction speed, improved qual-ity, and rapid return of investment. Furthermore, there is a

Received 15 October 2011. Revision accepted 6 June 2012. Published at www.nrcresearchpress.com/cjce on 17 August 2012.

M. Moghadam and M. Al-Hussein. Construction Engineering and Management, Hole School of Construction, Department of Civil andEnvironmental Engineering, University of Alberta, Edmonton, AB T6G 2W2, Canada.S. Al-Jibouri. Construction Management and Engineering, Department of Construction Management and Engineering, University ofTwente, 7500 AE Enschede, The Netherlands.A. Telyas. Kullman Buildings Corporation, One Kullman Corporate Campus, Lebanon, NJ 08833-2163, USA.

Corresponding author: Mansooreh Moghadam (e-mail: mansoore@ualberta.ca).1This paper is one of a selection of papers in this Special Issue on Construction Engineering and Management.

1053

Can. J. Civ. Eng. 39: 1053–1061 (2012) doi:10.1139/L2012-077 Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

noticeable motivation for owners and developers to use mod-ular building in affordable housing where speed of construc-tion is related to the economy of production scale. Hencethere is a significant need to evaluate the planning processfrom factory fabrication to construction and erection phasesinvolving crane operation. There is a significant amount ofresearch for optimizing mobile crane selection (Al-Husseinet al. 2005; Wu et al. 2011) and mobile crane operation thatleads to modular building field operation optimization (Ole-arczyk et al. 2009; Hasan et al. 2010; Hermann et al. 2010).However, little research has been conducted in the area oftower crane utilization, which is required in complex con-struction projects to lift modules in high-rise modular build-ings (Tam et al. 2001; Hasan and Al-Hussein 2010; Huang etal. 2011).In general, once the construction stage of a modular build-

ing has started, changes to the design and construction planare costly. Existing planning and scheduling tools are de-signed to support mainly the current practice of onsite stick-built construction, without considering modular building’sspecific needs. The application of such practice for modularbuilding could jeopardize a successful project outcome dueto the lack of a detailed construction process evaluation sys-tem. For modular construction, it is essential to create a real-istic simulation model to evaluate construction sequencesduring the planning phase and control the installation processonsite. Compared with typical construction projects, modularbuildings are characterized by repetitive processes that re-quire complex analysis. Additionally, the need to incorporatediverse information that is required to evaluate the design,cost, schedule, resources, quality, and safety complicates theprocedure even further. This work proposes modeling usingsimulation to help analyse the process and reveal possibleconflicts and errors for complicated processes.Simulation is a computer-based tool that models, in scaled

format, a real life system or process and facilitates the exami-nation and evaluation of different scenarios to choose thebest possible option. Before the introduction of CYCLONE(CYCLic Operation NEtworks) by Halpin in 1973, simula-tion was not common in construction (Huang and Halpin1994). CYCLONE uses graphical representations to modelvarious tasks in the construction process. The tool’s outputtypically consists of charts and diagrams. Simulation waslimited to research applications before AbouRizk and Hajjar(1998) presented a framework that customized simulation forindustry. They used Special Purpose Simulation (SPS) as atool to build systems for specific purposes. Later they im-proved their approach and developed a construction simula-tion system called Simphony, which provides a set ofpredefined elements representing construction requirements(Hajjar and AbouRizk 2002). Specifically in the area of mod-ular building, extensive research has been done to simulateconstruction operations. Mohsen et al. (2008) used Simphonyto create a simulation model that analyses the constructionprocess and predicts productivity and project duration. Themodel analysed scenarios and considered possible errors,which was effective in crane and crew utilization as well aspreventing delays. Hasan et al. (2010) simulated the opera-tion of a tower crane using an integrated dynamic modelingsystem and lean concepts. The result proved that the model

provides schedule satisfaction, reduces required resources,and decreases the construction cost.Although simulation tools are considered to be efficient in

evaluating construction processes, the project managementteam often views them as a “black box” that can only beunderstood by a highly skilled manager. The gap betweenthe specialist and management team leads to misunderstand-ing regarding simulation results and may result in incorrectinterpretations (Al-Hussein et al. 2006). The project manage-ment team is unwilling to make decisions based on currentsimulation outputs, and statistically based charts and dia-grams, because they do not provide adequate information re-lated to construction process requirements (Kamat andMartinez 2000). Visualization is a more popular constructiontechnique since it fosters better understanding of the con-struction process and performance. However, to be effectivefor decision making purposes, such a model is required to belinked to project information. As such, Kamat and Martinez(2001) developed a general purpose 3D text file driven visu-alization system named dynamic construction visualizer(DCV) that allows visualization of construction operationsand enables construction planners to obtain more realisticfeedback from simulation analysis. Huang et al. (2007) andLi et al. (2008) created a construction virtual prototyping(CVP) system that integrates visualization and simulation.The system can generate and modify 3D models of buildingcomponents, construction equipment and labour, and tempo-rary works. Although collecting input data for this VP systemis a time consuming process, it allows project teams to checkconstructability through visualized 3D models of projects. Liet al. (2009) used VP to optimize construction planningschedules by analyzing resource allocation and planning withintegrated construction models, resource models, constructionplanning schedules and site-layout plans. Staub-French et al.(2008) described a two-way relationship between 3D CADsoftware and software implementation of linear planning anddescribed the consistency of product representation in CADand scheduling models. They improved project schedulingby modifying construction sequences through a 4D CADmodel and validated the project completeness by connectingthe model to 3D objects and activities.In a different manner, Golparvar-Fard and Peña-Mora

(2007); Golparvar-Fard et al. (2009a, 2009b); and Kamat etal. (2011) developed visualization of construction sites viathe four-dimensional (4D) geometric model. They developedan automated vision-based approach that monitors as-builtprogress and compares it to the as-planned construction prog-ress. The system matches as-planned and as-built views witha selection of features where the as-built photographs are en-hanced and augmented with the 4D as-planned model.The objective of this paper is to introduce an integrated

method referred to as a post simulation visualization (PSV)model for modular building construction with emphasis onschedule efficiency. The proposed PSV model is developedusing the advantages of both simulation and visualization,where critical information such as the 3D model, time con-straints, and resource demand are incorporated into the sys-tem. The proposed model has been tested on the modularconstruction of a 34-storey building that is planned to beconstructed in Brooklyn, N.Y., USA.

1054 Can. J. Civ. Eng. Vol. 39, 2012

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

2. Proposed PSV methodologyThe proposed methodology follows the concept repre-

sented in Fig. 1. The purpose of the model is to integrate vis-ualization with project information to create a simulation ofproject operations. Input data including project managementrequirements and specifications; activities’ logic and dura-tion; and resource demands, which consist of equipment, ma-terial, and labour, are all collected in the first step. The mainpart of the system is focused on linking the information. Theprocess performs satisfying the project criteria and limita-tions, including modules’ specifications such as weight andsize; crane capacity for lifting purposes; trailer and transpor-tation limits from storage area to lifting zone; deadlines andtime constraints; structural constraints, such as the locationof shear walls and stairs; law and regulations; and targetedcost.The main process begins with the creation of a project in-

formation database. The activities are then put together in adefined sequence and duration to create the project schedule.Resources, such as material, equipment, and labour, requiredby individual activities are added to the database. The infor-mation gained from the above steps is used to create a simu-lation model in Simphony.Net 3.5 © environment. Thesimulation output in the form of an ASCII file, a binary textfile imported from Simphony, is then linked to 3D animationsemi-automatically. Within the 3D animation environment(3D Studio Max) the high level simulation model is trans-formed to a micro level representation in frames/second. Todevelop the PSV model, 3D Studio Max’s scripting language,MaxScripts, is used. As inputs of the PSV model, two sour-ces are required: the 3D model library of PSV and the simu-lation model output that stores the spatial configuration ofthe construction process along with performance time. ThePSV model imports 3D models from the 3D library includingmodels of the tower crane, modules, resources, and the 3Dsite, and assembles them in 3D Studio Max. Then the 3Danimation engine uses the data from the simulation outputfile to create the key frames. The model’s output is a meas-urement of the productivity of labour and the main equip-ment such as crane, project schedule analysis, safety andquality control, and evaluation of potential scenarios for theconstruction operations.

3. Post simulation visualization (PSV)implementationThe proposed PSV model depicts the simulation of the

construction process in detail, producing and displaying proj-ect information such as schedule and production rate simulta-neously for evaluation purposes, whereas the proposedsystem has the capacity to be linked to all possible projectinformation. The model needs to be flexible to deal withchanges in project scope and to present a complex construc-tion process in a simple way. The proposed PSV model hasbeen tested on a construction plan for a challenging 34-storeymodular project. The prospective project is to be constructedin Brooklyn, N.Y., USA and will be the highest modularbuilding ever built. Visualization of the simulated processproved to be an effective tool in communicating the valueand simplicity of the minute-by-minute schedule. The simula-tion result comparison between the initial model before ap-

plying PSV and the final version after running PSV severaltimes, indicated significant improvements in the constructionschedule, productivity, levelling of resources, and reductionsin idle time. Furthermore, a comparison was conducted be-tween various construction scenarios. Moreover, the projectmanagement team confirmed the considerable impact of thePSV model on decreasing the activities’ duration, eliminatingerrors and reworks, and identifying the best potential con-struction scenarios using visualization of the simulated proj-ect during the planning phase.Figure 2 shows a snapshot from the model’s output for the

case study that presents the module lifting process. The shut-tle truck transports the modules from the storage area to thelifting zone and the crane lifts the modules to place them intheir final locations. The schedule runs at the same windowto display the most probable duration of each task. Theschedule was produced using the critical path method (CPM)technique then a simulation model was created to evaluatethe CPM and program evaluation and review technique(PERT) schedule based on predetermined optimistic, pessi-mistic, and most likely project durations. Clocks on the left-hand side of the snapshot present the duration of criticalequipment including trucks and the crane, and critical crews,including erection and welding. These clocks track workingtime, work flow, and idleness duration for each aforemen-tioned instance separately and provide an evaluation for pro-ductivity analysis.This research contributed to (1) assigning production track-

ing clocks to each critical resource (crane, truck, hookingcrew, scaffolding crew, welders, alignment crew, and erectioncrew) as sampled in Fig. 2; and (2) changing the concept ofconstruction planning to match the factory requirements. Thework moves to the location where the station is assigned toeach worker crew (for example, the erection crew stays onthe platform waiting for the crane, while the crane operatorwaits for the truck which is transporting the module, thenthe crew hook the module to the crane which lifts the moduleand moves it to the installation zone). Similarly, stations areassigned to the alignment crew, welding crew, and cappingexterior finishing crew and work comes to their stations.Scheduling for the time, space, and movement of the workersis not possible without utilization of lean, simulation, andvisualization.

4. Case studyAffordable housing has been a long sought after, yet sel-

dom realized, goal. The very subsidy designed to assist mod-erate income renters accelerates housing inflation, and reducesaffordability across every income range, and thus the elusivegoal remains unmet. In an attempt to create affordable hous-ing, in effect, to increase the supply of housing stock in oneof the world’s costliest markets, the development team leadingthe Atlantic Yards Project of Forest City Ratner aimed toidentify innovation which could affect the supply of affordablehousing. The team realized that incremental improvementwould not suffice and that a substantial re-arrangement of thestatus quo was necessary. Upon an intensive review of keycost drivers and a search for alternative construction meth-ods, Kullman Buildings Corp. (KBC), a leader in industrial-ized construction, was selected to study the technical and

Moghadam et al. 1055

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

Fig. 1. Proposed methodology of post simulation visualization model.

Fig. 2. Output of PSV model on the case study presenting module lifting operation.

1056 Can. J. Civ. Eng. Vol. 39, 2012

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

economic feasibility of fabricating the building at its fac-tory. Tall modular buildings have been built as high as 24-Storeys in Europe to great success, but few have beenerected in North America, and certainly none at the pro-posed height. Kullman assembled a world-class team of de-signers, researchers and structural engineers, including aresearch team from the University of Alberta, to explorethis mandate and prove its technical and economical viabil-ity. As noted, the stated mandate is quite ambitious inbudget and scope. A 34-storey tower to be built entirely ofprefabricated modules has never been accomplished any-where in the world. Ultimately, a workable design solutionto this challenge was selected that relied on a factory pro-duction line to produce 950 modules requiring very littlefield work other than assembly onsite. Adopting such a sol-ution, the project would take less than a year to completeand save substantial cost. The 34-storey modular tower canbe assembled, lifted into position, stacked and connected,and built in such a short duration only by tracking thetime, resource, and equipment utilization in detail. ThePSV model provides a visualized minute-by-minute scheduleincluding a separate production clock for all the resourcesthat are automated and can be updated for each simulationrun.This project faced a number of challenges due to its loca-

tion in one of the busiest streets of Brooklyn, N.Y., USA,and its budget and time of construction which are cut tohalf. The trucks are expected to deliver the modules fromthe factory to the storage area near the construction site andfrom the storage area to the lifting area onsite where thetower crane is planned to lift two modules per hour. To suc-ceed in this plan, a pull schedule was utilized which added tothe challenges of resource utilization in the supply chain. An-other challenge arose after adding the scaffolding activities tothe project schedule which increased the duration of instal-ling 950 modules to 83 working days from 56 working days.However, changes to the process eventually reduced the over-all time required for the project completion. The sequence ofactivities, including module delivery by trucks and lifting bycrane; lifting of modules, shear walls, and scaffolding inter-mittently; and arrangement of module sequence based onfloor layout, created a dynamic schedule which was added tothe long list of challenges in this project.

4.1. Building simulation modelTo simulate the construction operation, a model was devel-

oped in the Simphony.Net 3.5 © environment as shown inFig. 3. The main focus of creating the simulation model wascrane activity, which is the most critical and schedule-driventask in modular building. In this model, crane activity wasdivided into proper work packages to create a non-stop workflow and efficient crane utilization. The simulation helpedidentify potential errors in the field operation and revealedpossible conflicts. Also, the effect of different module instal-lation sequence scenarios was analysed by the simulationmodel. The visualized model provided different scenarios,for which the simulation model calculated the correspondingimpact on the total project duration through defining prioritycriteria. Once a module is chosen to be assembled, it capturesthe resources and obtains a higher priority. Therefore, the pri-ority is defined as a function of the modules’ position in the

specific scenario. The priority for the module in the queueand for the module in progress (the module that has alreadybeen captured by the crane) is calculated satisfying eq. [1]and eq. [2], respectively (Mohamed et al. 2007).

½1� P ¼ F þ abs�min ðFÞ

�

½2� P ¼ ðES� TDÞ þmax ðPÞwhere P is the module priority; F is the free float for themodule = PS – ES – D; PS is the planned ship date; ES isthe early start date; D is the duration; and TD is today date.The simulation model output was total project duration.

Considering the probable duration distribution for each indi-vidual task enables the evaluation of project duration andlevel of confidence. The Simphony model output showedthat the project could be finished in 26 616.57 working mi-nutes (56 eight hour working days) with 95% confidence.The project duration result from the simulation model isshown in Table 1. The optimistic, pessimistic, and mostlikely duration for each task were estimated taking into ac-count the resource operation (mainly crane working hours).Task durations were determined based on the crane manufac-turer’s specifications to determine speed of operation andpast similar operation experiences (please see Olearczyk2010). The probability density function (PDF) and cumula-tive density function (CDF) charts for the model, as shownin Fig. 4, present the project duration distribution and levelof confidence. The PDF chart describes the likelihood forthe project duration, while the CDF presents the probabilitythat the project duration at each point has that value or less.This result is used to evaluate and aid the decision maker inarriving at a near optimum solution with reference to an opti-mization model of the tower crane location (which is outsideof this paper's scope). The optimization model evaluated thetechnically feasible location of the tower crane and found theshortest total travel distance for all the modules from thepicking and lifting location to the un-lifting point. Table 2summarizes the total project duration statistics. This mini-mum travel distance was synchronized in 8-h working daysand onsite installation of the whole project was achieved inonly 56 working days which is the minimum total project du-ration among all technically feasible solutions.

4.2. Schedule and resource allocationModule onsite installation requires a detailed construction

schedule. A breakdown of activities and required crews, aswell as a complete lifting and assembly schedule for all mod-ules was created. This analysis is based on the installation of950 structural steel modules for a 34-storey building. Thesimulation model showed that, on average, 16 modules perday could be installed from floors 2 to 20, 15 modules perday from floors 21 to 29, and 14 modules per day for the re-maining floors. This analysis showed that the bottleneck ac-tivity is the welding and bolting of modules at their finallocation. Outputs from the PSV model indicated that moduleerection is a function of crane capacity and utilization rate,and that typically 12 ft × 30 ft (3.6 m × 9 m) modules canbe erected at a rate of 2 per hour or 16 modules per day. Set-ting crews would have 8 h per day to bolt the modules toeach other. Mate-line connections include internally tying up

Moghadam et al. 1057

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

each module electrically and mechanically to the buildingservices as well as installing drywall in the mating areas. If a4-person crew is assigned to finish a module per day, thencompleting 16 modules a day requires 16 crews, or 64 peopleworking for approximately 3 months to finish the project asshown in Table 3. A typical schedule for onsite module in-stallation is shown Table 4.

The installation sequence per module was broken downinto 16 tasks, making use of 8 construction crews. Based onthe analysis of the model output, it was noticed that to keepthe installation sequence flowing, a minimum of two trailersare required onsite (loaded with ready modules) and (in thestaging area) the delivered modules must be moved to thelifting zone. Critical modules (modules next to shear walls,

Fig. 3. Part of simulation model created in Simphony.

Table 1. Output result from the Simphony model.

Name Graph Min Mean Max 5% 95%Project total duration(minute)

24 925.15 26 064.14 27 189.02 25 511.76 26 616.57

Fig. 4. PDF (a) and CDF (b) charts showing the simulation result.

1058 Can. J. Civ. Eng. Vol. 39, 2012

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

modules next to stairs and elevator shafts, and final cornermodules for every floor) will require 50% additional time forbolting and welding. Modules will be unhooked at their finallocation once they are aligned. Welding and bolting will con-tinue immediately thereafter. The tower crane will be tied-into the structure whenever necessary, including when climbingthe tower crane (adding mast sections) and installing bridgesbetween the structure and the tower crane. An estimated 56working days are required to complete the module onsite in-stallation, not including the time required for mobilization,assembly and disassembly of the tower crane, constructionsite set-up, and interior construction work as analysed previ-ously. The tower crane is the key resource in the assemblyprocess (Mohsen et al. 2008). The scheduling objective func-tion, as calculated in eq. [3], minimizes the tower crane idletime.

½3� f ¼ minX

ðW þ U þ SbÞh i

where W is the module wrapping time; U is the module un-wrapping time; and Sb is the spreader bars removal time.To reduce the tower crane idle time, unwrapping and weld-

ing operations need to be considered satisfying eqs. [4] and[5]

½4� U > SLþ Uhþ Sb

where SL is the securing and lining time and Uh is theunhooking time

½5� Wd > H þ Crþ U þ Sb

where H is the hooking time; Wd is the welding time; and Cris the crane operation time.

4.3. Potential scenarios evaluationThe developed PSV model is capable of producing a vari-

ety of outputs for analysis purposes. For instance, options forconnecting modules in the lower floor with one in the upperfloor were tested. Based on the evaluation of the PSV model,the best option was an oversized pop-rivet as shown inFig. 5a. Its installation is expected to be smooth and fastwith two workers making the connections. Essentially, thecorners of four modules are brought together and a plate isinstalled with 8 rivets. The rivets can also be used for liftingpurposes. To increase project productivity, the use of scaf-folding was suggested to reduce total project duration and la-bour idle time. As shown in Fig. 5b, scaffolding is attachedto the module to provide workers with access to the module’sfaçade. As the next module is lifted, the work on the priormodule is finished and the scaffolding is carried by the craneto the next module. Safety was controlled through using thePSV model for the entire construction operation. For exam-ple, as shown in Fig. 5c, the use of a ladder was not safeenough for labourers to hook or un-hook modules. Therefore,a replacement system was suggested and tested. As shownthrough the model, the system also increased labour workingspeed.Experts conducted scenario-based analysis through meet-

ings with different trades to evaluate potential scenarios. Forinstance, for hooking and unhooking modules, several scenar-ios were proposed and investigated based on criteria such asequipment efficiency and activity duration. The experts, in-cluding the erection crew and management at Kullman, ac-cepted the rivet (Fig. 5a) as the most effective option.Although this tool is made at an extra cost, it is still a morecost effective option compared to the other scenarios. The pro-posed result for the on-site construction process was reviewedand validated by the team involved in on-site work and theproject management team. All changes to the construction

Table 2. Summary statistics for project total duration.

Statistics PercentileMinimum 24925.1514 Minimum output value 5% 25511.75792Maximum 27189.01811 Maximum output value 10% 25639.69982Mean 26064.14241 µ = Sx/n 15% 25730.15543Std Dev 334.4575819 s = sqrt[s2] 20% 25791.67645Variance 111861.8741 s2 = S (xi - µ)2/(n - 1) 25% 25848.3896Skewness 0.01127267 S [(X-µ)/s]3 30% 25887.62913Kurtosis 3.094436921 S [(X-µ)/s]4 35% 25933.89473Median 26063.79792 The middle number in ascending order 40% 25980.92413Mode 26102.30269 The most repeated number 45% 26020.88949Left X 25511.75792 Left boundary 50% 26063.79792Left P 5% Left probability value 55% 26100.30683Right X 26616.57415 Right boundary 60% 26139.21917Right P 95% Right probability value 65% 26180.37703Diff X 1104.816234 The difference of X value to cover 90% of prob-

ability distribution70% 26226.13223

Diff P 90% 75% 26272.07795Chi-Sq 22.1340 The goodness of fit for normal distribution 80% 26348.63812Regr 0.227 Regression coefficient 85% 26425.09549# Error 0 90% 26496.64838Run time 00:00:05 Simulation duration 95% 26616.57415

Table 3. Mate-line crew specification.

Crew size 4.0 PersonsCrew capacity 1.0 modules/dayCrew needed 16.0 CrewsPersons needed 64.0 PersonsWeeks needed 11.8 Weeks

Moghadam et al. 1059

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

process were incorporated into the developed model. As aresult, the final analysis, including the proposed schedule,process evaluation, and resource allocation, was documentedby Kullman’s team and formed the basis of the constructionfor the project. It is important to mention that this project ison-hold due to legal challenges between developers andother parties.

5. Conclusion

There is a noticeable motivation for owners and developersto use modular building in affordable housing, where speedof construction is related to economy of production scale.Hence there is a significant need to evaluate the planningprocess of modular construction from factory fabrication to

Table 4. Typical construction schedule for onsite module installation (Module No. 1 Level floor 2).

Activity description Location Dependency

Time (min) Cumulative(average min)Pessimistic Average Optimistic

Trailer at loading zone with module Lifting zone 10 8 6 0Unwrap module Lifting zone 1FS 10 8 6 0Hook-up module Lifting zone 2FS 12 10 8 10Lifting Inspection Lifting zone 3FS 7 5 4 15Load module – hoist up Air 4FS 2.81 0.81 0.81 15.81Swing module Air 5FS 0.7 0.5 0.3 16.31Fold tarps Lifting zone 3FS 7 5 3 16.31Return trailer to loading zone Drop zone 5SS 10 8 6 16.31Hoist down module Air 6FS 0.7 0.5 0.3 16.81Align module and survey finallocation

Building 9FS 22 19 16 35.81

Bolt and weld module to structure Building 10FS 23 20 17 55.81Fireproofing partition walls Building 10FS 15 13 11 55.81Inspect installation Building 11FS,12FS 7 6 5 55.81Unhook module Building 10FS 2 1 1 55.81Swing back to loading zone Air 14FS 1.5 1 0.5 55.81Hoist down main line Air 15FS 1.24 0.24 0.30 55.81Module installation cycle time 68.21 55.81 46.41 55.81

Note: Module weight: 27 196 lbs (12 ton).

Fig. 5. Evaluation of construction operation scenarios: (a) proposed pop-rivet system for connections and lifting of modules, (b) use of scaf-folding to increase project productivity, and (c) proposed safety regulation led to a higher labours’ working pace.

1060 Can. J. Civ. Eng. Vol. 39, 2012

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.

construction and erection phases. In this paper, modeling us-ing simulation supported by visualization was proposed toanalyze the process and reveal possible conflicts and errors.An integrated method, referred to here as a post simulationvisualization (PSV) model for modular building construction,was introduced. The proposed PSV model is developed as atechnique using the advantages of both simulation and visual-ization. The model was tested on the construction of a pro-spective 34-storey modular building to be constructed inNew York. Visualization of the simulated process proved tobe an effective tool in communicating the value and simplic-ity of a minute-by-minute schedule. The focus of this projectwas improving the construction schedule and productivity,leveling resources, decreasing idleness time, and comparingvarious construction scenarios by applying the PSV modelduring the project planning phase. Issues related to productiv-ity of construction activities were synchronized to achieve on-site installation of the whole project in only 56 working days.

AcknowledgementsThe authors are sincerely grateful to Kullman Buildings

Corporation for providing required data for this research.Any opinions, findings, and conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reflect the views of teams involved in the AtlanticYards Project.

ReferencesAbouRizk, S.M., and Hajjar, D. 1998. A framework for applying

simulation in construction. Canadian Journal of Civil Engineering,25(3): 604–617. doi:10.1139/l97-123.

Al-Hussein, M., Alkass, S., and Moselhi, O. 2005. Optimizationalgorithm for selection and onsite location of mobile cranes.Journal of Construction Engineering and Management, 131(5):579–590. doi:10.1061/(ASCE)0733-9364(2005)131:5(579).

Al-Hussein, M., Niaz, M.A., Yu, H., and Kim, H. 2006. Integrating3D visualization and simulation for tower crane operations onconstruction sites. Automation in Construction, 15(5): 554–562.doi:10.1016/j.autcon.2005.07.007.

Golparvar-Fard, M., and Peña-Mora, F. 2007. Application ofVisualization Techniques for Construction Progress Monitoring.Computing in Civil Engineering, Pittsburgh, PA., USA. pp. 216–223.

Golparvar-Fard, M., Peña-Mora, F., and Savarese, S. 2009a. D4AR- A4-Dimensional augmented reality model for automating construc-tion progress data collection, processing and communication.Journal of Information Technology in Construction, 14: 129–153.

Golparvar-Fard, M., Peña-Mora, F., Arboleda, C.A., and Lee, S.H.2009b. Visualization of construction progress monitoring with 4Dsimulation model overlaid on time-lapsed photographs. Journal ofComputing in Civil Engineering, 23(6): 391–404. doi:10.1061/(ASCE)0887-3801(2009)23:6(391).

Hajjar, D., and AbouRizk, S.M. 2002. Unified modeling methodol-ogy for construction simulation. Journal of Construction Engi-neering and Management, 128(2): 174–185. doi:10.1061/(ASCE)0733-9364(2002)128:2(174).

Hasan, S., and Al-Hussein, M. 2010. Advanced simulation of towercrane operation utilizing system dynamics modeling and Leanprinciples, Winter Simulation Conference, Baltimore, MD, USA.pp. 3262–3271.

Hasan, S., Al-Hussein, M., Hermann, U.H., and Safouhi, H. 2010.Interactive and dynamic integrated module for mobile cranes

supporting system design. Journal of Construction Engineeringand Management, 136(2): 179–186. doi:10.1061/(ASCE)CO.1943-7862.0000121.

Hermann, U.R., Hendi, A., Olearczyk, J., and Al-Hussein, M. 2010.An Integrated System to Select, Position, and Simulate MobileCranes for Complex Industrial Projects, Construction ResearchCongress, Banff, AB, Canada. pp. 267–276.

Huang, R.-Y., and Halpin, D.W. 1994. Visual construction operationsimulation: the DISCO approach. Microcomputers in Civil En-gineering, 9(3): 175–184. doi:10.1111/j.1467-8667.1994.tb00371.x.

Huang, T., Kong, C.W., Guo, H.L., Baldwin, A., and Li, H. 2007. Avirtual prototyping system for simulating construction processes.Automation in Construction, 16(5): 576–585. doi:10.1016/j.autcon.2006.09.007.

Huang, C., Wong, C.K., and Tam, C.M. 2011. Optimization of towercrane and material supply locations in a high-rise building site bymixed-integer linear programming. Automation in Construction,20(5): 571–580. doi:10.1016/j.autcon.2010.11.023.

Kamat, V.R., and Martinez, J.C. 2000. 3D Visualization of SimulatedConstruction Operations. Winter Simulation Conference, Orlando,FL., USA. Vol. 2. pp. 1933–1937.

Kamat, V.R., and Martinez, J.C. 2001. Visualizing SimulatedConstruction Operations in 3D. Journal of Computing in CivilEngineering, 15(4): 329–337. doi:10.1061/(ASCE)0887-3801(2001)15:4(329).

Kamat, V.R., Martinez, J.C., Fischer, M., Golparvar-Fard, M., Peña-Mora, F., and Savarese, S. 2011. Research in visualizationtechniques for field construction. Journal of ConstructionEngineering Engineering and Management, 137(10): 853–862.doi:10.1061/(ASCE)CO.1943-7862.0000262.

Li, H., Huang, T., Kong, C.W., Guo, H.L., Baldwin, A., Chan, N.,and Wong, J. 2008. Integrating design and construction throughvirtual prototyping. Automation in Construction, 17(8): 915–922.doi:10.1016/j.autcon.2008.02.016.

Li, H., Chan, N., Huang, T., Guo, H.L., Lu, W., and Skitmore, M.2009. Optimizing construction planning schedules by virtualprototyping enabled resource analysis. Automation in Construc-tion, 18(7): 912–918. doi:10.1016/j.autcon.2009.04.002.

Mohamed, Y., Borrego, D., Francisco, L., Al-Hussein, M., AbouRizk,S., and Hermann, U. 2007. Simulation-based scheduling ofmodule assembly yards: case study. Engineering, Construction,and Architectural Management, 14(3): 293–311. doi:10.1108/09699980710744926.

Mohsen, O.M., Knytl, P.J., Abdulaal, B., Olearczyk, J., and Al-Hussein, M. 2008. Simulation of modular building construction,Winter Simulation Conference, Austin, TX, USA. pp. 2471–2478.

Olearczyk, J. 2010. Crane lifting operation planning and lifted objectspatial trajectory analysis, Ph.D. thesis, Department of Civil andEnvironmental Engineering, University of Alberta, Edmonton,AB, Canada.

Olearczyk, J., Al-Hussein, M., Bouferguene, A., and Telyas, A. 2009.Virtual Construction Automation for Modular Assembly Opera-tions, Construction Research Congress, Seattle, WA, pp. 406–415.

Staub-French, S., Russell, A.D., and Tran, N. 2008. Linear schedulingand 4D visualization. Journal of Computing in Civil Engineering,22(3): 192–205.

Tam, C.M., Tong, T.K.L., and Chan, W.K.W. 2001. Geneticalgorithm for optimizing supply locations around tower crane.Journal of Construction Engineering and Management, 127(4):315–321. doi:10.1061/(ASCE)0733-9364(2001)127:4(315).

Wu, D., Lin, Y., Wang, X., Wang, X., and Gao, S. 2011. Algorithm ofcrane selection for heavy lifts. Journal of Computing in CivilEngineering, 25(1): 57–65. doi:10.1061/(ASCE)CP.1943-5487.0000065.

Moghadam et al. 1061

Published by NRC Research Press

Can

. J. C

iv. E

ng. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

ON

CO

RD

IA U

NIV

on

11/2

6/14

For

pers

onal

use

onl

y.