BIM-Integrated Construction Operation Simulation for ...

sustainability Article

BIM-Integrated Construction Operation Simulation for Just-In-Time Production Management WoonSeong Jeong 1 , Soowon Chang 2 , JeongWook Son 1, * and June-Seong Yi 1 1 2

*

Department of Architectural Engineering, Ewha Womans University, 52, Ewhayeodae-Gil, Seodaemun-Gu, Seoul 03760, Korea; [email protected] (W.J.); [email protected] (J.-S.Y.) School of Building Construction, Georgia Institute of Technology North Avenue, Atlanta, GA 30332, USA; [email protected] Correspondence: [email protected]; Tel.: +82-2-3277-3577

Academic Editors: Vivian W. Y. Tam and Marc A. Rosen Received: 11 October 2016; Accepted: 27 October 2016; Published: 29 October 2016

Abstract: Traditional construction planning, which depends on historical data and heuristic modification, prevents the integration of managerial details such as productivity dynamics. Specifically, the distance between planning and execution brings cost overruns and duration extensions. To minimize variations, this research presents a Building Information Modeling (BIM)-integrated simulation framework for predicting productivity dynamics at the construction planning phase. To develop this framework, we examined critical factors affecting productivity at the operational level, and then forecast the productivity dynamics. The resulting plan includes specific commands for retrieving the required information from BIM and executing operation simulations. It consists of the following steps: (1) preparing a BIM model to produce input data; (2) composing a construction simulation at the operational level; and (3) obtaining productivity dynamics from the BIM-integrated simulation. To validate our framework, we applied it to a structural steel model; this was due to the significance of steel erections. By integrating BIM with construction operation simulations, we were able to create reliable construction plans that adapted to project changes. Our results show that the developed framework facilitates the reliable prediction of productivity dynamics, and can contribute to improved schedule reliability, optimized resource allocation, cost savings associated with buffers, and reduced material waste. Keywords: productivity dynamics; building information modeling; computer simulation; Just-In-Time; lean construction; data reuse

1. Introduction Due to the increasing size and complexity of construction projects, traditional construction planning is no longer sufficient for producing workable plans that incorporate all necessary project details, such as design complexities, learning curves, and the coordination process. Such traditional construction methods require construction managers to utilize data from previous projects and make heuristic adjustments to establish on-site construction plans. However, with the radical shifts now being seen, construction planning that refers to previous projects’ data cannot ensure the predicted level of productivity. Since construction projects have become more complex [1], managers’ heuristics cannot encompass all of the necessary managerial and operational details. On-site work now suffers from constant modifications and changes to project conditions [2]. Such inadequate planning can result in delays [3] and cost overruns stemming from on-site operational problems. This variability between planned and actual performance results in managerial inefficiency and, ultimately, lower-quality outcomes. Such differences lead to the supply of materials not coinciding with demand on the construction site; construction managers must then wrestle with an excess or lack Sustainability 2016, 8, 1106; doi:10.3390/su8111106

www.mdpi.com/journal/sustainability

Sustainability 2016, 8, 1106

2 of 25

of materials. If the actual performance is greater than expected, materials become scarce and labor and equipment are wasted. On the other hand, if the actual performance is lower than expected, materials become excessive; construction costs then increase due to interest accrual and inventory management. Thus, reliable planning is extremely important to managerial efficiency and waste reduction. However, since construction projects are inherently dynamic and complex, the work productivity on site can vary daily, according to the type and number of mitigating factors [4,5]. Productivity dynamics can proliferate such variations between planned and actual production, resulting in an exacerbation of on-site problems. Since productivity on construction sites is dynamic, it is challenging to develop sufficiently reliable construction plans. Each construction project is unique and complex. First, projects include numerous risks and uncertainty [6]. For instance, developments differ in location, design, level of skilled labor required, and team composition. The newness of the construction environment also adds unknowns to the execution process, so that planning based on historical data cannot guarantee the expected performance. This persistent newness results in frequent modifications to the planned schedule. Also, a variety of operational and managerial factors influence productivity. Construction work involves multiple processes and the complex interactions of a wide variety of components that are connected by nonlinear relationships [5]. Moreover, since a construction project involves numerous stakeholders (including owners, designers, contractors, and the government), coordination of all of the relevant participants can be challenging [7]. Finally, construction projects operate under numerous constraints, in dynamic environments, and require the coordination of multiple tasks [8]. Such complexity adds unpredictability, and prudent planning is required. Reliable predictions of productivity at the operational level minimize uncertainties, facilitate efficiency, and decrease waste in terms of time, cost, and materials. However, “construction companies still lack the ability to properly plan, estimate, and execute projects in a consistent, efficient, and reliable manner” [2]. Existing planning and control techniques are insufficient for predicting reliable and adequate on-site performance. However, computer simulations can be used to improve a plan’s reliability, by explicitly incorporating performance factors and mutual causal relationships. Such simulations can be utilized to quantify and validate the efficiency of the construction process [6]. In addition, simulation models represent the: (1) overall logic of the multitude of activities required to construct a building; (2) resources involved in performing the work (e.g., labor, equipment, management, etc.); and (3) environment in which the project is being built (e.g., site conditions, labor pool, market situation, etc.) [9]. Using computer simulation tools allows construction managers to build consistent plans that consider critical productivity factors. Furthermore, computer simulations automate the planning process, allowing for the integration of Building Information Models (BIM) that contain all of the information necessary for development. BIM design applications are more than just design tools; most BIM design applications also interface with other applications, allowing for energy analyses, cost estimations, and so on [10]. With the increasing amount of information available and BIM’s improved process annotations, information visualization has become central to the overall construction process [10]. In this context, integrating construction operation simulations with BIM would facilitate time and cost efficiency, and generally streamline the process. In this regard, our research developed a BIM-integrated simulation framework for predicting reliable productivity dynamics by considering the factors that affect productivity at the operational level. Our BIM-integrated planning framework alleviates the difficulties that currently plague reliable construction planning, and allows for managerial efficiency as well as technical advancement. Above all, since BIM is a digital representation of the functional and physical characteristics of a building [11], it allows for the early incorporation of the unique characteristics of a building into the planning process. Also, since simulation models represent the comprehensive production process, including the dynamic and complex interaction of sub-processes [12], our framework allows us to synthetically consider diverse factors and repeatedly run construction project in a virtual space,


3 of 25

without risk. In the BIM-integrated simulation framework suggested here, simulations were fed data on the building elements that had already been generated by the BIM tool. The complete simulation performed the construction operation in a virtual world, and generated accurate forecasts of productivity dynamics that were more reliable than those developed by traditional methods. 2. Literature Review This chapter reviews current studies on the use of computer simulation in construction, lean construction practices, and BIM applications used in our BIM-integrated simulation framework for Just-In-Time production planning. First of all, whereas computer simulation techniques can now be combined with advanced tools to analyze construction projects, difficulties with modeling simulations and unreliable data usage continue to inhibit the application of simulations in the construction management field. Additionally, while the construction industry has improved in many ways ranging from management practices to technical developments, to apply lean principles, lean construction should be emphasized (including all parties along the supply chain). While BIM’s benefits to the construction industry are clear, its utilization in the construction management domain remains narrow. 2.1. Computer Simulation in Construction Increasingly, simulation models are being used as problem-solving and decision-making tools in many industries, especially manufacturing [13]. Simulation has also long been an established method for analyzing both production and logistics [14]. Particularly, the manufacturing industry has successfully implemented simulations to improve production processes [15]. The popularity of simulation stems from its ability to make models of complex systems [16]. Simulations offer realistic representations of the interactions among a variety of systems’ diverse components [16]. Despite these advantages, in the construction industry, the use of simulation for planning purposes has been largely limited to academic research. This is because simulation modeling is a time-consuming and error-prone process [12,17,18]. Certain researchers have argued that 4D simulations should be considered appropriate for use in the construction industry. However, 4D simulations only allow for the visualization of construction schedules [19]. In this research, we used 4D models instead of 4D simulations; references to “simulations” should be understood as including discrete-event simulations, system dynamics, agent-based simulations, and multi-method simulations. Currently, schedules used to establish 4D models are created via traditional construction planning methods, using historical data and heuristic adjustments. Thus, 4D models cannot offer either optimal or reliable plans; they remain on the level of visualization, not planning. If schedules are not developed in a reliable manner, construction managers can only use 4D models as visual communication tools and not as facilitators of planning, analyzing, and decision-making. To address the current limitations in this area, we have employed computer simulations as a means of developing construction plans and providing reliable schedules to 4D models. There has been a substantial body of research on applying computer simulation technology to the construction field. The first simulation approach developed in this area was called CYCLONE [20], which can be used to model simple cyclic networks. For example, micro-CYCLONE simulations were used to analyze concrete batch plant operations according to resource combinations and distances [21]. However, this system was too simple to accurately model recourses, and thus tended to manipulate them in the model [9]. Then, with the development of object-oriented process simulation, simulation tools were improved to create more practical models with better user interfaces and the flexibility to adjust the model’s scope; as a result, reliability improved. AbouRizk and Hajjar [22] introduced Simphony as a simulation language. Exploiting the possibilities of this new simulation tool, researchers suggested a production-based framework using simulations to establish ideal project execution plans; they also discussed the requirements for extending the scope of such simulations [2]. Subsequently, researchers have attempted to integrate simulation technology with other tools [9]. One example


4 of 25

was the integration of 3D-CAD (Computer-Aided Design) with computer simulations to support decision making during construction [23]. Other researchers used discrete-event simulation for planning by designing construction operations and then visualizing the simulated operations in 3D [24]. With the advancements that have been made with modeling tools, the potential applications of computer simulation have increased [9]. For instance, simulation models can now simulate and analyze construction operations, and then support construction scheduling. Ultimately, they can predict the utilization of resources and identify logistics bottlenecks [25]. However, to utilize computer simulations in the construction industry, we first need to solve limitations such as difficulties with modeling from simulations and unreliable use of data. This research is an effort to minimize the effects of these limitations and convince practitioners of the merits of using simulations in the construction industry. If researchers built standard simulations of each construction operation’s work, practitioners could use those simulations for planning by only modifying the parameters salient to the particular project. Additionally, by using BIM-provided information and integrating BIM with simulations, the reliability of the simulation’s data usage would be greatly improved. Therefore, this research suggests a new methodology for establishing reliable construction plans via a convenient and practical BIM-integrated simulation framework. 2.2. Lean Construction Lean production, which tries to eliminate any waste of materials, time, or effort from a production process, was developed by Toyota (a Japanese car manufacturing company) [26,27]. It stemmed from the need to create a system of production that responded to customers’ orders (i.e., not mass production). This approach greatly increased Toyota’s productivity. Consequently, many industries have attempted to adopt the concept. In the construction industry, lean production was introduced in the mid-1990s [27]. The major aims are the minimization of waste and maximization of value. Various tools, methods, and techniques are employed to accomplish these goals. Applying lean thinking to the construction field has led to the development of a variety of planning and control systems [28]. For example, Mao and Zhang [3] established a construction re-engineering process that combines lean principles and computer simulation techniques. Based on such principles, their research classified construction activities into conversion and flow activities instead of the value-added/non-value-added classification system that was standard; the resulting re-engineering process avoids confusion in classification [3]. In this context, the application of lean principles in the construction field can improve the efficiency of construction planning. The most important lean principle is the minimization and eventual elimination of wasted time, money, equipment, etc. [29,30]. On construction sites, there are many kinds of waste, such as the time crews spend waiting to begin a project or reworking incorrectly completed tasks, and inefficient handling of materials, inventory and workspace [28]. According to parade game theory [31], these types of waste increase with variations in flow, which in construction affects the entire supply chain. Specifically, variations in the actual and expected outcomes result in the need for buffers to manage variability in orders. However, buffers mean wasted resources spent on management and control, even when the buffers are not absolutely necessary. These, the buffers increase the production cycle and extent of the project’s duration and budget [31,32]. If managers fail to reliably predict production in the field, they cannot help but accept an excess or lack of stock. The alternative is to control their inventory by change-orders. As a means of preparing for this type of situation, suppliers employ some number of buffers to shield against change-orders. If this unreliable and unpredictable situation continues, participants along the supply chain cannot achieve suitable levels of cost- and time-efficiency; efficiency along the entire extent of the supply chain fails. Reliability in production allows all contractors to manage their work with a minimum of buffers, and implement Just-In-Time production. Reliable production achieves minimum waste by delivering the right quantity at the right time, maximizing the value for the client by minimizing cost overruns and duration extensions.


5 of 25

Reliable construction planning is essential to eliminating waste in construction projects. However, it is difficult to accurately predict the needs of a particular project, because of their inherent uniqueness and complexity. In addition, productivity cannot be predicted at a normal or static rate because it involves dynamic interactions and complicated variables that come together in a variety of complex systems [33]. Productivity dynamics generated from new labor, a unique learning curve, or changes in the site’s conditions can all lead to waste. Therefore, construction managers should consider such variables when planning. To accomplish this goal, we developed a BIM-integrated simulation framework. This framework will assist in the prediction of productivity dynamics, contribute to minimizing inventory on site, and increase reliability with regards to orders. The result will be a decrease in buffers and a realization of the key principles of lean construction. 2.3. Application of BIM BIM is a remarkable technology regularly employed in the Architecture/Engineering/Construction (AEC) industry [34]. It was developed as a tool for engineers to use to generate and manage building information and facilitate three-dimensional design [35]. BIM applications are rapidly becoming more common in the construction industry because they are useful for reducing cost and time, and enhancing project quality [36]. The major merit of this technology is the visual enhancements it provides, which are useful means of facilitating communication among stakeholders. In addition to providing a means of visualizing 3D geometric expressions, BIM incorporates a variety of process data useful for analyzing constructability. Furthermore, in BIM, architectural design is integrated with energy simulations to efficiently and accurately apply simulation input [34]. Geometric information is also integrated into structural simulations to analyze structural safety; detailed models are produced that reflect the uniqueness and authenticity of a building without excessive geometric simplifications [37]. Likewise, BIM is utilized in many other aspects of construction, such as analyses of building energy, structure, and constructability. Rapid improvements in BIM technology have made building performance evaluations easier and quicker to complete; for example, energy consumption [34,38] and structural design analyses [37], and structural evaluations during construction [39] use BIM data to obtain the details necessary for a thorough investigation of a building’s performance. Construction management also uses BIM to assist in visual communication among all key parties, the integration necessary to obtain new managerial knowledge, and automation that improves managerial efficiency. Yet in spite of these important uses, in the construction management field BIM data continue to be utilized for only simple and partial purposes. For instance, according to a study on project scheduling via the integration of BIM with construction process simulations [1,19], researchers only used BIM data to obtain the quantities of materials needed. Although quantity takeoff data is necessary when determining the duration of the construction process [25], on its own it is not sufficient for reliable construction planning. Specifically, operational and managerial data are needed, such as workable hours, amount of skilled labor, level of difficulty, and so on. 3. Methodology 3.1. Productivity Dynamics In the past, scheduling for the construction industry has been based on average historical data collected from similar projects and heuristic adjustments made by construction managers. Construction managers multiply the quantity of materials by unit of time, based on historical data, and then calculate the total duration of the work. Then, the quantity of materials is multiplied by duration; as a result, daily production and productivity is established. Afterwards, managers heuristically adjust their construction production plans in accordance with initial delays, learning effects, and the need for re-work. In actuality, though, such plans tend to be quite different from execution. Since construction work follows an outdoor, project-based production process, the environment cannot be controlled and


6 of 25

variability is prevalent [40]. This variability leads to cost overruns and on-site delays (see Figure 1). For instance, if production is greater than expected, resources are wasted; this waste leads to unexpected costs and work delays (see Figure 1). On the other hand, if production is lower than expected, established buffers are unnecessary; this extra inventory leads to unexpected managerial costs and non-value-adding activities that waste resources [40] (see Figure 1). The goals of lean construction can be achieved by reducing or reliably predicting output variability [40]. Thus, in order to minimize the cost overruns and delay extensions caused by differences in plan and execution, this Sustainability 2016, 8, a 1106 6 of 25 planning. research paper presents BIM-integrated simulation framework for reliable production This research will lead to all parties along the supply chain achieving the expected level of efficiency framework for reliable production planning. This research will lead to all parties along the supply and reliable production plans for thelevel construction a result, the waste money, equipment, chain achieving the expected of efficiencysite; andas reliable production plansof fortime, the construction Sustainability 2016, 8, 1106 6 of 25 site; as a be result, the waste of time, money, equipment, and materials will be minimized. and materials will minimized. framework for reliable production planning. This research will lead to all parties along the supply chain achieving the expected level of efficiency and reliable production plans for the construction site; as a result, the waste of time, money, equipment, and materials will be minimized.

A diagram of productivity dynamics. FigureFigure 1. A 1. diagram of productivity dynamics.

3.2. BIM-Integrated Simulation Framework

3.2. BIM-Integrated Simulation Framework

Above, we explained why reliable production planning is important to achieving lean Figure 1. A diagram of productivity dynamics.

it increases the efficiency of all parties involved in the construction project. Above,construction; we explained why reliable production planning is important to achieving lean Consequently, in this work we argue for the use of computer simulations as a production process 3.2. BIM-Integrated Simulation Framework construction;assessment it increases of alldata parties involved in the section, construction project.a BIMConsequently, tool the and efficiency BIM as a reliable source. In the following we will present in this work integrated we arguesimulation forexplained the use of computer simulations as Subsequently, a production assessment Above, we why reliable production planning is important to demonstrate achieving lean framework for construction planning. weprocess will the tool and construction; increases efficiency of section, all parties in thea construction project.simulation methodology’s usefulness in establishing accurate productivity BIM as a reliable dataitsource. In the the following weinvolved willdynamics. present BIM-integrated Consequently, in this work we argue for the use of computer simulations as a production process framework 3.2.1. for Design construction planning. Subsequently, we will demonstrate the methodology’s BIM-Integrated Simulation assessment toolofand BIM as a reliable data source. In the following section, we will present a BIMusefulnessintegrated in establishing accurate productivity dynamics. simulation framework for construction planning. Subsequently, we will demonstrate the Construction managers usually develop their construction plans according to historical data and methodology’s usefulness establishing accurate productivity dynamics. heuristic adjustments (seeinFigure 2).

3.2.1. Design of BIM-Integrated Simulation

3.2.1. Design of BIM-Integrated Simulation

Construction managers usually develop their construction plans according to historical data and Construction managers usually develop their construction plans according to historical data and heuristic adjustments (see Figure 2). heuristic adjustments (see Figure 2).

Figure 2. Traditional construction planning process.

Since construction projects often differ in terms of design, site, and environment, historical data cannot guarantee the expected level of performance. Also, because a considerable number of factors influence construction productivity, managers’ heuristics cannot incorporate all managerial and Figure 2. Traditional construction planning process. operational details.Figure Therefore, this traditional construction planning planning method cannot produce reliable 2. Traditional construction process. plans that properly consider a project’s uniqueness and complexity. To remedy this, we have Since construction projects often differ in terms of design, site, and environment, historical data designed a set of BIM-integrated simulation procedures that can be customized to produce accurate guarantee the expected level of performance. Also, because a considerable number of factors Sincecannot construction projects often differ synthetic planning (see Figures 2 and 3). in terms of design, site, and environment, historical influence construction productivity, managers’ heuristics cannot incorporate all managerial and data cannot guarantee the expected level of performance. Also, because a considerable number of operational details. Therefore, this traditional construction planning method cannot produce reliable factors influence construction productivity, managers’ cannotToincorporate and plans that properly consider a project’s uniquenessheuristics and complexity. remedy this,all wemanagerial have designed a set of BIM-integrated simulation procedures that can be customized to produce accurate synthetic planning (see Figures 2 and 3).


7 of 25

operational details. Therefore, this traditional construction planning method cannot produce reliable plans that properly consider a project’s uniqueness and complexity. To remedy this, we have designed a set of BIM-integrated simulation procedures that can be customized to produce accurate synthetic planning (see Figures 28,and 3). Sustainability 2016, 2016, 8, 1106 Sustainability 1106 7 of 25 7 of 25

Figure 3. Proposed construction planning process using the BIM-integrated simulation framework.

FigureFigure 3. Proposed construction planning theBIM-integrated BIM-integrated simulation framework. 3. Proposed construction planningprocess process using using the simulation framework. First, we prepared a BIM model that included building information such as materials, geometry,

we so prepared a we BIM model that included building information such as materials, geometry, size,prepared and on. Then, developed commands using BIM Application Programming Interface (API) First, First, we a BIM model that included building information such as materials, geometry, (C#)so toon. extract thewe BIM data and commands translate that information into resource data for the Interface computer (API) size, and Then, developed using BIM Application Programming size, and so on. Then, we developed commands using BIM Application Programming Interface simulation. These data were then input into constructioninto operation simulation. Finally, we (C#) to extract the BIM data and translate thattheinformation resource data for the computer (API) (C#) to extract the BIM data and and translate that ininformation intoplanning, resource data forand the computer obtained productivity dynamics used our production scheduling, simulation. These data were then input intothem the construction operation simulation. Finally, we resource allocation. simulation. These data were then input into the construction operation simulation. Finally, weand obtained obtained productivity dynamics and used them in our production planning, scheduling, productivity dynamics and used them in our production planning, scheduling, and resource allocation. resource allocation. 3.2.2. BIM-Integrated Simulation Framework

We developed a BIM-integrated simulation framework for predicting productivity dynamics 3.2.2. BIM-Integrated Simulation Framework 3.2.2. BIM-Integrated Simulation Framework

and facilitating reliable production management. BIM offers reliable data for construction process

simulations. developing commands using theframework BIM API, wefor were able to extract BIM data dynamics and dynamics We developed a BIM-integrated simulation framework productivity We developed aBy BIM-integrated simulation forpredicting predicting productivity translate it into building data thatmanagement. could then be input into thereliable Using BIM will allow and facilitating reliable production BIM data for for construction process and facilitating reliable production management. BIMoffers offers simulation. reliable data construction process construction managers tocommands consider construction projects’ unique characteristics such as materials, simulations. By developing using the BIM API, we were able to extract BIM data and simulations. By developing the BIM API, weconstruction were able to extract BIM data and geometry, and quantitycommands and quality of using work needed. This will save managers both time translate it into building data that could then be input into the simulation. Using BIM will allow translate itand into building data that could then be input into the simulation. Using BIM will allow money. construction managers to consider construction projects’ unique characteristics such as materials, The simulation consisted of the overall construction work process, critical factors affecting geometry, construction managers to consider construction projects’ unique characteristics such as materials, geometry, and quantity and quality work needed. This setting will save construction managers both productivity, work resources, BIMofdata, and managerial information. By using Anylogic 7 time and quantity and quality of work needed. This will save construction managers both time and money. and money. (Anylogic Company: St. Petersburg, Russian, 2013), a remarkable simulation technology already in The simulation consisted the overall construction work process, critical factors affecting The simulation consisted of the overall construction process, critical affecting use in many industries, weof were able to synthetically considerwork the complex conditions offactors a standard construction project and a substantial amount of data. By using this BIM-integrated simulation productivity, resources, BIM data,and andmanagerial managerial setting By using Anylogic 7 productivity, workwork resources, BIM data, settinginformation. information. By using Anylogic 7 framework, construction managersRussian, will be able to a develop reliablesimulation productivity dynamics already from (Anylogic Company: St. Petersburg, 2013), remarkable technology in in use (Anylogic Company: St. Petersburg, Russian, 2013), a remarkable simulation technology already data (see Figure 4); asable a result, variations between planned and actual productionofwill be use industries, inaccurate many industries, we were to synthetically consider the complex conditions a standard in many minimized. we were able to synthetically consider the complex conditions of a standard construction project and a substantial amount of data. By using this BIM-integrated simulation construction project and a substantial amount of data. By using this BIM-integrated simulation framework, construction managers will be able to develop reliable productivity dynamics from framework, construction managers will be able to develop reliable productivity dynamics from accurate accurate data (see Figure 4); as a result, variations between planned and actual production will be data (see Figure 4); as a result, variations between planned and actual production will be minimized. minimized.

Figure 4. BIM-integrated simulation framework.



Sustainability 2016, 8, 1106 Sustainability 2016, 8, 1106

8 of 25 8 of 25

Developmentof ofBIM-Integrated BIM-IntegratedSimulation Simulation 4. Development section, we we explain explainthe thespecific specificprocedures proceduresfor forapplying applyingaaBIM-integrated BIM-integratedsimulation simulationtoto In this section, a a construction project. construction project. 4.1. Preprocessing Preprocessing aa BIM 4.1. BIM Model Model BIM models models contain containconstruction constructioninformation informationsuch suchasasa abuilding’s building’s geometry, size, type, BIM geometry, size, type, andand so so on shown in Figure 5. To generate the BIM model for this research, we used Autodesk on shown in Figure 5. To generate the BIM model for this research, we used Autodesk RevitRevit 2015 2015 (Autodesk: San Rafael, CA, USA, 2015). Autodesk provides powerful which (Autodesk: San Rafael, CA, USA, 2015). Autodesk RevitRevit provides powerful .NET.NET API,API, which we we employed to extend the information embodied in the BIM model. We then retrieved the employed to extend the information embodied in the BIM model. We then retrieved the inputinput data data relating to construction the construction process simulation programming the.NET .NETframework frameworkin in the the C# relating to the process simulation by by programming the C# programming language. programming language.

Figure 5. Autodesk Revit model combining the project’s information. Figure 5. Autodesk Revit model combining the project’s information.

4.2. BIM2SIM Approach 4.2. BIM2SIM Approach We developed commands to extract and change BIM data in order to input the resource data We developed commands to extract and change BIM data in order to input the resource data needed for the construction process simulation. We used the .NET framework language C# to needed for the construction process simulation. We used the .NET framework language C# to compose compose the BIM2SIM commands. These commands first accessed the BIM information and the BIM2SIM commands. These commands first accessed the BIM information and extracted the BIM extracted the BIM data; then, they translated the BIM data into simulation input data. The BIM2SIM data; then, they translated the BIM data into simulation input data. The BIM2SIM approach is depicted approach is depicted in Figure 6. The following procedures were required to extract the BIM data for in Figure 6. The following procedures were required to extract the BIM data for our construction our construction operation simulation. operation simulation.  Produced a pre-built BIM model Produced pre-built BIM model • Developeda code for extracting data from BIM Developed for extracting data BIM • Constructedcode a dataset sorted for thefrom simulation • Translated that datasetsorted into simulation input data Constructed a dataset for the simulation • Translated that dataset into simulation input data


9 of 25 9 of 25


9 of 25

Figure 6. Diagram of the BIM2SIM process. process.

4.3. Construction Construction Operation Operation Simulation Simulation 4.3. Figure 6. Diagram of the BIM2SIM process. After preparing preparinga pre-built a pre-built model and composing commands using BIM’s API, we After BIMBIM model and composing commands using BIM’s API, we developed 4.3. Construction Operation Simulation developed a construction operational simulation shown in Figure 7. To build a reliable simulation, a construction operational simulation shown in Figure 7. To build a reliable simulation, we first preparing a pre-built BIM model and composing commands using BIM’s API, we we firstAfter scrutinized various construction processes by conducting a thorough literature review, scrutinized various construction processes by conducting a thorough literature review, interviewing developed a construction operational simulation shown in Figure 7. To build a reliable simulation, interviewing and practitioners, and analyzing Then, we identified theinfluence critical factors that practitioners, analyzing textbooks. Then, textbooks. we identified the critical factors that operational we first scrutinized various construction processes by conducting a thorough literature review, influence operational work productivity. Using the information collected from our analysis, we built work productivity. Using the information collected from our analysis, we built a conceptual simulation interviewing practitioners, and textbooks. Then, we identified the critical factors thatan a conceptual simulation as aanalyzing scheme, listed the critical factors, and developed model as a scheme, listedmodel the critical productivity factors, andproductivity developed an operational simulation influence operational work productivity. Using the information collected from our analysis, we built operational simulation using the Thetool resulting toolinstrument is the first and only using the Anylogic software. TheAnylogic resultingsoftware. simulation is the simulation first and only to bring a conceptual simulation model as a scheme, listed thedynamics, critical productivity factors, simulations and developed an instrument to bring together discrete events, system and agent-based in one together discrete events, system dynamics, and agent-based simulations in one model development operational simulation using the Anylogic software. The resulting simulation tool is the first and only model development environment. Using the tool, Anylogic simulation tool, we followed thedevelop steps listed environment. Using the Anylogic simulation we followed the steps listed below to and instrument to bring together discrete events, system dynamics, and agent-based simulations in one below to develop and conduct the final construction operation simulation. conduct the final construction operation simulation. model development environment. Using the Anylogic simulation tool, we followed the steps listed  Defined the work work procedures for theconstruction constructionoperation operation to develop and conduct the final simulation. • below Defined the procedures for the construction operation  Determined the critical factors affecting work productivity •  Determined thework critical factors affecting work productivity Defined the procedures for the construction operation  Developed the simulation •  Developed the simulation Determined the critical factors affecting work productivity  Extracted the building data from BIM to conduct simulation Developed the simulation •  Extracted the building data from BIM to conduct simulation   Implemented the computer simulation Extracted the data from BIM to conduct simulation • Implemented thebuilding computer simulation  Implemented the computer simulation

Figure 7. Diagram of the simulation development process. Figure Diagram thesimulation simulationdevelopment development process. Figure 7. 7. Diagram ofofthe


10 of 25 10 of 25

5. Application Applicationfor forBIM-Integrated BIM-IntegratedSimulation SimulationMethod Method In this section, we discuss the application of the BIM-integrated simulation method to a steel erection project, in order to examine the applicability of the method method for for actual actual construction construction projects. projects. Steel erection is one of the most important processes on the construction site; it has critical effects on a project’s duration and cost. Consequently, steel erection work is one of the most actively-managed areas; accuracy is essential to reliable production planning. Therefore, by applying this method to steel erection work, we were able to validate the applicability of the method and minimize any excess money and time that would need to be spent on this step step of of the the construction. construction. 5.1. Preprocessing BIM Data Used for Planning We prepared BIMBIM model for anfor actual project inproject Seoul. in TheSeoul. building’s We prepareda steel-structure a steel-structure model an construction actual construction The area was approximately 390 m2 , and site area was approximately 650 m2650 . The had building’s area was approximately 390 the m2, and the site area was approximately m2.building The building a basement andand fourfour above-ground floors. The structure had a basement above-ground floors. The structurewas wassteel-reinforced steel-reinforcedconcrete. concrete. We only elements of of the structure structure to to test test our our method method for for steel steel erection erection work work (see (see Figure Figure 8). 8). modeled the steel elements The BIM model offered all of the necessary information, such as the number of columns and beams, size, BIM model offered all of the necessary information, such as the number of columns and beams, type of eachofmaterial, geometric location,location, and so on. These dataThese were data then used theused construction size, type each material, geometric and so on. werefor then for the simulation model. construction simulation model.

Figure steel model. model. Figure 8. 8. BIM BIM model model used used to to test test the the structural structural steel

5.2. Implementing Commands Using BIM’s API to Extract Building Information 5.2. Implementing Commands Using BIM’s API to Extract Building Information After preparing the BIM model, we developed commands using BIM’s API and the C# .NET After preparing the BIM model, we developed commands using BIM’s API and the C# .NET framework languages; these commands were imported using the add-in tool from Autodesk Revit framework languages; these commands were imported using the add-in tool from Autodesk Revit 2015. The commands allowed engineers to obtain the required information embodied in the BIM 2015. The commands allowed engineers to obtain the required information embodied in the BIM model. The commands translated the extracted building information into input resource data for use model. The commands translated the extracted building information into input resource data for use in the construction operation simulation. The procedures used to retrieve the building information in the construction operation simulation. The procedures used to retrieve the building information can be found in Figure 9. This process saved time, money, and effort that would otherwise have been can be found in Figure 9. This process saved time, money, and effort that would otherwise have been needed to reproduce the building data already in the BIM model. needed to reproduce the building data already in the BIM model.


11 of 25


11 of 25

Figure 9. Data retrieval process from the BIM model, and the input data preparation process. Figure 9. Data retrieval process from the BIM model, and the input data preparation process.

5.3. Development of the Construction Operation Simulation 5.3. Development of the Construction Operation Simulation We composed a steel erection work simulation via a discrete-event simulation approach because We composed a steel erection work simulation via a discrete-event simulation approach because steel erection work is generally performed as a sequence of events over time. However, we noticed steel erection work is generally performed as a sequence of events over time. However, we noticed that steel construction requires installed location factors such as height; thus, we interpreted steel to that steel construction requires installed location factors such as height; thus, we interpreted steel to be a simple agent. We also conducted a literature review, expert interviews, and a textbook analysis be a simple agent. We also conducted a literature review, expert interviews, and a textbook analysis to determine a reliable construction operation process and appropriately used the retrieved building to determine a reliable construction operation process and appropriately used the retrieved building data in the simulation environment. To properly express the steel erection work process, we first data in the simulation environment. To properly express the steel erection work process, we first composed a simulation outline to serve as a blueprint. Then, we identified critical factors affecting composed a simulation outline to serve as a blueprint. Then, we identified critical factors affecting work productivity by scrutinizing this type of work, interviewing experts, and reviewing the relevant work productivity by scrutinizing this type of work, interviewing experts, and reviewing the relevant literature. The Anylogic software was used to build the model. We arranged the simulation libraries literature. The Anylogic software was used to build the model. We arranged the simulation libraries and parameters into processes and resources, according to their properties. In the following section, and parameters into processes and resources, according to their properties. In the following section, we present our process for developing a steel erection work simulation on a construction site. we present our process for developing a steel erection work simulation on a construction site. 5.3.1. Composing the Simulation Outline 5.3.1. Composing the Simulation Outline After a literature review, interviews with experts, and a textbook analysis, we built our scheme After a literature review, interviews with experts, and a textbook analysis, we built our scheme for steel erection work on a construction site (see Figure 10). The work was divided into four sectors for steel erection work on a construction site (see Figure 10). The work was divided into four sectors and fifteen operational processes. The first sector involved transporting materials to the construction and fifteen operational processes. The first sector involved transporting materials to the construction site. The second was the temporary erection of the columns. The third sector was the temporary site. The second was the temporary erection of the columns. The third sector was the temporary erection of the beams. The final sector included installation of all erection materials. There were four erection of the beams. The final sector included installation of all erection materials. There were processes included in the first phase: (1) transporting the steel to the site; (2) moving the steel to the four processes included in the first phase: (1) transporting the steel to the site; (2) moving the steel to stockyard; (3) unloading the steel; and (4) maintaining the steel in the stockyard. In the second phase, the stockyard; (3) unloading the steel; and (4) maintaining the steel in the stockyard. In the second there were also four processes: (1) preparing the columns to be lifted; (2) the lifting columns; (3) phase, there were also four processes: (1) preparing the columns to be lifted; (2) the lifting columns; arranging the columns to be correctly placed; and (4) temporarily erecting the columns. In the third (3) arranging the columns to be correctly placed; and (4) temporarily erecting the columns. In the third phase, there were five processes: (1) installing the safety device; (2) preparing the beam materials to phase, there were five processes: (1) installing the safety device; (2) preparing the beam materials to be be lifted; (3) lifting the beams; (4) arranging the beams to be correctly placed; and (5) temporarily lifted; (3) lifting the beams; (4) arranging the beams to be correctly placed; and (5) temporarily erecting erecting the beams/girders. The final phase was divided into two processes: (1) gauging the the beams/girders. The final phase was divided into two processes: (1) gauging the verticality; and verticality; and (2) erecting all materials. (2) erecting all materials.


12 of 25 12 of 25

Figure Figure10. 10.Composing Composingthe thesimulation simulationblueprint. blueprint.

5.3.2. Finding the Factors Critical to the Construction Operation 5.3.2. Finding the Factors Critical to the Construction Operation We identified the critical factors that influence the productivity of steel erection work on We identified the critical factors that influence the productivity of steel erection work on construction sites. construction sites.ToToverify verifythe thereliability reliabilityofofour ourconclusions, conclusions,we we analyzed analyzed textbooks, textbooks, interviewed interviewed specialists, and reviewed thethe relevant literature on on thethe topic. According to research in this field,field, the specialists, and reviewed relevant literature topic. According to research in this factors that impact steel work productivity are extremely complex and diverse, and include the the factors that impact steel work productivity are extremely complex and diverse, and include the constructability and completeness constructability and completenessofofthe thedesign, design,responsibility responsibilityofofthe theworkers, workers,surrounding surroundingconditions, conditions, construction period, site size, structure type, supply of the materials, height of the work, and soso on.on. construction period, site size, structure type, supply of the materials, height of the work, and Among foundsix sixthat thatwere wererepeatedly repeatedly stressed experts. (1) Amongthese theseand andother other factors, factors, we we found stressed by by thethe experts. (1) The The number of workable days. Workable days are days in which it is possible to further the construction number of workable days. Workable days are days in which it is possible to further the construction project; this exclude project; this excludedays dayswith withbad badweather weatherororduring duringwhich whichunexpected unexpectedaccidents accidentsoccur. occur. This This factor factor has a significant impact on productivity; it is directly connected to the time resource input value; (2) has a significant impact on productivity; it is directly connected to the time resource input value; Variety of erection materials used in the same area. The more materials needed in a single area, the more (2) Variety of erection materials used in the same area. The more materials needed in a single area, the more time the project consumes; time the project consumes;(3) (3)The Thecapacity capacityofofthe thestockyard. stockyard.This Thisfactor factorinfluences influencesthe thematerials’ materials’ supply supply cycle; (4) The height of the project. If the steel erection work is performed above the fifteenth floor, cycle; (4) The height of the project. If the steel erection work is performed above the fifteenth floor, logistics logistics of placement so complex andspent the time spent lifting sothat prolonged that of placement become become so complex and the time lifting becomes so becomes prolonged productivity productivity is always reduced; (5) Access to a tower crane. Steel erection work requires the use of a is always reduced; (5) Access to a tower crane. Steel erection work requires the use of a tower crane. tower crane. access factor is a critical factor to productivity; Workers’ of skill. are Thus, easy Thus, accesseasy is a critical to productivity; (6) Workers’ (6) level of skill. level Workers areWorkers diverse and diverse and unpredictable, and theirand efficiency and knowledge critical to productivity. unpredictable, and their efficiency knowledge is critical to is productivity. 5.3.3. Steel Erection WorkSimulation Simulation 5.3.3. Steel Erection Work Using Anylogic, created a simulation model on a simulation the Using Anylogic, wewe created a simulation model basedbased on a simulation outlineoutline and theand aboveabove-discussed critical factors. The model was prepared as a discrete-event simulation (see Figure 11). discussed critical factors. The model was prepared as a discrete-event simulation (see Figure 11). Associations amongthe thecritical criticalfactors factorsand andsimulation simulationlibraries librarieswere wereorganized organized according according to to their their Associations among relation. As mentioned above, we divided the steel erection work into four sectors and expressed relation. As mentioned above, we divided the steel erection work into four sectors and expressed each differentline lineininthe thesimulation simulationmodel model(see (see Figure Figure 11). 11). However, However, the the specific specific simulation simulation each asasa adifferent processes went far beyond the defined steps in the simulation outline, due to the dummy steps processes went far beyond the defined steps in the simulation outline, due to the dummy steps necessary to meaningfully control the flow of materials. The critical factors were defined as: necessary to meaningfully control the flow of materials. The critical factors were defined as: ‘schedule’,‘numbersOfErect’, ‘numbersOfErect’,‘capaOfStockYard’, ‘capaOfStockYard’,‘heightOfErect’, ‘heightOfErect’,‘availableTC’, ‘availableTC’,and and ‘workersSkill’, ‘workersSkill’, ‘schedule’, which translated to workable days, number of erection materials, capacity of the stockyard, height of of which translated to workable days, number of erection materials, capacity of the stockyard, height the project, access to a tower crane, and skill of the workers, respectively. The schedule and parameter the project, access to a tower crane, and skill of the workers, respectively. The schedule and parameter simulationcomponents componentsininthe theAnylogic Anylogicsoftware softwarewere wereused usedto to incorporate incorporate these these factors factors into into the the simulation simulation model. simulation model.


13 of 25


13 of 25 13 of 25

Figure 11. Steel erection work simulation model in Anylogic. Figure 11. Steel erection work simulation model in Anylogic. Figure 11. Steel erection work simulation model in Anylogic.

Labor and equipment were defined as resources in Anylogic (see Figure 12). We initially divided We initially Labor defined as resources Anylogic (seewelding Figure 12). We dividedto labor into and five equipment categories: were one foreman, two bolting in workers, three workers, one worker fivehooking categories: oneforeman, foreman, two bolting workers, three welding workers, one worker labor intothe five categories: one two bolting workers, three welding workers, one worker toof support materials, and one signal man to support the lifting. The initial number toworkers support the hooking materials, and one signal man to support the lifting. The initial number of support the hooking materials, and one signal man to support the lifting. The initial number was determined by analyzing the standard crew used in steel erection work. Additionally, workers was determined Additionally, the and standard erection work.isAdditionally, we included one forklift, by oneanalyzing tower crane, smallcrew tools used used in forsteel bolting. A forklift used to move we included oneunder forklift, crane,itand toolsinused for bolting. because A forklift is capacity used to move steel materials 30 one tons;tower however, wassmall not used our simulation the of the it was not used in our simulation because the capacity of the steel materials under 30 tons; however, forklift was not considered critical to productivity. A tower crane moves one steel beam at a time. forklift was not considered critical to productivity. A tower crane moves one steel beam at a time. A tower crane moves onetools steelwere beamrequired at a time. Small tools are dummy resources and were included only to illustrate that for tools are dummy resources and were included only to illustrate that tools were required for the Small tools are dummy resources and were included only to illustrate that tools were required for the bolt workers. boltbolt workers. the workers.

Figure 12. Types and number of resources needed for steel erection work. Figure needed for for steel steel erection erection work. work. Figure 12. 12. Types Types and and number number of of resources resources needed

5.3.4. Simulation Libraries and Parameters 5.3.4. Libraries and Parameters 5.3.4. Simulation Simulation Libraries Parameters We determined that aand number of components and parameters were necessary for steel erection We determined that site. number of components components and were parameters were necessarythe forprocess steel erection erection work a construction Both ‘Source’ and ‘Sink’ essential to necessary building model Weondetermined that aa number of and parameters were for steel work on a construction site. Both ‘Source’ and ‘Sink’ were essential to building the process model because ‘Source’ provided a starting point and ‘Sink’ an end point. Despite the volume of steel work on a construction site. Both ‘Source’ and ‘Sink’ were essential to building the process model because ‘Source’ provided a starting point and ‘Sink’ an end point. Despite the volume of steel materials included in the process model, ‘carrying materials’ sector was expressed as of a single because ‘Source’ provided a starting pointthe and ‘Sink’ an end point. Despite the volume steel materials included the process model, the ‘carrying sector expressed as aused single entity because it was allprocess loaded onto a the single truck. materials’ To materials’ bind the steel to awas single entity, we the materials included ininthe model, ‘carrying sector was expressed as a single entity entity because it was all loaded onto a single truck. To bind the steel to a single entity, we used the ‘Batch’ component. ‘MoveTo’ usedtruck. to express a truck entering into a entity, stockyard space.the When the because it was all loaded onto was a single To bind the steel to a single we used ‘Batch’ ‘Batch’ component. ‘MoveTo’ wastoused to express a entering truck entering into a stockyard space. When the steel materials arrive at theused stockyard, workers must spend time them; expressed this component. ‘MoveTo’ was express a truck into aunloading stockyard space.we When the steel steel arrive the stockyard, workers must spend time unloading them; we expressed this step materials as a arrive ‘Delay’ Toworkers allocate a forklift before after the unloading process, westep added materials atcomponent. theat stockyard, must spend timeand unloading them; we expressed this as step as a ‘Delay’ component. To allocate a forklift before and after the unloading process, we added ‘Release’ components. If materials needed to bethe detached from their delivery truck, we a‘Seize’ ‘Delay’and component. To allocate a forklift before and after unloading process, we added ‘Seize’ ‘Seize’ and them ‘Release’ components. materials needed to be detached from delivery truck, weto identified individually via Ifthe ‘Unbatch’ We their used thetheir ‘Queue’ component and ‘Release’ components. If materials needed to becomponent. detached from delivery truck, we identified identified them individually via the ‘Unbatch’ component. used component the ‘Queue’tocomponent to demonstrate when steel incomponent. the stockyard until it was required for erection. them individually via theremained ‘Unbatch’ We used theWe ‘Queue’ demonstrate demonstrate when steel remained in the stockyard until it was required for erection. theremained second sector, ‘temporary column’for task, we selected the columns among the when In steel in thethe stockyard untilerecting it was required erection. In the in second sector, the erecting column’ we selected the columns among inventory the stockyard via‘temporary the ‘Pickup’ component. Thetask, columns could then be hooked to thethe lift, inventory in the stockyard via thespot, ‘Pickup’ columns could thensector, be hooked to the lift, lifted, arranged in the correct and component. temporarilyThe erected. In the third the ‘temporary lifted, arranged in the correct spot, and temporarily erected. In the third sector, the ‘temporary


14 of 25

In the second sector, the ‘temporary erecting column’ task, we selected the columns among the inventory in the stockyard via the ‘Pickup’ component. The columns could then be hooked to the lift, lifted, arranged in the correct spot, and temporarily erected. In the third sector, the ‘temporary erecting beam’ task, we followed a similar process. We first selected the beam materials via the ‘Pickup’ component, and then installed the safety equipment, hooked the beams to the lift, lifted the beams, arranged the beams on the correct spots, and then temporarily erected them. Finally, the ‘erecting all materials’ sector followed a relatively simple set of two steps: checking the verticality of all of the materials, and firmly erecting all of the installed steel materials. In order to erect the beams after the columns were installed, we added a ‘Hold’ component between the first and second sectors. The blocking condition for ‘Hold’ was whether the number of columns in the batch was the same as the number of beams in the batch. From this component, we were able to prevent steel materials from jumping onto a ‘temporary erecting beam’ sector before the columns were erected. In Table 1, the specific simulation components and our process of naming are described. Table 1. Simulation steps for steel erection work. Simulation Components

Naming

Source

source

Starting point of the process model

Batch

batch

Binding steels to one entity because requested steels are entered in one truck

MoveTo

truckEnter

MoveTo

driveToUnload

Seize

seizeSt

Description

Entering a truck Moving truck to a stockyard space Allocating steels to forklift for unloading

Delay

unloading

Release

releaseSt

Releasing steels from forklift after unloading

Unbatch

unbatch

Detaching steels from a truck

Queue

managSt

Hold

hold

Unloading step

A stockyard to load steals Holding steel entities to pass to erecting beams before erecting columns are completed

Pickup

pickupCol

Selecting the number of columns planned to be erected

Unbatch

unbatchCol

Selected one entity is divided to individual columns.

Service

hookCol

Service

liftColumn

Service

erectCol

Arranging columns on right point and erecting columns temporarily

Batch

batchCol

Binding erected columns as one entity to move onto next step when columns stretched to three stories are erected

Pickup

pickupBe

Selecting the number of beams planned to be erected

Unbatch

unbatchBe

Dividing selected one entity into individual beams.

Delay

equipSafety

Preparing safety materials

Service

hookBeam

Hooking/Wiring beam to tower crane

Service

liftBeam

Service

erectBeam

Hooking/Wiring column to tower crane Lifting columns one by one

Lifting beams one by one Arranging beams on right point and erecting beams temporarily

Batch

batchBe

Binding erected beams as one entity to move on next step when beams are erected

Delay

checkAll

Checking the vertical states of temporary erected materials

Service

erectAll

Installing all erected materials permanently

Sink

sink

Ending point of the process model

We presumed that 1 s in the simulation equated to 6 min in real time. We then defined the simulation input values and assumptions for each simulation step. Among these steps, we determined those directly related to productivity: ‘hookCol’, ‘liftColumn’, ‘erectCol’, ‘equipSafety’, ‘hookBeam’, liftBeam’, erectBeam’, ‘checkAll’, and ‘erectAll’. To determine the reliable productivity output provided by the simulation, we needed an accurate way of defining these nine steps. Through a literature review


15 of 25

and interviews with experts, we configured dependable values and an appropriate distribution for the steps presented in Table 2. Table 2. Input values and assumptions. Steps

Input Values

Description or Assumptions

source

1 entity every 5 s

Source depends on the schedule; this schedule will be defined after predicting reliable production plan

batch

numberOfCol + numberOfBeam (+dummySteel)

We assumed that columns and beams of each level are entered in one truck

truckEnter

None (Dummy step)

This step means that a truck enters a site. The speed can be limited during simulation

driveToUnload

None (Dummy step)

This step means that a truck moves in front of the stock yard for unloading. The speed can be limited during simulation

seizeSt

Resource sets: forklift

This step allocates steels to the resource (forklift).

* triangular(1, 5, 4)

This step is a dummy step because it is not directly related to measure productivity. Then, we assumed input values depending on various cases.

releaseSt

Release: all seized resources

This step releases steels from the resource (forklift)

unbatch

None

This step is the dummy step for recognizing steel entities individually by detaching them from one truck

managSt

Capacity = ‘capaOfStockYard’

Its capacity is adjusted by the ‘capaOfStockYard’ parameter

Blocking condition: hold.out.count() != batchBe.out.count();

This step holds steel materials for allocating to the proper sequence

pickupCol

Quantity: numberOfCol

This step picks up columns to perform ‘temporary erecting columns work’

unbatchCol

None

This step is the dummy step for recognizing steel entities individually

hookCol

triangular(0.583, 1.167, 0.833)

We assume hooking column material one by one. Also, it requires the same amount of time with hookBeam

liftColumn

triangular(0.767, 1.067, 0.917)

This input distribution is acquired from analysis of previous literatures

erectCol

triangular(3.667, 6.167, 4.167)

This input distribution is acquired by cross-checking of previous literatures

batchCol

Size: numberOfCol

We bound columns on one floor plan into one batch

pickupBe

Quantity: numberOfBeam

This step picks up beams to perform ‘temporary erecting beams work’

unbatchBe

None

This step is the dummy step for recognizing steel entities individually

equipSafety

1.667

expert interview

hookBeam

triangular(0.183, 0.44, 0.258)

We assume hooking a beam material one by one because of preventing torsion

liftBeam

triangular(0.769, 1.069, 0.919)

This input distribution is acquired from analysis of previous literatures

erectBeam

triangular(3.731, 4.000, 3.869)

This input distribution is acquired by cross-checking of previous literatures

unloading

hold

batchBe

Size: numberOfBeam

We bound beams on one floor plan into one batch

checkAll

80

expert interview

erectAll

numberOfBeam * triangular(3.731, 4.000, 3.869)

The number of erecting materials installed is same as the number of temporary erecting beams

None

Ending point

sink

* 1 s in simulation = 6 min in real time. * triangular(min, max, mode).

Whereas the input distribution of the construction simulation fit the beta distribution [41], it was more difficult to obtain an amount of data sufficient to extract the beta distribution. Since triangular distribution is both simple and intuitive [42,43], we applied it as input data retrieved, based on data from previous literature (see Table 2). The triangular input data used for critical simulation steps were determined by cross-checking the previous literature on this topic. The values are presented in Table 3.


16 of 25

were determined by cross-checking the previous literature on this topic. The values are presented in Sustainability 16 of 25 Table 3. 2016, 8, 1106 Table 3. Triangular input data for critical simulation steps. Table 3. Triangular input data for critical simulation steps.

Steps hookCol liftColumn erectCol hookBeam liftBeam erectBeam erectAll

Duration (Minutes in Real Time *) DurationMost (Minutes in Real Time *) Minimum Likely (Mode) Maximum Steps Most Likely (Mode) Maximum 3.500Minimum 5.000 7.000 4.600 5.500 6.400 hookCol 3.500 5.000 7.000 liftColumn22.000 4.600 5.500 6.400 37.000 25.000 erectCol 1.100 22.000 37.000 25.000 1.550 2.640 hookBeam 1.100 1.550 2.640 4.613 6.413 5.513 liftBeam 4.613 6.413 5.513 24.000 23.213 erectBeam22.388 22.388 24.000 23.213 24.000 23.213 erectAll 22.388 22.388 24.000 23.213 * 1* s1 in = 66 min minininreal realtime. time. s insimulation simulation =

Steel Agent Agent Modeling Modeling 5.3.5. Steel

In the steel erection work model using discrete-event simulation (process-centric models), steel was used as an object in the process flow. We needed the amount of materials and the height of the erection, which whichwere wereobtainable obtainable from the BIM model. we obtained this information, we from the BIM model. After After we obtained this information, we classified classified the steelaccording materialstoaccording to their required orderprocess in the flow workshown process in the steel materials their required order in the work in flow Figureshown 13. First, Figure 13. the First, we sorted columns and Then, the steel was divided according separate we sorted columns andthe beams. Then, thebeams. steel was divided according to separate floorstobecause it floors because it neededaccording to be installed to In thethe floor order. In the model presented Figure needed to be installed to theaccording floor order. model presented in Figure 8, thein columns 8, the only columns were on only on level, the basement level, so we one created state for columns. were installed theinstalled basement so we created only stateonly for one columns. Conversely, Conversely, forbecause the beams because needed to be on allTo sixclassify floors. we preparedwe sixprepared states forsix thestates beams they neededthey to be installed oninstalled all six floors. To classify input values, two we created two parameters, ‘baselevel’ and ‘referenceLevel’. After these input these values, we created parameters, ‘baselevel’ and ‘referenceLevel’. After developing developing the for commands forthe extracting the BIM information, webase stored thevalue base of level of the the commands extracting BIM information, we stored the level thevalue columns in columns inand ‘baselevel’ and thelevel reference of the beams in ‘referenceLevel’. Since only columns ‘baselevel’ the reference value level of thevalue beams in ‘referenceLevel’. Since columns had only had a base and beams had a reference level, these two parameters all that were a base level andlevel beams only hadonly a reference level, these two parameters were all were that were needed. needed. Additionally, we reflected the height by dividing thefloor. state by floor. Additionally, we reflected the height factor byfactor dividing the state by

Figure Figure 13. 13. Steel Steel agent agent model model composed composed using using Anylogic. Anylogic.


17 of 25


17 of 25

5.4. Running Running the the BIM-Integrated BIM-Integrated Simulation Simulation 5.4. Within the Within the Autodesk Autodesk Revit Revit 2015 2015 environment, environment, we we opened opened the the prepared prepared BIM BIM model model and and added added in the commands described above. By executing the process, we obtained the simulation in the commands described above. By executing the process, we obtained the simulation input input data data extracted from Then, we we opened opened the the extracted from the the BIM BIM model model and and translated translated it it into into the the simulation simulation data data format. format. Then, Anylogic simulation application and imported the simulation input data. We also imported the steel Anylogic simulation application and imported the simulation input data. We also imported the steel work was added to to thethe steel work process simulation. In the work process processmodel. model.Next, Next,the theinput inputdata datamodel model was added steel work process simulation. In Anylogic simulation environment, different files can be combined if they are written in an Anylogic the Anylogic simulation environment, different files can be combined if they are written in an simulationsimulation format such as alp.such We then integrated resource the input data from BIMdata intofrom the steel Anylogic format as alp. We thenthe integrated resource input BIMwork into process simulation, performed the steel erection work process simulation, and measured the column the steel work process simulation, performed the steel erection work process simulation, and and beam erection productivities and the overall productivity of the erection. The entire process is measured the column and beam erection productivities and the overall productivity of the erection. presented in Figure 12. When we measured the productivity, we were able to obtain the optimal The entire process is presented in Figure 12. When we measured the productivity, we were able to resource combination by adjusting the input data. Consequently, method enabled us to obtain theinput optimal resource input combination by adjusting the input data. the Consequently, the method establish optimal resources and production plans that considered the construction project’s uniqueness enabled us to establish optimal resources and production plans that considered the construction and complexity. Figure 14complexity. shows the entire project’s uniqueness and Figureprocesses. 14 shows the entire processes.

Figure 14. Processes running the BIM-integrated simulation method.

5.4.1. 5.4.1. Model Model Input Input and and Assumptions Assumptions This section describes made when when we we applied applied This section describes the the model model input input data data and and assumptions assumptions that that were were made the BIM-integrated simulation to the actual steel installation work. Based on experts’ opinions, we the BIM-integrated simulation to the actual steel installation work. Based on experts’ opinions, determined thethe simulation steps critical toto the we determined simulation steps critical thesteel steelerection erectionwork worksimulation. simulation.These These steps steps included included hooking, lifting, erecting, installing safety equipment, and checking verticality. We analyzed the the hooking, lifting, erecting, installing safety equipment, and checking verticality. We analyzed model’s input values for these steps, found historical values in a previously-published case study, model’s input values for these steps, found historical values in a previously-published case study, cross-checked the data, then decided on minimum, maximum, and most likely values. This process was validated via the expert interviews.


18 of 25

cross-checked the data, then decided on minimum, maximum, and most likely values. This process was validated via the expert interviews. We also assumed that there was one foreman who worked on the ground, one signalman for the signal, one worker who hooked the materials to the tower crane, two workers who erected the structure, and two workers who welded between the columns. We identified these values via a textbook analysis, expert interviews, and a review of the published literature on this topic. We presumed access to a forklift and tower crane, and analyzed the required use of this equipment by employing the basic analysis tools in the Anylogic simulation software. Also, the amount of materials the forklift and tower crane could move could be modified during the simulation because we considered them adjustable managerial values. 5.4.2. Role of Critical Factors and Input Assumptions As mentioned above, we identified six critical factors affecting productivity for steel erection work: the number of workable days, amount of materials, capacity of the stockyard, height of the erection, availability of a tower crane, and skill of the workers. These critical factors were divided into those that were and were not obtainable from BIM. Then, we determined how to consider unobtainable factors in the simulation model. The method we followed for reflecting all critical factors is presented in Table 4. Table 4. Method for reflecting critical factors in the simulation model. BIM Information Obtainable

Unobtainable

Critical Factors

Method to Reflect Critical Factors in the Simulation Model (Technical Method/Assumptions)

The number of erecting materials

The factor could be obtained from BIM model by executing commands

The capacity of stockyard

The factor represented the number of capable stored steel materials in the stockyard, and it was considered as the managerial factor, so this value could be changed while running the simulation

The workable day

The factor was unpredictable. Thus, if we predicted productivity, we could draw proper workable days

The level of erecting height

We could obtain the height of erecting, but could not obtain the level of erecting height

The availability of using the tower crane

We first supposed that there is one tower crane in a construction site. By calculating the utilization of tower crane resource, we could deduce proper utilization of the tower crane for steel erection work

The skill of workers

The factor was also unpredictable. However, by supposing workers’ skill, we could draw proper combination of workers’ skill for steel erection work

We assumed that the height of the erection increased until Floor 3, then decreased to the roof; that is, the level of the columns in the basement was 1.2, while the levels of the erecting beams on the first, second, third, fourth, fifth, and sixth floors were 1.1, 1.0, 0.9, 1.2, 1.3, and 1.3, respectively. The production values were divided into these values to reflect the predicted steel work productivity. When we added the BIM information by using the above-mentioned commands, we classified the steel materials by their height; and then, each level of height was used to predict productivity. Additionally, since workers’ levels of skill can be unpredictable, we extrapolated the optimal combination of skilled workers through an analysis of the various scenarios. Specifically, since the required number of workers will differ according to their skill, we applied a different number of workers to each process, compared their level of utilization, and extrapolated the optimal combination. 5.4.3. Results from Running the BIM-Integrated Simulation Model We executed the BIM-integrated simulation model and built graphs of the work productivity based on the model input values and critical factors determined from the above-mentioned assumptions. The amount of materials needed for each level was determined by using BIM’s API commands.


19 of 25

We assumed that the capacity of the stockyard was more than the sum of the number of columns and Sustainability 2016,to 8, avoid 1106 restrictions from a limited stockyard. The difficulty of the work as the 19height of 25 beams in order increased was a productivity-impeding factor. The entire setup of all the critical factors is presented in as the Table 5. height increased was a productivity-impeding factor. The entire setup of all the critical factors is presented in Table 5. Table 5. Simulation running conditions. Table 5. Simulation running conditions. Setup Data

Setup Data The number of The number of erecting materials erecting materials The capacity The capacity of of stockyard stockyard The workable day The workable day The level level of The of erecting height erecting height The availability of The availability of using tower crane using tower crane



Base Case

Base Case Column: 18

Column: 18

Level 1

Level 1 Beam: 32

Beam: 32

Level 2

Level 2 Beam: 45

Beam: 45

Level 3

Level 3

Beam: 45

Beam: 45

Level 4

Level 4

Beam: 35

Beam: 35

Level 5

Level 5

Beam: 25

Beam: 25

Level 6

Level 6 Beam: 8

Beam: 8

100

100 Extrapolation from the results of productivity on site

Extrapolation from the results of productivity on site 1.2

1.2

1.1

1.1

1.0

1.0

0.9

0.9

1.2

1.2

1.3

1.3

1.3

1.3

Assumed one tower crane in the construction site. Extrapolated from the results of productivity Assumed one tower crane in the construction site. Extrapolated from the results of on site

productivity on site

Maintain the initial number of workers Maintain the initial number of workers Foreman Foreman Signal man Signal Workers forman welding Workersfor forwelding bolting Workers Workers for hooking

Workers for bolting Workers for hooking

1

11 12 22 1 2 1

We performed our BIM-integrated simulation according to the simulation’s running conditions performed our BIM-integrated simulation to the simulation’s running shownWe in Figures 15–22. The installation sequenceaccording was according to height; that is, inconditions the first set in Figures 15–22. installation was according to height; thatinstallation is, in the first set of ofshown erections, the beams onThe Levels 1, 2, andsequence 3 were installed sequentially after of the first the beams 1, 2, and 3 were installed after installation the first set seterections, of columns. Then,on in Levels the second set of erections, thesequentially beams on Levels 4, 5, and 6ofwere installed of columns.after Then, in the second setsecond of erections, beams The on Levels 4, 5, and were installed sequentially installation of the set of the columns. productivity of 6the beams varied sequentially after installation of the second set of columns. The productivity of the beams varied according to the erection height. The difficulty of the work increased according to the following floor according to the erection height. The difficulty of the work increased according to the following floor levels: 3, 2, 1, 4, 5, 6. Likewise, productivity increased according to the same sequence. While we levels: 3, 2, 1, 4,the 5, 6. Likewise, productivity according to the same sequence. While weofsupposed supposed that standard height was 1.0,increased we observed a difference if the difficulty level work was that the standard height was 1.0, we observed a difference if the difficulty level of work was varied more. varied more.

Figure column erection erection(89.187 (89.187ssininsimulation simulationtime). time). Figure15. 15.Productivity Productivity of of the the first first column

Figure16. 16.Productivity Productivity of of the the second second column Figure column erection erection(88.818 (88.818ssininsimulation simulationtime). time).

Sustainability 2016, 8, 1106 Sustainability 2016, 2016, 8, 8, 1106 1106 Sustainability Sustainability 2016, 8, 1106 Sustainability 2016, 8, 1106 Sustainability 2016, 8, 1106

Figure 17.Productivity Productivityof ofbeams beams in in reference reference level level 111(131.244 (131.244 ininsimulation simulation time). Figure 17. level (131.244sssin simulationtime). time). Figure 17. Productivity of beams in Figure 17. Productivity of beams in reference level 1 (131.244 s in simulation time). Figure 17. Productivity of beams in reference level 1 (131.244 s in simulation time). Figure 17. Productivity of beams in reference level 1 (131.244 s in simulation time).

Figure 18. 18. Productivity Productivity of of beams beams in in reference reference level level 2 (190.287 ss in in simulation time). time). Figure Figure level 222(190.287 (190.287ssininsimulation simulationtime). time). Figure18. 18.Productivity Productivityof ofbeams beams in in reference level (190.287 simulation Figure Figure 18. 18. Productivity Productivity of of beams beams in in reference reference level level 22 (190.287 (190.287 ss in in simulation simulation time). time).

Figure 19. 19. Productivity Productivity of of beams beams in in reference reference level level 33 (190.000 (190.000 ss in in simulation simulation time). time). Figure Figure 19. Productivity of beams in reference level 3 (190.000 s in simulation time). Figure 19. Productivity of beams in reference level 3 (190.000s sininsimulation simulation time). Figure Figure 19. 19. Productivity Productivity of of beams beams in in reference reference level level 33 (190.000 (190.000 s in simulation time). time).

Figure 20. 20. Productivity Productivity of of beams beams in in reference reference level level 44 (191.196 (191.196 ss in in simulation simulation time). time). Figure Figure 20. Productivity of beams in reference level 4 (191.196 s in simulation time). Figure 20. Productivity of beams in reference level time). Figure level444(191.196 (191.196sssin simulation time). Figure20. 20.Productivity Productivityof ofbeams beams in in reference level (191.196 ininsimulation simulation time).

Figure 21. 21. Productivity Productivity of of beams beams in in reference reference level level 55 (22.400 (22.400 ss in in simulation simulation time). time). Figure Figure 21. Productivity of beams in reference level 5 (22.400 s in simulation time). Figure 21. Productivity of beams reference Figure level555(22.400 (22.400sssin simulationtime). time). Figure21. 21.Productivity Productivityof ofbeams beams in in reference reference level level (22.400 ininsimulation simulation time).

20 of 25 20 of of 25 25 20 20 of 25 20 of 25 20 of 25


21 of 25 21 of 25

Figure22. 22.Productivity Productivityof ofbeams beams in in reference reference level 66 (7.540 Figure (7.540ssin insimulation simulationtime). time).

To allocate the production periods, we compared the times needed to complete each level of the To allocate the production periods, we compared the times needed to complete each level of the erection work, as shown in Table 6. The values in Table 6 can also be seen in the graphs displayed in erection work, as shown in Table 6. The values in Table 6 can also be seen in the graphs displayed in Figures 15–22. The total production time broadly depended upon the amount of erection materials. Figures 15–22. The total production time broadly depended upon the amount of erection materials. According to a comparison of the Levels 1 and 2 beam erection data, although Level 1 was more According to atotal comparison of time the Levels 1 andbecause 2 beamthe erection althoughmaterials Level 1 was was less. more difficult, the production was shorter amountdata, of necessary difficult, the total production time was shorter because the amount of necessary materials was less. In this regard, we found that the amount of materials has a greater influence on the production Inschedule this regard, thewe amount of the materials hasinadifficulty greater influence onifthe thanwe thefound height.that Thus, assumed variation to be 0.1, so theproduction variation schedule than the height. Thus, we assumed the variation in difficulty to be 0.1, so if the variation increased, productivity would be controlled by height. increased, productivity would be controlled by height. Table 6. Time to complete erection work. Table 6. Time to complete erection work.

Results Base Case Level 1 Level 2 Level 3 The number of Column: 32 1 Beam: Beam: Results Base 18 Case Beam: Level Level45 2 Level 45 3 erecting materials The number of The level of erecting Column: 18 Beam: 32 Beam: 45 Beam: 45 1.2 1.1 1.0 0.9 erecting materials height The level erecting height Around 1.2 1.1 1.0 0.9 Time forof completion Around Around Around Time for completion Around Around Around (in Simulation) 80 s 130 s 190 s 190 s Around 80 s 130 s 190 s 190 s Time(in forSimulation) completion 480 780 1140 1140 Time for(min) completion (min) 480 780 1140 1140 Time Timefor for completion completion (h) 6.0 13.0 19.0 19.0 6.0 13.0 19.0 19.0 (h) 1 s in simulation = 6 min in real time.

Level 4

Level 5

Level 6

Beam: Level 435

Beam: Level 525

Beam: Level 64

Beam: 35

1.2

1.2 Around Around 150 s 150 s

900

Beam: 25

1.3

Beam: 4

1.3 Around 98 s

1.3 Around 20 s

Around 98 s

588

1.3

Around 20 s

120

900

588

120

15.0

9.8

2.0

15.0

9.8

2.0

1 s in simulation = 6 min in real time.

6. Sensitivity Analysis 6. Sensitivity Analysis This section discusses the sensitivity analysis that tested the model’s reaction to changes in This section discusses the sensitivity analysis that tested the model’s reaction to changes in the the input parameters. We analyzed four critical factors: the amount of erection materials, stockyard input parameters. We analyzed four critical factors: the amount of erection materials, stockyard capacity, height of the erection, and availability of a tower crane. Though there were a total of six factors capacity, height of the erection, and availability of a tower crane. Though there were a total of six found to be critical, workable days and workers’ skill could not be defined or assumed before the factors found to be critical, workable days and workers’ skill could not be defined or assumed before simulation. These two factors can only be obtained by extrapolation from simulation results and/or the simulation. These two factors can only be obtained by extrapolation from simulation results analysis of previous research. Among the four impact factors, we analyzed the amount of erection and/or analysis of previous research. Among the four impact factors, we analyzed the amount of materials capacity of the stockyard theybecause could bethey modified involved in the erectionand materials and capacity of thebecause stockyard could by be individuals modified by individuals construction project participants. For example, materials could be reduced by the strategic structural involved in the construction project participants. For example, materials could be reduced by the design of a structural structuraldesign designer, the capacity of a and stockyard could of beaadjusted bycould a strategic floor or strategic of aand structural designer, the capacity stockyard be adjusted path plan. Our sensitivity analysis considered these point of flexibility. The setup and results this by a strategic floor or path plan. Our sensitivity analysis considered these point of flexibility.of The analysis are presented in Table 7 and visualized in Figure 23. setup and results of this analysis are presented in Table 7 and visualized in Figure 23. Traditionally, construction managers havehave accepted that productivity decreases when when the amount Traditionally, construction managers accepted that productivity decreases the of amount erectionof materials However,However, accordingaccording to our sensitivity analysis, analysis, productivity decreased erection increases. materials increases. to our sensitivity productivity but then recovered the amount of erection On the other decreased but thenwhen recovered when the amount materials of erectionregularly materialsincreased. regularly increased. On hand, the when amount was disturb productivity, decreased along with the along extension the otherthe hand, when theenough amounttowas enough to disturbitproductivity, it decreased withofthe extension of the capacity ofWith the stockyard. regards toplanning, construction this be analysis canto capacity of the stockyard. regards toWith construction thisplanning, analysis can applied be applied determine the proper amount of erection sizeyard of the stock yard needed. determine thetoproper amount of erection materials and materials size of theand stock needed.

Sustainability2016, 2016,8, 8,1106 1106 Sustainability

22of of25 25 22

Table 7. Sensitivity analysis setup and results. Table 7. Sensitivity analysis setup and results. The Number of Erecting Materials of Erecting Materials Productivity (↘) 12 The Number 14 16 18 20 22 0.258 0.237 0.235 0.250 0.276 0.30724 Productivity (&) 17 12 14 16 18 20 22 34 0.258 0.258 0.237 0.231 0.233 17 0.237 0.235 0.250 0.276 0.235 0.307 0.239 0.339 34 0.237 0.231 0.233 0.235 0.235 0.239 0.239 0.238 51 0.258 0.259 0.232 0.235 0.237 51 0.259 0.232 0.235 0.237 0.235 0.239 0.236 The capacity of stockyard 68 0.257 0.232 0.235 0.237 0.235 0.237 The capacity 68 0.257 0.232 0.235 0.237 0.235 0.237 0.236 of stockyard 85 0.258 0.258 0.236 0.235 0.234 0.235 0.239 85 0.236 0.235 0.234 0.235 0.239 0.235 102 0.258 0.258 0.236 0.238 0.003 102 0.236 0.238 0.003 0.235 0.235 0.237 0.237 0.235 119 0.232 0.235 0.237 0.238 0.238 0.240 0.240 0.236 119 0.258 0.258 0.232 0.235 0.237

24 0.339 0.238 0.236 0.236 0.235 0.235 0.236

Figure23. 23.Visual Visualresults resultsof ofsensitivity sensitivityanalysis. analysis. Figure

7. 7.Conclusions Conclusions By Byexecuting executingthe theBIM-integrated BIM-integratedsimulation, simulation, we wewere wereable ableto toobtain obtainaadynamic dynamicproductivity productivityplan plan and and calculate calculate the the project’s project’s per-hour per-hour rate rate of of production. production. Compared Compared to to the the traditional traditional construction construction planning planning method, method,this this system systemestablished establishedaareliable reliableconstruction constructionplan planfrom fromaabottom-up bottom-upapproach, approach, meaning meaning that that we we considered considered operational operational work work and and the the relationships relationshipsamong amongthe the various variousactivities, activities, activities and resources, and resources themselves. This activity-based construction planning activities and resources, and resources themselves. This activity-based construction planningmethod method improves improvesreliability reliabilitybecause becausethe theresults resultsare are much much closer closer to to the the actual actual schedule schedule and and the the construction construction plan plan can can be be adjusted adjusted whenever whenever the the BIM BIM model model is is modified. modified. Also, Also, when when delays delays occur occur or or resources resources change, using this BIM-integrated simulation framework is more flexible because the work is change, using this BIM-integrated simulation framework is more flexible because the work isbroken broken down downinto into more more detailed detailed activities. activities. This This research research presented presented the the development development and and application application of of the the BIM-integrated BIM-integrated simulation simulation framework for reliable construction planning. From this framework, the operational level framework for reliable construction planning. From this framework, the operational level productivity productivity is measurable, canwhole-project be used for whole-project planning. While has is measurable, which meanswhich it can means be usedit for planning. While BIM has beenBIM used in been used in various construction fields ranging from predicting the energy performance of buildings various construction fields ranging from predicting the energy performance of buildings to analyzing to analyzing structural safety,isBIM usagelimited is relatively in the management construction management field. structural safety, BIM usage relatively in thelimited construction field. This research This research presented a method capable of utilization expandinginBIM’s utilization management in the construction presented a method capable of expanding BIM’s the construction domain. management domain. Also, since BIM information can now be reused by construction engineers Also, since BIM information can now be reused by construction engineers establishing construction establishing construction plans, the money, framework and effort be thatspent would otherwise plans, the framework saves time, andsaves efforttime, that money, would otherwise reproducing be spent reproducing construction project data. construction project data.


23 of 25

Furthermore, the developed framework presents a new perspective on construction planning by considering construction projects’ uniqueness and complexity. Whenever the BIM model is modified during the work’s progress, users can easily regenerate construction plans because BIM is integrated with the construction work simulation. Moreover, in the simulation environment, users can synthetically manipulate complex factors and resources. Since the current planning method uses historical data and heuristic adjustments, such plans depend on historical sample data and managers’ knowledge and skill. That is, these construction plans cannot ensure consistency or accuracy. However, our framework facilitates more reliable planning by using the quantitative BIM-integrated simulation method. The end result is more reliable construction plans. While the BIM-integrated simulation method created for this research is a technical breakthrough that represents a vast improvement in management efficiency, there were some limitations to this research. First, researchers should validate this framework by applying it to various test cases. Second, the framework should be employed in diverse types of construction beyond that of the steel erection work exampled here. Finally, the simulation results should be made visual via a BIM model; doing so would extend the applicability of this framework. Acknowledgments: This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (No. NRF-2016R1A2B4015977). Author Contributions: All authors read and approved this manuscript. All authors contributed to this research, discussed the results and implications, and were involved in the manuscript’s development during all stages. WoonSeong Jeong led the implementation process of the integration framework. Soowon Chang implemented the case study and its experiments. JeongWook Son provided the main idea for this research and developed the integration framework as a principal investigator. June-Seong Yi discussed the main idea of this research. Conflicts of Interest: The authors declare no conflict of interest

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13.

AbouRizk, S.; Mohamed, Y. Optimal construction project planning. In Proceedings of the Winter Simulation Conference 2002, San Diego, CA, USA, 8–11 December 2002. Mikulakova, E.; König, M.; Tauscher, E.; Beucke, K. Knowledge-based schedule generation and evaluation. Adv. Eng. Inf. 2010, 24, 389–403. [CrossRef] Cooke, B.; Williams, P. Construction Planning, Programming and Control; John Wiley & Sons: New York, NY, USA, 2013. Lee, S.H.; Pena-Mora, F.; Park, M. Dynamic planning and control methodology for strategic and operational construction project management. Autom. Construct. 2006, 15, 84–97. [CrossRef] Alvanchi, A.; Lee, S.; AbouRizk, S. Dynamics of working hours in construction. J. Construct. Eng. Manag. 2011, 138, 66–77. [CrossRef] Mao, X.; Zhang, X. Construction process reengineering by integrating lean principles and computer simulation techniques. J. Construct. Eng. Manag. 2008, 134, 371–381. [CrossRef] Tucker, R.L. Management of construction productivity. J. Manag. Eng. 1986, 2, 148–156. [CrossRef] Hao, Q.; Shen, W.; Xue, Y.; Wang, S. Task network-based project dynamic scheduling and schedule coordination. Adv. Eng. Inf. 2010, 24, 417–427. [CrossRef] AbouRizk, S. Role of simulation in construction engineering and management. J. Construct. Eng. Manag. 2010, 136, 1140–1153. [CrossRef] Eastman, C.; Teicholz, P.; Sacks, R. BIM Handbook: A Guide to Building Information Modeling for Owners, Managers, Designers, Engineers and Contractors; John Wiley & Sons: New York, NY, USA, 2011. National Institute of Building Sciences. National Building Information Model Standard. Available online: https://www.nationalbimstandard.org (accessed on 11 October 2016). Berner, F.; Kochkine, V.; Habenicht, I.; Spieckermann, S.; Väth, C. Simulation in manufacturing planning of buildings. In Proceedings of the 2013 Winter Simulation Conference on Simulation: Making Decisions in a Complex World; IEEE Press: Piscataway, NJ, USA, 2013; pp. 3306–3317. Sargent, R.G. Verification and Validation of Simulation Models. Available online: http://www.informs-sim. org/wsc11papers/016.pdf (accessed on 11 October 2016).


14.

15.

16. 17. 18.

19.

20. 21.

22. 23.

24. 25.

26. 27. 28. 29. 30. 31. 32. 33. 34.

35. 36.

24 of 25

Rabe, M.; Spieckermann, S.; Wenzel, S. A new procedure model for verification and validation in production and logistics simulation. In Proceedings of the 40th Conference on Winter Simulation; IEEE Press: Piscataway, NJ, USA, 2008; pp. 1717–1726. König, M.; Beißert, U.; Steinhauer, D.; Bargstädt, H.-J. Constraint-based Simulation of Outfitting Processes in Shipbuilding and Civil Engineering. Available online: http://www.simofit.com/Paper_EUROSIM_2007_ koenig_beissert_steinhauer_bargstaedt.pdf (accessed on 11 October 2016). Fu, J. A Microscopic Simulation Model for Earthmoving Operations. Available online: http://transp-or.epfl. ch/heart/2012/latsis2012_submission_27.pdf (accessed on 11 October 2016). Hajjar, D.; AbouRizk, S.M. Unified modeling methodology for construction simulation. J. Construct. Eng. Manag. 2002, 128, 174–185. [CrossRef] Wu, I.-C.; Borrmann, A.; Beißert, U.; König, M.; Rank, E. Bridge construction schedule generation with pattern-based construction methods and constraint-based simulation. Adv. Eng. Inf. 2010, 24, 379–388. [CrossRef] Wang, W.-C.; Weng, S.-W.; Wang, S.-H.; Chen, C.-Y. Integrating building information models with construction process simulations for project scheduling support. Autom. Construct. 2014, 37, 68–80. [CrossRef] Halpin, D.W. CYCLONE-method for modeling job site processes. J. Construct. Div. 1977, 103, 489–499. Zayed, T.M.; Halpin, D.W. Construction engineering I: Simulation as a tool for resource management. In Proceedings of the 32nd Conference on Winter Simulation; Society for Computer Simulation International: San Diego, CA, USA, 2000; pp. 1897–1906. AbouRizk, S.M.; Hajjar, D. A framework for applying simulation in construction. Can. J. Civ. Eng. 1998, 25, 604–617. [CrossRef] Xu, J.; AbouRizk, S. Product-based model representation for integrating 3D CAD with computer simulation. In Proceedings of the IEEE 1999 Winter on Simulation Conference Proceedings, Phoenix, AZ, USA, 5–8 December 1999; IEEE Press: Piscataway, NJ, USA; pp. 971–977. Kamat, V.R.; Martinez, J.C. Validating complex construction simulation models using 3D visualization. Syst. Anal. Model. Simul. 2003, 43, 455–467. [CrossRef] Konig, M.; Koch, C.; Habenicht, I.; Spieckermann, S. Intelligent BIM-based construction scheduling using discrete event simulation. In Proceedings of the IEEE 2012 Winter Simulation Conference (WSC), Berlin, Germany, 9–12 December 2012; IEEE Press: Piscataway, NJ, USA; pp. 1–12. Howell, G.A. What is lean construction 1999. In Proceedings of the IGLC-7, Berkeley, CA, USA, 26–28 July 1999; pp. 1–10. Koskela, L.; Howell, G.; Ballard, G.; Tommelein, I. The foundations of lean construction. In Design and Construction: Building in Value; Butterworth-Heinemann: Oxford, UK, 2002; pp. 211–226. Sacks, R.; Radosavljevic, M.; Barak, R. Requirements for building information modeling based lean production management systems for construction. Autom. Construct. 2010, 19, 641–655. [CrossRef] Melles, B. What do We Mean by Lean Production in Construction? Taylor & Francis: Oxfordshire, UK, 1997. Shingo, S. Non-Stock Production: The Shingo System of Continuous Improvement; CRC Press: Boca Raton, FL, USA, 1988. Tommelein, I.D.; Riley, D.R.; Howell, G.A. Parade game: Impact of work flow variability on trade performance. J. Construct. Eng. Manag. 1999, 125, 304–310. [CrossRef] Tommelein, I.D.; Weissenberger, M. More just-in-time: Location of buffers in structural steel supply and construction processes. In Proceedings of the IGLC-7, Berkeley, CA, USA, 26–28 July 1999; pp. 109–120. Radosavljevi´c, M.; Horner, R.M.W. The evidence of complex variability in construction labour productivity. Construct. Manag. Econ. 2002, 20, 3–12. [CrossRef] Yan, W.; Clayton, M.; Haberl, J.; Jeong, W.; Kim, J.; Kota, S.; Bermudez Alcocer, J.; Dixit, M. Interfacing BIM with Building Thermal and Daylighting Modeling. Available online: http://www.ibpsa.org/proceedings/ BS2013/p_1374.pdf (accessed on 11 October 2016). Lee, G.; Sacks, R.; Eastman, C. Specifying parametric building object behavior (BOB) for a building information modeling system. Autom. Construct. 2006, 15, 758–776. [CrossRef] Ku, K.; Taiebat, M. BIM experiences and expectations: The constructors’ perspective. Int. J. Construct. Educ. Res. 2011, 7, 175–197. [CrossRef]


37.

38. 39.

40. 41. 42. 43.

25 of 25

Barazzetti, L.; Banfi, F.; Brumana, R.; Gusmeroli, G.; Previtali, M.; Schiantarelli, G. Cloud-to-BIM-to-FEM: Structural simulation with accurate historic BIM from laser scans. Simul. Model. Pract. Theory 2015, 57, 71–87. [CrossRef] Azhar, S.; Brown, J. BIM for sustainability analyses. Int. J. Construct. Educ. Res. 2009, 5, 276–292. [CrossRef] Hu, Z.; Zhang, J. BIM-and 4D-based integrated solution of analysis and management for conflicts and structural safety problems during construction: 2. Development and site trials. Autom. Construct. 2011, 20, 167–180. [CrossRef] Forbes, L.H.; Ahmed, S.M. Modern Construction: Lean Project Delivery and Integrated Practices; CRC Press: Boca Raton, FL, USA, 2010. AbouRizk, S.M.; Halpin, D.W.; Wilson, J.R. Fitting beta distributions based on sample data. J. Construct. Eng. Manag. 1994, 120, 288–305. [CrossRef] Moder, J.J.; Phillips, C.R.; Davis, E.W. Project Management with CPM, PERT and Precedence Diagramming; Van Nostrand Reinhold Company: New York, NY, USA, 1983. Hong, T.; Hastak, M. Simulation study on construction process of FRP bridge deck panels. Autom. Construct. 2007, 16, 620–631. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).