By Martin Fischer, Kathleen McKinney Liston, and the Paperless Design Project Team at Walt Disney Imagineering

This document outlines the functionality needed for a 4D environment to serve the needs of the Paperless Design Project at Walt Disney Imagineering (WDI). This report was originally written for Paperless Design Team members and Research and Development personnel at WDI to understand our "wish-list" of functionality needed to support 4D environments. This document has been edited to reflect a more general perspective. Ultimately, this document should serve as a roadmap for the development of next generation CAD, project management, and concurrent engineering tools. We do not envision that all the functionality discussed below needs to be provided in one product, but any products used in this process should provide the "hooks" to enable the representation and linking of the information required for 4D modeling.

The report is organized into the following sections:

1. Report Summary
2. Vision of 4D Technologies in AEC Practice
3. Paperless Design Project Goals and Process for Evaluating 4D Technologies
4. Description of 4D Modeling Tasks
5. Functionality Required to Perform 4D Modeling Tasks
6. Implementation Challenges
7. Appendix: Purpose of 4D Models and Modeling Issues

We welcome any comments or feedback on this report.

• manage and minimize risk throughout all stages of a construction project
• effectively communicate the design, schedule, and other project data
• rapidly explore design and construction alternatives

• the use of these technologies to support the design and planning of construction projects within the WDI organization
• the architecture and functionality of next-generation 3D and 4D technologies

This report describes the results of the initial evaluation of the use of 3D/4D technologies for the planning of a WDI construction project. Some of the specific benefits of the 3D/4D modeling process realized to date are:

• clear communication of the sequence of construction to all project participants
• improved understanding of design and schedule
• ability to receive feedback from a broad range of project stakeholders

The limitations of current 4D technologies are:

• slow generation of alternatives
• no computer-based risk analysis
• minimal integration with other project data

Based on this initial evaluation, the Paperless Design Team has compiled an initial "wish-list" of functionality that is needed to address these limitations and support long-term project goals. This functionality wish-list is organized into three areas and are summarized:

• Functionality needed to Generate and Manipulate Model-Based Project Data, specifically functionality to:
• generate and manipulate object-models of the construction project
• generate and manipulate associations between those objects
• take apart and re-organize the project model
• Functionality needed to Represent Model-Based Project Data, specifically functionality to:
  • represent the 4D model as a set of objects
  • maintain multiple representations of a project
• Functionality needed to Visualize Model-Based Project Data, specifically functionality to:
  • view the object-models in a real-time 3D environment
  • view project data and relationships between project data within the 3D environment
  • provide a "4D viewer" accessible to all project participants

The remainder of this report elaborates this wish-list and the rationale for needing this functionality. We first describe the vision of 4D technologies and summarize the 4D modeling process performed during the Paperless Design Project. We then explain how planners perform 4D modeling tasks today, i.e., how they generate, update and maintain, and visualize and distribute 4D models, and outline how we envision planners will perform 4D modeling tasks in the future. We then discuss the functionality that is needed in a next-generation 4D environment to support these tasks. Finally, we conclude with the challenges we face to implement this functionality.

Planners, designers, and engineers will use 4D technologies to analyze and visualize many aspects of a construction project, from the 3D design of a project to the sequence of construction to the relationships between schedule, cost and resource availability data. Planners will create, update and maintain, and deliver a 4D object model throughout a design/construction project. The 4D technologies will enable planners to generate various views of this 4D object model to clearly communicate the spatial and temporal aspects of construction schedules to all project participants. 4D technologies will no longer be used to simply animate the sequence of construction but will be used to communicate a wide range of project data much more clearly and efficiently than possible today. Planners, designers, and engineers will use 4D environments to visually relate data much like the way engineers use gradated color 3D models to visualize the stresses on structures.

This vision implies that 4D technologies will provide a view of a project database. This project database will store and maintain the representation of building components and construction activities, their inter-relationships, and relationships to other project data. This database will be designed to support concurrent engineering of the facility and its delivery process by supporting multiple representations of project data, multiple ways of organizing and relating the data, domain-specific views of the data, and standard representations of these data so that multiple applications can interpret the data. This project database will not only support the use of 4D technologies throughout the project life-cycle but also support other design and construction technologies (Figure 1).

Figure 1: Vision of Project Database showing a shared database that stores project components that are generated, accessed, and organized into multiple project models and views

To fully maximize the benefits of 4D technologies, planners, designers, and engineers will need to change how they produce project data. They will focus their efforts on producing and evaluating "models" instead of drawings. To build 4D models that accompany a project throughout its design, planning, and construction phases, users need to be able to take these 3D models apart and put them back together quickly and in very flexible, interactive ways. They need to represent the design not only as graphical drawing objects, but also as intelligent and open information objects that can be linked to other project information. Today’s CAD tools are mainly focused on the production of construction drawings and do not offer the functionality needed for 4D modeling. They do not offer an intuitive and easy-to-use object-oriented environment that allows users to build a graphical and informational model of the project that can be taken apart and put back together in new ways, and that can be shared easily with the various disciplines involved in the project life cycle.

The goal of the first phase of the Paperless Design Project was to understand the issues with respect to using 4D technologies and to propose a plan for implementing 4D technologies in the short and long-term. This includes specifying the contents of a project database from the 4D perspective and specifying the functionality for generating and accessing the content in the project database (Figure 2). Future projects within R&D at WDI will consider the structure and contents of the project database from other perspectives. Much of the functionality needed to support 4D technologies, however, are needed to support other design technologies.

Figure 2: Focus of Paperless Design Project

The Paperless Design Project Team selected a construction project that presented potential construction problems due to the complexity of the site. The project team included representatives from the design team, construction team, virtual reality group, 4D consultants, and CAD analysts. The goal of the team was to produce a 3D and 4D model of the selected construction project for visualization in a Virtual Reality Cave, on the desktop, and in a web-browser (Figure 3).

During the three-month project, commercial and prototype technologies were used. The 3D models were produced by modelers in the VR group who were most proficient with Alias Wavefront and Multigen. These tools also enabled the most efficient transfer of the 3D model into the Cave environment.

Following is a brief description of the various 4D modeling processes performed.

1. Production of 4D model using commercial tools
2. Generation of a 4D model using commercial tools and customized add-ons
3. Generation of a 4D model using prototype tools

Figure 3: Inputs, Outputs, Controls, and Tools for Paperless Design Project

3.1 Production of 4D model using commercial tools

The first 4D modeling process utilized commercially available software: Alias Wavefront, AutoCAD, Primavera, and Jacobus Schedule Simulator. This process involved the following steps (See Figure 4):

1. Generating the 3D model in Alias Wavefront
2. Breaking the 3D model into twenty-three "chunks" to represent the construction breakdown of the project
3. Exporting these chunks from Alias as DXF files
4. Importing the twenty-three DXF files into AutoCAD
5. Organizing the components in each file into two layers. One layer represented the structure and another layer represented the spaces in between the structure and was used to represent the "torquing of the structure."
6. Merging all of the DWG files into one drawing
7. Exporting the single DWG file as a "JSM" file or Jacobus 3D model file.
8. Importing the JSM file into Jacobus
9. Importing the schedule file (generated in Primavera) into Jacobus
10. Linking the CAD components to the schedule activities to generate the 4D model
11. Playing the 4D model within the Jacobus environment

This process was tedious and presented the following problems:

• project participants could only view the 4D model within the Jacobus environment, unless screen captures were made of each day or specified period/time.
• the 4D model information could not be transferred to the cave environment
• most updates to the 4D model required starting at step 1

Figure 4: 4D Modeling Process with Commercial Tools

3.2 Generation of a 4D model using commercial tools and customized add-ons

The second process utilized commercial tools with add-ons. These customizations were:

• a visual-basic macro in Excel. This macro creates a dialog box that lists construction activities and lists 3D components (from 3D Studio Max) and enables the planner to manually link one by one the activities and components or automatically link components and activities.
• a set of MaxScripts in 3D Studio Max. These maxscripts take information sent from Excel and create animation information. For example, a message is sent from Excel to set the start and finish dates of a particular component. The maxscript creates animation keys to turn on the visibility of the component at the start date and animation keys to change material attributes of the component to reflect its state of construction, e.g., making the component red for the duration of its construction.

This process involved (Figure 5 and detailed here):

1. Exporting the AutoCAD file created in step 4 of the first process to 3D Studio Max
2. Exporting the schedule information from Primavera to Excel
3. Manually listing the name of the CAD component associated with each activity in the Excel spreadsheet
4. Running the macro within Excel to import the list of 3D components from 3D Studio Max and generate a list of the construction activities
5. Linking the activities to 3D components
6. Rendering the 4D visualization in 3D Studio Max
7. Exporting the Excel data in ASCII format to the CAVE

This process had the following benefits:

• the 4D visualization was fully-rendered
• all project participants could view the 4D visualization because the movie file could be posted on-line in Quicktime or AVI format

and the following problems:

• any changes that had to be made required performing the entire set of 4D modeling tasks again (starting at Step 1, Process 1)
• the ouput was not fully interactive. Although planners could change the temporal state of the 4D visualization the planners could not interact with the spatial data, e.g., change the viewpoint of the visualization.

Figure 5: 4D Modeling Process with Commercial Tools+

3.3 Generation of a 4D model using prototype tools

The third process involved developing a set of 4D prototype tools. These included:

• a Java 4D Application. This application provides the functionality to import the names of CAD components and schedule information, link activities to CAD components, and save and update the 4D model.
• a Java 4D Applet that communicates with a VRML world in a web browser. This applet provides the functionality to interactively view a 4D visualization in a browser.
• a Java3D application. This application provides the functionality to generate, view, and store the 4D model in a desktop environment.
• Cave 4D functions. These functions provide the functionality to import the 4D associations exported from the Java 4D application and to use these associations to view the 4D visualization in the cave.

The 4D modeling process with these tools involves (Figure 6):

1. Importing the schedule data into the Java 4D application
2. Importing the CAD data into the Java 4D application
3. Organizing the CAD data into approriate groupings (if needed)
4. Linking the components to the activities
5. Assigning "action-types"
6. Exporting 4D model to CAVE and to Java 4D applet

The process had the following benefits:

• updates and changes could easily be made without re-performing the entire 4D modeling process
• re-grouping of CAD components is possible within the 4D environment
• on-line viewing is possible in the 4D environment

and the following problems (Figure 7):

• data is in ASCII format
• links to original data are not maintained
• schedule is not hierarchical

Figure 6: 4D Modeling Process with Prototype Tools

Figure 7: Diagram showing Current Process and Current Problems

4. Tasks to Generate, Visualize, Update, and Distribute 4D Models

The execution and evalution of these processes as well as our experience gained on other 4D modeling projects at Stanford University allowed us to develop a set of 4D modeling tasks and outline the corresponding functionaliity. These tasks and the related functionality are presented next.

In this section we describe the 4D modeling tasks planners can perform today with commercial 3D and 4D tools, contrast those with how we envision planners performing 4D modeling tasks, and describe the functionality 4D technologies need to provide to support these tasks. We have organized these taks into three main groups of tasks:

1. Tasks to Generate and Manipulate 4D Models
2. Tasks to Visualize and Evaluate 4D Models
3. Tasks to Deliver and Distribute 4D Models

4.1 Tasks to Generate 4D Models

Today, the generation of a 4D model involves three main tasks, which we elaborate in the following sections:

1. generating a 3D model that represents the spatial aspects of the construction project at the appropriate level of detail
2. generating a schedule
3. relating components in the 3D model with components in the schedule

4.1.1 Generating a 3D model

4.1.2 Generating a Schedule

4.1.3 Relating 3D model Components to Construction Activities

4.2 Tasks to Update and Maintain 4D Model

4.3 Tasks to Distribute and Visualize the 4D Model

5.0 Functionality

This vision of the kinds of tasks planners will be able to perform with 4D technologies requires functionality in three areas:

1. Interaction: Generating and Manipulating 4D models
2. Visualization: Viewing 4D models
3. Representation: Storing 4D models

5.1 Interaction Functionality

Next-generation 4D technologies will enable planners to generate 4D models in a more interactive way. Planners will continue to generate 4D models from existing 3D models and schedules as well as generate schedules interactively within the 4D environment or automatically generate a schedule from a 3D model. 4D tools need to support a variety of interactions with the 3D model content, the schedule content, and the relationships between that content and other construction project data. Specifically they need to provide the following types of interaction functionality:

(a) 3D model interactions:

• Rotate and move viewpoint(required because, as the project gets built in the 4D visualization, parts of the 3D model that will be built later might become obstructed by already constructed components)
• Zoom (required because a schedule and 3D model might be at different levels of detail for certain time frames, or the user might be interested in more or less detail for a particular time frame)
• Turn classes of objects on and off or make them transparent during a 4D simulation (this is important for buildings, e.g., where the installation of slabs and roofs obstructs the view of the construction activities that go on inside)

(b) Schedule interactions:

• Support the typical temporal sequence relationships (finish-start, start-start, finish-finish) between activities and lag times for the relationships.
• Change activity dates. For example, a user should be able to slide an activity bar and see the repercussions of changed start dates on other activities, milestones, and space needs.
• Adjust activity durations (required because a contractor might add or take away resources to adjust activity durations as necessary, or the initial production rate assumptions might have been too high or too low).
• Resequence activities (change the schedule logic in the bar chart, network or object view). This is required because materials might be delivered early or late, necessitating the development of new 4D alternatives that minimize the impact of this schedule change. The 4D model should help assess the spatial (e.g., availability of access and work space) and temporal implications of schedule changes, thus limiting the need for further schedule changes due to oversights when making the first schedule changes.
• Enter activity name. The tool should offer a default activity name based on the name(s) of the components.
• Propose default relationships for a new activity (e.g., to its neighboring activity of the same type, to the activity that acts on the components below).
• Support the specification of spatial constraints or relationships between work elements (e.g., specify that an activity should always be 100 m ahead of a succeeding activity).
• Link lag times to spatial areas to represent lag times as spatial constraints (e.g., concrete in a certain area needs to cure, or no other activity should be within 50 m due to safety concerns). We envision that the user can again use the graphical models to specify these relationships or work with the (hierarchical) list of names.

(c) 4D model interactions

• Link several components to an activity.
• Link several activities to a component.
• Do not restrict the linking of activities and components to certain branches of the component and activity hierarchies (if those exist), e.g., it must be possible to link two components from different branches of a hierarchy to an activity.
• Capture the rationale for the 4D associations, either through manual processes or intelligent construction "agents."
• Allow the user to enter other information for the newly created 4D object (an activity linked to a 3D component), such as information about the reason for the association
• Regroup the activities and 3D components. Initial assumptions about weekly or daily progress might be off, now requiring or allowing a smaller or larger buffer or lag between activities in a work sequence, which in turn requires a new grouping of activities and 3D entities, since the scope for the activities has changed. The software needs to highlight the physical scope of work for an activity or a set of activities in the 3D model and allow the user to change the associations between 3D components and activities in the object or graphic view.
• Switch between and display several levels of detail. This requires that the graphical views of the schedule and CAD models adjust themselves depending on the level of detail chosen by the user.
• Attach other types of data, e.g., cost for activities, and corresponding visualization mechanisms, e.g., S-curves that can be displayed as the 4D model is running.

5.2 Visualization Functionality

One of the key parts of our vision is that planners will be able to visualize many aspects of the construction project within a 4D environment. In the following table we list the types of content that planners need to visualize and the correlating functionality that 4D technologies need to provide to support the visualization of that content. We also list some of the challenges in implementing this functionality.

Table 1. Examples of additional information a 4D model could display.

6.0 Implementation Challenges

Implementing this wish-list of functionality is a challenge due to the current make-up of software in the AEC industry. There are no standard representations of 3D CAD components, construction activities, and no standard communication protocols between programs. Thus, at a functional level, creating relationships between data in a useful and accessible way is a formidable task. At a conceptual level, the conceptual framework of these tools also precludes implementing robust 4D technologies that support large complex projects. In this section we highlight some of these functional and conceptual limitations.

Functional Challenges due to Current Make-Up of AEC Software

To overcome these limitations, there might be work-arounds or options already available today, which are unknown to us. In some areas, especially when we make reference to the new Architectural Desktop software, the discussion is based on a very limited understanding on our side of the available functionality, and we are offering our thoughts based on the demonstration we saw at Autodesk on August 21, 1998 and our ten years of experience with object-oriented software.

1. The current representation of building components is geometry-based, inaccessible, and not customizable.

Today, the only way to access "objects" in AutoCAD is through ARX or VB. This route allows us to access limited information about objects in the CAD model and to create new objects with attribute information. Unfortunately, these options do not enable the user to create a true 4D object model since the relationships between CAD objects in AutoCAD and activity objects must be generated and stored in a separate application. For example, standard building component classes can be extended as well as new ones created. However, the current object model in AutoCAD focuses on geometric attributes. It provides insufficient support to define relationships between components and activities, and it is not straightforward to access all the information about an object. In particular we need to be able to create and access methods that enable us to relate these components to construction activities. In summary, we need object-oriented representations of components that are customizable, editable, and accessible.

2. Models are not hierarchical.

AutoCAD allows the designer to create one breakdown of the geometric information into a set of layers. As mentioned throughout this document, this one breakdown is not sufficient to support the multiple perspectives on the 3D model necessary for 4D modeling. Furthermore, the layers give the user only an "all or nothing" option to display geometric information. Users do not always want to see all the detail, and they need to be able to reorganize the information that is shown and that can be selected and grouped. A hierarchical model would enable users to work at the appropriate level of detail, i.e., instead of either seeing all or none of the information they could now select the desired level of detail. Users should not be restricted to work at the same level of detail for all parts of the 3D model. For example, they should be able to work at the most detailed level for the foundations in a certain area of the project and view the other foundations and the structural system with less detail. This would allow them to see the context of their current work without their desktop being cluttered with unnecessary graphical objects. The "drawing methods" that are part of objects in Architectural Desktop allow a user to customize the display of geometric information to some degree. However, because the detail is generated by methods that are part of a particular object it can become difficult to generate views and groupings of the detail from a perspective other than the parent object. For example, a footing object might have a method that displays the footing as a simple 3D box in one view and with the reinforcement and formwork detail in another view. While this enables a user to show detail in some areas and hide it in others it makes it difficult to collect and group all of the footing reinforcement between gridlines A and G for association with a construction activity. In the example, the user wanted to switch the view of the information from a footing perspective to a reinforcement perspective. A hierarchical, fully object-oriented decomposition of the 3D objects would give users more flexibility to create, interact, and manage information about a project.

3. Views of models are mainly graphical.

In addition to the graphical view of the model, users should be able to navigate the component hierarchies easily. They should be able to inspect the graphic and information content of a CAD model and its components. They need to be able to browse the hierarchy and click on an object or group of objects and see the graphic model and bring up a list of all the objects’ attributes and relationships. Architectural Desktop begins to address this limitation.

4. It is difficult to convert AutoCAD files to a real-time 3D environment.

Currently, text files are the only option to export an AutoCAD model to another application. Today, we need to export the 3D AutoCAD model to other applications because of limitation (1) and because of the need to work in an environment that renders objects in real-time. It is difficult to export more than the graphic geometry, which then makes it extremely cumbersome to manipulate the 3D model for 4D modeling. An open and more accessible hierarchical object model is needed that facilitates the exchange of the graphic and information content of a 3D model and its components. The hierarchy and all other information needs to be maintained when the files are exported to other applications.

• interaction "metaphors" to interact with 4D models (e.g., for red-lining)
• development of a general tool set that supports project-specific work

• who owns the project database and who develops the project database
• how to utilize and extend current industry standard efforts such as STEP and IAI

• whether the display needs to happen as the 4D simulation/visualization progresses
• whether an asynchronous display is feasible and acceptable
• to what extent the display should be user-selected or system-driven
• large displays vs. desktop displays

Given the time pressures of many projects today and the limited space typically available on site, the overall goal for the 4D modeling process is to create a dynamic 4D model that allows project management to resequence activities rapidly when necessary while reducing the risk of running out of space. Hence the 4D model should allow users to play with and respond to scenarios like the ones listed below. These are examples of decisions a 4D model should support and insights it should provide.

• Communicate schedules clearly to all stakeholders to ensure that everyone is on the same page and to solicit everybody's input in a timely fashion.
• Try different activity sequences to lower risks of delays and unproductive work.
• Show relationships between on-site and off-site production and make sure that on-site production needs are communicated clearly to off-site fabrication centers.
• Respond to availability problems of pre-assembled steel components. If supply of pre-assembled components differs from that required in the schedule activities need to be rescheduled to understand the ramifications on space use.
• Determine overlap possible and necessary for crews and work. E.g., show the relationship between foundation work and structural work. How far ahead of the vertical steel erection should the footing and steel column work be so that the crews and equipment do not hinder each other and affect the work in adjacent areas? Issues to consider are the required and achievable early concrete strength of concrete footings which determines curing time for footings and the minimum lag time to vertical steel erection, quantity of vertical steel to be erected in each area before torquing and bolting, access needs for other work, etc.
• Ensure that laydown areas will be available close to the location offinal installation for each discipline without hindering the work of a discipline and other work.

• It is important to agree on the appropriate time step (i.e., should the schedule simulation show construction month by month or week by week) so that the schedule and CAD models are prepared at the right level of detail.
• Determine a time frame or a project area where construction is likely to be particularly complex and difficult to schedule. Identify the worries of project managers. This will help in determining the level of detail needed in the CAD and schedule models. With today’s CAD tools it is far easier to combine detail than to add detail. Hence our recommendation at this point is to err on the side of too much detai for initial model building.
• Discuss level of visual detail in model. For construction simulation, the CAD model does not necessarily have to contain a lot of visual detail. Hence, the 3D CAD model can often be simplified in comparison to an architectural 3D model if the sole purpose of the model is to support a 4D modeling effort.