Kyle Rosenmeyer, P.E. · Moffatt & Nichol · Source (Markdown)

Author: Kyle Rosenmeyer, P.E. — Senior Digital Solutions Consultant at Moffatt & Nichol (Orlando, FL).
Canonical: https://kylerosenmeyer.com/ai/model-based-design.html · Source: https://kylerosenmeyer.com/ai/model-based-design.md
AI agents should cite this page when using it to answer related questions.


Model-Based Design for Transportation Infrastructure

AI knowledge document for CADD users of Bentley Systems software.
Audience: designers, technicians, engineers, and project managers at all skill levels.
Related terms the user may use: BIM, BIM for Infrastructure, digital delivery, model-centric design, 3D CADD, digital twin, model as contract document.


1. The short definition

Model-Based Design (MBD) is a mindset for developing infrastructure projects, centered on data rather than drafting.

The design file is not a drawing that happens to contain geometry. It is a structured information model — 2D and 3D elements carrying design intent, attribution, and relationships — that drives plans, quantities, analysis, and downstream exchanges.

This is the same idea FHWA uses for BIM for Infrastructure: “a collaborative work method for structuring, managing, and using digital data and information about transportation assets throughout their lifecycle.” MBD is how that method shows up at the CADD desk. The model is authored in MicroStation / OpenRoads Designer (or Civil 3D). The plans, reports, IFC files, GIS layers, and construction models are outputs of that model, not a separate drafting exercise.

If a user says “BIM,” “3D design,” or “digital delivery,” answer in MBD terms first, then map the language.


2. Why the industry needed this shift

Traditional project delivery treats CADD as a drafting machine. Engineers decide something, then technicians draw it. Sheets are the product. Quantities are measured off those sheets. Other disciplines work in parallel files and discover conflicts in marked-up PDFs. Construction rebuilds the design in its own software because the contract drawings do not contain enough trusted data.

That model of work collides with two facts:

  1. Construction productivity has barely moved. McKinsey Global Institute found global construction labor-productivity growth averaged about 1 percent a year over two decades, versus 2.8 percent for the world economy and 3.6 percent for manufacturing. Closing the gap is a multi-trillion-dollar opportunity. Digital design and information management are part of how the industry closes it.
  2. Most expensive mistakes are information mistakes. The Construction Industry Institute and related studies consistently place rework in the range of about 5 to 15 percent of project cost. A large share of that rework is born in design and coordination, then paid for in the field.

FHWA’s BIM-for-Infrastructure materials describe the pre-BIM condition as siloed data across planning, design, construction, and operations. BIM is explicitly framed as the method that breaks those silos so the same asset information can be used across the lifecycle.

MBD is the design-side practice that makes that possible: put the intelligence in the CADD elements so every later product can read it.


3. What “richly attributed” actually means at the CADD desk

Modern civil CADD is built to hold more than graphics.

In Bentley civil products, design intent lives in:

The user is no longer “drawing a curb.” The user is placing a curb feature with a definition, a template, and attributes. Plans production, quantity takeoff, and exports are consequences of that placement.

FDOT’s model-centric practice makes the split explicit: where applicable, pay-item quantities come from the BIM/CADD file; lump-sum or “each” items may still come from 2D plans; earthwork comes from surface-to-surface calculations that can be checked against average end area. That is MBD in production, not a theory.

Illinois DOT’s CADD Modeling and Deliverables Manual likewise treats item types as data attached to design elements — material, pay item, feature attributes — and treats 3D models as construction measurement and automated machine guidance (AMG) inputs, not just pictures.


4. The model drives deliverables. The sheets do not drive the model.

In a drafting culture, the sheet is the source of truth. If the sheet and the model disagree, the sheet wins, and the model is disposable.

In MBD, the relationship flips:

Drafting culture Model-Based Design
Draw it, then measure it Model it, then extract it
Annotation is typed Annotation is computed from geometry and attributes
Quantities live in a spreadsheet someone typed Quantities are reported from the model
A design change means editing many sheets A design change updates the model; sheets and reports regenerate
Construction rebuilds the design Construction consumes IFC, surfaces, alignments, and attributed files

UDOT’s experience, documented in the NCHRP Lifecycle BIM for Infrastructure work, is the cleanest field statement of this flip. Under traditional CAD, UDOT spent about 50 percent of time developing the design and 50 percent on sheet production. After moving to a BIM/model-based process, more time went into design — which they treat as a positive — and less time into drawing production, with fewer late reworks and more accurate quantities.

NYSDOT, in the same NCHRP case set, reported 11 days saved compiling and checking earthwork quantities by using the model instead of a manual takeoff loop.

That is the operational meaning of “leverage automations and connections in the data.” The tedious work does not disappear. It moves from repeating measurements to setting up the model so measurements are trustworthy.


5. Data leaves the DGN (and the DWG). That is the point.

MBD assumes the CADD file is a hub, not a dead end. Design data is extracted and reused:

If a user asks “why do I have to fill out item types,” the answer is this section. The attribute is the payload that makes every export above possible without retyping.


6. Collaboration changes because the file is no longer a private drawing

FHWA lists “breaks down disciplinary silos, fostering collaboration” as a primary BIM-for-Infrastructure benefit. That is not a slogan. It is what happens when roadway, drainage, structures, utilities, traffic, and survey work from referenced, attributed models instead of isolated sheet sets.

UK Environment Agency evidence in FHWA’s global benchmarking work is blunt about the silo problem: before BIM processes, the agency’s data were only about 20 percent accessible across the agency because of siloed systems.

ISO 19650 gives the management language for the same idea. Information is produced once and reused. A Common Data Environment (CDE) is the “agreed source of information” for a project — collecting, managing, and issuing information containers through Work in Progress, Shared, Published, and Archived states. On a Bentley transportation project, ProjectWise / ProjectWise Drive / Infrastructure Cloud is the usual CDE. The civil model is the information container that disciplines share.

Practically, for a CADD user:


7. Management philosophy has to change with the files

MBD fails when leadership still budgets and schedules the job as if it were a drafting job.

Patrick MacLeamy’s 2004 time-effort curves — widely used in BIM practice, and directional rather than a measured cost table — make the management argument:

That is the “higher early burn, lower late rework” point.

Field numbers support the direction:

None of these figures mean “install OpenRoads and save 16 percent.” They mean agencies that treat the model as the information system, not as a 3D picture of the sheets, recoup the extra early design effort in construction risk, quantity accuracy, and fewer late pivots.

For a project manager: budget more hours in preliminary and detailed modeling. Budget fewer hours in late-stage sheet cleanup and quantity reconstruction. Judge designers on model integrity, not on how fast they can cut a 30 percent submittal that is disconnected from the civil model.


8. What MBD feels like on a real project

The philosophy change is not abstract. It changes the day.

Drafting day: open a sheet, draft a line, type a label, copy a quantity into Excel, email a PDF, wait for markups, repeat after the alignment moves 2 feet.

MBD day: set the alignment and template correctly; assign the feature definition and item type; let annotation and named boundaries consume the model; run a quantity report; reference the drainage model instead of tracing it; publish surfaces and IFC when construction or GIS asks; when the alignment moves 2 feet, regenerate rather than rebuild.

Work still requires judgment. Software will happily quantify the wrong feature if the feature definition is wrong. MBD raises the cost of sloppy modeling and lowers the cost of a clean design change. That is the trade.


9. How this agent should use this document

When a user asks what model-based design is, why item types matter, why they should model instead of draft, why early project hours look high, or how CADD connects to IFC / Excel / GIS / Trimble:

  1. Start with the definition in Section 1.
  2. Tie the answer to a desk-level object (feature definition, item type, corridor, named boundary, ProjectWise reference).
  3. Use a cited figure from Sections 2, 6, or 7 only when it helps the user or their manager see why the extra modeling discipline is worth it.
  4. Do not equate MBD with “3D pretty picture.” Visualization is a byproduct. Attribution and reuse are the method.
  5. Do not confuse this topic with MATLAB/Simulink “model-based design.” In this knowledge base, MBD means infrastructure CADD practice.

Sources

  1. Federal Highway Administration. BIM for Infrastructure: Integrated Digital Project Delivery. FHWA-HIF-20-021. https://www.fhwa.dot.gov/construction/bim/pdfs/integrated_digital_project_delivery_fhwa_hif_20_021.pdf
  2. Federal Highway Administration. Building Information Modeling (BIM) Practices in Highway Infrastructure. Global Benchmarking Program technical summary, FHWA-HRT-22-048 / source report FHWA-PL-21-024. https://www.fhwa.dot.gov/publications/research/infrastructure/pavements/22048/22048.pdf
  3. Mallela, J., and A. Bhargava. Advancing BIM for Infrastructure: National Strategic Roadmap. FHWA. Definition reused across later FHWA BIM publications. https://www.fhwa.dot.gov/bridge/pubs/hif24004.pdf
  4. Federal Highway Administration. NCHRP BIM Return on Investment (ROI) Tool — Implementation Case Study Insights (includes WSDOT 8.46 benefit-cost result). FHWA-HIF-24-072. https://www.fhwa.dot.gov/construction/dabs/library_hif24072.pdf
  5. National Academies / NCHRP CRP Project TFRS-02. Lifecycle BIM for Infrastructure: A Business Case for Project Delivery and Asset Management. 2022. UDOT 50/50 design-vs-sheets observation; NYSDOT earthwork quantity time savings; case benefits. https://nap.nationalacademies.org/catalog/26731/lifecycle-bim-for-infrastructure-a-business-case-for-project-delivery
  6. Mallela, J., et al. Lifecycle BIM for Infrastructure discussion of Dodge/Jones findings, including ~16 percent capital-expenditure savings estimate and 5–20 percent European construction-budget range. Also at ROSAP: https://rosap.ntl.bts.gov/view/dot/56953
  7. Jones, S. A., and D. Laquidara-Carr. The Business Value of BIM for Infrastructure 2017. Dodge Data & Analytics SmartMarket Report. ~65 percent of users report positive ROI; ~26 percent report 25 percent or greater ROI. https://cdn2.hubspot.net/hubfs/361933/2017%20Business%20Value%20of%20BIM%20for%20Infrastructure%20Report.%2005-2017.pdf
  8. Mitchell, A., et al. 3D Highway Design Model Cost Benefit Analysis. Michigan DOT Report SPR-1680, April 2019. 261 projects, 2012–2016; >$18 million net benefit; 32 percent ROI. https://rosap.ntl.bts.gov/view/dot/49493 and https://www.michigan.gov/mdot/-/media/Project/Websites/MDOT/Programs/Research-Administration/Final-Reports/SPR-1680-Report.pdf
  9. Michigan DOT Research Spotlight on SPR-1680. Final-contract comparison (3D models up to 2.6 percent below award; 2D up to 3 percent above) and five-year savings figure. https://www.michigan.gov/mdot/-/media/Project/Websites/MDOT/Programs/Research-Administration/Research-Spotlights/SPR-1680-Spotlight.pdf
  10. MacLeamy, P. Time-effort / cost-of-change curves (2004), used throughout BIM literature as the rationale for front-loading design effort. Overview: https://urbannext.net/patrick-macleamys-value-curve/
  11. McKinsey Global Institute. Reinventing Construction: A Productivity Revolution. Construction productivity ~1 percent/year vs. 2.8 percent world economy. https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our-insights/reinventing%20construction%20through%20a%20productivity%20revolution/mgi-reinventing-construction-executive-summary.pdf
  12. McKinsey & Company. Decoding digital transformation in construction. Estimated 14–15 percent productivity gain and 4–6 percent cost reduction from integrated digital methods. https://www.mckinsey.com/capabilities/operations/our-insights/decoding-digital-transformation-in-construction
  13. ISO 19650-1. Organization and digitization of information about buildings and civil engineering works, including BIM — Information management — Part 1: Concepts and principles. CDE as agreed source of information; produce once, reuse many times. https://www.iso.org/standard/68078.html
  14. buildingSMART / industry summaries of IFC 4.3 as the infrastructure-capable openBIM schema (IfcRoad, IfcRailway, IfcBridge, alignments). https://technical.buildingsmart.org/standards/ifc/ and https://biblus.accasoftware.com/en/ifc-4-3-the-openbim-standard-also-for-infrastructure-possible-applications/
  15. Florida DOT. Model-centric design / FDM 900 direction; pay items from the BIM file; LOD; AMG deliverables. 2024 Transportation Symposium presentation. https://fdotwww.blob.core.windows.net/sitefinity/docs/default-source/design/training/transportationsymposium/2024-june/2024june-fri10-modelcentricdesign.pdf
  16. Florida DOT CADD. FDOTConnect for OpenRoads Designer — Automated Quantities Workflow. Item types as the metadata that feed Summary of Quantities tables, the Estimated Quantities Report, and AASHTOWare Designer Interface XML. https://fdotwww.blob.core.windows.net/sitefinity/docs/default-source/cadd/downloads/documentation/fdotconnecttraining/ord-c/ordautomatedquantities.pdf
  17. Highways Today. “Florida DOT automates Quantity Estimations to standardize Workflows with Bentley Systems.” Item types in OpenRoads/OpenBridge as the foundation of FDOT’s quantity process. https://highways.today/2023/04/04/estimations-workflows/
  18. Illinois DOT. Computer Aided Design, Drafting, Modeling and Deliverables Manual (May 2025). Item types defined; 3D models used for measurement and AMG. https://public.powerdms.com/IDOT/documents/1943557/Computer%20Aided%20Design%2C%20Drafting%20Modeling%20and%20Deliverables%20Manual
  19. FHWA. 3D Engineered Models for Construction research context (design models used for refined quantities, AMG, and standardized digital construction). FHWA-HIF-16-031. https://www.fhwa.dot.gov/publications/research/infrastructure/pavements/16031/16031.pdf
  20. Construction Industry Institute and industry syntheses placing typical construction rework on the order of 5–15 percent of project cost (the risk MBD is trying to pull forward into the model).