This paper summarizes the use of Building Information Modeling as a vehicle to teach first principles of construction and structures to undergraduate students in architecture and engineering. Detailed virtual construction is used to explain the logic of structure and to illustrate the sequence of construction. Two techniques within Revit—the structural beam system and the curtain system—are utilized to introduce the concept that every part of a building’s assembly has structural responsibility and that the logic of structure and construction are inextricably linked to the logic of design. The paper argues that students who fully understand the logic of structure and construction through detailed Building Information Modeling efforts are better prepared to integrate that logic into design.
Architectural and Engineering educators across the United States are altering curricula to address the industry’s shift to a 3-D Building Information Modeling paradigm. This recalibration represents a unique opportunity to bridge an age-old chasm between compositional design and structural logic that are often taught independently from one another. The consequence of a disintegrated approach to teaching is that a fundamental aspect of architectural design—that structure and construction detailing are integral to design vocabulary—is typically lost on the student. To an inexperienced designer, design tends to follow a process that moves from abstract ideas to details as a one-way operation. Every tool within an architectural education must help students understand that discovery at the detail level can re-inform an early gesture, and that all tools are part of the iterative design process. Essays such as “The Tell-the-Tale-Detail” by Marco Frascari (Frascari, 1984) or exposure to the work and early design sketches of architects like Carlo Scarpa or Helmut Jahn (Herbert, 1993) demonstrate this concept.
In parallel, the elegance, efficiency and availability of digital fabrication have made it an expectation in industry rather than a luxury. This digital fabrication revolution has likewise exploded over the past two decades with students expecting access to laser cutting, CNC milling and 3-D printing operations for expressing their ideas. With more and more of the construction industry heading in this direction, schools of architecture must provide strong experiences for students to incorporate these digital tools and parametric design as an integral part of their basic disciplinary education. (Woodbury 2010).
The pedagogical approach presented capitalizes on the benefits of Problem-Based Learning (PBL) (Eberlein et al. 2008.) and utilizes detailed Building Information Modeling and parametric design to help students internalize first principles and apply those principles to the detailed modeling of building assemblies. The approach recognizes that BIM practiced at the level of modeling individual components of construction (Harfmann, 2004) will soon be a universal expectation as the culmination of a building design process. The strategy presented herein articulates a practical methodology using the logic of the structural beam system and the curtain system in Revit to move students with little to no construction or parametric computer modeling background toward thinking in parametric terms and internalizing first principles of structure, construction, and design drivers, thus preparing them for advanced level coursework and the future BIM-to-digital fabrication practice reality.
The overarching goal of the course is to continually reinforce the concept of modeling relationships and the layers of construction as structural systems. This exposes students to the power and potential of parametric modeling with the expectation that construction and structural logic will inform design logic during the process of design. In-class lectures use two tools in Revit to illustrate the logic of structure and sequence of construction for all layers of the building assembly. In most cases, this results in two, three or even four separate layers of beam systems or curtain systems to illustrate the hierarchy of structure and logic of construction.
Context And Objectives
The introductory course in the logic of structure through parametric modeling is taught to four distinct populations of students across two different colleges. The students, both undergraduate and a small number of graduate students, are all at the beginning stages of their architectural or engineering training and are described as follows:
- The largest population of about 90 students are sophomore architecture majors in the College of Design, Architecture, Art, and Planning, (DAAP) whose first-year experience includes mostly manual skill development, history, and design fundamentals.
- The second largest population of about 50 students are the Architectural Engineering majors who come from the College of Engineering and Applied Science (CEAS.) As an engineering degree, their first-year experience is dominated by math and science-based coursework such as Calculus, Physics and Chemistry as well as foundation courses in engineering.
- The third largest population of 45 Interior Design sophomores also comes from DAAP who share the firstyear foundation experience with the undergraduate students in architecture.
- Finally there are a small number of about 5 to 7 graduate architecture students in the M.Arch 1 program. The M.Arch 1 program is an extended, accredited graduate program for students with little no background in architecture. These students take the class during their first year in the graduate program along with an immersive studio and skills sequence.
Both the colleges that contain the programs in Architecture, Interior Design and Engineering at the University of Cincinnati participate in a mandatory cooperative education program where, beginning in the second half of the undergraduate year and the second year of the M.Arch 1 graduate program, students oscillate between 15 week academic terms and 15 weeks of professional practice employment. For undergraduates, this pattern continues until their senior capstone year resulting in an accumulation of more than one full year of practice experience. The graduate students have a total of three cooperative education semesters by the time they graduate with their degree. A diagram illustrating this curricular pattern for the undergraduates is shown in Figure 1.
Figure 1: Diagram illustrating cooperative education context within which the introduction course exists
The context of the cooperative education curriculum is of critical importance in this effort because the introductory structures/ construction class is the first “serious” technical course students take within their disciplines the semester before they enter their first coop employment experience. This relationship with our cooperative education employers has provided years of feedback and created the impetus to tailor this class to provide a reasonable level of structural/construction knowledge and experience with software so that students can be placed on meaningful projects during their first professional employment term.
The three core learning objectives developed are a direct response to having four distinct populations in the class and the need to simultaneously prepare each one of the populations for their first cooperative education experience. These learning objectives are:
- Introduce students to first principles and hierarchy of structure and how structural knowledge can influence design.
- Introduce students to basic materials, methods, and logic of construction as it relates to design.
- Introduce the concepts of parametric Building Information Modeling and Virtual Design & Construction.
With the learning objectives and the pedagogical instrument in place, the specific framework for orchestrating the path through the structural and construction content was developed. The overarching strategy of the course is to continually reinforce the concept of modeling relationships rather than simply modeling components of construction as static elements. This relationshipbuilding exposes students to the power and potential of parametric modeling with the expectation that construction and structural logic will inform design logic in the studio. The subsections are loosely organized according to the sequence in which they are introduced in the class. Classes typically alternate between lecture and demonstration on one day followed by an in-class “lab” the next where students follow the modeling process on their own laptops.
Moment of Inertia and Material Distribution
The first content area covered in the class is simple wood frame construction, which provides a relatively straightforward introduction to the complexity of structure and construction. On the first day of entering the class, students are asked to “walk the plank”—a 12’ long 2” x 8” supported at either end by a 4” x 4” block. This experience launches the discussion about the designer’s responsibility for the distribution of material and the patterns of that distribution. After a brief foray into calculating the Moment of Inertia for simple rectangular beams to illustrate how the depth of material is related to its capacity to resist bending, the focus shifts to modeling this behavior parametrically. The “plank” and visualization of Moment of Inertia used in class is shown in Figure 2.
Figure 2: Introduction to the concept of Moment of Inertia
All second-year students have access to introductory texts, in this case The Architect’s Studio Companion (Allen, 2007), and are familiar with rules of thumb for assigning depth to structure. This behavior is modeled as a simple rectangular beam that grows in depth proportionally to the length of the span using L/15 as the rule.
Since this course is obligated to introduce students to more depth with respect to structure and construction, the rules of thumb are set aside in favor of a more accurate approach to assigning depth to structure. Since this is quite simple for wood construction, we use the Southern Pine Span Tables (http://www.southernpine. com/span-tables.asp) to size and model wood joist behavior. As a point of departure, we consider and model a typical 16” on-center deployment and discuss how the different spacing and grades of lumber in the table affect the joist’s ability to span. The if-then statement that drives the depth parametrically based on the length utilizing values extracted from the Southern Pine table is shown in Figure 3.
Figure 3: IF-THEN-ELSE formula driving joist depth based on Southern Pine Span Table =IF(L< 9’1”,5.5”,IF(L< 11’11”,7.5”,IF(L< 14’8”,9.25”,IF (L< 17’2”,11.25”,0.1))))
In the formula, “L” represents the length of the joist so if the span (L) is less than 9’1” then the joist depth equals 5.5”, else, if the span (L) is less than 11’ 11” then the joist depth is 7.5” and so on. An abnormally small value is entered if the span exceeds the 17’ – 2” maximum for the 2 x 12. If the student attempts to model a joist exceeding the maximum length, the conditional statement prevents the software from generating the joist. By experimenting with the dynamically adjusting joist, the student learns that the design must somehow address material distribution according to the principles of structure. The “honest” expressions in the use of both the rule-ofthumb and dimension lumber joist structures in a triangular shaped space for instance, would require a varying bearing condition that becomes obvious when the structural system is placed relative to a foundation as seen in Figure 4. This experimentation provides students with a graphic, diagrammatic, and intuitive understanding of a fundamental principle of beam theory and its relationship to design.
One of the basics of structural logic is the principle of structural hierarchy—the idea that one item sheds its load to another then that element may shed its load to another and so on. To model this first principle, a simple plank is modeled as a family in Revit then imported and placed one at a time in the project (the way that a carpenter might install them.) Once the concept of placing these planks individually is established, the concept of a structural beam system is introduced. The planks are placed within the boundaries of a drawn beam system and placed at a regular spacing equal to the width of the plank to create a continuous deck surface. Students see immediately that the deck can quickly be modified by changing the shape of the beam system boundary and the planks will regenerate within the modified design. Once the planks are modeled, the beam system necessary to support the low Moment of Inertia planks is created below the planks. These joists are likewise spaced at regular intervals and can even incorporate the dynamically adjusting joists with varying depths if desired. Finally, the beams that support these joists are modeled to provide the full picture of the layering and hierarchy of structure that is typical in simple designs. Each of the systems are modeled individually on separate levels to explode the views as well as maintain easy individual editability of each layer of structure. This same technique is used for modeling tongue and groove flooring mounted over sleepers resting on hollow core concrete. To maintain relationships, each layer and boundary of each beam system is modeled and linked to a common set of reference lines. Figures 5 illustrates the layering of structural beam systems as a hierarchy of structure in exploded axonometric form
Figure 4: Varying joists depths based on length as part of structural logic
Figure 5: Exploded axonometric showing the introduction of layered beam systems
Envelopes and Facades as a Hierarchy of Beam Systems
All construction obeys the laws of physics and as such, the logic of structure applies, whether it is a series of joists resisting gravity loads or a series of girts that withstand wind load. While it may seem unintuitive to model girts using the same strategy for a system of joists, the first principles and rules for distributing material apply to both. The same logic applies to the application of furring strips for drywall applied over a concrete masonry wall or stringers for suspending a ceiling structure. As the thickness of the final layer increases, its Moment of Inertia and capacity to span increases, which in turn allows the spacing of the girts, furring strips or stringers to increase. This increase in girt spacing also directly increases girt size so the exercise of exploring options gives students a first-hand experience with material distribution. From this, students quickly develop intuitive understanding of the inverse relationship between the size of members and the necessary number of members. They also see that this choice has a visual affect that may or may not be consistent with design aspirations. Figure 6 illustrates a structural logic for a girt system that spans between columns to support a simple, vertical metal panel exterior.
Figure 6: Two different bays with girt systems for supporting metal panel cladding
As shown in the figure, girts in this case are simple sections that dynamically adjust their depth based on span. Various spacing can be explored as well, resulting in girts with a different cross section that also vary in depth with changes in span. As the spacing between columns changes the girts grow in depth accordingly. The bay on the right shows panels with deeper ribs and closely spaced girts spanning between columns that are 10’ apart. The bay on the left shows panels with shallower ribs and larger, less frequent girts spanning between columns spaced 16’ apart. Since the line of the exterior envelope has already been related to the grid line that controls the position of the columns, it is an easy task to adjust the distance between them to accommodate the combined thicknesses of envelope and girt.
Using the Curtain Grid System as Structural Logic in the Exploration of Form
Toward the end of the semester, students are introduced to the curtain grid system as another way of describing a pattern of construction. After a brief introduction to the logic of the curtain wall system in Revit and the basic modeling of a simple glass and mullion assembly, the instruction immediately shifts to developing an understanding of the curtain wall system as a hierarchy of structure, where the glass acts in the capacity of decking, mullions acting essentially as joists or beams and subsequent girts or columns acting like beams or girders. This is a natural progression from the modeling of girts and solid cladding elements as structural beam systems but uses the logic of the curtain grid system and the ability to map the grid system to a conceptual mass to explore the relationship between structure and form.
Once it is understood that mullions in one direction spanning between mullions in the perpendicular direction are behaving like joists spanning between beams, but in the vertical direction, we apply this logic to the curtain grid system that is mapped onto a conceptual mass form. The conceptual mass form created, illustrated in Figure 7a, is a simple conoid with different sized arch forms at either end of a vaulted space.
Figure 7a: Conoid conceptual mass form
Figure 7b: Curtain grid mapped onto form
Once the conoid mass family is loaded into the Revit project, the custom curtain grid system is mapped onto its surface as seen in Figure 7b. The grid at this point has no members, only the logic of a grid that serves as a placeholder for the placement of elements. Using the fundamental understanding of material distribution gained from repeated lessons on the hierarchy of structure and the relationship between span and depth of members, a “mullion” that is 24 inches wide by 50 inches deep is created and assigned to the cross direction of the curtain grid of the mass to emulate the arch ribs of a vault. “Mullions” that serve as purlins are created at a more frequent pattern and are assigned to the perpendicular direction to illustrate the hierarchy of structure. While the result in this case is a segmented arch for the ribs, the net effect offers an acceptable visualization of the design. Figure 8 illustrates two views of the curtain grid associated with the conceptual mass and the custom mullions that emulate ribs for the vaulted space.
Figure 8: Curtain grid on conceptual mass and large mullions emulating the ribs of a vault
One advantage of using the strategy of mapping a curtain grid onto a conceptual mass family is that one can easily alter the conceptual mass in response to new constraints or suggestions, then reload the mass into the project and simply reassign the curtain grid to the new form. The curtain grid will quickly adjust all the elements accordingly and the new solution can be assessed and visualized. If the mass is created with instance parameters, this adjustment can be made very easily by entering new values that will recalculate the height of the vaults, the span the length of the vault, etc. allowing for rapid exploration of alternative forms. Likewise, the “mullions” that serve as the ribs can also be parameterized allowing the adjustment in their size as well for rapid adjustments to their proportions.
Summary and Conclusion
The best method for illustrating the effects of this pedagogical approach is through assessing student projects. The project shown in Figure 9 is one student’s final submission of an 11-week project in which they design and detail the assembly of a simple structure consisting of three forms and three different construction types. Students are required to incorporate structural systems of wood joists and beams in one geometry, hollow core concrete planks bearing on masonry walls in another geometry and steel framing with open web steel joists and a curtain wall glass envelope in the third geometry. The submissions are cumulative, beginning with the wood frame portion followed by submission of the masonry bearing portion of their house. The third submission is a pre-final submission that includes all three geometries and all three types of construction with a first attempt to resolve the connection of all three geometries with the hallway. The third submission is returned with redline comments and suggestions that students incorporate into their final submission. The output for the final submission includes overall 3-D images of entire model as well as traditional plans, building sections, elevations and detailed wall sections of the wood frame, masonry bearing and steel frame portions of their structure. Success is measured by how well students model their constructs and assemblies with reasonable sizing, spacing and detailing of members for each of the geometries.
Figure 9: Final submission of component model (courtesy Troy Newell)
Admittedly, for experienced designers and architects the imagery and concepts presented herein are not extraordinary, remarkable, or new. However, the opportunity to combine principles of structure and construction with digital modeling and parametric design directly correlates to the expectations of practice today and provides a good platform to introduce beginning students to the logic of structure and its relationship to design. Digital modeling allows students to test and explore a multitude of construction conditions, especially because components generated according to parameters are adaptable to a variety of relationships. Furthermore, within the scope of BIM, students are encouraged to design and employ their own parametric components. This can help them understand that the artistry of design results not only from a response to form and aesthetics, but also from a command of data, parameters, construction, structural principles, and the relationships among these elements.