top of page

SOWETO, SOUTH AFRICA

OSEA 2021

Architectural Design Award

First Place Entry

Designing the Structural System for a Fire Station

Structure Axon w Explode.png

WARNING: Lots of reading ahead!

It turns out that creating the preliminary structural designs for an entire building requires A LOT of work!

Structural design is all about location, location, location!

As explained on the page dedicated to this project's architectural design, the primary building material chosen for the Orlando West Fire Station was the standard, fired clay masonry unit commonly referred to as a "brick." In the spirit of challenge and learning something new, I decided to use the same material for a vast portion of the building's structural system as well.

While the load bearing capabilities of a mass masonry wall were never a matter of great concern, the decision to provide the fire station with a full span, semi-intensive vegetative roof system meant that the roof structure that carried its immense load to the stout walls needed to be equally robust. In all, this structural design project allowed me to explore new avenues of problem solving as well as alternative perspectives on how integral the structural system is to ongoing architectural changes throughout a project's duration.

First Things First: What Loads Need Supported?

**

While the values for such data as tabulated seismic accelerations or local maximum wind speeds are surely available to engineers in South Africa in very similar fashion to a resource like the ASCE 7-10 for real design projects, all values that I deduced throughout the semester came from comparable standards used in the U.S. for the sake of practice and due to the academic nature of the project. In essence, it would not have made enough of a difference in the final product to use South African resources in their native metric units, undergo a series of unit conversions for the sake of calculations, and then undo it all for local member sizing and the like. As such, the following loads and calculations were conducted using standard resources for structural design in the United States.

Dead and Live Gravity Loads:

Of course, in order to make the "most correct" structural system for this building, or any building for that matter, a structural engineer would have to quantify the material weight and composition of every cubic inch that the building itself occupies. However, for the sake of efficiency and consistency, the "worst case scenarios" are often used as a baseline for the overall building design. The following wall sections provided the most feasible load values that the structural system would have to withstand:

Design Loads_Page_1.png
Design Loads_Page_2.png

As these simple calculations demonstrate, it is usually safer to find a "total value" for both Dead and Live loads and then round up to a number that will be easier to use throughout the calculations to come. While this is not done according to any specified procedure, it does provide another (small) factor of safety for the design of the structure.

Directional Wind Loads:

A noteworthy aspect of this building's design is how it seems to "grow" out of the ground from east to west. As such, the wind load from the east is negligible for design purposes. For the remaining cardinal directions, the ASCE 7-10 "Directional Method" was employed to calculate the peak wind pressure loads at applicable heights along the building's exterior walls.

Directional Seismic Loads:

Seismic loads are caused by the building's inertia in the event of seismic movement. In other words, the weight of the building itself is used to calculate just how much of an oscillatory force is applied throughout the finished structure when the ground shakes.

In the calculations below, the total weight of the building structure and the exterior facade is calculated after the steel frame for the roof is preliminarily sized. Although this structure weight may not reflect the exact value of the final design, the built-in factors of safety within the Base Shear (V) calculations allow the final value to be used.

What To To With All That Load: Structural System Design

Composite Roof Deck/Frame Design

As can be seen above, the design dead and live loads for the large vegetative roof with pedestrian access are a whopping 200 lb/ft  and 100 lb/ft  respectively. Compared to a typical roof load for a commercial building, these loads are significantly higher and, thus, require an overhauled structural system.

The most practical solution for such loads is a system of composite beams that employ rolled steel for the tension face and concrete for the compression face of the deck. This system plays off of the inherent strengths of both of these materials and allows for much larger loads to be carried to the load bearing walls that surround the roof.

As can be seen in the AISC Steel Construction Manual, Chapter I for design of a composite beam for flexure, the process is not entirely dissimilar from a standard steel beam for the same purpose. In most cases, the design is controlled by deflection criteria, which can create an iterative loop if camber is not desired. However, for the sake of this project, I decided to create a spreadsheet in Excel that would do most of the calculation work as long as I supplied the correct information from various Beam properties tables throughout the AISC Manual.

 

Below, a couple of examples of this spreadsheet can be seen. After nearly 23 hours of work in Excel, this spreadsheet is able to determine whether or not a beam is adequate for composite design given any tributary width as well as if a girder is adequate for composite or non-composite design based on different equal spacings of the beams that frame into it. As can be seen on the framing plan, nearly every bay width is different for the column grid to best accommodate the architectural plan. As a result, most beams that frame into a column have asymmetric tributary widths. This spreadsheet was invaluable for correctly calculating with this in mind.

2

2

A Foray Into Clay Masonry Design For Structure

The decision to utilize clay masonry for the building's structure along with its exterior facade was initially made with hesitation thanks to the scarcity with which it is used in the United States anymore. However, because I had recently completed a course about structural design of Concrete Masonry Units, the theory was still fresh in my mind and I was keen to accept the challenge. As it turns out, the design of concrete masonry and clay masonry are very similar and, with the exception of a few variable design factors and coefficients, their application is nearly identical. 

In order to utilize clay masonry as the primary structural material for the exterior walls of the building, two primary force resistance scenarios were to be considered:

combined axial and flexural loading from the weight of the roof and the prevailing lateral force between wind and seismic effects and shear resistance along the length of the walls themselves. Because the pure axial load capacity of a double-wythe, fully grouted cavity wall was never a matter of concern, it was decided that the combined loading scenario should be explored for the most extreme ground-to-roof height of the superstructure.

Beam-Columns, but with Bricks!

Masonry of any kind, be it stone, concrete, or clay, is not well known for its tensile strength. Quite the contrary, in fact. Once a load bearing, exterior masonry wall reaches a sufficient height, it is subjected to both the weight of that which it supports (axial load) as well as a wind load spread across its entire area that registers as a non-uniform, distributed load in force-per-area units (flexural load). In order to compensate for the flexural forces, masonry can be reinforced with steel much like a poured concrete structure. While open cells are utilized for steel placement in concrete masonry, "solid" clay masonry assemblies often make use of a fully grouted gap between two or more wythes of stacked units.

Could It Be Any More Shear-Resistant?

Even in modern structures composed primarily of steel or timber, masonry is still commonly employed for shear wall elements, both internal and external. Because of its nature as a dense, "massive" material, reinforced masonry is ideal for developing shear along any significant length. As it so happens, this building is literally one large shear wall after another by circumstance of the chosen structural material. For clay masonry, the point of concern when designing for shear is often the mortar joints, as there is more likelihood of slip failure between the mortar and the bricks as opposed to a shear fracture within the bricks themselves.

For this project, I decided to conduct a thought experiment of sorts. Even though every exterior wall of the building is undoubtedly able to carry shear to the foundation by way of its mass, I was curious to know if I could successfully design one section of one wall to carry the entire shear load of that wall segment without accounting for the additive shear development of the adjacent lengths. Below are the calculations for this scenario:

Another [Hole] in the Wall - Masonry Spandrel Design

Clay masonry is not the first choice of material for spanning a large gap due to its less-than-desirable qualities in flexure. The longest span that any of the brick was faced with in the design of this project was the 14 foot openings for the apparatus bay to accommodate entry and exit for fire engines. Fortunately, these openings are located at the western extreme of the main building volume which is also where the full-height masonry walls are tallest.

If the entirety of the wall above the opening is deemed to be the effective "member" in flexure, it allows for the "beam" depth to be over ten feet or 120 inches! Deflection is often the governing design criteria for a masonry flexure member. Maximum deflection is also calculated using beam depth. So, a very deep beam is very resistant to deflection, even over a span of 14 feet. Below are the arguably comical calculations that verify the ridiculous effective beam depth and its subsequent lack of failure:

The Pros and Cons of Technology: RISA 3D

Unlike the straightforward and linear process by which the National Corvette Museum structural systems were designed in RISA 3D, I tested my own abilities and patience by applying the most significant gravity loads on a non-horizontal roof plane. For all of the wonderous capabilities of RISA 3D and RISA Floor, their inability to analyze a sloped floor or roof plane "out of the box" can prove quite vexing. Similarly, RISA 3D imagines that all good engineers will design their buildings practically and efficiently using rectangular walls with floor and roof planes that are parallel. So, imagine my surprise when the default wall tool could not be used to create the main structural components of my fire station in a concise manner...

But, like any good engineer, the problem solving required to tackle this particular issue proved to be an invaluable learning experience in the end. If "walls" with assigned structural and material properties cannot be utilized successfully, the software offers a supplementary tool known as "Plates." These rigid, polygonal planes can be assigned material thicknesses, structural properties, and can be generated anywhere in space as long as three nodes exist.

Unfortunately, these plates cannot be perforated once placed. So, in order to accurately model a building with a rectangular plan, a sloped roof, and regular openings for windows and doors, every single corner's position must be known and a node must be placed at it. Once the nodes are placed, Plates can be created. The model will only run an analysis if every single node is correctly placed in-plane and if every single plate is connected to its neighboring plate at the exact corner node locations.

After much toiling and enough error messages to make a novice programmer look like a pro... behold the Orlando West Fire Station structural model made of 498 Plates:

So You Told Your Structural Analysis Software to Analyze Structure...

The power of such a tool as RISA 3D should never be undervalued, especially when it is wielded by a real pro. After the model is created and is compliant with the software's rules for analysis, the true magic can begin! While the Plates were a pain to create, they proved to be very handy for the application of loads.

Using a superimposed area load for the roof's gravity loads, both dead and live, as well as an area load for all lateral wind loads was a breeze with RISA 3D. The new interface is far more user friendly. Simply point, click, input a few numbers that were previously calculated, and voila! As for applying the lateral seismic loads... the Plates once more reared their frustrating head and challenged my patience as a designer. As can be seen in the above model, there are clusters of smaller plates in the middle of the high and low roof planes. This may seem strange at first, but they are created as such so that the centroids of the overall planes as well as a ±5% eccentricity in both major directions could be modeled as distinct nodes. These nodes were used as application points for the prevailing seismic loads as previously calculated.

The gravity, lateral wind, and lateral seismic loads applied to the model resulted in:

... And Then It Gave You a Bunch of Numbers.

For all the graphic bravado that the RISA modeling phase carries, the final result can often feel somewhat lacking in comparison. I iterated and reiterated over and over, checked and double checked property inputs, and finally it came time to click the elusive "Solve" button for the final time. No more errors, no more "Unable to Complete."

Wouldn't you know it, it's still not done! All of that work and, ultimately, the software throws a handful of numbers and shapes at you and tells you to resolve once they're implemented. So that's what you do and finally finally the golden results show their face! Since I happen to appreciate the inherent value that numbers have, I felt properly rewarded for my efforts when I was eventually granted the countless spreadsheets known as "Results."

There is one particular spreadsheet that I was most interested in, as it would allow me to continue onto one of the most important parts of the structural design process. The following calculations show the individual ground reactions for use in design of the building's foundation!

Factored Load Reactions

Service Load Reactions

For both sets of calculations above, each line of the spreadsheet is representative of a single point located at the corner of a plate. These are organized per length of exterior wall that they are located within. Then, the total reactions per wall length are summed and divided by the length of the wall to produce a per-wall lb/ft load. 

These lb/ft loads were used for the design of the strip footings that would be located beneath the loadbearing masonry walls.

Getting Down and Dirt-y: Foundation Design

Strip Footings To Support Loadbearing Masonry Walls? Really?

Yes! Thanks to a generous studio professor and local soils reports from the Soweto area of Johannesburg, the site soil for this project was assigned a bearing capacity of a whopping 4,000 psf! (See Structural Notes page in Drawings below.) As such, it was deemed that, even for a relatively massive building such as this, it was not necessary to design a deep foundation system with piers or piles.

For the sake of "simplicity," it was determined that only the worst case scenario in terms of ground reactions was to be used for footing design around the entire perimeter of the building. Below are the calculations for just such a case:

Oh Yeah... Those Sneaky Steel Columns Need Support Too.

Although not extensively mentioned thus far, there is a single "spine" of steel columns that *nearly* bisects the main volume of the building longitudinally. As with the strip footings, the worst case scenario of loading from all of the columns was used to design a single size of footing. This footing size was designated for every isolated footing.

It All Comes Down To This.

Everything that you have just read above is really just a brief summary of the semester-long project's highlight reel. As mentioned at the top, it truly was a lot of work to design a preliminary structural system for a medium-sized fire station. The fruits of all of this labor are as follows: The Drawings!

If anyone has made it this far and is a true glutton for punishment... the link below will open a 275 Page Project Manual that goes into far greater detail about every single step in the whole process of this project. Also, here is the link to the flip side of the coin: the Architectural Design Part of the Project!

bottom of page