You’re going to upgrade your facility with a new, rooftop packaged system or air conditioner. The inspectors are coming, they’re going to do some measurements, and ask ‘if I put 12 tons of metal on this roof, what’s going to happen?’ For the common person, two thoughts probably spring to mind: ‘this ends in tears’ or ‘that roof’s been there for 50 years, and it’ll be there for 50 more.’ For the scientist and engineer, we look to numbers to find out.
We’ll start with the concept of load. A load is anything pressing down on our structure. We have two types of load: dead load and live load. The dead-load is essentially the structure itself. It’s the weight of the rafters, walls, structure, everything that is the building. Live load is everything we put on that structure, everything that the user could move or have moved.
In working on the building, we need to first understand the forces it puts on itself. You find out what materials are in play, calculate approximations of their weight, and then you can say “these two stretches of roofing material are applying a force of 3 kN on the two beams under them.” Then we can say “each of these beams is putting down 1.5kN + their weight in KN on the walls/supports holding them.” We essentially build a map of where the forces are all going, and how big those forces are.
The same measurements and calculations need to be worked out for things in the building. In most cases, this only really a concern with the roof, but in some types of structures or installation plans, it may be necessary to perform calculations for the rest of the building. You need to work out how heavy the existing rooftop equipment is and apply it to the model.
A key part of these calculations is how weight will be distributed and spread out over the rest of the structure and further, how the structure will respond. The load you put on the structure will act as a lever to the beams supporting it. Leverage is a thing at play.
If you put a weight directly over a supporting wall, then nearly all of that weight goes directly into the wall and the support beams will basically receive no load. The farther from the wall you put that weight, the greater force it’ll have against the beam. This is why we have supports in the middle of structures. If you stretched a single, massive beam across a structure, the weight in the center of just the beam on its own could bend and break it. You might be able to demonstrate this with a stick. Bend it in half, break it, eventually the shorter chunks you have left will be too small to bend and break in your grasp.
We need to work out where our weight is going and what impact that will have on the remainder of the structure. For simplicity’s sake, we’ll say that any load we put on a beam will be distributed evening. An even chunk of it will go on everything touching that beam. If the load spans beams, then it evenly distributes between those beams and everything those beams touch. Your engineer will probably also calculate what impact that load has based on WHERE on the beam and supports it falls, mainly to determine if a beam is too long to tolerate the weight on it.
All the forces are now known, with a reasonable margin of error to say ‘we calculated this, but in case we’re wrong, we’ll increase the weight on the structure/consider it weaker than it is.’ The battle is half over. NOW we need to know about what this place is made of and just how strong it is.
Consider that most skyscrapers and tall structures today use steel and concrete, which are harder to make than something like lumber. Wood is an incredible material, but it doesn’t have the same strength as steel and concrete. Beyond that, the exact formulation of metal and concrete will change its strength.
Depending on the paperwork available about the building, there may be some testing involved to find out exactly what metals its made of and how well those materials have aged. A structure with significant rust may require a deeper examination to determine what, if any strength remains.
Answering these questions about materials will tell us the last major piece of the puzzle: these structures hold X now, they impart that force in Y places, and we believe each part can handle Z total minus these few tons for safety.
Other key factors in this also include how materials are joined, what bolts are used, and even how tightly those bolts are cranked down. These numbers, materials, and forces tell us what a building will do.
Don’t Skip the Math
You might ask ‘is this necessary’ and the answer there is yes. It is essential. When you skip out on doing the math and the measurements, then it ends in tears, vast legal expenses, prison time, and bad PR. There are any number of engineering disasters we could point to as key examples. let’s look at one where bad communication and lack of calculation caused deaths and hundreds of millions in damages.
The Hyatt Regency Walkway
In 1980, the Hyatt Regency of Kansas City had elevated walkways that ran across the upper floors in their lobby. These were basically bridges suspended from the ceiling. In July of 1981 the second and fourth floor walkways collapsed, killing 114 and injuring 216 others. The legal payouts were at least $386 Million, accounting for inflation.
The investigation found that the bridge collapse largely resulted from design flaws that would have been caught if the necessary calculations were done. As designed originally, the bridge held only 60% of what building codes required. Revisions to these designs were made with a mere phonecall and no math to determine if they were even possible. The walkway that was built could hold 30% of the load put on it. It was just strong enough to hold up for a short time, but gradually deteriorating until it collapsed.
Always Use Science
Remember, plan your upgrades ahead of time, understand the basics involved, and everything will go fine. Your engineers have learned from past engineering mistakes. If they tell you “your building will collapse in a massive disaster if you even try this,” they know what they’re doing. There’s no such thing as too much margin for error.