Kennywood begins adding new columns to Steel Curtain roller coaster

Walt S:

Out of curiosity, does anyone know if the structure for a steel coaster has to be signed off by a civil engineer / architect in the same manner that the steel frame for a building needs to be signed off on?

It appears most of the manufacturers rely on Ingenieurburo Stengel to do structural design/analysis.

https://www.rcstengel.com/e.../ueber-uns

I would certainly assume a PE eventually signs off on the design after analysis and all that is complete. Given lives at risk and all that, I can't imagine a case in which they would not. I worked for an engineering consulting firm at one point that did all types of mechanical, civil, HVAC, heavy industrial, etc... And basically the entire civil and structural group were required to have their PE.

Rick_UK:

It is incredible how stuff like this happens in a world of CAD, but good to see them fixing it. I am sure some people would like it to have been done quicker, but perhaps that wasn't possible.

I wouldn't be surprised if Steel Curtain's issue (whatever it is) came about because of computer-aided design. I doubt the structure is one where forces can reasonably be checked by hand, so the designer had to rely on computer-generated results, which would have been dependent on assumptions defined by the engineer or program defaults. If those assumptions weren't reasonable, the model used couldn't accurately predict the structure's behavior.

Nothing you're saying makes sense. No one is modeling anything with a slide rule.

His point is perfectly valid.

In structural analysis, the output is only as good as your inputs. How you constrain the joints, the mesh of the finite element model, loading, how you model the welded joints, material properties, etc...

Changing some of those conditions can make significant changes to the stresses that result in the analysis.

It's not a fully automated thing by any means.

I worked for a premier structural engineering firm responsible for analyzing many of the most well-known structures in the country. Our computer models were always validated using engineering fundamentals (i.e., by hand) to confirm the output. Sometimes this was as simple as "Does the computer know how much this structure weighs?" Other times it was much more complicated than that. But if the computer and the hand calcs weren't reasonably similar, we went back to the drawing board.

If S&S or whoever was responsible for Steel Curtain's structure modeled the ride but didn't validate the inputs/outputs with engineering fundamentals, it doesn't matter how powerful their software is. Garbage in = garbage out.

In the infamous words of one of our former Chief Engineers (Big 3 Full Size Truck)....

"All models are wrong. Some are useful."

Later,
EV

The sway of the structure in the video at first does not appear to be anything that would not be expected on a coaster, but long after the train passes the structure continued to sway. This is unlike other coasters, at least from what I have ever observed. The supports seem to be too vertical. With the trains momentum the structure does not seem to provide the support in the direction of the stress.

There's deflection, and there's oscillation. Some amount of deflection seems reasonable (though too much is another sap of kinetic energy). Oscillation I would not consider desirable or normal. I don't have a technical reasoning, other than I wouldn't want something structural flapping around in the breeze.