One dead, nine injured in roller coaster derailment at Grona Lund

I love this. As a mechanical engineer, this discussion is really tickling me in all the right places. Hopefully I can provide some insight.

What's missing from this discussion, and is probably something that we'll never know in any particular case: what is the expected service life of the product? Because as an engineer, that is one of the fundamental inputs to the design, along with the load conditions. A roller coaster is a major investment and a permanent installation, so the answer could very well be "forever," or "until people stop riding it." Of course, the park would like to be able to operate the ride as long as it is popular and worth the investment needed for operations and upkeep. I think we all recognize that when a park closes a coaster because it has "reached the end of its service life," that's really code for "no one rides it anymore, so we're done dealing with it, and by the way we could really use that land to build something better." It's why Kings Island's Vortex gets dismantled while Magnum (same manufacturer, only two years newer) has no end in sight.

The thing you are trying to avoid is fatigue failure, where at some point a small crack develops in the surface of a material, and over time repeated stress cycles cause the crack to propagate, until the leftover material is no longer strong enough to handle the load, which leads to immediate, catastrophic fracture. This is different than something breaking due to excessive load; think gripping a pencil with both hands and bending it until it snaps. While this is fairly easy to calculate and predict, fatigue life is notoriously difficult to predict. In the automotive industry, or really in any industry where parts at produced at high volume and used in the same application (i.e. the load conditions are well understood and consistent), empirical data can be used to predict fatigue failure. But in our case, with low-volume parts used in different ways and in different environments, it's much more difficult to calculate accurately. And because fatigue failure generally doesn't come with any kind of warning--no squeak, rattle or vibration--it is absolutely conceivable that a coaster train could run without issue for 20 years and then fail catastrophically, and this could be the result of a design flaw, poor maintenance, or a combination of the two. Another issue to contend with is corrosion, but remember that many corrosion protection measures (specifying the right material and material treatment, i.e. stainless steel, galvanized or zinc plated steel, or even the type of paint) are part of the design.

Because most of what we're concerned with here is made of steel, the engineer could potentially design whatever it is--the track, the wheel carrier, the chassis--for infinite life, where the material strength is so much higher than the stress level that it will never fail. You could stand on the end of an I-beam and hop up and down a trillion times and the thing won't break. But often times there are constraints, such as size, weight and cost, that preclude designing for infinite life, so you must then prescribe a refurbishment schedule based on how long the parts are expected to last. So if we tell you to replace X part after Y hours, and you don't do that, and then that part fails later on, well then we're off the hook. Complicating matters is the fact that not only are service parts very expensive, but as we've all seen coaster companies come and go, and so it can be difficult and very costly to maintain a ride that has outlived the company that built it.

On the maintenance side, a few things stick out to me, in addition to the refurbishment/replacement of worn out parts I've already mentioned. You've got fastener checking/tightening, where a loose or missing bolt can cause additional load to be transferred to a component that wasn't designed to handle it. You've got lubrication, where wheel bearings need to be oiled and the couplings between the cars (and often times the track itself) need to be greased. And you've got regular full tear-downs with non-destructive testing (NDT). For the most part, what that refers to is looking for those fatigue cracks I talked about earlier, using methods like liquid penetrant testing and magnetic particle inspection. But NDT is not a guarantee that a crack will be discovered, because a) you're still relying on a human to identify the crack visually, and b) cracks can be hard to distinguish from normal surface imperfections, especially in cast parts or welded joints. Also, it is more difficult to inspect the inside of hollow tubes than the outside surface, but these areas can still see significant stress and are more prone to corrosion, as we learned from the Ohio State Fair accident a few years ago.

So there you go, just some things I was thinking about as I read through this thread.

Thanks for the excellent and thorough explanation, Bakeman. Definitely makes things clearer.

Had not known about this accident before - yeesh.

https://www.reddit.com/r/ro..._big_bend/

That's pretty nuts. Two trains collided on the lift? Which caused a car to completely derail and fall to the ground?

As an intern at SFOT in 1995, a supervisor told me about a Big Bend accident where 'a car fell off/hit a roof then the ground'. Wasn't sure to believe it or not. Looks like he was right, ouch. He also said something about those wrap-around bogies/weel casings were aluminum alloy, and bent easily.

Sometime in the late 70s, Schwarzkopf changed out the original designs for a bulkier ones, at least on Revolution.

All parts have different service lives, and can be changed out indefinitely. Schwarzkopf coasters - the trains and track, are lots of pieces that can come apart and can be replaced... whereas all those old Arrows are completely welded together, no moving parts (just tons of vibration).

I'd imagine the time and cost to preserve and properly maintain a vintage Schwarzkopf coaster is great, but some parks do it because it's worth it to them. Whizzer at Great America has 5 trains but only runs 2 or 3 while the others are constantly being refurbished. Hopefully some new parts are going where needed, that ride is more popular than ever. sooperdooperLoooper at Hersheypark is great example too, and with it's 3rd new set of trains.

sooperdooperLooper. Thank you for that, lol.

Demon lost an entire rear axle and didn't derail. It's safety inherent in Arrow's design. I'm not sure it was hyperbole in the Reynolds book to say that on the Matterhorn train (which was the prototype for all of Arrow's non-suspended train designs) all the wheels could fall off of the thing and it would still go down the track. That's not to say you can't get that kind of safety out of a design with the wheels outboard of the track; Arrow did some pretty clever things with the suspended coaster such that, again, if the wheel carrier assembly completely separates from the train, the train still won't derail.

Schwarzkopf's designs lack that level of redundancy.

--Dave Althoff, Jr.