Stockton Rush says “we’ve got that covered!”
I have been doing and overseeing design, analysis, qualification/acceptance testing, and non-destructive inspection (NDI) of fiber-resin composite structures using ‘graphite’/carbon, aramid, glass, polyethylene, and other fibers for over twenty years. Fiber properties, and especially the different types and grades of carbon fiber are well understood and characterized. The failure modes of carbon fiber structures can be quite complex compared to those manufactured from isentropic metals like alloy steel or even grain-dependent materials like many aluminum alloys, but to say they are “poorly understood” or that “only very recently have we developed the tools to even inspect it properly” are completely wrong. Modern finite element tools have specific elements and construction methods to represent composite layups and apply specific failure criteria to results. NDI using ultrasonic (pulse-echo and phased array), thermographic, acoustic emission, radiographic/CT have been around for a couple of decades or more, and more advanced methods like eddy current, miliwave, laser shearography, lamb-vibe, and ERM are mature enough for production use. The real problem is that many of these methods take a lot of time and experience to use correctly, and highly trained (Level 3) inspectors in these areas are difficult to find and hard to keep. Commerical aircraft manufacturers have been slow to adopt composite structures because they are very conservative and there is a big investment of time and tooling to develop these to a production level (and often questionable value).
Carbon fiber would not “be an excellent material for suspension bridge cables” for the reasons previously mentioned despite the tensile strength they offer. Specifically, the propensity to fail without warning and in a potentially catastrophic way, as well as their vulnerability to moisture and UV light. The selection of a material or any other design element isn’t a matter of a single “figure of merit” but a consideration of many applicable factors such as damage tolerance, inspectability of failure mechanisms, manufacturability, cost and availability, et cetera. I’ve seen many design decisions where people elected to build a structure out of carbon fiber composite in the belief that it would substantially reduce weight or provide a higher load capability than aluminum or another choice only to discover that after going through prototype/pilot manufacturing that the amount of reinforcement in joints and stress concentrations, repairs to delamination, difficulties in controlling manufacturing tolerances and defects, et cetera that there was little to no weight savings or increase in load capacity and at greater expense and more rejects, too. Composites definitely have their place in engineered products but they are not the be-all of materials even when high strength is an important criteria.
Stranger
A couple of decades is recent when it comes to civil engineering (or commercial aerospace). As you admit, it isn’t long enough to have a generation of technicians available that are trained in it.
I find it funny that you’re simultaneously claiming that CF composites aren’t poorly understood, but also that they can catastrophically fail without warning. Those things can’t both be true. Well-understood materials don’t fail without warning. And that understanding is distributed well enough that people don’t regularly misdesign things so that they are prone to random failures.
For all their problems, Boeing seems to have worked out the bugs with the 787 wing/fuselage and the craft are not failing without warning. Despite the fact that the amount of the amount of flexure is far higher than an aircraft than it would be on a bridge, and any issues with lamination or otherwise would be amplified.
CFRP cores are already in use in power transmission lines, and again the amount of flex they experience is greater than on a long bridge. The advantages are exactly as with the bridge: higher tensile strength for a given weight, so you can put the towers farther apart, increase the cross-section of the conductors, reduce sag, etc.
This is not a generic “make it out of CF to make the thing stronger.” The figure of merit is tensile strength, and CF does extremely well there. It’s even more well-suited there than aircraft components, since the load is along just a single axis.
You two might be talking past each other or meaning different things, but the way I read it, this isn’t true. Brittle materials can be well understood, but they fail suddenly. Ductile materials have a lot of deformation before failure and therefore have a clear warning beforehand. This concept is drilled home in reinforced concrete design, whose failure can be either brittle or ductile depending on if the concrete or reinforcement fails first. It is vastly better for the reinforcement to fail because there are a lot of visual cues beforehand and remedial action can be taken.
It’s well understood when they fail, though. No one avoids glass because it fails suddenly when you exceed the ultimate strength. You design the structure to not reach that point.
You may well be right that I’ve misinterpreted his statement, but if that’s the case then it makes even less sense. Materials with brittle failure modes are commonplace. You take it into account when you design the thing. Yes, ductile failure modes can be handy, but ideally you never get there even when the material allows it.
Incidentally, there is at least one CFRP cable stayed bridge out there:
https://www.researchgate.net/figure/Stork-Bridge-in-Winterthur-Switzerland-with-CFRP-stay-cables_fig2_257803214
It’s pretty small, but it’s been standing since 1996.
You absolutely avoid brittle materials / composites because they fail suddenly. Sometimes the benefits outweigh the downsides, but just because you can predict an ultimate load doesn’t mean there’s not a preference for a safer failure mechanism. I mean, structural engineering is a mature field with decades of testing and experience, but buildings still fail. And if that’s the case the manner in which they fail matters.
I’ve designed concrete beams w/ composite rebar for their corrosion resistance so I’m not saying a ductile failure mode is mandatory, but “just design it not to fail” isn’t sufficient in that profession.
Interesting bridge. Cable stayed vs suspension though. I’d wonder about the bearing of the cable on the supports for a suspension bridge, but I doubt it’s an insurmountable task.
I mean no one avoids the material completely. If you need glass for the application, you use it. And if you need CF, you use that. Here, it’s impossible to build a sufficiently long bridge without a high tensile strength material. Steel (for the cables at least) simply won’t cut it. And high tensile strength materials tend to have brittle failure. One might have an overall preference for ductile materials, but it’s engineering–you don’t always get what you want.
There are certainly many complexities. The “cable” doesn’t necessarily have to be a cable, though. It could be stacked ribbons, for example, giving them more flex and more surface area for the supports.
Also, I’m saying “suspension bridge” in only the loosest sense. It wouldn’t necessarily look like a normal suspension bridge.
That was me, by the way:
That one was even more speculative than the discussion here.
Some more examples in this paper:
That covers FRP (fiber reinforced polymer) bridges in all their varieties, whether glass/aramid/CF reinforced or whether it was used for the main structural elements or just cables.
The main impediment seems to be cost and expertise, which partly comes from the lack of standardization.
Another paper on the subject of super-long bridges (up to 10 km) using FRP:
What I said is that carbon fiber can fail in tension without warning because it has very low elongation before failure and, as @Snarky_Kong noted, it fails in a brittle manner. This gives no opportunity to observe or measure a failing condition (due to degradation, overload, or just a mistake in design analysis) before catastrophic failure occurs. This isn’t about carbon fiber not being well characterized; in fact, I have an entire book on my bookshelf about carbon, glass, and aramid fiber (published in 1982, to boot, and yet still quite relevant to materials in use today) that extensively details the properties, failure mechanics, and vulnerabilities of different classes of these fibers as well as the methods used to manufacture, inspect, and test them. It is just the fact that when it fails it does so abruptly, so it makes a poor choice of material where a direct tensile failure would be catastrophic in a variable loading condition, or in an application where you want to have a useful degree of elasticity. Carbon fiber is great in fiber/resin composites where the loads are well characterized or where the stress state can be managed to ensure high margin to fiber rupture, and where high stiffness-to-weight is a key consideration, which makes it well suited for many aircraft and space vehicle structures. In fact, most carbon fiber composite structures other than pressure vessels don’t fail due to tensile stress causing fiber rupture; they fail in interlaminar tension or interlaminar shear where the matrix that bonds the fiber layers together fails first.
You seem to have the view that tensile strength is the only “figure of merit” in the design of a structural or mechanical system. This is, frankly, a very facile and uninformed view of how materials are selected and used. In many cases, you want or even require a controlled amount of elasticity or flexure, or to have a particular shear modulus, or the ability to fail ‘gracefully’ in ductile mode to prevent abrupt, uninspectable failures or sudden release of stored energy. (The last is literally a design requirement in most aerospace structures safety standards and pressurized vessel design criteria.) This is why materials handbooks don’t just list tensile modulus and ultimate strength; they have entire tables of different mechanical, metallurgical, electrical, thermal, and chemical/corrosive properties as well as extensive information on fatigue life, notch strength, and crack propagation behavior, and how different methods of processing, joining, coating, heat treating, et cetera impact strength, stiffness, and failure mechanics. All of this (at least the factors pertinent to the manufacturing and application) are taken into consideration rather than just picking the strongest material possible.
You would never put glass or a brittle metal into a direct tensile condition, or at least not without reinforcement or some secondary load path if it fails. In the case of a structure like a bridge, it will have to survive many millions of cycles of traffic with highly variable loading conditions which can increase over time as the traffic volume increases or vehicles get heavier, under adverse environments, with often inadequate maintenance and often cursory inspection. The reason that we don’t have bridge and other civil structure failures more frequently than we do despite a lack of support for infrastructure maintenance is because these structures are built to very conservative margins using resilient materials and redundant load paths, and designed such that when they do start failing that they show cracks or obvious ductility well before they experience catastrophic failure.
Stranger
That paper is discusses using fiber reinforcement as rebars/tendons in concrete, decking, griders, parapet elements, but when discussing cables it specifically identifies aramid (Kevlar©) fiber, which unlike carbon fiber has a very high elasticity and elongation prior to failure.
This paper does discuss investigation of carbon fiber in cable stays; however, it also notes:
However, the high and continually-increasing cost of carbon fibers limits their applications in new structures, especially in large-scale construction, such as bridges. Additionally, the sensitivity of CFRP cables to wind load is difficult to control due to their extremely light weight and high strength [12]. Considering the limitations of CFRP, the feasibility of other FRP materials as stay cables was investigated [10], including that of aramid FRP (AFRP), glass FRP (GFRP), and the newly-developed basalt FRP (BFRP).
As elucidated here, there are considerations other than just tensile strength that dictate what materials to use in an application regardless of the advantage of greatest tensile strength.
Stranger
Some of the examples in the first paper used CFRP cables (including the aforementioned Stork Bridge). But yes, the review covered multiple uses of FRP.
Clearly, cost is an obstacle and I find the second paper’s suggestion of hybrid basalt- and carbon- FRP cabling intriguing. That is likely to have even greater problems with a lack of experience, but perhaps given that a several-kilometer bridge is at least a 10-digit and maybe an 11-digit project, one could justify the required one-off R&D, training, etc. needed.
How does this differ from any other elevated highway?
I read in the paper this morning that the proposed bridge would have an unsupported suspension span of 2 miles, longer than any current one. Two mile bridges are not that unusual, but a 2 mile unsupported span has not existed. Why can’t they use cable stays from each end, I wonder?
This article mentions the longest cable-stayed bridge has an 1100m center span.
Which suggests to me that the engineering stretch needed to do this cable-stayed is even greater than to do this as a suspension bridge.
That article also has a pretty good subsection compare-and-contrasting those two bridge designs.
The stays are at an angle to the road bed so a portion of the load is horizontal, not vertical. Basically wasted tension in the cable. The longer the span, either the shallower the angle of the cable, and more inefficient, or the taller the towers, also inefficient.
Also, I suppose at some point the horizontal compression load induced in the roadbed would buckle the structure.
Do they use twisted wire cables for modern bridges? The Brooklyn Bridge is built using non-twisted, parallel 1/4” diameter strands. They are threaded at the ends and coupled with small diameter, long nuts. Layed side by side in place and bundled and wrapped, then the bundles are made into larger bundles, etc. Each strand hangs in it’s own natural arc.
The Mackinac suspension bridge has the longest span between anchorages in the western hemisphere.
By design, the bridge deck can sway up to 35 feet so the bridge does not break. Apparently this is so disconcerting to drivers who have not used it that they have a driver assistance program. A bridge employee will drive your car across the bridge with you as the passenger.
That would seem a bit much to manage for the Sicily bridge. Maybe more modern tech can do it better.
A suspension bridge does not induce compression in the roadbed. The cables pull up on the road, for a cable stay bridge, a portion of the load in the cables is horizontal, which is what induces the compression.
Ah…gotcha.
Nevertheless, the bridge deck will need to sway some. The question is how much? The balance between freaking out drivers and cost to build the bridge.
In Chicago, where I live, we have many drawbridges. When you walk across them you can feel them bounce as traffic moves across. It’s no big deal at all but still a little surprising (and they are vastly smaller than a bridge like this).
ETA: I get our drawbridges are an entirely different thing than a bridge like the Sicily bridge. Just noting that when you look at them they seem very solid and it is weird when on one to feel it move and move more than you’d think.
Steel is flexible. So’s aluminum. Good thing for both. Nervous fliers ought not watch the wings in turbulence.
A certain amount of bucking up & down in a long-span suspension bridge is cool. Much torsional motion is gonna scare the horses drivers.
I’ve walked the Golden Gate. As have millions of tourists. It can be a ride in addition to the small vertical amplitude vibrations from traffic. It’s a lively structure under live loads.
See also
Heave is not a bad thing in a bridge. Sway & surge are the scary ones.