WM_S2014_24

We ready Nikolas? Okay, so I had mentioned the King Street Bridge failure in Melbourne Australia, and I was looking through a book on design of weldments for this course and came across the design that caused the failure on the King Street Bridge that I mentioned before. And here's the girder which is holding up the bridge, and they had — it was a very big girder, I mean I don't, they didn't give me the dimensions here, but you had a plate and another plate and they basically put fillet welds along here. And so on one side you have a fillet weld and on the other side they had what they call a doubler plate. You need a little more stiffness on the flange so they put a doubler plate on, and I had mentioned to you they had welded all the way around, I'd remember that. But basically this is like a cruciform joint almost. You've got two welds on this side, you've got this very stiff very thick plate, got a lot of restraint here, nothing can bend or bow very much, and they put this weld across here. Well this weld across here was doing nothing, it wasn't carrying any load, okay. And so that's why today the codes don't allow you to wrap all the way around, because usually it's not adding any strength. It might be convenient for some things, but in any case they wrapped all the way around. They got a hydrogen crack because of all this restraint at that location, and that grew over time in fatigue because it's on the tension side of the beam, and eventually got long enough and you got a little fracture and the bridge came down. And people all over the world were upset, people don't like bridges falling down, okay. So that just illustrates some of the things that we've been talking about.

Getting back to aluminum, we talked about not-heat-treatable aluminum alloys, kind of the 5000 series and a few others. But it's not too hard to have enough alloying element in the weld metal that when it solidifies, the weld metal has sufficient alloy addition that you can get 100% joint efficiency. That means if you make a tensile bar it will stay on the base material and it will fail — the weld metal, you get equal-matching or over-matching weld metal, which is what you usually want, okay. Then there's the heat-treatable aluminum alloys. The difference here is, these alloys might be 30 ksi type of strength, these can be 50, 60, or 70 ksi strength. There's double the strength. But the precipitation, the heat-treatable ones are precipitation-hardened, which means when you heat them, the heat of welding over-ages the material in the heat-affected zone, you lose your strength. And if you actually plot the hardness across the weld, so this will be the weld, this is the hardness — weld, heat-affected zone, and base metal. So you do a micro-hardness traverse in the weld, you can alloy, make it fairly strong, and then in the heat-affected zone you're gonna have a softening, and in the base metal you'll have reasonable strength. So you're going to get softening, and you can lose up to 40% of your strength. And so you don't get — if you're using a strength-limited design, you'd lose the advantage of the heat-treatable alloy. So we have to do some things to take care of that, and there are a lot of things that people in aluminum do to try to make it possible to weld and take advantage of the high strength of the heat-treatable aluminum alloys.

But in addition, I had mentioned to you that aluminum alloys tend to be prone to distortion and high residual stresses, because aluminum has a 6% volume shrinkage on solidification. And so they do a lot of things to try to reduce the stresses at the welds in aluminum, more so than we have to fool around with in steels most of the time. So here's an example out of an aluminum design handbook where I have two different thickness plates. And in this case you make the weld — you also taper the thicker plate, rather than just having, you saw in them — no, were Street into [um, the King Street Bridge] — they basically don't do anything, just have a square into the steel, which creates a stress concentration. To alleviate the stress concentration on aluminum, because it's easier to machine, will often bevel that joint so it tapers down to get a more uniform stress distribution across that joint, okay, to improve fatigue resistance. And fatigue resistance is more of a problem in aluminum alloys than steel. Steel has a fatigue limit, aluminum doesn't, at high cycles.

Now another thing that they do — this is, instead of trying to weld two plates together because you can't get full strength, you sometimes can put two plates on top of each other. And this is, they call it a practical design. It increases the weight, it messes up the geometry of the whole structure, you may not always have time to do it. But essentially you can put a lot of weld metal in, you get a lot more area than you get with just the simple little seam between two plates like we do as a butt weld in steels. This is not a good detail. Even though you don't need all the weld metal that's over here at A and D, you end up with ends, and I'll show you later, the ends of the welds often have their own types of stress concentrators. You can do it on just one side of the plate if the loads aren't very heavy. You can do a really nice one, cost a little bit more, where you bevel the edges and distribute the stresses. Or you can just simply overlap the two plates and put a full fillet weld on either side, so you actually get a big double weld in a lap joint like this — as long as you're not going to be worried about fatigue cracking down this little slot that you have between the two plates, or corrosion that might get in there or something like that.

So there's a lot of different designs that people fool around with. I'll show you some more of how we pay more attention and spend more money in aluminum. But there's a reason why we can spend more money in aluminum. Anybody know, those of you that took class last semester? It's lighter, but it's also more expensive. Aluminum is about five times the expense of steel, so they're always more expensive structures, okay. So you can afford to spend more time worrying about the joints, because you've got apparently more stress, you know. Remember I talked about the value of a pound saved? In automobiles, two dollars a pound; in aircraft, $200 a pound; in spacecraft, twenty thousand dollars a pound. Well aluminum structures are five times the cost per equivalent volume or size as steel, and therefore you can fool around a little more with forming or machining your edges and prepping them and whatnot, and you need to.

But here's a way to look at — you're trying to make some sort of box section with an internal — well, you can't make this one, you don't have access inside to get in there to weld that. You can design that on a computer, but you can't make it in practice, okay. So here's one where instead of having the U-shaped piece overlap, you have it as an internal insert, and now it's — it's recommended, although you might end up leaving these off and just having the top two without the bottom two, because even these might be hard to get at, the inaccessible ones over here are really hard to get at, because you've got to kind of turn the corner here. You might be accessed from this side, but there might not. But a simpler one is to get rid of half the welds by just forming the whole thing, and aluminum is relatively easy to form compared to steel, so far as that goes. Here are some things where they basically scarf the joints and make overlaps, and try to make simple fillet welds as opposed to butt welds and things like that. We see they have a fillet weld over here, they basically have a curved joint. You can have a lot of aluminum extrusions and stuff that you can't afford to do in steel with its high melting point. And they actually machine the edges to make a nice bevel to get better penetration. Yep?

Student: [inaudible question]

Potentially, yes. You could, yes, okay. I mean, it depends, yeah, it depends on what you're doing, what the parts, stresses, things like that. But I mean, actually as I look at that, I don't think that's as good an example as when I just copied it out of the book now.

Well here's some other joint designs, of just intersections. This is kind of sheet metal welding — simple butt joint, fillet joint, T joint. Here you got a corner joint, but you might want to try to get both sides. The lap joint again — you form, and you make a weld up here. If you have access to the backside, you can put a weld underneath there, and you get double the weld. If you're doing heat-treatable alloys in aluminum, you may have to do such things.

I didn't talk to you about the air tanks on the cement trucks today? Well let me tell you the story of that. So they're aluminum, two of them, aluminum tank. Well here's the dome on the end, here's the cylinder, and they put a backing bar in here. This is typical so that you can put the weld from the inside — cuz it's circular, I mean from the outside because it's circular. You could do things like — this is a simple butt weld, like we haven't seal it with no problem, but in a little gnome [dome] you're liable to use a sleeve and essentially double the welds, okay, a sleeve on the outside. But on this one, basically it's just a simple tank and cylinder, and there's a story that goes with this type of joint.

There was a company that made cement trucks, and if you drive around a cement truck, when you deliver your cement it splatters all over everything including the cement truck. And so they carry about a 200-gallon tank of water on the cement truck with a little hose so you can hose off the cement that you spill on everything. They also typically carry a gallon or two of sulfuric acid or hydrochloric acid, muriatic acid, because that's even better at removing dried-on cement, okay. Except it's a little corrosive, so far as that goes, so they don't — and they don't carry that much of it. But muriatic acid is the thing that you go to the hardware store and it'll tell you that — something used to clean cement. Why? Because hydrochloric acid will react with the calcium carbonate and allow you to dissolve it, as far as that goes.

I had a situation once where a cement truck was coming through Connecticut, going over a bridge, all of a sudden the guy lost his steering and drove the cement truck off the bridge and killed himself, okay. And the truck wasn't in very good shape either, you know — take 60 tons off a bridge and drop it a hundred feet, you know, it's not good for the truck. And the question was, why did he just drive off the bridge? Well it turns out, in the steering gear, that it basically had some gears, and they had corroded, because he had been using the muriatic acid to clean off the underside of his truck and he had been dissolving away his gears, okay. Some of the stuff was leaking through the seals and stuff, and you know, that's why he lost his steering. So anyway, on cement trucks you have to clean things off. They try to use water — certainly when cement is wet and hasn't hardened you can use water, it's only when it's hardened you have to use the muriatic acid.

So in any case, they had these 200-gallon tanks they're gonna put on these cement trucks. They wanted to keep the weight down, even though you're carrying all this heavy cement. There are load limits on the highways, and so they wanted to make them out of aluminum, okay. So this company — very big company, made military hardware and stuff, but in fact they kind of went around buying up little mom-and-pop vehicle manufacturers. A little company over here in Wisconsin might make fire trucks. They buy the basic truck from General Motors or Ford, which is just the cab and the steel beams, and then they just slap on some back end of the truck and turn it into a fire truck, okay. And so it, for two hundred thousand dollars, another one builds ambulances, another one builds police cars, or you know, modified cars for police cars. Another one builds cement trucks. So this company in Iowa that had been making cement trucks was purchased by this bigger company.

The bigger company had a very sophisticated military division that made a lot of our — well, they had — when they were trying to replace the Humvees and stuff in around Iraq, this company had like a forty-billion-dollar contract over a couple of years to make all the MRAPs, okay. Big company. But there's sort of a world apart from the military division and all the 50 or 60 other little mom-and-pop organizations. And for these 50 or 60 other little mom-and-pop organizations, at one time they had one engineer for all 50 or 60. Oh, you're making vehicles here, you know how many thousands of engineers General Motors would have, or Ford. Well that's why these little moms-and-pops could compete with General Motors or Ford or somebody in making cement trucks. They could buy the basic, you know, the engine from Cummins or Caterpillar, and they could slap it onto a frame that they'd buy from somebody else, and they can buy the hydraulics from somebody else. They didn't have to have an engineer, they had all their suppliers do the engineering for them. They go to a hydraulic supplier and say, yeah we need something that'll do this, and the hydraulic supplier would sort of engineer it for him and give it to him. But they weren't really engineering the whole vehicle. There was no one engineering the whole vehicle for this — these little moms-and-pops. They had a guy in Missouri or somewhere, but he basically retired, and they hired a consultant for a while, and he probably had plenty of work to do, I guess, and then he went somewhere or something. And so they had a lawyer, okay, and he decided to do the engineering, okay.

So it turns out, they could make these tanks, and they just made a nice simple little tank — just a cylinder of aluminum, and you had a dome and it had some ports, you know, you had to bring the — and a flange on it, you had to bring the water in somehow, right. And they put a little weld here and they put a weld around here, just like you see over there. And they weld this up, and they didn't really — I mean, the guys in the shop, well, they've been farmers, okay, they know how to weld. And so they made a bunch of these. And then someone thought, well, maybe we ought to check and see — the lawyer thought maybe we ought to see if there's any requirements on the design, some specifications for this. Turns out the cement trucks are regulated by the Department of Transportation, because they have to go over the highways, right, obviously. So he goes to the US Department of Transportation, he says, we got this 55 or — this 200-gallon pressure vessel, you know, it's about twice the size of a home hot water heater tank, right, and do we have to do anything special on it? And the D.O.T. gives him an opinion, says well, that's not part of the vehicle that you use when you're rolling down the highway. That's something you use when you're stopped. That's sort of — it's not, we don't consider it part of the vehicle, and so we don't regulate that, okay. It's like kind of carrying a toolbox on your truck, we don't regulate toolboxes, okay. And the guy says, oh fine.

And then he looks himself, this lawyer, in the boiler pressure vessel code, and it says that — there are some exclusions in there, and the way he interpreted it, this even though it's a pressure vessel — it would be pressurized to 55 psi, cause they're going to use the air brake pressure. There's a compressor on the trucks with air brakes that runs at 55 psi, they're gonna have 55 psi to push the water out of here so you can hose down things. And 55 psi sort of typical pressure in a garden hose. So that worked great, they didn't have to have the extra weight of a compressor, they could just tie into something that's already on the motor of the truck. And they built these things, and they didn't go and look at what the boiler pressure vessel code said. The boiler and pressure vessel code said, hey, if you got a hole in a plate, you just got a stress concentration factor of three, you have to use a doubler plate. That's required by the boiler pressure vessel code. But this lawyer determined, well, in his engineering judgment, okay, you don't need a doubler plate, because the boiler and pressure vessel code doesn't apply, this is part of the exclusion. Well, we're not excluding science here, okay, we're excluding who regulates what. And he says, well, D.O.T. doesn't regulate me, and I'm part of the exclusion of the boiler pressure vessel code, we'll just keep building these things without any doubler plates, without any reinforcements where the boiler and pressure vessel code — so we're gonna ignore good engineering practice, okay, we're just gonna build these as cheaply as we can, and we're gonna put them out there.

Well they did, they built a hundred thousand of them and they put them out there. And aluminum, particularly in certain environments, can corrode, and they did sort of have a warning that if you — well, they said if you ever develop a leak, send it back to us and we'll fix it, okay, it's kind of what they said. But a lot of these shops have got their own welders or mechanics, and so people were welding on these things all the time. Well, if it had been come under a boiler and pressure vessel code, no one can weld on it unless they have an ASME stamp, which means they're certified as an authorized — American Society of Mechanical Engineers repair station. Well these, you know, cement truck companies don't have ASME certified welders or qualified welders. But these people didn't really know about that, and so they're welding on these things all the time.

So one of them comes back for about its third time being welded on, because it started leaking, and the guy did the weld repair, and then he takes some compressed air to check and make sure there's no more leaks. And he's supposed to be testing with five psi air, which is the typical way to test an oil tank for someone's home or something, because five psi, you're not generating enough pressure, even if this thing does break open, it's probably not gonna kill you. Well for some reason he didn't have a regulator on there, and he was using shop air, 100 psi, and he fills it up. And frankly, he should've realized, it took a while to get up to 100 psi. Maybe he kind of had left it on there, something — went, did something else. But in any case, the thing blows. He's standing right towards the end. Cap comes off, and they find him a hundred yards away up against the dumpster in four pieces, okay. It's pretty gruesome, so far as that goes.

So a couple of us go down. One guy from the National — one guy who had served on the boiler pressure vessel committee — looked at it, I looked at it, and we could see it was corroded and stuff, but it was basically an overload failure. And so we knew there wasn't five psi air, and when they later confirmed that. But we said, well wait a second, there's no doubler plates here, what's going on? And that's when we uncovered all this stuff about the lawyer had been doing the engineering, okay, and he was reading the letter of the law, not the common sense of what's good engineering practice. Well, my friend from California, who used to be on the boiler pressure vessel Code Committee, he said — was a professional engineer — he knew that this type of vessel was actually required to be inspected in Pennsylvania, which is where this occurred. The boiler pressure vessel code is really written not as a federal law, but every — all the 50 states have incorporated the boiler pressure vessel code into the law in one way or another. And so he says, under Pennsylvania law, it doesn't matter whether it has to have an ASME stamp. State of Pennsylvania says any pressurized vessel — not the ones, including the ones, that are not regulated by ASME — has to be inspected by someone on what they call the national board, the national board. I can't remember the full name, but it's the National Board of Inspectors. And there's probably a few hundred of these people who spend their lives going around and doing independent inspections for insurance companies and other people. We used to have them come around MIT to the compressed air tanks every year. They'd have to take the, you know, shut it down, take a nipple off, take a flashlight and a little mirror, look down inside, and make sure they hadn't corroded the welds. Something changed in Massachusetts laws and that never gets done anymore. So when you're standing around a big compressor tank at MIT, be careful, okay. It may have been ASME inspected for a while, but in any case — and they do blow up every time, even if they're built ASME. This one wasn't built ASME.

And Roger said, I got a problem, these should be inspected under Pennsylvania law. And so he had an ethical responsibility to go and tell some of his friends who were on the National Board. The National Board is not the same as the ASME board, a pressure vessel board, but he knew all these people, because the two of these sort of fit together. ASME code is for building the vessel, the national board code is for inspecting the vessels after they're in service. So he knew all these people. So he goes to — he's at some meeting and he had a cocktail party or whatever, he tells his friend about — there are some of these trucks out there and they haven't been inspected on a regular basis, and people are doing repair welds on them, and he thought that these should come under the national board, and should be required under many state laws to be inspected whether they gotten an ASME stamp or not — they don't have to have an ASME stamp in many states. And the guy from the national board, he says, well, how many are there? And Roger says, well, about a hundred thousand. And the guy said, you know, a hundred thousand of these would take a hundred inspectors. Well, they only had a couple hundred inspectors, they didn't have enough people on the National Board inspectors to be able to take care of this problem, okay.

So in any case, now, how could they get into this problem? Well, it got even worse, because when these things were brand new, they actually had a failure. They were doing hydro test, pressurizing them brand new tanks in Iowa, and one guy was using — we don't know exactly what pressure air — but one of them blew up and took his legs off, he became — and that was before this guy got killed in Pennsylvania. And the company had looked at it, and the attorney, using his best engineering judgment, decided it was sort of a one-of-a-kind problem, okay. It had nothing to do — I mean, well, he didn't even know what a doubler plate was. So this is a story to tell you a story, but there is a moral to it, in the sense that I see this once or twice a year in some failure I'm doing, where people have decided they don't want to pay the overhead of engineering, engineers cost too much, and you can, you know, we got guys in the shop who can design this, okay. And that's how you get in trouble a lot of times, when you don't have any technical input to your designs.

Anyway, here are some other aluminum designs for welds. In this case they put intermittent welds on a plate on top of a plate, so you're really making a T-shaped plate, but you've really got this flange here that gives you lots of weld area. This is good. Better, I'd invest in terms — here they've scarfed the edges to reduce the stress concentrations and taper the stress at the end. Here are some other things where you end up with bigger weld areas than you would often end up with if you were welding things in steel, okay.

Here's actually the more common one. In steel, if you were trying to weld a pipe onto the web of a steel I-beam, you just put the pipe up there, cut it at an angle, and put a little fillet weld around the tube. In aluminum, because you can't get the full strength, you actually weld on a gusset plate, and then on the gusset plate you slit the tube, and you end up with four long fillet welds, whereas before you would have just had a simple little elliptical weld going on steel going around the tube. A lot more expensive, but it's a more expensive structure, so far as that goes. Here again, scarfs, so you taper your stresses, things that we don't usually have to do in steel. So the design of aluminum weldments certainly requires a lot more thought for the stress and load paths than we usually have to worry about in steel — you just sort of slap it together and it usually is okay.

Now, there are a number of types of defects that you can have in aluminum, but it turns out most of them, or many of them, are not terribly harmful. We do have a problem with porosity and hydrogen. And here's the solubility of hydrogen as a function of temperature. With steel when it melts it has a very high solubility. What's the units? Cubic centimeters of hydrogen per 100 grams of aluminum — that's one of the things we use in steel, a measure of how many cubic centimeters are dissolved in 100 grams. Anyway, so there you go from 0.5 down to — semi-log plot — you go down to a factor of 10, basically more than a factor of 10, factor 10, 12, or 15 drop in solubility of hydrogen. What that does is it gives you porosity. So you will never have in a little weld — there's always more, decompose the moisture into hydrogen and oxygen, you're gonna get a few tenths of a percent hydrogen in the weld metal. Did this one actually tell me, this one, they had argon gas with no hydrogen, there's still a little bit of something because there's moisture on the surface, adsorbed moisture on the surface of the metal before you weld. Here you got a quarter percent hydrogen in the shielding gas, here you got one percent hydrogen in the shielding gas, and you get all this porosity.

Well it turns out, there's an x-ray code standard by ASTM for aluminum castings and welds. And it's basically just a bunch of x-ray radiographs. I bought my copy about 15 years ago and paid about $300 for it. You go buy it today, you'll pay about $2,000 for the exact same thing. Why? Because the code writing bodies have decided this is a good profit maker, okay, they've jumped the prices of all these things. So I guess it was a good investment, if I can, do anyone else want one of the copies, and I was willing to pay for it. Anyway, you say, well gee, that porosity, if you saw that kind of porosity in steel, it would be absolutely rejectable. In fact, this type of porosity probably is gonna be rejected in aluminum, but not really with good reason, just because it's sort of objectionable, it shows a complete lack of care in making the weld.

If you actually look at — I mean, in steel you can get hydrogen embrittlement. There is no embrittlement with hydrogen in aluminum, there's only porosity. And people have done extensive studies looking at the percentage of porosity on the fracture surface versus the tensile strength. So this is full strength 50 ksi, and if I have up to 40% porosity I lose 40% of my strength. So it's just — aluminum is a fairly ductile metal, no hydrogen embrittlement, and my strength reduction is directly proportional to the volume fraction of porosity. I can have 5% porosity, I lose 5% of my strength. It's not usually a big deal. However, when some failure occurs in aluminum, so there's porosity there, okay. So yeah, what — you do lose more of your elongation, it's a little steeper loss in elongation, so it doesn't stretch as far.

So one time I had a guy who was in a tree stand somewhere — well, actually it must have been somewhere near White Plains New York, because that's where the trial was. And it was an aluminum cast aluminum turnbuckle that was one of the things that was holding this tree stand. You know, a tree stand is, you know, some guy's gonna go out there with a bow and arrow or a rifle, go out and shoot little deers, shoot Bambi, okay. He sits in the tree freezing in the fall just so he can shoot Bambi, okay. So anyway, so this guy fell from the tree stand, hurt himself, and the turnbuckle was broken. So someone looked on the fracture surface and they found four pores. Whew, I should be concerned, I guess.

So I actually went to trial, and it was sort of interesting. I was in the witness box, and it was sort of a table that was bigger than this table — it's about twice the size of this table — and I was trying to explain, because the other side had gotten up and put these pores in the scanning electron microscope and made them about the size of basketballs, right, okay, which is a typical trick. And I had to explain to the jury how big these pores were. And I did explain, in words, this kind of graph that the loss in strength is roughly proportional to the area of the pores. You know, if I had five percent area loss I'd have five percent strength reduction. And I said, if I took this table and considered it the size of the fracture surface — magnified fracture surface — it was the equivalent of putting four ping-pong balls on the table. And I had done the calculation, okay, for the size of — that was the way I got the message across. And we won the case, because the jury — I had pointed out four ping-pong balls on something the size of a desk was not enough loss in area, it was probably less than two percent of the strength. This turnbuckle still had 98 percent of its strength, that's not why the turnbuckle failed. Actually the turnbuckle was also bent, and so it turns out he probably fell in a way that caused the turnbuckle to bend, and turnbuckles aren't designed to take a big bending load. So anyway. But a lot of this is when you get into —

Student: [inaudible question about explaining to juries]

Well, you don't have to be a jury, just explaining it to a layperson, you have to come up with something that they can understand. They look at this thing, a picture in the scanning electron microscope — things down to something they can relate. And I actually find most of my chemistry examples come out of the kitchen, okay. You usually find some analogy, the testing kitchen cooking.

If we look at other defects in aluminum welding — because aluminum has the six percent volume shrinkage — when you stop welding, then you have a little casting right here at the end of the weld, that shrinking right in the center, you'll get enough shrinkage that you'll develop what we call crater cracks. So the little decreased crater is — everything shrinking around it, large tensile stresses, and you'll develop cracks. So here's an end crack. Whenever you finish the weld, you often — well, if you don't have the right technique, really, you'll get a crater crack. What you should do, if the guy's welding along and moving along, at the end before they turn the power off, they should stay steady, they should stop, and melt extra aluminum so you form a mound and don't leave a concave crater, okay. You put extra metal in there, and now when the thing solidifies, it has enough — in casting we call it, has a riser head to feed molten metal to fill the region that's solidifying and shrinking, okay. So you're just making a little casting, you just have to create a riser, a little head of molten aluminum. So this is actually poor technique to end up with crater cracks. It's easy enough to get rid of, it's just welder technique, okay.

Another thing is, aluminum has problems in that it has a very tenacious aluminum oxide skin. That's what gives it good corrosion resistance, but that aluminum oxide skin is actually — it's actually a hydrated aluminum oxide, so it doesn't really melt at 2,000 C, maybe it only melts at 1500 C. But given that the aluminum alloy melts at 550 to 660 C, depending on how pure it is or how highly alloyed it is, that's a big temperature difference. And so when you melt aluminum and there's very much air around, you'll form this oxide skin, and the aluminum doesn't really flow, it just gets this oxide skin, and you end up with something looks like —