SSW_S2013_01

Okay so I know a few of you anyway. You are Nikolai, and you're on course of course two, yeah, as a graduate student okay. And Shannon, I know you, I know Steve, and you are Tina. In which department? Course three. A senior or — right, graduate student. Okay, who's your thesis advisor? Okay. And your thesis advisor? Okay. Well, I know Professor Slocum from way back okay.

So I'm Tom Eagar and you already met Simone Belmar. I apologize I wasn't here for the first day of class. I had to go to Livingston, Louisiana. It's not a lot in Livingston, Louisiana, but there is this half-billion-dollar facility which is the National Science Foundation's largest scientific project ever, where they're trying to measure gravity waves from space to prove Einstein's theory of relativity, general theory of relativity. So in any case, they have another half-billion-dollar facility out in the state of Washington, and the two of these together are called the Laser Interferometric Gravitational Observatory [LIGO]. And so I actually got to see the detector, which is a clean room building about size of a football field, full of all this stainless steel high vacuum equipment and stuff. It was sort of neat. It's more fun than coming to class, I'm sorry okay.

Anyway, but hopefully we will be able to learn several things and enjoy the time we have together. I'm sure that Dr. Belmar told you that 3.370 is in a modular format. He will be lecturing, has been lecturing on material selection. I will be doing the module on really on joining technology, and I'll talk about that a little bit more, and not doing so much on the fusion module right now. But there are currently seven modules, you have to do three. And so you should hopefully understand that the third module you can take any three. In fact, there are people who are taking, not coming to class. There's a student at Wellesley and she's in economics, and she decided she wanted to take an engineering class before she graduated, so she's taking this class. Because you don't have to know anything to take this class, you just have to come and listen to Tom Eagar tell stories, right.

Anyway, you actually do have to do something. You either have to do a term paper, a problem set — if you do two of the welding and the joining modules you can do the problem set — or do a presentation. And I'm getting to the point where presentations are sort of new in the last five years. I've actually been teaching the welding and joining stuff for 28 years now, sort of changed a little bit. Some of those notes go back 28 years that I just handed out. But in any case, let's look at those handouts.

The first page right on here is some comments I just took from student evaluations of years past. "A good overview of joining processes. If you want something in depth you're not going to get it here." By "in depth" I think that student meant we're not going to solve a lot of differential equations. I solve one differential equation just to prove I'm an MIT professor. But in fact, it used to be that no one had to take this course, and I got to teach it for fun, and the people who took it were taking it for fun. Unfortunately some people have to take it for various reasons now.

And they said — next person said they had no background but they'd learned something. Well, I think that's success okay. But one comment too, because you still talk about the modules, right. I was going through that yesterday actually, said that we're covering most of what is involved with structural materials. Well, I was going to actually stand point your property for a fracture, because I was trying to go through kind of the design of the overall process for structural materials. Right, it almost corresponds one-to-one with the different — the seven different modules of the class.

Yeah, what happened is when I got back from my sabbatical in 1985 in Japan, I had finally gotten tenure and I decided I could teach a course that no one cared about that I wanted to teach. Because tenure means a number of things. One, tenure means never having to say you're sorry, it means never having to write the first draft, it means you could actually start teaching something you like, as opposed to what you're assigned. So anyway, I started teaching welding and joining because that was my research area. But it turns out a few years ago Professor Shu was chairman of the graduate committee, and he came to me, said, Tom, we want you to teach more than just welding and joining. So Dr. Belmar and I have been adding modules on various things, and I think he's told you what the modules are. You need to figure out which modules you want to take. You already took one of these versions, right? Yep. So you can take — you could take the course twice. It's now has numbers of 3.370, 3.371, 3.372, fall, spring, and summer. So you can take it twice, and if we ever get up to 12 modules you can take it four times if you want to hear my stories multiple times.

But they change a little, you know, it's sort of like one of these video games that can have different endings or something, or you know, mysteries or something, and you can figure out differences. That's right. So anyway, you can read some of these things. About that, the next thing actually for this module and the next module on welding, I actually did produce a thick syllabus. I mean what it is that I went through and outlined my notes once upon a time. So this was back when it was 3.37. And then I actually went through and gave you a whole list of all the handouts. This is back when I was trying to be diligent and all this other stuff.

The next thing in here is there's my World Trade Center paper, which we probably won't go over. But the next thing, chapter 16: if you really like doing problem sets, there is a ten-question, ten-problem set at the back of this on page — starts on page 16-29. This was supposed to be chapter 16 in a book that was being written by the mechanical engineering department 20 years ago. It turns out they never wrote the book, but I wrote my chapter. And actually when they went and sent it out for review, my chapter was the only one that got a good review. But anyway, this chapter 16, if you don't want to read all this stuff, just read the Cliffs Notes version, which is chapter 16. So you know, that could be your whole reading assignment for the term if you wanted to do as little as possible with this course.

And you could do the problem set or you can do a presentation, which should be a ten-minute presentation, no more than about ten overheads, on a topic of your choosing. Ideally it should be on something had to do with joining, but sometimes people do casting and other things. I mentioned Professor Shu told me three or four years ago, expand the course. So they're now modules on casting and — I think Dr. Belmar gave you the list of the different modules. You're going to have to pick one of those modules. We will lecture every day we can, five days a week. And there are people who are taking the course who can't come on Tuesdays and Thursdays. Fine, they can watch the video. I've been videotaping my classes since 1992, and I found it was good for the students, because you can watch the movie if you missed the class, right. I always remember the student who said he used to put it on while he's fixing dinner. He'd watch the video while he's fixing dinner. Shows you, you know, what other class can you do that, right.

In any case, any questions on the requirement? Do three of the modules. One will be on video. We will lecture every day of the week. We will probably, if we're fortunate, finish up by spring break in terms of our two-thirds of the modules, if you're taking the two live modules — his module and my module. And we will probably ask you to start after spring break, we'll probably start to schedule the presentations. So if you're smart, you can get rid of this course and be done with it by April 15th, and so you can do your taxes, and then you can worry on your other courses in May when everybody's loading up on you, piling on. You can be done with this, right. So any questions?

If not, I'm going to start talking about welding and joining, which is actually the way I got my tenure okay. By the way, if you want to come and talk to me about what modules you might want to take, I'd be happy to sit down with you and discuss whatever your interests are. You said you're working with Professor Deb Senor [Devasenapathi?], with Professor Slocum. And so if there's something in your thesis or something — and I think I've talked to Shannon about things before, so I mean, but I'm happy to talk with any of you to help you pick modules. And Steve, you don't have a whole lot of choice, you're kind of stuck doing three out of four, right, because you already did three. Anyway, you must do three out of seven.

Welding and joining. Why bother to study welding and joining? Well, that's a good question. Because the way they used to control what courses you would take in the materials department is they had something called — they called a general exam. And I had to go through this, I had to take the general exam. But even when I was a graduate student, I used to call it the specific exam. Because it wasn't a general exam where they could ask you any question from anywhere in the field of materials. You knew that certain professors would always write questions for the general exam, and if you took their course, it was easy to answer the question.

The only problem with that is when I took the general exams I remember — my thesis advisor's thesis advisor was John Wulff, and he had the office that I'm in right now. And I knew they had graded the exam, and I went down — Bob Rose, my advisor, who was where Sam Allen is now, had gone to lunch, and I wanted to know how I did. So I knock on Professor Wulff's door and say, excuse me Professor Wulff — what do you want, he was an old German crusty guy. I said I was wondering if you could tell me if you know the results of the general exam. He says, pass by the skin of your teeth. Which, you have to understand, pass was all I was interested in, the skin of the teeth didn't matter a whole lot okay. But in fact I passed by the skin of my teeth. One of the reasons I did is there was a question on casting, and I answered that as if — and I'd taken Professor Flemings' course — I answered that question as if Professor Flemings had written the question. But in fact Professor Russell had written the question. And so I just bombed the question because I gave Flemings' answer to Russell's question. That's why I call them the specific exams. If you would take a particular course, you could, you know, set of courses, you could pass the exam. And that's why I have somewhat contempt for the whole general exam process. But I got through it, and many of you will go through it and you will also learn to have contempt for the process.

In any case, I used to write questions for the exam, but one of the problems was they didn't require my course for the exam, and so even if I wrote questions, the students didn't take my course unless they wanted to take my course. So I had to do something a little different. Anyway, why study welding and joining? Well, I studied it because I could get tenure that way. And there's another story to that I won't take the time on, but it has to do with the guy who I replaced, a guy named Bob Moravian [Mehrabian]. He didn't get tenure. Now he went on to be president Carnegie Mellon University, and then he went on to be chairman of Allegheny Teledyne Industries and stuff. So he did a lot better than if he stayed here, at least financially.

By the way, does anybody know who the next president of Carnegie Mellon is going to be? Suresh. Yeah, Subra [Suresh]. You know — how, Chris, you got that information? Because I was at the NSF meeting beginning of the week with these guys who were doing the largest project. They got that information to me and I gave it to Professor [Belmonte] the other day. So, but anyway, so it hasn't really been widely disseminated yet, but it is public.

Anyway, why study welding and joining? One thing is it's integrative. Virtually everything you've ever learned in any subject — art, science, physics, chemistry, civil engineering, mechanical engineering, materials engineering, ocean engineering — you can apply a lot of these things to welding joining. And I actually enjoyed doing it as a way to get my tenure, because I could work on anything I wanted okay. If I wanted to understand how fluorescent lights work, I could go study arcs. Because a fluorescent light is nothing more than a low vacuum arc okay, it's a low pressure arc. The welding arcs are what we call high pressure arcs or atmospheric arcs. Anyway, so it's integrative.

And as a result, you don't have to have any prerequisites for this. But in fact, this is sort of — I see it as sort of a capstone subject to your engineering education. You're not going to know everything from everywhere. I don't know everything from everywhere, but I'll tell you what I know and more. That's actually an old Harold Edgerton story. Freshman came up to him — you know who Harold Edgerton was, father of the strobe, you go up on the fourth floor of building eight okay. And he was actually a wonderful guy, I worked, did my thesis right down the hall from him and got to know him pretty well. And he told the story once of this young freshman came to him, she says, oh Professor Edgerton, I'm so excited to come to your laboratory, I want you to teach me everything you know and more. Okay, so I will teach you everything I know and more. And I don't expect you to have background in everything. One of my goals in this course is to help you understand that you actually know everything from high school physics and chemistry, freshman physics and chemistry or engineering, to actually solve a lot of problems on your own okay. You haven't really learned integrated though, I don't think most students haven't learned to integrate it.

So welding and joining is integrative okay. It's improved what we can make. Really the field of welding and joining goes back thousands of years, from artisans making jewelry and religious objects and things like that, to people in the Middle Ages making weapons okay, to today. Really starting somewhere in the 1880s, people started doing a lot of, a lot more welding and joining. I mean before that, metals welding was with some blacksmith forging something together okay, or someone might be soldering or brazing, but we didn't have arc welding. Humphry Davy discovered [the arc] in 1805, but no one had a source of electricity until Edison, Westinghouse came along and provided that for us in the 1880s. Once we had a source of electricity, they started using just bare wires to run arcs, and they made lousy looking welds, full of, full of pores and Swiss cheese, until people improved on it and improved on it.

The ships that were built in World War One were all riveted okay. No one would dare build a critical structure by arc welding. Around World War One, the guy who is chairman of the American — of the Emergency Fleet Committee for World War One was a guy named Comfort Adams. He was a professor at Harvard, and he was an electro engineering professor. And in fact, when he saw the need for greater productivity in building ships, in 1920 he actually started the American Welding Society, which actually started in 1921. So he's the father of the American Welding Society. So far as that goes.

But it's improved the size of what we can build okay. Can you imagine building a nuclear reactor with an eight-inch-thick wall and having to rivet it together okay? Couldn't be done. So you can build larger things. It's reduced the cost. Those over-riveted structures had a certain reliability all their own, and a lot of them are still standing, but they were a lot more expensive per ton, or per bridge foot, or building story, or whatever metric you want to use. It's reduced the cost of things, and in general it's improved the reliability in life. That's not always true.

In fact, to continue with the ship analogy, for World War Two they actually had a huge shipbuilding program. It's called Liberty ships. Anybody know anything about Liberty ships? Never heard about them okay. Well, here's the report issued in July 1946 on the design and methods of construction of welded steel merchant vessels. So just like in 1917 when Comfort Adams was asked to head a committee to build ships to get the boys over there, as they remember the song from 1917, "over there, oh the Yanks are coming." Anyway, so the Yanks actually had to go not only to Europe, but they had to go all over the Pacific, and so they built Liberty ships. They called them Liberty and T1 tankers. Oops, I just hit the wrong button, sorry.

So this is the report because they had a number of failures, and I'll show you some of the failures. But the total number of ships they built in World War Two was 4,694. They were laying a keel and floating the ship finished, full-size ship, in two weeks in the middle of World War Two. Pretty amazing. And a lot of it was because they were welding them together rather than riveting them together okay. The total number of ships reporting casualties was 1,442 okay. That's a pretty high — 25 percent, it's more than 25, almost a third, it's pretty high. Total number of fractures — there's almost, there's actually more than one fracture per ship, and these are not small fractures, these are big fractures.

And in fact, what some of these things look like — there's the famous picture, if you study fracture mechanics, of the Schenectady sitting at harbor, and it had a crack went all the way around. And they didn't sell that one out in the ocean. However, they did sail some ships, like the SS Manhattan did make it out into the ocean, and they did the same thing. And so after World War Two the US Navy commissioned a study to try to understand why these ships were failing. And it's partly because they were welded as opposed to riveted. What — can anybody figure out why you get a crack that goes all the way around if you weld, whereas with a rivet, riveted ship, the cracks don't go all the way around?

Student: [Suggests redundancy.]

No, it wasn't redundancy. It has to do with — cracks tend to run all the way to the end of the plate, and if you have a weld that's continuous, well, the crack goes all the way around. There's nothing to stop the crack once it starts to run okay. Turns out most of these cracks started at welding defects okay. And so there were three places after World War Two where big research programs were started to understand what happened to the ships.

One was a guy named Richard Weck, who became Sir Richard Weck. Was riding his bike from Cambridge through a town called Abingdon close to Cambridge, and he found some land there, and he decided he wanted to build what used to be called the British Welding Institute. Now it's called The Welding Institute. And short of the Soviet, former Soviet Union, it's become the largest welding research center in the world. And they studied — Richard Weck funded it. A guy who replaced him as president of it — Richard Weck was just a jerk, I mean he's passed away now, but why not say bad things about people who are dead and can't fight back. But Alan Wells actually was doing the research on brittle fracture of these ships, and Alan was a great gentleman okay, wrote a letter from my tenure too. But Alan Wells was a wonderful man, and he became head of the British Welding Institute.

The other place that was important in all this was the US Navy — Naval Research Lab. It was a guy [Pellini] who did all the work down there, and we'll talk about Pellini at some point. But Pellini, after he finished at the Naval Research Lab, he came to a place called MIT in his retirement from the Navy, and he wrote things about all the work he did at the Navy research labs. And the third place where they did a lot of important work on brittle fracture was MIT okay. Morris Cohen and Ben Averbach had projects to look at brittle fracture of steel. So most of what we know about brittle fracture of steels came originally from those three places okay, as a result of going from riveted to welded ships okay.

Historically, there are other important things. In 1930 the first really critical structure made by welding, arc welding, was — they called it the Big Inch pipeline. It was a 30-inch pipeline that went from Louisiana to New Jersey. It was the first all-welded pipeline, and this was a gas pipeline, or maybe it might have been an oil pipeline, I don't remember. Anyway, but you know, people were concerned, if this thing breaks, you know, what's going to happen. Before that, if they had pipelines, they didn't go that far okay. I said, increase the size. And they were — they might have been riveted, I mean going to California, I've been to some penstocks built around 1900, and these were all forged and riveted together. Great big pipes okay, still in use okay, so far as that goes.

So anyway, it's improved the size, reduced the cost, improved the reliability in life in most cases. There's a relatively new process called friction stir welding. It's been developed by the — it was invented sort of at the British Welding Institute in the late 70s. And someone started a firm and they were going to build a small business jet by friction stir welding aluminum and titanium. And I thought, oh I don't know if I'd like that, because, see what happened to the Liberty ships, if you get a crack started in the fuselage of a friction stir welded jet — which might reduce the cost okay, it's not as good. When you can have your crack on this SS Manhattan in the middle of North Atlantic and most of the people can get their lifeboats. But when you're up there flying in the air and you get a crack and it runs all the way around, it's not a good day for the pilots and passengers okay. So it turns out fortunately that company went bankrupt before they came out with it.

But generally it's — we've improved the reliability in life. You don't have the crevice corrosion that you have in joints, and it's allowed you to build things at higher stress levels and things like that okay. Another reason why welding is important —

Student: [Says "ubiquitous."]

No one knows what ubiquitous means. It's everywhere. Exactly. You get that from your Latin or what? No, it means it's everywhere. And what I always like to say — think of the largest standalone product that can be used that doesn't contain a joint. Now it's not a chair, because look at all the joints in that chair. It's not a table. So what can you come up with? Typically students would come up with needles or a nail. A what? An I-beam. Well, an I-beam is not really useful in and of itself, but it's okay, an I-beam is pretty good. I mean Frank McClintock, who is a professor of mechanical engineering, used to talk about this 20 years ago, and he, this really started bothering him, and so he thought about it and thought about it for a while. We were on a trip together to Idaho, to go to the national lab out there, he finally came with — he said an anvil okay, came up with an anvil right.

Yeah, so over the years students have come up with various things, and you can talk about things, the largest single manufactured product that doesn't stand alone, it's just a paperweight by itself. But the largest casting in the world, which has been later turned into a forging, is a seven-hundred-thousand-ton — uh, seven-hundred-ton chunk of steel that's cast to seven hundred tons and then it's forged, and then it's machined, and it's a generator rotor for a large electrical utility. It's not standalone, because otherwise, without all the copper windings and everything, it's just sort of a big paperweight. But you know, that's the heaviest single product I know that doesn't have a joint okay.

Now in fact, I was asked when I went to work for Bethlehem Steel in the 1970s to figure out how to take two forgings and weld them together. Because the heat treatment technology to make that generator rotor forging — there were only two places in the United States that could do it: U.S. Steel, Bethlehem Steel. There was one place in Japan, one place in France, and there were some places in the former Soviet Union. But back then it wasn't the former Soviet Union, it wasn't part of the free world, and so we couldn't get it there. And the growth of electricity was so fast that we were going to exceed the capacity. And even in the mid-70s, to replace those heat treatment machining centers to build these 700-ton — to work with these 700-ton castings was going to cost $2 billion for the manufacturing plant. So they wanted to know, well, could we take some of our smaller plants and weld the two forgings together? Well, now you're talking about welding two pieces that are like three or four foot square weld, right. It's not an easy problem. And I didn't bother to solve it. From my left, it came here. No, it turns out other things came along that changed the technology, and they got around, didn't need to do that.

But why did those facilities exist? They existed because the U.S. government built them at the beginning of the 20th century to make battleship gun barrels, which are long tubular things that need to be heat-treated. No company would ever have invested in that type of manufacturing equipment based on generator rotor forging. It wasn't a big enough business. Once you had the facility, and someone else had paid for it, and they no longer needed battleship gun barrels, they found another use for it was to make generator rotor forgings. But we never would have, we never could have grown the electricity business that we have today if we hadn't started out building battleship gun barrels. And those facilities still exist, but we don't use them as much.

It's ubiquitous. You can go to very small things like semiconductor chips. So this is a — what do you call this, I can't remember right now. Anyway, but this is a semiconductor chip, it's a practice chip from Intel, from like almost 20 years ago. And I think I have one of them that — in here — that actually has the — it was the original Pentium from — no, looks like I've lost — oh, here's my original Pentium. It's got a little ink dot on it, so it's a bad Pentium chip. But that's, you know, that's 20-year-old semiconductor technology. But they still use this type of bonding to do the lead connection. You've got about 400 leads going in and out of that chip. So that's all part of bonding technology, and we will talk about that in this module of this semester.

So I sometimes say it's ubiquitous from chips to ships. That's amazing. It applies to all manufactured products. And even some of the simplest products — I mean one student said a frying pan. Well, now I have to ask, what type of frying pan? If it's an aluminum frying pan, you don't have aluminum handles, they conduct heat too well, right. So you usually rivet on a plastic handle or something. It was a cast iron frying pan, yeah, okay, that's single, you know, no welds. But even a milk jug has got a joint in it. You know, blow-molded plastic, but then it's got a mechanical joint to — you know, got to be able to get in there somehow.

Another reason we like to worry about when we're talking about manufacturing is joining is a relatively large fraction. And that's really manufactured cost okay. Why is that, when I say relatively large fraction? I think Dr. Belmar gave you some data the other day — what was this aircraft thing you had in your slide, that the material cost of that component was three percent? Yes. And if you will go through it, this was the whole life cycle cost. But just the manufactured cost, my kind of rule of thumb, and it's actually in this paper, which we can — you can pass around — this is when I used to teach material selection. Now he's teaching material selection. But just like I had chapter 16 was my Cliff Notes version of welding joining, this five-page or four-page document is Tom Eagar's material selection condensed okay. Everything you need to know about material selection. And in here there's a little table that says — actually, I have a bigger table, let me put up the bigger table, it's easier to see, if I can find it.

That's not much bigger. Here it is. So this should be the same table because I probably stole the table out of the other one from my lecture notes here. Materials are 10 to 20 percent of the manufactured cost. Welding and joining can be equal to the material cost. All the cost of the materials, you know, welding is going to be equal to it. Other fabrication — what's other fabrication? Well, you got to machine it, you got to roll it, forge it, whatever you're going to do. Non-destructive evaluation — you've got to test it, make sure you did it right okay. And that can be equal to your material cost. And everything, your G&A, which is, you know, your engineering to design it in, your administrative costs, because you have to have a human resources department okay, and all this other stuff. And if you add all this up, you might make ten percent profit, or you might lose ten percent depending on how you do everything.

But frankly, materials is not a very big cost of the overall manufactured product. But the thing is, welding and joining is a fair fraction of it. But it comes at the end of the process. It's where you've already got your machining and your forging and your heat treatment and all these other things are already built into the part, it's almost ready to go, you've just assembled it. You weld it, you screw it up, and you got scrap. So you lost all the rest of that except NDE, right. So when things fail in manufacturing and welding, you've lost an awful lot of value okay. Because the cost of scrap is high, it comes at the end of the process.

Another reason to study all this is we have many failures. So why was I down looking at the NSF's — why is Caltech paying me to go down to Louisiana and eat barbecue or whatever — we had to look at their billion-dollar project. Because they have these four-kilometer-long, four-foot-diameter stainless steel ultra-high vacuum systems. And it's been operating for 12 years under an ultra-high vacuum, and come Thanksgiving they discovered they had a leak. It wasn't a very big leak. So they, along this four-kilometer module, they finally figured out by gas dynamics — because they could see the pressure rise at one end versus the other end and other places — they found, found the leak, took the insulation off the vessel. Because it's on there for to bake out the vacuum system, and also to damp the vibrations and stuff.

This is the most precise measurement ever made. It's one part in 10 to the 23rd, and they're undergoing $100 million upgrade so it'll be one part in 10 to the 24th, which is essentially at the one of the quantum limits. It's trying to measure movement of this beam tube that would be coming from a supernova explosion in space because of gravity waves according to Einstein's general theory. They can actually measure where the moon is, because this four-kilometer beam tube — it's a Michelson interferometer, it's got a big laser inside — it actually changes one millimeter in length depending on the location of the moon okay. A four-kilometer piece of steel — moon's gravity will change the length. Now it does it over six or eight hours. They're looking at gravity waves from space that might have a periodicity of one second okay, from a supernova explosion.

Anyway, when they had to align this thing, it was a feat of civil engineering. It turns out the Italians actually did something apparently similar, they had to build a linear accelerator. But you had to align this four-kilometer object so the laser could go straight down it within about one centimeter at the other end okay. And the whole thing — anyway, you have to keep the sun off of it or it'll change lengths and distort and everything. Turns out they didn't lay it horizontal with the ground, because the curvature of the earth changes by four inches when you go from one end to the other. They actually had — they had to account for the fact that they had to account for the curvature of the earth where they built this thing.

So why am I down there? Well, Professor Latanision, who used to be here in corrosion — they found the leak, and what was in the leak? Some mice had built a nest, and they had been urinating and defecating on the stainless steel, and the stainless steel corroded. And here's this billion-dollar facility and a few mice have jeopardized this facility okay. And so I have some of the world's top physicists who are hoping to win the Nobel Prize, and they are now worrying about the social behavior of mice okay, as part of their trying to win the Nobel Prize.

Anyway, yesterday I spoke about the little pressure vessel for the tomato paste, yeah, and a bunch of stress, falling, crack on the outside, chlorides in the insulation. Oh yes. On tomato paste — here, I haven't heard that story. It was a food processing plant. It's one of your jobs. Oh no, okay, that wasn't — it was, yes, it was actually — it was Manwich, it was, they were making Manwich okay. But it was actually in the steam boiler. Yeah, they had a hot water tank that was 304 stainless steel, and it was covered with insulation, and they got corrosion under the insulation, because they had to go in there and clean as a food plant, and they had to go in and clean it with Clorox and everything else. And what's better to destroy the corrosion protection of stainless steel than Clorox in an oxygenating chloride environment. Yes. Okay, so you told them about that. Okay, yes, it was Manwich, not tomato paste. I may have told you tomato paste, but anyway, this is tomato paste with spices, they cost three times as much.

Anyway, so it's a relatively large fraction of the overall cost. Many failures. The space shuttle — when the space shuttle blew up in 1986 or whatever, one of my students came rushing in, oh, did you hear the news, the space shuttle just blew up. And this is before the internet, so things didn't pass quite as quick. And my first comment was, I bet it failed at a joint. Actually I said, I bet it failed at a weld. I was wrong, it didn't fail at a weld, but it did fail at a joint, it was an O-ring seal okay. And so it turns out when something fails, it has a very high likelihood of failing at a joint.

The first year I was on the faculty that we had a guy in fatigue and fracture named Professor Reggie Pelloux. And he was teaching fatigue and fracture, and some of the graduate students who came to me said, do you know what Professor Pelloux was saying in his class? I said what? He says, "something won't fail unless it's been welded." And I thought, I'm in welding. And then I thought another five seconds, I said, oh, I will have a job for the rest of my life okay. Because they will fail if they've been welded.

Why do they fail at the welds? Anybody can think of a reason? Yes, degradation of material properties. Many welds fail because you just can't get the same properties. High-strength aluminum alloys, very difficult to weld them and get 100 percent joint, what we call joint efficiency. You make a little tensile bar and you just put a weld across — 6061 aluminum, one of the most common, you might get 50 percent joint efficiency compared to the heat-treated properties. That's not true so much in steel. We actually can get 100 percent joint efficiency in steel.

And then material selection, or my version of material selection — which actually brings me to another handout, called The Future of Metals. Back in the early 90s everybody was telling — well, the ceramists were telling the world that everything would be made of ceramics in the future. In fact, in the beginning of the world okay — and Shannon could attest to this, she's in archaeology — everything was made out of ceramics. I mean, they made pottery, and they made, you know, things like that. But in cement a few years later. But we got away from ceramics, and one of the reasons we got away from ceramics is got a lousy fracture toughness okay. And the ceramists didn't understand that. I mean, this was the 1980s, right. You know what did they know about fracture toughness in the ceramics business in 1980 or 85? Actually they knew nothing. And that was one of the tragedies. But they spent hundreds of millions of dollars, they were trying to build all-ceramic engines, internal combustion engines, and all of this. And the metals business said, huh. You know, it made no sense, but they were getting all the research money. Oh, we don't have any corrosion — you don't, okay. Actually ceramics do corrode okay. Oh, it takes a little chloride, most of the ceramics are toast, and the critical flaw size is about, you know, smaller than the width of a piece of paper.

Anyway, where was I going with all that? I will talk a lot about metals, and if you look at this article on the future of metals you'll see that 95 pounds out of 100 pounds of metal made in the world is steel. And you already started talking about some of that, didn't you? Okay, well, I just talked in general that steel was great for manufacturing. For joining it's just cheap, it's easy to weld, relatively easy to weld, and you can get 100 percent joint efficiency.

So joint efficiency is one thing. Another reason they tend to fail at welds is because welds are usually placed at the most highly stressed location of this — of the structure. Think about it, you put them at the corners right. Well, that's the easiest place to put them. Now it turns out in aluminum we don't put the joints at the corners, we actually put nodes and knuckles and things in there. And so we actually will tend to move the welds away from the corners and the edges and stuff. I talked about gusset plates. Oh, you talked about gusset plates okay. So you use gusset plates okay, and you use longer welds because you can't get 100 percent joint efficiency, then you can get the equivalent of structural 100 percent efficiency by using gusset plates.

On a gusset plate — if you have a tube, a tubular joint, and I want to join it to another tube, I actually will put a slot in it okay. I'll put my slot in the top, and I'll have a gusset plate, and I just cut a slot in the tube, and the gusset plate might look like this — let me do it like this. And if I want to have another tube like this coming in, I have another gusset plate. And I can have that length of weld along the edges of that slot on the tube. I get tremendous length of weld on the overlap. It's not very efficient from a space point of view, but I can get 100 percent joint strength okay.

If you go out here to — MIT used to have, the physics department had something called Millstone, and they gave the — this great big radar dome thing to the — they gave it to the Air Force. They gave it to MIT, had it as a physics experiment, kind of ran out its course, and they gave it away. An MIT Lincoln Laboratory took a $20 million project, which I think turned into a $100 million project with overruns, so they could do 300-gigahertz radar into space. And why do they want to do that? There's all kinds of little things flying up there in space in these little satellites, and the Russians didn't always send us a photograph and a spec sheet on each one of these satellites.

So now, it used to be we could just sort of see a blob up there that might be the size of a basketball or something. Now you can see, you can read Adidas or something on it, right. I didn't say that. Okay, oh I did, okay. But now with 300-gigahertz imaging you can actually see the antenna and things like that. I've actually seen some of these pictures or simulations of these pictures. But great big aluminum structure the size of a football field had to be made lightweight. Dr. Belmar worked for the firm that helped design this thing for Lincoln Lab. But it's — and it had to be accurate within like three millimeters on the surface, or less, something like that, over a hundred yards.

They could handle some misalignment right at the beginning, because they can adjust for it. But then when they want to change the orientation of where they were looking at, just the weight would distort it, right. They only had so much we could accommodate, life, because originally — yeah, they had active sensors and things. I got involved in it because they went out for bids, and no welding shop would bid on it, because the tolerances were just beyond what anyone could imagine. So we actually had to find shops that would, and kind of come up with fabrication techniques that would make it possible. Yeah, by a firm that does a lot of work for General Dynamics Electric Boat okay, the people who build nuclear submarines.

In any case, there are many failures. This quote from Professor Pelloux of "something won't fail unless it's been welded" — I've been using it my course for years, and in 1992 I was invited to give the Houdremont Lecture of this International Institute of Welding up in Montreal. So I had a room full of 500 of the world's top welding engineers, and I was giving the opening lecture for this conference. And I put up "something won't fail unless it's been welded," and I had Art Pelloux as the author of this. It was just dead silence among all these engineers okay, and then slowly a little ripple of laughter went through the room. But I thought, oh — but I kind of, I like it. Well, there's a reason why they fail at joints.

Welds aren't always the best material. We go to great pains to do mechanical working of metals and heat treatment, and then if we do fusion welding of it, we wipe all that out okay. So it's hard to get 100 percent efficiency, and then we put them at the most highly stressed locations. And then many times, unfortunately, the quality control is not there, and they don't always put in good welds okay. And I was just talking to a welding engineer the other day and said, you know, we generally know how to do it, it's usually someone — we know how to do it, but the people actually doing it don't always know how to do it. So they screw it up, and that's why I got a job okay.

I think we've kind of run out of time. I will finish up tomorrow with the last couple of things on why we need to study welding, and then we'll get into cold welding and space environments and things like that. So if you have any questions, stop me, I'd much rather tell a story than follow my lecture outline.