WIE_F2015_11

We will start presentations a week from Monday. This Monday is a holiday, this next Monday is a holiday, right, for Columbus Day. But this is the schedule that Jerry put together as of yesterday afternoon. This should be what's on Stellar. People want a copy, she made copies, yeah, you can pass them around. If you don't want them, leave them in the back of your scrap paper.

Anybody have any questions on the presentations? You've had fewer questions, maybe the word's getting around. You just, you only have ten minutes, but any topic you want. Hopefully you will pick something watching, you can pay if you think you want. Some people have done cooking because they like to cook, there's a lot of material science of cooking. In fact Dr. Taylor, who works with me, is thinking about doing an IAP course on the material science of bread, okay. And we'll have to get ourselves some little ovens and some mixers, and so it'll be a lab component to this, and you can eat what you cook, which may or may not be good.

In any case, the next week on Tuesday I will lecture and finish up for my part, and on Wednesday Professor Slocum, who has apologized because he thought it was classes at ten o'clock, yeah, she was sitting in his office that day, but there was a miscommunication. Anyway, he's going to come in and talk about his combined pumped water and desalinization idea, where in certain parts of the world where you have an ocean next to a mountain, you could do pumped water storage on the side of the mountain. And then you could have the high pressure of the water go through a reverse osmosis process and you basically get both desalinization, good or low salt water for use for a population, and you could easily do things for a couple of million people.

And he's looked at the different areas of the world that this can occur, and he's done some of the engineering analysis. It's a new proposal and he just came up with this summer, so it's a complex problem. And it's a way to make the whole process of energy conversion and water desalinization sort of combine them, and for tremendous energies in cost savings. So it's a complex engineering system, a new type of design that he'd like to talk about. I wanted him to come in because I consider him one of the three brightest people I've ever met, all of whom that I met have been from MIT.

There's another person that I would have included, and did include four years, in my top four. And he was chief technology officer at United Technologies. And then we got into a debate once as part of a committee — wasn't actually me getting in a debate, someone, I said something and someone else agreed with me. Actually what I said is, there's only three things that increase the wealth in the world. And I said this, it's mining, agriculture, and manufacturing. Everything else is just redistribution of the wealth. If you start thinking about, you go to the dentist, okay, they're not producing anything new in terms of material goods or anything, they might be improving your health but you're sort of redistributing the money. Financial people on Wall Street, what did they ever produce? Well, except depressions, okay. They're just shifting paper around.

Anyway, another person, the research fellow at Intel argued with me that I should add knowledge as a fourth area. But mining, which means oil business or coal or, you know, getting out of the ground, you're getting something that can be made into something. Agriculture, you're growing something that's useful, food. And manufacturing, you're taking things and putting them together, adding a lot of value, and all those things increase wealth. Whereas, you know, religious clergy, doctors, lawyers, most of these people, just teachers, just sort of redistributing the wealth, right. And it's not always redistributed very fairly either.

Well anyway, so we got in a big argument. To him wealth was money, okay. He couldn't see a broader vision of wealth. So I took him off my list of smartest people. But he was someone who was very thoughtful, one of the only people I've ever talked to where I had a hard time, just in a casual conversation, keeping up. And it wasn't that he was incoherent, it was he was about three steps ahead of where ever we were in the conversation. It was hard to track his conversation, he's a brilliant guy. He was on the board of directors of Draper Labs and other things. But I just couldn't figure out why he could only define wealth as money. But a lot of people define success as how much money you make, right. And that's not true, okay. It might be for some people, but I'm not — if I wanted to make a lot of money I'd become a financial analyst on Wall Street and then steal it from everybody, okay. Anyway, that's what they do.

And they're actually now, people, Bernanke, back to, you know, the head of the Federal Reserve, he just came out, hello, people saw it a week ago, he actually said yeah, there should have been some criminal complaints against these people. And his point was the Fed doesn't have the authority to charge people criminally, which is true, but he was head of the Federal Reserve. Now if he said that, the question is, did he go to the Justice Department or the FBI to suggest that some of these people be investigated? Well, that's another question. Anyway, these are kind of ethics questions.

So we've been talking about engineering and how it involves complexity, ambiguity, uncertainty, and safety. And so today I want to talk a little bit about safety and safety factors. There's a guy, Henry Petroski, I think I mentioned him before, who wrote a book called To Engineer is Human: The Role of Failure in Successful Design. Petroski is a civil engineer, I think he's retired now, living in Maine, but he was a civil engineer at Duke University. And he wrote a number of books. He wrote the book on the pencil, a whole book on the pencil. This was before he wrote Engineer is Human. Then the one on the paperclip, had a chapter on the paperclip. He's kind of a historian of things. But basically as he looked at things, he, and I don't know that he's the first person who ever came up with this, but he wrote a very, wrote a good book, particularly for civil engineers because that's where the examples come from, that we progress by failures, okay.

To give you an example, the very first research contract ever given by the federal government was given in like 1836 to the Franklin Institute, to try to figure out why the cast-iron boilers in the river boats on the Ohio River — we hadn't gone past the Mississippi in 1836, we were still on this side of the Mississippi — but on the Ohio River boats were blowing up and killing people, okay. And no one understood why. We didn't have fracture mechanics, we didn't know about toughness, we did, we barely knew about strength of materials, okay, in 1836. So that was one of the first contracts, was the first contract then.

And I didn't bring a picture of it, I guess I could have. The thing that sort of helped precipitate the boiler and pressure vessel code, which is the American Society of Mechanical Engineers, sells a code on telling you what you have to do, what are the rules if you're going to design and build a pressure vessel. Well in 1836 we had boilers, okay, and this — the boiler pressure vessel code didn't come along until like 1905, I think, was when they call it the beginnings of it. I don't think the first edition came out until 1916, but in 2005 they wanted to celebrate the 100th anniversary. But anyway, the boiler and pressure vessel code is about this thick. It comes out and gets renewed every three years, and there are supplements every six months. And for only seventeen thousand dollars you can have a hard copy print, and I think for sixteen thousand dollars you can have a digital version, okay. And it gives you all — for the three years it gives you all the updates and stuff. And there are people who spend their life — yep?

Student: [inaudible question]

Yes, there are some people who spend their entire life on that code. If you're designing nuclear reactors, okay, there's a pressure vessel there and you better know how to do it. Oh they do.

I brought a smaller code with me. [Tom produces a copy of the structural welding code.] This is the, since you asked the question, the — for welding, the boiler and pressure vessel code is one of the primary codes, but it only applies to boilers and pressure vessels. For bridges and buildings you have the structural welding code, okay, for steel. There's actually one for aluminum and other things. But the first couple of hundred pages are actually the code, which tells you what you shall do, what you may do. And these do not have force of law until they've been adopted by some government agency and then they do have force of law. This one has force of law virtually around the country for any bridge or building, because every state, every city, like New York City has their own set of codes, they'll write this into their codes, say if you're going to do welding you must follow AWS D1.1.

It turns out the federal government doesn't mandate safety for boilers and pressure vessels, but every one of the 50 states has adopted some part of the ASME code. But there are people — I mean I have a good friend who's, well he's probably in his 80s now, but for 30 years he was on the boiler and pressure vessel code main committee. There are probably close to a thousand engineers who attend these meetings and help update the codes, and most of it's volunteer work. And it used to be I could buy this in the 1990s for two hundred dollars, now it's six hundred or five hundred, okay. Many of the societies have learned there's gold in them thar hills, because if you're going to be in business the law requires you to follow the code, and you have to follow the current edition of the code. This one comes out every three years, okay, just like the boiler and pressure vessel code. They tried a few years ago to bring a new one of these out every year, because you get a whole bunch extra profit, right. But people yelled and screamed enough, they went back to every three years.

But in any case, does that answer your question? There are people who really do it. There's no one who knows the whole tax code in the United States, okay, it's like a hundred thousand pages. But these codes, there are plenty of designers who can pick out exactly what's there. And if you want to get a certified welding engineer degree, are not degree, certificate, you have to pick a code. You can pick any code you want among a group of five or six, and they will examine you on that code, okay. That's part of the whole test, but part of the test is to show that you are familiar in some detail with the code. It takes hours and hours to study these codes, they're not written to be conversational like reading a novel, okay. Yep?

Student: [inaudible question, apparently about how often the code is updated]

It doesn't, it's just updated, okay. They update it with these committee meetings, and someone had some failure somewhere, or some fight. And for example, I can look in this code, and this is common with a number of codes, any new edition will be in underlines, so that paragraph didn't exist in the previous edition, okay. Well the other rest of these two pages there's nothing underlined, okay. And even if it's only one word they'll underline it. So someone who uses the code every day, they get a new version of the code and they'll see the new stuff.

Also, to finish answering your question, is there any commentary? Yeah, there's 200 pages here of the code, and then there's a commentary which is about another 200 pages. And the commentary is written by the people who wrote the code, and it tells you, it interprets more than the code says directly. Here's the commentary. [Tom flips to commentary section.] See, bring down, see 4.5. So they changed, you can see they changed the five, that was the numbering change, okay, but it is a change, okay, they added something. No, it looks like they took out 4.1, okay. Whatever, 4 point, you go back to the old code and you see it changed. So if this one has a commentary that comes with it. The boiler and pressure vessel code, there are people who write books and sell them, okay, because they know the code in detail better than most people do, and it will help explain things. The codes are pretty dry reading, okay, so far as that goes, but they have force of law.

And they actually have a module on codes and standards. And I mentioned once that I was teaching this to someone down at the National — what used to be the National Bureau of Standards [NIST]. And she was head of, she was actually, started at MIT as an undergraduate in physics. And of course MIT won't admit undergraduates to graduate school, so she went to Harvard for a doctoral degree. And she was heading up like a hundred fifty people at NIST. And I mentioned to her I was teaching a course on codes and standards. She said, "You are?" She had never heard of anyone in academia teaching a course on codes and standards. I mean, I don't know, well, we have any civil engineers, and we have one or two civil engineers in the course, but I never heard anything about codes and standards as a student. But you know, as an engineer when I go out and do my consulting, I deal with codes and standards all the time.

I mean the last two days I was out in Calgary dealing with a problem of, you know, big steel transmission towers that have 500 kVa bringing electricity to the country and stuff. Well, one company in India sold a hundred thousand tons to build a bunch of these towers that are going to be in Alberta. If you figure it out, a hundred thousand, there are 111 towers and there's a hundred thousand tons, that means each one of these really big towers weighs a thousand tons of steel, which is about one and a half million dollars just for the steel. And that means erecting one of these towers is probably two or three million dollars per tower. So next time you're driving along the highway and you see these big towers, that's two or three million dollars per tower, okay, for the really big ones.

Anyway, I had to go back through my knowledge of the code and help advise people on what the contract said we had to do and what it didn't say we had to do. Because they were in a big fight of, did the buyer and the seller — the buyer and seller in a big fight about a fifty-million-dollar fight, that's the value of their claims, to decide whether they carried out their contract properly. One of the things that I know well, because I've run across it many times, is section 6.5, just to give you an example. And I was quoting it yesterday.

So there are sections in the code, and I showed you some before, they talked about the responsibilities of an engineer. Here's contractor responsibilities, obligations of the contractors, build it, okay. Visual one's talking about inspection, in here this is section six, is inspection. You got to inspect it, and then it points out there's some engineering judgment if there's faulty welding, okay. But really what I was getting to them is, they're arguing over the inspection. People in India would do the inspection, and there'd be people from Canada who were placed in the plant in India. Hey, you've got a hundred million dollars worth of product, you often will put your own owner's inspector right in the facility. And if it's around the world and you have to send the inspector there to live there for three years, so be it, okay. There are whole organizations that do that all the time.

You've heard of some of them. You've heard of Lloyd's Register, which you used to know as Lloyd's of London. Lloyd's of London now is a bunch of insurance companies, but Lloyd's Register basically would put someone in a shipyard — this goes back hundreds of years — the owner would hire someone to be in the shipyard to look over the shoulder of the people building the ship. And we now have American Bureau of Shipping, Lloyd's Register, DNV, which is Det Norske Veritas, and the French have Bureau Veritas, okay. And these are just quality control, quality assurance people looking over the shoulders of the people doing it to make sure they do a good quality job, okay. When you're buying a hundred million dollars worth of steel, you might as well have someone inspect it, right.

But it says, if you ask for inspection that wasn't in the original contract agreement, you can have it, the contractor must give it to you, but you as the owner have to pay for it. You didn't put it in the purchase order to pay for it to begin with, it's extra, and you're going to have to pay for it. Unless they can prove attempt to defraud or gross non-conformance — you're just doing a terrible job, you were careless and you didn't even care about being careful, or you're actually committing fraud. Those things happen, but in general — and so, you know, there's a whole law that's built up around contracts and things. Because like, if you've read about the government contracts, this contract's over budget by three hundred percent, right. The Big Dig was over by, what, seven hundred percent or something from the original estimate. And this leads to lots of disputes, okay.

In fact I'm — the Big Dig, they brought me in. What was it, they had some hangers — oh, that, yeah, they had some hangers in the tunnels they were holding the roof up. This is before — I was also involved in the one that collapsed and killed the woman, but this was before that, while they were still under construction. And they, some of the big steel bolts that were going to hold up the roof were supposed to be galvanized, that is supposed to coat them with zinc in a particular way. And the contractor had done it out of sequence, and the question is whether the threads on the bolts — these are big bolts like an inch and a half in diameter — whether they were going to rust in service.

Well these are all supposed to be hot-dip galvanized. Hot-dip galvanized, you just dunk it in a bath of molten zinc and you get a pretty thick layer, like two or three thousandths of an inch. That's the guardrails along the highways and stuff, that's hot-dip. Or the paint buckets, not paint buckets, but the galvanized buckets that, you know, on a farm they just carry water, or have a trough and the cows and the pigs drink water out of the hot-dip galvanized bucket. There's also electro-galvanized, and if I looked — a lot of this is aluminum, but if you saw any steel tubing it's zinc-plated. Well that's one-tenth as thick, okay.

So I was in this meeting, and we're — this was going to bankrupt this little company that had made these bolts for the Big Dig. I mean if they had to go in and undo, you know, dis-install — this is what some of the engineers, the Big Dig, were saying. And they said, "We're going to make you uninstall all this stuff and redo it." And this little company was sort of a mom-and-pop company, they were going to go bankrupt under this. And so we had a little meeting, and I was hired to go in and kind of talk about whether this was going to really affect the corrosion up there in the roof, no, in the ceiling. And we were meeting with the chief engineer of the Big Dig, and the underling engineers were telling me, "Wait, we put it in this back and you need to do this, you did it out of sequence and blah."

And so we go in the meeting and I said, you know, I've seen some of the construction, you've got some electro-galvanized up there for the electrical stuff. And, you know, I talked about the throwing power of zinc and how you have ten times as much on hot-dip galvanized, and the fact that they had a little area, a quarter of an inch wide, that didn't have some zinc on it would all be protected. There was plenty of zinc to protect it, because compared to the electro-galvanized you had ten times as much, okay. And I said, you know, if anything's going to start rusting up there it would be the electro-galvanized, because it has less zinc, okay. And I said, you're going to start seeing rust from the electro-galvanized a lot sooner than you're going to see it from these. And you can see all these engineers of the Big Dig, and I said, "Oh, you guys already have rust!" They had to admit, the electro-galvanized was already rusty, they hadn't even finished building it, okay.

So okay, we ended up, we know, of coming up with a fix that wouldn't bankrupt the company. They didn't have to take everything down and everything else, we came over a fix that would make the underling engineers happy. And basically the chief engineer was just sort of the arbiter between me and the guys who were trying to play hardball, "And this is what the contract says." Well, you know, there's always things that go wrong, and then an engineer, it says in the code, the engineer has to exert some judgment on whether it's important or not. You can have non-conformances, and then you can have waivers, or you can have new ways to fix it. In fact that's what this fifty-million-dollar dispute in Canada is all about. There were certain managers in Quebec who, "Oh, we weren't, we want exactly what we paid for," whether it was important or not, okay. Some things are cosmetic and hidden, and ain't nobody above the ceiling, which is basically what happened in the Big Dig.

So this is where engineering is a lot different than just being a research scientist at MIT. They teach you how to be a research scientist, right, as if that's what people do. Ninety percent or ninety-five percent of the engineers are not — what I call the wannabe scientists — but there are people who actually go out trying to make something work and solve the problem, and trying to decide when something goes wrong what do you do about it, okay. I used to work for a division of Johnson & Johnson that made medical instruments, and they'd have a manufacturing hiccup and the question was, they'd come to me and they'd say, "Do we have to throw out all this half-million dollars worth of parts we made, or can we do a fix?" And I'd have to help them through and decide whether there was a fix or whether you had to throw it out. Sometimes I said, "Sorry, you got a half million dollars worth of scrap." Sometimes I said, "Well no, they can do this, but you have to, you know, you have to get into the details." And you can't always write down equations for this, but you've got to be able to justify it.

And there's a number of times that I've been asked to go to Washington and talk to the regulatory agencies, in part because I'm at a university, I'm sort of a neutral party, okay, about these things. And sometimes there's a lot of money involved. I remember one day I didn't have to go to Washington because right up here by Burlington Mall the Federal Aviation Administration has — they have offices around the country, but these offices have to approve parts manufacturing if it's not the original, with that, for everybody.

If you want to sell an aircraft in the United States, unless you're going to limit it to flying it yourself or your family, you can — if you're flying it for yourself or your family the regulations say you can build anything you want, because, you're, it's okay for you to kill you or your family. But if you're going to sell it for other people to fly in it, you've got to follow our rules, I'm not kidding, okay. There was once a — I'm really going afield here, but there was once a, some people who wanted to be able to fly at Mach 2.5, same as like an F-14, okay. But they, first of all the government wouldn't sell them an F-14, but also it would probably cost 70 million dollars. Although some of these people, such as, well, Harrison Ford, but some of these movie stars and stuff, they own their own helicopters. I mean, when I went through helicopter flight ground school, Harrison Ford had been there at Bell the week before learning how to fly his helicopter, okay.

But there are some of these people who would like to fly at Mach 2.5. Well, you could design — you can design your own private aircraft, but the rules say you have to build half of it yourself. So they set up this factory in Reno, Nevada, and you could buy one of these for like 10 million dollars, and you would have to come in and turn the screws and be able to say that you built half of it, even if you could, you know, afford someone else to do it, the rules say you have to do it. So they, anyway, so they're trying to qualify this, and this was basically a — an aerospace engineer at Pratt and Whitney, a General Electric — and they kind of built an aluminum frame and put a seat in front, and then they had wings, okay.

And they had lost their first test pilot over the Nevada desert. And I got a call at six o'clock one morning, I was at work, and this NTSB accident investigator says, "Tom, I want you to look at some parts." I said, "Sure, what do you want, you want to come by?" He says, "I'm down here at 77 Mass — I can be there in five minutes." I said, "Oh okay, I'll meet you in 8-140," all right, what, 33. Anyway I go down there, he shows me these parts, where the owner of the company had died — he became the test pilot when the first test pilot died, and now he was dead. And his widow wanted to understand why her husband had died.

Anyway, they had like a hundred million dollars — you had to put a two-million-dollar down payment, even though this thing had not even been proved out, to buy this super-duper plane that you could kill yourself in. And anyway he brings by these parts, and I look at them in about 50 minutes, I said, "There's nothing here, this is all overload failure. So tell me more about this accident." He said, "Well, they think they had wing flutter, they don't know exactly. He's at 40,000 feet, he's going Mach 2, and they think that wing flutter, and if you get wing flutter that starts getting into an unstable mode, they think that maybe one of the lead weights on the end of the wing flaps had come off." So the wing flaps are the things that allow you to tilt and stuff, but one of the lead weights was there to dampen some of the vibration, and if it had come off it would be unsteady, be asymmetric weight, it wouldn't be balanced. And they think that's what caused the wing to rip off.

And the wreckage was spread out over 50 miles of the Nevada desert, okay. Because you're at 40,000 feet going Mach 2, you know, the first parts that fall off land here, and you know, later parts laying somewhere else, so it's hard to get everything. I said, "Well, why don't you try to get me the end of the wings?" Well they brought in the two tips of the wings about three months later, and sure enough, that adhesively bonded a one-pound lead weight, and I could look at it. I didn't, we didn't have the lead bit because it was gone, but I can look at the adhesive joints. They didn't get enough adhesive here. And why they didn't put a screw in there to join the lead to the aluminum is beyond me. But to me that was the probable cause. I mean, that was helped out by an aerodynamics guy, but I said, "Yeah, you got bad adhesive joint." Anyway, the whole company folded, all the people put in their two-million-dollar down payment, they lost it.

These things go on all the time. There was a guy I know, passed away now, but, actually no, it was another guy I know who actually bought into this too, and it was only — it was an all friction-stir-welded titanium skin plane, like a little Cessna, but it was, you know, it's going to be very light and stuff. And you had to pay a million dollars down before the thing had even been designed, okay. So there's lots of these little, well, I call them scams, but they wouldn't call them scams. Anyway, when you're doing these things there's lots of uncertainties.

And as Petroski goes through and talks about the factor of safety, this is chapter nine, you know you'll have it on Stellar, but it says, "Safety in numbers. While engineers can learn from structural mistakes what not to do, they do not necessarily learn from successes how to do anything but repeat what they already did." And this is his thesis, that we learn from failures. You know, he uses the Tacoma Narrows Bridge and other things. And we have something called a factor of safety, where if we think the load, the required load is going to be a hundred pounds, we make it so it will support 150 pounds. That gives us a 1.5 factor of safety, fifty percent extra, because there's uncertainties out there in what we design.

And so it turns out back in 1849, well, one of the things they can do — he talks about proof testing. If we want to have a rope that we're going to lift something, and we might need to lift 1400 pounds, we might design a rope and test it at three thousand pounds. And if it doesn't fail, we just proved that one rope will hold 3,000 pounds. And the chances of it failing if you load it again at 1400 pretty small. And proof load testing is done all the time. In fact I ran into that the last two days out in Calgary. And you do careful inspections, but you need to know how to define failure. And it might seem obvious, it broke, well that's a simple, but sometimes if you have a fatigue crack that gets to a certain size, then it has failed because you don't want to put it out there any longer. And that's what we do all the time on an aircraft, we inspect them, we know what size cracks are acceptable. We do it in bridges and buildings.

That's actually why I brought this thing in. [Tom holds up the structural welding code again.] In this 200 pages, they're very explicit, in chapter — in chapter 6 it tells you, if you're going to do ultrasonic testing what size flaw you can accept. If you're going to find porosity, well if you find one pore that's easy. What happens if you find a cluster, and how close can they be in a cluster before you say that's too much? The code has developed historically parameters that tell you, this is acceptable, this is not. If it's not, it's going to be classified as a defect. But just because you find an imperfection doesn't mean it's a defect, okay.

Now in 1849, the Royals [Royal] Commission in Britain asked a bunch of bridge designers how big safety factors should they have. Well this is when they're designing bridges. If they had — they designed a girder, they hadn't really worked out all the formulas like we have now. And they might just put the girder across two pieces and just load it up with 10,000 pounds of something and see if it breaks. And they asked, "What type of safety factor should you use?" And they asked some of the big names in the industry at the time, like Charles Fox the engineer, the Crystal Palace and other people. And they came up and they said, "Well, I think three." Another one said seven. The Commission came up and said, "Well, we're going to use a safety factor of six." So in 1840 and 1850 we were using structural safety factors of six.

Now you have to remember that is — we're using all kinds of material, steel that didn't have much toughness, cast iron that was very brittle, just because inherently it's brittle. They talk about the Washington Monument, well that's a, that's a different topic I'll get to later. But actually he talks about the Washington Monument in crushing has a safety factor of 9. And then Washington Monument is designed in the 1870s, okay.

In any case, there are safety factors. And if you go look at the codes today, the factors of safety — bridges in the UK in 1850 had a safety factor of six, the ASME boiler and pressure vessel code, which is one of the most sophisticated codes in the country and includes nuclear reactors as division three and things like that, for ductile iron, not gray iron which is very brittle, ductile iron which is almost as good as steel, you have to have a safety factor of five even today, because it's a casting and castings can have big imperfections. Whereas wrought materials, you've squeezed out the imperfections, and so steel it might be a safety factor of two. But the boiler and pressure vessel code, it's this thick because when you design something, it will tell you what you have to do in the design and what's factor of safety above your anticipated load, okay.

The structural welding code basically used to have two safety factors. Dynamic loading, which is a bridge where you have trucks and buses and cars going over it and it vibrates, and that's 2.0. Anybody know why the structural welding code would use 2.0 for dynamic or cyclical fatigue loading? It's because steel actually — if you look at the old way of looking at stress concentration or at fatigue life — steel has something called a fatigue, this is not the way we look at it now, this is the way we looked at 50, 60 years ago, but we had what we called an SN curve. [Tom draws on board.] S is stress and N is number of cycles, and actually it's log N, so it's ten thousand, a million, okay, you're going on a log scale.

But you'll have a stress that would be your uniaxial tensile stress, and you come down — and I kind of went too far — but at about fifty percent, well, 57 would be down here, you get to the infinite life, okay. Steel under constant amplitude loading will have infinite life. If you have a lot of variable amplitude that will then also have variable life. But anyway, that's one of the reasons they use the factor of two, whereas on a static design like a bridge, building, there's not anything going over the building that's going to cause it to vibrate like a bridge. And so I have static — and static design is 1.67, five thirds.

Now if I go through the old structural welding code from 15 years ago, they would require the welds to have twice 1.67, or 3.3, as a safety factor. Why? Because just like the castings in the cast iron, the welds can have porosity, they can have imperfections that are larger than you expect to the base metal. So the weld metal has to have twice the safety factor that you would have in the base metal, okay. That's the way we used to do it, and we still do it sometimes. You can do it either way.

OSHA, if someone's going to be up more than six feet in the air, OSHA by law requires 4.0 on safety. Ladder I go to the hardware store and buy, a ladder will have a 4.0 safety factor. If it's rated for 225 pounds of person and tools on there, you can put 900 pounds on it, okay. The odds of that failing are pretty slim, as 4.0 safety factor is really big.

Now it turns out tree stands, you know, hunters who want to go out and kill themselves — actually they want to kill something else, but sometimes they kill themselves — tree stands are only a factor of two, okay. OSHA cares about scaffolding for construction, but OSHA doesn't regulate tree stands. It's sort of like build-your-own-plane, kill yourself, knock yourself out. But if you're going to be doing it, you're going to have workmen building a building, you've got to have them with safety harnesses and things like that.

In fact I was watching on the construction of the nano lab, I kind of come up an elevator there from the parking lot or parking garage, and they had a guy, and he was just a little bit more, he was like seven feet off the ground, but he had a harness, you got a D-ring, and he had a cable so if he fell he wouldn't fall more than six feet. Why? Because they've learned, if you weigh 200 pounds and you fall six feet, that's 1,200 foot-pounds. I did that calculation in my head, okay. And if you look at historically when people fall, they can get fairly serious injuries when you have 1,200 foot-pounds, okay. 1,200 foot-pounds is a lot of energy.

The X-33 spaceplane started out with a 2.0 safety factor. They had some manufacturing problems, this was a fifty-million-dollar structure, and they knew they had some big defects in it. So NASA sharpened their pencils, said, "Oh, we still got 5% safety factor above," okay. So they went, they tested it, and it actually passed until they started to warm it up, and the thermal stresses from going from liquid hydrogen temperatures to room temperatures, it failed because it didn't have enough safety factor, okay. But if it had been manufactured properly it may not have failed, okay.

I put in, the New Orleans levees — I hear about 1880 my great-grandfather worked, he was a civil engineer working on the levees — and I read somewhere after Katrina that the safety factor on the levees was 1.3. That 1.3 is one of the lowest safety factors I know of, okay. Typically safety factors are around 1.67 to 2. When it's like hoisting equipment above someone's head, or someone more than six feet off the ground, they'll go to 4. Some cases when you're working with castings and pressure vessels where the consequences of failure are pretty severe, blowing up a building or a city block, safety factor of five.

Anybody have an idea of what the safety factors might be on a helicopter? No, it's actually very high. I was surprised. And actually it's a trick question, I said safety factors, there's not one safety factor, okay. You're weight-critical, and so they're going to look at how critical the failure would be. What is your answer? Okay, most of the things I've seen are around 10, okay.

Now some things are not, and some things are critical, like the mast. There's no redundancy on the mast. If the mast cracks, you're going down, okay. And so I don't even know what the safety factor is, I suspect it might be something like three. The problem is the mast is big steel, okay, it is heavy, and so they can't use a factor of ten. But most of the structural stuff, you know, the spars, you know, for the — when I've done the safety factor calculations on these things when I've had a failure, it's up around ten. But these are vibration machines, they sometimes cause helicopters fatigue machine, and so, and the consequences of failures, it's very easy to kill yourself, particularly if you lose your rotor. There's not much safety after that. If you actually still have the mast of the rotor you can do an autorotation. Anybody know what autorotation is? Yep?

Student: [explanation of autorotation]

Yeah. So if your engine's working, you're spinning the blades and that allows you to have lift. But if your engine quits working, you basically — if you have some forward momentum — if you're just hovering you're dead, okay. But if you have some forward momentum, you can put those rotor disks, the rotor blades, into the wind, your forward momentum, and the air coming through will cause them to freewheel, and they will give you some lift, okay. So I did an autorotation once, and it's not much different than a regular landing if you have enough forward momentum.

There's a little part of the envelope of height versus forward velocity where you can autorotate, and there's an area which is called the dead man's zone, for obvious reason, where you can't autorotate. But helicopters are not as unsafe as you might think. I was surprised when I started doing some calculations, they have so many parts and they're vibrating, they actually have quite an amazing safety factor for given the situation.

So now what's happened in recent years — that's why I brought these two books — the old way was called allowable stress design. We did that from the 1920s probably up through around 2000. And the first nine editions of the American Institute of Steel Construction Manual of Steel Construction — and it's a really exciting manual, you need to own a couple of these books, I mean, one for home, one for work — you know, tables of what the sectional properties of an I-beam or a channel or something like that. But it also has some sections in here about design requirements.

And what we did in the old days under ASD, allowable stress design, we basically said we're going to use 1.67 or 2.0. So the structural welding code used to say, before 2000, you use ASD — they didn't even call it ASD. You basically didn't know what the stresses were going to be, you're going to give sixty-seven percent or a hundred percent extra capability in your design, depending on whether you are statically loaded. And it doesn't matter if the stresses are coming from north, south, east, or west, top or bottom, you're going to just use the same safety factor. And historically we found 1.67 statically loaded, 2.0 fatigue loaded, was good enough, okay. We moved all the way down from six in 1850 to one hundred fifty years later we were down to two or less.

But then in the 1950s, this engineer in the Soviet Union said, "Now this doesn't make any sense, we're using too much material." And as an engineer he wanted to do it better. He said, "We can — if we know whether it's tension, compression, shear, bending, there are different types of loads, you can write down the stress tensor, okay, and you can find each one of those in the stress tensor, rather than just assuming the loading can come from any direction. We often know what direction it's going to come from, and we can do a better job design job." And he came up with what is load and resistance factor design.

The AISC manual is now the Manual of Steel Construction, this one's in its third edition. Actually I think it's in the 5th edition. The 9th edition was the last ASD design. If I go to the 2010 welding code, it will say that I can use either allowable stress design — there across the top — or load and resistance factor design, okay. And this has a safety factor of 1.67, this has a resistance factor. And you have to know a lot more about the loading on your structure. But if you know more about the loading on your structure, you can save about one third of the weight. So there's a lot of money here, okay.

If you're talking about these transmission towers out in Alberta, you're talking a hundred thousand tons, you know, you could be saving 30,000 tons, you know, you're talking tens of millions of dollars worth of material. Now it turns out that there have been some interesting failures because of this stuff, and maybe — sense it's towards the end, I'm going to talk about on Tuesday. I mean, there are books now written on this. You know, we'll look at some of the introduction of this, Steel Structure Design ASD/LRFD, and this guy compares the two, and he shows that if you know more — but the problem is, that I'll talk to you about next week, is it's a more complex system.

In general you need to — most companies have proprietary computer programs to do the LRFD. ASD we could all do on a sheet of paper with pencil, okay. Well, actually that's why we did it then, they got to a PowerPoint, the computers were doing that in the 1990s. But now LRFD, it's just the American Society of Civil Engineers document 360 is a little too complex. This is where you have to ask, well, do you have to have some expert who really knows the whole thing? And the answer is with LRFD, yes you do. Or you have some computer program that tends to be a surrogate expert for them, because most engineers can't do it. I can't do LRFD okay on my own, okay. I haven't studied it. Some people coming out of school can, but most of them, unless they're doctoral-level civil engineers, are going to be doing something that's sort of a, they're going to be using some canned program. And there are some interesting failures that come about because of canned programs, okay. So I'll talk about some of that. Okay, I'll see you next week, have a good weekend.