Engineering Ethics: Notes on "To Engineer is Human" aka. "When Engineering Fails"

• We take successes for granted, but failures are headline news.

Notes mainly from/while watching the movie

Factors in Engineering Failure

Instructions from teacher: Cite at least 2 examples for the factor that you choose.

• Cycle of success and failure
  • Shuttle take off (it had worked so many times before)
  • Failure is the key to making new engineering discoveries
  • We work and innovate until something does go wrong
  • All Accidents lead to more successful designs
  • B25 Bomber. Empire State Building
  • Safe designs evolve into unsafe ones because people think that since it worked, we can do it with less material (e.g. on bridge)
  • Brooklyn Bridge learned from other failure[s]
  • Engineers continued to build longer and more thin bridges
  • Galloping Gurdy: Limiting factor wind.
  • Other bridge had ugly steel added
  • We take successful designs and evolve them into failure. Then we start making them stronger.
• Anticipation of failure mode
  • Challenge is to predict fatal detail and prevent them in design.
  • Bridge models were used
  • Stevenson's tubes worked well.
  • Safety is only assured after all possible modes of failure are considered.
  • Robert Stevenson models
  • Antonio Penizze lbrar (?)
  • You have to imagine how something will fail.
  • Pyramids: angles were build at less steep angles to prevent failure
  • Gothic chapels
  • We need to anticipate how something might fail and take steps to avoid
  • Governing mode of failure: what fails first
  • St Lawrence Bridge - 75 men killed (while building?). Was eventually finished
  • Bending is a combination of push and pull
  • Failure mode: Failure by torsion
  • Hotel in Kansas City: Sky walks from ceiling. 2 Skywalks collapse. Hundreds injured. Bolt pulled through. Design was changed.
  • Failure mode: Variability: Things that go wrong later on.
    • Knives had cracks, looked different. Cracks turned out to be benign.
    • Eyeglasses cracked and broke.
  • Failure mode: Fatigue
    • Cracking grows. Can lead to catastrophic failure.
    • Paper clip breaks after repeated use.
    • Big Ben clock shaft twisted millions of times before breaking.
    • It is impossible to predict exactly when something will break
    • Experiments help you learn, but even better:
  • Factor of safety
    • needs to hold 10 tons? Then make it hold 20 tons to give it a factor of safety of 2.
    • 6 is adequate safety factor
  • Proof testing
    • Bridge was weighted before opening
    • Success: London 1851 Crystal tower. Joseph Paxton won out of 245 other designs. From work on greenhouses. He thought through all (failure modes?). Cast iron were all tested (more likely to fail). Wrought iron were spot tested. Soldiers walked of a part. Created to be dismantled. Lasted until 1936 when it was burned down.
    • More successes:
      • 1st iron bridge (still stands?)
      • Brooklyn Bridge
      • Empire State Building (build rather quickly)
      • Paris Tower
      • (they?) used factors of safety to handle unexpected risks
    • Modern buildings are constructed to look good.
    • New Materials, new techniques.
  • Maintenance:
    • Will it continue to function?
    • Bridge against sea air
    • Bad maintenance can cause failure
    • Using fork lifts to service airplane engine instead of taking engine off
    • 1983 Connecticut bridge
• Computer Engineered Disaster
  • Slide rules helped give a feel for scale.
  • Using computer gives illusion of looking at everything
  • You shouldn't think that the computer will find all (modes of failure?).
  • Computer can give false sense of security
  • Roof in basketball stadium failed because roof had ice and snow. (That structure would have never even been attempted without the aid of computers. They didn't think of checking for what would happen when snow or ice built up on the roof).
  • Computer Aided Catastrophe
  • Higher risk, less trust of computer.
  • Non-technical side can ignore
• Engineers must be cautious and continue to worry and lose sleep over stuff. That's what engineering is all about.

Notes from discussing the movie 2 days later in class

Intro and instructions

• Engineers participate in projects that lead to problems.
• There is risk involved.
• "You can't put a price on human life."
• We risk our lives when we drive our car.
• Acceptable risk (what is it?)
• Sovereignty of God comes into this.
• Try to be realistic about trade offs
• Exodus 21:28,33 - Bring verses into essay if you like. Start putting some spiritual thought into this.
• Books: "To Engineer is Human". "The evolution of useful things".
• Ethics section from training manual of PE exam. EIT (Engineer in training) for 4 years, then PE exam. Engineer who faked data about slab repair and lost his engineering liscense because of it.
• Write report on movie. Report has to be from movie.

Factors (from class discussion)

1. Cycle of success and failure

Examples

• If you built something that worked. You would build on it until it failed. Then you would would fix that problem and (learn from the failure?)
• Bridges: Each design they wanted longer and smaller
• Ballroom Kansas City Catwalk (better for the next factor): Rod didn't go all the way from floor to ceiling
• Tacoma Narrows
• Pyramids: Goal was as high as possible (steeper angle gets you higher).
• Drive to make things better and better eventually leads to failure.

How does this apply to you?

• If you have a design and want to expand upon one part, you need to think through how the changes will affect the rest of the design.
• Do research and find how similar have failed.

2. Anticipating Failure Mode

• Looking for things that could go wrong.
• Governing Failure Mode
• Overbuild with safety factor

Examples

• Tacoma Narrows: didn't think about it twisting and oscillating at resonate frequency
• Beam Kansas City: Design changed messed it up. Was hard to make the rod long. The rod was fine and didn't break. Pressure mass doubled and thus it failed. Was not checked.
• Plane that lost engine:
  • Caused by fatigue.
  • Maintenance procedure: Fork lift shouldn't have been used.
  • Should there have been more clear instructions like "Don't step on wing?"
• Stadium that got snowed on (Also in category 3):
  • There are loads (load charts) for snow loads
  • They could have checked.
  • You need to anticipate how something might fail to prevent
  • You almost have to test something to be able to address and fix it.
  • You can make scale models. But we don't really know until the real one is built
  • Use factor of safety
• Power Parachute
• FMEA: Failure Mode Engineering Analysis
• Step up for old lady:
  • For rehabilitation. Out of PVC.
  • Humility helps. It is arrogant to think that since we did calculations it would work.
  • "We didn't think of racking"

  How does it affect you

• Motivates you to test better. You want to test more. Student mentions boat stern that failed.
• If you anticipate failure, you won't ever blindly trust your calculations. You will have a sense of humility that you might have missed something.
• Wake up in cold sweat. Do everything you can to prevent failure.
• Test above and beyond what is expected.
• Build Factors of Safety
  • Factor of safety = (Design Load) / (Expected Load)
  • To make a school desk not fail when a 300lb person stands on it: FS = (600lb / 300lb) = A factor of safety of 2.
  • Think about it material was wet, shoddy, or improperly built. A factor of safety makes it still OK even if everything does not go perfectly.
  • Bathrooms guard against vandalism by using different bolts
  • You have to consider cost
    • You could save money by using a 1.5 Factor of safety.
    • If FS=10 then priced out of market. You've made a shop desk, not a school desk.
    • Trade off: Important to think about what level of risk is acceptable

Computer aided disaster

• Computers add a feeling of extra safety because of so many calculations
• They give you a false sense of security that they've covered everything
• Before computers: It gave an engineer a better feel for the numbers
• With Computers: We don't feel numbers 20 vs 200
• We starting to get wise about to to use computers good (well)
• Finite Element Analysis: Used to be wrong 50% of time. Depends on how you set it up.
• You often need to test physically
• Understand software limitations
• Software may not check for all variables
• Basketball arena that fell under pressure of snow: They couldn't have tried without computers. They just didn't test for buckling (they could have).
• Example from Calc:
  • dpert (?) 3cm = 0.03m J(?) = F * 0perp(?)
  • Could be bigger or smaller. Mesurment could be off 10%
  • Significant Figures (Sig figs) are important: 0.254329071 N*m The "329071" means nothing. This is not any bigger: 0.254427898 N * m

How does it affect you?

• Know limits of software
• More than pull and push, torque too. Check for all.
• Analyze as many ways as possible. Think of other ways.
• Cross checking: you get 3,000 RPM (for lego NXT mindstorms motor testing). You need to do reality checks to make sure things make sense. Convert 3,000 RPM (rev/min) to RPS (rev/sec) = 150 rev/s (way too high).
• Make sure things make sense.
• If some small engine needs a radiator the size of a wall (according to the computer), you probably calculated something wrong.

Bottom Line: Humility and trust in God is called for. Accept responsibility for failure and move on.