When Things Break


Published on

This is the presentation file that generally accompanies the short course I teach on failures. For more info on the course, http://is.gd/things_break

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

When Things Break

  1. 1. When Things Break Dr. Ron Graham, editor a Clarity Strategic production
  2. 2. Contents  What is an engineering failure?  What are some examples?  What lessons can be learned?  How can failures be avoided?  Can engineered systems be truly “safe?”  Oh GOD! I've had a failure! What do I DO??? The conscientious, effective engineer is a virtuous engineer. - Samuel Florman
  3. 3. What is an engineering failure?  failure (n) == malfunction + loss of opportunity  risk (n) == the chance of something going wrong  hazard (n) == what happens if something goes wrong  Murphy (n) == the guy who says something will go wrong  bug (n) == what makes software not work as advertised
  4. 4. What are some examples?  Tacoma Narrows Bridge  Space Shuttle Challenger  Kansas City Hyatt Regency skywalk  Union Carbide Bhopal, India  THERAC-25 radiation device These are just a few of the popular examples – those students have really responded to in the past.
  5. 5. What can be learned?  Better design techniques  Enhanced safety precautions  More rigorous testing (and simulation)  More enlightened management  Discovery of new failure modes (not just what initiates the failure, but also what propagates it)  Better estimates of cost and risk For we can demonstrate by geometry that the large machine is not proportionately stronger than the small. - Galileo
  6. 6. How can failures be avoided?  Design  Redundancy and spare parts  Watch out for discontinuities and interfaces  Problems of scale (not just changing size)  Operation  Massive manned tests  Training and retraining (with updated manuals)  Carefully designed rules for alarms
  7. 7. How can failures be avoided?  Management  Exercise controls  Employ verification and validation  Pay attention to systems engineering  Maintenance  Examine repair v. replacement  Have procedures that disconnect energy  Examine self- and remote-test capability
  8. 8. How can failures be avoided?  Materials  Use materials well within their load limits  Production  Be sure equipment works properly and operators are qualified  Use inspection and testing to separate defective components (especially at fastened joints)  Adhere to relevant codes
  9. 9. Can engineered systems be truly “safe?”  Make sure development and support staff are not all the same people  Complete, timely, readable diagnostics  Backup “human-in-the-loop” options  Guard against power failure, surges, and EMI  Give yourself enough duty cycle for multiple tasks
  10. 10. Can engineered systems be truly “safe?” (Life or property at risk)  Redundant fastening to prevent collapse  Containment of explosions or hazardous flows  Adequate shielding  Sufficient alarms  Buffer zone between system and neighbors  Make dangerous releases or shrapnel dissipate or reach a low-energy state quickly
  11. 11. Oh GOD! I've had a failure! What do I DO???  Get hold of yourself  Prepare for a failure investigation – and LEARN from it!  Be prepared for a critical look at  Your experience base  Safety margins  Self-limiting phenomena  Worst-case environments  The importance of observation
  12. 12. Sources  Casey, Set Phasers on Stun  Florman, The Civilized Engineer  Florman, The Existential Pleasures of Engineering  Jones, Engineering Materials 3: Materials Failure Analysis – Case Studies and Design Implications  Kepner, The New Rational Manager  Kletz, What Went Wrong?  Kletz, An Engineer's View of Human Error  Kletz, Learning From Accidents  LaPierre and Moro, Five Past Midnight in Bhopal  Levy and Salvadori, Why Buildings Fall Down  Lovell and Kluger, Apollo 13  Murray and Cox, Apollo: The Race to the Moon  Nishida, Failure Analysis in Engineering Applications  Peterson, Fatal Defect: Chasing Killer Computer Bugs  Petroski, To Engineer is Human  Petroski, Design Paradigms: Case Histories of Error and Judgment in Engineering  Rogers et. al. "Report to the President by the Presidential Commission on the  Space Shuttle Challenger Accident." Washington DC, June 1986.  Vaughan, The Challenger Launch Decision
  13. 13. An online example: Union Carbide Bhopal Accident Causal factors include three protective systems out of service:  refrigeration system out of service because of operating costs  high temperature alarm was reset high  scrubber inadequately sized for large release, under repair  flare stack out of commission; inlet line under repair Read MUCH more here: http://is.gd/UCBhopal