Dear BS, Thanks for the questions. I have put in some answers below. I might remark that many of your questions are "advanced", i.e., I would not expect to get them from people who have been working in the area for only a few years. Re: Rewrite of The Phoenix Handbook. I rewrite about once a year when business is slow. This year business has not been slow since April so I'm remiss. When I rewrite I take the previous version and change all the stuff where I've changed my mind or learned something new. Since I'm doing a workshop for ANS in June I'm trying to get the rewrite done by May. What's good? What's not? How do I get the points across better. Thanks and Happy New Year, Bill Corcoran ENTP 63 ___________________________________________ Nuclear Safety Review Concepts promotes the control of risk through organizational learning from adverse experience. For our training cards point your browser at www.exxelinc.com/tools.htm. Contact: William R. Corcoran, Ph.D., P.E. Nuclear Safety Review Concepts 21 Broadleaf Circle Windsor CT 06095-1634 USA 860-285-8779, Fax: 860-285-0012 preferred alternate -------------------------------------------------------------------------------- From: Salot, William Sent: Tuesday, June 27, 2000 8:38 AM To: 'DR. WILLIAM R CORCORAN' Subject: RE: The Firebird Forum goes after The Occurrence Pyramid-s COMMENTS & QUESTIONS ON THE FIREBIRD FORUM I am a 70 year-old reliability engineer, with 47 years in the chemical processing industry and a long-time interest in "cause analysis". I just finished reading through every issue of the Firebird Forum for the second time. It left me both awed by the editor’s loquacious tendencies and delighted by his gems of wisdom. They also gave me a feel for the "Phoenix Approach", but perhaps not the grasp I might get from reading The Phoenix Handbook. Q. This leads me to inquire how I might obtain a copy of the handbook without leaving the premises. A. I sent you one by e-mail in PDF. I will send any member of this group a "personal use only" copy if they agree to read it and send me some feedback. Here are some questions I hope the handbook will answer for me: Vol. 1, No. 1, March 1, 1998, defined "root cause" as "A member of that small set of causal factors, which, if removed, will prevent similar events (or will greatly decrease their consequences) and which are not caused in turn by more important underlying causes." Q. Does "A member" imply that there is only one root cause in "that small set", or does it imply that every member "of that small set" could be a root cause? A. Every member of that small set is a root cause. Q. Does the existence of "causal factors, which, if removed, will . . ." imply that there can be "causal factors, which, if removed, will" not? A. Yes. Many events are "over determined" (in the sense that there are more causal factors than you need just to have the event) and you need to correct more to keep them from recurring. Q. What is the Firebird definition of "set of causal factors"? Does such a set exist at a single point in time, or does it span a sequence of events? A. "A set of causal factors" is meaningful with respect to an event or with respect to a group of events. Q. What determines whether "underlying causes" are "more important" or less important than their effects? A. Their meaning for the future of the organization and the business significance of them. For example, a cause of widespread environmental contamination in NC last year was a hurricane. The underlying causes of the hurricane are probably not of interest to the NC DEP, but the underlying causes of the vulnerability of the storage facilities might be. Q. Are noncorrectable "causal factors" acceptable as root causes? A. "Root causes" are neither acceptable or unacceptable. They either are or they aren't. I would accept a hurricane as one of the root causes of an environmental release, but it is not the only one. The fact that hazardous material storage facilities were not designed to the building code was probably a causal factor, but not a root cause. Q. Vol. 1, No. 2, April 1998, shows a matrix relating the values of "Behaviors" versus "Root Causes". Does this imply that "Behaviors" are never "Root Causes", and that they therefore always have underlying causes? A. Yes. Sometimes you don't need to know what the underlying causes are to get a fix. E.g., you don't need to know why people speed into toll plazas to get them to stop speeding into toll plazas. Just install rumble strips to inflict pain on the infractors. Rumble strips are just about 100% effective when installed as directed. Vol. 1, No. 3, May 1998, shows two "dendograms". Q. Is "dendograms" misspelled? The prefix "dendro" means "resembling a tree". A. This has been pointed out to me. Thank you very much. I try to respond well to criticism, especially when it is deserved. It should be "dendrogram". Q. What is the purpose of the last "dendogram"? (It erroneously implies that the five "root cause types" are acceptable per se as root causes.) A. You are so correct. Categories never cause anything. Behaviors and conditions do. The five categories at the bottom of the Why Staircase can't really be at the bottom. There must be exact behaviors and conditions in one or more of the five categories and they in turn have causes that may or may not be worth pursuing. Vol. 1, No. 4, June 1998, lists seven reasons for ineffective or insufficiently effective corrective actions. "Underlying causes" are mentioned, but not root causes. Q. Since by the definition, discussed above, root causes are "more important" than are their underlying causes, shouldn’t corrective actions instead be directed primarily at the root causes? A. See the discussion of speeding into toll plazas. Generally, if you can correct a root cause you will get a lot of benefit, but often you can't. I am reminded of a mix-up case I investigated last summer. A root cause of the mix-up is that people do what they are used to doing. That can't be fixed. You can get people to be alert to signs that they are doing the wrong thing, though. Q. Shouldn’t corrective actions also be directed at causal factors that are neither root causes nor underlying causes; for example, physical causes? (See Vol. 2, No. 1, January 1999.) A. If you want to be businesslike you will fix all the causes that are cost-effective to fix and decide to live with the rest. You need to know what they are to make intelligent choices. Q. What determines how many corrective actions are needed on a given problem? A. It depends on what the problem reveals about the organization. For example, the TMI meltdown resulted in billions of dollars of hardware and organizational fixes in nuclear power plants all over the world. Q. Vol. 1, No. 5, July 1998, mentions "MWBA". What does it stand for? A. Management by Wandering About (a Tom Peters favorite) Q. Vol. 1, No. 7, September 1998, in the response to "letter to the editor", mentions "cognitive modeling". What is it? A. Cognitive modeling is just building models in your head that you use to solve problems. E.g., if you have a good cognitive model of your car's cooling system you will be able to figure out that you can reduce engine temperature by turning the heater up high and opening the windows!! (I did this in a snow storm when I couldn't pull off the highway. the "root cause" was snow blocking the radiator!!) Q. Vol. 1, No. 8, October 1998, in the "letter to the editor", mentions "risk evolutions". What are they? A. The term was "elevated risk evolution". Some things that we have to do are more risky than others. These are "elevated risk evolutions". For example, (pun intended) every time you get up on a ladder you are entering into an elevated risk evolution. You are well advised to review your ladder safety training before you do so. Actually you would be well advised to think about an alternate success path, e.g., hiring a professional. I reviewed a Collective Significance Analysis of ladder accidents for an electric company. It concluded that the ladder is the most hazardous piece of equipment the company used. this company has 500 thousand volt lines!! Q. Vol. 2, No. 3, April 1999, has a lot to say about "plant evolutions". What are they? A. A plant evolution is just a controlled way of changing the plant state, e.g., starting it up. Q. Vol. 2, No. 5, June 1999, in the "letter to the editor", mentions "many levels of why staircase". Why hasn’t a more detailed discussion of the "why staircase" been included in any issue of the Firebird Forum so far? A. Sloth, indolence, procrastination, paying work, grandchildren, etc. No excuse, Sir. Q. Vol. 2, No. 7, August 1999, presents a valuable checklist for reviewing an RCA report. Since such a broad scope can be an excuse for verbose reports, shouldn’t another checklist item be added to cover the organization and conciseness of the report (both written and oral) for the benefit of the decision-makers? A. A report needs to be as big as it needs to be to do what it needs to do. This is a business issue. Q. Vol. 3, No. 1, January 2000, shows a photograph of a distinguished looking gentleman. Who is it? A. My hero, the Nobel Prize Winner for the photoelectric effect and the inventor of the nuclear reactor, Dr. Enrico Fermi. Thanks for asking. Vol. 3, No. 2, February 2000, provides an enlightening digest of The Phoenix Handbook. Q. Principle 1. Considering the importance of consequences, would it be equally appropriate to treat each consequence as a separate event, with a common triggering event? A. That is basically what I do for big events. Q. Principle 7. I see the "Why Staircase" in Vol. 3, No. 4, April 2000, but what does a "Why Staircase Tree" look like? A. The "dendrograms" of a previous issue. Q. Principle 10. Can corrective actions address behavior? If so, would it also be correct to say that a corrective action has not been taken "until behavior has been changed"? If so, what is the real difference between corrective actions and lessons learned? A. There can be some overlap, but in general corrective actions need not result in learning whereas lessons to be learned do. Vol. 3, No. 4, April 2000, is an excellent technical paper on personnel error. In the section entitled The Bottom Line, "classic causal factor analyses" are discussed. Q.How can I obtain the "INPO and NRC products" mentioned? A. NRC products are available on their website www.nrc.gov. It will take a while to learn your way around. Seriously, if you don't like the NRC website tell your congressperson and senator. INPO is a secret, members only organization. If you don't like their website, tell your local nuclear utility CEO. Q. Doesn’t the Phoenix methodology include an equivalent procedure? A. The short answer is no. Call 860-285-8779 for the long answer. Vol. 3, No. 5, Special Edition, April 2000, is an impressive partial Phoenix analysis of the Tokaimura Accident. Q. In contrast, what would be an example of the least significant incident that would justify the Phoenix approach? A. Anything that would be worth more than about $50, 000 to fix. Q. How could I obtain a copy of a Phoenix analysis of such an minor example? A. There are several in The Phoenix Handbook. The oil flooding incident is my favorite. Bill, Thanks ever so much for the questions. Ever since Socrates, the Q. and A. format has been useful.