Why 'Plan B' often works out badly

[MechAg94]

http://redtape.msnbc.com/2011/03/why-plan-b-often-works-out-badly.html
This is a decent article on design and engineering back up systems and why they often are not adequate or don't work.  There are some good nuggests of truth in there.  Much is probably familiar to many of you.  I could say I see these compromises every day, but I think everyone sees them but may not recognize them.  Automobiles have a lot of good examples of compromises between safety and cost. 

I am not sure I completely agree with them on preventative maintenance being a thing of the past in all areas.  However, it is common to push off more expensive PM work or not have enough redundancy to the point that PM work gets delayed waiting for outages.

In a world of just-in-time manufacturing and corporate penny-pinching, this is easier said than done, warned Neumann. It's hard to get companies to spend money on Plan B when they are cutting thing so close on plan A.

Companies aren't willing to spend infinite dollars to prepare for seemingly unlikely events.  I would also add that consumers are mostly not willing to do so either even when it comes to things like nuclear power.

[TMM]

and here i was thinking you were talking about birth control. hmm, interesting article. brings up some interesting points.

tmm

[Hawkmoon]

I am not sure I completely agree with them on preventative maintenance being a thing of the past in all areas.  However, it is common to push off more expensive PM work or not have enough redundancy to the point that PM work gets delayed waiting for outages.

Semantics. Preventive maintenance has simply been renamed, to "deferred maintenance." As such, it gets deferred until something breaks or explodes.

A guy I used to work with nailed it, IMHO. He used to say that, "Deferred maintenance is a euphemism for intentional neglect."

That's a different animal from backup systems and redundancy. There's no question that risk management involves balancing cost against risk against consequences. The problems arise when those in the decision process adopt an "It hasn't happened, therefore it won't happen" perspective. The example of the flood design is both a good one and a bad one. Those of us "in the trenches" for designs of things subject to floods typically sit around and comment that we've had three 50-year flood events in the past five years. For anything with a projected life span of fifty years, the designers in the trenches are absolutely going to go for the 100-year design, because they know the 50-year event is not scheduled to occur next in 2061. The expression is only a probability, not a prediction. It means that past history suggests that floods of that intensity occur only once in fifty years -- on average. But if you're the designer and you know it's been 48 years since the last one, you need to figure there will be at least one in your structure's 50-year life span. Heck, even if the last one was a month ago -- there's STILL going to be another (statistically) within the projected 50-year life span.

It's the pencil pushers who look at the numbers and decide that the 50-year design is adequate, and that the project "can't afford the luxury" of the 100-year design.

In the northeast of the U.S. we had regular examples of such thinking recently. Most municipalities used up their snow removal budgets for the entire YEAR in the first storm of the winter, because the past few years have been light on snow and the geniuses who make up the budgets looked at the numbers and said, 'You only spent $___ last year, and every department has to cut back, so this year you can only have ($___ - 5%)." The fact that the Director of Public Works has no control over how much snow falls, or how fast, is ignored.

Like the old margarine commercial used to say, "It's not nice to fool Mother nature."

[Pharmacology]

and here i was thinking you were talking about birth control. hmm, interesting article. brings up some interesting points.

tmm

Same here,    I'll bet tons of plan B is being dispensed around the nation's beaches

[Headless Thompson Gunner]

He lost me in the first paragraph.

Engineers used to talk about guarding against the “single point of failure” when designing critical systems like aircraft control systems or nuclear power plants. But rarely does one mistake or event cause a catastrophe. As we’ve seen in Japan, disaster is usually a function of multiple mistakes and a string of bad luck, often called an “event cascade” or “propagating failures.”

One mistake or event rarely causes catastrophe precisely because we put so much effort into designing away single points of failure, you dummy.  It ain't because we're shortsighted and stupid and only ever think about single point failure events.  It's usually possible to design away single point failures, and we're pretty good at it, which leaves compound failures as the only ones that are likely to manifest in the real world.  

The more robust we make things, the more elaborate, complicated, and spectacular the failures are going to be, if/when failures do eventually occur.

[MechAg94]

He lost me in the first paragraph.

One mistake or event rarely causes catastrophe precisely because we put so much effort into designing away single points of failure, you dummy.  It ain't because we're shortsighted and stupid and only ever think about single point failure events.  It's usually possible to design away single point failures, and we're pretty good at it, which leaves compound failures as the only ones that are likely to manifest in the real world.  

The more robust we make things, the more elaborate, complicated, and spectacular the failures are going to be, if/when failures do eventually occur.

i think he was referring to something I have heard and seen before is that most all events have more than one cause.  I'm sure you have heard it also.  It is not always design or safeguard failures either.  It can also be failures in operations, maintenance, or just acts of God piled on top of other stuff.  If you have the information after the fact, you can almost always put your finger on certain things and say if we had done THIS, it wouldn't have happened.  Usually there are several contributing factors.  Design issues are often one of them, but not always.  There are also lots of things you just don't recognize as issues until you build it and try to run it.

Anyway, that idea is common in Health and Safety circles looking at accidents and injuries also.  It is often more than one failure or mistake leading to the injury.  Often, people were at risk for some time and either were taking shortcuts or didn't recognize and/or fix the hazard.

Of course, you can go way overboard also.  Our company has recently been getting big on process risk management and some of the changes and safeguards they want seem pretty far fetched.  It is easy to get carried away.