Subject: Disasters at (petro)chemical plants. One example:
“The most famous of all temporary modifications is the temporary pipe installed in the Nypro Factory at Flixborough, UK, in 1974. It failed two months later, causing the release of about 50 tons of hot cyclohexane. The cyclohexane mixed with the air and exploded, killing 28 people and destroying the plant.… No professionally qualified engineer was in the plant at the time the temporary pipe was built. The men who designed and built it (design is hardly the word because the only drawing was a full-scale sketch in chalk on the workshop floor) did not know how to design large pipes required to operate at high temperatures (150 degrees C) and gauge pressures (150 psi or 10 bar). Very few engineers have the specialized knowledge to design highly stressed piping. But in addition, the engineers at Flixborough did not know that design by experts was necessary. They did not know what they did not know.”
Lessons: “A high price has been paid for the information in this book: many persons killed and billions of dollars worth of equipment destroyed. You get this information for the price of the book. It will be the best bargain you have ever recieved if you use the information to prevent similar incidents at your plant.”
There are numerous lessons here, even for people who aren’t (petro)chemical engineers—I’m a C++ programmer. Guidance like “It is no use telling the operator to be more careful. We have to recognize that the possibility of an error—forgetting to open the valve—is inherent in the work situation. If we want to prevent an error, we must change the work situation. That is, change the design and/or the method of operation—the hardware and/or the software.” is highly portable (just replace “open the valve” with “free the pointer” and there you go).
A memorable anecdote concerned removing a dent from a storage tank. It had been filled with a warm vapor and accidentally sealed, and when the vapor cooled it reduced in volume, resulting in predictable (but slight) denting.
The plant engineer had to stop workers from attaching high pressure pumps to the tank, since that would have resulted in a burst tank. Instead, he attached a couple of feet of garden hose to a high point on the tank, and dribbled in water, which pushed out the dent.
What Went Wrong?: Case Histories Of Process Plant Disasters, Fourth Edition by Trevor Kletz
Subject: Disasters at (petro)chemical plants. One example:
“The most famous of all temporary modifications is the temporary pipe installed in the Nypro Factory at Flixborough, UK, in 1974. It failed two months later, causing the release of about 50 tons of hot cyclohexane. The cyclohexane mixed with the air and exploded, killing 28 people and destroying the plant.… No professionally qualified engineer was in the plant at the time the temporary pipe was built. The men who designed and built it (design is hardly the word because the only drawing was a full-scale sketch in chalk on the workshop floor) did not know how to design large pipes required to operate at high temperatures (150 degrees C) and gauge pressures (150 psi or 10 bar). Very few engineers have the specialized knowledge to design highly stressed piping. But in addition, the engineers at Flixborough did not know that design by experts was necessary. They did not know what they did not know.”
Lessons: “A high price has been paid for the information in this book: many persons killed and billions of dollars worth of equipment destroyed. You get this information for the price of the book. It will be the best bargain you have ever recieved if you use the information to prevent similar incidents at your plant.”
There are numerous lessons here, even for people who aren’t (petro)chemical engineers—I’m a C++ programmer. Guidance like “It is no use telling the operator to be more careful. We have to recognize that the possibility of an error—forgetting to open the valve—is inherent in the work situation. If we want to prevent an error, we must change the work situation. That is, change the design and/or the method of operation—the hardware and/or the software.” is highly portable (just replace “open the valve” with “free the pointer” and there you go).
Still Going Wrong!: Case Histories Of Process Plant Disasters And How They Could Have Been Avoided is its sequel.
I created an account just to vote this one up.
A memorable anecdote concerned removing a dent from a storage tank. It had been filled with a warm vapor and accidentally sealed, and when the vapor cooled it reduced in volume, resulting in predictable (but slight) denting.
The plant engineer had to stop workers from attaching high pressure pumps to the tank, since that would have resulted in a burst tank. Instead, he attached a couple of feet of garden hose to a high point on the tank, and dribbled in water, which pushed out the dent.
Because, you know, hydraulics.
(At least, it blew my mind, man)