Chemical Process Safety:
Learning from Case Histories
by R.E. Sanders
Butterworth-Heinemann, 1999, 289 pages

ometimes seemingly simple, mundane changes in a petrochemical plant can result in disaster. In 1989, a repairman at the Phillips 66 plant in Pasadena, Texas connected two compressed air hoses in a reversed position. This mistake resulted in an explosion of ethylene and isobutane and destroyed what had been a very impressive facility, resulting in the deaths of 23 people. In 1974, a 20-inch bypass pipe containing expansion joints was used to create a temporary repair to a reactor at a plant in Flixborough, England that manufactured caprolactam, a chemical used to make Nylon. The engineers never anticipated that the pressurized gas would create lateral shear forces in the expansion joints that would cause the pipe to buckle. When it did, cyclohexane vapor was released and exploded with the force of 15 tons of TNT, killing 28 people.

Less dramatic, but often equally unexpected, are collapses of liquid storage tanks. Typical storage tanks are not designed to hold more than a few psi, and often collapse because of condensation or expansion of vapor. Large tanks have been destroyed by overpressures of less than one pound per square inch.

This book documents some of the well-known and less-well known disasters in petrochemical plants and discusses their causes, which often result from well-meaning attempts to prevent another problem. The author notes that 40 percent of all major process plant losses are due to piping failures. This was the case with Flixborough, Three Mile Island, Bhopal, and probably also at Mexico City in 1984, when four LPG spheres, each containing 630,000 gallons of LPG, ruptured. The resulting boiling liquid expanding vapor explosion (BLEVE) killed 542 people in the worst petrochemical disaster of all time in the Western hemisphere.

Less frequent are disasters caused by lack of knowledge of chemistry. Chemical reactions between chlorine and steel, which react vigorously at temperatures above 250°C, have created spectacular fires when operators allowed chlorine to flow into a pipe too soon after it had been welded. One explosion occurred in a nitric acid plant when a well-meaning instrument repairman, no doubt trying to improve the performance of a measuring instrument, replaced water in a nitric acid bubbler with glycerin. And on one occasion the entire population of ducks in a pond was wiped out when a worker failed to realize that an herbicide used to clean the pond contained a surfactant. The detergent wet the ducks' feathers, causing them to sink.

For those of us who have gotten into the habit of pounding on malfunctioning safety release valves with a wrench to stop them from venting, this book, or one of the many excellent books by the legendary Trevor Kletz, should be required reading. In my opinion, this book is superior to Trevor Kletz's books, such as “What Went Wrong”, because Trevor Kletz tends to be relatively light in detail and heavy on drawing conclusions--often putting them in all capitals (perhaps an occupational hazard that someone will someday write a book about).

Chemical Process Safety is oriented toward petrochemical plant engineers and managers, with little discussion of chemicals or chemical reactions used in industry and no discussion of cryogenics, nuclear energy, or other important technologies. Perhaps the most important message is that a post-hoc "blame the operator" approach that assigns blame to individuals is nothing less than a formula for failure. As Sanders says, we must design the procedures, the equipment, and the controls to make unanticipated equipment failure and operator mistakes as difficult as possible. The second half of the book contains many helpful tips and checklists for investigating incidents, managing a process change, and ensuring OSHA compliance.

May 7, 2006