Monday, 17 August 2015

Flixborough 1974

At tea-time on Saturday 1st June 1974 a seeping cloud of cyclohexane vapour reached a distant ignition source and deflagrated. As the leak and the ignition occurred in the middle of a chemical production facility the explosion rapidly propagated through the plant and killed or injured almost everyone at work at the time.  It went down in the history of health and safety, chemical engineering and North Lincolnshire as the Flixborough Disaster.  That was the end of Flixborough's two generational record of being thankful for sparing its inhabitants from sudden death.

Flixborough was a massive chemical campus owned and operated by NyPro, a joint venture between Dutch State Mines (DSM) and the British National Coal Board (NCB) to produce caprolactam, an essential precursor in the manufacture of Nylon.  The plant had initially been set up to make fertiliser from the coking-furnace byproducts of a nearby steel factory. When the price of Nylon went up and the price of fertiliser came down, one line of the system was re-engineered to process cyclohexane, a colourless and highly flammable [think octane] organic compound with a distinctive 'cleaner' smell. The oxidation by compressed air was carried out at about 8 atmospheres [0.86MPa] and 155oC but it was still quite inefficient. The cyclohexane was passed through a cascade of six pressurised reactors [L above] each holding 25 tons of hot seeth - it boils at 80oC at atmospheric pressure: see Gay-Lussac's Law for the relationship between temperature and pressure.  The 'wort' went down the cascade under gravity and each pass converted about 6% of cyclohexane into cyclohexanone which was bled off.  The 94% unreacted cyclohexane was sent back to the start for another go-through.

In April 1974, routine inspection and audit showed that the cascade was leaking and stripping off the lagging of reactor 5 showed a crack 2m long its the outer shell. Time was money, the plant was way behind on its production targets, the government had capped the price of caprolactam to help the nylon industry and management called for a quick-fix.  The chemical engineers decided to by-pass reactor #5 with some pipework between #4 and #6.  They couldn't find any 700mm pipe in the yard so they bodged in some available 500mm, to accommodate the fall between #4 and #6 a dog-leg drop [cartooned in red in the annotated photo above L] was mitred into the tube.  To the chemical engineers who made up the  caprolactam working group and most of the senior management it looked okay on paper.  So they shifted reactor 5 out of the way and installed the bypass supported by scaffolding. But NyPro's mechanical engineer had retired at the beginning of the year and not been replaced, so there was nobody on site really qualified to critically evaluate the effects of turbulent flow, fluid-hammer, and joint-loading: the bread and butter of engineering engineers.  Anyway, the fix-up worked, so the chemical engineers thought they knew as much as they needed to know about the system.

A few weeks later, the cyclohexane line was switched off to fix more niggling leaks, inspected, repaired, relagged and  . . .  powered up on the morning of 1st of June. On a Saturday, there were far fewer people about the plant, including 18 shut up up in the central control-room. So nobody noticed the distinctive smell of leaking cyclohexane until it reached a furnace in another distant part of the plant and >!WHOOOMPH!< the explosion destroyed the reactors and ripped through the factory breaking and shattering and causing a cascade of supplementary fires and explosions. Nobody knows exactly what happened because everyone in the control room was killed and all their log books were burned to cinders. The only survivors were those who made a quick exit in the right direction immediately on hearing the alarm and/or the first bang. Every window within 2km was broken and structural damage was sustained as far away as Scunthorp 5km distant. 28 dead, 36 injured (burns and flying glass, mostly), a handful of workers shaken but alive. Forensic engineers moved in even before the fires were finally extinguished (things were still smoking on 10th of June) and tried to piece together what had occurred by the position and sustained damage of the parts, which were scattered far and wide. The Rotor of fan #9 [L above], for example was blown out of the plant and into the waste ground. Two hypotheses surfaced: the 20 inch theory which laid the blame on failure in some part of the the complex 500mm bypass OR the 8 inch theory based on finding a rupture in a 200mm pipe which originally joined two nearby separators; the explosion that broke the narrow pipe shook out the 500mm reactor #5 bypass or its Heath-Robinson scaffolding support. There is, inevitably, a  minority opinion that the investigation was a stitch-up to exonerate the company, limit the damage and get everything back into production asap. One member of the investigating committee was 'let go' when he insisted on bringing management and protocols before the board of enquiry.  My father, after he retired from the navy at age 50, secured a sales manager job in another British industrial combine.  His stories of management incompetence in his firm had him occasionally incandescent: ". . . if those buggers had been that stupid and careless aboard any ship of mine, I'd have taken their stripes and sent 'em down to the fo'c's'le". So the idea of men [it was always men back then] drawing a comfortable salary to make cavalier decisions beyond the level of their competence is not entirely without plausibility.

Nevertheless, the Board of Enquiry did find that "All engineers should learn at least the elements of other branches of engineering than their own".  Amen to that, specialization is for insects.  Here's another piece of advice: when management announce that they are going to re-start the production line after some down-time . . . that's the day to pull a sickie.

No comments:

Post a Comment