The ambulance screamed as it zipped by our company van. A few days later someone read that a contractor had been injured at the refinery. We could still see the banner hanging outside a trailer from the day before: "2-million man-hours without an accident." At least we got to keep our baseball caps, our reward for reaching a meaningless safety goal.
Safety departments put lots of stock in minimizing reported lost-time injuries. Reducing them certainly is of value. Often calamities occur at places with a bad reputation for injuries. Take, for example, the August 2008 fire at Bayer's Institute, W.Va., plant: the U.S. Occupational Safety and Health Administration cited years of accidents there.
However, keeping minor injuries low doesn't necessarily keep you safe from major accidents. When I worked at SCM Chemicals we were proud of our low injury rate. That didn't prevent a heater accident that eventually killed one worker and injured others. Perhaps we were lulled into a false sense of security.
We need a proactive way to: identify and investigate near-misses; disseminate information and include lessons learned into re-training; pinpoint improvements in equipment, processes and practices; and share insights among plants and, perhaps, competitors and regulators to reduce production risk. As safety guru Trevor Kletz stressed last month in "Bhopal Leaves a Lasting Legacy," "Chemical makers should set up systematic procedures — rather than rely on memory — to recall lessons of the past, lessons for which we have paid a high price in deaths and injuries as well as money." You'd be wise to check out the 10 steps he suggests.
SCM's competition, DuPont, also suffered heater-coil failures. Yet the companies made no attempt to address the common problem — why? Do we really need to compete on safety? Here's another question: why wasn't OSHA interested in solving this problem?
You could argue that OSHA's 29 CFR 1910 covers all this — but we need to take it much farther. Engineers must be engaged in studies resulting from hazard and operability studies (HAZOPs) and process hazard analyses (PHAs). When I was at SCM I was dismayed to find that only about 10% of follow-up studies for the first chlorine HAZOP were completed — two years later! Is this a warning sign? You bet! But nobody recognized it.
The U.K. safety regulator, the Health and Safety Executive (HSE), warns in "HSE Human Factors Briefing Note No. 3, Humans and Risk": "Before each major accident, there was a series of similar accidents, near-misses and other failures." It cites as root causes: ineffective HAZOPs and PHAs; operators using inappropriate or poorly designed equipment; inadequate indications of process condition; and management ignoring signs of trouble. Maybe it's time to change the definition of near-miss.
Let's start with HAZOPs and PHAs. How do you know they're effective? By volume of studies and changes produced or completion rate? Has anyone ever done a risk assessment, before and after a HAZOP, to judge whether hazards actually were reduced?
BP's Texas City refinery used the "What-if" HAZOP method for the first and subsequent HAZOP reviews of its raffinate splitter column that figured in a March 2005 explosion that killed 15 people. Even with this stripped down version of a HAZOP study, a flare stack was repeatedly recommended and repeatedly ignored by management — that's a red flag. (For more details on contributing factors to this accident, see "Panel Blasts BP Safety Practices," www.ChemicalProcessing.com/industrynews/2007/003.html.)
How much you rely on personal protective equipment (PPE) can provide another important safety metric. Have you identified all situations and places at your site where PPE is a key line of defense? Maybe it's time we stopped blaming operators for not using PPE or wearing faulty equipment, and found ways to avoid the need for PPE.
Maintenance plays a critical role in safety. Frequent shutdowns usually mean an increased risk to people and equipment. Creating an index that summarizes hours of emergency downtime and routine downtime might be useful in assessing catastrophic risk. Another measure might be to consider the risk factor of particular maintenance and production tasks. How many bumps and scrapes occur? Has this type of equipment ever contributed to a critical accident? How often have you been one mistake away from a catastrophe? When I worked at a bioproducts plant it took at least two fires before anyone thought to improve operator startup procedure. Another red flag!
A bit of serious effort can lead to meaningful metrics.
Dirk Willard is a contributing editor for Chemical Processing. You can e-mail him at firstname.lastname@example.org.