Foil HAZOP Failings

Appropriate action on most items flagged in a hazard and operability (HAZOP) evaluation too often doesn’t take place before the next HAZOP. Some key items from the checklist will roll into projects and process studies but the vast majority will get ignored. That’s the harsh reality that I’ve witnessed over and over again for more than thirty years. For instance, two years after a HAZOP, I found plant staff hadn’t touched 90% of the items from a chlorine checklist. (By the way, I was able to reduce that number to 10% in only three months.) Why does this happen?

[pullquote]

Well, the most obvious reason is personnel are busy and management doesn’t have the resources. The HAZOP follow-ups actions just join the existing, usually substantial, backlog of work. Unfortunately, there’s little incentive to tackle them. After all, promotions seldom depend on clearing up backlogs. You don’t see people listing this as an achievement on their resumés, do you?

In addition, safety management deficiencies also usually contribute.

For instance, consider the way planning typically takes place for a HAZOP. Even though every covered process must undergo a HAZOP every five years, many sites wait far too long to prepare, resulting in a mad rush at the end; people suffer HAZOP fatigue. No wonder the will to complete tasks uncovered during the review usually dissipates.

It’s easier to update drawings, files and procedures continually rather than every five years. Why not assign the projects department or the individual process engineer involved in a project with the responsibility for updating the drawings? I can brag here because I always took this up as a personal responsibility. However, I understand the reluctance of many engineers; getting the drawings caught up after years of neglect by lazy engineers can pose a challenge. Management sometimes doesn’t help; engineers get re-assigned or leave because of downturns. Managers are hesitant to hand these efforts to experienced engineers — who likely will see them as grunt work — and, so, wind up forgetting to assign them. Yet, there’s a good alternative, one that I’ve mentioned before: entrust engineers new to a site or fresh from college to clear up HAZOP items, update drawings, etc.; it’s an excellent way for them to become site-savvy.

Another valuable approach is creating a current material balance — one based on actual laboratory data. This is more important than updated process and instrumentation drawings. Too often, balances are based on models or simplifications that don’t reflect reality — and, thus, may endanger people. Lab measurements can reveal flaws in models.

Oversight is another failure. Companies may proclaim safety as their top priority but don’t always act that way. Why don’t safety managers demand meetings to review progress? Why don’t corporations support better oversight? Probably because of lack of foresight, as is apparent in the Arkema accident reported in the New York Times.

The Times article noted that the company failed to update contingency plans. I would bet that procedures and other key operating plans hadn’t been revised in many years either. This touches upon an issue that afflicts many sites: inexperienced chemical engineers, quite appropriately, are reluctant to change things they don’t understand and don’t have mentors to bring them up to speed. As a result, updating gets pushed back or ignored.

Another problem is the lack of practical experience among regulatory staff. Part of this stems from prejudice in industry hiring practices: try getting a job in industry once you’ve served as a regulator! Ideally, safety inspectors should know as much about the unit operations at a facility as the engineering staff. Industry should promote this by encouraging a free flow of personnel between industry and regulatory bodies. Experience benefits industry; familiarity breaks down walls. Of course, some might worry about potential conflicts of interest. However, state and federal agencies now review documents almost simultaneously, minimizing the risk. Moreover, public agencies must share information via the Freedom of Information Act.

Regulators could help by auditing progress towards completion of HAZOP items. However, this never is done — even at sites where fatal accidents occur. Regulators should review progress every few years for these sites and others they believe deserve attention based on risk, past performance and other factors. I’d suggest regulators consider recent takeovers and downsizings. My experience is that sites in turmoil tend to have more accidents than those that are stable.