Numerous safety experts, including Trevor Kletz, will tell you to conduct safety reviews even for minor process changes. However, if your process and instrumentation diagrams (P&IDs) are wrong, such a review may do more harm than good.
I bet erroneous drawings have contributed to every major accident in our industry — but few appreciate this unspoken risk. I couldn't find any case histories during an afternoon of searching. The Occupational Safety and Health Administration (OSHA) levies heavy fines for inaccurate P&IDs, but only after an accident.
The entire process industry suffers from poor history collection. P&IDs are incomplete, out-of-date and in error for several reasons:
1) No one is assigned responsibility to maintain them. 2) Company standards for markups, equipment, tag names, lines, etc. don't conform to industrial standards. 3) No funds are budgeted to keep the P&IDs current. 4) Engineers aren't familiar with the process the P&IDs cover.
Any company without a clear procedure for maintaining as-built drawings, in my opinion, is setting itself up for an accident. Require the change champion to update the drawings, establish a review committee and get it done. At one company, the hazards and operability (HAZOP) team did as-builts — but only for 25% of the P&IDs in an area!
The industrial-standard color coding for marking up a drawing is: red for new, green for delete and blue for comment. I've filled an extra D-size drawing with blue comments. Never use orange and yellow except during the review process; drafters use yellow to show what they did.
Your company, if it deviates from standards, puts you and your operators at risk. Contractors waste thousands of your dollars changing drawings; for example, redoing all the tags on a distillation column with 25 transmitters will cost you more than $2,000. That's a high price for individuality and there's a hidden cost — the time anyone reading your drawings must spend learning your standard.
Letting P&IDs get seriously outdated fosters engineers becoming disconnected from the operators who know the process. What good is a HAZOP study of a process that only exists on paper?
To ensure that major new projects, e.g., those stemming from operational change requests, get properly captured on the P&IDs, consider requiring the project engineers or team to update the drawings while they're designing the process. Involve safety and environmental as well as production in this review. In addition, periodically create a history of the design process from memos and markups — make a summary mandatory. Another option might be to automate part of the design process and procedure preparation. This could reduce the effort involved in history collection. Several software packages are available.
Ideally, such software should produce P&IDs in writable PDF format; only permit the drafter to make the final changes. The approval document should include a series of questions to narrow approval to those most familiar with particular drawings. A pull-down list, resulting from these questions, should enable selecting the most appropriate primary, secondary and backup reviewers. Allow these individuals to request additional reviewers. Set a time limit for reviewers, corrections, responses and other correspondence. Insist that requests for change contain documentation justifying the change. Include calculations for review. Of course, once the review process is completed, write a summary history explaining the reason for the P&ID change and the outcome of the review process. Store this with the rest of the process changes.
These procedures won't keep P&IDs completely free of errors. Day-to-day operational changes such as job plans and maintenance work will generate scores of minor changes every month.
To (hopefully) catch errors before they occur, two review processes will have to work overtime.
First, a HAZOP team should audit all P&IDs in preparation for upcoming HAZOP reviews. In addition, it should survey plant areas to identify where to focus attention.
Second, a reliability team should continually scan the historical record for errors. Do vessel nameplates match the drawings? Do marshalling panels contain new additions? Who added a plant vent connection near the new tower? Why is stainless steel hose instead of hard pipe connecting a cleaning tank? This team's goal is to generate a dialog between process, design and production to produce a history where none existed. Because reliability goes hand-in-hand with safety, turn over any finds to safety and process engineering.
This two-pronged approach might not suffice in environments where change is happening too fast. When procedures and drawings begin to look like a 2nd grade class got hold of them, it's time to enforce a policy of a 2nd HAZOP. As always, safety must adapt to conditions at hand.
DIRK WILLARD is a Chemical Processing contributing editor. You can e-mail him at firstname.lastname@example.org.