Management at the consulting firm was confident we could do the analysis within budget. After all, 40 hours were allowed for process engineering for each pressure safety valve (PSV). We had a snappy spreadsheet didn't we? It would be easy money.
In one project after another, the spreadsheet failed. It was clunky, poorly understood by its users, and based on unsupported assumptions. Ordinary engineers hated it. Time over-runs of 150% were typical.
My first big PSV survey assignment in 1997 taught me that such an analysis needed more time. Then, my job was to develop the reasoning for the relief scenarios thrown together during an expansion of an entire plant. It took four months of numerical analysis and plenty of extra hours. I enjoyed it immensely. Now, imagine doing the work as a contractor with 40 hours per relief — fat chance. So, how long should it take?
Let's consider contractor work first. A simple PSV analysis done without the aid of an expert would require about 50 hours of process engineering time, ten hours of peer review, 32 hours for the project manager, eight hours of mechanical engineering time, 14 hours apiece for reviews by a licensed engineer or department head and the corporate engineers. This assumes no vendor interaction. I estimate a total cost of $27,000/valve. If vendor interaction is necessary, add at least another eight hours. Drawings further increase the cost. Preparing isometrics and updating process and instrumentation drawings (P&IDs) add another 16 hours for the drafter and 16 hours for the mechanical engineer in addition to another six hours of process engineering. This translates to a total contract cost of $38,000/valve. With an expert having a hands-on role in doing the work, the cost for a simple PSV analysis would drop to $25,000/valve.
If the operating company has an in-house expert qualified to do PSV analysis, the simple PSV would take about 50 hours of process engineering and ten hours of peer review, and incur little of the overhead associated with contractors. Assuming similar billing rates and no vendor interaction, I estimate a cost of $20,000/valve.
You might wonder why so many hours are necessary. First, there's the field walk-downs to confirm the drawings. Then, you must assess each scenario to determine which one requires the largest relief. Such an analysis can cover fire, in-breathing (pump-out), out-breathing (pump-in), hydrostatic, control valve failure, chemical reaction, heating or cooling, steam-out and many other scenarios. Fire exposure isn't always the worst case. If the isometrics or P&IDs are awful, it takes time to correct them.
Unfortunately, most relief applications aren't simple. Complications include: unusual vessels such as a buried sump, interconnecting vents that impose back-pressure, independent equipment connected to the same header, two-phase flow as in a vessel jacket or coil, and variable heat loads.
Then, there're situational issues, such as dealing with an inexperienced client. A faulty hazard and operability study (HAZOP) could define the relief scenario but mislead the analyst. Changing regulations or corporate standards can muddy the water.
Recently, I was involved in analyzing a concrete sump built in 1942. A structural engineer was required to certify the pressure and vacuum rating for the conservation vents. We had to sort out a variety of issues: it was an unregistered vessel, the vent design was unworkable, and only the top face was exposed to fire. Eventually, after a seesaw battle, we were able to rule out a fire case. We had to contend not only with the tricky analysis but also with an ambitious schedule. Pressure to present something to the client meant the design was re-done many times while we waited on the structural engineer. These complications and the political atmosphere of the task quadrupled the actual design hours from the 40 hours estimated.
Let me recommend a few ideas to avoid costly over-runs. First, do the work based on time and material; if your contractor goes broke working for you, you'll have to hire your own experts. Second, consider a staged approach. Pay for an initial analysis to eliminate relief scenarios, perhaps one by one. If the drawings are poor, provide money to prepare proper drawings separately. If an extensive layers of protection analysis (LOPA) or HAZOP is required, break the payments into funds for the safety meetings and those for updating the PSV design. Another option is a larger over-run allowance for the work, say, 50% for a low-bid contract.
Predicting an exact cost of a PSV analysis may not be practical. Perhaps, a flexible funding approach can work best. Don't let safety be hostage to funding.
DIRK WILLARD is a Chemical Processing Contributing Editor. You can e-mail him at firstname.lastname@example.org.