The vaporizer was a pain to operate. It was ordered during a plant expansion. Probably nobody really gave much thought to how it would be incorporated into the plant control scheme. This was before that watershed year 1991, when OSHA began enforcing 29 CFR 1910, which mandated hazard assessment.
A relay box operated the vaporizer and one of the relays was burned out. The plant was down while we figured it out.
So, what do you think caused this ongoing maintenance nightmare? Let's go back to the ordering process for vaporizer. Didn't anyone consider how difficult it would be to work with the relay box? Was maintenance consulted during ordering? Did company standards exist and, if so, why weren't they followed? One value of the management of change process is that it fosters broader participation on projects. Including skid equipment designers obviously makes a lot of sense, too.
A vaporizer is fairly simple compared to units such as distillation columns, crystallization systems and dryers. Their long lead times are a crude gauge of their sophistication. A typical crystallization setup includes a couple dozen pumps, scores of vessels, screens and centrifuges, as well as instruments. The instrument list alone will be 15 pages long, raising potential for thousands of mistakes. How will you avoid show-stopping errors like the wrong thermowell material for an acid scrubber? That happened to me and caused a day of downtime and some embarrassment.
You've got to get organized! Here's a short shopping list to consider: 1) a delivery schedule (coordinate it with a pay schedule); 2) an alarm-trip summary; 3) a spare parts list and maintenance schedule; 4) requirements for special training and an operating and maintenance manual; 5) materials standards — you'll want to plan coupon tests for new materials; 6) pipe standards; 7) quality assurance procedures; 8) battery limits for the new equipment; and 9) recommendations for specialty testing equipment, if required. For sophisticated systems add a list of subcontractors — try to keep suppliers local and don't spread motors, pumps and gearboxes among a dozen vendors — that will cost you extra later.
Vendors of complicated systems like dryers should have extensively tested their units long before you can order them. However, for a risky project, build a prototype and run the bugs out of it or identify your worst problems. When I was in the TiO2 business we knew from our prototype tests the flow control valves for a new reactor would be a problem. We front-loaded a lot of engineering time to solve that problem. During another plant expansion we developed a list of maintenance priorities and rolled them into the project.
Another challenge with complex process systems is the number of choices — each of which may drastically affect economic feasibility downstream. This probably isn't a good time to cut low-bid contracts with your equipment vendors. You'll want them on your side, not watching the clock. Assess choices in the lab and via simulators.
Constantly evaluate the feasibility of the project — keep looking backward. But don't leave your vendor hanging when you find an improvement worth including. A delay may add to your costs or lead to contractors and vendors assigning their people to billable projects. Plan for a meeting at least every two weeks with the vendor for complicated projects even when it seems like the system should be well on its way.
You must be involved in every phase of a vendor's work, setting standards but taking advantage of its expertise whenever appropriate. A good project engineer asks tough formulated questions. Consider asking one question that's rarely posed: "How can I make your job easier?" Be prepared for an enlightening response.
What you mustn't do is constrain the design process. Follow Patton's rule: "Never tell a man how to do his job. Let him surprise you with his ingenuity." You can learn a lot from your interaction with a vendor.
I once worked on a plant expansion for a customer accustomed to getting everything he asked for, regardless of practicality; his favorite phrase was, "Do it my way or you're fired." We completed the project but communication suffered.
And, one last thing, test everything — in the shop before delivery and then in the field. Did you know that 18% of instruments fail on delivery and another 18% fail during testing? ("Instrument Engineers' Handbook," 4th ed., Vol.1, p. 136, Table 1.11a.) I have seen a similar failure rate in the field. Even if the failure rate for ubiquitous equipment like pumps is only one-tenth that, it's still worth testing them.
Dirk Willard is a Chemical Processing Contributing Editor. You can e-mail him at firstname.lastname@example.org