"There's the good, there's the bad, and there's the ugly," began Dan Schwab, project manager for automation projects at BP Whiting, who together with Len Laskowski, principal technical consultant for Emerson Process Management, explained the importance of properly calculated safety integrity levels (SIL) in a presentation at this week's Emerson Global Users Exchange.
"The good means you did your SIL verification calculations early: You have a feasible design that proves that the required SIL can be achieved, describes the project with a proper scope and adequate funds, and positively impacts lifecycle costs. You get an award and maybe even a raise."
"The bad," he went on, "is doing your SIL verification calculations late in the project, and you find that the design is not feasible, and you have to redesign to attain the required SIL, with the potential for significant change orders and schedule delays. There's also a potential for substandard design, and the lifecycle costs are negatively impacted. You don't win that award, and you probably won't get that raise."
"The ugly," Schwab continued, "is when you do your SIL verification calcs just prior to start-up. The required SIL may not be achieved, and the design has to be completely redone, because you've installed inadequate protections. Additional mitigating measures cost more and take lots of time. You may have to demolish the inadequate systems, piping and controls that you just installed, and that will impact lifecycle costs very negatively. Get your resume out and find a different line of work."
Laskowski provided a fire-hose of definition and detail about safety integrity levels, how to produce a properly SIL-rated system—and have the SIL verification calculations to back it up. He described in detail what you need and how to perform SIL verification calculations properly.
"You need a lot of information before you start doing your SIL calculations," Laskowski said. "You need to know the number and types of instruments you will be using and what their SIL ratings are generically. You need to know the voting architecture, and you need to decide which devices are going to be safety-critical."
"And there's even more," he went on. "You need to know what the demand rate will be, and whether you're designing to a SIL or to a SIL based on the residual risk factors (RRF). You need to know the valve fail positions and whether you will have energy to run wildcard devices like those that energize to trip."
"You need to decide what the proof test interval needs to be, and you better remember that if you set your intervals too close together, you will certainly get push-back from operations because they don't want to shut down, frankly, ever."
When should you do your SIL calculations? "As early as possible," Schwab said. "For brownfield modernization and expansion projects, you should do it in front-end engineering design (FEED). For new grassroots facilities you may not have the information until early detailed design, but do them as early as you can. Any high-risk areas should be evaluated in FEED."
Using a recent project at Whiting as backdrop, Schwab and Laskowski went through the iterative process necessary to do a real-world SIL verification calculation project and what the effects of the calculations were on the system as installed. "With an early start and successive iterations," Laskowski said, "you can deliver your project successfully on budget and schedule, with the lowest overall cost of ownership."