Process Pop Quiz 21-01: SIS/SIL System Review for Overfilling a Tower--Answers in comments
ERIC PARVIN P.E. P.Eng.
Owner / Chief Process Engineer, Expert Witness at Parv Process Consulting & Engineering
ANSWER: Mr. Sonu Singh below has provided a great response to the question below. So spoiler alert--read the comments to see the answer to our pop quiz below. Thank you!
Process Pop Quiz--our first one for 2021. I'll try to keep up this year and do one a month.
This one--while PSV related, is not about PSVs but about process safety and using your process engineering skills.
You have a crude unit with cold feed from tanks, to pumps to get to the crude unit where you pre-heat and provide 2 phase flow to a preflash tower. Let's assume the percent vapor is approximately 10% by mass just to keep things simple.
The tower is small enough that overfilling the tower in under 10 minutes is credible if you close off the liquid out of the bottom of the tower, and the PSV evaluation found that the PSV is significantly undersized as well as the flare header, flare KO drum, and more. This scenario is extremely costly to execute a PSV / flare solution (greater than $2MM US).
The refinery and operations representatives decide to install a pressure differential system on the tower to monitor the liquid level across the entirety of the tower at 4 key spots--the bottom, the high liquid level for the bottom sump, about half way between the feed nozzle and the top of the tower, as well as the overhead nozzle of the tower.
The tower is approximately 50 ft. tall, and you can assume at operating conditions the liquid in the bottom of the column has a specific gravity of 0.7. There's no reboiler--just some minor reflux that is relatively small.
Note: The level transmitters--YOU CAN ASSUME THEY ARE CALIBRATED CORRECTLY for the operating conditions, so no replies please that you need to check it. Let's assume the instrument technicians actually know what they are doing and are worth their salt (I find they usually are).
The project team decides to install a high integrity protection system (HIPS) with a SIS/SIL level of 2. The system is designed that upon detecting a pressure differential on the level tap locations of 10 psi or more, then it will necessarily turn off the pumps in the tank farm and stop flow to the unit. The differential will be measured on the highest differential recorded among all the level tap locations described above so they have redundancy and can comply with ISA standards.
For a 50 ft. tall tower with a 0.7 specific gravity, the engineering calculations show that if the tower does reach the tap between the feed and the top of the tower, the height of the tap is about 35 ft, so the pressure differential to the bottom tap is 10.6 psi (35 / 2.31 x 0.7 = 10.6). For the tap above the top tray, let's say it's 50 ft. from that tap to the bottom tap, so it's a 15 psi differential. The system should trip and prevent a complete overfill and overpressure. The normal pressure drop across the trays for this system is ~3 psi for the normal liquid level, and ~2 psi for the tray pressure drop. Your level of tolerance between the normal operating pressure of ~ 5 psi and the trip of 10 psi is "tight", but nobody has ever seen the pressure drop across the tower in other normal operations exceed 7 psi, so there's believed to be adequate gap to prevent the unit from tripping by bad measurements.
You are the official "Master and Commander" of the site and are charged with approving the MOC to install the SIS/SIL system for the Process Safety Department. The project manager is looking for any advice and wants to make sure the system will work. It's estimated to be a $350,000 US project, but is far better than the $2MM US price of upgrading the flare system.
Pop Quiz: Will this work?
Pop Quiz: If you find a problem, how do you tactfully tell everyone and advise them that they just wasted 6 months of engineering effort?
Part of the pop quiz is the communication of the technical details without insulting everyone's work (engineers do so poorly at this sometimes).
Answers late this week or during the weekend.
Principal Facilities Engineer at Mesa Applied Technologies, LLC
4 年Personally, I don't like using dP gauges/transmitters for critical shutdowns. Too many things can go wrong with them, like plugging etc., even with diaphragm seals. Call me old fashioned but I'd use the differentail system as a guide but for the LSHH I'd go with either a float or displacer switch, especially one actually inserted in the vessel itself. And I'd also make sure the float/displacer would trigger at a significantly lower SG than 0.7 ... more like 0.5 or so.
Senior Process Engineer
4 年HIPPs vs Flare modifications and HIPPs should work, I would go with this option. However the question is how much the company has already spent on the engineering, 6 months worth of engineering? When you add that to the $350K probably you come very close to $1MM, that's savings of $1MM, not bad.
Associate Director- Technology & Design
4 年ERIC PARVIN P.E. P.Eng. 1) I am assuming the SIL study is done right and SIL2 is sufficient to lower the risk. 2) As mentioned in the post the liquid will be still flashing in the column so the S.G is going to be less than 0.7 this need to be considered in the calculations for LT pressure taps and set points.
Senior Process Engineer
4 年If the system fails once in the life of the plant which are the consequences? Overpressure with a psv and flare system not designed for the same with a high risk of loss of containment, liquid carry over up to the flare, HC burning rain......fatalities. the reliability of the system dependes also of it certification and maintenance during the life of the plant, as well as operators training, which results in important or significant operational costs. The 350000$ project will be finally much more than the same. Better to provide inherent safe design : a column designed for the pressure at fully liquid condition....if now possible relief valves and flare system adequate designed for the same..