How can such a little gauge play such a BIG role in process safety? It is actually quite simple, yet so many relief system designs fail to incorporate them. These gauges are BOTH critical to process safety AND compliance with PSM/RMP. In this article I will explain what a "tell-tale" gauge is and what purpose it serves, as well as document their mandate for those relief systems that utilize dual in-line relief systems.
The most widely used RAGAGEP for pressure vessels and their relief systems is ASME Section VIII; so lets look at Section UG-127 and see what ASME says we need when we have a Rupture Disc (RD) between the pressure vessel (PV) and the pressure relief valve...
UG-127(NONRECLOSING PRESSURE RELIEF DEVICES)
(4) the space between a rupture disk device and a pressure relief valve shall be provided with a pressure gage, a try cock, free vent, or suitable telltale indicator. This arrangement permits detection of disk rupture or leakage53.
Footnote 53 - Users are warned that a rupture disk will not burst at its design pressure if back pressure builds up in the space between the disk and the pressure relief valve which will occur should leakage develop in the rupture disk due to corrosion or other cause.
So as we can see from the code, the purpose of this requirement is so that ANY amount of leakage by the RD is recognized ASAP. Failure to recognize this leakage could have catastrophic consequences, since any back pressure on the disc will quickly incapacitate the relief system from being able to fully relieve the pressures. NOT a good thing, seeing how our relief system is our LAST LINE of defense in process safety.
A couple of KEY POINTS to consider when your Relief system has this type of arrangement:
1) the gauge used between the RD and PRV/PSV, should be an actual tell-tale gauge and NOT just any ordinary pressure gauge. As shown above a tell-tale gauge will "register" any pressure and indicate this pressure rise with the "tell-tale" stick. So even if there is no pressure in the space when the operator does his/her rounds, he/she will be able to see that the "tell-tale"stick has moved. A regular pressure gauge is not capable of this, as its stick returns to zero if the pressure in the space goes back to zero. As a side note this becomes EXTREMELY important in batch processing and I will provide a first hand example of a personal lesson learned.
2) the gauge used between the RD and PRV/PSV, MUST be on the daily/shift rounds. Recognizing that the RD is leaking or has blown is an ABSOLUTE CRITICAL path to process safety in these types of relief systems. This is so important that most new systems are designed to have pressure transmitters in the space that will alarm, providing IMMEDIATE notification of a pressure rise in this space. Many processes I work with treat this pressure alarm as a potential EMERGENCY SHUTDOWN TRIGGER; I am in NO WAY saying you have to do this, I am just trying to stress the significance of this process deviation.
3) the gauge (or pressure transmitter) used between the RD and PRV/PSV, MUST be in the facility's Mechanical Integrity inspection/testing program.
4) this design REQUIRES that personnel operating/maintaining the process be TRAINED to recognize this deviation. As I said in #2 above, operators need to be checking these gauges on shift/daily basis (if there is no instrumentation monitoring the space). Because this involves the operators, they need to be trained that ANY PRESSURE in this space is a "process deviation" and thus we need to have "steps to correct" this deviation. YES, this means that this type of design may indeed call for an operating procedure. Call it what you want, but pressure in the space is in FACT a process deviation that must be "corrected" and this calls for "procedures to correct...". As I stated above, there are several clients that have established this deviation as a "trigger" for shutdown - not all classify this as an emergency shutdown, but most do.
As I stated earlier I learned early on that the simple use of a gauge with a "tell-tale" stick can make a world of difference, especially in batch processes, where there is NOT a constant pressure in the process. One of my processes was highly corrosive so the relief system utilized this dual system in order to protect the PSV from the highly corrosive environment seen in the process vessel(s). We had our tell-tale indicator as required, but used regular pressure gauges. We had these gauges on the daily rounds for operator to log their readings. This was done like clock work; however, the times the operators made their rounds ranged over two (2) 12-hour shifts (another lesson learned on this topic later). So in short, operators were logging their readings on vessels that were empty and open to the atmosphere! So you can guess what the gauges were reading. One day we had a surprise and we came damn near to launching one of our PVs as the PSV was so corroded that it only opened around 5% and even the RD did not fully open due to the back pressure on it. In other words we GOT VERY LUCKY that we were running a very small batch in one of our largest reactors and the Relief System on that reactor was way over sized for that particular batch. Had the event occurred with a batch normally run in that vessel, there is no telling what could have happened. So we began our investigation with every kind of engineer you can imagine and it all boiled down to the fact that although our operators were making their rounds, over 95% of those inspections were done when the vessel just happened to be at atmospheric pressure. You ask... "What are the odds of that happening?" Quite high if there are no controls as to when these gauge inspections take place. During batches operators were tasked with running the process and running 2-3 vessels at a time. So they would do their rounds BETWEEN batches when the vessel was EMPTY and OPEN!!!
SPECIAL NOTE: We NEVER sent our PSVs off to be bench tested. We just changed them out and used our own in-house and UNTRAINED personnel to do this work. We sent out over 50 PSVs after this incident after finding the RDs had been compromised and ALL the tested PSVs had significant issues! So when the code says "atleast every five years" you better never ASSUME that your RDs are working as planned without some type of VERIFICATION!
SIMPLE and CHEAP fix
We spent around $35 per gauge and installed gauges with "tell-tale" sticks. Now when operators made their rounds the stick would let them know if that space saw any pressure during their batch. And of course we could then change out the disc BEFORE the next batch! I would also like to say that we discovered that our system had a design flaw in the materials of construction of the RD, which led us to change our RD design basis (using an MOC of course). We found that the disc, although had an annual change out schedule per manufacturer, were last only several months. Due to down-time costs due to changing the RDs out more often we changed to a new RD so that we could get more up-time and improved safety. This is how MI data can play a role in IMPROVING production AND process safety.
PLEASE NOTE: we could have made only administrative changes and required the gauge inspections be done during batch runs, but chose the route we felt best.
The two potential OSHA citations for those not following these guidelines:
1910.119(d)(3)(i)(D) Relief system design and design basis;
1910.119(d)(3)(ii) The employer shall document that equipment complies with recognized and generally accepted good engineering practices.