I asked some ChemEs and MEs the following questions about relief valve testing. We come across some RV programs and practices that we feel fall short of following any RAGAGEPs and thought it would be helpful to hear from engineers who manage these programs on a daily basis. Here is my scenario, and their responses follow.
Hypothetical Situation:
I have a PSM-covered process with a refrigerant as my highly hazardous chemical. I practice removing all my RVs every five years, throwing them in a dumpster, and installing new ones. The practice has shown that it is “cheaper” to replace the old RVs rather than have them tested with the chances they may fail and I would have to buy new ones anyway (e.g., the cost of a new RV is less than the price of testing and a new RV). The old RV gets no type of inspection, and neither does the vent pipe to the RV nor the vent pipe on the discharge of the RV. I do this every five years, as per an Industry Trade Association safety bulletin that OSHA has deemed a Recognized and Generally Accepted Good Engineering Practice (RAGAGEP). These industry bulletins (2 different ones from the same organization) state the following regarding the inspection and testing of RVs:
6.5.4 Pressure relief valves shall be replaced with a recalibrated valve or cartridge at intervals not exceeding five years.
and
4.9.7 Pressure-relief valves discharging to the atmosphere should be replaced or inspected, cleaned and tested every five years of service. Testing should be done by an authorized testing facility.
The questions for this group are…
- in your RV program, do you base your inspection frequency on past inspection data, not to exceed five years or do you have them set up on a five-year schedule, come hell or high water?
- do your employees who REMOVE RVs get formal training on what to look for when an RV is removed (e.g., inspect the inlets and outlets of the RV, including the inlet and outlet piping, look for corrosion, etc.)?
- If your process is considered a “clean process” with no potential for bridging to occur at the inlet of the RV does this matter in the inspection of the vent piping?
- Do you have a maintenance procedure that your employees follow when working with an RV?
- Do your employees who INSTALL an RV after an inspection/testing receive formal training on how to install the RV?
- If an RV lifts for any reason, either prematurely or for an overpressure event, do you REMOVE the RV from service ASAP or leave that RV in service until its five-year anniversary?
- Do you allow an RV to be removed from service and thrown away without any type of formally documented inspection, even if it is just a visual inspection by a trained worker?
- Do you feel that your RV program MUST have data on each valve to justify its inspection frequency if it allows for the maximum service life of the RV based on the manufacturer’s recommendation?
- Does my practice of removal and installation of a new valve without any inspection documented meet a Recognized and Generally Accepted Engineering Practice like ASME Section VIII? I have nothing on hand but a purchase order to show that the valve purchased and installed was the right set point and scfm.
I know this is asking a lot, but I need some help from engineers who manage RV programs on processes containing a Highly Hazardous Chemical.
Response #1:
When PSM first came out, several major chemical industry clients had every PRV on their sites (and I mean every single one) rebuilt yearly. It didn’t matter if it was on a chemical process that used a listed chemical or on a steam system or on a refrigeration machine, or even if the process was operating near the PSD set pressure or if it was operating at a small fraction of the set pressure. “Experience” had shown that “some” of their chemical process PRVs started leaking after a couple of years, so they got into the habit of rebuilding everything when labor was cheap. When labor costs started creeping up they rationalized continuing the practice as the paperwork associated with having to report a fugitive emission was so monumental. As labor costs continued escalating, they decided they needed to get smarter with their maintenance practices.
One of the tasks I did for them was to help them set up a PSD database where they could track the wear shown on the PRVs during the rebuilds so that after a few years, they could start making informed decisions. I also made recommendations regarding the frequency of their rebuild cycles. Basically, we left anything that interfaced directly with a chemical that was on the PSM list or for which they would have to report a fugitive emission at an annual rebuild until they had a few years of historical data accumulated. We moved any high-pressure steam PRVs to at least a biannual frequency and all low-pressure steam and refrigerant PRVs to a four-year frequency. We didn’t immediately push out to a five-year frequency to allow for the inevitable schedule slips. So the PSM folks would have at least one data point for their periodic system re-evaluation.
Almost every chemical industry client I have worked with has their own maintenance people remove the PSDs but have an outside contractor inspect, rebuild, and certify the PRVs. Very few have maintained their own “valve shop” since the late 1980s, and even those that have usually subcontracted the work out to a contractor located on-site. When they have a PRV contractor on-site, they remove and re-install the PRVs.
Whoever installs or removes the PSDs should be trained in how to handle them and protect the associated pressure vessel connection and vent piping. That does not mean they must be trained in PRV maintenance or testing.
Usually, when a chemical process PRV lifts, it responds to a process upset. If the process is shut down in response to the upset, the PRV is removed for rebuilding at that time. If the process continues after the upset, most of my clients remove the PRV from service at the next scheduled outage. A steam PRV may lift due to poor thermal control or upsets in the steam supply system. Since this can happen quite often, most of my clients allow steam PRVs to remain in service until the next scheduled outage after they start noticing steam leakage or excessive condensate discharge from the vent weep hole.
None of my clients would throw away a PRV without any form of documentation. The accountants wouldn’t let them, even if the engineers didn’t want to know the condition of the PRV at the time of disposal/scrapping. Plus, if the PRV has been exposed to a listed chemical, it probably has to be cleaned before it can be scrapped.
Yes. Most PRV bodies are intended for decades of service (I once had to deal with a refrig machine that had on it PRVs from the 1940s, and the PRVs hadn’t needed to be rebuilt for over 20 years). Unless you are dealing with small, inexpensive PRVs intended to be disposable, rebuilding PRVs and fine-tuning your preventive maintenance schedule to keep just ahead of the need for break-down maintenance will save your company money.
If you have no intention of ever re-using a particular PRV, there is no reason I am aware of to inspect it for wear upon disposal. (If you aren’t keeping a database of rupture disk erosion when you remove them from service, you won’t inspect it would you?) That aside, you do need to have in your engineering files calculations and drawings that document the PSD system’s design basis. Then when a PSD arrives, you should have some QA process established that verifies that the received PSD matches the design basis. You should not exclusively rely on a PO for documentation as sometimes the wrong PSD is delivered or the wrong PSD was ordered in the first place. You also should have some QA process verifying that the PSD installed on the pressure vessel was intended for that PV.
Response #2
These questions cover all possible technical information for PRV (Pressure Relief Valve) inspection and testing. Not possible to answer in a short space.
I was a technical administrator for PRVs in a large Petrochemical company and can share my experience.
Firstly, just throwing away a PRV after five years is very dangerous. (Not cost; safety is the consideration here) I will tell you why. RV is the last line of defense. Or it should, I repeat, SHOULD open to release the excess pressure, set for MAWP (Maximum Allowable Working Pressure), which is normally higher than the design pressure) to protect the vessel on which it is mounted only for this purpose. It is not part of your protection systems, which are supposed to operate before the design pressure is reached and before the last line of defense has to operate.
One of the most important purposes of inspecting and testing a PSV is to confirm that it is ready to open when the pressure exceeds the maximum pressure set as per ASME Code. Before testing, the PSV has to be removed without disturbing it and moved carefully in the vertical position to the shop. Without any cleaning or repair, it should be tested in as-received condition. The purpose is to verify that it would have opened in situ at the set pressure. If it did not open, it means that if an overpressure had happened in operating condition, the chance of the vessel failing and leaking or bursting is real! The last line of defense would have failed! And if it was a hazardous fluid, imagine the consequence!
I think this principle will explain all your other questions with the same logic.
About the frequency, I will stress that it has nothing to do with whether the PRV is critical or non-critical. All are supposed to be critical as per the ASME mandate. But frequency is determined by the probability of the PRV failure, getting stuck due to corrosion products or process deposits. The frequency can be a minimum of 2 years if the service is a process and corrosive or with deposits which can prevent the PRV from opening; 3 years if it is processed but noncorrosive and clean, and 5 years for nonprocessed service like water. (Only as an example)
Lastly, if during as-received testing (with a frequency of 3 years as an example) the PRV failed, it is essential to reduce the next frequency to the previous test frequency when it passed or half the present frequency since it was proved that if the pressure protection systems would have failed within three years. If pressure increased above the set pressure, the vessel would have failed. On the other hand, we can increase the frequency if, on two consecutive tests, the PRV passed the test.
As per RCM (Reliability Centered Maintenance) methodology, the failure of PRV (Or PSV- I have covered both as PRV) is classified as “Hidden Failure,” and hidden failure prevention is possible only by verifying by testing in actual simulated conditions whether it passed or failed. If failed, reduce the frequency to be sure next time. Such tests are also done for instrument protective systems, which fail as hidden failures.