Effective maintenance procedures are vital to the safe, efficient operation of processing plants.
Maintenance engineers and supervisors from refinery and petrochemical plants around the world met last May to discuss the latest practices and technology in maintenance engineering.
A major portion of the National Petroleum Refiners Association's annual Refinery and Petrochemical Plant Maintenance Conference, May 19-22, in San Antonio, followed a question and answer format.
Here are excerpts from the engineering Q & A session regarding predictive/preventive maintenance, rotating machinery care, and equipment inspection techniques.
In this session, delegates exchanged experiences and new ideas to keep their processing operations running at optimal capability. A panel of experts (see box) answered presubmitted questions, after which further questions and discussion were invited from the floor.
What techniques are you using to monitor reciprocating compressors?
SHAFFNER:
We currently monitor suction and discharge pressures and temperatures to indicate our valve and piston performance. On our newer compressors we are installing packing leak-rate monitors and rod-drop indicators. On our engine-driven compressors, we run pressure volume (PV) analyses quarterly and also do monthly brake mean effective pressure (BMEP) analyses to balance the fuel injectors.REYNOLDS:
We measure pressures and temperatures for each cylinder to generate PV diagrams. We monitor vibrations, take acoustic readings on valves, and analyze lube oils. We are now installing resistance temperature detecter thermocouples and valve covers for our machines and are starting to do some rod-drop measurements on our critical machines.CHAMBERS:
Our predominant monitoring tool for reciprocating compressors is a routine beta performance analysis by in-house personnel. We also collect vibration data on the driver, crankcase, and cylinders; we analyze the lube and seal oil; and monitor discharge temperatures for signs of valve damage.FRANKLIN:
We also monitor the process conditions. We use continuous rod-drop monitoring, monthly ferrography lube oil analysis, and vibration analysis.GUMPERT:
On our reciprocating compressors we monitor the oil monthly by use of ferrography and do monthly vibration readings using an IRD 890 data logger.SAMI ASAD
(Abu Dhabi National Oil Co.): My question is about beta analysis. Is this what we call recip trap equipment?SHAFFNER:
The pressure volume transducers make a graph that shows the actual volume and pressure in the cylinder as it is moving back and forth. From analysis, you can determine the performance of the compressor's reciprocating elements.CHAMBERS:
I believe they are one and the same. Recip Trap is a trade name for one of the beta analyzers.On what equipment and at what frequency do you use ferrography?
CHAMBERS:
As mentioned a moment ago, we use ferrography to analyze our oil from our compressors. We also use ferrography on large pieces of critical equipment that utilize a forced lubrication system. Most samples are taken on a monthly basis.FRANKLIN:
We use ferrography monthly on all our reciprocating compressors, most of our centrifugal compressors, and several of our vacuum pumps. We also will do ferrography to verify problems we think we see, using vibration. Moving into turnarounds, we will use ferrography to check a wider range of equipment to see if there is anything additional we need to do during the turnaround.GUMPERT:
We take oil samples monthly on the gear boxes for our reactor mixers, all our compressors (centrifugal and reciprocating), and our high-pressure reciprocating pumps used for well injection. We have found that on the high-pressure pumps, vibration works better than ferrography.We found that the Wilson Snyder high-pressure reciprocating pumps had a little magnet on the troughs which was picking up filings and metallic particles giving a false ferrography reading. Vibration data showed -bad bearings.
REYNOLDS:
We do lube oil testing on all our critical machinery, which includes viscosity, total acid number, flash point, and water content, on a monthly to quarterly basis, depending on location. But we include ferrography only on an exception basis, where specific problems have been encountered.For routine measurements, our experience indicates that it reveals more information than is useful, causing unnecessary concerns where no problems exist.
SHAFFNER:
We use oil ferrography on a monthly basis on our critical service compressors. Mainly what we look at is the wear particle analysis, to determine if we have an imminent bearing failure. Also, we look at the water content and viscosity to make sure our frequency on oil changes is appropriate.TONY SILIBERTI
(Aero-state Chemicals): We have been doing ferrography for a period of time now. I have really seen very little in the way of meaningful results. I have not seen any reliable trending data come out of this. I am just wondering about the experience of others. Have you been able to get reproducible data? How do you feel about using ferrography as a diagnostic tool?SHAFFNER:
We feel that it is a legitimate tool. One of the problems you get into with it is, if you do not have a good sample location and a good sampling procedure, then it is hard to get good data. It is a lot harder to get meaningful data from sump-lubricated equipment.GUMPERT:
Again, sampling is the most important part of ferrography. We definitely keep the same persons sampling all the time. We put in extra special sample points, especially on our centrifugal compressors, on the common drain before it goes to the sump. We feel some confidence.What is the latest technique used for packing high-pressure reciprocating pumps (discharge pressure over 2,000 psi)?
KIRK BARRINGTON
(Newfoundland Processing Ltd.): Known as the "Smart Seal" because of the self-adjusting nature of the packing, this sealing arrangement (Fig. 1) has been operating successfully in excess of 9 months without replacement, whereas previous seal designs rarely lasted more than 1-2 weeks before severe leakage developed.The seals operate 24 hr a day, 7 days a week, at pressures of 2,600-3,000 psi. The pump is a triplex horizontal design powered by a 125-hp driver and driven through a gearbox. Rod diameters are 2 5/8 in. with an 8 in. stroke.
Components consist of a spring-loaded bronze carrier, six rows of Teflon packing separated by Teflon antiextrusion rings, followed by a cast iron spacer. Springs are retained inside the carrier by a stainless steel washer. The cylinder nut is tightened to fully compress the springs, which are installed away from the direction of pressure.
As the packing compresses and seats itself, the springs allow for compensation as the packing adjusts itself automatically. To further extend packing and piston rod life an oil/air mist is introduced onto the rod surface at the point where the rod enters the cylinder. The clearance between the rod and the cast iron spacer acts as an oil trap. Product temperatures vary from 100 F. to 190 F.
What is the present trend of using variable speed drives in the refinery and petrochemical industries?
FRANKLIN:
Where we have had problems with mechanical variable speed drives, particularly in dusty atmospheres, we have gone to variable frequency and have had great success with it. Also, in applications where mechanical changers are only used occasionally, they tend to wear in that set position, causing problems. We have also had good luck with variable frequency there.In the past we have had some success, where we have had problems with slurries and the use of control valves, using variable frequency to control flow instead of a valve.
CHAMBERS:
We have found variable speed drives to be useful for varying heavy loads or situations where a satisfactory means of flow control is not available using more conventional technology. Examples might include cooling tower fans, forced draft and induced-draft combustion air blowers, and waste water pump services. Except for such specialized applications, we have not found the energy benefits associated with variable speed drives to outweigh their inherent complexity.One plant reports the installation of several adjustable frequency drives on alternating current (ac) motors. They have worked well for them and require minimal maintenance. In some cases they have hydraulic and variable speed ac units to compare side by side. They see both improved reliability and improved housekeeping with the variable speed units.
GUMPERT:
Typically at our plant we use variable speed drives on our cooling tower fans. That is the only application. At some of our other facilities we have used variable speed drives to replace direct current (dc) drive applications, which are a lot more costly motor-wise and drive-wise.REYNOLDS:
We are increasing the number of new variable speed drives (VSDs) on our new projects. With the renewed emphasis on energy efficiency, we expect the applications to continue increasing.Currently new projects include about 10-15% VSDS, and most of those are specific critical control applications, including close temperature control on air coolers, eliminating control valves on slurry pumps, some cooling water tower fans, mixers, conveyors, and extruders. In general, we do not believe current economics favor much retroactive conversion.
TONY SILIBERTI
(Aero-state Chemicals): I have some questions about vibrations on things like vertical pumps and right-angle gear boxes. In particular, I have a right-angle gear box on which I have experienced rather high levels of vibration. This is the only one I have.I would like to know whether this is something I can normally expect from this piece of equipment, or is that something I should be concerned about. Has anyone had experience with this type of machinery?
GUMPERT:
Typically, we take a look at what the trends are. We figure that is more important than trying to establish a good base reading. Naturally, you expect horizontal pumps to be lower. Maybe that right-angle drive is OK where it is, as long as you monitor the trends.On what equipment, other than hydroprocess reactors, have you used acoustic emission testing?
REYNOLDS:
We use acoustic emission (AE) testing on an infrequent basis as a global inspection technique for most pressure equipment, including pressure vessels, spheres, columns, and tanks, and more frequently on railroad cars. Each case in the refinery is separately justified, with the economics of the test often determined by the cost of access and setup, rather than the test itself.We carried out three cooldown AE tests on catalytic reforming unit (CRU) reactor spheres, with limited success. In one case, we had two different companies monitor the cooldown of a single reactor sphere, producing widely different results, neither of which correlated well with the cracking that we found with magnetic particle testing (MPT) in a followup inspection. This has caused us to pull back and do some AE application development work to understand how cooldown stresses relate to AE testing.
This work is still under way, but I am hopeful the results will lead us to productively expanding our use of AE, in combination with other NDE and cooldown tests.
The American Petroleum Institute is currently evaluating the results of an AE survey of the industry to answer this question in more detail, and I expect those results to be ready for the fall meeting in Houston.
SHAFFNER:
We have used AE on reformer reactors, hydrodesulfurizer reactors, also LPG spheres and bullet tanks. So far we have not really located any major problems with any of our equipment, so it is hard to tell if we have missed anything or not.CHAMBERS:
We have had some good experience with acoustic emissions testing to check for cracks in LPG spheres and bullets which may have been exposed to amine or caustic. We have also recently used acoustic emissions testing on hydrotreater heaters and piping.GUMPERT:
We have done acoustic emission testing on one tank. It was to verify that there was a leak in the tank bottom. Results showed that there was a leak and the approximate location, so it helped us there.At one of our Texas plants, AE was used to try to locate a leak in a large cooling coil in one of our process reactors. No leak was found in the coil. It turned out that the leak was in the coil of the sister reactor, and the leaking material was entering the drain system via some common piping.
So the AE-tested coil showed no leak and revealed to us that the leak was in the other reactor. We found the leak in the other coil visually.
IAN MCCONNEL
(Kuwait Petroleum Europoort B.V.): We have used AE extensively over the last 3 years. There is also a European AE users group which is working with most of the oil companies in Europe. We have been careful about our approach. We started using it on LPG spheres and correlated with destructive tests. We are now quite confident that AE on spheres correlates well with the real situation.We are cynical about the use of AE on tank bottoms. However, we have just done AE tests on a large crude tank, and we are now going to open it and clean it and try to see how that correlates. We use AE to follow up on hydrogen-induced cracking (HIC).
We recently took two vessels out of service for replacement, did destructive tests on the presence of HIC, and this was exactly where the AE said it was. So we are expecting to use AE more often, and we are feeling more confident about the results it gives.
What is the maximum interval between internal pressure vessel inspections? What basis do you use for justifying an increase in inspection interval (e.g., past 10 years)?
REYNOLDS:
In jurisdictions that do not set arbitrarily short, time-based intervals, we follow API 510 inspection code. Per that code, the maximum interval for the inspection of pressure vessels in corrosive service is half life, or 10 years. So we do not go past the 10 year maximum. However per the code, that interval can be for internal inspection or on stream inspection.In cases where our records and experience indicate that the corrosion rates are fairly uniform and fairly low, we may conduct only an on stream inspection at the 10-year interval, including extensive external ultrasound testing (UT) data taking. In other cases, where our records or experience indicate that corrosion rates may be nonuniform, unpredictable, or localized, or where other problems like environmental cracking are an issue, we do inspections at half life or more frequently.
To stay within the API 510 pressure vessel code, the only justification for going beyond 10 years without an inspection would be if it was specifically defined as a noncorrosive service by that code.
SHAFFNER:
Our refineries are typically on a 5-year internal inspection interval. We try to balance past performance and jurisdictional requirements. In some areas the interval is shorter. As Mr. Reynolds said, we never go past the 10 years either. If we do have a service where we know it is corrosive, we will shorten the interval. We may be into those particular pieces of equipment on 2-year intervals.One consideration is that if your relief valve inspection is going along with your pressure vessel inspection, you need to ensure your relief valves can go that long between inspections also.
GUMPERT:
We have plants in various states. Some states like Pennsylvania and Arkansas require regimented inspection intervals. In the others, we do follow API guidelines for the internal and/or external inspections.In one of our facilities, we have one or two vessels that have never been entered, and they are in desulfurized dry hydrocarbon service, known to be noncorrosive. Our corporate corrosion materials department favors an absolute limit of 10 years between internal inspections.
We use that interval only if the equipment has a history of noncorrosive service. So we are definitely staying within the 10-year limit, as far as internals. Typically, we do inspect our vessels every 2 or 3 years internally.
K. SUBRAMANIAN
(BHP Petroleum): Does the API code directly or indirectly cover the shell and tube heat exchanger type? The primary concern we have is when you have not tested within 10 years. When we pull out the tube bundle, do we need to hydrotest the tube bundle also?REYNOLDS:
The API 510 code is mostly concerned with the pressure boundary of a pressure vessel. So therefore it is only indirectly applicable to the tube bundle. I would say, if failure of that tube bundle could, in any way, threaten the integrity of the pressure boundary, then it would also fall under the API 510 code and need to be inspected.SHAFFNER:
When we inspect our exchangers, we always pull the bundles and clean the shells out, inspect the shells, and hydro the shells and tube bundles at the same time.REYNOLDS:
We classify our bundles A, B, or C. A is a critical bundle, which if it were to fail, may result in a safety or environmental incident. Class C, on the other hand, are those bundles that have relatively low economic impact. B Class is for all other bundles, which, if they were to fail, would result in significant economic loss. The classification of the bundle affects the scheduled inspection interval.JIM WINN
(Allied Signal Inc.): Concerning relief valve inspections on these pressure vessels, is it pretty much the standard now to pretest your relief valves before you repair them, as well as testing them afterward?SHAFFNER:
We prepop all our relief valves before we disassemble them. Basically, what we are trying to do is build a data base.REYNOLDS:
We, too, require a prepop of all of our relief valves when they are removed for service. We feel that if you do not prepop them, then you do not have very good information on how to adjust the interval, or properly set it for the next time.GUMPERT:
We previously did not do the prepop. We are starting to do that, because again we are trying to get some good intervals for the OSHA 1910 consideration, so all our specs now require a prepop, assuming it is in clean service. We have been cautioned by our repair shop that if they are in extremely dirty service, you can cause a lot more damage by prepopping the relief valves.RANDY CALLOWAY
(Southwestern Refining Co.): Do any of the panelists have aluminum alkyl tanks in service, aluminum alkyl storage vessels (catalyst for a Dimersol reactor)? What is your frequency of inspection on that vessel?REYNOLDS:
I believe we have vessels in that service, but I can't speak specifically to the time interval on the vessel itself. It would be at half life in corrosive service or 10 years max if it were in relatively noncorrosive service. But I do not have any first-hand experience with that service.SCOTT SAVAGE
(Amoco Oil Co.): I would like the panel to comment on internal inspections, specifically in the case of tray towers. Do you require all the tray manways to be open, or do you take that on a case-by-case basis?SHAFFNER:
We typically try to enter a section of a tower where we realize we have more corrosion potential. And typically we will not tray the entire tower. An exception is a small tower where you have only one or two manways to go in and out of. But on the very large towers, we typically will not open the manways up in the entire tower.GUMPERT:
We do it on a case-by-case basis, depending on the kind of repair history we have seen.V.R. KRISHNASWAMI
(Abu Dhabi National Oil Co.): What are the frequencies of testing of safety relief valves, such as in crude towers, vacuum towers, hydrodesulfurizer units, etc.?REYNOLDS:
Typically, we do ours at every turnaround. In normal refining service, we do not let our relief valves go over 5 years without some sort of service, and typically it is less than that, if there is any kind of fouling or corrosion or concern that the valve might not pop at the right pressure.SAMI ASAD
(Abu Dhabi National Oil Co.): I have a question about inspection of heavy wall reactors by ultrasonic and X-ray. What is the interval for removing insulation and doing full external inspection. Is it done every 5 years, every 10 years, or is there any experience preference with that?REYNOLDS:
We have not yet completely stripped a large hydroprocess reactor for external inspection. But we are definitely stripping a portion of it, say in the 1020% range, to look at the areas that we think may be the most susceptible to cracking, if cracking were to occur, such as those areas most susceptible to triaxial stresses.We do magnetic particle and external ultrasonic scanning on a spot basis.
SHAFFNER:
We will typically strip the nozzles and check the nozzle wells and also we will take spot checks on the hemi heads on both ends. Again, if you do some acoustic emission testing that points out that you may have a crack in some area, then we normally go back and strip insulation in those areas and check there also.FRANK A. BALCER
(Chevron U.S.A. Inc): Do the members of the panel have the ability to block in your safety valves for removal when testing, when the plant is in service, or do any of you people do what we call, live relief jobs, where we pull a valve off the relief header hot?SHAFFNER:
We have both situations. On our newer units we have installed block valves and bypass valves beneath the relief valves. Typically, when the units are operating, those are left chained open. What we will do in the event we have to remove a relief valve, we block them in and station a man there to watch the pressure gauge with control over the bypass valve.We will take that relief valve down and we will work on it on an emergency basis and then replace it right away. Once that is done, we open the bottom block valve back up to the flare header and chain and lock the bottom block valve and the flare header valve open.
REYNOLDS:
Our experience is very much the same as Mr. Shaffner mentioned, only we do that now on a scheduled basis, not just a high priority basis. We have made a real effort to try to get that kind of routine maintenance out of our turnaround, so that we do as much of it on stream as we safely can. But at the same time, it is not inexpensive because it does require a person to stand by while the valve is out of service.Copyright 1992 Oil & Gas Journal. All Rights Reserved.
