NPRA MAINTENANCE Q&A-1 ROTARY MACHINERY MAINTENANCE IS KEY PROCESS PLANT ACTIVITY

Dec. 17, 1990
Keeping rotating machinery, such as compressors, pumps, turbines, and blowers, in top operating condition is a key activity that helps maintain efficient process operations in refineries and petrochemical plants. At the most recent National Petroleum Refiners Association annual Refinery and Petrochemical Plant Maintenance Conference, May 23-25 in San Antonio maintenance engineers and managers discussed a wide range of rotating machinery maintenance topics.

Keeping rotating machinery, such as compressors, pumps, turbines, and blowers, in top operating condition is a key activity that helps maintain efficient process operations in refineries and petrochemical plants.

At the most recent National Petroleum Refiners Association annual Refinery and Petrochemical Plant Maintenance Conference, May 23-25 in San Antonio maintenance engineers and managers discussed a wide range of rotating machinery maintenance topics.

This annual conference draws operating company maintenance personnel from all parts of the world. Several technical papers that cover maintenance management and maintenance engineering topics are first presented at morning sessions. Then question and answer sessions follow in the afternoon.

In the question and answer sessions, moderated by Terrance S. Higgins, assistant technical director of NPRA, a panel of maintenance managers and engineers answers preselected questions from the industry. The audience is then invited to ask additional questions or provide additional information to the panel's answers.

During the Q&A session on rotating equipment, topics relating to bearings, seals, materials of construction, and preventive and predictive maintenance techniques for rotating equipment prompted extensive discussion.

WHAT ARE SUITABLE BEARINGS FOR VERTICAL PUMPS IN HIGH VAPOR PRESSURE SERVICE?

ROCK: We found two bearings materials that work fairly well depending upon what the service is. We found that carbon bearings work fairly well in some services. But mainly we have been using a graphalloy bearing material in our pumps and have been getting good life out of it.

CASADA: We have had similar poor experience with the high leaded bronze bearings that used to come in this type of pump. We did experience problems with them and have since switched to the graphalloy-type bushing material. We have attained excellent bearing life with this material.

EBATA: For bearings, one of our plants has been using babbitt-filled carbon bushings exclusively for the past 3 years. Prior to this material, various materials were used including bronze, oil impregnated bronze, and plain carbon. Plain carbon bushings were the least successful, often cracking and breaking if the pump became vapor locked.

The metal filled carbon bushings withstand vapor locking effects and do not appear to cause any scoring or damage to the shaft.

Surface hardened shaft sleeves are used in a number of our plants with no bearing problems.

WHAT PROBLEMS HAVE BEEN ENCOUNTERED IN GAS EXPANDERS USING GAS WITH EROSIVE FINES? WHAT ARE RECOMMENDED CHOICES OF BLADE MATERIALS?

ROCK: Our California refinery has a single-stage expander driving the main air blower in the FCC unit. As one might expect, the erosion comes from catalyst fines entrained in the flue gas. As originally installed, there was only a protective coating on the leading edge of the blades. After about 1 year of service, we found that we were getting erosion on the trailing edge and also at the root of the blade. We changed the blades out.

Currently we are running a Waspaloy blade that has a chromium carbide coating on all surfaces. We are getting good service life out of the blades. We do change them out during turnarounds which are basically every 3 years. The primary erosion that we are seeing at present is on the outer edge and that is primarily caused by catalyst building up between the shroud and the blades that actually generated contact from time to time. When we do change out the blades we have found that we have lost about one half to three quarters of an inch of the blades. But generally we see no change in the efficiency of the expander through the course of the run.

To remove the catalyst from the shroud, we generally go through a thermal spalling. We pull feed and drop the temperature. To see when we need to do this we watch the motor generator set up on our blower. We watch for it to start generating power which indicates that the clearances are decreasing. That is where the extra power comes from; that is when we spall.

We have tried other techniques of spalling, such as walnut shelling, and things of that type. Generally what we found is that when we use these techniques, the catalyst stays on the blades longer and does not come off evenly. You go through higher vibration periods for prolonged periods of time. With thermal spalling it all comes off fairly evenly, fairly quickly. You do have a high vibration blip but it generally comes right back down and we continue running.

SCHULZ: We operate a two-stage expander. Blade erosion and tip rubs have been the problems we have seen. We divide the blade erosion into two types: primary erosion that we attribute to larger particles in the main flow path, and secondary erosion that occurs from smaller particles and in the flow eddy paths. We think the amount of erosion is related to the amount of catalyst fines in the gas and to off-design operation.

We considered the blade coating to be of more importance in erosion resistance than the blade material. A suggested blade coating is a chromium carbide type coating - Union Carbide LC1C. Formally tungsten carbide type coatings were used, but they had problems due to differential thermal expansion. We think that Waspaloy blades are fairly standard for this service,

CASADA: We have had similar experiences with blades. We did, however, have a four-stage machine which had problems handling catalyst carry-over to the point that we ended up taking the machine out and putting in a single-stage unit. We have had catalyst build-up on the case and have gone to thermal spalling similar to Ultramar.

Stator blade materials which we have utilized have been Haynes Stellite 31 (a cobalt base alloy) with a chromium carbide coating. On the four-stage unit that we had, we used a Nimonic 90 base material (a nickel base alloy) with no surface coating.

Rotor blade material on machines in current use, are Waspaloy (nickel base alloy) with a chromium carbide coating. The use of the chromium carbide coating has appeared to decrease leading edge erosion to only a polishing effect.

EBATA: We have also experienced blade tip erosion coupled with thermal cracking on our catalytic cracker blower expander. Blades had to be replaced every 2 years. The erosion was due to the blade tips in contact with catalyst build-up on the stationary face. While we could solve the erosion problem by eliminating catalyst build-up through the on-line injection of walnut shells, this did not prevent the thermal cracking. The original blade material, A286, was replaced this past shutdown with Waspaloy, which has much better thermal tensile properties. Blade tips were surface hardened. It is my understanding that most companies with this problem are using the Waspaloy.

(Waspaloy-Haynes International Alloy: Ni, 58%; Cr, 19%; Co, 14%;

Nimonic 90-Inco Alloys International: Ni, 55%; Cr, 19%; Co, 18%. Stellite 31 Studi Deloyo Alloys: Ni, 10.5%; Cr, 25%, Co, 52%.)

BAZIL BURGESS (Hill Petroleum Co.): We have had a similar experience as the panel on expander blades. We are also experiencing erosion at the root of the blade. If we do a walnut hull blast about once a month, we can manage to keep the build-up off of the shroud and we do not have to worry about a thermal cycle. We had a couple of real shocking experiences with a thermal cycle and we try to avoid it.

ROBERT STEVENSON (Petro-Canada Products Co.): Would you explain in a little more detail please, your process for thermal shocking or spalling of the rotors?

SCHULZ: I am not familiar with the details of the procedure, although I know that to get the full spall we have to take oil out of the unit.

HIGGINS: The 1986 NPRA Process Q&A transcript, referred to a paper presented at the Seventh Annual Katalystics FCC Symposium. I believe it specified a procedure.

BAZIL BURGESS (Hill Petroleum Co.): Mr. Higgins, I can describe the procedure that we use. Our cat cracker has a charge of about 60,000 bbl. We normally reduce that to about 20,000-25,000 bbl and that helps us reduce the temperature about 200 F./hr. We would also add quench spray, which we put in the regenerator plenum chamber to maintain that temperature rate. We will come down 400 F.

This has been successful for us in taking catalyst off the shroud. (Procedures have previously been specified by Dave Linden of Ingersoll Rand.)

FRED VITALE (Mobil Oil Co.): Does the panel have any experience with blade tip refurbishment or blade refurbishment?

ROCK: We would not do it. It is not worth the effort nor the risk. We put new blades in every time.

FRED VITALE (Mobil Oil Co.): Even after only a 2-year run?

ROCK: We run 3 years and lose basically half to three quarters of an inch but we will not repair them, at least not at present.

WHAT METHODS HAVE BEEN USED TO SUCCESSFULLY DETECT COMPRESSOR VALVE LEAKAGE?

FRAZIER: we use two methods to detect compressor valve leakage. First, on a daily basis the operators check the valve caps with a hand-held, non-contact type pyrometer with a digital readout to detect differences in temperatures that would arise if you had leakage.

Another method we use from time to time is infrared thermal imaging of the valves to detect differences in temperatures.

The other successful method we use, on about a 90-day cycle, utilizes an engine/compressor analyzer which produces pressure-volume and pressure-time curves for the cylinders. These curves can be used to determine how much leakage you have and which valves are leaking.

If you have multi-valve cylinders, you can use an ultrasonic trace versus crank angle to pinpoint which valve is leaking.

SCHULZ: We use high temperature differences across the cylinder, valve cap temperature, and periodic ultrasonic noise/engine analysis.

CASADA: Conoco has used primarily the temperature method as well, but we have also used ultrasonics, the engine analyzers, RTDs embedded in the valve caps, and things of that nature.

EBATA: Most of our plants also use temperature as a means of detecting compressor valve problems. At one of our plants we are also using a combination of temperature and ultrasonic vibration measurement as a means of trending and predicting which valve should be watched or is in need of replacement. When the ultrasonic vibration level exceeds a specific value, data to produce a graph of ultrasonic vibration level versus time are taken by a data collector for further analysis. This combination of temperature and vibration has a success rate greater than 70%. The success rate, we found, depends on the type of transducer attachment (channel-lock pliers for the best results); transducer frequency range, and gain (varying degrees of success); the proximity of a transducer to the valve; and using the same test point at all times. We are also using this ultrasonic technique in our steam trap monitoring program where our success rate in predicting a faulty trap is 95%.

MIKE GLOVER (Exxon Chemical Co.): What has been the panel's experience with high frequency detection in regards to valve cap temperatures in determining whether the problem is the valve itself or an unloader?

CASADA: We have used the ultrasonics a little bit but have not had very much success with it in the compressor systems. We have used it to a degree in some other areas for tracking down relief valve leaks and things of that nature.

WHAT METHODS HAVE BEEN USED TO REDUCE OR ELIMINATE FUGITIVE EMISSIONS FROM PUMP AND COMPRESSOR SEATS?

FRAZIER: For centrifugal pumps we have gone to tandem seals with either a buffer fluid between the seals or a sensor to detect leakage in the seals. The reciprocating pumps and the reciprocating compressors are still a problem. Currently we are just monitoring the leakages and performing routine maintenance but this is probably not going to be acceptable very long. I think the problem right now is determining what kind of buffer gas or fluid to use and how to contain it and possibly work it off to a flare.

CASADA: We have done the same, whereby we have installed tandem seals with AP] Plan 52 set up where you have a little buffer tank in between with a buffer fluid. We have done that primarily on the new source compliance situations. On the rest of our existing equipment we have gone the monitoring route. On the compressors we have used vents to the flare.

ROCK: Our experience is similar. We are using tandem seals with buffer pots in the case of centrifugal pumps. On centrifugal compressors we have seal oil pots vented to the flare system. In the case of reciprocating machines, particularly at our California facility, we have nitrogen purged packing boxes again vented to the flare system. All other centrifugal pumps that do not have tandem seals get monitored once a shift. The idea is obviously to minimize the amount of time between the actual problem and repair.

KENNETH JACKSON (Mobil Oil Corp.): We have similar problems with fugitive emissions on our seals. I was wondering if any of the panelists have considered going with the magnetic drive pumps and can talk about some of the problems you might have had.

ROCK: We were considering trying magnetic drive pumps, but have not used one as yet. We are considering having a test unit in some particular service to see how well they work. I think most of the experience on them is from Europe.

MELVIN C. MCDANIEL (Texaco Chemical Co.): We have looked at the magnetic drives on a smaller scale. I was curious if the panel had had any experience with the magnetic drives, and what about magnetic drives on larger pumps?

CASADA: We have seen them but to my knowledge within Conoco, we do not have any.

C.C. WILSON: (Ethyl Corp.): We have approximately 10 magnetic drive pumps of various sizes and services. Limitations associated with magnetic drive pumps seem to be size and being incapable of handling high temperature fluids. The largest of the pumps we have in service is 40 hp. The most severe service, from the standpoint of temperature, is Dowtherm at 575 F. We have had some of these pumps in service for approximately 3 years and have had no mechanical failures.

WHAT METHODS ARE BEING USED TO PREVENT OR REMOVE CONTAMINANTS FROM EQUIPMENT TUBE OR SEAL OIL SYSTEMS?

ROCK: At Ultramar most of our experience has been predictive in nature. We take lube oil samples from our critical rotating equipment once a month and run them for physical properties, contaminant levels, particle analysis, etc. Generally speaking in the case of uncontaminated oil, we change it out. We do not try to filter it or refurbish it. The only exception is if we run into water in which case we have, at times, brought in contractors and centrifuged it.

WHAT PUMP AND SEALING ARRANGEMENTS HAVE BEEN USED SUCCESSFULLY TO HANDLE HIGH TEMPERATURE/VISCOUS MATERIALS SUCH AS ASPHALT, BITUMEN, AND RESID?

CASADA: We have primarily used metal bellow seals with silicon carbide versus silicon-carbide as far as faces. We have used in some of our pumps what is called a 11 canned faced design" whereby the silicon-carbide is not a solid silicon carbide piece, but is actually just a piece of silicon carbide held within a metallic can. The preferred method of flushing is a self-flush. In some situations where you cannot go with a self-flush because of contaminants or particulates, we have gone to an external flush and a dual seal arrangement.

EBATA: Most of our plants transfer asphalt with gear, screw, or steam driven simplex/duplex pumps. Worthington gear pumps are the most common used. Most of our pumps have packing, and some of our plants have been successful with carbon-filled Teflon packing. This has proven to have the longest life with minimal shaft sleeve wear. Metal bellows type mechanical seals having a "two hard face design" have also proven to be successful provided an anti-choke sleeve for low pressure steam quench under the bellows is installed.

SCHULZ: In several of our more critical applications we used double seals with external oil circulation at a higher pressure than the stuffing box. For these several applications the mean time between failures is about 5,000 hr. Otherwise we use a single metal bellows seal with a clean external flush, the API Plan 32. The mean time between failures for these applications is about 4,000 hr. In some cases we have also used the API Plan 23 in which the seal flush is recirculated through an external seal cooler with a pumping ring to reduce the flush requirements.

WHAT IS THE PANEL'S EXPERIENCE WITH MAGNETIC BEARINGS?

CASADA: We do not have any experience with magnetic bearings. It appears from looking at some of the industry information on magnetic bearings, that they have been under scrutiny since around the mid- 1970s. There are only two applications that I could find from the literature where magnetic bearings were in actual refinery service. Those two were in the same refinery. They were in a wet gas service on wet gas compressors. There has apparently been a fair amount of experience along natural gas pipelines using magnetic bearings. But again, most of the installations are relatively new.

JOHN POPE:(Shell Oil Co.): On the magnetic drive pumps, we have used a 25-hp Ingersoll Rand unit under a test program for about a year, with good success, i.e., no failures.

WHAT HAS BEEN THE PANEL'S EXPERIENCE WITH CENTRALIZED LUBE SYSTEMS?

ROCK: At Ultramar we basically have two systems that I would classify as central lube oil systems, one being those associated with critical pumps and compressors.

Generally we have had good experience with them, with one notable exception. We had one compressor-lube oil system that was supplied originally with centrifugal pumps, as opposed to your normal (PD) positive displacement pumps. When the pumps were switched, there would be a momentary dip in pressure which was enough to trip the system. We installed a nitrogen bladder-type buffer system to solve the problem.

The other centralized systems that we have in the plant are oil mist systems. These are basically multiple oil mist generators that generate oil mist to a header that goes to numerous bearings as secondary lubrication. We have had good success in using oil mist systems, and have really had no problem with the oil mist systems themselves.

MELVIN C. MCCDANIEL (Texaco Chemical Co.): Has there been any great concern about the environment with emissions from the oil mist system.

ROCK: We have not run into environmental concern, other than some of them make a mess at times.

WHAT ARE MORE SUBTLE CAUSES OF VIBRATION IN CENTRIFUGAL OR AXIAL FLOW COMPRESSORS BEYOND BALANCE, ALIGNMENT, AND CLEARANCE PROBLEMS?

ROCK: As you might guess, most of the problems that we run across in the vibration-related area generally do end up being alignment, balance, and clearance-related problems. Some of the more subtle things that we have run across or that we have looked for include things like cracked shafts. We have looked and found instances of it in some cases. We have found piping and structural resonance-related problems, rotor stack shifting problems, and soft foot problems.

We are now involved in a case where we are looking at coupling spring rates, in terms of axial vibration.

FRAZIER: Some years ago, I had an experience with an electric-driven unit where the electric motor had a stacked rotor. The rotor had been sent back to the manufacturer for rebuilding. When it was placed back in the unit, we could never find its magnetic center. Therefore, it put an axial load on the unit it was driving, resulting in a vibration. I think this is rather common with stacked rotors. They are very difficult to rebuild.

WHAT HAS BEEN YOUR EXPERIENCE WITH DRY-TYPE SEALS FOR COMPRESSORS? WHAT ARE THEIR LIMITATIONS?

CASADA: Our experience at Conoco is very limited with this type of sealing arrangement, as far as I could find. We do have a few Sundyne compressors that are running with a dry-type seal. However, I believe there are quite a number of companies that are using these.

As far as some of the limitations, you must have a buffer gas that is clean. You cannot have particulates or problems of that nature. Usually an external gas is used, other than the process stream, or the process stream is filtered significantly-down to 5 [L, or smaller. There are also pressure and speed considerations. As pressure or speed increase, leakage across the seal faces also increases. There are machines up to 10,000 rpm and pressures up to 2,000 psi.

Another limitation or item to be considered is the dampening effects of the oil system when changing from oil to the dry-type seals. The dry seals will no longer have some of the viscous dampening that you get with seal oil systems.

The last thing is just one of physical size. The dry-type seals require a tandem arrangement, which requires some modification to the compressor. If the space is not available to install this, then you have a problem.

EBATA: Our experience is limited, as well. We have recently installed a dry gas seal on our new isomerization unit recycle compressor. This is a low-pressure application, and the unit is about to be placed on stream. Running the gas seal with nitrogen so far has proven successful. We see no reason why it will not work with the on-stream gas.

JIM FOUGERE (PetroCanada Products): What type of filters do you use on the buffer gas for dry seals?

HUGH WEST (Shell Canada Ltd.): Over the past 4 years we have modified or bought seven compressors with dry gas seals, up to a speed of 22,000 rpm on two pipeline machine. We have a wet gas blower that has magnetic bearings, as well as a single dry gas seal. Our most recent experience was on a recycle hydrocracker machine which operates at 1,800 psi. We had quite a learning experience at these higher pressures.

I would like to note to everyone that the single most important thing to be careful of is the filtration of the gas that you are using for your purge. In our case, in two different situations, we use two different gases. One is the recycle from the discharge of the unit itself. If your gas can condense on its way from the discharge back through the filter back to your seal, be very careful because these are dry gas seals as 10,000 or 12,000 rpm, and if they get hit with some liquid, they literally explode.

One thing we have found that works well to get rid of liquid or make sure it does not happen is to install a coalescing filter. Units are available that will filter down to 0.3 m. So if your electric heat tracing or steam tracing on your purge does not work, at least the coalescing filter will catch it, and the drain off the bottom can be orificed and run to flare. We had six failures in that particular one over a short period of time, until we figured out all the bugs that were in it. It is now operating very satisfactorily, and we are quite happy. We are in the process right now of purchasing a large number of new centrifugals, and we expect to have dry gas seals for all of them.

Copyright 1990 Oil & Gas Journal. All Rights Reserved.