ft has been said that since the publication of API 613, Second Edition, the reliability of gears used in compressor drives has moved up several orders. This edition of API 613 mandated a conservative rating formula that was widely adopted in North America and to a more limited extent globally. It is credited with a significant reduction of new installation gear failures. Prior to that time, the feeling among operating companies was that reliability and gear drives were not compatible combinations .


Expanders have not been the essence of reliability. It is not that the expander design in itself has any significant problems. The problems for the most part seem to be related to the application. Most of the failures have been the result of the expander ingesting foreign substances, such as the catalyst in a catalytic cracking unit heat recovery application. Unlike the expansion section of the gas turbine, the inlet temperature is not as high, therefore, temperature is not a significant factor in reliability reduction.

Another vulnerable area of concern is the result of the design requirement for high temperature service. To accommodate the thermal growth potential, the various sections of the expander are quite compliant. As a result, piping and forces need to be kept to a minimum.



The process application can have a significant impact on the reliability of a compressor. The reliability aspects take on two paths: the nature of the gas being compressed and the correct compressor selection for the service.

When fitted with appropriate resistant materials, the compressors can reliably compress corrosive gases. The key is the appropriate resistance. This may be achieved to varying degrees of success. Most material corrosion resistance is temperature dependent. It is important to understand the compressor temperature profile. This includes making allowance for off-design performance. In the extreme, some gases can reach a very high level of corrosion by crossing a critical temperature point, which is more commonly called burning. Thankfully, most compressor services are not that dramatic. The point is that during the application or selection phase, the subject of corrosion must be fully explored. Factors such as the presence of moisture, or degree thereof, is significant for many of the corrosive gases.

Gases that tend to polymerize with temperature must be recognized and, as mentioned for corrosion, temperature limits must be imposed at the application phase to prevent problems later in operation. The only compressor, of the types covered, that is not as sensitive to the presence of polymers is the helical-lobe compressor. It should be recognized that even this machine does have limits.

The proper compressor selection for the application is probably the most important single factor to achieving long-term reliable operation. For all the factors that make the centrifugal one of the most reliable compressors, a misapplication will negate these most dramatically.

One of the most common pitfalls for misapplication is the size of the compressor itself. Because of the centrifugal compressor's good record as a reliable machine, there is a tendency to attempt to use this compressor for all applications. The centrifugal compressor does have a lower size limit. Application ranges were covered in Chapter 1. The lower limits are not sharply defined lines, and are heavily influenced by the specific application. For example, a multistage centrifugal with an inlet capacity of 800 acfm, on the lower end of the application range, may perform reliably on a single component clean gas and fail to operate well on a dirty or corrosive gas.

Gas sonic velocities must be kept in mind when selecting the compressor type. The helical-lobe compressor is tolerant of many items that can plague the other types, but some caution must be exercised with low sonic velocity gases. The nature of the operation of this compressor tends to produce a wide spectrum of pulsations. The higher molecular weight gases, due to their lower sonic velocity, are prone to excite acoustic and mechanical resonances in the piping associated with the machine. An interesting symptom of the presence of resonant responses is that the bolls tend to loosen and fall out. This certainly is not a good situation. The application must make certain that the appropriate pulsation treatment is used. Serious vibration problems have been experienced and reliability reduced in applications where the acoustic problem was not appropriately addressed.

Reciprocating compressor pulsations were covered in Chapter 3, but need to be mentioned with the discussions on reliability. Problems with reciprocating compressor pulsations and the potential for acoustic and mechanical resonances are very similar to those experienced with heii-cal-lobe compressors. The significant difference is the frequencies are much lower and the number of discrete frequencies per compressor are much less. However, piping vibrations can occur and there is always a danger that pipe breakage can occur. Aside from the piping acoustic and mechanical resonances, the presence of standing waves in the piping at the cylinder can reduce the output of the compressor. When the valve action is influenced by the standing wave, serious problems can occur.

For higher molecular weight gases, the centrifugal tip speeds must be reduced appropriately. The operating range of the compressor will be reduced significantly if operated at high Mach numbers. This is another example of the influence of speed in reliability.

High pressure operation, while generally quite successful, is another criteria that must be carefully considered. The increased pressure contributes to a higher density. As mentioned before, the higher density will increase the destabilizing forces on the rotor and can lead to rotor stabili ty problems. This is a problem unique to the centrifugal; however, in the helical-lobe compressor, high differential pressure contributes to rotor deflection, which, if carried to extreme, can certainly cause early failures. In axial compressors, high pressure, high density increases the bending load of the blading, which raises the stresses and potentially will shorten blade life and contribute to early failures.


Since many of the compressors covered in the earlier chapters fall into the category of custom-engineered units, the influence of experience on reliability is worth a few words. Experience may well be considered as being "viewed through the eyes of the beholder." How can one establish a base of experience if the unit under discussion is a custom unit?

With custom-engineered compressors, since exact duplicates are rare and prototyping of the exact unit is not feasible, the first consideration is the history of the component parts. Another aspect to consider is whether the application is an interpolation of existing designs or an extension or extrapolation. The latter, while a fact of life, needs to be understood at an early stage of the overall plant design so that the appropriate risk factors may be considered prior to the final commitment. Failure to properly evaluate the risk of an extension has resulted in serious compressor and driver problems, with long, expensive outages.

The critical components should be listed and checked to see that prior usage is within the range of experience. It should be determined if the impellers have been stage tested and whether this performance has been reflected in actual field experience. Prior use history of the blading geometry being proposed should be reviewed. This should include the chord length and blade length and, hopefully, experience with both longer and shorter applications. The rotor length (bearing span), which applies to all rotary and dynamic compressors, should have a successful history on similar frame size machines. All parameters relative to the list of critical items should be carefully scrutinized.

The supplier's general experience should be used to round out the review. Has the vendor used this corrosive gas at these concentrations' Does the supplier have experience with the specified gas at the trace levels of reactive gas, preferably at both higher and lower concentrations? An example of the trace gas found in hydrocarbons is H2S. It is very important that the supplier has experience at the anticipated pressure or density levels. The successful operation with gases in the low Mach number range should be demonstrated by the referencing of previous installations. The compression temperatures should be well within the experience range, because this is particularly critical for reliability with reciprocating compressor valves. Hopefully, these few examples are adequate to guide the reader in the proper direction, as a more comprehensive list would become quite lengthy and may not cover all the area needing review.


General Comments

It should be stated that operation should also be included in the list of considerations. It cannot be stressed enough that proven operating procedures be used. Maintenance should be performed in a timely manner as required. The manufacturer's instruction book is a good starting point for information on the last two items. The words "starting point" are important in that to this advice should be added the other considerations needed to form a well-established plant operating discipline that takes into account process, past experience, and industry knowledge. A machinery monitoring system should be in place to aid in early detection and diagnosis of possible problems, should they occur.

Gas Considerations

The use of drums for compressors was discussed earlier in the chapter. At that time, the recommendation was made that the drum be dry. This is still the prime consideration. However, if liquid carryover must be a way of life, then the question of particle size must be addressed. For the cen trifugal compressor, a rule of thumb with the usual disclaimers and exceptions is a particle size of 10 microns is safe, 100 microns marginal, and 1000 microns too large. The most significant disclaimer for this rule is that the gas with particles is acting in the erosion mode and not corrosion mode. The rule can be applied to the reciprocating compressor as well, with the note that the reciprocating compressor is less tolerant of liquid than the centrifugal. The helical-lobe compressor is somewhat of an exception in that it can better tolerate liquid particles. Liquids that are near the flash point and vaporize on impact are more prone to cause problems than particles that move through the compressor as mist or flash due to the heat of compression in the gas path. None of the compressors can tolerate liquid in large slugs.

Fouling and dirty gases have been discussed several times. These must again be mentioned in the realm of operations. While the potential of dirty gases should be considered and provided for in the design of the installation, it is not that unusual to discover this was not done. One reason may be that "off-design" operation that causes carryover of foreign material either as liquids or solids is not anticipated. While solutions to this type of problem must involve both engineering and operations, it is up to operations to recognize the problem. Some fouling problems are difficult to recognize. An example that illustrates this case consisted of a gas with extremely fine particles that were able to pass through the inlet filter and were not in themselves considered a problem. Superimposed on the gas was moisture, not high enough to cause concern in itself. As the gas was compressed, and in the application, intercooled, the moisture became saturated. The moisture mixed with the apparent harmless particles to form a paste. The paste was deposited on the blade surfaces of the downstream compressor stage and dried by the heat of compression. A classic case of fouling took place due to factors that were known but considered benign. If one can draw a moral from this example, it would be that apparent insignificant factors when taken singularly may not be so insignificant when occurring in unique combination.

Operating Envelope

All compressors have an operating envelope. The size is dependent on the compressor type and the application. If the appropriate design considerations were implemented, then operation within the envelope should be problem-free. The first problem stems from a communication gap. Engi neering and the equipment supplier do not always adequately communicate the bounds of the operating envelope to operations. Operations cannot be expected to operate inside an envelope that is not well-defined. Of course, operations may consciously choose to operate outside this envelope due to production pressures. It is important to consider the risks involved with this mode of operation. It may only reduce efficiency and no significant prob lem with equipment reliability may occur. For this, the only consideration is operating economics that may or may not be significant. However, accompanying "off-design" operation outside the envelope are factors thai may well impact reliability and the life of the equipment. It is relatively universal that all of the compressors will tend to increase in operating temperature when operating "off-design." This was illustrated in the previous chapters where performance was discussed. In most cases, the operating temperature has an upper limit after which deterioration may take place. Significant changes in pressure may well cause problems by exceeding deflection ratings. In dynamic compressors, molecular weight swings can cause unexpected surging. In the rotary compressors, changing the pressure ratio will cause over or under compression, which at the least, causes efficiency problems and, at worst, accentuates an acoustic response. These are a few examples to help illustrate the point; however, many other possible scenarios do exist. Reliable operation may well hinge on the ability to recognize and properly design for these eventualities.

System Components


Lubrication is a rather basic requirement for any operating machinery, which certainly includes compressors. Because it is so basic, it seems rather intuitive that reliability would be highly dependent on it [4|. In some cases, such as the basic rotary compressors covered in Chapter 4, rolling element bearings are standard and many of these require grease lubrication. It is important to remember that while the bearings should be greased at regular intervals, too much is just as serious as not enough.

Most of the compressors are of the pressure lubricated type and need some sort of lubrication system. For critical service, the API 614 lubrication system should be used. These systems were covered in Chapter 8. Since lubrication systems are key to the reliable compressor operation, a few things important to a reliable lubrication system seem in order. First and foremost is to keep the system as simple and basic as practical. Parts count is also a factor at this point. Direct-operated regulators (control valves) should be used whenever they are technically feasible. Computer control adds needless complication, particularly if the system is a basic lubrication system. It does not appear to add value to take the measured lube system pressures as transmitted signals to the control room. First it adds transmitters, usually in multiple form. Then the signals are put into the plant computer for processing, which requires someone to write the appropriate control code. The output signal is sent back to the lube oil console, where it has to be converted to a pneumatic signal to operate a control valve. If the parts count parameter has any merit, it should certainly be valid in this example. This is contrasted to using a direct-operated control valve that operates directly from a lube oil pressure sensing line to position the valve to control the lube oil pressure. Combined lube and seal systems sometimes become complex. With the dry gas seal, covered in Chapter 5, the need for these complex systems should be minimized. The rotary screw pump should be used on all but the large sized systems. A centrifugal pump is sometimes used when at the upper end of the positive displacement rotary screw pump range. Because of the shape of the centrifugal pump curve, selection is somewhat more critical and should be backed by proven field experience. The upper and lower viscosity range is important for the selection of the rotary screw pump as well as the centrifugal. An operations function should be mentioned at this point. Lubricating oil condition monitoring is a vital step to ensure good quality oil is sent to the compressor. Finally, shop testing is an important step in the procurement of a reliable lubrication system. While it may not be as exciting as a full-load compressor performance test, it may well influence the reliability of that very compressor train.


Couplings are discussed with reliability partly because in the past, as the single-train compressor installations were going through early growing pains, couplings were certainly a part of the pain. Not too many words need be used here, but possibly a bit of a reminder of the past will avoid problems in the future.

Not all that long ago, the main drive coupling of choice for compressor trains was the gear coupling. At first, the gear coupling was not a large contributor of problems. One guess is that on the early smaller trains, the couplings were oversized for the service. With oversizing, the lubrication quality, which was not good, did not cause any significant problem. It could also be said that any problems were not recognized because they were masked by all the other problems being experienced at the time. As the power density was increased, oversizing was slowly diminished, if for no other reason than economics, as the larger couplings were of a size that represented a significant cost to the equipment supplier. Also, the past oversizing had been more of a catalog rating thing and definitely was not recognized as such. The catalogs list the various sizes of couplings in a given manufacturer's line. The number of different sizes is limited by economics and the range of torque transmission coverage. Each catalog rating will cover a range of applications. The smaller sizes, when selected from the catalog, normally will fall somewhere between two successive sizes. It is normal to pick the larger size. This inherently oversizes the coupling. However, on the smaller sizes, this was not considered a problem because the cost benefit outweighs a custom design. As the power ratings went up, economics dictated a coupling selection sized closer to the application, resulting in the lubrication quality becoming more of a major factor. The fact that continuously lubricated couplings tended to centrifuge out the residual impurities in the oil (sludging) became quite significant. The presence of sludge prevented the lubricant from reaching the sliding area of the coupling and properly performing its function. Also the impact of misalignment became a factor. Highly loaded couplings with marginal lubrication tended not to operate well at high misalignment levels. The sensitivity to misalignment brought about better alignment techniques and helped minimize misalignment problems. Also, to aid in maintenance and alignment, longer spacers were used. Grease lubrication was improved, with the hopes of solving the sludging problem, but it never has achieved the run time being demanded of the ever larger trains. The advent of the flexible metallic element coupling solved the lubrication problem. It did not need lubrication. The flexible element coupling does require a discipline in machine-to-machine alignment, but given reasonable misalignments functions quite reliably.



Quality and reliability are not free, but poor quality and reliability cost more than good quality [2J. Costs include excessive down time and asso ciated loss of production, cost of repairs both in time and materials, and endless meetings. The list could go on but, hopefully, the point is made.

There are two general concepts of quality:

• Quality of design

• Quality of conformance

The type of compressor selected sets the quality of design. The API-based compressor is generally regarded as a higher quality than a catalog blower. Most of the chapter to this point has attempted to emphasize the many factors involved in design. Quality of conformance is a measure of how well the compressor met the specification. One of these measures is the achievement of the desired "run time." Another measure is the inspection for conformance of the parts to the drawings.

The role of quality in reliability would seem obvious, and yet at times has been rather elusive. While it seems intuitively correct, it is difficult to measure. Since much of the equipment discussed in this book is built as a custom engineered product, the classic statistical methods do not readily apply. Even for the smaller, more standardized rotary units discussed in Chapter 4, the production runs are not high, keeping the sample size too small for a classical statistical analysis. Run adjustments are difficult if the run is complete before the data can be analyzed. However, modified methods have been developed that do provide useful statistical information. These data can be used to determine a machine tool's capability, which must be known for proper machine selection to match the required precision of a part. The information can also be used to test for continuous improvement in the work process.

The advent of ISO 9000 certification has helped bring the quality issue to a focus. While ISO 9000 emphasis is on documentation of methodology and performance to the documented methods, the discipline that results is a direct step toward improvements in reliability. Out of this seems to come a goal of continuous improvement. This is a sharp contrast to the high growth era that took place in the early 1980s when poor quality on the part of the equipment suppliers was a significant factor in the lack of reliability at the field operation level.

Manufacturing Tolerances

One measure of quality is the conformation of the compressor parts to drawing dimensions. While it is customary to assign tolerances to a given dimension and base acceptance on the part being within the tolerance, there is a better way. The alternative requires a shift from the old paradigm and has been difficult to achieve in many of the compressor manufacturing locations. It has been proven in the Japanese automobile indus try. The concept is to make the part to the target dimension with measurements based on the deviation from target.

The benefits are many, though the basic cost justification is more elusive. The first benefit is that assembly is easier as the variations between parts is lower and less selective fitting is required. It inherently makes for better part interchangability. For the maintenance people, parts should fit without rework or hand fitting.

The inherent contribution to reliability should be compressor assemblies that truly reflect dimensions intended by the designer. This in turn should make the compressor perform to the level intended. It should offset the desire on the part of maintenance people to improvise when parts don't fit. This improvising can certainly become a problem as the original design intent is not normally known or understood at the field maintenance level.


The preceding discussion covered quite a number of topics relative to compressor reliability. A brief summary of the items covered may help the reader recap some of the thoughts presented. Issues that commonly play a major role in the search for reliability of compressors include:

• The concept of a robust design, in the compressor component and in the installation

• The hazards of liquid carryover and the need for suction drums

• The proper application of check valves

• Problems caused by pipe strain

• The significance of proper application, selection and sizing

• The impact of high molecular weight gases

• Corrosive and dirty gas effects

• The role of experience in avoiding problems

• Operating procedures

• Ancillary systems, such as lube oil systems and couplings

• Impact of quality

It is hoped that this material will give the reader an appreciation for the many different factors that interact to cause a compressor installation to be successful or a failure. It is sincerely hoped that when all the various aspects are critically addressed that any proposed installation will achieve the high reliability that a user may expect.


1. Abernethy, R. B, Weibull Analysis for Improved Reliability and Maintainability for the Chemical Industries, 3rd International Conference and Exhibition on Improving Reliability in Petroleum Refineries and Chemical Plants, Organized by Gulf Publishing Co. and Hydrocarbon Processing, Houston, TX, November 1994.

2. Coombs, C. F., Ireson, W. G. and Moss,, R. Y., Handbook of Reliability Engineering and Management, New York: McGraw-Hill, 1996.

3. Sohre, J. S., Tubomachinery Analysis and Protection, Proceedings of the First Turbomachinery Symposium, Texas A&M University, College Station, TX, 1970, pp. 1-9.

4. Sohre, J. S., Reliability—Evaluation for Troubleshooting of High-Speed Turbomachinery, ASME Petroleum Mechanical Engineering Conference, 1970.

5. Smith, J. B., Predicting Future Failure Risk with Weibull Analysis, First International Conference on Improving Reliability in Petroleum Refineries and Chemical Plants and Natural Gas Plants, Organized by Gulf Publishing Co. and Hydrocarbon Processing, Houston, TX, November, 1992.

6. Corley, J. E., Troubleshooting Turbomachinery Problems Using a Statistical Analysis of Failure Data, Proceedings of the 19th Turbomachinery Symposium, Texas A&M University, College Station, TX, 1990, pp. 149- 158.

7. Brown, R. N., ASM Handbook: Friction and Wear Technology, "Friction and Wear of Compressors," Vol 18, ASM International, pp. 602-608.

Living Off The Grid

Living Off The Grid

Get All The Support And Guidance You Need To Be A Success At Living Off The Grid. This Book Is One Of The Most Valuable Resources In The World When It Comes To When Living Within The Grid Is Not Making Sense Anymore.

Get My Free Ebook

Post a comment