Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information
Petroleum Engineering Handbook
Larry W. Lake, Editor-in-Chief
Volume III – Facilities and Construction Engineering
Kenneth E. Arnold, Editor
Copyright 2006, Society of Petroleum Engineers
Chapter 7 – Compressors
This chapter provides an overview of the primary categories of natural gas compressor services and a description of the different classifications and types of compressors available to the industry. Adiabatic and polytropic compression theory are discussed with supporting definition of terminology. Specific topics relating to compression theory include power requirement, isentropic exponent, compressibility factor, intercooling, adiabatic and polytropic efficiency, actual and standard volume flow rates, mass flow rates, inlet and discharge pressures, inlet and discharge temperatures, and adiabatic and polytropic head. Major components and construction features of centrifugal and reciprocating compressors are emphasized. Installation, safety, and maintenance considerations also are discussed.
For centrifugal compressors, the performance characteristic curve is presented with emphasis on process control of capacity by speed variation, suction throttling, or variable inlet guide vanes. Process control to avoid operation in a damaging surge condition is also addressed.
The discussion on reciprocating compressors includes a description of process configuration for multistage units, as well as an explanation of the concepts of speed control, inlet throttling, recycling, pressure relief, blowdown, and distance piece venting and draining.
- 1 Overview
- 2 Classification and Types
- 3 Compression Theory
- 3.1 Isentropic (Adiabatic) Compression
- 3.2 Polytropic Compression
- 3.3 Head
- 3.4 Adiabatic (Isentropic) Efficiency
- 3.5 Polytropic Efficiency
- 3.6 Compressibility Factory
- 3.7 Flow (Capacity)
- 3.8 Compression Ratio
- 3.9 Intercooling
- 3.10 Power Requirement
- 3.11 Compressor Selection
- 3.12 Number of Stages of Compression
- 4 Centrifugal Compressors
- 5 Reciprocating Compressors
- 6 Summary
- 7 Nomenclature
- 8 Superscripts
- 9 General References
- 10 SI Metric Conversion Factors
Compressors used in the oil and gas industry are divided into six groups according to their intended service. These are flash gas compressors, gas lift compressors, reinjection compressors, booster compressors, vapor-recovery compressors, and casinghead compressors.
Flash Gas Compressors. Flash gas compressors are used in oil handling facilities to compress gas that is "flashed" from a hydrocarbon liquid when the liquid flows from a higher pressure to a lower pressure separator. Flash gas compressors typically handle low flow rates and produce high compression ratios.
Gas Lift Compressors. Gas lift compressors are frequently used in oil handling facilities where compression of formation gases and gas lift gas is required. Gas lift compressor duty is frequently of low to medium throughput with high compression ratios. Many gas lift compressors are installed on offshore facilities.
Reinjection Compressors. The reinjection of natural gas is employed to increase or to maintain oil production. Reinjection compressors can be required to deliver gas at discharge pressures in excess of 10,000 psi. Reinjection compressors also are used for underground storage of natural gas. Compressors, applied to these services, have large compression ratios, high power requirements, and low volume flow rates.
Booster Compressors. Gas transmission through pipelines results in pressure drop because of friction losses. Booster compressors are used to restore the pressure drop from these losses. Selection of these compressors involves evaluating the economic trade-off of distance between pipeline boosting stations and life-cycle cost of each compressor station. Booster compressors also are used in fields that are experiencing pressure decline. Most centrifugal pipeline booster compressors are gas turbine driven, although the use of variable-speed motor drives is becoming more prevalent. Low-speed integral gas engine reciprocating compressors also are used for gas transmission applications. Booster compressors typically are designed for high throughput rates and low compression ratio. Many booster applications can be configured in a single-stage centrifugal compressor.
Vapor Recovery Compressors. Vapor recovery compressors are used to gather gas from tanks and other low-pressure equipment in the facility. Often the gas from a vapor recovery compressor is routed to a flash gas, gas lift, or booster compressor for further compression. Low suction pressures, high compression ratios, and low gas throughput rates characterize these compressors.
Casinghead Compressors. Casinghead compressors are usually used with electric submersible pumps and rod pumps where formation gas is required to be separated downhole and then transported through the annulus. Often the compressor discharge is routed to either a booster or flash gas compressor or to a low-pressure gathering system. Like vapor recovery compressors, casinghead compressors operate with low suction pressures, high compression ratios, and low gas throughput rates.
Classification and Types
Compressors are classified into two major categories—positive displacement and dynamic (kinetic) compressors. Positive displacement compressors are further divided into reciprocating and rotary types. Reciprocating compressors are positive-displacement machines in which the compressing and displacing element is a piston having a reciprocating motion within a cylinder. Rotary compressors are positive-displacement machines in which the compressing and displacement are affected by the positive action of rotating elements. (Note: For positive displacement compressors, the focus of this chapter is on reciprocating compressors.)
Dynamic compressors are continuous-flow machines in which a rapidly rotating element accelerates the gas as it passes through the element, converting the velocity head into pressure, partially in the rotating element and partially in stationary diffusers or blades. Dynamic compressors are further divided into centrifugal, axial-flow, and mixed-flow types. (Note: Out of these three types of dynamic compressors, only centrifugal compressors are addressed in this text.)
There are two types of reciprocating compressors—high speed and low speed. The high-speed category also is referred to as "separable," and the low-speed category also is known as "integral."
The American Petroleum Institute (API) has produced two industry standards, API Standard 11P and API Standard 618, which are frequently employed to govern the design and manufacture of reciprocating compressors.
Separable CompressorsThe term "separable" is used because this category of reciprocating compressors is separate from its driver. Either an engine or an electric motor usually drives a separable compressor. Often a gearbox is required in the compression train. Operating speed is typically between 900 and 1,800 rpm.
Separable units are skid mounted and self-contained. They are easy to install, offer a relatively small initial cost, are easily moved to different sites, and are available in sizes appropriate for field gathering—both onshore and offshore. However, separable compressors have higher maintenance costs than integral compressors.
Fig. 7.1 is a cross section of a typical separable compressor. Fig. 7.2 shows a separable engine-driven compressor package.
Integral CompressorsThe term "integral" is used because the power cylinders that drive the compressor are mounted integrally with the frame containing the compressor cylinders. Integral units run at speeds of between 200 and 600 rpm. They are commonly used in gas plants and pipeline service where fuel efficiency and long life are critical. Integral compressors may be equipped with two to ten compressor cylinders with power ranging from 140 to 12,000 hp.
Integral compressors offer high efficiency over a wide range of operating conditions and require less maintenance than the separable units. However, integral units usually must be field-erected and require heavy foundations and a high degree of vibration and pulsation suppression. They have the highest initial installation cost.
Fig. 7.3 is a cross section of a typical integral compressor. Fig. 7.4 shows an integral compressor package.
Rotary Positive Displacement Compressors
The two most common types of rotary positive displacement compressors are vane compressors and screw compressors.
Vane Compressors. The vane-type compressor consists of a cylindrical rotor with longitudinal slots in which radial sliding vanes are fitted. The rotor is positioned eccentrically within a cylindrical housing. The spaces between adjacent vanes form pockets of decreasing volume from a fixed inlet port to a fixed discharge port. Compressor inlet and discharge valves are not employed in the design. A vane compressor always compresses the gas to the design pressure defined by the manufacturer, regardless of the pressure in the system in which the compressor is discharging.
Screw Compressors. The screw compressor, also known as a helical-lobe or spiral-lobe compressor, is a positive displacement rotary design that compresses gas between intermeshing helical lobes and chambers in the compressor housing. Screw compressors do not use valves. Their compression ratio is determined by the wrap angle of the lobes and the location of the opening edges of the discharge port. Of the various compressor types, screw compressors are best able to accommodate liquid carryover.
Centrifugal CompressorsIn a centrifugal compressor, energy is transferred from a set of rotating impeller blades to the gas. The designation "centrifugal" implies that the gas flow is radial, and the energy transfer is caused from a change in the centrifugal forces acting on the gas. Centrifugal compressors deliver high flow capacity per unit of installed space and weight, have good reliability, and require significantly less maintenance than reciprocating compressors. However, the performance characteristic of centrifugal compressors is more easily affected by changes in gas conditions than is the performance of reciprocating compressors.
The physical size (diameter) of a centrifugal compressor is determined by the volumetric flow rate at the inlet. The compression ratio (or head) determines the number of stages (length). The rotating speed of a centrifugal compressor is an inverse function of diameter to maintain a desired peripheral speed at the outer diameters of the impellers regardless of the physical size of the compressor. Very large (i.e., high-volume) flow compressors may operate at speeds as low as 3,000 rpm. Conversely, low-volume flow compressors may operate at speeds up to 30,000 rpm. Power requirement is related to mass flow, head, and efficiency. Depending on the particular application, centrifugal compressor powers can range from as low as 500 hp (400 kW) to more than 50,000 hp (40 MW).
At low volume flow rates, the width of the gas passages in a centrifugal compressor becomes narrow, and the effects of friction become significant, resulting in reduced efficiency. For this reason reciprocating compressors often are more appropriate for low-volume flow applications. For further discussion of this subject, see Sec. 7.3.11.
The API has produced an industry standard, API Standard 617, which is frequently used to govern the design and manufacture of centrifugal compressors. A typical centrifugal compressor package is shown in Fig. 7.5. The compressor shown is mounted on a single baseplate and is driven by an electric motor.
Both positive displacement and dynamic compressors are governed by a few basic principles derived from the laws of thermodynamics. See the chapter on thermodynamics in the General Engineering volume of this Handbook for a basic discussion of thermodynamics. This section defines terminology and discusses the operating principles essential for understanding compressor design, operation, and maintenance.
Isentropic (Adiabatic) Compression
An adiabatic process is one in which no heat is added or removed from the system. Adiabatic compression is expressed by
where k = Cp/Cv = ratio of specific heats, dimensionless.
Although compressors are designed to remove as much heat as possible, some heat gain is inevitable. Nevertheless, the adiabatic compression cycle is rather closely approached by most positive displacement compressors and is generally the base to which they are referred.
A polytropic process is one in which changes in gas characteristics during compression are considered. Dynamic compressors generally follow the polytropic cycle as defined by the formula
where n = polytropic exponent.
The polytropic exponent n is experimentally determined for a given type of machine and may be lower or higher than the adiabatic exponent k. Because the value of n changes during the compression process, an average value is used.
When inlet and discharge pressures and temperatures are known, the polytropic exponent can be determined from the relationship
Head is simply the work expressed in foot pounds per pound of gas or N-m/kg. At a given compressor speed and capacity, the head developed by a centrifugal compressor is the same regardless of the nature of the gas being compressed. The pressure rise produced by the given amount of head varies with the density of the gas.
Isentropic (Adiabatic) Head. In an isentropic compression process, head is calculated by Eq. 7.4.
|His||=||isentropic head, ft-lbf/lbm,|
|zavg||=||average compressibility factor, dimensionless,|
|Ts||=||suction temperature, °R,|
|S||=||gas specific-gravity (standard atmospheric air = 1.00),|
|Pd||=||discharge pressure, psia,|
|Ps||=||suction pressure, psia.|
|Hp||=||polytropic head, ft-lbf/lbm,|
Adiabatic (Isentropic) Efficiency
Adiabatic efficiency is defined as the ratio of work output for an ideal isentropic compression process to the work input to develop the required head.
For a given compressor operating point, the actual or predicted isentropic efficiency can be calculated with Eq. 7.6.
|Ts||=||suction temperature, °R,|
|Td||=||discharge temperature (actual or predicted), °R,|
|k||=||ratio of specific heats, Cp/Cv.|
|N||=||number of moles,|
|R||=||constant for a specific gas,|
In reality, all gases deviate from the ideal gas laws to some degree. This deviation is defined as a compressibility factor, z , applied as a multiplier to the basic formula. Therefore, Eq. 7.8 is modified to include the compressibility factor as shown next.
Compressor flow (capacity) can be specified in three ways: mass (weight) flow, standard volume flow, and actual (inlet) volume flow.
Mass (Weight) Flow. Mass flow is expressed as mass per unit of time, most often pounds-mass per minute (lbm/min) or kilograms per minute (kg/min). Mass flow is a specific value independent of gas properties and compressor inlet conditions. Mass flow can be specified on either a wet (water vapor included) or dry basis.
Standard Volume Flow. Standard volume flow is the most common term used by the industry to describe volumetric flow because it is independent of actual gas pressures or temperatures. It is the volume per unit of time using pressures and temperatures that have been corrected to "standard" conditions. These conditions apply to pressure, temperature, molecular weight, and compressibility. The standards must be known and held constant. For purposes of this text, the standard conditions used are
|molecular weight||=||MW of subject gas.|
Standard volume flow is usually dry and expressed in millions of standard cubic feet per day (MMScf/D).
Actual (Inlet) Volume Flow. Actual volume flow is defined as the amount of volume per unit of time at the inlet to the compressor. Actual volume flow is normally expressed in actual cubic feet per minute (ACFM) or actual cubic meters per hour (m3/hr). When gas composition and pressure and temperature are known, the specification of actual volume is appropriate because the fundamental performance characteristic of the compressor is sensitive only to actual volume flow at the inlet (see Sec. 7.4.2).
Mass flow can be converted to actual volume flow with Eq. 7.10.
|W||=||mass flow, lbm/min.,|
|R||=||universal gas constant = 1,545,|
|Ts||=||suction temperature, °R,|
|zs||=||compressibility at inlet,|
|Ps||=||absolute suction pressure, psia.|
Compression ratio, Rc, is simply the absolute discharge pressure divided by the absolute suction pressure. As expressed in Eq. 7.3, temperature ratio increases with pressure ratio. Temperature limits related to the mechanical design of compressors often will dictate the maximum pressure ratio that can be achieved in a stage of compression. (Refer to Sec. 7.3.9 for a related discussion on intercooling.)
Where large pressure ratios are needed, splitting the compression duty into one or more stages with intercooling between stages can be the most energy efficient arrangement. The energy savings must be compared with the capital and maintenance investment necessary to provide the cooling. In addition to the thermodynamic benefit, intercooled compression systems result in lower discharge temperatures, which reduce the need for special compressor materials.
The total power requirement of a compressor for a given duty is the sum of the gas power and the friction power. The gas power is directly proportional to head and mass flow and inversely proportional to efficiency. Mechanical losses in the bearings and, to a lesser extent, in the seals are the primary source of friction power.
For centrifugal compressors, the gas power can be calculated as
|GHP||=||gas power, horsepower,|
|W||=||mass flow, lbm/min.,|
|Hp||=||polytropic head, ft-lbf/lbm.|
|P1||=||inlet pressure, psia,|
|V1||=||inlet volume, ACFM,|
|P2||=||discharge pressure, psia,|
|CE||=||compression efficiency (assume 0.85 for estimating purposes).|
Compressor SelectionProper selection of the compressor type and number of stages can be accomplished only after considering a number of factors. (For the purposes of this chapter, discussion is limited to centrifugal vs. reciprocating compressors.) Basic information needed for the proper selection includes volume and mass flow of gas to be compressed, suction pressure, discharge pressure, suction temperature, gas specific gravity, and available types of drivers.
The required volume flow and discharge pressure define a point on a graphic representation of compressor coverage, as shown in Fig. 7.6. Examination of this chart reveals that, in general, centrifugal compressors are appropriate for high flow applications, and reciprocating compressors are better suited to low flow rates.
Number of Stages of Compression
Using the specified overall pressure ratio and suction temperature (and an assumed efficiency), the discharge temperature for compression of gas with a known k value in a single stage can be estimated by rewriting Eq. 7.7.
|T2||=||estimated absolute discharge temperature, °R,|
|T1||=||specified absolute suction temperature, °R,|
|P1||=||specified absolute suction pressure, psia,|
|P2||=||specified absolute discharge pressure, psia,|
|k||=||ratio of specific heats,|
|ηp||=||assumed polytropic efficiency,|
|≈||0.72 to 0.85 for centrifugal compressors,|
|≈||1.00 for reciprocating compressors.|
If the single-stage discharge temperature is too high (typical limit is 300 to 350 °F), it is necessary to configure the compression equipment in more than one stage. Calculating the compression ratio per stage with Eq. 7.15 does the evaluation of a multistage design.
|Rsect||=||compression ratio per section,|
|n||=||number of sections.|
Using the previous equations and prudent assumptions, it is possible to determine the minimum number of stages required to accomplish a given overall compression ratio without exceeding temperature limits.
Multistage centrifugal compressors can be arranged in a variety of flowpath configurations employing from one to ten impellers, depending on the head required for the process duty. When intercooling is not needed, the arrangement is usually a straight-through (inline) configuration. For applications that require intercooling, the resulting two-section compressor may be configured in either an inline (compound) or back-to-back arrangement. For high-flow/low-head applications, a double-flow configuration is sometimes employed. In a double-flow arrangement, half of the flow enters the compressor through an inlet connection at each end of the casing and exits the casing through a common discharge connection in the center. All of these configurations described are beam-type designs in which the impellers are located between the radial bearings.
Single-stage centrifugal compressors may be configured as a beam design or with an overhung impeller arrangement. In the overhung configuration, the impeller is located at the nondrive end of the shaft (outboard of the nondrive end radial bearing).
Major ComponentsThe major components of various centrifugal compressor flowpath configurations are illustrated in Figs. 7.7 through 7.10. This section describes the major elements of centrifugal compressors.
Case (Casing or Housing). The case (casing or housing) is the pressure-containing component of the compressor. The case houses the stationary internal components and the compressor rotor. Bearings are attached to the case to provide both radial and axial support of the rotor. The case also contains nozzles with inlet and discharge flange connections to introduce flow into and extract flow from the compressor. The flange connections must be properly sized to limit the gas velocity as necessary. The case is manufactured in one of two basic types: horizontally or vertically split. Construction can be cast (iron or steel), forged, or fabricated by welding.
Horizontally (Axially) Split Case. A horizontally split case is split parallel to the axis of the rotor. The upper half of the case is bolted and doweled to the lower half. Access to the internals of the compressor for inspection and maintenance is facilitated with this case design (especially when the process piping connections are located on the bottom half of the case). The horizontally split design is inherently pressure-limited to prevent gas leakage at the case split joint.
Vertically (Radially) Split Case. This case is split perpendicular to the axis of the rotor. Heads (end covers) are installed at both ends for pressure containment. The vertically split case configuration is capable of handling higher pressures than the horizontally split type. The rotor and stationary internals are assembled as a cylindrical inner bundle that is inserted axially through one end of the case. Inspection and maintenance of a radially split centrifugal compressor require that the inner bundle be removed for disassembly. Removal of the inner bundle requires that sufficient space be provided in the layout of the compressor installation.
Rotor Assembly. The compressor rotor is fundamentally an assembly of impellers mounted on a steel shaft. Miscellaneous hardware such as a thrust balance drum (balance piston), impeller spacers, seal sleeves, a thrust disc, and one or two couplings are additional rotor components. A typical compressor rotor is pictured in Fig. 7.11.
The impellers impart velocity to the gas with blades that are attached to a rotating disc. The impeller blades are forward-leaning, radial, or backward-leaning (with respect to the direction of rotation) depending on the desired performance characteristic curve. Backward-leaning blades tend to provide the widest operating range with good efficiency. They are the most commonly used blade shape. Proper sizing of the impeller flow channels is determined by the volumetric flow rate to control gas velocities through the impeller. This means that, in a multistage compressor, the impellers must be properly sized for peak performance and properly matched to accommodate the volumetric flow rate reduction through the compressor. Impellers can be of the open type without a cover plate or the closed type that incorporates a cover plate attached to the blades. Most multistage compressors use the closed-type impeller design. Impeller construction can be riveted, brazed, electron beam welded, or welded conventionally. For most applications, high-strength alloy steel is selected for the impeller material. Stainless steel is often the material of choice for use in corrosive environments. Because the impellers rotate at high speeds, centrifugal stresses are an important design consideration, and high-strength steels are required for the impeller material. For gases containing hydrogen sulfide, it is necessary to limit the impeller material’s hardness (and therefore strength) to resist stress corrosion.
Multistage centrifugal compressor rotors have natural resonant frequencies that must be outside the operating speed range. Rotordynamic design considerations can limit the maximum number of stages per case or, stated another way, limit the maximum speed for a given number of stages.
Stationary Components. After the gas enters the compressor through the inlet nozzle, it must be directed to the inlet of the first stage impeller in a way that uniformly distributes the flow to the impeller at a desired velocity. A system of internal stationary components is designed to deliver the gas to the first impeller with minimal pressure drop. Stationary inlet guide vanes are normally positioned adjacent to the impeller inlet. Variation of the inlet guide vane angles can be employed to adjust the flow capacity of the compressor’s performance characteristic curve. However, a variable inlet guide vane system introduces mechanical complexity as well as additional sealing considerations (see the subsection on Flow Control in Sec. 7.4.5).
The gas exits the impeller at high velocity and enters a diffuser passage. The diffuser is an important part of the stationary flowpath that usually comprises two parallel walls forming a radial flow channel. In the diffuser, the gas velocity decreases and dynamic pressure is converted to static pressure. Diffusers can be either vaneless or vaned. After exiting the diffuser passage, the flow encounters a return bend, which creates a 180-degree turn in the direction of flow (i.e., from radially outward to radially inward). Following the return bend, the flow enters a vaned return channel that directs the flow inward to the next impeller. The function of the return channel is (in the same manner as the first stage inlet system) to uniformly deliver flow to each impeller with minimal losses. The inlet guide vanes are located at the exit of the return channel. The components that form the return channel are called "diaphragms," and the diffuser passages are the spaces between adjacent diaphragms. Inlet guide vanes can be attached to a separate piece fitted into the diaphragm or an integral part of the diaphragm.
Following the last stage impeller, the gas must be collected and delivered to the discharge flange. The stationary component, typically used for this purpose, is a discharge volute. The volute must be well matched to the discharge nozzle to minimize pressure losses. Discharge nozzle velocities also must be kept within limits to avoid excessive noise levels. All of the stationary components previously described play an important role in overall compressor performance.
Bearings and Seals. Centrifugal compressors are equipped with two radial (journal) bearings to support the rotor weight and position the rotor concentrically within the stationary elements of the compressor. One thrust bearing also is used to ensure that the compressor rotor is maintained in its desired axial position. The thrust bearing usually is a "double-acting," tilt-pad design installed at both sides of a rotating thrust disc. Proper rotor axial position is thereby assured regardless of the direction of the net axial pressure forces acting on the rotor.
Two distinct categories of compressor seals are used: internal seals and shaft seals. Internal seals minimize internal recirculation losses between stages and across the thrust balance drum. Labyrinth type seals are customarily used for this purpose to maximize operating efficiency. Shaft seals are required to seal the gas inside the compressor at the point where the compressor rotor shaft penetrates the case. This vital sealing function is necessary to prevent escape of process gas to the environment surrounding the compressor. Dry gas seals are the most commonly used type of shaft seal. Liquid film seals are sometimes used.
Labyrinth Seals. Labyrinth-type seals are used to minimize recirculation losses within the compressor. A labyrinth seal consists of a number of teeth (knife-edges) that can be either stationary or rotating. Stationary labyrinth teeth are fitted to the compressor stationary components very close to the compressor rotor (see Fig. 7.12). Sealing action is the result of flow resistance caused by repeated throttling across the labyrinth teeth. Labyrinth seals are designed so that one of the two adjacent parts (labyrinth teeth and rotor) is relatively soft. The softer material yields on contact without damage to the harder material. Compressor manufacturers select labyrinth seal clearances that are as tight as practical to minimize leakage while avoiding heavy rubbing with the rotor.
Dry Gas Seals. Beginning in the late 1980s, the compressor industry began to embrace the application of dry gas seal technology to the critical function of shaft sealing. The seal consists of a rotating disc running very close to a stationary ring. The rotating disc face contains special grooves that generate an axial ("lift") force during rotation. The stationary ring is backed by a quantity of coil springs that force it tightly against the rotating disc when the compressor is at rest. The lift force compresses the coil springs slightly, resulting in the very small running clearance between the two faces. This small clearance effectively limits gas leakage from the compressor seals. The small amount of gas leakage exits the compressor through auxiliary seal piping, where it is then either sent to a flare system or to some other recovery system. Usually the two compressor seals (inlet and discharge ends of the compressor) are subjected to the gas suction pressure. A thrust balance line (see further discussion in the section on bearings) subjects the discharge end dry gas seal to inlet pressure, thereby avoiding the need to seal the higher discharge pressure.
Dry gas seals require clean and dry gas for reliable operation. Seal gas is normally taken from the compressor discharge and then cooled and filtered as part of an external seal gas processing system. A seal reference pressure is measured just inboard of the dry gas seal, and a pressure regulating valve supplies the seal gas to the sealing faces at a pressure slightly above the reference pressure. This system ensures that the seals are not exposed to untreated process gas containing liquids or particulate matter that could damage the seals. Although dry gas seals are relatively expensive, their auxiliary system is less complex, physically smaller, and less expensive than the auxiliary system required by the predecessor liquid film design.
Liquified Film Seals. Liquid film seals can be of the bushing type or mechanical contact type. The bushing type is a very simple and rugged design that incorporates two adjacent seal rings (bushings) at each end of the compressor. A sealing fluid is introduced into the space between the seal rings at a pressure slightly above the process gas pressure inboard from the inner ring. The pressure differential across the inner ring is assured by an overhead seal oil tank pressurized by compressor suction pressure. The elevation above the compressor of the oil level in the tank assures the required seal ring pressure differential. For almost all centrifugal compressors equipped with liquid film seals, the sealing fluid is the same light turbine oil as that used to lubricate the bearings. Therefore, the auxiliary seal oil system needed to supply the seal oil can be combined with (or separate from) the auxiliary lube oil system.
The inner seal ring is designed to minimize oil leakage into the process side. Inner seal leakage (also called sour oil leakage) mixes with the process gas and is drained from the compressor as an oil/gas mixture. A labyrinth seal inboard of the sour oil drain is installed to prevent seal oil from contaminating the process gas. The oil/gas mixture drains into a degassing tank where the gas is removed so that the oil can be sent to a seal oil reservoir for reuse.
The outer seal ring breaks serve to inhibit flow as pressure is reduced to an atmospheric drain. This drain is common with the bearing oil drain when a combined oil system is used. When the lube and seal oil systems are separate, a buffered labyrinth seal is placed between the lube and seal oil drains to ensure that there is no oil carryover from one system to the other.
Mechanical contact seals employ a stationary carbon ring against a rotating seal face. Oil is also used as the sealing medium in mechanical contact seals. The sealing oil is introduced by a pressure-regulating valve that is maintained at 25 to 40 psi above the seal reference pressure. One advantage of mechanical contact seals is a significantly reduced sour oil leakage compared with the bushing design. Unlike oil film seals, mechanical contact seals can be supplied with a feature that allows the compressor to maintain case pressure during shutdown without requiring that the auxiliary seal oil system be operating. Mechanical contact seals, however, are relatively complex.
Bearings. The radial bearings most often used in centrifugal compressors are the tilting pad type and are continuously lubricated with light turbine oil. Before tilting pad designs, sleeve type bearings were commonly used. The tilting pad bearing design provides rotor-dynamic characteristics that help assure smooth and reliable mechanical operation. Radial bearings are sized to be large enough to support the rotor weight, yet small enough to operate at sufficiently low peripheral speeds required to limit operating temperature to acceptable levels. Some centrifugal compressors are equipped with magnetic radial bearings. These bearings suspend the rotor by electromagnetic force to center the rotor within an air gap at the bearing. Use of magnetic bearings eliminates the need for an auxiliary lube oil system; however, the magnetic bearing control system also requires cooling.
The pressure rise in each of the stages of a centrifugal compressor creates an axial thrust force that acts toward the inlet end of the compressor. Depending on the overall pressure rise in the compressor, these thrust forces can be significant. An inline configuration employs a thrust balance drum (balance piston) to generate a thrust force to oppose ("balance") the sum of the impeller thrust forces. Located at the discharge end of the compressor, the balance piston is a simple disc-shaped element installed on the compressor shaft and equipped with a seal around its outer diameter. The space adjacent to the outboard face of the balance piston is subjected to compressor suction pressure as created by an auxiliary thrust balance line. The inboard surface of the balance piston is subjected to what is essentially the compressor discharge pressure. The resulting pressure differential across the balance piston creates an axial force toward the discharge end, thus opposing the impeller thrust forces. Proper selection of the balance piston diameter results in small net thrust force and allows use of a reasonably small thrust bearing to absorb the residual thrust forces and maintain proper rotor axial positioning.
Like the radial bearing, the thrust bearing is usually a tilting-pad design lubricated with light turbine oil. Some thrust bearing designs employ a system of leveling blocks behind each tilting pad to ensure uniform load distribution. As with the radial bearings, magnetic thrust bearings also are available.
Compressor PerformanceThe performance characteristic of a centrifugal compressor is graphically presented in the form of a family of curves that collectively are known as a performance map or operating envelope. An example of a performance map is given in Fig. 7.13. In the example map, the inlet volume flow is plotted along the x-axis and the head or pressure ratio is plotted along the y-axis. The "approximate surge limit" depicted at the left side of the map defines the minimum flow necessary to avoid a potentially damaging surge condition (see Sec. 7.4.3). At the extreme right portion of the map is the "stonewall" (choke) limit (see Sec. 7.4.4). Each of the family of curves from the surge limit to the stonewall represents the flow vs. pressure characteristic at a given compressor speed. The slope of the curve varies with the number of stages, becoming steeper with an increasing number of stages. The elliptical curves (dashed lines) denote compressor efficiency. The design point is at 100% speed, and the compressor components are selected so that the design point has a safe margin from surge and stonewall, as well as optimum efficiency.
The surge limit defines the flow at which, for a given speed, the operation of the compressor becomes unstable. At flow rates below the surge limit the characteristic curve actually droops toward zero flow after having reached its maximum point at the surge limit. Because operation below the surge limit is unstable, this portion of the curve is not shown in Fig. 7.13. When the flow is reduced below the surge limit, the pressure at the discharge of the compressor exceeds the pressure-making capability of the compressor, causing a momentary reversal of flow. When this flow reversal occurs, the pressure of the discharge system is reduced, allowing the compressor to resume delivering flow until the discharge pressure again increases, and the surge cycle repeats. Surging usually creates a clearly audible noise. Prolonged operation in this unstable mode can cause serious mechanical damage to the compressor. When operating in a surge condition, the compressor discharge temperature increases significantly and the compressor experiences erratic and severe vibration levels that can cause mechanical damage—particularly to the internal seals.
A compressor can be brought out of surge in a number of ways. The most obvious is to increase flow (see the subsection on Antisurge Valves in Sec. 7.4.5). Decreasing discharge pressure and/or increasing speed are other ways to move out of a surge condition.
Compressor manufacturers usually perform an aerodynamic performance test before delivering the compressor. Determination of the compressor’s actual surge limit is a very important aspect of the manufacturer’s shop testing program.
The stonewall limit of the performance curve defines the flow at which the gas velocity at one of the impellers approaches the velocity of sound for the gas at the conditions within the compressor where this sonic condition is first encountered. At the stonewall (or choke) flow the pressure vs. volume curve becomes essentially vertical, and it is not possible to develop head or pressure at any greater flow. When the required operating flow exceeds the stonewall limit, the only remedy is to reconfigure the compressor with impellers (and matched stationary hardware) designed for larger flow rates.
Process InstallationA centrifugal compressor may be configured with one of a variety of process connection arrangements. For grade-mounted installations, the process connections are most often positioned on the upper half of the casing with the process piping connected from above the compressor. In some installations, horizontal (side) connections are employed. The horizontal connection arrangement is frequently used in booster compressors for gas transmission. A radial split case design (see the subsection on Case in Sec. 7.4.1) is preferred for these two arrangements. Another arrangement is the mezzanine-mounted configuration. With this type of installation, the compressor connections are on the lower half of the casing, and the process piping is connected from underneath the compressor. If the operating pressures are sufficiently low, an axial split case design is appropriate (see the subsection on cases). To achieve optimum performance, it is necessary to install the compressor with a sufficiently long straight section of inlet piping upstream of the compressor inlet flange. Most compressor manufacturers require that the length of this straight section be at least two times the inlet flange diameter.
The compressor must be well integrated into the entire process so that start-up, operation, and shutdown can be safely controlled. The next section provides a description of control concepts and the process equipment required. (Refer to Fig. 7.14 for a typical process flow diagram.) The individual elements of the control and safety system illustrated in Fig. 7.13 are discussed in this section.
Flow Control. Most compression processes require the compressor to deliver a relatively constant discharge pressure over a range of capacities. However, the centrifugal compressor characteristic curve from Fig. 7.13 shows that the pressure ratio, in fact, varies continuously with flow. The process can control either suction or discharge pressure. If one is fixed, the other will vary as dictated by the compressor characteristic curve. The three methods of maintaining a constant discharge pressure for varying capacity are discussed next.
Speed Control. Centrifugal compressor drivers are either of the fixed or variable speed type. Most steam or gas turbines and those electric motors equipped with a variable frequency drive system are all available as variable speed drivers. For a given discharge pressure, compressor capacity may be increased by merely increasing the speed of rotation. Conversely, capacity may be decreased by reducing compressor speed. Capacity control by speed variation is the most effective way to maximize the operating flexibility of a centrifugal compressor.
Suction Throttle Valves. A fixed-speed motor is often the least expensive driver for a centrifugal compressor. When designing a centrifugal compressor driven by a fixed-speed motor, it is necessary to establish the speed based on the operating condition that requires the largest capacity for the required discharge pressure. When operating at lower capacities, the compressor inherently delivers a greater discharge pressure (for a given process suction pressure) than desired. The solution to this problem is to install a throttle valve at the inlet of the compressor. Suction pressure reduction by throttling increases the pressure ratio required to deliver a given discharge pressure. The economic trade-off for this method of capacity control is additional compressor power vs. additional capital expenditure for a variable speed driver.
Variable Inlet Guide Vanes. As discussed in the subsection on Stationary Components in Sec. 7.4.1, the compressor performance characteristic curve can be adjusted by changing the direction of the flow of gas into the impeller. When a system of variable inlet guide vanes is employed, it is possible to adjust the inlet guide vane angles to maintain a desired discharge pressure over a range of capacity. Practical design limitations make it difficult to install variable vanes at all stages other than the first stage. For single stage compressors, this method of control is sometimes quite effective. However, for multistage compressors, the range of control is less effective and becomes even less so with increasing numbers of stages.
Antisurge Valves. As discussed in Sec. 7.4.3, avoiding surge is extremely important. The installation of an antisurge (recycle) valve and its associated control devices is required. The antisurge valve is located in a recycle line connecting the compressor discharge to the inlet. For multisection compressors, it is good practice to install a separate recycle line with an antisurge valve for each of the compressor sections. Instrumentation is required to measure the flow to each section, and a surge controller must initiate the opening of the recycle valve when reduced capacity approaches the surge limit. The capacity at which the antisurge valve begins to open is usually set to be about 10% larger than the actual surge limit.
For variable speed compressors, the surge limit curve (see Fig. 7.13) defines the relationship between the surge limit and the operating speed. The logic programmed into the antisurge controller maintains the 10% safety margin, regardless of speed. This can be depicted graphically by a line parallel to the surge limit curve and is typically called the "control line."
The gas recycled through the antisurge valve also must be cooled because its source is the compressor discharge. If uncooled, the suction temperature will increase by mixing the hotter recycled gas with the main process inlet gas.
Flare Valve. The flare valve protects upstream equipment from overpressurization that may occur because of a flow increase and prevents overloading of the compressor driver. For a constant discharge pressure system, an increase in flow results in an increase in suction pressure. Higher suction pressures deliver more mass flow and, therefore, increase the power required to operate the compressor. The presence of a suction throttle valve also can contribute to an increase in pressure upstream from the compressor. Thus, flare valves are particularly important in installations with inlet throttling.
Shutdown Valve. Shutdown valves are installed at both the suction and discharge to enable the compressor to be isolated during shutdown periods. To satisfy safety concerns, the shutdown valves should be located outside any building or enclosure. Automatic control of the shutdown valves is usually employed.
Blowdown Valve. At shutdown, after the shutdown valves have isolated the compressor, the pressure in the compressor settles out to a level determined by a variety of factors. A blowdown valve is used to depressurize the compressor upon shutdown. Automatic control of the blowdown valve is recommended for high-risk locations and for compressors that are fitted with liquid film seals. When liquid film seals are employed, the compressor must be depressurized before the overhead seal tanks have been drained.
Discharge Check Valve. Placement of a check valve at the discharge of each section of compression can minimize or eliminate backflow through the compressor. Should backflow occur, it is possible for the compressor to experience potentially damaging reverse rotation. The presence of discharge check valves also provides the benefit of isolating each of the antisurge recycle loops (see the subsection on Antisurge Valves above).
Relief Valve. The compressor develops its maximum pressure ratio when operating at both its maximum continuous speed and the surge control capacity. If the suction pressure increases for any reason, the discharge pressure correspondingly increases to the value given by the performance map for the speed and capacity in question. A pressure relief valve is installed to protect against overpressurization of downstream equipment by the compressor.
Purge Valve. Before startup, it is necessary to purge air from the compressor and piping system. A purge valve is installed as a bypass to the suction shutdown valve for this purpose. Purging must be done with a low flow rate to prevent the purge gas from initiating compressor rotation. For this reason, the purge valve is small.
Discharge Coolers. A discharge cooler (after cooler) is required if the temperature of the gas at the compressor discharge exceeds that required for the next step in the process.
Suction Scrubbers. Erosion of compressor components can be caused by ingestion of excessive liquid. To prevent erosion damage, suction scrubbers are installed to remove liquids that condense in the gas suction line because of cooling or that result from an upstream-process upset resulting in liquid carryover to the gas suction line.
Vent Valve. A manual vent valve is installed between the compressor discharge and the discharge check valve to allow the compressor to be isolated from the vent header for maintenance. Once the compressor is shut down and blown down to the vent header, the blowdown valve can be closed and the vent valve opened. If the blowdown valve were kept open, there is a possibility that gas in the vent header would flow into the compressor system, endangering the maintenance operation.
Safety and Monitoring Devices. Centrifugal compressors are equipped with instrumentation to monitor mechanical health. Vibration monitoring is accomplished by eddy current probes installed at each of the compressor bearings. Vibration amplitude is measured at each radial bearing, and the axial position of the rotor is measured at the thrust disc or shaft end. The trend of radial vibration amplitude provides insight into the condition of the compressor regarding rotor balance and alignment. When a problem arises, the vibration frequency spectrum can also be analyzed to provide useful diagnostic information. The axial position probe monitors the state of thrust bearing wear. Each of the bearings is also fitted with temperature-sensing devices. By trending the thrust bearing pad temperatures, it is possible to discern the condition of the internal seals because changes in seal condition affect thrust loads and, therefore, bearing temperature. Alarm and shutdown settings for high bearing vibration and temperature are established in the compressor control system.
External to the compressor are numerous other alarm and shutdown safeguards. As a minimum, low lube oil pressure, low seal gas pressure differential, overspeed, high discharge gas temperature, high and low suction and discharge pressures, and high liquid level in the suction scrubber are monitored and will initiate a shutdown when necessary.
When properly designed, operated, and protected, centrifugal compressors are capable of long sustained runs with very little maintenance. The components most prone to wear are the bearing pads and internal labyrinth seals. Fouling of the internal surfaces can occur in some services causing a degradation of performance. The vibration and bearing temperature monitoring instrumentation, described in the subsection on safety and monitoring devices, provides valuable information to the operator about the probable condition of the compressor bearings. Excessive wear of the internal labyrinth seals can occur when the compressor experiences high vibration excursions from process upsets or operation in a surge condition. Worn internal seals cause a degradation of compressor performance similar to that caused by fouling.
Unless there is an identified problem with the compressor, maintenance is generally carried out during planned turnarounds. As a minimum, the easily accessible compressor bearings and shaft endseals are inspected and replaced with spares, if necessary. Complete disassembly is required to inspect the compressor internals. The auxiliary lube and seal systems require maintenance of miscellaneous items such as seal gas filters, lube oil pump seals, oil filters, etc.
Major ComponentsReciprocating compressors are available in a variety of designs and arrangements. Major components in a typical reciprocating compressor are shown in Fig. 7.15.
Frame. The frame is a heavy, rugged housing containing all the rotating parts and on which the cylinder and crosshead guide is mounted. Compressor manufacturers rate frames for a maximum continuous horsepower and frame load (see the subsection on Rod Load in Sec. 7.5.2).
Separable compressors are usually arranged in a balanced-opposed configuration characterized by an adjacent pair of crank throws that are 180 degrees out of phase and separated by only a crank web. The cranks are arranged so that the motion of each piston is balanced by the motion of an opposing piston.
Integral compressors typically have compressor and engine-power cylinders mounted on the same frame and are driven by the same crankshaft. Cylinders in integral compressors are usually arranged on only one side of the frame (i.e., not balanced-opposed).
Cylinder. The cylinder is a pressure vessel that contains the gas in the compression cycle. Single-acting cylinders compress gas in only one direction of piston travel. They can be either head end or crank end. Double-acting cylinders compress gas in both directions of piston travel (see Fig. 7.16). Most reciprocating compressors use double-acting cylinders.
Choice of cylinder material is determined by operating pressure. Cast iron is normally used for pressures up to 1,000 psi. Nodular iron is used for pressures up to 1,500 psi. Cast steel is usually used for pressures between 1,500 and 2,500 psi. Forged steel is selected for cylinder operating pressures greater than 2,500 psi.
A cylinder’s maximum allowable working pressure (MAWP) should be rated at least 10% greater than the design discharge pressure (minimum 25 psi). The additional pressure rating allows a high-pressure safety sensor (PSH) to be set above the design discharge pressure, and for a relief valve (PSV) to be set at a pressure above the PSH.
Wear compatibility of the rubbing parts (piston rings and cylinder bore, piston rod and seal rings, etc.) is also a criterion for selecting materials. Cylinders experience wear at the point of contact with the piston rings. In horizontal arrangements, cylinder wear is greatest at the bottom because of piston weight. Thermoplastic rings and rider bands are used in most reciprocating compressors to reduce such wear.
Cylinders are frequently supplied with liners to reduce reconditioning costs. Liners are pressed or shrunk in place to ensure that they do not slip. Replacement of a cylinder liner is much less expensive than replacing an entire cylinder. In addition, performance can be adjusted to new requirements by changing the inside diameter of the liner. However, cylinder liners increase the clearance between the valve and the piston, diminish the effectiveness of jacket cooling, and decrease compressor capacity from a given diameter.
Distance Piece. The distance piece provides separation between the compressor cylinder and the compressor frame. Fig. 7.17 illustrates API Standard 11P and API Standard 618 distance pieces. Distance pieces can be contained in either a single- or double-compartment arrangement. In the single-compartment design, the space between the cylinder packing and the diaphragm is lengthened so that no part of the rod enters both the crankcase and cylinder stuffing box. Oil migrates between the cylinder and the crankcase. If oil contamination is a concern, an oil slinger can be provided to prevent packing lube oil from entering the compressor frame. For toxic service, a two-compartment design may be used. No part of the rod enters both the crankcase and the compartment adjacent to the gas cylinder.
The packing case should be vented to the first stage suction or a vent gas system. Distance pieces contain a vent to evacuate additional leaking process gas from the packing. The diaphragm and packing are designed to keep gas from entering the crankcase. Effective venting is required to ensure that the process gas does not contaminate crankcase oil.
Each compressor should be equipped with a separate vent and drain system for distance pieces and packing. Distance piece and packing vents should be piped into an open vent system that terminates outside and above the compressor enclosure at least 25 ft horizontally from the engine exhaust. The distance piece drain should be piped into a separate sump that can be manually drained. The sump should be vented outside and above the compressor enclosure. Lube oil from the sump can be mixed with crude oil or, under certain circumstances, must be transported for disposal or recycling.
Crankshaft. The crankshaft rotates around the frame axis and drives the connecting rod, piston rod, and piston (see Fig. 7.18).
Connecting Rod. The connecting rod connects the crankshaft to the crosshead pin.
Crosshead. The crosshead converts the rotating motion of the connecting rod to a linear, oscillating motion that drives the piston.
Piston Rod. The piston rod connects the crosshead to the piston.
Piston. The piston is located at the end of the piston rod and acts as the movable barrier in the compressor cylinder. Selection of material is based on strength, weight, and compatibility with the gas being compressed. The piston is usually made of a lightweight material such as aluminum or from cast iron or steel with a hollow center for weight reduction. Thermoplastic wear (or rider) bands often are fitted to pistons to increase ring life and reduce the risk of piston-to-cylinder contact. Cast iron usually provides a satisfactorily low friction characteristic, eliminating the need for separate wear bands.
Wear bands distribute the weight of the piston along the bottom of the cylinder or liner wall. Piston rings minimize the leakage of gas between the piston and the cylinder or liner bore. Piston rings are made of a softer material than the cylinder or liner wall and are replaced at regular maintenance intervals. As the piston passes the lubricator feed hole in the cylinder wall, the piston ring gathers oil and distributes it over the length of the stroke.
Bearings. Bearings located throughout the compressor frame assure proper radial and axial positioning of compressor components. Main bearings are fitted in the frame to properly position the crankshaft. Crank pin bearings are located between the crankshaft and each connecting rod. Wrist pin bearings are located between each connecting rod and crosshead pin. Crosshead bearings are located at the top and bottom of each crosshead.
Most of the bearings in reciprocating compressors are hydrodynamic lubricated bearings. Pressurized oil is supplied to each bearing through oil supply grooves on the bearing surface. The grooves are sized to ensure adequate oil flow to prevent overheating.
Piston rod packing provides the dynamic seal between the cylinder and the piston rod. The packing consists of a series of non-metallic rings mounted in a case and bolted to the cylinder. The packing rings work in pairs and are designed for automatic wear compensation. Because each pair of rings accommodates a limited amount of pressure differential, multiple pairs are required depending on the pressure required by the application. To safely vent gas leakage through the packing, the vent port is usually located between the two outer ring assemblies (see the subsection on Distance Piece in Sec. 7.5.1).
Auxiliary connections to the packing may be required for cooling water, lubricating oil, nitrogen purging, venting, and temperature measurement. Lubrication must be finely filtered to avoid damage that would result from small particulate matter entering the case. The lubricating oil is normally injected into the second ring assembly, with pressure moving the oil along the shaft.
Compressor Valves. The essential function of compressor valves is to permit gas flow in the desired direction and to block all flow in the opposite (undesired) direction. Each operating end of a compressor cylinder must have two sets of valves. The set of inlet (suction) valves admits gas into the cylinder. The set of discharge valves is used to evacuate compressed gas from the cylinder. The compressor manufacturer normally specifies valve type and size.
Plate valves constructed from rings connected by webs into a single plate are a common valve type. Depending on the sealing plate material, plate valves are capable of handling pressures as high as 15,000 psi, differential pressures to 10,000 psi, speeds to 2,000 rpm, and temperatures to 500°F. Plate valves do not perform well in the presence of liquids.
Concentric ring valves are capable of handling pressures to 15,000 psi, differential pressures to 10,000 psi, speeds to 2,000 rpm, and temperatures to 500°F. Advantages of concentric ring valves include moderate parts cost, low repair cost, and the ability to handle liquids better than plate valves.
Poppet-style valves generally provide performance that is superior to both plate and concentric ring valves. The poppet style uses separate, round poppets to seat against holes in the valve seat. This type of valve offers high lift and low pressure drop, resulting in higher fuel efficiency. Poppet valves are widely used in pipeline, gas conditioning, and processing facilities. Metallic poppets work well at pressures to 3,000 psi, differential pressures to 1,400 psi, speeds to 450 rpm, and temperatures to 500°F. Thermoplastic poppets can be applied to applications with pressures to 3,000 psi, differential pressures to 1,500 psi, speeds to 720 rpm, and temperatures to 400°F.
Most compressors have valves mounted in the cylinders. A relatively new design concept places the valves in the piston. The valve-in-piston design (Fig. 7.19) operates with low valve velocities and provides longer life cycles and reduced maintenance time.
Compressor capacity and horsepower are affected by piston displacement and cylinder clearance. The flow capacity of a given cylinder is a function of piston displacement and volumetric efficiency. Volumetric efficiency is a function of cylinder clearance, compression ratio and the properties of the gas being compressed. Compressor capacity may be calculated with any of the following three equations.
|qa||=||inlet capacity of the cylinder at actual inlet conditions, Acf/min,|
|PD||=||piston displacement, Acf/min,|
|qg||=||inlet capacity of the cylinder, scf/min,|
|Qg||=||inlet capacity of the cylinder, MMscf/D.|
Piston Displacement. Piston displacement is defined as the actual volume of the cylinder swept by the piston per unit of time. Displacement is commonly expressed in actual cubic feet per minute (Acf/min). Calculation of the piston displacement is a straightforward procedure that depends on the type of compressor configuration. Single-acting cylinders can have either head-end or crank-end displacement. Eqs. 7.19 and 7.20 are used to calculate displacement of single-acting cylinders.
Single-Acting Cylinder (Head-End Displacement)
|PD||=||piston displacement, Acf/min,|
|N||=||compressor speed, rpm,|
|dc||=||cylinder diameter, in.|
|PD||=||piston displacement, Acf/min,|
|N||=||compressor speed, rpm,|
|dc||=||cylinder diameter, in.,|
|dr||=||rod diameter, in.|
|PD||=||piston displacement, Acf/min,|
|N||=||compressor speed, rpm,|
|dc||=||cylinder diameter, in.,|
|dr||=||rod diameter, in.|
The methods used to change piston displacement include changing compressor speed, removing or deactivating suction valves in a double-acting cylinder, and changing the diameter of the cylinder liner and piston.
Unloading one end can significantly reduce the capacity of a double-acting cylinder. The best method to unload a cylinder is to deactivate or remove the suction valves from one end to prevent that end from compressing gas. Depending on the frequency of unloading and the molecular weight of the gas, a port or plug unloader is the next best method of unloading a cylinder. A doughnut replaces one suction valve of three or more valves per corner, and only one unloading device is required per cylinder end. With concentric ring-type valves, it is possible to place a plug unloader in the center of a suction valve for unloading. Depending on the molecular weight of the gas, both port and plug unloaders reduce BHP/MMscf/D and significantly improve the reliability of the unloading system.
If the suction valve is held open with finger depressors during the compression stroke, gas will flow through the open valve back into the suction gas passage, and no gas will be discharged from the end of the cylinder containing the unloaded suction valve. Deactivation of the valves may be performed manually while the compressor is shut down or by using a valve unloader or lifter while the compressor is in operation. Control of the valve unloader can be manual or automatic by a diaphragm that unloads the compressor using a suction pressure sensor. Diaphragm actuators are more reliable than manual lifters or unloaders.
Unloading both ends of the same cylinder may cause the cylinder to overheat; thus, it is best to unload only one end of a double-acting compressor cylinder. In most cases, it is preferable to remove the suction valve when unloading the head end of a cylinder to assure load reversal in the rods. (See the subsection on Rod Load in Sec. 7.5.2.)
Clearance Volume. Clearance volume is the space remaining in the compressor cylinder at the end of the stroke. Clearance is made up of spaces in valve recesses and the space between the piston and cylinder end. At the completion of each compression stroke, the compressed gas trapped in the clearance space expands against the piston and adds to the force of the return stroke. Fig. 7.20 is a pressure vs. volume (P-V) diagram illustrating the effect of clearance.
Expansion of the gas trapped in the clearance space occurs before the suction valve opens to admit new gas into the cylinder. As a result, part of the piston displacement occurs before the suction valve opens. The compression process in reciprocating compressors is nearly isentropic, so the energy required to compress the gas in the clearance space is recovered when the gas expands at the end of the compression stroke. For this reason, changes in the clearance space do not affect the compressor power.
Clearance volume is expressed as a percentage of piston-swept volume using one of the following configuration-dependent equations.
Single-Acting Cylinder (Head-End Clearance).
|%C||=||cylinder clearance, %,|
|CHE||=||head-end clearance, in.3,|
|dc||=||cylinder inside diameter, in.,|
|S||=||stroke length, in.|
|%C||=||cylinder clearance, %,|
|CCE||=||crank-end clearance, in.3,|
|dc||=||cylinder inside diameter, in.,|
|dr||=||rod diameter, in.,|
|S||=||stroke length, in.|
|%C||=||cylinder clearance, %,|
|CHE||=||head-end clearance, in.3,|
|CCE||=||crank-end clearance, in.3,|
|dc||=||cylinder inside diameter, in.,|
|dr||=||rod diameter, in.,|
|S||=||stroke length, in.|
Clearance can be added to a cylinder as fixed-volume clearance pockets, variable clearance pockets, or split-valve yokes.
A fixed-volume clearance pocket is normally a volume bottle permanently attached to the cylinder. Fixed volume also can be added by a side-passage clearance plug consisting of a flange with a variable length plug inserted into a passage built into the side of the cylinder. A fixed-volume clearance pocket may be continuously open or may be controlled to be either open or closed. Control can be by manual hand wheel or automatic actuator. An actuator control allows the clearance pocket to be opened or closed from outside the cylinder while the compressor is in operation.
Variable clearance pockets allow a variable amount of clearance to be added to the cylinder and can be attached to either the head end or crank end of the cylinder. Most often, variable clearance pockets are attached to the head end, as shown in Fig. 7.21.
Excessive clearance in a compressor cylinder can cause slamming of the discharge valves. If too much clearance is present, no gas will be discharged. Rapid overheating can occur because no cool suction gas enters the cylinder.
Volumetric Efficiency. Volumetric efficiency is the ratio of the actual volume of gas (Acf/min) drawn into the cylinder to the piston displacement (cf/min). This ratio is less than unity because of three fundamental effects. First, the gas is heated during admission to the cylinder. Second, there is leakage past valves and piston rings. And third, there is re-expansion of the gas trapped in the clearance volume from the previous stroke. Of these three, re-expansion has, by far, the greatest effect on volumetric efficiency.
Compressor manufacturers have not reached consensus on an appropriate calculation method because measurement of these effects is extremely difficult. Recognizing this, the next approximate equation may be used to estimate volumetric efficiency.
|C||=||cylinder clearance, % of piston-swept volume,|
|Zs||=||inlet compressibility factor,|
|Zd||=||discharge compressibility factor,|
|dr||=||rod diameter, in.,|
|k||=||ratio of specific heats, Cp/Cv,|
|L||=||slippage of gas past piston rings, % (1% for high-speed separable, 5% for nonlubricated compressors and 4% for propane service),|
|96||=||allowance for losses because of pressure drop in valves.|
Rod Load. Rod loads consist of gas loads caused by pressure and inertia loads that result from acceleration and deceleration of the piston, piston rod, crosshead, and approximately one-third of the connecting rod weight. Manufacturers specify a maximum rod load to protect the compressor because overloading the rods can severely damage the compressor. Loads must be evaluated for normal operating conditions and also during upset conditions. Rod loading must be reviewed at minimum suction pressure and relief valve pressure to assure an adequate safety margin.
Rod load reversals must be of sufficient magnitude to provide lubrication to the crosshead pin bushing. The bushings are lubricated by the pumping action of the opening and closing of bearing clearance that occurs when the rod load reverses from tension to compression. Operation without rod reversals also can severely damage the compressor.
Rod loads for the various compressor configurations are calculated with the following equations.
Single-Acting Cylinder (Head End).
Single-Acting Cylinder (Crank End)
|RLc||=||rod load in compression, lbf,|
|RLt||=||rod load in tension, lbf,|
|ap||=||cross-section area of piston, in.2,|
|ar||=||cross-section area of rod, in.2,|
|Pd||=||discharge pressure, psia,|
|Ps||=||suction pressure, psia,|
|Pu||=||pressure in unloaded end, psia.|
Other Performance Factors. Suction Pressure. At constant discharge pressure with compression ratios greater than 2.0, the compression ratio decreases as the suction pressure increases. A decrease in compression ratio reduces the power requirement per unit of flow. The capacity of the cylinder, however, increases with suction pressure at a faster rate, resulting in an overall increase in power. To avoid overloading the driver, additional clearance must be added to reduce cylinder capacity.
Suction Temperature. Cylinder capacity is inversely proportional to absolute suction temperature. As temperature decreases, more standard cubic feet fill the cylinder. Thus, a 10°F reduction in suction temperature increases compressor mass flow by almost 2%. Precooling the gas can be an effective way to increase cylinder capacity.
Discharge Pressure. Changes in discharge pressure have little effect on cylinder capacity. Volumetric efficiency varies slightly with compression ratio, and the required power is directly proportional to the change in compression ratio.
Ratio of Specific Heats (k). An increase in k value produces an increase in the volumetric efficiency as defined by Eq. 7.25. Thus, a given compressor cylinder has a higher actual capacity when compressing natural gas (k = 1.25), compared with its capacity when compressing propane (k = 1.15). The higher capacity, when compressing natural gas compared to propane, results in greater power consumption as well.
Speed. Cylinder capacity is directly proportional to compressor speed. It is common practice to adjust compressor speed (within reasonable limits) to maintain desired suction pressure. Reduction of driver speed lowers fuel consumption and operating costs.
Performance Maps. Performance maps can be developed for a specific compressor with base conditions held constant. Fig. 7.22 illustrates that as suction pressure increases, both inlet flow rate and power increase for constant discharge pressure and temperature. At very low ratios, the power may actually decrease with increasing suction pressure.
Process InstallationThe compressor is an integral part of a complete compression system. Fig. 7.23 is a typical process flow diagram for a reciprocating compressor installation.
Recycle Valve. Compressor suction pressure decreases as the flow rate decreases until the gas expands to satisfy the flow required by the cylinder. The increase in compression ratio caused by reduction in suction pressure results in an increase in discharge temperature. Thus, the recycle valve in the system must be set to prevent low suction pressure from creating excessive discharge temperature. In addition, rod load limits may dictate the minimum acceptable suction pressure for a compressor installation. Where possible, recycle valve should be downstream of gas coolers.
Blowdown Valve. The blowdown valve relieves trapped pressure when the compressor is shut down for maintenance. Valve control is typically automatic but is sometimes manual at some small, onshore compressor installations.
Suction Scrubber. Ingestion of liquids into the compressor through the inlet gas stream can cause damage to the compressor internals. For this reason, an adequately sized suction scrubber with provisions for draining is required. The scrubber may be part of the pulsation control when properly planned (see Sec. 7.5.4). If the inlet stream is near saturation, horizontally-oriented cylinders and bottom-connected discharge nozzles are recommended.
Relief Valves. Pressure relief valves set at a margin of 10% above the highest stage discharge pressure, or a minimum of 15 to 25 psi, provide static pressure protection for piping and coolers. Relief valve setting should never exceed the cylinder maximum allowable working pressure (see the subsection on cylinders in Sec. 7.5.1). Caution should be taken to ensure that all suction side gas piping, cylinders, and relief valves are rated for settle-out pressures in closed-loop refrigeration or low gas temperature services.
The flow of gas through a reciprocating compressor inherently produces pulsation because the suction and discharge valves are not open for the entire compression stroke. Pulsation damping is needed to create a more uniform flow through the compressor to assure uniform loading and to reduce piping vibration levels.
Pulsation Control Devices. If long, straight runs of piping of the same diameter as the compressor cylinder line connection can be provided, and the stage power is less than 150 hp, separate volume bottles or pulsation vessels may not be required. For most applications, volume bottles or pulsation vessels with internal baffles and/or choke tubes should be located as close to the cylinder as possible for optimum valve reliability. The addition of orifices at key locations in the piping can also reduce piping pulsations. Several different bottle-sizing formulae are available. Typical bottle sizes are five to ten times the cylinder swept volume.
Pulsation Design. Digital piping pulsation analysis is a relatively low-cost method to ensure that a piping system is designed to meet acceptable pulsation levels (typically 2 to 7% peak to peak). The piping system layout must identify locations and volumes of knockout drums, bottles, coolers, and relief valves. The analysis should include the first major vessel or volume upstream and downstream of the compressor. Double-acting and single-acting (if applicable) operating conditions should be analyzed.
Imbalance of the rotating elements in the compressor cause mechanical vibration. Counterweights on the crankshaft and arranging cylinders in pairs on both sides of the crankshaft (in plain view) can minimize but not eliminate imbalance forces. Thus, there will always be mechanical vibrators that must be taken into account in foundation design.
Piping Vibration. The compressor process gas piping must be properly designed and installed to avoid problems associated with excessive vibration. It is important that the natural frequency of all pipe spans is greater than the compressor pulsation frequency. The compressor pulsation frequency is calculated with Eq. 7.32.
|fp||=||compressor pulsation frequency, cycles/sec,|
|N||=||compressor speed, rpm,|
|=||1 for single-acting cylinder|
|=||2 (for double-acting cylinder).|
Piping should be securely tied using short pipe spans that are not uniform in length. Adequate pulsation damping helps prevent piping-related vibration problems.
Foundation Design. For large integral compressors, or for compressors installed on complex structures or soft soils, it is best to perform a dynamic design using the imbalance forces provided by the manufacturer. The details of performing such an analysis are beyond the scope of this handbook.
For high-speed compressors installed in areas with soils that can support a pickup truck, the following rules are useful.
- Weight of concrete foundation should be at least three to five times the equipment weight.
- Use soil bearing for a design that is less than 50% of that allowable for static conditions.
- It is generally better to increase length and/or width rather than depth to meet weight requirements.
- For rectangular block, at least 40% of height (but not less than 18 in.) should be embedded in undisturbed soil.
- Concrete should be poured into a "neat" excavation without formed side faces.
The heat of compression and friction between the piston rings and the cylinder add heat to the cylinder. Removing some of this heat is beneficial to the performance and reliability of the compressor in several ways. Cylinder cooling reduces losses in capacity and power caused by suction gas preheating. It also removes heat from the gas, thereby lowering the discharge temperature of the gas. Cylinder cooling also promotes better lubrication for longer life and reduced maintenance. When water is used as the cooling medium, uniform temperatures are maintained around the cylinder’s entire circumference, reducing chances for thermal distortion of the cylinder.
Care must be taken to avoid condensation that can result from excessive cooling. This can be assured by maintaining the cylinder jacket coolant temperature at least 10°F above the suction gas temperature.
Insufficient cooling can lead to reduced capacity and fouling of the cylinders. For this reason, it is recommended that the cylinder not be more than 30°F above the suction gas temperature.
Cooling Systems. Air Cooled. Air-cooled systems are used for small throughputs and low heat loads. Cooling fins provide a sufficient surface area to cool the cylinder.
Static. Static systems are sometimes used on small compressors to assist air-cooled systems. Cooling fluid functions as a static heat sink and acts more as a heat stabilizer than a cooling system. Some heat is transferred from the system by conduction to the atmosphere.
Thermosiphon. The driving force for a thermosiphon derives from the change in density of the cooling fluid from the hot to cold sections of the system. API Standard 618 permits use of this system when discharge gas temperatures are below 210°F or when temperature rise across the cylinder is less than 150°F.
Pressurized. Pressurized cooling systems are the most common. In locations where cooling water is not available, a self-contained, closed cooling fluid system may be used. The system consists of a circulating pump, surge tank, and a fan-cooled radiator or air-to-liquid heat exchanger. The radiator may have multiple sections—one for cylinder coolant, one for cooling lube oil, and one (or more) for cooling discharge gas. The cooling fluid is either water or a mixture of water and ethylene glycol. The crankshaft usually drives the circulating pump.
Frame Lubrication. The frame lubrication system delivers oil to the frame bearings, connecting rod bearings, and crosshead shoes. Some frame lubrication systems also supply oil to the packing and cylinders. For most reciprocating compressors, the lubrication system is integral with the frame.
Splash Lubrication. Splash lubrication systems distribute lubricating oil by the splashing of the crank through the lubricant surface in the pump. Dippers may be attached to the crankshaft to enhance the effect. Splash systems are used on small, horizontal, single-stage compressors with power demands up to 100 hp. Low initial cost and minimal operator attendance are the two main advantages of splash systems. The main disadvantages are that splash systems are limited to small frame sizes and that the oil cannot be filtered.
Pressurized Lubrication. The most common type of frame lubrication is the pressurized system. Oil enters passages drilled into the crankshaft and flows through the main shaft and crank pin bearings. A pressurized lubrication system consists of the components discussed next.
Main Oil Pump. The main oil pump is driven by the crankshaft or may be separately driven. It is typically sized to deliver 110% of the maximum anticipated flow rate. When speed reduction is used for capacity control, care must be taken to ensure that this pump provides adequate lubrication at the minimum operating speed.
Auxiliary Pump (Optional). An auxiliary pump is provided to back up the main pump. The auxiliary pump is usually driven by an electric motor and is designed to start automatically when oil supply pressure falls below a specified level.
Prelube Pump (Optional). A prelube pump supplies oil to the bearings before the compressor is started. This assures that the bearings are not dry at startup. Because this function is provided by the auxiliary pump, a prelube pump is required only when the system does not have an auxiliary pump.
Oil Cooler. The oil cooler ensures that the temperature of the oil supply to the bearings does not exceed the maximum value required to protect the bearings from wear. A typical maximum oil supply temperature is 120°F. Jacket cooling water in a shell and tube heat exchanger is often used to cool the lubricating oil.
Oil Filters. Oil filters protect the bearings by removing particulates from the lubricating oil. Some systems are equipped with dual, full-flow oil filters with transfer valves. Transfer valves allow switching from one filter to the other so that the filters can be cleaned without shutting down the compressor.
Overhead Tank. The overhead tank provides oil to the bearings if a pump fails. The oil from the overhead tank is gravity-fed to the bearings. The tank must be sized to provide oil until the compressor has completely shut down. The tank is usually equipped with a level indicator.
Piping. The components of the lubrication system are connected by piping. Cleanliness and corrosion resistance are important considerations. Galvanized piping should be avoided because of possible corrosion. Carbon steel piping should be pickled or mechanically cleaned and coated with a rust inhibitor. Stainless steel piping should be used downstream of the filters. The piping system should be designed to avoid any pockets in which dirt or debris could accumulate. Socket welded piping should be avoided for this reason. Before initial startup, the lube oil system should be flushed with lube oil at approximately 170°F. A 200-mesh screen should be added to the system, and flushing should continue until the mesh is clean. Safety instrumentation should include a crankcase low oil level switch, a low oil pressure shutdown switch, and a high oil temperature switch.
For compressors with integral engine drivers, it is recommended that the compressor and driver be lubricated with separate systems so that combustion gases from the engine do not contaminate the lube oil. In this case, the packing and cylinder lubrication is provided by the compressor lubricating system. For installations in very cold environments, immersion or in-line heaters and special lubricating oils should be considered.
Cylinder and Packing Lubrication. The quantity of oil required to lubricate the packing and cylinders is small when compared with the bearing oil requirements. While the quantity is small, the oil pressure necessary to supply oil at the packing and cylinders is high. A small plunger pump (force-feed lubricator) is used at each stage of compression. Divider blocks are used to distribute the flow of oil between the cylinders and the packing. The oil can be supplied from either the frame lubricating system or from an overhead tank. Compatibility of the oil with the process gas must be checked to protect against contamination.
In the preceding pages, we’ve reviewed the fundamental designs of compressors and compressor components used throughout the natural gas industry. Such equipment is an integral part of that industry, maximizing the ability of producers and distributors to access, recover, and deliver energy resources to meet demand anywhere in the world.
The diversity of design and application of compression equipment reflects not only the specialized needs of the industry, but also the adaptability of compression engineers to fulfill industry requirements. Because many compressors are designed to meet the rigid performance requirements of their owners ("engineered to order"), engineers have ample opportunity to apply their skills.
Although this chapter has been confined to basic principles, many dramatic design innovations have kept pace with industry demands for greater capacity, efficiency, safety, and environmentally acceptable performance under a remarkably broad range of operating conditions. In fact, it is a rarity to find conditions under which a properly engineered compressor cannot function as designed.
|ap||=||cross-sectional area of piston, in.2|
|ar||=||cross-sectional area of rod, in.2|
|C||=||cylinder clearance, %|
|CCE||=||crank-end clearance, in.3|
|CHE||=||head-end clearance, in.3|
|dc||=||cylinder diameter, in.|
|dr||=||rod diameter, in.|
|fp||=||compressor pulsation frequency, cycles/sec|
|His||=||isentropic head, ft-lbf/lbm|
|Hp||=||polytropic head, ft-lbf/lbm|
|k||=||ratio of specific heats|
|L||=||slippage of gas past piston rings, %|
|N||=||compressor speed, rpm|
|P1||=||specified absolute suction pressure, psia,|
|P2||=||specified absolute discharge pressure, psia|
|Pd||=||discharge pressure, psia|
|PD||=||piston displacement, Acf/min|
|Ps||=||absolute suction pressure, psia|
|Pu||=||pressure in unloaded end, psia|
|qa||=||inlet capacity of the cylinder at actual inlet conditions, Acf/min|
|qg||=||inlet capacity of the cylinder, scf/min|
|Qg||=||inlet capacity of the cylinder, MMscf/D|
|RLc||=||rod load in compression, lbf|
|RLt||=||rod load in tension, lbf|
|S||=||gas specific gravity (standard atmospheric air = 1.00)|
|T1||=||specified absolute suction temperature, °R|
|T2||=||estimated absolute discharge temperature, °R|
|Td||=||discharge temperature (actual or predicted), °R|
|Ts||=||suction temperature, °R|
|V1, V2||=||volumes at pressures 1 and 2|
|W||=||mass flow, lbm/min|
|zavg||=||average compressibility factor, dimensionless|
|zs||=||compressibility at inlet|
|Zd||=||discharge compressibility factor|
|Zs||=||inlet compressibility factor|
|ηp||=||assumed polytropic efficiency|
|k||=||ratio of specific heats, Cp/Cv|
API Standard 617, Centrifugal Compressors for Petroleum, Chemical, and Gas Service Industries, sixth edition. 1995. Washington, DC: API.
API Standard 618, Reciprocating Compressors for Petroleum, Chemical, and Gas Industry Services, fourth edition. 1995. Washington, DC: API.
Beyer, R.W. 1991. Reciprocating Compressor Performance and Sizing Fundamentals. New York City: Dresser Rand Company, Painted Post.
Brown, R.N. 1997. Compressors: Selection and Sizing. Houston, Texas: Gulf Publishing Co.
Engineering Data Book. 1987. Tulsa, Oklahoma: Gas Processors Suppliers Association.
Lapina, R.P. 1983. Manual for Estimating Centrifugal Compressor Performance. Houston, Texas: Gulf Publishing Co.
Loomis, A.W. 1980. Compressed Air and Gas Data. Ingersoll Rand: Woodcliff Lake, New Jersey.
Pichot, P. 1986. Compressor Application Engineering. Houston, Texas: Gulf Publishing Co.
SI Metric Conversion Factors
|ft||×||3.048*||E – 01||=||m|
|°F||(°F – 32)/1.8||=||°C|
|hp||×||7.460 43||E – 01||=||kW|
|in.||×||2.54*||E + 00||=||cm|
|in.2||×||6.451 6||E + 00||=||cm2|
|in.3||×||1.638 706||E + 01||=||cm3|
|lbf||×||4.448 222||E + 00||=||N|
|lbm||×||4.535 924||E – 01||=||kg|
|psi||×||6.894 757||E + 00||=||kPa|
Conversion factor is exact.