Improving Machinery Reliability 3 Episode 4 ppt

The compressor speed 1X excitation would excite the first torsional natural frequency at 1,005 rpm and the second natural frequency at 3,619 rpm on the gas turbine.. 112 Improving Mac

Trang 1

Machinery Reliability Audits arid Reviews 109 EQUIVALENT SYSTEM WODE GRAPES

WODE XO 1 31.80 HZ ieo7.e3 CPW

1

WODE XO 2 114.49 HZ 6869.44 CPY

WODE XQ 3 307.04 HZ 18476.23 CPW

WODZ WO 4 3 8 2 2 9 HZ 2 ~ 3 7 1 7 CPW

Figure 3-23 Torsional natural frequencies and mode shapes

An interference diagram for the turbine-driven compressor with a gear box is given

in Figure 3-24 The rated speeds are 5,670 rpm for the gas turbine and 10,762 rpm for the compressolr In this system, excitation at 1X and 2X the gas turbine and compressor speeds are possible The 1X excitation of gas turbine speed excites the first critical speed at 1,907 rpm; however, it will not reach the second natural frequency at 6,869 rpm since maximum speed would be less than 6,000 rpm The compressor speed (1X) excitation would excite the first torsional natural frequency at 1,005 rpm and the second natural frequency at 3,619 rpm on the gas turbine

Once the system has been modeled and the natural frequencies have been determined, the forcing functions should be applied The forcing functions represent dynamic torques applied at locations in the system that are likely to generate torque variations Identification of all possible sources of vibration is an important step in diagnosing an existing vibration problem or avoiding problems at the design stage The most likely sources of dynamic torques include reciprocating engines, gears, fans, turbines, compressors, pumps, motors (synchronous and induction), couplings, fluid interaction (pulsations), and load variations

Trang 2

110 Improving Machine9 Reliability

Figure 3-24 Interference diagram for gas turbine-compressor train

To evaluate the stresses at resonance, the torsional excitation must be applied to the system For systems with gear boxes, a torque modulation of 1%, zero-peak is a

representative torque value that has proven to be appropriate for most cases As a

rule of thumb, excitations at the higher orders for gears are inversely proportional to

the order numbers: the second order excitation is 0.5%, the third is 0.33%, etc

The torque excitation should be applied at the appropriate location and the torsion-

al stresses calculated on the resonant frequency and at the running speed An exam-

ple of the stress calculations of the second natural frequency resonance is given in

Table 3-5 It shows that a 1 % torque excitation on the bull gear would cause a maximum torsional stress of 4,179 psi p-p in shaft 9, which is the compressor shaft between the coupling and the first-stage impeller The dynamic torque modulation across the couplings is calculated for the applied input modulation For this mode, the maximum torsional vibrations occur across the compressor coupling and the dynamic torque modulation was 2,626 ft-lb

Variable Speed Drives Units that use a variable speed drive in conjunction with an

electric motor will have excitation torques at the running speed frequencies, and at several multiples, depending upon the design of the variable speed d r i ~ e 2 ~ Figure 3-25

gives an interference diagram for one such system It is difficult to remove all coinci-

Trang 3

Machinery Reliability Audits and Reviews 111

Table 3-5

Torsional Stress Calculations at the Second Torsional Natural Frequency

for 1 % Excitation at the Bull Gear Dynamic torques (lob zero-peak) applied at the bull gear

Maximum resultant torsional stresses at the 2nd torsional resonance 6,869.44 cpm

Stress psi P-P

Gear mesh I.334.81

Dynamic torque variation 1,393.16

859.02 723.70 582.18 564.75 434.23 422.73

41 1.05

3 14.94 215.28 98.88

2.00

I 50

3.00 3.00 3.00 3.00 1.50 1.50

I .50

I .50 1.50 1.50 1.50 1.50 1.50

I 50

7.83 51.71

35 I 68

65.27 4,004.43 4,179.49 1,288.52 1,085.55 873.27 847.12

65 1.35 634.09

6 16.57

472.41 322.93 148.32

Trang 4

112 Improving Machinery Reliability

dence of resonances with the excitation sources over a wide speed range; therefore, stress calculations must be made to evaluate the adequacy of the system response

Reciprocating Machinery For reciprocating units such as compressors, pumps, or

engines, the harmonic excitation torques must be calculated and applied at the appropriate shaft location to calculate the stresses.25

Allowable Torsional Stresses The calculated torsional stresses must be compared

to applicable criteria The allowable values given by Military Standard 164 are appropriate for most rotating equipment The allowable zero-peak endurance limit is equal to the ultimate tensile strength divided by 25 The MIL Spec uses this as a global derating factor rather than calculating on the basis of individual factors accounting for keyways, surface roughness effects, and the like When comparing calculated stresses to this value, the appropriate stress concentration factor and a safety factor must be used Generally, a safety factor of 2 is used for fatigue analysis When these factors are used, it can be shown that fairly low levels of torsional stress can cause failures, especially when it is observed that the standard keyway (USA

Standard ANSI B17.1) has a stress concentration factor of 3

We should not lose sight of the fact that process machinery is expected to live much longer than military hardware, and that our process machinery manufacturer has, perhaps:

1 no S-N curves

2 no intention of applying individual derating factors for either known stress rais-

3 no interest in determining coupling and misalignment-induced stress adders, etc.,

It would thus be reasonable to use a global derating value of 75, and, indeed, world-class turbomachinery manufacturers such as Elliott, Dresser-Rand, Mannes- mann-Demag, Sulzer, Mitsubishi, and surely many others, have both the experience and analytical capability to virtually guarantee unlimited life of turbomachinery shafts operating at relatively much higher mean torsional stresses A typical example would be steam turbine shafts with tensile strengths of 105,000 psi (ult.) and steady- state torsional stresses of 10,900 psi, where this latter stress simply uses the standard calculation formula

ers or unknown superimposed stresses

Nevertheless, 2,,,,/75 is not at all unreasonable for machines built by the “other” manufacturers A midwestern U.S plant uses rotary blowers direct-coupled to 200

hp, 1800 rpm motors The blowers came with 2% inch diameter shafts that had an ultimate strength in tension of 80,000 psi Although nominal stresses are thus only

2281 psi, the plant experienced many shaft failures with derating values as high as

80,000/2281 = 35 A typical torsional stress allowable thus becomes the ultimate tensile strength divided by 75

Trang 5

Transient Torsional Analysis

After the steady state analysis is made, a transient analysis should be made to evaluate the startup stresses and allowable number of startups for synchronous motor systems.22,26 The transient analysis refers to the conditions on startup, which are con- tinually changing because of the increasing torque and speed of the system When a synchronous motor starts, an excitation is imposed upon the torsional system due to field slippage As the motor increases in speed, the torsional excitation frequency decreases from twice power line frequency (typically 120 Hz) linearly with speed toward zero During this startup, the torsional system will be excited at several of its resonant frequencies if they are between 0 and 120 Hz, as shown in Figure 3-26 The response amplitudes and shaft stresses depend upon the resonant frequencies, the average and pulsating torque when the system passes through these resonant frequencies, the damping in the system, and the load torques The startup analyses can

be made for loaded or unloaded operation The transient response is also affected by the starting acceleration rate of the motor For slow motor startups, the system will stay at a resonant frequency for a longer period of time, allowing stresses to be amplified If acceleration is rapid, passing through the resonance quickly will mini- mize the amplitude increase at resonant frequencies

Synchronous motors develop a strong oscillating starting torque because of slippage between the rotor and stator fields Although this is only a transient excitation, the pulsating torque can be strong enough to exceed the torsional endurance limit of

Figure 3-26 Torsional stresses introduced into motor shaft during synchronous motor startup

Trang 6

the shaft For this reason the transient stresses must be calculated and compared to the endurance limit stress It is not necessary that the transient stresses be less than the endurance limit stress; however, the stresses must be sufficiently low to allow an acceptable number of starts If the transient stresses exceed the endurance limit, the cumulative fatigue concept is applied to the stresses in excess of the endurance limit stress to determine how many starts can be allowed for the system

Cumulative fatigue theory is used to estimate the number of cycles a certain stress level can be endured before shaft failure would occur This is based upon a plot of stress versus number of cycles (S-N curve), which defines the stress conditions at which a failure should occur The S-N curve is based upon actual tests of specimens

of a particular type of metal and defines the stress levels at which failures have occurred in these test specimens These S-N curves are available for most types of shafting materials Using the appropriate curve, the allowed number of cycles for a particular stress can be determined It is possible to calculate the number of total startups that can be made with the system before a shaft failure is predicted Since the stress levels vary both in amplitude and frequency, a more complex calculation must be made to determine the fraction of the total fatigue that has occurred The stress levels for each cycle are analyzed to determine the percentage of cumulative fatigue and the allowable number of startups can then be determined

The calculation of the allowable number of starts is strongly dependent upon the stress versus cycles to failure curve and whether torsional stresses higher than the torsional yield are allowed In the design stage it is preferable to design the system such that the introduced torsional stresses do not exceed yield This can usually be accomplished through appropriate coupling changes

Impeller and Blade Responses

A design audit should also include an assessment of the potential excitation of blade or impeller natural frequencies Several papers document such

The impeller and blade response analysis should include:

1 The blade and impeller natural frequencies

2 The mode shapes

3 Interference diagram indicating potential excitation mechanisms and the natural frequencies

The interference diagram, which gives the blade and impeller natural frequencies and the various potential excitation mechanisms, is the key to prevention of failures The resonances should be sufficiently removed from the major excitations in the operating speed range

In the design stage, it is possible to calculate the natural frequencies and mode shapes using finite element method [FEM] computer programs However, the accu- racy of predictions depends to a great extent upon the experience of the analyst and the complexity of the system

Since the blades and impellers will usually be available in advance of the rotor assembly, the most accurate natural frequency and mode shape data can be obtained

Trang 7

Machinery Reliability Audits arid Reviews 115

from shaker tests or by modal analysis methods The modal analysis technique uses a two-channel analyzer and an impact hammer and accelerometer to determine the nat-

ural frequencies and mode shapes For example, the natural frequencies and mode shapes of a centrifugal impeller were measured using modal analysis techniques (Figure 3-27) When these frequencies were compared to values determined from a shaker study, good correlation was obtained The mode shape for the two-diameter mode is given in Figure 3-28

An interference diagram for this impeller is given in Figure 3-29 Note that potential excitation mechanisms include vane passage frequency (1 5X) and two times vane passage frequency (3OX)

It is sometimes impossible to completely avoid all interferences over a wide speed range, since there are so many natural frequencies For most systems, in order for a failure to Q C C U ~ , several things usually occur together First, there must be a mechanical natural frequency Second, there must be a definite excitation frequency, such as vane passing or diffuser vane frequency Third, there must be some acoustical resonant frequency that amplifies the energy generated; and fourth, there must be the appropriate phase relationship that causes the pulsation to cause a shaking force on the impeller or blade The best way to avoid such problems is to avoid coincidence

of the resonances with the excitation mechanisms

Figure 3-27 Natural frequencies of centrifugal impeller

Trang 8

Figure 3-28 Two-diameter mode shape for centrifugal impeller at 1,360 Hz determined

by modal analysis tests

Trang 9

Muchinery Reliability Audits arid Reviews 117 Pulsations

Pulsations can cause problems in rotating equipment as well as reciprocating ma~hinery.~'.~* Pulsation resonances occur in piping systems and are a function of the fluid properties and the piping, compressor, or pump geometry

Pulsations can cause premature surge in centrifugal compressors and pumps if the generated pulses, such as from stage stall,I6 match one of the pulsation resonances of the system The potential pulsating excitation mechanisms for piping systems are the running speed component and its multiples, vane, and blade passing frequency and those caused by flow excited (Strouhal frequency) p h e n ~ m e n a ' ~

In the design stage, the acoustical natural frequencies of piping systems can be calculated using either digital2s or analog modeling proced~res.~' A model of a piping system analyzed on a digital computer is given in Figure 3-30 The predicted

Trang 10

118 Improving Machitiery Relinbiliry

pulsations in the reciprocating pump system at selected locations are given in Figure

3-31 These pulsation levels define the energy in the pump and the piping shaking

forces, and can be used to define the necessary piping supports and span lengths to achieve acceptable vibration levels The program can be used to redesign the piping

to reduce the pulsations to acceptable levels

Summary and Conclusions

Rigorous rotordynamic analyses must be thought of as additional insurance that the machinery will run without major problems Many of the analysis procedures and

computer programs that have been developed are being used by both the manufactur-

er and by consultants who offer these design audit services As with many computer

programs, the interpretation of the computer results is dependent upon the skill and

2

Figure 3-31 Predicted pulsations in pump and piping system for pump speed range of 170-260 rpm

Trang 11

experience of the analyst The manufacturer produces many machines that have no problems and thus has the confidence that the machinery will run successfully The independent consultant usually is asked to look only at machinery with some kind of prloblem Therefore, consultants probably look at the analysis from a slightly different viewpoint The goal of the manufacturer, the user and the independent consultant who make audits is the same: everyone wants a trouble-free machine

The independent audit generally occurs after the manufacturer has finalized his

drawings In many cases where problems were found during the audit, it turned out

that although the manufacturer had made an analysis in the initial stages, some dimensions were changed during the manufacturing phase, causing significant change in the calculated responses If the independent audit is made, any differ- ences between the manufacturer’s and the consultant’s calculations can be resolved before installation

Guidelines as to when an independent audit should be obtained are influenced by many factors Generally audits should be performed on:

1 , New prototype machines that are extrapolations in horsepower, pressure, number of impellers, bearing span, or incorporate new concepts

2 Machines which, if unreliable, would cause costly downtime

3 Machines that are not spared (no backups)

4 Expensive machines and installations in which the cost of the audit is insignifi- cant compared to total cost

Typical costs for analyses are difficult to specify since the scope of the work depends upon the adequacy of the supplied information, the complexity of the machinery, and the number of parametric variation analyses required If accurate drawings and system information can be supplied to independent consultants who perform these studies, then accurate cost estimates can be given Exxon’s experience may, however, give the most powerful incentive to studies of this type: It has been estimated that for every dollar spent on machinery reliability verification before

machinery commissioning, ten dollars were returned at plant startup, and one hun- dred dollars were saved during the life of the plant

Failure Statistics for Centrifugal Compressors

For any type of machinery considered, knowledge of failure causes and downtime statistics allows us to determine which components merit closer review Also, properly kept records could alert the review engineer to equipment types or models which should be avoided In some cases, failure statistics might provide key input to

a definitive specification In other words, “you learn from the mistakes of others.”

All of this presupposes that “others” saw fit to record their experiences If an engi-

neer has these data available he will no doubt use them before selecting machinery Whenever specific data are lacking, the review engineer has to resort to general statistics These statistics are typically available in the form given for centrifugal

compressors, Table 3-6 or in the form of availability tables (see Chapter 4) Probable

Trang 12

Table 3-6 Typical Distribution of Unscheduled Downtime Events for

Major Turbocompressors in Process Plants

Approximate number of shutdowns per train per year: 2

Estimated Estimated Average Downtime

.I6

.12 I 2 .06 02 40

Table 3-6 shows that rotor and shaft distress rank highest in downtime hours per year per train Blade or impeller problems rank next, followed by motor failures Obviously, centrifugal compressor reliability audits and follow-up reviews should concentrate on these areas first

Failure Statistics for Steam Turbines

Failure statistics for special-purpose steam turbines are often separated into those for impulse turbines and those for reaction turbines One such important statistic is represented in Figure 3-32 Of practical interest are primarily those failures which have a reasonable probability of occurring within 80,000 operating hours The author’s experience would indicate that each turbine type is acceptable as long as rigorous selection criteria are applied First and foremost of these would be an inves- tigation of vendor experience, blade stresses, and blade vibration behavior

Operating data accrued after 1970 support the belief that rotor blades furnished with impulse turbines rank about even with reaction-turbine blading A prerequisite

to operation would be that they were subjected to user design audits and pre-com-

missioning reliability reviews

Interesting statistics have also been quoted by a large North American industrial equipment insurance company Over a number of years, this company has insured an average of 6,353 steam turbines During this time period, one out of every 186 steam

Trang 13

Machinery Reliability Audits and Reviews 121

Versicherungs-AG, Munich, Germany Transcription courtesy of Siemens-America.)

turbines experienced failures which were serious enough to require disbursements from the insurance carrier

These failure figures appear to cover only the most serious events At 0.03 events per year per steam-turbine driver, the insurer’s statistics might lead us to believe that

we could relax our audit and review efforts However, we must keep in mind that our efforts are aimed not only at eliminating major wrecks but nuisance trips, excessive downtime, startup delays, and frequent maintenance as well

It should also be recognized that an awareness of failure causes is necessary for the effective implementation of machinery reliability audits For instance, bearing distress in steam turbines, large electric motors, and associated connected equipment

is often caused by the action of stray electric currents This type of damage is best eliminated by the up-front installation of current leak-off brushes, Figure 3-33 A well-designed leak-off system can be monitored for effectiveness and serviced with the equipment operating

Failure Statistics for Gas Turbines

In May, 1980, Allianz Versicherungs-AG of Munich, Germany, released a report

on the failures of modern industrial gas turbines.* Figure 3-34 gives the distribution

*Leopold, J., “Erfahrungen Mit Stationaren Gasturbinen Moderner Bauart,” Der Mascliinrtischarlen,

Vol 53, No 5, 1980

Trang 14

122 Ittprovittg Macltittery Reliability

Figure 3-33 Shaft riding brushes in turbine-generator application (Courtesy of Sohre Turbomachinery Inc., Ware, Massachusetts.)

of primary failure causes for industrial gas turbines from 1970 to 1979 Figure 3-35 shows the component damage distribution for the same machines

Gas turbines have been found to experience more frequent failures than steam turbines Quoting again the insurance company mentioned earlier, we would expect one serious failure per 26 gas turbines per year (They reported 20 failures per year out of

an average population of 520 gas turbines.) For gas-turbine-driven generators, their statistics show 4.3 driver failures for every failure of the driven equipment

Typical primary failure causes are reported in Table 3-7

It should be noted that the failure cause distribution given by U.S insurance carriers differs from that reported by Allianz in Figure 3-35 It is very difficult to weigh the significance of this observation-especially when we are being told that an identical series of failure reports submitted to both insurance carriers was coded quite differently by the two companies

Failure Statistics for Centrifugal Pumps

Many petrochemical plants assemble some data on pump failure causes, but few

of them take the time to make sure that pertinent findings reflect in the next issue of

Trang 15

Machinery Reliability Audits arid Reviews 123

Figure 3-34 Distribution of gas turbine component damage (Courtesy Der Maschinen-

schaden, No 153, 1980.)

Figure 3-35 Distribution of primary causes of failure for industrial gas turbines (Cour- tesy Der Maschinenschaden, NO 153, 7980.)

Trang 16

Table 3-7 Primary Causes of Gas Turbine Failures in the USA

Primary Failure Cause, Faulty Design Operating Bad

Gas Turbines Driving: Maintenance Problems Problems Repairs Other

Once the failure causes are broken down into such elements as O-ring failures, bearing overload failures, incorrect metallurgy, etc., it may become possible to trans- late failure analysis results into better specifications For instance, swelling of Viton O-rings in services containing certain chemicals may lead to the specification of Buna or PTFE elastomers; thorough analysis of bearing overload failures could point the way toward requiring pump vendors to provide bearing load values at off-design conditions; and frequent erosion damage in a given pumping service could lead to specifying hard-metal overlays

Nevertheless, Table 3-8 does represent fairly accurately the pump failure picture

of large U.S refineries As can be seen, mechanical-seal failures seem to predomi- nate Similarly, the tabulation summary given in Table 3-9 points to mechanical-seal distress as an area of concern at four overseas plants, also

Table 3-8 Typical Yearly Repair Summary, Centrifugal Pumps

Centrifugal pumps installed: 2560

Centrifugal pumps operating at any given time: 1252 (average)

Total pumps repaired in 1979: 768

Pumps repaired at site location: 382

Pumps repaired at own shops: 267

Pumps repaired at outside shops: 119

Trang 17

Machinery Reliability Audits and Reviews 125

Table 3-9 Primary Failure Causes, Centrifugal Pumps

(Four Overseas Plants)

Auditing and Reviewing Centrifugal Compressors

There is little difference in how an experienced engineer approaches audit and review tasks for centrifugal compressors as opposed to those for turbines, gears, and other machinery In each case he must obtain drawings and other technical data from equipment vendors He will then review all pertinent documentation for consistency, safety, compliance with specifications, etc., and document all areas requiring follow-

up Of course, he will also initiate and review the rotordynamic design audits described earlier in this chapter

Centrifugal compressor documentation requirements probably exceed those of most other machinery with the possible exception of large mechanical drive steam turbines A listing of relevant documentation is contained in API Standard 617 Using the API tabulation facilitates outlining the items recommended for review:

1 Certified dimensional outline drawing, including:

a Journal-bearing clearances

b Rotor float

c Labyrinth, packing, and seal clearances

d Axial position of impellers relative to guide vanes

e List of connections

Journal-bearing clearances may be required for rotor sensitivity studies Bearing dimensions allow rapid calculation of bearing loading and serve to screen for potential oil whirl

Labyrinth, packing, and seal clearances may be too tight for normal process operation The vendor may attempt to show good efficiency (low recirculation) during shop performance tests

Trang 18

Axial position of impellers relative to guide vanes needs to be reviewed in List of connections may uncover dimensional mismatching with purchaser’s conjunction with rotor float dimension Is rubbing likely to occur?

lines, excessive flow velocities, omission of specified injection points, etc

2 Cross-sectional drawing and bill of materials

These documents are primarily used for verification of impeller dimensions and internal porting, visualization of maintenance access, materials selection, and assessment of number of spare parts needed, etc A copy of this drawing and the bill of materials should also be forwarded to responsible maintenance personnel

a Axial position from active thrust-collar face to:

3 Rotor assembly drawing, including:

1 Each impeller

2 Each radial probe

3 Each journal-bearing centerline

4 High-pressure side of balance drum

b Thrust-collar assembly details, including:

I Collar-shaft fit with tolerance

2 Concentricity (or run-out) tolerance

3, Required torque for locknut

4 Surface-finish requirements for collar faces

5 Preheat method and temperature requirements for “shrunk-on” collar installation

c Balance drum details, including:

Thrust-collar assembly details are to be analyzed for non-fretting engagement and field maintenance feasibility Hydraulic fit is preferred

Balance-drum details are needed for rotor dynamics analyses and maintenance reviews

Dimensioned shaft ends for coupling mountings allow calculation of stress levels, margins of safety, uprateability, and coupling maintenance

The bill of materials is used to compare component designs and materials being released for fabrication Again, the bill of materials will allow definition

of spare parts requirements

4 Thrust-bearing assembly drawing and bill of materials

These are used to verify thrust-bearing size and capacity They are important

if directed oil-spray lubrication has been specified, and are essential maintenance information

Trang 19

Machinery Relicibility A d i t s and Reviews 127

5 Journal-bearing assembly drawing and bill of materials

Bearing dimensions are required to calculate bearing loading and rotor dynamic behavior, and for maintenance records

6 Seal assembly drawing and bill of materials

These are required to compare seal dimensions, clearances, and tolerances with similar data from seals operating properly under essentially identical operating conditions

7 Coupling assembly drawing and bill of materials

These are used for calculations verifying load-carrying capacity, mass moment of inertia, overhung weight, shaft-fit criteria, dimensional compatibil- ity between driver and driven equipment, material selection, match-marking, assembly and disassembly provisions, and spare parts availability

a Steady-state and transient gas or oil flows and pressures

b Control, alarm, and trip settings

c Heat loads

d Utility requirements including electrical, water, and air

e Pipe and valve sizes

f Bill of materials

Oil or gas flows and pressures must change as a function of gas-pressure and compressor-speed changes The review must verify that the seal gas or seal oil supply can accommodate all anticipated requirements for a given compressor This would include operation during run-in on air

Control, alarm, and trip settings are required for operating and maintenance manuals, as well as for initial field implementation by the contractor

Heat loads are required for capacity checks on oil coolers

Utilities requirements are required for proper sizing of switchgear, steam Pipe and valve sizes are employed in calculations verifying that maximum

8 Seal oil or seal gas schematic, including:

lines, eic

acceptable flow velocities are not needed

9 Seal oil or seal gas assembly drawings and list of connections

These are required for contractor’s (purchaser’s) connecting design

10 Seal oil or seal gas component drawings and data, including:

a Pumps and drivers, or motive gas suppliers:

1 Certified dimensional outline drawing

2 Cross section and bill of materials

3 Mechanical seal drawing and bill of materials

4 Priced spare parts list and recommendations

5 Instruction and operating manuals

6 Completed data forms for pumps and drivers

Trang 20

c Coolers and filters:

1 Fabrication drawings

3 Completed data form for cooler(s)

Overhead tank, main reservoir, and drain tanks (degassing tank, sour seal oil reservoir) must comply with specifications Should overhead tanks be given thermal insulation?

Coolers must be suitable for heating the seal oil during oil flushing opera-

tions Are they sized to cool the oil flow resulting from more than one pump operation? Can filters be fully drained? Do they have vent provisions? What is their collapsing pressure? What kind of cartridges do they accept? Specifica- tion compliance must be ascertained

Is instrumentation accessible? Can it be checked, calibrated, or replaced

without causing a shutdown? Is it properly identified? Are controllers and transmitters located at optimum locations for rapid sensing and control? Are switches of sound design and are they manufactured by a reputable company? Control valves sized right? Gauges made of acceptable metallurgy? Proper ranges?

a Steady-state and transient oil flows and pressures

b Control, alarm, and trip settings

c Heat loads

d Utility requirements including electrical, water, air, steam, and nitrogen

e Pipe and valve sizes

f Bill of materials

Are steady-state and transient flows within capability of pumps and accumulator? Will pumps and accumulators satisfy turbine hydraulic transients? Accumulator maintainable?

11 Lube oil schematic, including:

Are control, alarm, and trip settings tabulated?

Do heat loads have to be accommodated by fouled coolers?

Utility requirements are needed to allow plant design to proceed in such areas as electrical protective devices, water supply lines, and nitrogen supply for blanketing of reservoir Steam requirements must be identified for turbine- driven pumps

Pipe and valve sizes need to be checked to determine acceptable flow velocity The bill of materials should be reviewed to identify both inexpensive and

hard-to-obtain components It should be reviewed also by maintenance per-

sonnel Are O-rings, rolling element bearings, etc., identified so as to allow purchase from the actual manufacturers of these components?

Trang 21

12 Lube oil assembly drawing and list of connections

13 Lube oil component drawings and data, including:

1 Certified dimensional outline drawing

2 Cross section and bill of materials

3 Mechanical seal drawing and bill of materials

4 Performance curves for centrifugal pumps

5 Priced spare parts list and recommendations

6 Instruction and operating manuals

7 Completed data forms for pumps and drivers

1 Fabrication drawings

2 Maximum, minimum, and normal liquid levels in reservoir

3 Completed data form for cooler(s)

of pumpage

Instruction and operating manuals are intended for future incorporation in owner’s “Mechanical Procedures Manual” and conventional plant operating manuals

These are required for contractor’s (purchaser’s) connecting design

a Pumps and drivers:

b Coolers, filters, and reservoir:

14 Electrical and instrumentation schematics and bill of materials

The machinery review engineer should be given responsibility for obtaining these data and forwarding them to his electricalhnstrument engineering coun- terparts for review and comment

15 Electrical and instrumentation arrangement drawing and list of connections Same as item 14 At the completion of reviews by electrical/instrument engineering personnel, the final arrangement will be implemented by the contractor

16 Polytropic head and polytropic efficiency versus ICFM curves for each section

or casing on multiple section or casing units in addition to composite curves at 80%, 90%, loo%, and 105% of rated speed

Note; Request information on probable location of surge line for various molecular weight gases, as required

These are important data for future uprate and general performance verification studies, and can be used for purchaser’s check on vendor’s predicted performance

Trang 22

17 Discharge pressure and brake horsepower versus ICFM curves at rated conditions for each section or casing on multiple section or casing units in addition

to composite curves at 80%, 90%, loo%, and 105% of rated speed

For variable molecular weight (MW) gases, curves also shall be furnished at maximum and minimum MW For air compressors, curves also shall be furnished at three additional specified inlet temperatures

18 “Pressure above suction pressure behind the balance drum” versus “unit loading of the thrust shoes,” both in pounds per square inch (bar), using rated conditions as the curve basis

The curve shall extend from a pressure equal to suction pressure behind the drum to a pressure corresponding to at least 500 pounds per square inch (35 ata) unit loading on the thrust shoes Balance drum OD, effective balance drum area, and expected and maximum recommended allowable pressure behind the balance drum shall be shown on the curve sheet

Will balance drum labyrinth wear cause overloading of the thrust bearing? What happens when fouling (polymerization) occurs in the balance line? Is the design safe for a wide range of suction pressures?

19 Speed versus starting torque curve

Will the motor be designed to safely start the compressor? Even more important for gas turbine drivers!

a Number of vanes-each impeller

b Number of vanes-each guide vane

c Number of teeth-gear-type couplings

These are required for machine signature “real time” on line diagnostic or spectrum analysis They will allow identification of relevant frequencies, and possibly be useful in determining which component has undergone deteriora- tion Refer also to the illustrative example in Chapter 1, “How to Deal with the Typical API Data Sheet.”

a Method used

b Graphical display of bearing and support stiffness and its effect on critical

c Graphical display of rotor response to unbalance

d Graphical display of overhung moment and its effect on critical speed Reviews will identify whether there is risk of operating too close to critical speed, or whether rotor is likely to vibrate at the slightest sign of unbalance If gear couplings are used, the effective (instantaneous) overhung moment may change as a function of tooth loading or tooth friction The probability of encountering critical speed problems as a function of gear-coupling deteriora- tion can be investigated by examining graphical displays of effective overhung moment versus critical speed

a Method used

b Graphical display of mass-elastic system

c Tabulation identifying the mass-moment torsional stiffness for each compo-

20 Vibration analysis data, including:

21 Lateral critical analysis, including:

speeds

22 Torsional critical speed analysis for all motor and gear units, including:

nent in the mass-elastic system

Định dạng
Số trang	45
Dung lượng	1,14 MB