Improving Machinery Reliability 3 Episode 3 pot

Vendor Selection and Bid Coriditionirig 67 Mechanical Seal Selection and Evaluation Pump-application engineers will generally agree that seal and seal environmental system selection on

64 Improving Machinery Reliability

Sneak analysis identifies the proper and improper operation of a system’s hard- ware and software The analysis provides a systematic, consistent, and thorough review of the system’s current and logic paths, down to the individual statement and component level Sneak analysis is not restricted to critical functions (as are many other analyses) but analyzes the complete system, function by function The analysis will identify the cause and recommend a solution to a sneak condition, design con- cern, or document error before the problem occurs If left undetected, sneak condi- tions occurring during testing usually result in program delays to the project

The sneak analysis process generates detailed functionally oriented patterns, called network trees, of the circuitry and software that can be reviewed individually

or in groups to understand the system These network trees not only make sneak analysis possible, but are a powerful tool in reducing the cost and improving the quality of other reliability and safety analyses, such as Failure Mode and Effects Analysis, Hazard Analysis, Fault Tree Analysis, Common Cause Failure Analysis, or Mean-Time-Between-Failure Analysis, which may be specified on a project

The major benefits derived from the performance of sneak analysis are:

1 Savings of overall project dollars

2 Increased confidence in system safety, reliability, and operability through inde-

3 Fewer system development delays

pendent design verification

Identification of sneak conditions early in the project life cycle can provide cost savings as a result of changing a circuit or logic path on paper-rather than changing actual hardware or software Empirical data obtained after performing a sneak analy- sis demonstrates that increased reliability and operability of the system occur when corrections for identified sneak conditions are made

The broad principles governing the selection of major machinery vendors were stated earlier in the chapter How these principles can be applied most advantageous-

ly is again best illustrated in a typical example Take pumps, for instance

An up-to-date edition of Thomas Register of American Manufacturers will list dozens of pages of pump manufacturers Their detailed product listing contains another 100 or so pages ranging from “Pumps, Acid,” to “Pumps, Wine.” Even after reducing the potential bidder’s list to manufacturers of refinery-type centrifugal pumps, there remain some 20-30 vendors who could be invited to bid on a given project or for a given pumping service Were all of them to be considered, much time and money would be spent on preparing bid specifications, providing the neces- sary vendor liaison, and finally evaluating the profusion of bids received I

The need to limit bidding to a few capable vendors is quite evident But what con-

stitutes capable vendors? What criteria should be applied to narrow down the selec- tion to manageable size? How many bids are manageable? This segment of our text

Vendor Selection and Bid Conditioning 65

attempts to give guidance in this regard It explains selection procedures that have

given satisfactory results in a large number of major refinery and chemical plant construction projects More important, though, it shows how vendor selection and equipment selection criteria interact and must be given simultaneous consideration

Standardization Within What Limits?

A petrochemical plant obviously would not find it practical to purchase equipment from boo many manufacturers Spare parts identification, procurement, and ware- housing are expensive and leave margin for error Also, it would be progressively more costly and difficult to train mechanical workforces for full proficiency in too many equipment types or models

On the other hand, standardizing on too few manufacturers may deprive the user

of optimally selected equipment There is obviously no single manufacturer of cen- trifugal pumps who can lay claim to products that are consistently more efficient, easier to maintain, and more rugged than competitive equipment

Experience shows that two or three vendors could adequately cover the on-site pumping services of a typical petrochemical plant Five or six manufacturers should

be invited to submit bids, and two or three of these subsequently selected for con- tract award Off-site pumps, frequently of ANSI or ISO-type, could be selected from one additional vendor among those who had been invited to bid

Highly specialized centrifugal pumps for critical services, such as high-pressure boiler feedwater or pipeline supply pumps, may have to be purchased from the most experienced source, regardless of whether the vendor is among those selected for on- site and off-site pumps Purchase of these special pumps should be handled separate-

ly on an individual basis, rather than being lumped with other pumps

Assessing Vendor Experience

Three principal characteristics identify a capable, experienced vendor:

1 He is in a position to provide extensive experience listings for equipment offered

2 His marketing personnel are thoroughly supported by engineering departments Both groups are willing to provide technical data beyond those that are custom- arily submitted with routine proposals

3 His centrifugal pumps enjoy a reputation for sound design and infrequent main- tenance requirements

With a large: project involving 200-300 or more pumps, it is often necessary to delegate to the contractor’s equipment engineers the responsibility of verifying ven- dor experience with other users For certain critical services (e.g., high-pressure boil-

er feedwater pumps, large cooling-water pumps, multistage pipeline or feed pumps, etc.), the owner’s engineer would be well advised to check for himself When

66 Improving Machinery Reliability

procuring pumps that are required to comply with the standards of the American Petroleum Institute (Le,, API 610), a capable vendor will make a diligent effort to fill

in all of the data requirements of the API specification sheet However, the real depth of his technical know-how will show in the way he explains exceptions taken

to API 610 or to supplementary user’s specifications Most users are willing to waive some specification requirements if the vendor is able to offer sound engineering rea- sons, but only the best qualified centrifugal pump vendors-those from whom you want to purchase-can state their reasons convincingly

Concentrating on Problem Applications

In assessing vendor experience, the engineer responsible for vendor selection should concentrate on pumping services that have a history of being troublesome The approaches proposed by the various bidders for solving typical problem applica- tions may differ drastically and allow rapid separation of sound proposals from potentially troublesome ones

One of the most common problems in pump application is insufficient net positive suction head available Often this is not realized until vendors’ proposals have been solicited and some vendors have failed to meet the conditions

When NPSH is insufficient, the plant design should be reviewed to determine whether decreasing the length or increasing the size of the suction line is feasible Raising suction-vessel elevations is another possibility, but economics may dictate selection of pumps designed especially for low NPSH When NPSH availability is limited, the pump vendors, of course, will offer double suction pumps when possi- ble Also to be considered are inducer-type impellers, but it must be realized that inducers may have a limited flow range relative to a non-inducer-type pump Also, inducers should never be used in erosive services Pump performance deteriorates rapidly as the effectiveness of the inducer is reduced by erosion, and cavitation begins to take place

Another solution to NPSH problems is to use a vertical deep-well pump In order to minimize maintenance problems with this type of pump, each proposed pump should

be checked for applicable service experience, with attention to exact model numbers, similar pumpage characteristics, and process conditions High-head, low-capacity ser- vices also pose problems to engineers selecting pumps In general, four types of pumps are available for this application: multistage horizontal centrifugal, multistage vertical centrifugal, reciprocating, and single-stage high-speed centrifugal For process units in continuous duty, experience has shown that multistage vertical and reciprocating-type pumps require more maintenance than the other two types If a pump must handle a wide range of specific gravities, or if the fluid is particularly vis- cous, a reciprocating pump may be the only answer If there is limited NPSH avail- able, a multistage vertical pump may be required But for the bulk of high-head, low- capacity applications, serious consideration should be given to high-speed pumps, with the multi-stage horizontal centrifugal pump a good second choice

Vendor Selection and Bid Coriditionirig 67

Mechanical Seal Selection and Evaluation

Pump-application engineers will generally agree that seal and seal environmental system selection on many pumps is becoming more complex and time consuming than pump selection.’2 Still, an average of only 10%-35% of the allotted pump-engi- neering time is generally spent on the mechanical seal system.13 Looking at the cost

of seal failures and such consequences as major fires, release of toxic materials, and unit downtimes, there is reason to believe that more time should be spent on seal sys-

tems design But before we decide who should spend this time, we should examine

the various selection practices prevalent in the petrochemical industry

Seal selections made entirely by the pump vendor have generally proven to be least reliable The pump vendor is concerned that his competitor will underbid him, and thinks that the engineer selecting the pump will only look at the initial, installed cost without giving credit to the potential run-length extension and maintenance cost avoidance of superior seal components or seal system designs Consequently, the least expensive seal is often selected, leaving plant operations or maintenance bur- dened with an inherently weak seal Furthermore, pump manufacturers seldom receive experience feedback on seals furnished with their pumps Seal selection by the pump vendor alone should thus be discouraged

Contractor’s or user’s standards have generally been applied with somewhat high-

er success Unfortunately, many of these standards are full of generalities and give little guidance on specific requirements Very often the stated requirements do not separate barely acceptable from truly successful seal systems Lack of specific guid- ance in an otherwise well-intended specification may significantly impair its useful- ness and deprive the user of a low-risk sealing system

Optimum seal selection practices should make extensive use of vendor experi- ence These practices must encourage the seal vendor to use his own gland design and to recommend seal systems, not just seals To properly advise the user, seal ven- dors require full information on product composition, process conditions, crystalliza- tion temperatures, solids entrainment, and the like All of these data are highly rele-

vant if proper selection is to be ensured, and withholding data for “security” reasons may cost the user dearly Optimum seal selection consists of the following steps:

All relevant data must be disclosed to the seal vendor If security is truly a valid concern, disclosure should be preceded by signing confidentiality agreements

At least three and preferably four major seal manufacturers with strong and capa- ble representation in the user’s geographic location should be invited to submit bids The user should screen and verify vendor capability by such criteria as ability

to furnish engineered seal components, e.g., special pumping screws instead of

ineffective pumping rings, and by vendor’s willingness to stock appropriate spares

in the user’s geographic area

The bid invitation should clearly state that the user is interested in buying a seal

68 Improving Machinery Reliability

The bid invitation should instruct the vendor to propose a minimum of seal types For example, better-than-necessary seal-face materials should be offered for some services, or bellows seals should be considered for an expanded range of applica- tions in order to reduce the user’s spare parts inventory

Proposals obtained from the various bidders must be examined critically for tech- nical differences, and pertinent deviations should be noted for follow-up by the user’s resident expert These technical differences must be investigated by request- ing that vendors submit experience data These data should include the names of other users who may be contacted to obtain verification of satisfactory experience The user’s resident expert should make these contacts

Whenever possible, seal selection and evaluation procedures should make use of in-house feedback as to principal reasons for seal failures at that location These data should be discussed with the vendor and solutions proposed

The user’s priorities should be clearly established and transmitted to potential seal vendors Safety, long life, standardization, and ease of maintenance all precede cost in order of importance

These last topics merit further elaboration Experience shows that the emphasis should clearly be towards standardization without sacrificing reliability Seals should

be cartridge seals that are simple to install and maintain They should not be vulnera- ble to minor deviations in pump operation or properties of pumpage Indeed, a given seal should suit the needs of many pumping services so as to reduce spare parts inven- tory, streamline training requirements and leave little room for maintenance errors These goals can be met with a well-defined selection strategy Such a strategy may well result in the added bonus of greatly increased seal reliability and consider- ably lower maintenance expenditures for most petrochemical plants Many major manufacturers of mechanical seals have indicated their support and willingness to cooperate in implementing our selection strategy Specifically, this strategy requires the development of bid request information that asks the seal vendor to identify opti- mum seal configurations Principal features of optimum seals are highlighted in the following pages; for additional details see Chapter 13

Development of Bid Request Package

The development of bid request information is a key element in a selection strate-

gy leading to the procurement of mechanical seals that comply with the user needs previously outlined Bid requests must be forwarded to several experienced seal manufacturers and must clearly state that the manufacturer should submit cost pro- posals for only those services where his seal selection is backed by solid experience His inability to furnish seals for some services should in no way disqualify him from submitting bids for those services where he is competent to provide a good product

To begin with, the user should assemble API data sheets for the various pumps that need mechanical seals

This pump tabulation must include and disclose all of the fluid properties and operating parameters known to the user With disclosure thus going beyond the typi-

Vendor Selection and Bid Conditioning 69

cal contents of API data sheets, the seal manufacturer can be requested to list the

“operating windows” within which the proposed seal will function reliably in the pump tentatively selected It is to be understood that the “operating window” refers

to the actual seal with the seal materials, balance ratio, flush plan, stuffing box lay-

out, etc., selected and disclosed by the vendor Notice also that the pumps are desig- nated “tentatively selected” because even a capable seal vendor may be unable to offer his optimum seal for a given pump Should this be the case, it might be appro- priate to consider selecting a more suitable pump model for the intended service

As mentioned earlier, it is recommended to mail the bid request information to three or four capable mechanical seal manufacturers Their response should be criti- cally analyzed to ferret out significant deviations among the various proposals These deviations may range from materials of construction to different API flush plans and from differences in basic configuration to differences in application phi- losophy of stationary vs rotating seal members Reconciling the deviations or differ- ences will assist the engineer responsible for final selection in determining whose seal offer has the best chance of meeting the user criteria highlighted earlier

Desirable Design Features identified

Evaluation of the various bids is made easier by recognizing desirable design fea- tures incorporated in mechanical seals Some of these merit closer consideration

Cartridge Construction Cartridge seals are designed for rapid installation on and removal from pump shafts The cartridge seal is an arrangement of seal components

on a shaft sleeve and in a seal gland constituting a single unit that is usually assem- bled and pre-set at the factory Both bellows and spring-type seals can be cartridge- arranged if the pump stuffing box is large enough

Cartridge seal units offer major maintenance advantages Replacement is rapid and there is far less risk of assembly error and assembly damage than with conven-

tional mechanical seal mounting API 682 requires that cartridge seals be selected

for typical pumps in petrochemical plants

Silicon Carbide Hard Face Material Heat generated at the seal faces must be rapidly conducted away if fluid vaporization and resulting problems are to be avoid-

ed Depending on service conditions and pump design, either the rotating or station- ary seal ring must be counted on to dissipate as much frictional heat as possible High thermal conductivity and hardness make silicon carbide the preferred seal face material in many of the more severe applications

Placement of 0-Rings An advantageous seal design recognizes that O-ring life can be reduced by close proximity to the heat source, by swelling due to chemical attack, and by operation i n a dynamic mode, especially in the presence of erosive materials Going to PTFE chevrons or wedges may allow operation at higher temper- atures and reduced risk of chemical attack, but will lead to fretting of metal surfaces

in contact with it Bellows seals eliminate many of these problems by using static

70 Improving Machinery Reliability

secondary sealing Seals with spring loaded running faces are forced to use dynamic means of secondary sealing which could, in some instances, be more prone to fail- ure This potential problem can be overcome by selecting "stationary" and/or gas seal designs In a stationary seal, the spring-loaded face is not rotating

Mechanical Design Considerations Important differences can exist in the

mechanical designs of competing vendors For instance, execution E of Figure 2-7 shows the method of clamping rotating hard face ( 1 ) against shaft sleeve ( 2 ) of a sta- tionary bellows seal assembly This clamping method ensures perpendicularity between shaft centerline and rotating seal faces However, this clamping method also invites distortion at the running faces Execution F tends to avoid distortion by mounting the rotating hard face in a resilient backing ring ( 3 ) and by pulling mount-

ing ring (4) against collar (5)

Two different clamping methods are shown in Figure 2-8 Both of these methods

were devised to eliminate distortion of running faces However, in execution G the collar is set-screwed to the shaft or shaft sleeve, whereas in execution H the mating ring carrier is set-screwed to the shaft or shaft sleeve Experience shows that perpen-

MATING RING CARRIER

Figure 2-8 Mechanical seal face clamp-

ing methods devised to reduce risk of

distorting rotating face

Vendor Selection and Bid Conditioning 71

dicularity and seal setting accuracy are more difficult to achieve with the clamping method indicated as execution G in Figure 2-8, which requires very careful adjust- ment of cap screws inserted through the collar

Another interesting difference exists in the carbon holders of the more conven- tional mechanical seals The upper half of Figure 2-9 shows a lock ring ( 1 ) executed with a set screw (2) that engages a slot in carbon holder (3) Under certain service conditions, contact between set screw and slot may cause a wear pattern that may prevent proper seal operation Moreover, tightening of the set screw can distort the relatively thin lock ring and cause contact or interference between lock ring OD and carbon holder ID The construction features shown in the lower half of Figure 2-9 would tend to eliminate both of these potential problems by providing an axially ori- ented drive pin (4) and a considerably heavier lock ring (1) Both designs shown in Figure 2-9 deserve credit for presenting relatively smooth, low turbulence surfaces to

their respective fluid environment This is largely accomplished by locating the springs on the atmospheric side of the seal

Whenever possible, seals should avoid having the spring (or springs) immersed in the fluid Figure 2-10 shows two of several seals that probably give excellent service

in many services and applications These seals use one or more stationary springs and incorporate several desirable features: a cartridge arrangement for ease of instal- lation; the single non-rotating spring shown with the design in the top half of Figure 2-10 is arranged to operate away from the product; a bronze spring retainer ( 1 1 )

serves as a throttle bushing; and the relatively clean profile inside the stuffing box

Figure 2-9 Two conventional mechanical seals with carbon holder driven by modified set screw (upper half) and horizontal pin (lower half)

72 Improving Machinery Reliability

reduces seal drag (Note that rotating components are identified with even, and sta- tionary components with odd numbers.)

Except for using several springs, the seal design shown in the bottom half of Fig- ure 2-10 appears to be quite similar to the design shown in the top half Both seals incorporate the desirable features of many similar stationary seal designs:

Self-squaring faces This feature may result in appreciably better seal life for pumps with excessive shaft deflection or pumps operating with nominal shaft deflection at high speeds

Non-flexing springs Spring life extension and long-term, uniform pressure can be expected

Pre-assembled cartridge construction These seals can be shipped with the gland plate in place No field measurements or settings will be required

However, a closer look will show functional differences in the arrangement of the O-ring (3) It could be argued that progressive wear of the seal faces shown in the upper half of Figure 2-10 will cause the O-ring to make sliding contact with a clean portion of part (9), whereas advancing the stationary seal face in the lower half of Figure 2-10 will cause the O-ring (3) to slide over a wetted and potentially contami- nated portion of the gland plate (15)

Figure 2-10 Two different stationary seals with non-rotating springs located away from pumped liquid

Vendor Selection and Bid Conditioning 73

Functionally similar features can also be found in an intermediate range of bel- lows seals as shown in Figures 2-1 l and 2-12 Rotating bellows seals tend to be self- cleaning by virtue of centrifugal action As mentioned earlier, they do not incorpo- rate sliding (dynamic) elastomers Instead, they use static secondary sealing means

Figure 2-1 1 Bellows seal per API 682 (Courtesy Flowserve Corporation.)

\

Figure 2-1 2 Cartridge-mounted bellows-type mechanical seal (Courtesy Borg-Warner

Seals.)

74 Improving Machinery Reliability

Properly designed stationary seals have given outstanding service in abrasive slur-

ry applications Figure 2-13 shows one such seal installed in a cone-shaped environ- ment This short “tapered stuffing box” promotes the outward flow of abrasive parti- cles and at all times allows new, fresh fluid to contact the balanced single-acting, dead-ended seal It should be noted that long tapers will require vortex-breaker ribs;

these are usually cast or welded into the housing bore

Generally, stationary-type seals or gas seals should be preferred in applications encountering face peripheral velocities in excess of 4,500 feet (1,372 meters) per minute, serious shaft deflection, or coking after the pumpage has crossed the seal faces When applying bellows seals in light hydrocarbons, we should look for design features that prevent torsional windup of the bellows in case the seal faces undergo slip-stick motion relative to each other Also, the seal face balance line of bellows seals may shift when applied without adequate forethought in light hydrocarbon ser- vices If the design does not alleviate these concerns, the user may favor spring, and especially gas-type mechanical seals for light hydrocarbon services

Bid Comparison

After the bids are received, they must be tabulated and compared The reviewing engineer should look for significant differences among the competing bids and should determine which offer incorporates most of the desirable design features Special features beyond those specified by the purchaser may have been proposed by some bidders and would deserve extra credit for reducing the risk of catastrophic failure incidents Each special feature must be given a separate assessment of value Alternatively, the purchaser may decide that bidder X’s offer is less expensive than,

but nevertheless technically equal to the offer made by bidder Y He may now wish

to upgrade his selection by asking X to furnish the seals with optional, although not previously specified features

Using optimized seal selection procedures as outlined requires close cooperation between seal user and seal vendor This should not be too difficult to achieve and the benefits to both parties should be quite evident In this cooperative effort, the user has the responsibility of disclosing to the vendor fluid properties, application para- meters, permissible leakage rates, and even maintenance practices This disclosure will allow the seal vendor to obtain a better feel for user know-how and sophistica- tion The vendor can then effectively plan ways to assist the user He can remind the

Figure 2-13 Stationary seal with

successful experience record in

(Courtesy Burgmar% Seals Amer- ica, Houston, Texas.)

Vendor Selection arid Bid Conditioning 75

user that there are perhaps elements of pump design which will influence the success

of even the best available seal system In some cases, the seal manufacturer will have

to deal directly with the pump manufacturer in efforts to define such pump modifica- tions and might be necessary to accommodate optimum seal configurations

The results of these efforts are shown in Reference 13, which documents signifi- cant increases in seal mean time between failures at three British oil refineries This

is plotted later in Figure 5-4

Gas-Lubricated Mechanical Seals for Pumps The gas sealing technology used

in gas compressors has been successfully applied to the emission-free sealing of liq- uid pumps since the early 1990s Gas-lubricated seals are thus an option to be weighed against both conventional seals and such higher-cost alternatives as canned motor and/or magnetically driven sealless pumps While sealless pumps have their place, the potential purchaser should carefully study anticipated installation and operating and maintenance costs before making an informed choice

Gas-lubricated seals employ a pressurized gas supply system and are available as single or double seals Figure 2-14 shows a double mechanical gas seal in the unvented seal chamber of a conventional centrifugal pump The seal or barrier gas, typically nitrogen, is introduced at a pressure of 0.2 Mpa (3psi) higher than the product pressure Since many single-type gas seals find application as safety seals, the reader is referred

to Chapter 13, pages 550-558, which describe these seals in greater detail

Good Proposal Data Lead to Comprehensive Bid Tabulation

As mentioned earlier, it is essential that completed API data sheets be submitted

with each proposed pump The proposal package must also include performance curves and typical pump cross-sectional drawings In addition to these, the vendor

must state minimum allowable flow and NPSH required (NPSHR) for the entire

capacity range Because minimum allowable flow could be governed either by ther- mal or mechanical considerations, the vendor should be asked to specify his basis With these data and any notable exceptions given by the various bidders, a com- prehensive bid tabulation can now be constructed Careful review will narrow this bid tabulation to two or three principal manufacturers, as illustrated in Figure 2- 15

These would be the manufacturers whose equipment more consistently offered high performance by demonstrating such features as low risk suction specific speeds and operation near the best efficiency point (BEP) (For a more detailed treatment of this

topic, refer to Chapter 3, “‘Reliability Review for Centrifugal Pumps.”)

Engineers involved in pump-failure analysis have long suspected that cavitation erosion can plague even pumps in pumping circuits where the available net positive

suction head (NPSHA) exceeds the manufacturer’s certified required value, NPSHR

NPSHR tests by vendors are principally concerned with determining the performance

drop-off point A drop of 3% in total dynamic head is usually considered indicative of cavitation, and whatever NPSH is available at that point is though to be the NPSH

required by the pump However, it should be realized that significant mechanical damage may be encountered by long-term operation of some pumps with inadequate

76 Improving Machineiy Reliability

Buffer gas IN

Figure 2-14 Gas-lubricated mechanical seal (double seal) (Courtesy of Burgmann Seals America, Houston, Texas.)

margin between NPSHA and NPSHR These findings have prompted many users to

develop and use tentative guidelines that are largely empirical in nature.14 The publi- cation of quantitative correlations of suction recirculation as a function of proximity

to best-efficiency-point (BEP) flow for pumps with different construction and perfor- mance characteristics had to wait until late 1981 Details are given in Chapter 3

Bid Tabulation and Bid Conditioning: An Overview

Bid tabulations represent the culmination of the engineer's specification efforts

Proposals are submitted by the various vendors and pertinent items are written up side by side to facilitate easy comparison To the experienced engineer, the bid tabu-

Vendor Selection arid Bid Conditioning 77

Figure 2-15 Bid tabulation for pumps

78 Iniproving Machinery Reliability

lation is far more than a mere comparison: it is a master checklist of relevant items, some of which can have a profound influence on equipment safety, reliability, main- tainability, and availability Bid tabulations highlight the strengths and weaknesses

of certain proposals They indicate which items are acceptable or unacceptable, which items require follow-up, or which items require that the vendor be challenged outright to explain inconsistencies, design extrapolations, or oversights

Bid conditioning refers to the assignment of dollar credits and debits for strengths

and weaknesses, pluses and minuses, in a vendor’s offer Credits and debits may be assigned to reflect differences in contractually defined operating efficiencies, com- ponent strength differences, uprate capability, tolerance for occasional overloading, better or inferior maintainability, likelihood of delivery delays, extra expediting and inspection efforts, field service capabilities, spare parts requirements, etc

Bid-conditioning efforts for large unspared turboniachinery can be extensive and time consuming Inevitably, though, the exercise will be worth the effort Figure 2-

16 shows a fictitious bid tabulation sheet for centrifugal compressors Figure 2- 17 explains some of the many entries for which credits and debits must be assigned, or which must at least be considered before a final “conditioned” value can be calculat-

ed for each of the three offers

We have alluded to cost considerations other than pure bids prices With progres-

sively higher energy costs, pump efficiency becomes a major factor In some loca- tions, the yearly cost (in 1998) of one horsepower now often exceeds $500, and one might have justifiable concern that quoted efficiencies could be falsely inflated Mere consideration of quoted efficiencies should be replaced by certified test-stand efficiencies, field feedback, and perhaps contractual penalty clauses Credits and debits for efficiency deviations must compare the future value of money and the anticipated operating life of pumps Depending on the rate of return acceptable for energy conservation on the project, the value of one horsepower saved may justify

an incremental investment of several times $500-perhaps $4,000 in 1998 dollars Credits can be assigned also to recognize lower maintenance costs If it is possible

to make such an assessment based on comparisons of repair costs, the credit or debit number can be used outright If only repair frequencies are available for comparison,

it should be remembered that average pump repairs often cost in excess of $7,000 per event (including shop labor, materials, field labor, and burden) in 1998

There is value also in demonstrably better, heavier, more easily groutable base- plates for large horizontal centrifugal pumps Again, a rule of thumb: $1,200 Unlike a bid tabulation sheet listing only API data and cost, bid conditioning

serves to bring all offers on the same common denominator by assessing all relevant

parameters The bid tabulation, shown in Figure 2- 15, illustrates this supplementary

input It can be seen that more often than not, the best centrifugal pump is neither the most expensive nor the least expensive one on the bid slate The best pump manufac-

competition And the most capable reliability engineer is one who heeds the advice given by John Ruskin: “There is hardly anything in the world that some man cannot make a little worse and sell a little cheaper, and the man who buys on price alone is this man’s lawful prey.”

Vendor Selection und Bid Conditioning 79

Absorbed Power a t Normal PT (BHP Cornpressor)

O r i v e r Output Power P Norniai P l (BHP O r i v e r )

Max D i s c h Temp @ Max Cant Speed (‘F)

Heaviest Maintenance Load ( I b )

Base P l a t e Comnon f o r Turbine & Compressors

Common Turb & Camp Lube System Mfd b y

SeparateICormion Lube & Seal Systein

Seal Type

Overhead Bladder TanklCapacity (Gal I o n s )

Overhead Seal Rundown TankICapacity

Lube O i l Rundown TanklCapacity ( G a l l o n s )

Method o f F a b r i c a t i o n o f I m p e l l e r s

Thrust B e a r i n g Mfr.1Type

Thrust Hearing L o a d i n g l R a t i n g (PSI)

Journal Bearings Mfr.1Type

Mechanical Running Tests

Closed Loop Performance Test

I n l e t Mach Number (each wheel)

O i f f u s e r Width (each stage)

I m p e l l e r Width (each stage)

Shaft Oiam a t I m p e l l e r s (nominal, inches)

Shaft S t r e s s a t Coupling, p s i

Coupling Hare, inches

Couplinq R a t i n q (HPIIOO rpm) & S F

A

-

x - 100 Hor

650 A-111 5-111

I M P - X

C I ALU

13280

6380 Yes Vendor Conmon

F i l m

58

1000 Welded KTB

1251500

O r i o n

1401200 Vac Dehyd

Bendix 416 Yes

B e n t l y None Attached 7%

Attached Included Inc 1 uded None Included

C i t y H Submitted Attached

88

8

-

Y-101 Hor

16200

0 4 3

108 0-G3

680 A-I21 IMP-Y

C l ALU-TEF

14000

6000 Yes Vendor Comnon Mec h Con t

62 None

S e i f Oynac None Attached 7%

L a t e r

l n l u d e d None None

123 0-G4

670 A-811 5-1550 IMP-X

2201300 Waukesha

851500

C e n t r i f 2urnIG Yes

L a t e r

L a t e r Attached

80 Yes

Yes

Per Spec GWPJFSS 2.612.0 2.311.86

R

.7a1.81

11380 CB-C M13.2

Figure 2-16 Bid tabulation for centrifugal compressors

80 Improving Machinery Reliability

CONSlDEKATlONS FUR EVALUATING CENTKIFUGAL COMPRESSOR BIDS

Have a l l compressor niodels been s u c c e s s f u l l y operated elsewhere?

Does t l i e [machine w i t h fewer i m p e l l e r s produce t o o much head i n t h e second s e c t i o n ?

Compare q u o t e d d i a m e t e r s w i t h t i p s p e e d s S u i t a b l e f o r i m p e l l e r f a b r i c a t i o n method?

C o n s i s t e n t w i t h t i p speeds? Maximuin a l l o w e d c o n t i n u o u s speed?

t l i e vendor h a v e e x p e r i e n c e w i t h i n i p e l l e r s a t t h e s e t i p speeds?

How c a l c u l a t e d ?

Are a p p a r e n t e f f i c i e n c i e s and losses r e a l i s t i c ? How determined?

Are h i g h e r d n t i c i p d l e d d i s c h a r g e t e i n p e r a t u r e s (Vendor 8 ) d e t r i i n e n t a l t o process! C o s t l y ?

Why i s Vendor (I o f f e r i n y PTFE-Coated l a b y r i n t h I l i d t e r i d l ?

Does luwer Iiiaiiitendnce weight a l l o w i n s t a l l a t i o n o f less w p e n s i v e overliead

i l h y does Veiidor C propose t o indnufacture h i s own t h r u s t b e a r i n q s ?

Lhy does Vendor ti propuse t o iiianufacture h i s ohn j o u r n a l b e a r i n g s !

I n v e s t i g a t e r a t i n g b a s i s used O Y r d c l i vendor h'hy t h f d i f f e r e n c e ?

Is t h e c l a r i f i e r i n t e n d e d t o reiiiove o n l y w a t e r ? Water and I i g l i t t i y a r o c a r b o n s ?

l i r e t l l e f e ally d d V d n t a g e S l d i S a d V d 0 t n g F s t o b a r r e l COnStrUCtlOn f o r t h i s Sel-vlCe?

Are these vdlUeS c o i i s i s t e n t With t h ? r e l a t e d d d t d , above? U O f S

S u f f i c i e n t l y f a r f r o n i a n t i c i p a t e d o p e r a t i n g Speed r a n g e ?

llave you

llldde d CUlllpdfiSOll? DO VCiIdUT PtUpUSed V r l O C l t i e S dCCUrdteIy r e f l e C t S l d l e d gas

Are quoted l l i d t e r l d l S COllSlSterlL W i t h s p e c i f i e d l i i d t e r l d l S ? I n d U S l r y p r d C t l t e S ?

WOUlO Sepdl'dte l u b e dnd Sed1 011 S y S l W l S lldVe been ddVdlltdgC'OUS?

h h d t i 5 the C d l C u l d t l U n D d S i S Used t o a r r i v e a t 1000 qallOllS V S 2000 q d l l O n S

Could d IIIO~L! S u i t a b l e b a l a n c e p i s t u n d e s i g n be used b y Vendor C f o r t l l r u s t r e d u c t l o r l ?

D i s c u s s e x p e r i e n c e q u a l i f i c a t i o n s w i t h Vendors B and C as soon as p o s s i b l e

Vendor C n u s t SUbniIt f u l l A P I d a t a b e f o r e f u r t h e r c o n s i d e r a t i o n Can b e g l w n

Uoes he have soniethiny t o I i i d e ?

Are i m y e l l e r and d i f f u s e r diinensioris l a r g e enough f o r p o t e n t i a l l y f o u l i n q

Vendor Selection and Bid Conditioning 81

References

1 Staroselsby, N and Ladin, L., “Improved Surge Control for Centrifugal Com-

pressors,” Chemical Engineering, May 21, 1979

2 Staroselsky, N., “Better Efficiency and Reliability for Dynamic Compressors Operating in Parallel or in Series,” Unnumbered paper presented at ASME Ener-

gy Technology Conference and Exhibition, New Orleans, LA., February 3-7,

1980

3 Abbey, B and Bloch, H P., “Continuous Torque Measurement Boosts Machin-

ery Efficiency,” Hydrocarbon Processing, September 1977

4 Turner, B., “On-Steam Cleaning of Turbomachinery,” Proceedings of Second Turbomachinery Symposium, Texas A&M University, October 1973

5 Warren, G B and Howard, T W., “The Removal of Deposits from Steam Tur- bine Passages,” Transactions of the ASME, April 1947

6 Skrotzki, B G A and Vopat, W A., Stearn and Gas Turbines, McGraw-Hill

Book Company, Inc., New York, 1950

7 Howell, F W and McConomy, T A., “Turbine Blade Deposits,” Power Engi- neering, June 1967

8 Scott, J N., “Improving Turbocompressor Efficiency Via Performance Analysis Techniques,” ASME Paper No 77-GP-53, March 1977

9 Weich, B., “Axial Compressors For Large Blast Furnaces,” Sulzer Brothers, Ltd., Winterthur, Switzerland, April 1974

10 Prueftechnik, A G., D-8045 Ismaning, Germany

11 Bloch, H P., “How to Select a Centrifugal Pump Vendor,” Hydrocarbon Pro- cessing, June 1978

12 Bloch, H P., “A User’s View of Fluid Sealing Economics,” Paper presented at 45th Annual Meeting of Fluid Sealing Association, Sun Valley, Idaho, October

13 David, T J., “A Method of Improving Mechanical Seal Reliability,” Proceedings

of the Institution of Mechanical Engineers’ Fluid Machinery Ownership Costs Seminar, Manchester, England, September 16, 1992

14 Taylor, I., “The Most Persistent Pump-Application Problem for Petroleum and Power Engineers,” ASME Paper 77-Pet-5, September 1977

10-13, 1978

Chapter 3

Machinery Reliability

Audits Versus Reviews

For the purpose of this textbook, machinery reliability audit is defined as any rig- orous analysis of a vendor’s overall design after issuance of the purchase order and before commencement of equipment fabrication Reliability review is defined as a

less formal, on-going assessment of component or subsystem selection, design, exe- cution, or testing

Reliability audits tend to use outside resources for brief, concentrated efforts dur- ing the first two months after issuance of the purchase order

Reliability reviews are assigned to one or more experienced machinery engineers who would be involved in a project from the time specifications are written until the machinery leaves the vendor’s shop for shipment to the plant site

The primary purpose of the audit is to flush out deep-seated or fundamental design problems on major compressors and drivers A secondary purpose is to determine which design parameters should be subjected to non-routine computer analysis, and whether follow-up reviews should employ other than routine approaches

Machinery reliability reviews are aimed at ensuring compliance with all applica- ble specifications These reviews also judge the acceptability of certain deviations from applicable specifications Moreover, an experienced reliability review engineer will provide guidance on a host of items which either could not, or simply had not,

been specified in writing

Where to Concentrate Audit and Review Efforts

Reliable and efficient machinery is probably the most important factor ensuring profitable operation of process plants This contention becomes law in the petro- chemical industry where economic considerations often mandate the use of single, unspared machinery trains to support the entire operation of steam crackers produc- ing as much as 800,000 metric tons (-1.76 billion lbs) of ethylene per year When

82

Muchinery Reliability Audits and Reviews 83

plants in this size range experience emergency shutdowns of a few hours’ duration, flare losses alone can amount to $400,000 or more Evidently, the incentives to build reliability into the machinery installation are very high This is generally recognized

by contractors and plant owners who allocate funds and personnel to conduct relia- bility reviews before taking delivery of the machinery, during its installation, or even after the plant goes on stream

Of course, reliability assurance efforts made before delivery of the machinery are more cost effective than post-delivery or post-startup endeavors aimed toward the same goals However, the questions remain how to optimally conduct these efforts, how to man them, and which components or systems to subject to close scrutiny

This is where an analysis of available failure statistics will prove helpful A review

of the failure statistics of rotating machinery used in modern process plants will help determine where the company’s money should be spent for highest probable returns Moreover, failure statistics can often be used to determine the value of and justifica- tion for these efforts

Experience shows that a petrochemical project in the $800,000,000 range would optimally staff machinery reliability audits with four engineers for a four-month period, and machinery reliability reviews with two engineers for a period of 2-3

years The total cost of these efforts would be in the league of $800,000-$950,000 If this sounds like a lot of money, the reader may wish to contrast it with the value of a single startup delay day, say $550,000, or the cost of two unforeseen days of down- time-perhaps accompanied by the thunder of two tall flare stacks for the better por- tion of two days

Machinery reliability audits and reviews can be a tremendously worthwhile investment as long as they are performed by experienced engineers Of course, this presupposes that a perceptive project manager will see to it that the resulting recom- mendations are, in fact, implemented

Rotordynamic Design Audits*

By far the most prevalent and also most important design audit effort is focused

on turbomachinery rotordynamics It is in this area that design weaknesses can often

be spotted and appropriate changes implemented before the equipment leaves the manufacturer’s shop Large multinational petrochemical companies are sometimes staffed to handle these audits However, in most instances this audit task is entrusted

to independent consulting companies with the experience and technical resources to perform this critically important task soon after a purchase order has been issued

~ ~~ ~~~ ~~~~

~ ~ ~

“Source: J C Wachel, Engineering Dynamics Inc., San Antonio, Texas, USA Originally presented

at the 15th Turbomachinery Symposium, Texas A&M University, Corpus Christi, Texas, 1986

Reprinted by permission

84 Improving Machinery Reliability

Of course, rotordynamics design audits may be equally valuable for existing equipment with inherent design defects These defects may manifest themselves in a number of ways; they include sensitivity to unbalance and sensitivity to misalign- ment, and range all the way to frequent, unexplained downtime Moreover, retroac- tive rotordynamic design audits represent an excellent means of determining the merits of component upgrading in existing turbomachinery

On new equipment, the decision to perform a rotordynamic design audit is gener- ally based on the type of machine, the manufacturer’s experience with similar sizes,

speeds, etc., and the assessment of the benefits versus the cost of the analysis If it

could be assumed that nothing would go wrong, then the audit would not be needed However, statistics show that design and manufacturing problems do occur that result in considerable delay to projects Cook’ indicates that over half of the major

projects in the 1974-1984 time frame encountered a critical speed design problem

and/or high vibration near rated speed This study indicated that the delay time to correct design equipment error could be as high as 100 weeks For some perfor- mance-related problems, up to four years were needed to correct the difficulties Exxon Chemical Company statistics for the late 1970s and early 1980s indicated that approximately 22% of the unscheduled downtime events for major turbocom- pressors in process plants were caused by the rotor/shaft systems.2 Considering all

the unscheduled downtime causes which could be vibration-related, the percentage would be greater than 50% A study by an insurance company found that failures expected each year were about one out of every 186 for steam turbines, and one out

of every 26 for gas turbines

Data such as this and the author’s experience in troubleshooting vibration and fail- ure problems indicate that design audits can help prevent many of the problems causing unscheduled downtime, project delays, and/or failures by identifying poten- tial problem areas before manufacture

Another reason for performing an independent audit is the fact that the system may consist of used equipment In order to avoid any contractual liabilities, the man- ufacturer may not want to perform the rotordynamic calculations on the new system

or the changes that are being made

The following are major types of rotordynamic design audits that can be per- formed, and they are discussed in the following sections

1 Lateral Critical Speed Analyses

Critical speed map

Undamped natural frequencies

Undamped mode shapes

Bearing and seal stiffnesses and damping

Rotor response to unbalance

Pedestal and foundation effects on response

Machinery Reliability Audits and Reviews 85

Interference diagram

0 Coupling dynamic torques

Dynamic gear loads

0 Harmonic torque loads for reciprocating machinery

0 Allowable number off starts

4 Impeller and Blade Analyses

3 Transient Torsional Analyses

Lateral Critical Speed Analysis

The most common design audits are the lateral and torsional critical speed audits since they potentially offer the most benefits Experience indicates that many sys- tems have been installed with critical speeds in the running speed range and have run successfully for years before troubles are encountered This sometimes is difficult to

understand, but a design audit that considers the entire range of possible values for the shaft unbalances and bearing and seal parameters will usually indicate the possi- bility of a problem

A lateral critical speed audit should include these calculations:

1 Critical speed map

2 Undamped natural frequencies and mode shapes

3 Bearing stiffness and damping properties

4 Seal stiffness and damping properties

5 Rotor response to unbalance