Baker R.C. Flow Measurement Handbook: Industrial Designs, Operating Principles, Performance, and Applications

Подождите немного. Документ загружается.

3.3 NOTES ON THE SPECIFICATION FORM 45

1.0

0.1

0.01

0.001

0.1

0.01

0.001

\inlp

0.1

10 100

(a)

(b)

Figure 3.1. Friction variation (from Summers-Smith 1994; reproduced with permission of the

author and Professional Engineering Publishing): (a) Stribeck diagram for a journal bearing;

(b) Coefficient of friction as a function of specific film thickness.

C) where the clearance is slightly larger than the surface roughness dimension, and

the rapid increase in friction coefficient with decreasing viscosity occurs (region B)

due to the smallness of the clearance allowing interference between the roughness

on the opposing faces.

It, therefore, seems likely that low lubricity has been used to refer to situations

where

/Jin/p is

very small, where

is fluid viscosity,

n is

rotational speed, and p is load.

Many gases will require external lubrication in turbine meters, whereas a bearingless

flowmeter may be suitable for hydrofluoric acid and very dry liquids such as carbon

disulfide.

Some gases (e.g., oxygen) require special lubrication. Because it oxidizes steel

pipes with the danger of chunks of rust detaching from the pipe wall, a maximum

velocity is sometimes specified for oxygen flows.

Abrasiveness caused by solids in the flow will clearly be a problem for bearings

and will lead to wear unless the particulate matter is filtered out.

D SITE DETAILS

Much of the data requested would be clarified and reinforced by a site plan. Details

of access and environmental hazards are clearly essential. The access may be such

that the meter has to be man-handled into position, placing severe limitations on

weight.

E ELECTRICAL AND CONTROL DETAILS

Whether the proposed instrument becomes part of

sophisticated modern telemetry

system or is a stand-alone with local power - not to mention all the intermediate

variants - will make a lot of difference to meter specification. In addition, all safety

details for the site must be set out either here or under environmental hazards in D.

Appropriate documents on safety should be consulted.

F OTHER USER INFORMATION

In specifying accuracy, it is essential that the user does not overstate requirements or

that the manufacturer does not oversell the product. The accuracy should be realistic

when the total installation constraints are taken into account. Response time should

46 SPECIFICATION, SELECTION, AND AUDIT

be appropriate to the requirements of the plant. For a plant with, say, a dosing period

of 10 s, the response will need to be rapid enough to achieve the required dosing

precision. The instrument identified by the manufacturer should fit in with existing

plant schedules as far as possible and should be compatible with extreme conditions

(e.g., steam cleaning of hygienic process plant).

G MANUFACTURER'S PROPOSED METER AND REQUIREMENTS

This section allows the manufacturer to specify the meter that most nearly meets

the user's requirements. The manufacturer should make clear any ways in which the

meter will not meet the specification. The details of the calibration procedure for

the flowmeter should be spelled out.

The form may well have been supplied to several manufacturers, and this section

will provide the user with a comparison of the possibilities.

H ANY OTHER CONSIDERATIONS AND CONCLUSIONS

This is an opportunity for the manufacturer to add further information and for the

user to summarize the decision on selection.

3.4 FLOWMETER SELECTION SUMMARY TABLES

Table 3.1 is an update of the table in Baker (1988/9) resulting from the fuller infor-

mation included in this book. It lists the main types of flowmeter as set out in the

chapters of this book and attempts to encapsulate the key features.

For convenience, they are subdivided into momentum-, volume-, and mass-

sensing meters. This subdivision also emphasizes that, for instance, a meter respond-

ing to momentum will be sensitive to density change. Where a less common meter

of the same family has particular features that make it unique or particularly use-

ful in certain applications, these may be indicated in notes in the table and in the

discussion in subsequent chapters, where other points arising from this chapter will

be elaborated. This table, therefore, provides a first cut at the selection for the task

which has been identified by the specification questionnaire. The values given are

my best estimates at the time of writing, but they probably fail to reflect the inherent

variation in designs.

Nature of the Fluid to Be Metered

The first choice, liquid or gas, will cause a few types of meters to be eliminated. The

selection of suitable meters for slurries and two (or more)-phase flows, is severely

restricted at present. In Table 3.1, X indicates unsuitable, and * indicates suitable

except where some are more suitable (**) than others. A ? indicates that it is not

cohimon but might be appropriate in this application. The electromagnetic flow-

meter has found a valuable role in slurry flow measurement. However, the Coriolis

mass flowmeter is sometimes claimed to be suitable for slurry, and an ultrasonic flow-

meter has been specially designed for this purpose. The electromagnetic flowmeter is

also useful in two-phase flows when the continuous phase is conducting. However,

Venturis are usable, but with caution. The ultrasonic correlation flowmeter and the

doppler flow monitor may also be worth considering in such flows.

3.4 FLOWMETER SELECTION SUMMARY TABLES 47

Flowmeters with rotating parts will be less suitable for dirty abrasive fluids or for

fluids that do not lubricate adequately. Abrasion may also detract from the perfor-

mance of meters, such as the vortex meter. Cavitation should be avoided. The liquid

must be conducting for current commercial electromagnetic flowmeters.

Flowmeter Range and Specification

Accuracy is perhaps the most difficult to determine because both user and manu-

facturer are prone to overstate their requirements and capabilities. In Table 3.1,

***is for random error <

0.1%

rate,

**is for random error < 0.5% rate, and

*is for random error < 2% rate.

The estimates reflect my view of the accuracy of flowmeters. These values should

be achievable with a particular type of flowmeter. For a standard orifice the uncer-

tainty before calibration due to the discharge coefficient is 0.5-0.75%, and overall

uncertainty will be greater than this when errors in the measurement of dimensions,

pressure, and density are included. After calibration, the total uncertainty will be less

than this, and it is reasonable to assume that the random error will be less than 0.5%.

This appears to be confirmed by data. Some manufacturers may feel able to claim

higher values for the various meters than those in the table, but the reader should

ask for justification for such claims. Commercial flowmeter claims of better than

0.5%

indicate a level of accuracy that is probably only achieved in a few designs.

Some manufacturers may not always achieve the values indicated. The repeatability

of a variable area meter is often quoted as a percentage of full scale, and so may be

just outside the

rating for flow rates over, say, the top half of the range.

Precision relates to random error. It does not allow for changes in bias error

caused by changes in method of measurement, location, or long elapsed times. Thus,

if the reader is concerned about long-term repeatability or total uncertainty, these

estimates may not provide the most appropriate guide.

Diameter, temperature, and flow are my best estimates based on manufacturers'

literature. In all

cases,

the diligent reader can probably find examples of greater ranges

than those given, and manufacturers are always seeking to extend the capability of

their instruments. Note that, in the column for flow range, the values in parentheses

indicate kilograms per hour rather than cubic meters per hour.

The maximum pressure has not been included because, apart from exceptionally

high pressure applications, when one might turn to an orifice plate or a turbine me-

ter, there appear to be versions of most designs that are offered for normal industrial

ranges. In some cases, minimum pressures maybe important because cavitation may

be a danger, and the manufacturer should be consulted.

The pressure loss is given as H (high), M (medium), or

(low). Only meters with

a clear bore have been categorized as

because there is virtually no more pressure

loss than in a similar length of pipe. The orifice plate meter is H/M, and positive

displacement meter and vortex are H/M, although their relative losses will depend

on such things as the instrumentation being driven by the positive displacement

meter.

Table 3.1. Flowmeter selection summary tables

Flowmeter

type

Momentum"

Orifice

Venturi

Nozzles

Variable-area

Other

Venturi-like

Averaging

pitots

Laminar

Volume

Positive

displacement

Turbine

Vortex

Electromagnetic

Transit-time

U.S.

Mass

Multisensor

Wheatstone

bridge

Thermal

Angular mom.

Coriolis

Liquid

Gas

Key:

***,

Very high; **,

3 4 5 6

Other Typical

Slurry Two-Phase Accuracy Turndown'*

? ** 5:1

5:1

* ** 5:1

5:1

X 10:1

10:1

? ? 3:1

3:1

X ? 10:1

10:1

XX ? 20:1

b .*.

10;1

***

80:1

X *** 10:1

***

30:1

XX ** 20:1

20:1

** •• 100:1

•fb

-fb «• 20:1

30:1

XX ** 50:1

X * 15:1

50:1

XX * 7:1

X ** 100:1

* *

More suitable/high; *, Suitable/medium

Dia

Range (mm)

50-1,000

50-1,200

50-630

15-150

13-1,200

25-12,000

4-200

5-600

25-600

12-200

2-3,000

10-2,000

20-1,000

6-60

2-6

6-200

20-50

6-200

Temp

Range

(°C)

-200 to 350

up to 450

-100 to 450

-50 to 290

-40 to 65

-265 to 310

-10 to 50

-200 to 400

-50 to 190

-200 to 260

-50 to 260

-18 to 150

0to65

-50 to 300

-40 to 150

-240 to 200

Flow Ranged

/h (kg*))

1-3 x10

10-4 x 10

30-7,000

400-10

2-1.7 x 10

20-2.5

io-

-ioo

10-

-2,000

10-3 x 10"

200-6

up to 120

0.01-2,000

0.01-1,200

0.03-7000

0.01-25,000

3-2,000

50-10

10"

-10

3-10

0.04-10

(0.05-2.3

x 10

)

(0.002-100)

10~

-8000)

(100-4.5

x 10

)

(1-7

)

10 11

Pressure Sensitivity to

Loss

H/M

L/M

M/L

;

?/blank,

Uncertain/lower; X. Unsuitable.

Some proprietary devices offer special features: higher differential pressure

Some designs have been produced for slurries.

Multisensor systems have been developed specially for multiphase.

Typical estimates that

Larger or smaller sizes

linear characteristics.

Installation*

H/M

L/M

H/M

L(CTMF)

M/H(ITMF)

rmmmm

Initial

Cost

L/M

M/H

H/M

L/M

M/H

Notes

Concentric ISO orifice

with differential pressure

cell assumed

Does not include critical

nozzle.

Glass/plastic assumed;

higher ratings

for

steel.

Consult manufacturers.

Particularly appropriate

for pulsating flows.

Turndown given may be

achieved for rotary positive

displacement gas meter.

High precision instruments

assumed rather than

water meters, etc.

Fluidic flowmeter suitable

for

wide range of utility flows

Only available commercially

for conducting liquids.

Single beam are more

sensitive to installation.

Correlation and doppler

for two-phase flows.

Specialist instrument

especially

for

engine

testing.

Higher

flows

require CTFM

with bypass or ITMF.

Particularly suitable

for aircraft fuel flow.

Straight single tube has

essentially zero pressure

drop.

Gas ranges are

probably more limited.

Low;

M = Medium; H = High.

may be exceeded by using smart transducers and may be greater or less than the value attained by some designs.

may be available.

f Other ranges may be available.

Flow conditioning may be used in some applications to reduce this effect.

50 51

Table 3.2. Rough selection criteria

for

liquid and gas flowmeters

'•

Property of liquid

Clean

•••••••••

Slightly contaminated

•••••••••

Contaminated or abrasive

•••••••••

Heavily contaminated

•••••••••

Viscosity below 10 cSt

••••••••••••

Viscosity 10 to 40 cSt

••••••••••••

Viscosity over 40 cSt

••••••••••••

Diameter

<25 mm (<1 inch) ••• ••• ••• •••

25 to 500 mm (1 to 20 inches) ••• ••• ••• •••

• >

500 mm (>

inches) ••• ••• ••• •••

Type of flowmeter

Magnetic"

••//•//////•••/•••••////•/•••••••///

Mass flowmeter

////// //•••/

oooooo

Vortex • o / O / o / O • o / O

Ultrasonic transit time

//////

Differential pressure 000//////000//////000000000

Positive displacement

//////

Variable area //// ////

Turbines

////// /•••••

Property of gas

Clean

•••••••••

Slightly contaminated

•••••••••

Contaminated

•••••••••

High humidity

•••••••••

Density >60 kg/m

(3.75 lb/ft

)

••••••••••••

Density 1 to 60 kg/m

••••••••••••

(0.06 to 3.75 lb/ft

)

Density <lkg/m

(0.06 lb/ft

)

••••••••••••

Diameter

> <25 mm (<1 inch) ••• • • • ••• •••

25 to 500 mm (1 to 20 inches) ••• ••• ••• •••

>500 mm (>20 inches) ••• ••• ••• •••

Type of flowmeter

Vortex* / / o / / o / / o / / o

•

Differential pressure 000//////000//////000000000000000000

Positive displacement

//////

Variable area

/•//// //////

Turbines //0//0 //0//0 OOOO

•

Reproduced from Endress et al. (1989) with permission.

i Key: o = only under special conditions (e.g., special version, only certain diameters); • = Property of material;

•"

= suitable.

" Only if liquid is electrically conductive.

;

Suitable for steam measurement.

52 SPECIFICATION, SELECTION, AND AUDIT

Installation effects caused by upstream pipework will be a recurring theme in

most chapters. The positive displacement meter may be assumed to be insensitive to

upstream installation effects and is, therefore, rated L. A few others, also, have low

sensitivity to installation. However, most flowmeters are affected, and these have

been categorized for a fitting, such as a bend, at 5D upstream.

L, negligible effect;

M, <2% increase in uncertainty;

H, >2% increase in uncertainty.

These categorizations are my best estimates, and manufacturers may be prepared to

uprate them.

Most flowmeters are sensitive to pulsation over part of their range, with the

possible exception of the ultrasonic flowmeter and the laminar flowmeter, which

was specially designed to cope with pulsating gas flows.

The flowmeters with essentially a clear straight bore (nonintrusive) are the elec-

tromagnetic, ultrasonic, and some Coriolis designs. The only flowmeter that may

fail and block the line is the positive displacement flowmeter. All the others have

partial line blockage, although the venturi has such a smooth change in section that

its effect is probably of small concern.

Only the ultrasonic family offers clamp-on options (noninvasive) at present.

Response time will depend on fluid inertia, component inertia, or electrical

damping. The first two are inherent to the meter type and are not reducible. The

electrical response may be speeded up in some meters provided stability and noise

permit.

The installation of the flowmeter will also be affected by environmental

considerations, such as ambient temperature, humidity, exposure to weather, elec-

tromagnetic radiation, and vibration. These will need to be assessed in consultation

with particular manufacturers. The area safety and explosion classification should

also be checked.

Price and Cost

This is possibly the most important consideration and is difficult to tabulate because

the range of price for any design is wide and the data are always changing! Therefore,

I have attempted to rate the initial cost of the flowmeters as H (high), M (medium),

or L (low). However, the medium bracket tends to become a "catch-all" and covers,

for instance, electromagnetic and turbine flowmeters, which can actually have an

initial cost range of 4:1 for a 100-mm diameter design. For this reason, L/M or M/H

have been used where low or high cost versions are available.

Initial cost will, of course, include both purchase price and cost of installation.

There will also be ongoing costs associated with maintenance, energy loss due to the

presence of the flowmeter, and savings based on the information provided by the

flowmeter. If the flowmeter is clearly required, then the savings should outweigh the

costs.

This does not reduce the responsibility of the user to install the best and most

economical instrument for the job!

Table 3.2 is reproduced with permission from Endress et al. (1989) and gives a

rough guide to the flowmeters most likely to be useful, but starting with the nature

of the fluid and the pipe size.

3.5 OTHER GUIDES TO SELECTION AND SPECIFIC APPLICATIONS 53

3.5 OTHER GUIDES TO SELECTION AND SPECIFIC APPLICATIONS

General Selection Advice

Apart from my introductory guide (Baker 1988/9), the reader is referred to the British

Standard Guide (BS 7405:1991) to the Selection and Application of Flowmeters for a

review of each type of flowmeter and a guide to selection (cf. Furness's 1991 review of

BS 7405). Miller (1996) offers some guidance on selection, as does Hayward (1979),

which is still good value, even if a bit dated. The Institute of Petroleum guide (IoP

1982) gives a comprehensive list of information required.

Furness and Heritage (1989) gave a breakdown of considerations when selecting

a flowmeter, similar to those in the British Standard. In particular, they have used

the 11 categories of meter type set out in the Standard, although slightly modified

to include open channel flow measurement. This is an interesting subdivision, but

it is not one I have chosen to use in this book.

useful aspect of the directory is its

ability to provide the names of all flowmeter manufacturers. However, the names

change with great rapidity in this field, as company buy-outs take place.

A review of flowmeters by Ginesi (1991) covers well-trodden ground, but it also

is a reminder that some less common meters (e.g., the segmental wedge flowmeter

or the V-cone) may offer solutions.

Refinery Flow Measurement

Furness and Jelffs (1991) sought to make a case for using meters to measure flows

in and out of refineries rather than depending on tank transfer. They also mention

briefly some problems in flow metering.

• For gas metering, the gas often contains liquid.

• Provided there are no liquid particles, the orifice plate should retain its sharpness,

although they suggest that these meters seldom achieve an accuracy of better

than 4%.

• Flare gas metering poses problems of large turndown.

• Rotating meters are usually unsuitable for naphtha because external lubrication

is necessary.

• Black oils require positive displacement meters with heavy-duty rotors and large

clearances, possibly with heat tracing.

• On-line density measurement (they claimed) had a minimum uncertainty of

±0.3-0.5%.

• Most Coriolis meters (they claimed) are influenced by density in a way that can

be checked only by proving gravimetrically, and the measurement within the

meter of density is not sufficiently accurate to use a compact prover to obtain a

±0.2%

requirement.

They also refer to Shewhart Control Charts, which plot meter factor against

time,

with limits to indicate change that requires action by the plant

staff;

Moving

Average Charts, which detect long-term meter factor drift; and Cumulative Sum

(CUSUM) Charts, which indicate the occurrence of an event due to some assignable

cause.

54 SPECIFICATION, SELECTION, AND AUDIT

Water Industry

Fowles (1994) commented that the selection

an instrument

for

a required duty is a

very complex decision-making process.

the wrong selection

made,

it can

result

in abortive cost, erroneous information,

and

incoherent operational decisions.

addressing these problems, the user is often directed

American, British, European,

and international standards. However, such specifications are often

not

suited to

the

specific

and

often unique demands

the industry.

Reynolds (1994) discussed water industry specifications that have

led to

data

sheets

electromagnetic, ultrasonic, differential pressure, insertion flowmeters,

and ultrasonic level instruments.

The

data sheets consist, first,

technical specifi-

cation and option selection

be completed by the purchaser and, second, supplier-

provided information that will influence cost

ownership. A European

set of in-

strument specifications

also being established.

Nuclear Industry

Finlayson (1992) gave information

the use

flowmeters

one plant, which had

1,500 flow measurements:

For Radioactive For Nonradioactive

Type Liquid

Gas

Liquid

Gas

Electromagnetic

/ /

Ultrasonic

/ /

Differential pressure

/ / / /

Variable area

/ /

Turbine

/ / /

Positive displacement

For differential pressure,

the

paper gave diagrams

traps, vents, etc.,

and

listed

number

considerations.

• Stainless steel should

used

for

corrosive fluids

and

when behind shielding.

• All welded construction should

used

avoid leakage.

• All vents should

routed

to a

suitable drain

for

safety.

• Drain legs should take

the

shortest route and be valved before penetrating plant

shielding.

• Sensors should

housed

in a

shield bulge

or a

glove box.

• Drain cocks within

shield bulge should have extended spindles

to the

outside

of the bulge.

• Transmitters should

located outside

the

bulge

glove

box in a

lockable

en-

closure

equipment room.

The paper described some

of the

flowmeters

and

modifications needed

for the ap-

plication, including

time-of-flight ultrasonic one.

Cryogenic Applications

Rivetti

and

Martini (1990) reviewed various types

flowmeters

for

space applica-

tions and concluded that only

in a

few cases were they suitable

for

cryogenic service

3.7 FINAL COMMENTS 55

and that, even then, care needed to be taken over the use of materials and the sen-

sors.

The use of superfluid helium for space applications required new approaches.

They discussed specifically designed transducers of small size and weight with good

accuracy and reliability. Thermal transducers and turbine meters that used super-

conducting magnet suspension for the bearings appeared to be the most promising,

although much work still remains to be done.

3.6 DRAFT QUESTIONNAIRE FOR FLOWMETER AUDIT

In this chapter we have taken a detailed look at the questions that a user and a manu-

facturer need to answer before selecting a flowmeter for a particular application. We

have then looked at two selection tables, which do little more than indicate the

range of possible alternatives for most applications. We have also reviewed some

published guidance.

This final section sets out the essential information needed if the flowmeter in-

stalled is to be audited in the future. The objectives of such an audit should probably

establish one or more of the following:

• The suitability of the flowmeter for the task,

• The current state of the meter,

• The optimum specification for a new flowmeter, and

• The suitability of the flowmeter

;

s position for precise measurement and main-

tenance.

If in situ calibration is to be used, then the options need to be considered. The list

in Appendix 3.A.2, based on Baker and Smith (1990), offers a starting point for the

flowmeter owner. The list should be used either with the specification questionnaire

if such exists or to create such a questionnaire for future meter actions.

3.7 FINAL COMMENTS

To the company that uses flowmeters, note that the results of auditing flowmeters

may vary widely. The replacement of meters because of

age,

wear, or incorrect sizing

may result in more precise measurements provided installation constraints allow it.

However, this may not mean that the user reduces costs because correct measurement

may mean higher costs. One is tempted to suggest that, for statistical reasons, the

overall gain will be hardly worth the cost of the audit and replacement exercise. Of

course, the need for greater accuracy will exist if an absolute flow rate is needed for

control or fiscal reasons.

To the person responsible for purchase, make sure that you have specified the

requirements carefully, and do not be content until you are sure that you have found

a commercial device that matches your requirements in all points.

To the manufacturer, make sure that you have a clear understanding of the

following:

a. what you are offering and why it is superior to your competitors' designs;

56 SPECIFICATION, SELECTION, AND AUDIT

whether you expect to sell your products because you can offer a complete

range for

95%

of applications, or because you have the only device for particular

applications;

c. the weaknesses in your products that need to be addressed and the gaps in your

range that need to be filled.

To the research center, identify the applications that have not been met and the

technology most likely to meet them.

To the inventor, start from the application problem, look for features that may be

inherently flow-dependent, and see if they can be used as the basis for a flowmeter.

Remember that the power of modern signal processing should allow the mechanical

design to be very simple.

The development of an expert system to satisfy all parties would clearly be valu-

able,

but its production and updating may cost too much to make it worthwhile. I

should like to prove myself wrong about this one day.

APPENDIX

3.A

Specification and Audit Questionnaires

3.A.1 SPECIFICATION QUESTIONNAIRE

Response

A Information about the manufacturer

Al Manufacturer and address

A2 Contact person and position in company

A3 Company size and international standing

A4 Flow product range and turnover

A5 Other products

A6 Accreditation (BS 5750, ISO 9000, UK

NAMAS,

etc.)

B User's business and specific application

Bl Company's site address

B2 Contact person and position in company

B3 General description of application and telemetry links

B4 Site safety requirements

B5 Current general policy on instrument maintenance

C Fluid details

Cl Chemical composition, state (gas or liquid), and materials that

are incompatible with the chemical

C2 Purity of chemical and if contaminated or multicomponent,

the nature of other components, likelihood of

deposition

C3 Commonly known properties

C3.1 Density

C3.2 Viscosity (if non-Newtonian, state type)

C3.3 Conductivity