ASM Metals HandBook Vol. 17 - Nondestructive Evaluation and Quality Control

Подождите немного. Документ загружается.

Gas factor Gas

σ (σ - 1)

[(σv/μ) - 1]

Helium 0.17

-0.83 -0.93

Argon 1.3 0.3 0.85

Carbon dioxide

1.4 0.4 0.4

Water vapor 0.9 -0.1 -0.72

Hydrocarbons 3-10

2-9 1-10

Hydrogen 0.5 -0.5 -0.94

By spraying tracer gas over the outside surface of a vacuum system, leaks can be located by observing the ionization-gage

response. The gage reading should decrease in the presence of gages having a negative gas-factor value and should

increase for gases having a positive value.

The palladium-barrier gage is a modified ionization gage that has a palladium barrier in front of the ionization

chamber. Hydrogen is used as a tracer gas because it is the only gas that will permeate heated palladium into the

ionization chamber.

Thermal-Conductivity Gages. The thermal conductivity of a gas is a function of its mean free path; therefore, gages

measuring the thermal conductivity of the gas as it passes over a heated filament will indicate vacuum. Several types of

these gages, including Pirani gages, are available (see the discussion on thermal-conductivity gages in the section "Leak

Testing of Pressure Systems Without a Tracer Gas" in this article).

Gas chromatographs can be used as leak detectors in vacuum systems by condensing in a cold trap a sample of gas

drawn from the vacuum system with a vacuum pump. The cold trap is then isolated from the vacuum system with valves,

and the frozen gas is released to the chromatograph by warming the trap. The sensitivity of this instrument is limited

essentially by the volume of sample drawn into the cold trap, the volume being proportional to the length of time the gas

is drawn into the cold trap.

Residual gas analyzers and partial pressure analyzers are similar in operating principle to mass-spectrometer leak

detectors. They have the added advantage of being able to scan the molecular peaks of all gases within the sample

entering the instrument. This provides the leak-test operator with information as to whether the sample is predominantly

leakage or contamination from system surfaces, O-rings, and cleaning agent residues. They have the disadvantage of

having no integral pump system. They must be installed behind a pump system on the system being tested in order to

obtain any kind of reasonable response time if they are to be used for leak detection and location.

Leak Testing

Revised by Gerald L. Anderson, American Gas and Chemical Company

Choosing the Optimum Leak Testing Method

The three major factors that determine the choice of leak testing method are:

• The physical characteristics of the system and the tracer fluid

• The size of the anticipated leak

• The reason for conducting the test (that is, locating or detecting a leak or measuring a leak rate)

System and Tracer-Fluid Characteristics. The physical characteristics of the system play a large role in the

selection of leak detection methods. If a tracer fluid, either gas or liquid, is used, it must be nonreactive with the test item

or system components. In general, the items or systems should be leak tested with the actual working fluid. This

eliminates any errors that might be caused by converting from tracer-fluid leakage to working-fluid leakage. It is also

possible that tracer-fluid leakage will occur while working-fluid leakage will not, and vice versa.

If tracer gas is to be used, consideration should be given to the characteristics of the gas. In most cases, gases with high

diffusion rates and small molecular size, such as hydrogen and helium, are desirable. On the other hand, when probing the

surface of a container filled with tracer gas, it may be desirable to use a persistent gas or a gas with a low diffusion rate. A

persistent gas will remain in the leak area longer and may facilitate leak detection and location because of an increase in

tracer-gas concentration.

Leak Size. The method and instrument must match or be responsive to the size of the leak. The approximate working

ranges of several leak detectors or leak detection methods are given in Fig. 15. If leakage is too great, the leak detector

will be swamped. This is normally self-evident because the detector will go to full scale and stay there until it is removed

from the leakage source. Some combustible-gas detectors, however, may return to zero in the presence of a high

concentration of leakage. For example, if the vapor-to-air ratio of gasoline vapor becomes richer than about one part

vapor to nine parts air, the mixture is no longer combustible, and the catalytic detector may indicate zero or some nominal

value, despite the fact that high concentrations of gasoline vapor exist.

Fig. 15 Approximate working ranges of various leak detectors or leak detection methods

Small leaks may also produce no response on a given detector, thus eliminating any possibility of finding the leak. It

should not be assumed, however, that the leakage is zero. There is universal agreement that the terms zero leakage and no

leakage are inaccurate. Specifications or test procedures should define leakage with such terms as minimum detectable

leakage or a maximum leak rate.

Leak Location. If leak location is the purpose of the test, methods that include the use of probes or portable detectors

are necessary so that the surface of the test vessel can be scanned. In vacuum systems, tracer gas can be sprayed over the

vessel surface and its entry point detected by observing the leak detector reaction as the tracer-gas spray is moved along

the surface. In pressure systems, bubbles, immersion, liquid penetrants, and chemical indicators provide means of locating

leaks through visual observation. Figure 16 outlines various methods of leak detection and leakage rate measurement in

terms of methods, sensitivity, and cost.

Fig. 16 Decision-tree guide for the selection of leak testing methods. Source: Ref 1

Reference cited in this section

Leak Testing, in Nondestructive Testing Handbook, R.C. M

cMaster, Ed., Vol 1, 2nd ed., American Society

for Nondestructive Testing, 1982

Leak Testing

Revised by Gerald L. Anderson, American Gas and Chemical Company

System Response

The response of a leaking system to a leak is important in leak testing; if the response is too slow, it can be very time

consuming to locate the leak, or the leak can be missed. A number of factors affect system response, including size of the

leak, diffusion rate of the tracer gas, absorption and adsorption of the tracer gas, cleanup time of the detector, pumping

speed, and volume of the system. Attention must be given to these factors to minimize response time.

Leak Size. Theoretically, leak size does not affect system response time; it only affects the magnitude of the detector

readout. However, if some type of leakage accumulation (or mass loss) method is being employed, leak size will affect

response time because sufficient leakage must be accumulated (or lost) before the detector will respond.

The smallest leakage to which the detector will respond is termed the minimum detectable leakage and is measured by

observing the leak detector reaction to a calibrated or known leakage. When calculating the minimum detectable leakage

of a mass spectrometer, it is conventional practice that the noise and random-background count rate be doubled. The

sensitivity of the mass-spectrometer detector is defined by:

(Eq 11)

where S is the sensitivity (in atmosphere cubic centimeters per second per scale division), D is the total deflection

produced by the calibrated leak (in scale divisions), Q is the calibrated leakage flow (in atmosphere cubic centimeters per

second), and B

is the steady background reading (in scale divisions). In terms of S, the minimum detectable leakage

(MDL) can be expressed as:

MDL = 2(N + B

)(S)

(Eq 12)

where N is the scale deflection produced by instability or noise in system (in scale divisions) and B

is the random

background deflection (in scale divisions).

Tracer-Gas Diffusion Rate. In large, open vessels, gas diffusion is not a major problem. However, in tortuous

systems, one investigator found that even after 24 h neither helium nor Freon had evenly diffused throughout the system.

Absorption and adsorption of tracer gas are responsible for hang-up in the system; they can also materially

reduce tracer-gas concentration. Permeation also causes difficulties; for example, helium permeates through a 25 mm (1

in.) diam O-ring in about 1 h, thus producing a detectable reading in the system.

Detector cleanup time (the time required for the leak detector to purge itself of tracer gas) can range from a few

seconds to several minutes, thus affecting the time between searches for a leak. Cleanup time is usually defined as the

time it takes for the meter reading to decay to 37% of its maximum value.

Pumping Speed. It can be shown that a leak detector will exhibit a readout deflection that is 63% of its maximum

value at some characteristic time, t, known as the response time. Response time can be calculated as t = V/S, where V is

the system volume in liters and S is the pumping speed in liters per second. Increasing the pumping speed will decrease

the response time, but will introduce a corresponding decrease in the maximum signal. Increasing the pumping speed by a

factor of ten will reduce the readout deflection of the detector by a factor of ten, which may cause problems. Calculation

of the response time is useful for determining the length of time that a tracer gas must be sprayed over the external surface

of a vessel under vacuum in order to detect a leak. For example, if the vessel has a volume 1000 L and the leak detector

has a pumping speed of 5 L/s, then the suspected leak area should be sprayed with search gas for about 200 s (1000/5) for

the detector to reach 63% of its maximum value.

System Volume. The larger the volume of the system, the smaller the concentration of leakage. This can be alleviated

by closing off part of the system when possible. The enclosures used to trap the leakage should be as small as possible.

For example, if the leakage of tracer gas is 10

-5

atm cm

, if the enclosure volume minus the test unit volume is 10

and if the detector can detect 1 ppm of search gas, then the tracer gas will take about 28 h to reach detectable

concentration.

Leak Testing

Revised by Gerald L. Anderson, American Gas and Chemical Company

Measurement of Leak Rate Using Calibrated Leaks

Leak-rate measurement can be accomplished with almost any of the standard leak detecting methods in conjunction with

a calibrated or known leak. If the leakage is inward, it can be measured by tracer-gas concentration inside the vessel or

system vacuum. If the leakage in outward, it can be collected in an enclosure and the concentration rise within the

enclosure monitored; it can also be measured by placing the leak detector probe directly over the leak and isolating the

surrounding area with a suction cup. In addition, leaks can be measured by mass change methods.

To calibrate a leak testing instrument to measure leakage rates, its response to a known leakage must be found.

Furthermore, the instrument should be calibrated with the smallest leakage to which it will respond, preferably using the

fluid that will be utilized during testing, because most leak detectors respond differently to different fluids. Leak detectors

for measuring gross changes in mass or flow can be calibrated in a conventional manner by using dead-weight testers,

nozzles, orifices, and manometers. The more sensitive gas or tracer-gas detectors can be calibrated by injecting known

amounts of tracer gas into the system or by using standard or calibrated leaks.

A standard leak can be made by constructing a device with a small orifice or nozzle; by maintaining constant upstream

pressure, the flow will be constant. Another method is to maintain the upstream pressure at a level high enough to

produce flow of sonic velocity at the nozzle throat, thus making the leakage independent of upstream pressure. Very small

calibrated leaks are produced by flow through a capillary tube or by permeation through a membrane (normally glass).

The capillary type may or may not have a self-contained gas supply. There is an advantage to capillary-type leaks not

having a self-contained gas supply, because a variety of gases can be used with them, although a constant tracer-gas

pressure must be supplied. Capillary tubes may be purchased with leak rates ranging from about 10

-2

to 10

-7

atm cm

/s.

Specific leak sizes are also available; these calibrated leaks can normally be supplied to match any specified leakage

within about 20%. Permeation-type standard leaks are supplied with their own self-contained gas supply, usually helium.

Although the flow through the membrane is molecular and therefore pressure dependent, typical helium loss is only 10%

in 10 years. These leaks have a flow in the range of 10

-6

to 10

-10

atm cm

/s.

Leak Testing

Revised by Gerald L. Anderson, American Gas and Chemical Company

Common Errors in Leak Testing

The most common errors encountered in leak testing are:

• Use of a method that is too sensitive

• Use of only one leak-testing method

• Failure to control the environment in the vicinity of the leak

Because sensitive methods cannot be used in the presence of gross leaks, leak testing should be conducted in two or more

stages, beginning with gross testing and progressing to the more sensitive methods. This procedure saves both time

money.

Control of environment in the vicinity of the leak consists of a number of steps, depending on the test method employed.

If a tracer gas is being used, the atmosphere must be controlled to prevent contamination by the tracer gas that might be

erroneously reported as a leak. If sonic or ultrasonic methods are being used, background noise must be controlled. If

small leaks are present, airborne contaminants may cause temporary plugging of the leak. From example, breathing on a

molecular leak may cause it to plug due to water vapor in the breath. Particulate matter in the air may plug or partially

plug a leak. Table 5 lists the differences in leak tightness requirements for a variety of industrial products.

Table 5 Leak tightness requirements for various products

Leak testing term

Vessel or part

(material

contained)

Minimum detectable

leakage rate,

std cm

Large Truck (gravel) 1 × 10

Hourglass (sand) 1 × 10

Car window (air) 1 × 10

-1

Tuck (oil) 1 × 10

-2

Bucket (water) 1 × 10

-3

Gross

Storage tank (gasoline) 1 × 10

-4

Pipeline (gas) 1 × 10

-5

Small

Tanker (liquified natural gas)

1 × 10

-6

Storage tank (NH

) 1 × 10

-8

Fine

Heart pacemaker (gas) 1 × 10

-11

Leak Testing

Revised by Gerald L. Anderson, American Gas and Chemical Company

References

1. Leak Testing, in Nondestructive Testing Handbook,

R.C. McMaster, Ed., Vol 1, 2nd ed., American Society

for Nondestructive Testing, 1982

2. Encyclopedia of Materials Science and Engineering, Vol 4, MIT Press, 1986

Liquid Penetrant Inspection

Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Introduction

LIQUID PENETRANT INSPECTION is a nondestructive method of revealing discontinuities that are open to the

surfaces of solid and essentially nonporous materials. Indications of a wide spectrum of flaw sizes can be found regardless

of the configuration of the workpiece and regardless of flaw orientations. Liquid penetrants seep into various types of

minute surface openings by capillary action. Because of this, the process is well suited to the detection of all types of

surface cracks, laps, porosity, shrinkage areas, laminations, and similar discontinuities. It is extensively used for the

inspection of wrought and cast products of both ferrous and nonferrous metals, powder metallurgy parts, ceramics,

plastics, and glass objects.

In practice, the liquid penetrant process is relatively simple to utilize and control. The equipment used in liquid penetrant

inspection can vary from an arrangement of simple tanks containing penetrant, emulsifier, and developer to sophisticated

computer-controlled automated processing and inspection systems. Establishing procedures and standards for the

inspection of specific parts or products is critical for optimum end results.

The liquid penetrant method does not depend on ferromagnetism (as does, for example, magnetic particle inspection), and

the arrangement of the discontinuities is not a factor. The penetrant method is effective not only for detecting surface

flaws in non-magnetic metals but also for revealing surface flaws in a variety of other nonmagnetic materials. Liquid

penetrant inspection is also used to inspect items made from ferromagnetic steels; generally, its sensitivity is greater than

that of magnetic particle inspection.

The major limitation of liquid penetrant inspection is that it can detect only imperfections that are open to the surface;

some other method must be used for detecting subsurface flaws. Another factor that can limit the use of liquid penetrants

is surface roughness or porosity. Such surfaces produce excessive background and interfere with inspection.

Liquid Penetrant Inspection

Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Physical Principles

Liquid penetrant inspection depends mainly on a penetrant's effectively wetting the surface of a solid workpiece or

specimen, flowing over that surface to form a continuous and reasonably uniform coating, and then migrating into cavities

that are open to the surface. The cavities of interest are usually exceedingly small, often invisible to the unaided eye. The

ability of a given liquid to flow over a surface and enter surface cavities depends principally on the following:

• Cleanliness of the surface

• Configuration of the cavity

• Cleanliness of the cavity

• Size of surface opening of the cavity

• Surface tension of the liquid

• Ability of the liquid to wet the surface

• Contact angle of the liquid

The cohesive forces between molecules of a liquid cause surface tension. An example of the influence of surface tension

is the tendency of free liquid, such as a droplet of water, to contract into a sphere. In such a droplet, surface tension is

counterbalanced by the internal hydrostatic pressure of the liquid. When the liquid comes into contact with a solid

surface, the cohesive force responsible for surface tension competes with the adhesive force between the molecules of the

liquid and the solid surface. These forces jointly determine the contact angle, θ, between the liquid and the surface (Fig.

1). If θ is less than 90° (Fig. 1a), the liquid is said to wet the surface, or to have good wetting ability; if the angle is equal

to or greater than 90° (Fig. 1b and c), the wetting ability is considered poor.

Fig. 1 Wetting characteristics as evaluated by the angle, θ

, between a droplet of liquid and a solid surface.

Good wetting is obtained when θ< 90° (a); poor wetting, when θ 90° (b) and (c).

Closely related to wetting ability is the phenomenon of capillary rise or depression (Fig. 2). If the contact angle, θ,

between the liquid and the wall of the capillary tube is less than 90° (that is, if the liquid wets the tube wall), the liquid

meniscus in the tube is concave, and the liquid rises in the tube (Fig. 2a). If θ is equal to 90°, there is no capillary

depression or rise (Fig. 2b). If θ is greater than 90°, the liquid is depressed in the tube and does not wet the tube wall, and

the meniscus is convex (Fig. 2c). In capillary rise (Fig. 2a), the meniscus does not pull the liquid up the tube; rather, the

hydrostatic pressure immediately under the meniscus is reduced by the distribution of the surface tension in the concave

surface, and the liquid is pushed up the capillary tube by the hydraulically transmitted pressure of the atmosphere at the

free surface of the liquid outside the capillary tube. Figure 3 clearly shows the forces that cause liquid to rise in a capillary

tube.

Fig. 2 Rise or depression in small vertical capillary tubes, determined by the contact angle θ

, between a liquid

and the wall of a capillary tube. (a) θ< 90°: capillary rise. (b) θ= 90°: no capillary depression or rise. (c) θ

90°: capillary depression

Fig. 3 Schematic showing the forces involved in capi

llary rise: the downward force from weight of the liquid

column, and the upward force from surface tension along the meniscus perimeter

The height to which the liquid rises is directly proportional to the surface tension of the liquid and to the cosine of the

angle of contact, and it is inversely proportional to the density of the liquid and to the radius of the capillary tube. If, as in

Fig. 4, the capillary tube is closed, a wetting liquid will still rise in the tube; however, there will be extra pressure

resulting from the air and vapor compressed in the closed end of the tube, and the capillary rise will not be as great.

Fig. 4

Schematic showing that the rise and depression of liquids in closed capillary tubes are affected by the

compressed air entrapped in the closed end. Compare height of liquid in tube shown in Fig. 3

with that in the

tube here.

These examples of surface wetting and capillary rise illustrate the basic physical principles by which a penetrant enters

fine surface discontinuities even though the practical circumstances encountered in the use of liquid penetrants are more

complex than these examples may suggest. Cracks, for example, are not capillary tubes, but simulate the basic interaction

between a liquid and a solid surface, which is responsible for the migration of penetrants into fine surface cracks.

The viscosity of the liquid is not a factor in the basic equation of capillary rise. Viscosity is related to the rate at which a

liquid will flow under some applied unbalanced stress; in itself, viscosity has a negligible effect on penetrating ability. In

general, however, very viscous liquids are unsuitable as penetrants because they do not flow rapidly enough over the

surface of the workpiece; consequently, they require excessively long periods of time to migrate into fine flaws.

Another necessary property of a penetrant is its capability of dissolving an adequate amount of suitable fluorescent or

visible-dye compounds. Finally, the penetrant liquid must be removable with a suitable solvent/remover or emulsifier

without precipitating the dye.

Just as it is important that a penetrant enter surface flaws, it is also important that the penetrant be retained in the flaws

and emerge from the flaw after the superficial coating is removed from the surface and the developer is applied. It is

apparently a paradox that the same interaction between a liquid and a surface that causes the liquid to enter a fine opening

is also responsible for its emergence from that opening. The resolution of the paradox is simple: Once the surface has

been freed of excess penetrant it becomes accessible to the entrapped liquid, which, under the effect of the adhesive forces

between liquid and solid, spreads over the newly cleaned surface until an equilibrium distribution is attained (Fig. 5).