UNCLASSIFIED CENTRALDO

H LIBRARYELECTION

ORNL-2053

Metallurgy and CeramicsTID-4500 (llth ed.) Cff. #

A LABORATORY FOR THE HIGH-TEMPERATURE

CREEP TESTING OF METALS AND ALLOYS

IN CONTROLLED ENVIRONMENTS

D. A. Douglas filW. D. Manly

~"CENTRAL RESEARCH LIBRARYDOCUMENT COLLECTION

LIBRARY LOAN COPY

DO NOT TRANSFER TO ANOTHER PERSON

If you wish someone else to see this document,send in name with document and the library willarrange a loan.

OAK RIDGE NATIONAL LABORATORYOPERATED BY

UNION CARBIDE NUCLEAR COMPANYA Division of Union Carbide and Carbon Corporation

POST OFFICE BOX P • OAK RIDGE, TENNESSEE

UNCLASSIFIED

Printed in USA. Price30

cents. Available from the

Office of Technical Services

U. S, Department of Commerce

Washington 25, D. C.

LEGAL NOTICE

This report was prepared as on account of Government sponsored work. Neither the United States,

nor the Commission, nor any person acting on behalf of the Commission:

A, Makes any warranty or representation, express or implied, with respect to the accuracy,

completeness, or usefulness of the information contained tn this report, or that the use of

any information, apparatus, method, or process disclosed in this report may not infringe

privately owned rights; or

B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of

any information, apparatus, method, or process disclosed in this report.

As used in the above, "person acting on behalf of the Commission'* includes any employee or

contractor of the Commission to the extant that such employee or contractor prepares, handles

or distributes, or provides access to, any information pursuant to his employment or contract

with the Commission.

UNCLASSIFIED

ORNL-2053

Contract No. W-7405-eng-26

METALLURGY DIVISION

A LABORATORY FOR THE HIGH-TEMPERATURE CREEP TESTING

OF METALS AND ALLOYS IN CONTROLLED ENVIRONMENTS

D. A. DouglasW. D. Manly

DATEISSUEDL

SEP 1« 1956

OAK RIDGE NATIONAL LABORATORYOperated by

UNION CARBIDE NUCLEAR COMPANYA Division of Union Carbide and Carbon Corporation

Post Office Box POak Ridge, Tennessee

UNCLASSIFIED

MARTIN MARIETTA ENERGY SYSTEMS LIBRARIES

3 WSb D3SD1DD 5

UNCLASSIFIEDORNL-2053

Metallurgy and CeramicsTID-4500 (11th ed.)

INTERNAL DISTRIBUTION

1. C. E. Center 56.

2. Biology Library 57.

,3. Health Physics Library 58.

-4r5. Central Research Library 59.

^6. Reactor Experimentcil Engineering Library 60.

7-26. Laboratory Records Department 61.

27. Laboratory Records, ORNL R.C. 62.

28. A. M. Weinberg 63.

29. L. B. Emlet (K-25) 64.

30. J. P. Murray (Y-12) 65.

31. J. A. Swartout 66.

32. E. H. Taylor 67.

33. E. D. Shipley 68.

34. M. L. Nelson 69.

35. W. H. Jordan 70.

36. C. P. Keim 71.

37. R. S. Livingston 72.

38. R. R. Dickison 73.

39. S. C. Lind 74.

40. F. L. Culler 75.

41. A. H. Snell 76.

42. A. Hollaender 77.

43. M. T. Kelley 78.

44. K. Z. Morgan 79.

45. J. A. Lane 80.

46. A. S. Householder 81.

47. C. S. Harrill 82.

48. C. E. Winters 83.

49. D. S. Billington 84.

50. E. M. King 85.

51. D. W. Cardwell 86.

52. A. J. Miller 87.

53. D. D. Cowen 88.

54. R. A. Charpie

55. M. J. Skinner

G. E. BoydJ. H. Frye, Jr.W. D. ManlyJ. E. CunninghamG. M. Adamson, Jr.R. J. Beaver

E. S. Bomar, Jr.J. H. Coobs

J. H. DeVan

D. A. Douglas, Jr.R. J. GrayJ. P. Hammond

T. T. Hikido

M. R. Hill

E. E. Hoffman

L. K. Jetter

R. V. MeghreblianR. B. Oliver

P. Patriarca

M. L. Picklesimer

J. L. Scott

G. P. Smith

A. Toboada

J. C. Wilson

P. M. ReylingH. L. Yakel, Jr.E. Creutz (consultant)N. J. Grant (consultant)H. Leidheiser, Jr. (consultant)T. S. Shevlin (consultant)C. S. Smith (consultant)

E. E. Stansbury (consultant)ORNL - Y-12 Technical Library,

Document Reference Section

EXTERNAL DISTRIBUTION

89. R. F. Bacher, California Institute of Technology90. Division of Research and Development, AEC, ORO

91-766. Given distribution as shown in TID-4500 under Metallurgy and Ceramics category(300 copies - OTS)

DISTRIBUTION PAGE TO BE REMOVED IF REPORT IS GIVEN PUBLIC DISTRIBUTION

UNCLASSIFIED

UNCLASSIFIED

CONTENTS

Gaseous and Vacuum Creep Testing Machines 1

Specimen Design 5

Operational Problems 5

Liquid-Metal Creep-Testing Equipment 15

Tube Burst Testing Apparatus 23

Conclusions 23

Acknowledgments 29

UNCLASSIFIED

A LABORATORY FOR THE HIGH-TEMPERATURE CREEP TESTING OF

METALS AND ALLOYS IN CONTROLLED ENVIRONMENTS

D. A. Douglas W. D. Manly

The. determination of creep properties of metalsand alloys has become increasingly importantduring the past decade as interest has increasedin steam turbines, jet engines, and nuclear reactors,which require structural materials that are capableof withstanding increasingly higher and highertemperatures. Although much work has been donein creep testing and data have been made available for most of the known materials considered

practical for high-temperature operations, almostall data have been obtained in air, the environmentmost commonly encountered in engineering applications. In the field of nuclear engineering,however, structural materials at elevated temperatures must be compatible with gases, variousliquids, and liquid metals. It was realized, byW. D. Manly and R. B. Oliver, that the behaviorof metals in air would not necessarily reflecttheir behavior in these other environments; therefore a program was initiated for the specificinvestigation of the effect of various environmentson the high-temperature creep properties of metals.No commercial machines were available for testingspecimens in pure gases, in vacuum, or in aliquid-metal bath. Therefore it was necessaryto design and build such equipment at ORNL.

GASEOUS AND VACUUM CREEP

TESTING MACHINES

A testing chamber that is leaktight, free ofatmospheric diffusion, and has no internal sourcesof contamination is required for both the controlledatmosphere and the vacuum-environment work. Inorder to satisfy these requirements it was necessary to devise a strong, impervious, test chamber,which could be heated uniformly to any desiredtemperature and in which strain and temperaturemeasurements could be made on a stressed test

specimen. Figure 1 illustrates the final designof the chamber used for testing in gaseous environments; the chamber consists of a thick-walledmetal tube, water-jacketed on the ends to allowsufficient cooling for the brass bellows and forthe rubber O-rings utilized as seals. Side armswere provided to allow for optical measurement ofthe specimen extension. Around the core theheating elements were wound in such a way as

to compensate for the heat losses at the ends ofthe tube and along the side arms. External shuntswere provided to allow for adjustment of thevertical temperature gradients. Pull rods connected to and extending through the two bellowsmake it possible to introduce a load on the specimen inside the chamber. With the bottom pullrod anchored in place the specimen can be stressedby application of a load to the top bellows througha lever arm and a weight-pan system. Thermocouples are wired to the specimen, and the leadsare brought out to a junction box through rubberglands. Space considerations and the need toeliminate all internal sources of contamination

ruled out conventional extensometers utilizingelectronic signals. A method developed byBattelle Memorial Institute was used to opticallymeasure the extension of specimens during creepby means of a pair of dovetailed platinum alloystrips, which were attached over the gage lengthand referenced so that measurements could be

made with a microscope. Figure 2 shows thespecimen connected to the top bellows, withthe platinum strip and thermocouples in placeready for assembly in the test chamber. Figure 3is a view of the scribed platinum strip from whichthe extension measurements are made. A dial

gage attached to the pull rod outside the testchamber acts as a rough check on the micrometermicroscope readings. In the hands of experiencedpersonnel this system is capable of detectingstrain in the order of 5 x 10 in. Although thisoptical method of measuring extension introducesthe problem of human error, the prevention ofenvironmental contamination and the lack of main

tenance problems are thought to more than offsetthis disadvantage.

It is of course obvious that as a specimencreeps the lever arm will travel through an arc,which on this particular equipment is approximately 16 deg. The corresponding movement ofthe pull rod will cause a constantly varying forceto be exerted on the specimen by the bellows asit travels from a compressed position to an extended one. It is desirable to compensate forthis variable stress by exerting an equal varyingforce in the opposite direction so that the tare

Fig. 1. Creep Frames for Testing in Gaseous Environments.

Fig. 2. Assembly of Specimen, Extensometer, Thermocouples, and Pull Rods with the Top Bellows Flange.

Fig. 3. Gage of a Specimen with Platinum Strip Extensometer in Place.

on the system will be constant in any lever-armposition. This can be accomplished by counter-weighting the lever arm with a weight above thecenter line of the pivots (Fig. 4). The momentarm of this force will then vary with the angularposition of the lever arm, and, with a properadjustment of the height of the weight above thelever arm, the variable tare of the bellows canbe exactly counterbalanced. The proper calibration can be accurately determined by mountingstrain gages on a dummy specimen and assemblyingthe specimen in the test rig.

Figure 5 shows the slightly different setuprequired for the vacuum-testing work. An oildiffusion pump (VMF-80-06, 250 w, 150 cc) usedin conjunction with a mechanical fore pump provides a test atmosphere averaging 0.03 /x. Thevacuum is monitored during test with a Phillipsion gage. Considerable outgassing occurs duringthe heating-up period at the start of the test, andit is necessary to predicate the rate of heatingon the over-all capacity of the vacuum pumpingsystem.

In all environmental work the use of a helium

leak detector is a necessity in order to ensurethe integrity of the atmosphere. The standardprocedure is to make a careful check for leaks,immediately after assembly and again after thetest has reached temperature.

Temperature itself plays an important role inall types of creep testing. With the use of a Leeds& Northrup duration-adjust-type controller (DAT)and a Speedomax recorder, a set temperature levelin the range 1000 to 1800°F can be maintainedwithin ±1.5°F. Figure 6 is a view of one of theinstrument panel boards showing the location ofrecorders and controllers. The instrument on the

left is a 100-point precision indicator and isused to obtain a precise temperature reading fromeach thermocouple in the test. Above the panelboard is the alarm system, which gives both anaudible and visible signal whenever a seriousover or under temperature condition exists. Sincethe vacuum test is a particularly slow system torespond to temperature fluctuations, especiallyat low temperatures, some modification of theDAT unit was required. One important modificationconsisted in introducing a manual control forthe rate-of-droop correction to allow for a slowerreset rate. A complete discussion of the problemsinvolved in setting up the instrumentation for

this laboratory is given in a report by J. Lundholm,1a former instrument engineer at the Oak RidgeNational Laboratory, who worked out many of thecontrol difficulties.

SPECIMEN DESIGN

The selection of the type of specimen geometrywas considered most important, and considerablestudy was given to this problem before the finalconfiguration, as shown in Fig. 7, was decidedupon. There were several factors which led tothe selection of a flat sheet type of specimenrather than the more conventional round specimenfor the test program.

1. Early detection of significant trends, animportant factor in the investigation of environmental effects, was more apt to be achieved inthe sheet-type specimen because of its largesurface-area-to-volume ratio.

2. The pin-type connection in the sheet specimenwould, it was believed, tend to allow for generalself-alignment during loading of the specimen.Since screw joint connections could not be usedin tests in liquid metals because of galling andself-welding problems, it appeared obvious thatfor the sake of uniformity of the specimen designthe flat specimen with the pin connection was themore desirable.

3. The unit cost of machining the specimens, animportant consideration since it was envisionedthat hundreds of such specimens would be needed,was lower with the sheet-type specimen. With thesimple jig and the milling cutter, ground to theproper radius, shown in Figs. 8 and 9, it is possible to machine about 12 specimens of the sheettype in the length of time required to machine oneround bar specimen.

OPERATIONAL PROBLEMS

In comparing the environmental effect of puregases — in this investigation, hydrogen, nitrogen,and argon - with that of air, it is, of course,necessary to completely eliminate oxygen fromthe test gas. In the case of hydrogen the conventional method of using a platinum catalystcoupled with magnesium perchlorate drying tubes

J. Lundholm, Jr., Instrumentation for a PrecisionCreep Testing Laboratory, ORNL-1044 (Oct. 18, 1951);also, Proc. Instr. Soc. Am. 6, 43 (1951).

2E. G. Brush, Corrosion 11, 299T-302T (July 1955).

COUNTER WEIGHT

AC,AC = VARIABLE MOMENT ARM OF THE COUNTER WEIGHT

B = POINT OF ACTION OF VARYING BELLOWS FORCE

C = FULCRUM

D — POINT OF ACTION OF LOAD

Fig. 4. Diagram of a Lever-Arm System Counterweighted Above the Center Line.

UNCLASSIFIED

ORNL-LR-DWG 14673

Fig. 5. Creep Frames for Testing in Vacuum.

Fig. 6. Instrument Panel.

UNCLASSIFIED

ORNL-LR-DWG (4674

H in.

-4% in. 3V,6

~%in.^

( % in. 0.500 in. ±0.00tin. &

-V2 in. DRILL

Fig. 7. Sheet-Type Creep Specimen.

is sufficient to prevent the test pieces fromscaling or discoloring during test. Nitrogen,which is heavily contaminated with oxygen whenpurchased from commercial vendors, poses a moredifficult problem; it has to be bubbled through apacked column of a sodium-potassium alloy whichis liquid at room temperature in order for a largeportion of the oxygen to be removed. The gas isthen passed through a similar unit containingsodium at 250°C. This method of purifying nitrogenwill prevent the discoloration of a specimen duringtest. Argon, because of its inert nature, wasselected as an environment in which base-linedata could be determined. Helium would, ofcourse, provide an equally good environment;however, the extensive use of the helium leakdetector made it mandatory that the background ofhelium be kept at a low level to ensure the sensitivity of this test. Although argon can be procuredwith a very low level of contamination, it wasfound that test specimens would not retain abright tarnish-free surface unless additional purification was provided. Figure 10 is a designdrawing of one of the hot sodium purificationunits, which have been found to be capable ofcontinuous service for a number of years beforereplacement becomes necessary. Little is accomplished by this type of purification unit unlessextreme care is taken to ensure that the purity ofthe gas as it emerges from the unit is maintained

to the point where it is introduced into the testchamber, which, at this Laboratory, necessitatedthe use of 50 ft of connecting tubing. The originalinstallation of such tubing must be followed veryclosely to make sure that it is thoroughly cleanedand degreased and that no flux or other sources ofcontamination are introduced into the systemduring the final fabrication.

One environmental effect not under investigationbut which caused trouble on two occasions was

catastrophic oxidation. The original materialselected for the construction of the test chamber

core was type 316 stainless steel. This decisionwas based on the ease of fabrication and the high-temperature properties of this alloy. In the construction of the furnace around the core it is

necessary, in order to maintain the rigid temperature control required, to prevent chimney effectsalong the outer surface. As a result the airsurrounding the core is quite static, thus providingan ideal atmosphere for the initiation of catastrophic oxidation. This phenomenon is veryprone to occur in materials containing alloyingelements, whose oxides are quite volatile, such asvanadium or molybdenum. Type 316 stainlesssteel contains just enough molybdenum to makeits behavior unpredictable. Figures 11 and 12

A. deS. Brasunas and N. J. Grant, Trans. Am. Soc.Metals 44, 1117-1149 (1952).

I

ffi im

Si^4j L

Fig. 8. Jigs and Cutters for Machining Creep Specimens.

8

UNCLASSIFIEDPHOTO 9843

Fig. 9. Setup for Multiple Milling of Sheet-Type Creep Specimens.

LIQUID DEFLECTORS

LIQUID PUMP COLUMNS

LIQUID LEVEL

12

OUTLET LINE £

UNCLASSIFIED

ORNL-IR-DWG 14675

LIQUID LEVEL

LIQUID RETURN LINE

STRIP HEATERS AND INSULATION

NLET LINE

Fig. 10. Liquid-Metal Gas Purification Unit.

OJ

Fig. 11. Oxidized Metal Liner from Gaseous Test Chamber.

UNCLASSIFIED

Y-12547

14

unclassifiedY-15607

Fig. 12. Oxidized Metal Liner from Gaseous Test Chamber Near Water-cooled Head.

illustrate the deterioration of one chamber in which

the catastrophic oxidation reaction occurred. Although only two chambers, in over four years ofcontinuous operation, have suffered from thisdefect, all subsequent chambers have been constructed of type 309 stainless steel.

Another problem which became evident duringthe early part of the test program was that, forsubsize specimens or conventional specimenstested at low stresses, the inherent friction in thepivotal knife edges of the lever arm imposed astress variation of sufficient magnitude to warrantcorrection. An easy solution to this problemappeared to be the installation of some type ofdead-load device which would not disturb the

purity of the environment. Figure 13 is a photograph of the cylindrical shell, which can beattached to the bottom of the test chamber and

sealed with the standard rubber O-rings. Thisshell encases a weight pan and hydraulic jackand allows for the direct loading of the standardsheet type of specimens to about 3000 psi.Figure 14 shows the isoelastic spring used toconnect the bottom pull rod to the rod holding theweight pan; the strain gage attached to the reduced section of the rod enables the operator toload the specimen incrementally by lowering thehydraulic jack.

LIQUID-METAL CREEP-TESTING EQUIPMENT

In designing the liquid-metal creep-testing equipment it was necessary that the system be absolutely leaktight, not only to provide the proper testconditions but also to afford protection to operatingpersonnel. The main shift in design philosophyfrom that used with the gaseous creep apparatuswas the recognition of the complexities that wouldbe involved in trying to devise a bottom closureso that the pull rod could be anchored outsidethe liquid-metal container; therefore attention wasgiven to the possibility of anchoring the pull rodinside. Figure 15 shows the component parts ofthe final design. A capped, heavy-walled pipewith a flanged top acts as the container for theliquid metal. A smaller diameter pipe, perforatedto allow for circulation, is suspended from thetop flange and serves as an anchor for the bottompul l-rod. The specimen and the pul l-rod componentsfit inside- this "compression tube," which issufficiently large for its deformation under loadto be neglected. Figure 16 shows, schematically,

!

the specimen assembled in the test chamber. Alever-arm system is used to exert the load througha stainless steel bellows closure and can be

counterweighted similar to the gaseous type oftest apparatus. Thin-wall tubing is used as

Fig. 13. Cylinder Encasing Weights for Dead-Load Creep Tests.

15

UNCLASSIFIED

Y-15916

Fig. 14. Isoelastic Spring, Strain Gage, and Weights for Incremental Loading in Dead-Load Creep Tests.

16

fl5WWa*gJ=pK»*7*5

Fig. 15. Component Parts of Liquid-Metal Creep-testing Apparatus.

17

18

BELLOWS

INLET FOR COOLANT

LIQUID METAL

TEST SPECIMEN -

UNCLASSIFIED

ORNL-LR-DWG 14676

Fig. 16. Assembly Drawing of Liquid-Metal Creep-testing Apparatus.

protective shields for the thermocouples to preventtheir being attacked by the liquid metal.

The corrosive effects of liquid metals and theirvapors made it necessary to position the strain

measuring device outside the chamber. A 0.001-in.dial gage is used to sense the movement of thepull rod above the bellows. The errors inherentin a measuring system that is dependent on therelative movement of parts 12 in. or more fromthe actual gage of the specimen and the effectsof ambient temperature cycles on a bellows causeconsiderable scattering in the data for the first0.5% strain. The reproducibility of strain measurements above 0.5% is good, and reliable designdata can be obtained.

The liquid-metal test environment must be transferred from an outside container into the test

chamber through a completely closed system. Thelevel of this liquid medium must be carefullycontrolled so that the specimen will be entirelycovered but the chamber will not be overfilled. A

procedure has been developed which allows theoperator to establish the precise level of theliquid desired. An ordinary spark plug is weldedto a length of metal rod, and the rod is introducedinto the chamber through an appropriate seal inthe top flange. The length of the probe is cut atthe required level. A resistance meter is attachedto the spark plug and grounded through the frameof the machine. When the liquid metal reachesthe probe, the meter shows an instant response,as evidenced by the needle moving from infiniteresistance to essentially zero resistance. Additional control of the level of the liquid can beobtained by bringing the charge line, throughwhich the liquid is introduced into the chamber,down the inside wall to the required level. Afterthe signal has been observed on the resistancemeter and the flow of metal has been stopped, itis possible to reverse the pressure and force anyliquid above the outlet of the tube back into thecharging container. It has been found that differential pressure is the best motivating force foreffecting the transfer of the molten metals fromone container to another. Figure 17 is a photograph of the portable rig which contains the furnaceand controllers, for heating the metal, and the tanksof gas and regulators to control the flow of themolten metal from the charge pot into the testchamber. Swagelok fittings are used to connectthe transfer lines to the test chamber and to

disconnect them, thus facilitating the ease ofoperation as well as providing the necessaryleaktight integrity.

It has been found that small amounts of oxygenor carbon contamination of the liquid-metal bathcan produce substantial changes in the high-temperature properties of many materials.4 It istherefore necessary, before the test begins, tocompletely clean not only the test specimen butall component parts which are in contact withthe liquid environment. Although steam jets andacid dips have been tried, the most satisfactorycleaning method is to charge the assembled testchamber with the particular molten media of interestand allow a soaking period at the test temperature.After this batch is discharged, the actual testcharge can be introduced with the assurance thatall contaminants have been removed. Another veryimportant precaution which must be strictly adheredto in this type of testing is that all metal surfacesexposed to the liquid metal be of the same materialor at least be very similar in chemical composition.Otherwise, a corrosion phenomenon known as"dissimilar metal transfer" will occur with adetrimental effect either on the test specimen oron the accessory metal parts of the assembly.

Figure 18 is a drawing illustrating the test rigused for low-stress tests. In this machine a leverarm with a 1:1 ratio is employed. The lever armis made of magnesium and is suspended by reedsto reduce the tare and friction in the system. Therelatively large distance between the pull rod andcenter pivot of the lever arm makes it impracticalto employ the counterweighting system describedfor the other lever-arm machines. Instead, twosmall metal pots are partially filled with mercuryand interconnected by two lengths of tubing. Onepot is fastened to the front of the lever arm andthe other to the back. The displacement of themercury from one pot to the other in relation tothe position of the lever arm can be regulated byshifting the positions of the pots to automaticallycounterbalance the changing spring load of thebellows. Figure 19 is a view of a number of thetest machines now in operation.

C. B. Jackson (ed.), Liquid Metals Handbook, Sodium-NaK Supplement, TID-5277 (July 1, 1955).

R. N. Lyon (Ed.-in-Chief), Liquid Metals Handbook,2d ed. NAVEXOS P-733(rev.) (June 1952).

19

Fig. 17. Portable Rig for Charging and Discharging Liquid Metal in the Test Chambers.

20

10 12 3 4 5 6

SCALE IN INCHES

UNCLASSIFIED

ORNL-LR-DWG 14677

>THERMOCOUPLES

Fig. 18. Assembly Drawing Showing the Counterweighting System on a 1:1 Lever Arm.

21

Fig. 19. Six Liquid-Metal Creep-testing Frames.

TUBE-BURST TESTING APPARATUS

In the design of nuclear reactors many componentparts consist of pipe or tubing. The design engineermust, then, be concerned with the high-temperatureproperties of structures utilizing this configuration.The question naturally arises as to the reliabilityof data obtained from uniaxially stressed sheetspecimens as design criteria for internally pressurized tubes. In order to resolve this question, anapparatus was devised in which pipe or tubing couldbe tested in the environments of interest. Figure20 is a descriptive drawing showing the test chamber, the specimen, and the necessary accessories.The problems of measuring temperature, of fillingthe chamber, and of discharging the molten metalsare similar to those encountered in the lever-arm

creep machines. The specimen consists of a lengthof tubing with a reduced wall thickness along a 2-in.-gage section. One end is sealed, and pressureis introduced at the other end as a means of inducing the stress in the specimen. Under theseconditions the stress system is biaxial, with thetangential or hoop stress being twice the stress inthe axial direction. It is possible to modify thissetup as shown in Fig. 21 so that the specimen canbe stressed only in the tangential direction, only inthe axial direction, or in any simple ratio of thesetwo stress directions by changing the pressure oneither side of the piston. It is also possible tointroduce cyclic stresses by alternately pressurizing one side and then the other of the piston.The flexibility of the test is one of its mostattractive features. The main drawback to thistype of test is that there appears to be no feasibleway of measuring deformation during the test, sothat results must be compared on the basis of thetotal time to rupture. Since establishing the timeto rupture is very important, considerable effortwas made to determine the exact time that failureoccurs. The encasement of the specimen in a metalchamber containing a liquid metal or a pure gas restricts the possible methods of determining failure.However, by inserting a sensitive pressure switchin the line from the pressure source to the specimenit is possible to detect the drop in pressure whichoccurs as the high-pressure gas in the specimenleaks out through the intergranular cracks whichpropagate through the wall of the specimen. Figure22 is a photograph showing the test apparatus. Onthe left is the argon tank, which is the pressuresource. The small cylindrical can is the solenoid

valve which periodically isolates the specimen fromthe pressure source so that the pressure will bleeddown when failure occurs. Next in the line is the

pressure switch which signals failure of the specimen. When a drop of from 10 to 15 lb in pressureoccurs, the switch is activated and an electricalcircuit is opened. This cuts off the cycle timer andthe clock which is recording the test time. With thestoppage of the cycle timer the solenoid valve remains closed so that there is no excessive loss of

gas from the argon tank. Following the pressureswitch is a long-scale pressure gage which enablesthe operator to accurately establish the desiredpressure on the specimen. Figure 23 is a photograph of a section of the panel board showing someof the temperature controllers and recorders, theVariacs, the clock timers, and the Flexopulse cycletimers.

CONCLUSIONS

Through the use of the equipment described, design information similar to the curves shown in Fig.24 has been obtained for a number of materials,over a wide range of temperatures, and in many different environments. The reliability of the dataobtained in the gaseous and vacuum type of creepmachines compares favorably with that obtainedfrom the conventional apparatus used for testing inair. The data obtained in the liquid-metal test unitsshow considerable scatter at 0.5% strain and below.

This equipment should be modified to improve thesensitivity of the extensometer system for the measurement of strains of less than 1.0%.

The life of this equipment cannot be accuratelyestimated, since, after four years of almost continuous service, all the major component parts are stillin operation. A program of preventive maintenancehas ensured uninterrupted operation of the temperature-controlling instruments. The records indicatethat for the vacuum and gaseous test machinesduring the past four years only 2.0% of the availabletesting time that was lost was due either to thenecessary disassembly of completed tests and initiation of new tests or to maintenance and repair.About 5.0 to 7.0% of the available testing time waslost in the operation of the liquid-metal test machines.

It has been gratifying to the personnel in the Metallurgy Division to note that the equipment described here has been used as a model by a number

23

ANGLE VALVE

SIGHT FEED BUBBLER-

1 ° ' 2 3 4 5 6

SCALE IN INCHES

.ARGON OUTLET

PRESSURE LINE

UNCLASSIFIEDORNL-LR-DWG 14678

THERMOCOUPLES

SPECIMEN

Fig. 20. Assembly Drawing of Tube-Burst Testing Apparatus.

24

O-RING

COOLING WATER OUT

CONTAINER FOR TEST ENVIRONMENT-

Fig. 21. Assembly Drawing of Strain-cycling Apparatus.

UNCLASSIFIED

ORNL-LR-DWG 14679

PISTON PRESSURE

PISTON PRESSURE

-COOLING WATER IN

25

Fig. 22. Control Equipment Used for Tube-Burst Testing.

• s a• * - § i •

UNCLASSIFIEDY. 13950

°8 Q B

Fig. 23. Instrumentation for Tube-Burst Testing.

27

CO

20,000

10,000

°= 5000 —

2000

1000

100 200

TIME (hr)

Fig. 24. Typical Design Curves.

1000 2000

UNCLASSIFIED

ORNL-LR-DWG 14680

5000 10,000

of other laboratories who have become interested in

thi s type of testing.

ACKNOWLEDGMENTS

In the development of the various equipment forthis laboratory, ideas were gathered from manysources at the Oak Ridge National Laboratory andfrom other testing laboratories around the country.Thus it is impossible to credit any one individualwith the specific ideas which culminated in theequipment described. However, credit for the synthesis of all the ideas, and the evolution of thedesigns from their inception to their final form, be

longs chiefly to R. B. Oliver, who was at that timein charge of the Mechanical Testing Group. Considerable assistance on the difficult design problems was afforded by A. M. Tripp, W. A. Pate, andJ. M. Kasserman of the Engineering Department.The furnace used in the vacuum and gaseous creeprigs was designed by the Marshall Furnace Companyof Columbus, Ohio. J. Lundholm, formerly with theInstrument Department, was of invaluable assistance in solving some of the complex instrumentationproblems. Many improvements have been made inthe equipment over the past four years, and most ofthe modifications have resulted from proposals bythe excellent technician staff, which operates theequipment.

29