TECHNICAL HIGHLIGHT

An overview of thermo-mechanical test methods for CFCC tubes

Michael G. Jenkins

Evaluation of the thermo-mechanical properties and performance of CFCCs in the form of tubular specimens presents serious pitfalls and testing complications for even the most experienced empiricist. Among these complications is the fact that the tube itself represents a component with a somewhat complex stress state. Separation of a simple coupon from the tube to facilitate testing violates the structural integrity of the material inherent in the fabrication of the tube. Thus, one is left with evaluating the tube as the test specimen and extracting the material properties from the less-than-ideal stress and strain states.

As with most thermo-mechanical tests, the first step in test preparation is to determine the properties and performance of interest. Generally, the actual use conditions of the material play a significant role in determining which thermo-mechanical properties are to be evaluated. However, even though most applications involve multi-axial loading, in many cases properties for isotropic, homogeneous materials are obtained under uniaxial conditions and are used as input to design codes to predict component performance under these multi-axial conditions. Thus, when evaluating coupons of "ideal" material, uniaxial conditions are generally sufficient to characterize material response. Because tubes are components and CFCCs are anisotropic, non homogeneous materials, multi-axial test conditions may be required to extract the exact material response [1]. For example, for filament-wound tubes with the fibre direction 45° to the longitudinal axis of the tube, the hoop stress produced by an internal pressure will not produce a maximum stress in the fibre direction (see Fig. 1a). In this case, pure circumferential torsion will be required to produce a maximum stress in the fibre direction as shown in Fig. 1b.

FIG. 1 Stress states at the surfaces of internally pressurized and circumferentially torqued tubes.

Therefore, depending on the composite structure (fibre lay-up, number of plies, etc.) and final design application, test conditions for tubular specimens may require somewhat complicated multi-axial loading arrangements. The most complete loading arrangement is shown in Fig. 2 where the tube is loaded by an longitudinal load, a circumferential torque, and an internal pressure. Using this type of multi-axial loading, the stress state at the surface and through the tube wall thickness can be varied to achieve any direction of the principal stresses and maximum shear stresses.

FIG. 2 Complete loading arrangement for CFCC tubes.

Although stresses can be inferred from the applied loads and elastic stress relations, the determination of strain is somewhat more complicated. For homogeneous, isotropic materials the complete strain state can be experimentally determined at room temperature via three-element strain gage rosettes. However, for anisotropic, non homogeneous materials such as CFCCs, sufficiently large strain gages must be used to average strains over local strain "spikes" such as those occurring at crossovers of fibre bundles or those occurring near regions of porosity. In addition, strain gage rosettes are limited to rather low temperature (<200°C) applications, thus limiting the effectiveness of determining the complete strain state at elevated temperatures. Most high-temperature extensometers measure uniaxial displacements and thus are not amenable to determining the complete strain state if the principal strain directions are not known a priori. Thus, high-temperature Moiré interferometry [2] holds promise for providing the whole field strain information. The zero-thickness grating applied to the specimen provides usefulness of high temperature Moiré to temperatures of ~1400°C. However, application of this grating requires etching the specimen surface, which by necessity must be optically smooth. Therefore, application of the whole field technique of high-temperature Moiré interferometry may be limited by the typically rough surfaces of as-processed CFCCs.

Additional complications of testing CFCC tubes involve the actual transmittal of the test loads to the tubular test specimen. Internally-pressurized loads are applied through either high pressure gases [3] or through molten glasses [4]. Longitudinal axial loads and circumferential torques must be transmitted from the test machine to the CFCC tube via suitable specimen grips. Typically, the tube must be prepared first by suitable smoothing and shaping of the internal diameter followed by the insertion and bonding of an appropriate end plug (typically a metallic plug and epoxy or ceramic adhesive) [5]. The end plugs serve three primary purposes: 1) they provide additional radial strength and stability so that the tube ends will not collapse upon gripping, 2) they facilitate the formation of cylindrical tube ends with dimensional tolerance insufficient for gripping, and 3) they provide a measure of strengthening in the gripped section to increase the likelihood of gage section failures. Thus, it can be appreciated that as more complex stress states are striven for, the difficulty of performing tube testing increases appreciably.

As a first step in evaluating viable thermo-mechanical test methods for CFCC tubes, selected test facilities and capabilities for evaluating ceramic and ceramic composite tubes are summarized pictorially in Figs. 3-10. As with all such scenarios, the described facilities and test methods have advantages and disadvantages. As previously discussed, the choice of approach is dependent on the desired information. Thus, industrial participants must develop reasonable requests for material properties and performance parameters which are required for their design applications before appropriate thermo-mechanical test methods for evaluating tubes can be developed.

FIG. 3 - Tube testing at Pennsylvania State University (Center for Advanced Materials).

Primarily room- and high-temperature tests with internally pressurized tubes (burst testing). Various gases are used for the working medium from inert to combustion. No strain measurements appear to be made. Radial and hoop stresses (possibly axial with closed ends) with thermal and mechanical (pressure) cycling also possible.

FIG. 4 - Tube testing at Southern Research Institute. Primarily room- and high-temperature tests with longitudinally and torsionally-loaded, internally pressurized tube testing. Various gases are used for the working medium from inert to combustion. Normal and shear strain measurements are made.

FIG. 6- Tube testing at University of Dayton Research Institute. Primarily high temperature tests with internally-pressurized tubes. Molten glass is used as the working medium. No strain measurements appear to be made.

FIG. 7 - Tube testing at Virginia Polytechnic Institute and State University (Materials Response Group). Primarily room and high temperature tests with longitudinally and torsionally-loaded tubes. Normal and possibly shear strain measurements appear to be made. Various 2-D stress states possible with maximum stresses in axial and off axis directions although maximum stress in the hoop direction is not possible. Thermal and mechanical cycling are also possible.

FIG. 8 - ASTM D2290 Test Method for Apparent Tensile Strength of Ring or Tubular Plastics and Reinforced Plastic Parts by Split Disk Method. Primarily room temperature but may be extended to high-temperature tests of diametrally-loaded tube sections. Normal strain measurements may be made. Hoop stresses with thermal and mechanical (pressure) cycling also possible.

FIG. 9 - Tube testing using 'generic' O-ring and C-ring tests. Primarily room and high temperature tests of diametrally-loaded sections of tubes. Normal strain measurements are possible. Hoop stresses vary across the thickness since the primary loading mode is flexure. Thermal and mechanical cycling are possible.

FIG. 10 - ASTM Proposed Standard Test Method for (Inplane Shear) or (Transverse Compressive) or (Transverse Tensile) Properties of Hoop Wound Polymer Matrix Composite Cylinders. Primarily room temperature tests with longitudinally- or torsionally-loaded tubes. Normal and shear strain measurements appear to be made. Axial and torsional stresses with thermal and mechanical cycling may be possible.

REFERENCES

[1] W.W. Stinchcomb, K.L. Reifsnider, and T.J. Dunyak, "Investigation of Properties and Performance of Ceramic Composite Components," ORNL/Sub/87-SA946/02 CCMS-92-14, Virginia Tech Center for Composite Materials and Structures, Blacksburg, Virginia (1992).

[2] B.S.J. Kang, G.Z. Zhang, M.G. Jenkins, M.K. Ferber, and P. Ifju, "Development of Moiré Interferometry for In-Situ Material Surface Deformation Measurement at High Temperature," pp. 964-976 in Proceedings of the 1993 SEM "50th Anniversary" Spring Conference on Experimental Mechanics, Society for Experimental Mechanics, Bethel, Connecticut (1992).

[3] D.L. Shelleman, O.M. Jadaan, D.P. Butt, R.E. Tressler, J.R. Hellmann, and J.J. Mecholsky, Jr., "High Temperature Tube Burst Test Apparatus," J. Testing and Evaluation, 20 [4] 275-284 (1992).

[4] L. Chuck, Private Communication, University of Dayton Research Institute, Dayton, Ohio, (1993).

[5] K. Liao, J.J. Lesko, W.W. Stinchcomb, K.L. Reifsnider, and T.J. Dunyak,

"An Axial/Torsional Test Method for Ceramic Matrix Composite Tubular Specimens," Ceram. Eng. Scien. Proc. , 11 [9-10] (1993).

http://www.ms.ornl.gov/cfcc/tech/hilites/tubeht.htm
November 19, 1996 mgc@ornl.gov