Ultrasonic testing is extensively utilized for integrity verification of composite material. In advanced applications and for composite process control it is advantageous to perform non-contact ultrasonic measurements that do not require traditional ultrasonic coupling and contact requirements. The new non-contact ultrasonic testing methods using laser ultrasound and air/gas coupled ultrasonic transducers enable in-situ evaluation of composite process such as fiber placing. This hybrid ultrasonic configuration has been extended to composite applications that are not possible with the conventional ultrasonic methods. Non-contact ultrasonic methods are adapted for surface wave testing in MHz frequencies that enable defect detection in surface plies of the composite material. Surface wave ultrasonic methods are demonstrated to be sensitive to composite process defects and are affected by ply disbonds or lack of composite consolidation. Non-contact ultrasonic approach enables on-line process control. The non-contact test configuration has been packaged into a compact remote test head.
Key Words: Rayleigh waves, Ultrasonic Non-contact Sensors, Composite Fiber Placing, Nondestructive Evaluation, Laser Ultrasonic, In-process testing
Nondestructive evaluation (NDE) has played critical role in processing, testing and structural evaluation of composite materials. Many conventional and emerging NDE tools are used for NDE of composites. Ultrasonic methods have been demonstrated as very effective tool in characterization of anisotropic composite materials and structures (1,2). Most of the ultrasonic testing is performed on finished products and most of the conventional test device and ultrasonic transducers cannot be directly used for in-process applications (3-4). Figure 1 illustrates the range of ultrasonic sensors that have been adapted to composite inspection needs.
Fig 1: Non-contact ultrasonic sensors suitable for scanning of composites.
The need for robust, sensitive, and inexpensive ultrasonic inspection measurement methods is common to many NDE needs. In advanced manufacturing applications and for composite process control it is advantageous to perform non-contact ultrasonic measurements that do not require traditional ultrasonic coupling and mechanical contact. Ultrasonic testing configurations can utilize many wave types. Shown in figure 2 are types of ultrasonic stress-waves that can be present in material plate of thickness T.
Fig 2: Ultrasonic stress wave types present in the plate thickness T . Bulk waves shear and longitudinal wavelength must be much smaller than T . The plate geometry readily supports Rayleigh and Lamb waves that are harder to induce using conventional transducers and are not commonly used for NDE.
Generally, ultrasonic NDE is performed with ultrasonic wavelengths much smaller than T using transducer configurations generating longitudinal or shear waves. Structural and materials testing is possible using Lamb or Rayleigh (surface) waves. However, conventional contact and immersion transducers are not readily adaptable for surface or Lamb wave applications. Non-contact ultrasonic NDE sensor technologies using alternate wave types, namely laser ultrasonic and air-coupled ultrasonic have been demonstrated in the laboratory and now appear to be on the threshold of broad application utility. The development of non-contact sensor technologies is of special importance to process control and for defect detection in fiber placed composite structures (5-8).
The process of fiber placement of thermoplastic matrix composite structures requires real time process control and rugged nondestructive evaluation (NDE) sensors. The surface acoustical stress wave methods combined with remote laser/air coupled ultrasonic transducers are in development for on-line, in process control of fiber placing operations. Figure 3 is a diagram of the hybrid ultrasonic test configuration using laser ultrasound and air/gas coupled ultrasonic transducers. These advanced non-contact ultrasonic methods are adapted for surface wave testing in MHz frequencies that are not readily achievable using contact transducers. Non-contact ultrasonic configuration enables non-interfering defect detection and materials characterization for in-process inspection and on-line process control. Figure 4 is photograph of compact remote sensor test head designated as Hybrid Ultrasonic Remote Inspection System (HURISTM) developed for non-contact in process NDE. This ultrasonic test setup is usable for both, process control and in-process structural integrity assessment. Significant advantage of the above method is ability to examine large areas of material without a need for scanning. Currently, the demonstration HURISTM is operating at 1.3 MHz and operates in a pitch-catch surface wave mode across fibers.
Fig 3: Diagram of the hybrid ultrasonic test configuration using laser generation and air coupled detection. This configuration allows large separation between laser generation and air coupled detection heads and uses surface acoustic signals for non-contact inspection of composites. The set-up is adaptable for pulse echo or pitch-catch test configurations.
Fig 4: A photograph of the Hybrid Ultrasonic Remote Inspection System (HURISTM ) test head incorporating laser generation and air coupled detection of surface acoustic signals for composites materials. This system is suitable for totally remote in-process monitoring of composite fabrication.
The HURIS test head assembly is serviced by two umbilical links; electrical cables to Air coupled transducer and fiber optic link to supply laser light for signal generating. Currently the umbilical is set at 5 meters with practical stand off at approximately 3 meters. The test head allows for alignment of transducers, adjustment of transducer separation and precision stand off adjustment. Air coupled ultrasonic detectors also include angular adjustments that are critical for signal optimization. Currently, the assembly is about 10 lb. and can be mounted on extended arm using pre-machined slots. With the addition of the flexible umbilicals, the electronic and laser support instrumentation is removed from the process area. The frequency selection of ultrasound laser generation is achieved by formed multi-line array source that also enables control of directivity of the surface acoustic waves. Shown in Figure 5 is a surface acoustic wave signal acquired using hybrid configuration on fully consolidated Carbon/Peek fiber placed composite material. This digitally captured signal shows narrow frequency content and good signal to noise. This test set-up is suitable for many refinements and can be adapted to variety of part configurations and ultrasonic test requirements.
Figure 6 is an example of ultrasonic signal change due to poor consolidation of the fiber placed tows. The loss of surface acoustical wave signal intensity demonstrates the robustness of this approach similar to ultrasonic C-scan that is widely used by the industry for quality screening of composite structures. These tests used Nd:YAG Q-switched Laser at 1.06 µm, pulse duration 10ns, and average energy 15mJ/pulse irradiating the test piece via 10 lines mask. Laser light coupling to test head was via pure silica core optical fiber, doped silica cladding with 1500 µm core diameter or a fiber bundle. Generating head standoff distance is variable between 1cm and 5cm. Air coupled transducer detection standoff distance is variable between 2cm and 5cm.
Fig 5: Surface acoustic wave signal for a consolidated composite material.
|| b) Poor consolidation, two plies deep|
Fig 6: Surface acoustic wave signal for a region of poor consolidation due to processing detected in Carbon/Peek fiber placed composite.
Figure 7 is a graphic summary of surface acoustic signals measured across the GrafoilÒ inserts at different ply depths in the fiber place composite. These controlled disbond defects are triangular shape. The disbond characteristics of the flaws were verified via a conventional ultrasonic c-scan. The surface wave signals are attenuated due to presence of this mechanical discontinuity. The depth of the test sensitivity is dependent on the frequency of the surface wave. For fiber placing of composites, the tests need to detect disbonds at the most recent consolidation plies. This requirement translates to surface acoustic frequencies between 0.5 and 1.5 MHz. Figure 8 shows a signal intensity variance for the line scan across the ply disbonds. The loss of surface acoustical wave
Fig 7: Surface acoustical signal across the first to second ply and second to third ply disbond triangular defect embedded in the Gr/Peek fiber placed composite. The 500 kHz signal is attenuated by presence of ply disbonds.
signal intensity due to disbond condition further demonstrates the robustness of this approach and is similar to ultrasonic C-scan testing approach. For better quality screening of composite structures and more quantitative process control further signal processing is possible to associate signal loss with the types of defects. Discriminate between signals is desirable for the process control and for comprehensive engineering data-base summary of the finished composite component.
Fig 8: Surface acoustic wave tone-burst signal intensity plotted as function of position across the disbond region of the sample. Acoustical signals are generated across the fibers using multi-line laser source and air coupled transducer detector.
Using hybrid ultrasonic configuration, we have developed a remote ultrasonic inspection system using surface acoustic waves for subsurface inspection of in-line/in-process fiber placed composite materials. This system meets essential objectives for such composite inspection. This device is based on laser ultrasonic and air-coupled ultrasonic transduction methods and the technology is feasible for production implementation. Laser generation and air-coupled reception can be packaged as a compact remote sensor test head (Hybrid Ultrasonic Remote Inspection System -HURISTM;), for the development of non-contact ultrasonic test methodologies. The same test set-up is usable for both, process control and in-process material structural integrity assessment. Significant advantage of the above method is ability to examine large areas of material without a need for scanning. Laser ultrasonic trunsduction extends ultrasonic measurements to test configurations and wave-modes that are difficult to perform using conventional technology. The hybrid test configuration allows for truly non-contact and remote inspections and incorporates spatial modulation technique for generation of narrow-band ultrasonic surface waves.
- R.E. Green, Jr., "Special Problems Associated with Ultrasonic Wave Propagation in Anisotropic Materials", in Nondestructive Characterization of Materials V, T. Kishi et al. (Eds.), Iketani Science & Technology Foundation, Tokyo, Japan, pp. 469- 476 (1992).
- Boltz, E.S. and Green, R.E., "Ultrasonics in Thick Anisotropic Materials", Review of Progress in Quantitative NDE, Vol. 12B, pp. 1241-1248 (1993).
- B.B. Djordjevic, "Advanced Ultrasonic Probes for Scanning of Large Structures", Proc. Ultrasonic International 93 Vienna, Austria, July 1993, Pub. Butterworth Heinemann 1993
- B.B. Djordjevic, Henrique Reis, editors, G. Birnbaum, B. A. Auld, Technical editors, "Sensors for Materials Characterization, Processing, and Manufacturing" ASNT Topics on NDE Vol. 1, published by ASNT, Columbus OH, 1998.
- B. B. Djordjevic, R. E. Green, Jr., "Non-Contact Ultrasonic Techniques for Process Control of Composite Fabrication" Proc: NDE Applied to Composite Processing, St. Louis, MO October 1994, NTIAC 1994
- E.S. Boltz, C.M. Fortunko, M.A. Hamstad and M.C. Reneken, Review of Progress in Quantitative Nondestructive Evaluation, Vol. 14, Plenum Press New York, 1995, p967
- B. Boro Djordjevic, T. Berndt, M. Ehrlich, K. Baldwin, D. Palmer, S. Holmes "Advances in NDE for On-line Fiber Placement Process" Proc. SAMPE98, May 31-July 4 1998, Anaheim CA 1998
- F. Lanza di Scalea, T. Berndt, J. Spicer, B. B. Djordjevic, "Remote Laser Generation of Narrow -Band Surface Waves Through Optical Fibers" IEEE Tran. UFFC, V46, No. 6 November 1999
The author would like to thank Dr. F. Lanza diScalea, Dr. T. Berndt and M. Ehrlich in technical contribution to this work. Part of this work was funded by ONR, Mr. James Kelly, project manager.