NEW RESULTS IN THE APPLICATION OF E/M IMPEDANCE METHOD TO MACHINERY HEALTH MONITORING AND FAILURE PREVENTION

Transcription

1 3 rd Meeting of the Society for Machinery Failure Prevention Technology, April 2-22, 1999, Virginia Beach, VA NEW RESULTS IN THE APPLICATION OF E/M IMPEDANCE METHOD TO MACHINERY HEALTH MONITORING AND FAILURE PREVENTION Victor Giurgiutiu and Craig A. Rogers Department of Mechanical Engineering, University of South Carolina Columbia, SC, 29212, USA , ABSTRACT The electro-mechanical (E/M) impedance method for structural health monitoring, damage detection and failure prevention is a new technology utilizing the emitter-detector properties of active material sensors and the structural drive point mechanical impedance. In previous work, the authors showed that changes in the drive point mechanical impedance of a structure or machinery can be sensed in the form of changes in the apparent E/M impedance of the active material sensor. Subsequently, a number of situations have been considered for the use of E/M impedance method in conjunction with active material sensors to important health monitoring and failure prevention applications. The paper starts with a brief review of the electro-mechanical (E/M) impedance method for structural health monitoring and non-destructive evaluation. Experimental results from tests conducted on various pieces of structures and machinery are presented. The experimental set-up consisted of an HP 4194A impedance analyzer, Pentium computer, National Instruments data acquisition hardware and software, and an array of PZT active sensors. These proof-ofconcept demonstration experiments proved the suitability of the electromechanical (E/M) impedance method to a number of industrial applications. This paper brings together, in a unified format, a large body of knowledge that has been accumulated regarding the electro-mechanical (E/M) impedance method and indicates its suitability to failure prevention monitoring in complex machinery. The theoretical and practical issues posed by its further development and wide industrial applications are also addressed. Key Words: Damage detection; Health monitoring; Failure prevention; Electromechanical impedance; NDE; Non-destructive evaluation; Incipient damage; Progressive deterioration.

2 INTRODUCTION Machinery failure prevention is a complex activity that requires the synchronous interaction of several factors. Critical among them is the ability to detect the apparition and propagation of structural damage in vital machinery parts. The detection of damage (e.g., cracks, delaminations, disbonds, etc.) is crucial in any failure prevention technology. If damage were detected at an early stage, corrective measures could be taken and catastrophic failure can be prevented. Moreover, the affected structure can be locally repaired, and put back in service. By substituting timely repairs for costly replacements, important cost saving can be achieved in the machinery lifecycle. A reliable procedure for early damage detection will reduce the design uncertainties, will increase designer confidence, and will result in lower reserve factors, smaller weight, and reduced initial cost. It is apparent that the development of damage detection technologies plays a major role in the larger picture of preventing the failure of machinery and reducing the lifecycle cost. (Giurgiutiu and Rogers, 1998e) Today's damage detection technologies fall into two categories. On one hand (Figure 1), we find a multitude of active and passive scanning technologies that are essential during periodic inspections. On the other hand, there are a number of in-situ sensor technologies that could enable continuous health monitoring. The in-situ sensors and their accompanying technologies are of paramount importance for on-line health monitoring and failure prevention. Damage Detection Technologies Passive and Active Scanning 1. Ultrasonic probing 2. Eddy currents 3. Liquid penetrant 4. Thermography and Vibro-thermography. Magnetic particles and Magnaflux 6. Computer tomography 7. Laser ultrasound 8. Low power impulse radar In-situ Sensor Arrays 1. Vibration monitoring 2. Strain monitoring (electrical and fiber optics) 3. Peak-strain indicators 4. Acoustic emission. Dielectric response 6. Emitter-detector pairs 7. Electro-mechanical impedance Figure 1 Overview of damage detection technologies Vibration sensors (Figure 1), such as acceleration and velocity transducers, have been used for a long time to monitor the vibration level and its frequency spectrum at critical locations. By detecting changes in the vibration signature, the appearance of incipient damage can be inferred. The strain monitoring sensors (e.g., conventional strain gages or fiber optic sensors) are used as an alternative way of recording vibrations. The peak-strain at critical locations can be recorded with the special peak-strain gages recently developed by Strain Monitoring Systems, Inc (Thompson and Westermo, 1994). The acoustic emission sensors are another example of passive technology. They are able to pick-up the minute pops generated by the crack as it advances by tearing through the material. The

3 dielectric sensors are capable of passively detecting the structural changes taking place in a polymeric composite due to insufficient cure, or damage, or moisture absorption. The advent of active materials capable of deforming their shape and dimensions in response to electric, magnetic, and thermal fields has opened new options and opportunities in the field of sensor technologies for nondestructive evaluation (NDE) and health monitoring. It is now possible to advance from passive sensors (e.g., vibration pick-up, acoustic detection microphones, etc.) to active devices that can simultaneously interrogate the structure and listen to its response. Emitter-detector pairs of piezoelectric transducers have been used to send ultrasonic waves through the material and detect the incipient damage using wave signature (Keilers and Chang, 199). Point-wise structural impedance can be also measured by an array of piezoelectric wafer transducers (Rogers and Giurgiutiu, 1997). In this case, the processing of the electro-mechanical impedance spectrum of the transducers identifies the presence of incipient damage (Giurgiutiu and Rogers, 1997). These emerging new technologies pose the promise of identifying incipient damage well before it starts to affect the normal and safe operation of the machinery. ELECTROMECHANICAL (E/M) IMPEDANCE TECHNIQUE Consider a piezo-electric transducer wafer intimately bonded to the surface of a structural member. When excited by an alternating electric voltage, the piezoelectric transducer applies a local strain parallel to the surface (Giurgiutiu and Rogers, 1997). Thus, elastic waves are transmitted into the structure. The structure responds by presenting to the transducer the drive-point mechanical impedance Z str ( ω ) = iωme ( ω) + ce ( ω) ike ( ω) / ω. Through the mechanical coupling between the PZT transducer and the host structure, and through the electromechanical transduction inside the PZT transducer, the drive-point structural impedance directly reflects into the effective electrical impedance as seen at the transducer terminals (Figure 2). v() t = V sin( ωt) PZT wafer transducer i() t = I sin( ωt + φ) F(t) u( t) m e (ω) k e (ω) c e (ω) Figure 2 Electro-mechanical coupling between the PZT transducer and the structure. The electro-mechanical (E/M) impedance technique for health monitoring, damage detection, and NDE (Rogers and Giurgiutiu, 1997) utilizes the changes that take place in the drive-point structural impedance to identify incipient damage in the structure. The apparent electro-mechanical impedance of the piezotransducer as coupled to the host structure is given by

4 2 Zstr ( ω) Z( ω ) = iωc 1 κ 31 ZPZT( ) + Zstr ( ). (1) ω ω In Equation (1), Z( ω) is the equivalent electro-mechanical admittance as seen at the PZT transducer terminals, C is the zero-load capacitance of the PZT transducer, and κ 31 is the electro-mechanical cross coupling coefficient of the PZT transducer ( κ 31 = d13 / s11ε 33 ). The mechanical impedance of the structure is Z str, while that of the PZT transducer is Z PZT. As seen in equation (1), the interaction of the structural and transducer mechanical impedance modifies the effective electrical impedance as measured at the transducer terminals. This frequency dependent process is highly coupled with the internal state of the structure, as reflected in the drive-point mechanical impedance, Z str. The electro-mechanical impedance method is applied by scanning a predetermined frequency range in the hundreds of khz band and recording the complex impedance spectrum. By comparing the impedance spectra taken at various times during the service life of a structure, meaningful information can be extracted pertinent to structural degradation and the appearance of incipient damage. The frequency range must be high enough for the signal wavelength to be compatible with the defect size. 1 RECENT EXPERIMENTS WITH THE ELECTRO-MECHANICAL IMPEDANCE TECHNIQUE Several experiments have proven the ability of the E/M impedance technique to detect damage and localize its position in a variety of applications. (a) (b) Figure 3 High-frequency electro-mechanical impedance health monitoring testing of bolted joints: (a) Four shear lap joint tension specimens. (b) Close-up view of tone of the joints showing bolt-heads, washers, and the placement of two PZT active sensors (Giurgiutiu, Turner, and Rogers, 1999) High-frequency electro-mechanical impedance health monitoring testing of bolted joints A major challenge for the successful performance of damage detection tests is the need to create controlled-damage situation. In general, damage creation is

5 irreversible, and hence cannot be repeated on the same specimen. However, a special situation arises in the case of bolted joints. In bolted joints, damage can be created and eliminated by modifying the bolted joint parameters, such as the tension in the bolt, or the presence/absents of stiffening washers. Figure 3 presents experiments performed to correlate the E/M impedance readings with the presence of damage in the most common structural joint the bolted joint. Results of these investigations are shown in Figure 4. As the damage progresses, the RMS impedance changes. The undamaged condition bolt+nut+washer was taken as reference state. After removing the washer, the local stiffness of the joint decreased. Consequently, the frequency spectrum of the E/M impedance shifted, and some of the local impedance peaks disappeared. The recorded RMS change in the E/M impedance was very large (up to 14%). Further damage was induced by removing the bolt completely. Another significant change was observed in the value of E/M impedance. This experiment has proved beyond any reasonable doubt the E/M impedance can be directly correlated with the progression of structural damage. The attractiveness of the bolted joint specimen as a proof-ofconcept demonstration of damage detection lies in the ability to induce and then remove damage. Thus, in our case the experiments were repeated, and consistent results were obtained. 17 Impedance Changes for Lap Joint 3 (a) Re (Z) (Ohms) No Bolt Bolt Bolt + Washer Frequency (khz) (b) RMS Impedance Change RMS Impedance Change Comparisons for for M-Bond 2 Adhesive 16% 14% 12% Lap Joint 1 1% Lap Joint 2 Lap Joint 3 8% Lap Joint 4 6% 4% 2% % % Bolt, Nut, Bolt + Nut Nut Free Washer Figure 4 Results of the high-frequency (E/M) impedance health monitoring testing of bolted joints: (a) electro-mechanical impedance signatures for three structural health situations: no damage (bolt+washer); partial damage (bolt only); extensive damage (no bolt). (b) Correlation between RMS impedance change and specimen structural health (damage progression). (Giurgiutiu, Turner, and Rogers, 1999) High-frequency electro-mechanical impedance health monitoring testing of composite overlays on concrete substrates (civil infrastructure repairs/strengthening/rehabilitation) The composite repair of metallic and concrete structures is an area in which adhesive joining techniques are of great importance. Due to in-service degradation, a structure can become damaged and develop cracks. Short of complete replacement, the structure can be locally repaired through the application of composite patches consisting of advanced high-strength fibers

6 embedded in a polymeric resin. These composite patches act as crack stoppers. The load field around the repaired crack is redirected through the composite patch. Thus, the crack is arrested, and the crack growth is stopped. Composite overlays are thin sheets of fiber reinforced polymeric material (1/8-in to 1/4-in) adhesively bonded to conventional construction engineering materials. Candidate polymeric systems include polyester, vinylester, epoxy and phenolic. Fibers can be glass, carbon, Kevlar, or combinations thereof. Glass and Kevlar fibers come in a variety of forms including weaves and non-woven fabrics. Carbon fibers can be woven, but common usage relies on unidirectional prepregs. The composite may be applied as: (a) wet lay-up; or (b) precured panels; or (c) partially cured prepregs. For wet lay-up and prepreg systems, the adhesive is the polymeric resin itself. For precured rigid panels, separate adhesive material needs to be used. Structural upgrades with composite overlays offer considerable advantages in terms of weight, volume, labor cost, specific strength, etc. However, one critical issue raised by the structural engineers concerning the use of composites in infrastructure projects is the still unknown in-service durability of these new material systems. Their ability to safely perform after prolonged exposure to service loads and environmental factors must be ascertained before wide acceptance in the construction engineering community is attained. The E/M impedance technique has been evaluated as a potential health monitoring method for the composite overlays repairs, strengthening, upgrade and rehabilitation of nation s aging infrastructure. Figure shows the type of specimen used to correlate the E/M impedance readings with crack propagation in the bond between a composite overlay and a concrete infrastructure substrate. Figure 6 illustrates the correlation between crack propagation and E/M impedance reading as measured during these experiments. The specimen underwent controlled amounts of cracking in a DCB-type test performed in an MTS universal testing machines. A number of cracks of increasing length were generated (Figure 6b). The high-frequency E/M impedance spectrum (Figure 6a), Composite overlay Concrete substrate Support fixture PZT active sensors Reaction bolts (a) (b) Figure Test specimen developed at USC for testing disbond detection using the E/M impedance technique and piezoelectric active sensors: (a) side view showing support fixture, concrete brick and composite overlay; (b) bottom view showing retention bolts (Giurgiutiu, Whitley, and Rogers, 1999). as measure by the active transducers placed on the composite overlay, remained undisturbed until the crack front came into the very proximity of the transducer. The changes in the E/M impedance spectrum clearly detected the presence of the

7 crack. As the crack progressed, the E/M impedance spectrum of the sensors left behind the crack front remained, again, unchanged, while the sensors ahead of the crack tip became sensitive to the approaching front. Impe dance (R ez ) Im pe dan ce (R ez ) Im pe da nc e (R ez ) Baseline Signature Impedance Fr equency ( Hz) Freq uenc y (Hz) Crack 3, S en sor 1 Impedance Crack intiated Impedance Frequency (Hz) Impedance (ReZ) Impedance (ReZ) Im pe dance (ReZ) Baseline Signature Impedance Frequ ency (Hz) Crack intiated Impedance Fre que ncy ( Hz) Crack 3, Sensor 2 Impedance Freque ncy (Hz) Impedance (ReZ) Im p e dan ce (R e Z ) Im pedan ce (R ez) Baseline Signature Impedance Fre quency (H z) Crack intiate d Impedance Frequency (Hz) Crack 3, S en sor 3 Impedance Fre que nc y (H z) Crack Initiation 36.mm Crack 1 49mm Crack 2 61mm Crack 3 83.mm Crack 4 19.mm Crack 127mm Crack 6 18mm Figure 6 Results of the high-frequency electro-mechanical impedance health monitoring testing of composite overlay repairs of civil infrastructure systems indicate a direct correlation between the crack position and the E/M impedance spectrum. The E/M impedance spectra are arranged in three columns, corresponding to the active sensors 1, 2, and 3, as shown on the specimen sketch. As the crack advances from Crack 1 through crack 6, the sensor E/M spectra changes accordingly (Giurgiutiu, Whitley, and Rogers, 1999). Disbond Sensors for Adhesively-Bonded Rotor Blade Structures The E/M impedance technique has been used to detect disbonds between adhesively bonded structural elements such as the Apache 64H helicopter rotor blades. These rotor blades have a built-up construction consisting of preformed

8 sheet-metal members adhesively bonded with high-performance structural adhesive. In-service disbonds appear between the structural elements due to inflight vibrations. In our experiments, a rear rotor blade section was considered (Figure 7). The section was instrumented with several E/M disbond gauges surface mounted with standard strain-gauge installation materials and procedures. #3 disbond gauge: Main spar #1 disbond gauge: Trailing edge (a) Real Part of Impedance, Ohms 4 4 Damage Index at Location #1 DI = 39.4% 3 3 Disbonded 2 As received (b) Real Part of Impedance, Ohms Damage Index at sensor #3 DI = 286.1% Disbonded As received Frequency, khz Frequency, khz Figure 7 E/M impedance disbond gauges were placed on a rear rotor-blade section to detect delamination between the adhesively bonded structural elements. Comparison of the E/M impedance response curves measured for the as received and disbonded structure shows clear identification of the disbond: (a) spectrum of the disbond gauge at location #1; (b) spectrum of the disbond gauge at location #3. CONCLUSIONS The electro-mechanical impedance technique was presented as an emerging technique for in-situ health monitoring and NDE of complex machinery. Its fundamental principles were been presented and its benefits highlighted. Demonstration experiments conducted on a variety of applications (precision machinery parts, real-life aircraft junction, composite patch repairs, etc.) were presented and discussed. The importance of this method for on-line health monitoring and damage detection of complex machinery is apparent. Further investigation is warranted to strengthen the theoretical understanding of the method and to extend its applicability to other engineering areas.

9 ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of the following agencies: National Science Foundation through NSF/EPSCoR Cooperative Agreement No. EPS ; US Army Construction Engineering Research Laboratory Contract # DACA88-98-K-1; and SC National Guard through Agreement #1/1998. The authors would also like to express thanks to Shannon Whitley and Maurice Turner, Undergraduate Research Assistants, for their contribution to data collection and data processing. REFERENCES 1. Giurgiutiu, V., Reynolds, A., Rogers, C. A., Chao, Y. I., and Sutton, M. A., 1998, E/M Impedance Health Monitoring of Spot-welded Structural Joints, Adaptive Structures and Materials Systems, AD-Vol. 7, MD-Vol. 83, ASME, Aerospace and Materials Divisions, 1998, pp Giurgiutiu, V., and Rogers, C. A., 1998e, Adaptive NDE and Health Monitoring Methods for Civil Engineering Structures, Proceedings of the 1998 Structural Engineers World Congress, July 18-23, 1998, San Francisco, CA, paper # T176-4, abstract on pp. 449, full length paper on CD-ROM. 3. Giurgiutiu, V., and Rogers, C. A., 1998d, Principles and Applications of the Electro- Mechanical Impedance for Structural Health Monitoring and NDE, Proceedings of the 13 th US National Congress of Applied Mechanics, June 21-26, 1998, Gainesville, FL. 4. Giurgiutiu, V., and Rogers, C. A., 1998c, Application of the Electro-Mechanical (E/M) Impedance Method to Machinery Failure Prevention, Proceedings of the 2 nd Meeting of the Society for Machinery Failure Prevention Technology, March 3 April 2, 1998, Cavalier Hotel, Virginia Beach, VA, pp Giurgiutiu, V., and Rogers, C. A., 1998b, Recent Advancements in the Electro-Mechanical (E/M) Impedance Method for Structural Health Monitoring and NDE, Proceedings of the SPIE s th International Symposium on Smart Structures and Materials, 1- March 1998, Catamaran Resort Hotel, San Diego, CA, pp Giurgiutiu, V., Lyons, J., Petrou, M., Dutta, S., and Rogers, C. A., 1998a, Strength, Durability, and Health Monitoring of Composite Overlays on Civil Engineering Structures, Proceeding of the International Composites Expo ICE-98, Nashville, TN, January 19-21, 1998, pp. 13.D.1-13.D Giurgiutiu, V., and Rogers, C. A. 1997, "The electro-mechanical (E/M) impedance method for structural health monitoring and non-destructive evaluation", International Workshop on Structural Health Monitoring, Stanford University, CA, September 18-2, 1997, pp Rogers, C.A., and Giurgiutiu, V., 1997, Electro-Mechanical (E/M) Impedance Technique for Structural Health Monitoring and Nondestructive Evaluation, Mechanical Engineering Department, University of South Carolina, OTT Disclosure ID No of 7/2/ Giurgiutiu, V., Reynolds, A., and Rogers, C. A., 1998, Experimental Investigation of E/M Impedance Health Monitoring of Spot-Welded Structural Joints submitted for publication to the Journal of Intelligent Material Systems and Structures, July Giurgiutiu, V., Turner, M., and Rogers, C. A., 1999, High-Frequency Electro-Mechanical (E/M) Impedance Health Monitoring Testing of Bolted Structural Joints, to be submitted to the Journal of Intelligent Material Systems and Structures 11. Giurgiutiu, V., Whitley, Shannon, and Rogers, C. A., 1999, Health Monitoring of Composite Overlay Repairs of Civil Infrastructure Systems using the High-Frequency Electro-Mechanical (E/M) Impedance Technique in preparation for the Journal of Intelligent Material Systems and Structures