N00014-11-1-0271: Predictive Modeling of Structural Sensing for Navy Applications Dr. Victor Giurgiutiu University of South Carolina Laboratory for Active Materials and Smart Structures (LAMSS) ONR Structural Composites and NDE Program Review Wednesday August 20, 2014 Objective Global objective: develop the science for unprecedented SHM systems based on novel piezoelectric and optical sensing principles and methods Sensor development • Piezo-optical sensing • Resonance-enhanced AE sensing (the ring sensor) • High-rate multi-channel ultrasonic FBG sensing Structural state assessment • Predictive modeling of wave-damage interaction • Connect sensor signal to structural state Software development • Integrated SHM software package • Signal processing Specific objectives: (a) Predictive modeling of structural sensors response to changes in the structural state and/or the presence of structural flaws or damage (b) Development of novel guided-wave sensors combining piezo-optical sensing with mechanical resonance in 100 – 300 kHz ultrasonic range Hardware development • Combine piezo and optical sensing principles 2 Technical Approach Predictive modeling: • Develop fast analytical modeling of 2-D guided waves propagation in interaction with linear and nonlinear damage • Combine local FEM modeling of realistic damage with fast analytical modeling of the global wave propagation in a hybrid global-local (HGL) approach Develop new sensor concepts that combine piezo-optical sensing with resonant amplification: • Analytical and FEM modeling • Prototyping • Extensive testing for ultrasonic pitch-catch, impact detection, and acoustic emission detection 3 HGL Predictive Modeling of 2-D Wave Propagation in Interaction with Damage G1S (ω , RIN ) = −π i a 2κ PWAS (ω ) J1 (ξ S a ) N S (ξ S ) (1) S H1 (ξ RIN ) ∑ 2µ DS′ (ξ S ) ξS G1A (ω , RIN ) = −π i a 2κ PWAS (ω ) J1 (ξ A a ) N A (ξ A ) (1) A H1 (ξ RIN ) ∑ 2µ D′A (ξ A ) ξA Giurgiutiu V., [Structural Health Monitoring with Piezoelectric Wafer Active Sensors], Second Edition ∞ ur = ∑ an ( z ) H1(1) (ξ n r ) e − iωt n =1 Giurgiutiu V., [Structural Health Monitoring with Piezoelectric Wafer Active Sensors], Second Edition Glushkov E., et al. “ Lamb wave excitation and propagation in elastic plates with surface obstacles: proper choice of central frequencies. (2011) For comparison with laser measurements 4 Analytical Wave Propagation: WFR-2D Scatter Information Platform WFR Main Interface Real time sensing signal Parameter control panel Dispersion and tuning curves Excitation control T-PWAS properties Module Spatial Propagation Solver 5 Local FEM w NRB: WDICs Extraction ∞ Sensing boundary Lamb waves S0, A0: ur = ∑ an ( z ) H1(1) (ξ n r ) e − iωt n =1 ∞ Shear horizontal waves, SH0: uθ = ∑ bn ( z ) H 0( ) (ξ nSH r ) e − iωt 1 n =1 Separation of guided wave modes: uθT + uθB urT + urB urT − urB A0 SH 0 = ; uSC = ; uSC ; 2 2 2 S0 = uSC Local FEM analysis = WSC WTOTAL − WIN Pristine Relationship between incident and scattered waves: NRB Α − iϕ IN Β ⋅ CΑΒ (ω , θ ) e − iϕΑΒ (ω ,θ ) ⋅ H m(1) (ξ Β r ) = uIN e uSC (θ ) e−iϕSC (θ ) Β NRB Sensing nodes NRB Α CΑΒ (ω , θ ) e NRB Excitation nodes Damaged NRB WDICs: − iϕ ΑΒ (ω ,θ ) Β uSC (θ ) 1 − i∆ϕ θ Β = e ΑΒ ( ) ; ∆ϕ ΑΒ (θ ) = ϕ SC (θ ) − ϕ INΑ Α (1) Β uIN H m (ξ r ) Β uSC (θ ) 1 CΑΒ (ω , θ ) = Α (1) uIN H m (ξ Β r ) NRB NRB NRB ϕ ΑΒ (ω ,θ ) = ∆ϕ ΑΒ (θ ) − ∠ 1 H m(1) (ξ Β r ) −∠ H m(1) ( 0+ ) 1 Harmonic FEM analysis: all frequencies in one run! 6 Discussion of WDICs S0 incidence at 200 kHz 400 kHz A0 incidence at 200 kHz • • • • 300 kHz 200 kHz S0-incident wave results: Mode conversion is captured WDICs depend on direction WDICs depend on frequency Phase information is essential for correct constructive and destructive wave interaction A0-incident wave results – similar obs. 7 Detection Sensitivity vs. Frequency Example: Crack growth in a rivet hole No damage 1.6-mm cracks WDIC_pristine WDIC_damaged WDIC_crack = WDIC_damaged – WDIC_pristine CSS _crack ϕ SS _ crack 90 90 2 120 1.5 60 1 150 120 30 300 90 60 200 150 0.5 120 30 150 100 180 0 210 180 330 240 CSSH _ crack 400 0 210 300 180 330 240 270 2.5 60 2 1.5 1 0.5 300 400 300 60 200 150 30 30 100 330 240 270 90 120 0 210 ϕ SSH _ crack 180 0 210 330 240 300 300 270 270 2.5 2 1.5 60 1 150 2 30 0.5 180 0 210 Amplitude 90 120 X: 482 Y: 2.039 1.5 482 kHz -- Yes! X: 400 Y: 0.8081 1 330 0.5 240 X: 328 Y: 0.07469 300 270 0 0 200 X: 776 Y: 0.7684 328 kHz -- No! 600 400 Frequency (kHz) 800 1000 8 PIEZO-OPTICAL RING SENSOR 9 Piezo-Optical Ring Sensor -- Concept Enhance acousto-ultrasonic crack out of plane waves propagation detection Employ mechanical AE source resonance principles to FBG • Enhance sensitivity at desired frequencies • Filter out low-frequency noise and vibration Bestow omnidirectionality onto FBG optical sensing Permit design for specific frequency ranges T2 T1 T3 R 10 Ring Sensor - Overview Analyzed and designed three rings: 100 kHz, 200 kHz, 300 kHz (any desired frequency may be accommodated) Fabricated the 100 kHz and the 300 kHz Ring Sensors. Tested for: • • • • Resonance: E/M impedance and chirp FRF Pitch-catch: comparison with conventional FBG sensors and PWAS Impact detection (160-mg steel ball) Simulated acoustic emission (pencil lead breaks -- PLB) detection 100 kHz 8.00-mm dia 316 stainless steel Mill machining of internal ellipse shape $100 for one (USC machine shop) 100 kHz 300 kHz 300 kHz 4.35-mm dia 304 stainless steel EDM machining of internal ellipse shape $500 for two (Alpha Manuf. Inc.) 11 EMIS Tests: 300 kHz Ring Sensor 277 kHz fundamental resonance 2.4 10 Impedance real (Z) X: 277.1 Y: 245.7 2.3 10 50 100 150 250 200 Frequency (KHz) 300 350 400 Chirp FRF Tests: 100 kHz–Ring Sensor Linear chirp frequency-time variation PWAS FBG on side FBG through hole Frequency-Domain Time-Domain 102 kHz FFT 13 Pitch-Catch Tests Longitudinal path Transversal path FBG on plate is directional! FBG on Ring sensor is not directional! FBG bonded onto the plate FBG bonded onto the ring sensor A0 A0 Transversal path S0 Transversal path Longitudinal path S0 A0 Longitudinal path A0 14 Impact Detection: Plate PWAS vs. Ring Sensor PWAS Plate PWAS Plate PWAS: Time Domain Plate PWAS: Frequency Domain Ring PWAS Ring PWAS: Time Domain Ring PWAS: Frequency Domain Longitudinal Transverse 15 Impact Detection: Narrow-Band Ring Sensor Capability PWAS: Time Domain PWAS: Frequency Domain Longitudinal Transverse 160-mg small steel ball dropped from 100 mm height 16 AE Detection: Narrow-Band Ring Sensor Capability PWAS: Time Domain PWAS: Frequency Domain Longitudinal Transverse 0.5-mm pencil lead break (PLB) at 100 mm from sensor 17 FBG on Plate vs. FBG on Ring Sensor FBG Ring Sensor chirp response shows a narrow resonance peak FBG Ring Sensor is omnidirectional, whereas FBG on plate is only sensitive along its axis FBG Ring Sensor attenuates undesired frequencies FBG Ring Sensor can be designed for specific frequency range, including frequency range of acoustic emissions 18 Accomplishments Fast analytical modeling of 2-D guided waves propagation in interaction with linear and nonlinear damage Local FEM frequency-domain analysis of realistic damage to determine the damage interaction coefficients for efficient HGL of realistic structures Novel omnidirectional piezo-optical ring sensor for enhanced guided-wave SHM developed ab initio through analytical and FEM modeling, prototyping, and extensive testing for ultrasonic pitch-catch, impact detection, and acoustic emission 19 Recent Publications, Patents, Awards Giurgiutiu, V.; Roman, C.; Lin, Frankforter, E. (2014) “Omnidirectional Piezo-Optical Ring Sensor for Enhanced Guided Wave Structural Health Monitoring”, Smart Materials and Structures, under review, manuscript # SMS-101065 Shen, Y.; Giurgiutiu, V. (2014) “WaveFormRevealer: An analytical framework and predictive tool for the simulation of multi-modal guided wave propagation and interaction with damage”, doi: 10.1177/1475921714532986 Structural Health Monitoring – An International Journal, May 13, 2014 Shen, Y.; Giurgiutiu, V. (2014) ‘Predictive Modeling of Nonlinear Wave Propagation for Structural Health Monitoring with Piezoelectric Wafer Active Sensors”, Journal of Intelligent Material Systems and Structures, Vol. 25, No. 4, pp. 506-520 Gresil, M.; Giurgiutiu, V. (2013) “Guided Wave Propagation in Composite Laminates Using Piezoelectric Wafer Active Sensors”, Aeronautical Journal, Oct. 2013, Vol. 117, No. 1196, pp. 971995, Royal Aeronautical Society, London UK Lin, B.; Giurgiutiu, V. (2014) “Development of optical equipment for ultrasonic guided wave structural health monitoring”, SPIE Vol. 9062, paper number 9062-27 Shen, Y.; Giurgiutiu, V. (2014) “WFR-2D: an analytical model for PWAS-generated 2D ultrasonic guided wave propagation”, SPIE Vol. 9064, paper number 9064-40 Giurgiutiu, V.; Gresil, M.; Roman, C. (2013) Acousto-ultrasonic Sensor, US Patent application publication # US 2013/0129275A1 20 AM-MF Materials Opportunities Sensing and excitation capabilities could be embedded into the multifunctional material through additive manufacturing The predictive modeling methodology developed in this grant could be extended to simulate the behavior of AMMF materials: • Hybrid global local (HGL) would provide an efficient modeling tool • Anisotropy should be included in the analytical part of the HGL model • Multi-physics description could be used both directionally to send and transmit structurally interrogative signal • The HGL approach and tuning principles could be extended to energy harvesting AM-MF materials Incorporation of ‘intelligence’ through printed electronics is an essential future need for AM-MF materials. 21 Navy Relevance and Impact Increase in-service safety and reliability Assess structural state to predict operational capability Reduce lifecycle cost of naval assets 22 Significance and Originality The project has strong significance for the Navy efforts in developing structural health monitoring (SHM) methods and technologies. The predictive modeling part of the project would allow the SHM system designer to produce an optimal system and sensor installation with strong damage detection sensitivity and rejection of other confounding effects (loads, vibration, environment, etc.) The novel piezo-optical ring sensor is very original and has not been reported anywhere else yet. It bestows omnidirectionality to optical FBG sensors which are preferred in Navy applications. It enhances certain frequency band while inherently rejecting noise and vibration perturbations. 23 Scientific Merit The project has significant scientific merit. The predictive modeling part of the project has demonstrated a skillful combination of global analytical prediction of wave propagation in a 2-D medium with scatterers combined with a innovative use of local FEM mesh with non-reflective boundaries to determine the wave damage interaction coefficients for complicated damage shapes The novel ring sensor part of the project has skillfully combined analysis and CAD-FEM techniques to predict the behavior of this new sensor concept and then validated these predictions through carefully-conducted professional quality experiments. 24 Risk and Potential Impact The novel piezo-optical ring sensor is very innovative and interesting; it may have a strong impact on in-situ monitoring especially for impact detection and acoustic emission with FBG optical sensors The risk related to this novel sensor resides in the fact that it has only been proven in laboratory. Its implementation in actual applications will have to traverse the ‘valley of death’ between concept and a commercially viable product. To mitigate the risk and open the opportunity for the potential impact to materialize, it might be advisable to use SBIR/STTR or other small 6.2 funding in order to obtain an actual product that can be tested on actual structures in collaboration with Navy Labs. 25