Structural Wireless Delamination Sensor

Abstract

Composite materials, particularly laminated fibre reinforced polymer composites, have gained widespread acceptance in various industries due to their superior strength-to-weight ratio and corrosion resistance. The phenomenon of cracking between plies/laminae of such a layered composite is commonly referred to as delamination and occurs due to various reasons, such as corrosion and fatigue of the structure. Structural integrity can be enhanced by monitoring delamination and, ideally, to have a sensor that can continuously monitor the delamination extent. The delamination sensor proposed in this thesis (termed as Wireless Interlaminar Nano Sensor, or WINS) is an LC resonant circuit (resonant frequency fₛ = 1 MHz), and unlike prior sensors, is comprised solely of structural materials: a structural epoxy and carbon nanotubes (CNTs). The delamination crack causes a change in capacitance of the sensor leading to a change in its resonant frequency. The wireless sensor operation was demonstrated using an LC resonant circuit implemented on a printed circuit board, which is termed the sensor emulator (SE). A wireless sensing circuit and reader provided by Analog Devices Inc. was used for the initial measurements using the SE and later, the proof-of-concept (PoC WINS) devices. The PoC WINS device is a CNT-polymer nanocomposite based parallel plate capacitor, adhesively bonded between two composite laminates, and connected in parallel to the capacitor of the SE. The PoC WINS device was subjected to loading in the Mode-I configuration to induce delamination crack growth. The quality factor Q of the SE was varied (Q = 18, 3.2, 1.6, 0.8) by adding different external resistors, and a signal was acquired wirelessly for each value of Q as the delamination crack propagated. The wirelessly acquired signal was also sampled (sampling frequency Fₛ = 100 MHz) and analyzed to estimate the resonant frequency of the sensor. The effect of low sampling frequency was studied by downsampling of the acquired signal by a factor of 100. When Q was large (Q= 18), a change of∼2 kHz in the resonant frequency could be detected, corresponding to a change in capacitance of∼100 pF. At smaller values of Q∼1, challenges encountered in wireless signal acquisition were the too-rapid decay of the sensor signal and low signal-to-noise ratio (SNR). A wireless sensing circuit was designed and developed to enable signal acquisition at Q ≤1. The SE was used in the feedback system of a modified Armstrong oscillator (MAO) to obtain a sinusoidal signal of constant amplitude (∼1 V, SNR∼100 dB) even at Q = 0.8. The frequency (f_AO) of the signal wirelessly acquired from MAO is a non-linear function of the capacitance and the quality factor Q of the sensor and was observed to be in the range of 2 MHz. The MAO was tested for its performance using PoC WINS devices. It was observed that capturing the output signal for a duration of∼100 µs was sufficient for the accurate estimation of frequency (standard deviation∼3 Hz). At Q = 0.8 of the sensor, the MAO was able to detect a change in capacitance of 100 pF. To enable the use of low sampling rate (Fₛ = 1 MHz) for wireless signal acquisition, enhance the sensitivity of detecting change in capacitance, and provide direct readout of the change in capacitance of the sensor, the MAO was made part of another circuit termed MAO+. In the MAO+, mixer and filter circuits were used to modulate fAO from∼2 MHz to∼180 kHz and then to∼25 kHz, allowing the use of sampling frequency as low as 50 kHz to estimate the frequency. A phase-locked loop was made part of MAO+ which enabled direct readout of the change in capacitance of the sensor through a 4 1⁄2 digit digital display. The MAO+ was independently tested using PoC WINS devices and was able to detect a change in capacitance (at Q= 0.8 of the SE) of∼10 pF, corresponding to∼200 microns crack advance. This thesis presents the design, implementation, and operation of a wireless sensing circuit that allows signal acquisition at a low quality factor (Q ≤1) without compromising the SNR, demonstrating the first practical (wireless, made out of structural materials) delamination sensor for advanced composites.