TL;DR
This paper presents a novel thermography technique that uses a Digital Micromirror Device (DMD) to create complex, spatially and temporally modulated heat patterns on a target. This allows for the detection of subsurface defects, particularly those oriented perpendicular to the surface, which are difficult to find with traditional methods. The approach enables simultaneous interrogation of large areas and offers potential for system identification.
What Problem Does This Paper Address?
Traditional active thermography techniques, like pulse and lock-in thermography, typically involve uniform spatial illumination of a target. While effective for detecting defects like delaminations that are parallel to the surface, these methods struggle with defects oriented perpendicular to the surface. This is because they primarily generate thermal gradients that propagate through the material, rather than laterally across it. Previous attempts to address this involved scanning a pulsing laser or using multiple line sources with phase shifts, but these were often complex or limited in their ability to interrogate large areas simultaneously. This paper aims to overcome these limitations by using a DMD to create arbitrary, frequency-multiplexed heat sources, allowing for more versatile and efficient detection of challenging subsurface features.
Methods
The core innovation of this paper is the use of a Digital Micromirror Device (DMD) to spatially and temporally modulate a near-infrared (NIR) laser. The DMD, essentially a high-speed array of mirrors, can be programmed to create intricate heat patterns on the target surface. The key idea is to modulate different parts of the target with heat sources at different frequencies. This "frequency multiplexing" allows for the simultaneous interrogation of a large area.
The experimental setup involves a NIR laser homogenized and then projected onto the DMD. The DMD, controlled by a computer, then images the modulated laser onto the target. A thermal imager records the surface temperature response. The data analysis involves demodulating the thermal response at each specific drive frequency. This process effectively isolates the thermal signature of each individual modulated heat source, allowing for the analysis of thermal diffusion and the detection of anomalies caused by subsurface defects. The authors also developed a 2D finite difference model in MATLAB to simulate heat diffusion and validate their approach.
Datasets
Not discussed in the paper. The experiments were conducted on a piece of copper-plated FR-4 with a spray-painted black surface to enhance emissivity.
Key Insights and Ideas
- Frequency Multiplexing for Spatial Control: The DMD's ability to modulate different areas at different frequencies is the central breakthrough. This allows for simultaneous, localized heating and subsequent analysis, overcoming the limitations of uniform illumination.
- Detecting Perpendicular Defects: The technique shows promise for identifying defects oriented perpendicular to the surface, which are notoriously difficult to detect with conventional thermography. The lateral thermal diffusion patterns created by the modulated sources are sensitive to these types of anomalies.
- Arbitrary Heat Source Generation: The DMD offers unprecedented flexibility in creating complex heat source patterns, moving beyond simple pulses or uniform sinusoidal heating. This opens doors for more sophisticated interrogation strategies.
- System Identification Potential: Beyond defect detection, the ability to precisely control and analyze thermal responses at different frequencies suggests applications in material characterization and system identification.
Key Results
- Simulations: Modeling demonstrated that a line source excitation could reveal a simulated crack by impeding lateral thermal diffusion, whereas uniform illumination was blind to it. Simulations also showed that distinct regions of influence for line sources at different frequencies could be analyzed independently.
- Experiments: Preliminary experiments successfully demonstrated the ability to demodulate the thermal response at different frequencies. The amplitude traces indicated thermal superposition, meaning the responses from different sources didn't significantly interfere. However, the phase response showed some deviation from simulations, attributed to wider line sources and potential system misalignments.
- Logo Demonstration: A two-dimensional demodulation example successfully distinguished between two university logos projected at different frequencies, showcasing the technique's ability to resolve spatially distinct features.
Limitations and Future Work
- System Misalignment and Aberrations: The experimental setup suffered from misalignments between the laser, DMD, and target, leading to non-uniform illumination and ghost images. This affected the symmetry of amplitude decay and the accuracy of phase measurements.
- Thermal Imager Linearity: The thermal imager used was not radiometrically calibrated, though the authors note this is acceptable as long as the response is linear.
- Focusing Issues: The target had to be canted to allow observation by the thermal camera, leading to focusing errors and image spreading.
- Future Work: The authors propose improving the physical setup to minimize aberrations, potentially by using a more conventional beam path and a spatial filter. They also suggest exploring alternative demodulation approaches, such as a least-squares method, which could be more computationally efficient than FFT for a finite number of frequencies. Further research will focus on testing with simulated defects and exploring the system identification capabilities.
Summary updated