Loading...

CENTER FOR NON DESTRUCTIVE EVALUATION

RESEARCH THEMES

UBIQUITOUS SENSING

Redefining Material State Awareness

The current state of the art in fiber optic sensing relies on Fiber Bragg Grating and similar discrete sensing elements. While effective for localized measurements, they face scalability challenges in large-area monitoring due to increased complexity and cost. Continuous sensing provides a more efficient alternative, enabling measurements along the entire sensing medium. Existing distributed techniques, such as Rayleigh scattering, offer wide coverage but are limited in spatial resolution. Piezoresistive nanocomposites, such as GNP-doped PMMA, present a complementary approach by enabling conformal, surface-level sensing. SEM analysis reveals flake-like GNPs (~20–50 µm) forming partially interconnected conductive networks within the matrix. These networks are sensitive to mechanical deformation, where changes in inter-particle spacing lead to measurable variations in electrical resistance, enabling effective monitoring of damage evolution. Recent advances in fiber optic sensing and ultrasonic waveguide technologies at the Centre for Non-Destructive Evaluation (CNDE) further address these limitations by improving both resolution and operational range. Developments such as high-resolution distributed acoustic sensing (DAS) and waveguide-based ultrasonic measurements at temperatures up to 1400 °C significantly expand sensing capabilities. Together, these approaches point toward next-generation sensing paradigms that combine large-area coverage, high resolution, and enhanced environmental robustness, transforming materials and structures into truly ubiquitous sensing systems.

Structured Materials for Imaging

Pushing Imaging to Its Physical Limits

NDE 5.0 leverages engineered and structured materials, including metamaterials, to overcome classical resolution limits in ultrasonic imaging. These materials enable enhanced wave manipulation, sub-wavelength focusing, and improved defect sensitivity. Recent demonstrations of ultrasonic super-resolution imaging beyond the diffraction limit, with reported resolutions approaching λ/75 under controlled conditions, highlight the potential of material-assisted imaging for identifying early-stage damage and micro-scale defects that are inaccessible to conventional ultrasonic techniques. In bulk media, however, high-resolution evaluation at greater depths remains a fundamental challenge, particularly for detecting microscopic defects in critical domains such as quantum materials, nuclear systems, aviation, and biomedical applications. While electromagnetic methods like X-ray imaging provide high resolution, they suffer from limited penetration, high cost, and ionizing radiation. Ultrasound offers a cost-effective, non-ionizing alternative with better penetration, but conventional bulk ultrasonics is constrained by the diffraction limit (λ/2), limiting its ability to resolve micron-scale features. Advanced approaches such as scanning acoustic microscopy improve resolution but are restricted to surface inspection. To address this, metamaterial-based lenses (metalenses) enable recovery of evanescent waves for sub-wavelength imaging. Building on this, CNDE has developed micro-structured holey metalenses combined with sub-micron laser-based detection to achieve ~50 micron resolution at MHz frequencies, enabling deeper, high-resolution inspection beyond conventional limits.

Edge Intelligence & Soft-Sensing

Increasing Fidelity of Measurements

Numerical computational methods such as Finite Element, Finite Difference, Finite Volume, and Boundary Element methods are widely used for solving complex physical problems; however, their high computational cost and resource requirements often restrict them to offline execution. This work proposes a disruptive paradigm that translates these numerical models onto edge computing platforms, including GPUs and TPUs, enabling real-time and resource-efficient computation. Over the past two years, the CNDE group has advanced this approach and demonstrated more than seven orders of magnitude reduction in computational time for problems in wave propagation, phased array ultrasonic testing (PAUT), and infrared thermography, while using significantly lower computational resources. The core idea involves leveraging numerical simulations to train advanced AI models, which can then perform rapid inference directly at the sensor level, achieving efficient and scalable computation. This framework has already demonstrated up to 1000% extrapolation capability across both 2D and 3D problem domains. Building on this, edge-enabled intelligence is expected to enhance last-mile decision-making by providing real-time analytical capabilities at the sensing stage. Furthermore, physics-based soft sensing approaches are explored to improve data fidelity by augmenting sparse measurements and enabling reliable predictions in inaccessible or harsh environments. This integrated methodology constitutes the Data-driven Simulation-Assisted Physics-Learned AI (DPAI) approach developed at CNDE. In parallel, CNDE is actively exploring the use of quantum computing for early-stage signal classification, where quantum-enhanced learning models are being investigated to improve feature representation and enable faster, more accurate identification of defect signatures.

Multi-Modal Inspection

Enriching Information Boundaries

NDE 5.0 recognizes that no single inspection modality is sufficient for the reliable assessment of complex materials and structures. Multi-modal inspection integrates complementary techniques such as Digital X-ray, Phased Array Ultrasonic Testing (PAUT), Infrared Thermography (IRT), Acoustics, and Terahertz (THz) imaging and Time Domain Spectroscopy (TDS), supported by advanced data-fusion strategies to enhance information richness and reliability. By combining modalities with different penetration depths, sensitivities, and spatial resolutions, a more comprehensive understanding of defects and material properties can be achieved. At CNDE, this paradigm has been demonstrated through integrated Digital X-ray and PAUT analysis with automated data fusion, enabling improved defect characterization across multiple spatial scales. Extending this, recent work on Thermal Barrier Coatings (TBCs) integrates IRT and THz TDS, leveraging their complementary strengths in surface sensitivity and subsurface penetration. To address data scarcity and generalization challenges, CNDE employs simulation-assisted multimodal frameworks that combine physics-based models with experimental data. Advanced deep learning architectures, including attention-based models, enable effective fusion of heterogeneous data with varying spatial and temporal resolutions. This approach enhances defect detectability, improves property estimation, and reduces uncertainty, enabling more reliable and intelligent inspection systems aligned with the NDE 5.0 vision.

Remote and Pervasive Inspection

Improving Reliability and Safety

NDE 5.0 emphasizes improving safety and reliability through remote and pervasive inspection systems capable of operating in hazardous and inaccessible environments. These approaches significantly reduce human exposure while ensuring consistent, repeatable, and high-quality inspection outcomes. By enabling continuous and real-time monitoring, they also support early detection of defects and improved decision-making for critical assets. At CNDE, this vision is being realized through the development of integrated remote inspection platforms that combine advanced sensing, edge intelligence, and autonomous operation. Notably, CNDE has developed first-of-its-kind bio-inspired robotic systems and autonomous shop robot, demonstrating seamless integration of sensing, intelligence, and deployment in real-world environments. The focus is on creating unified inspection solutions where sensing, intelligence, and deployment capabilities are tightly integrated to deliver reliable and scalable NDE. Such systems enable last-mile inspection and decision-making, ensuring safe, efficient, and human-centric deployment of NDE technologies across diverse applications.

CENTER FOR

NON-DESTRUCTIVE EVALUATION

Email: cnde.in@gmail.com

Phone: 044 2257 5688

Room 312, Machine Design Section,

Indian Institute of Technology Madras

Chennai 600036 India

© 2024 COPYRIGHT BY CNDE IIT MADRAS.