Tarar, Hasan
(2025)
Detection and characterization of in-service damage in lattice structures using ultrasonics.
PhD thesis, University of Nottingham.
Abstract
This thesis investigates the detection, quantification, and localization of in-service damage in additively manufactured lattice structures using ultrasonic techniques. Lattice structures, characterized by their lightweight design and superior strength-to-weight ratio, have gained significant attention across aerospace, automotive, and biomedical industries. However, their intricate geometries and susceptibility to defects during additive manufacturing (AM) and operational use pose critical challenges to ensuring their structural integrity. The research aims to address these challenges by developing a robust methodology for structural health monitoring (SHM) in lattice structures, leveraging advanced ultrasonic testing and machine learning models.
The study focuses on strut-based lattice structures, which are particularly prone to damage such as cracking and breaking of struts, often initiated by inherent manufacturing anomalies like porosity, residual stresses, and delamination. Ultrasonic wave propagation within lattice geometries was analyzed to understand the interaction between high-frequency waves and structural discontinuities. Piezoelectric sensors, known for their precision and sensitivity, were deployed to generate and capture ultrasonic signals, enabling real-time damage detection. The methodology integrates numerical simulations and experimental setups to ensure comprehensive analysis and validation.
Key features were extracted from ultrasonic response signals using advanced signal processing techniques, including principal component analysis (PCA) and energy-based feature extraction. These features served as inputs to a machine learning neural network model, trained to classify the health states of the structures. The results demonstrated the ability of the proposed approach to accurately identify damage states, quantify damage severity, and localize damage zones within complex lattice structures. For damage detection, models achieved a high classification accuracy, distinguishing between healthy and damaged states with over 90\% precision across 2D and 3D lattice configurations. The quantification study showed reliable predictions for the extent of damage, particularly for groups of damaged struts and damaged cells, with models achieving consistent accuracy in these classifications. The developed localization methodologies, using multiple sensors and spatial mapping of damage in different zones of the structure, proved highly effective, achieving damage zone localization accuracy of up to 85\% for cases involving up to three damaged cells. Experimental validation using normalized datasets further affirmed the robustness of the methodology, with experimental predictions aligning closely with numerical simulations for 2D lattice structures.
The results of this research provide a detailed understanding of ultrasonic wave propagation in lattice structures and demonstrate the feasibility of using ultrasonic techniques for SHM in these complex geometries. The proposed methodology was validated through experimental work involving 2D and 3D lattice structures, with findings highlighting the efficacy of the approach in real-world applications.
The novelty of this work lies in adapting ultrasonic SHM techniques to the complex geometries of lattice structures, bridging the gap between existing methods for traditional materials and the unique challenges posed by AM designs. The findings not only advance the theoretical understanding of ultrasonic wave interactions in lattice structures but also provide practical tools for ensuring their structural integrity. This research establishes a foundation for industry standards in the non-destructive evaluation of AM lattice structures, with implications for improved safety, maintenance practices, and operational efficiency.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
