Optimization of prominent and panoramic isotropic-material structures for energy absorption through volume fraction and graded feature

• المؤلف الرئيسي: Ngin, Eric Cheong Soon
• مؤلفون آخرون: Dong Zhili
• التنسيق: Thesis-Doctor of Philosophy
• اللغة: English
• منشور في: Nanyang Technological University 2025
• الموضوعات: Computer and Information Science; Engineering; Energy absorption; Numerical modelling; Enhancement features
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/182985
• المؤسسة: Nanyang Technological University

In today’s competitive landscape, staying ahead requires not only the continuous development of new products but also the enhancement of existing systems to ensure relevance. In aviation, the pressing need for lightweight materials and effective health monitoring of aircraft structures, without compromising their mechanical properties, remains a focal point of research. Recent advancements have seen the use of additive manufacturing (AM) to create intricate structures that were previously unattainable, yielding materials with enhanced mechanical properties. Furthermore, the integration of advanced sensors and machine learning algorithms into aerospace systems has improved anomaly detection and analysis. Despite these progressions, significant areas still warrant exploration. This project addresses these gaps through three innovative contributions, detailed as follows. This thesis makes three pivotal contributions (1) addressing the intricacies in developing high-energy absorption structures by pioneering a new general design guideline through integration of features, paired with a thorough investigation of the energy absorption enhancing effects of volume fraction and graded features, using a combination of experiments and finite element validations. The experimental tests analyzed structural behavior under compression while simulations are designed to provide empirical validation of the findings. This is realised through (2) an epochal comprehensive review of all-inclusive energy absorption structures. And lastly, the development of real-time damage detection models (3) to predict damage location and severity in composites through embedded sensors network. A general design guideline for creation of high energy absorption structures is developed through a critical evaluation of energy absorption capacities across leading structural groups which encompass the bulk of energy absorption structures. Tube structures demonstrate superior energy absorption compared to other designs at both 60% and 80% volume fractions with all structures subjected to a fixed displacement distance of 9 mm, primarily due to the buckling of vertical walls which enhances their load-carrying capacity, though they display unstable deformation behavior. For 60% volume fraction, the square tube (STT) and circular tube (CTT) structures demonstrate 41.19% and 33.88% higher energy absorption (EA) and specific energy absorption (SEA), respectively, compared to the simple cubic (SCL) structure, which is the third highest in this category. For 80% volume fraction, the STT and CTT structures demonstrate 10.65% and 7.48% higher EA and SEA, respectively, compared to the re-entrant honeycomb (RHCH) structure, which ranks third in this category. In contrast, simple, body-centered, and face-centered cubic lattice structures, re-entrant honeycomb structures, and open-cell foam structures offer competitive performance due to their design, which facilitates rapid space filling during compression. This design allows for a shorter plateau region and early densification, resulting in stable deformation behavior. The "plateau region" refers to the phase in the material's stress-strain response where energy is absorbed at relatively constant stress due to progressive deformation while "densification" describes the phase where the material experiences significant compaction, leading to a rapid rise in stress as the structure's internal voids collapse and fill. These insights underscore the importance of structural design in optimizing both performance and stability in energy absorption applications by demonstrating the importance of design elements that promote efficient densification and stable deformation. A systematic investigation into the influence of volume fraction on the augmentation of energy absorption capabilities in structures has been meticulously conducted. The body-centered cubic (BCCL) lattice structure, re-entrant honeycomb (RHCH) structure, square tube (STT) design, and open-cell foam (OCFF) structure each exhibit lower initial energy absorption at a 20% volume fraction but show substantial improvements by 40% volume fraction, ultimately surpassing other structures at higher volumes. These structures are characterized by a slow initial space-filling design at lower volume fractions, but they fill more quickly at higher volume fractions, thus achieving early densification, which improves the energy absorption capacity of the structures. For example, BCCL structure exhibited the highest percentage increase in EA and SEA, with improvements of 1254.62% and 577.30%, respectively from 20% to 40% volume fraction. These values increased from 33.68 J and 6.27 kJ/kg at 20% to 456.23 J and 42.49 kJ/kg at 40%. In contrast, the octet-truss lattice (OTL) structure and diamond cubic lattice (DCL) structures showed relatively lower percentage increases in EA and SEA. Specifically, OTL increased by 399.64% in EA and 149.77% in SEA, while DCL increased by 396.34% in EA and 129.87% in SEA over the same volume fraction range. The overall improvements in energy absorption diminish with increasing volume fractions. These insights contribute to understanding that, beyond the common trend of diminishing improvements in energy absorption with increasing volume fractions, structures with lower initial energy absorption but significant gains at intermediate volumes tend to outperform others at higher volumes, highlighting the critical role of structural design in achieving efficient densification and enhanced load-carrying capacity. The impact of graded features on altering the deformation modes of structures to optimize energy absorption performance has been extensively analyzed. While graded features reduce energy absorption in most structures due to top-down deformation and delayed densification, they allow the open-cell foam with graded features (OCF-GF), diamond cubic with graded features (DC-GL) and octet-truss with graded features (OT-GL), to achieve early densification. This results in superior energy absorption at high compression displacements, mirroring the previously finding of the effect of volume fraction on energy absorption enhancement. These insights highlight that while the integration of graded features generally decreases energy absorption in some structures due to top-down deformation which causes delayed densification, structures which benefit from rapid deformation of the less dense top layer to achieve early densification of the structure, exhibit superior energy absorption at high compression displacements, demonstrating the significance of optimizing structural deformation behavior for early densification to enhance energy absorption performance. The total EA of OT-GL and DC-GL is 10.42% and 3.00% higher than that of OTL and DC-GL respectively, while the total EA of OCF-GF is only 0.06% lower than that of OCFF. With further compression, OCF-GF will also absorb more energy than its non-graded counterpart, OCFF. All published articles, totaling 10,439 articles, encompassing the domain of energy absorption structures have been extensively reviewed and systematically categorized to identify prominent enhancement features, to facilitate a representative comparison of the energy absorption capacities of widely recognized structures that encompass the bulk of energy absorption designs. A thorough literature review within the domain of energy absorption structures, leveraging the Web of Science core collection database and employing the keywords "energy absorption" and "structure," yields a staggering 10,439 articles covering all related publications. Initial assessment identifies 21 unique enhancement features prevalent in the literature. Categorizing these articles by these features returns a sum of articles for each distinct group of energy absorption structures. Subsequently, those with publications close to or exceeding 1,000 are painstakingly reviewed to determine the types of structures employed, while the groups representing a minority of energy absorption structures are dropped due to resource limitations. A series of rationalizations consolidates the remaining groups, associating similar structures and removing subsets. Ultimately, lattice, honeycomb, tube, foam, and graded structures are identified as the five most prominent enhancement features for energy absorption structures. The best-ranked structure based on energy absorption performance is the square tube (STT) structure, with an EA value of 2910.3581 J under simulated low velocity compression at 80% volume fraction while the lowest-ranked structure is the diamond cubic lattice (DCL) structure, with an EA value of 1907.6835 J under simulated low velocity compression at 80% volume fraction. This novel endeavor to systematically review, identify, categorize, and rank energy absorption structures based on their core enhancement features and the volume of journal articles published provides insight into the predominant trends and innovations in the field, guiding future research and development towards the most impactful and extensively studied structural designs. In addition, machine learning models capable of real-time prediction of damage location and severity in composite materials have been developed, utilizing data generated from an embedded sensor network within the composite structure, through the collaboration project with Rolls-Royce. The experimental impact tests conducted on a composite plate embedded with a network of piezoelectric sensors generated voltage data that underwent extensive post-processing. This included peak value selection and data normalization to refine the dataset. The normalization process allowed for the creation of an expanded dataset in which voltage readings were matched with key impact parameters, including impact location, impact loading force, and loading cycles. The refined dataset, incorporating both sensor voltage data and impact characteristics, was then used as input features for training and testing Random Forest (RF) and Artificial Neural Network (ANN) models, facilitating accurate prediction of impact parameters. Through meticulous hyperparameter tuning, both models achieve high accuracy in predicting impact damage localization and quantification. With a dataset comprising only 27 sets of impact location, impact loading force, and loading cycles, and the matching sensor voltage data, the RF model achieves an overall prediction accuracy of 75%, with specific accuracies of 100% for predicting x-coordinates of impact location, 66.67% for predicting y-coordinates of impact location, 50% for predicting impact force, and 83.33% for predicting loading cycles. The ANN model achieves an overall prediction accuracy of 62.5%, with 100% for predicting x-coordinates of impact location, 66.67% accuracy for predicting y-coordinates of impact location and 66.67% accuracy for impact force prediction, though it performs poorly for predicting loading cycles. The effective performance of these machine learning models highlights their potential for real-time monitoring of damage locations and severities. This research paves the way toward developing a structural smart skin for efficient and effective maintenance, marking a significant advancement in composite damage assessment and monitoring.