Aerospace composites are vulnerable to Barely Visible Impact Damage (BVID), which is difficult to detect as surface indications are minimal or absent. If undetected, such internal damage can degrade structural integrity and compromise safety, highlighting the need for reliable approaches to detect and characterise impacts. Despite significant advances in this topic, accurate impact energy reconstruction in composite structures remains challenging because the mapping between sensor data and impact parameters is often highly nonlinear, computationally demanding, and ill-posed. This inverse estimation problem is further complicated by limited sensor coverage, measurement noise, uncertain boundary conditions, and the dispersive nature of guided wave propagation in composite laminates. Physics-based models can describe fundamental structural dynamics but are computationally costly, sensitive to modelling assumptions, and difficult to apply to real-world systems, whereas data-driven models can approximate complex relationships efficiently but depend on large, high-quality datasets and often lack robustness or physical interpretability. To address these limitations, this study introduces the Physics-Informed Impact Identification (Phy-ID) framework, which integrates prior knowledge, data-driven inference, and sensor-based observations within a unified modelling strategy. The central objective is to stabilise impact parameter estimation by embedding physically motivated constraints directly into the model formulation rather than relying solely on statistical correlations. The framework builds on the hypothesis that embedding physics into machine learning architectures constrains the solution space towards physically consistent predictions and enhances accuracy, generalisation, and interpretability, even under limited or noisy data. Accordingly, Phy-ID introduces three forms of bias that make the learning process physics-informed: observational bias, derived from physics-based feature extraction using multi-domain signal representations; inductive bias, incorporated through the definition of model architecture to define key physical dependencies; and learning bias, formulated through hybrid loss functions enforcing both physical constraints and agreement between predicted and measured parameters. In this context, observational bias ensures that the input representation reflects impact mechanics and wave-structure interaction; inductive bias acts in the model to incorporate known governing relations and prior constraints; and learning bias regularises optimisation by penalising physically implausible responses. Together, these components integrate physics and data, improving stability under measurement uncertainty, reducing dependence on extensive experiments, and increasing the transparency of model outputs. The framework aims to provide robust and accurate predictions of impact energy while also enabling the inference of physical quantities relevant to impact characterisation. Moreover, by combining feature-level design, physics-guided constraints, and experimental data, Phy-ID addresses the trade-off between knowledge and data by compensating for practical limitations associated with incomplete or noisy datasets. In conclusion, Phy-ID establishes a structured, scalable, and data-efficient approach that bridges the gap between physics-based and data-driven methodologies, advancing reliable, explainable impact identification in SHM of aerospace composite structures.

Structural Health Monitoring (SHM) is crucial for detecting and assessing damage in engineering structures made of composite materials, which are widely used in critical applications such as aircraft and exhibit complex failure modes. The need to reduce maintenance costs, along with the demand for techniques capable of identifying imperceptible damage, highlights the importance of developing advanced SHM approaches. The objective of this work is to refine and advance a patented active structural health monitoring approach, enabling the detection and characterization of imperceptible damage in composite materials through high-frequency strain patterns. The technique combines piezoelectric transducers (PZTs) to excite the structural component and fiber Bragg grating (FBG) sensors to locally measure the resulting strain fields. The sensor signals are filtered to isolate the response caused by the excitation, removing contributions from temperature variations as well as from primary and secondary loads. The resulting strain maps are then analyzed using an artificial intelligence algorithm for damage detection, enabling accurate characterization of imperceptible defects. Preliminary evaluations suggest that the integration of FBG, PZT, and neural network analysis can detect and monitor subtle structural changes in composite materials. The technique is expected to enhance early-stage damage detection, provide insights into damage progression, and support predictive maintenance strategies when compared to conventional methods. This study demonstrates that the combination of advanced sensing technologies with neural networks represents a promising step toward more intelligent, adaptive, and cost-effective SHM systems. The approach has the potential to improve the reliability and efficiency of composite structures and provides a solid foundation for future research and practical implementation in real-world applications.

Composite materials are increasingly adopted in aerospace applications due to their high strength-to-weight ratio, which enables significant weight reduction, improved fuel efficiency, and lower CO₂ emissions. Ensuring the structural integrity of such materials is therefore critical, particularly in flight-critical components such as aircraft wings equipped with Ice Protection Systems (IPS). Within the framework of the EU project PLEIADES, this research presents an efficient and unified methodology for damage detection, identification, and prognostics of composite aerospace structures using vibrational signals from integrated piezoelectric (PIC) sensors.
The proposed approach employs an integrated framework comprising a 1D Timoshenko beam-based [1] electro-thermo-mechanical (ETM) analytical model and a multiparametric Unified Butterfly Optimisation Algorithm (UBOA) to characterise bulk and delamination damages in 3D finite element (FE) beams. The analytical ETM model simulates the vibratory response of electrothermal ice protection laminate comprising heaters sandwiched between fibreglass layers and vitrimer polymer with embedded sensors. The vibratory signal from the ETM model is then optimised by multiparametric UBOA technique which combines local exploitative and global explorative search strategies to synchronise damage detection, localisation and quantification.
In the prognostic phase, the analytical model is extended to assess the influence of damage on the static, dynamic and potentially thermo-mechanical response of the degraded beam. Validation against 3D FE simulations demonstrates that the proposed method achieves substantial computational savings while maintaining a high level of accuracy as shown in Figure 1. This unified framework therefore offers a promising basis for real-time Structural Health Monitoring (SHM) and damage prognostics of composite aircraft wing structures, supporting the goals of PLEIADES toward enhanced safety and reduced maintenance costs through smart sensing and integrated SHM-IPS systems. The above research is supported by project PLEIADES (Advancing Aerospace Composites through Induction Welding & New Vitrimeric Formulations Enhanced by Integrated Photonic Sensors, Providing Data to Digital Supply Chain, SHM, Maintenance) - European Union’s Horizon Europe research and innovation program under grant agreement no. 101192721.

Fatigue crack initiation and propagation remain critical challenges in structural engineering due to their potential to cause unexpected structural failures with severe safety risks and economic consequences. This study proposes a digital-twin-enabled approach for in-service fatigue crack detection and quantification, where real-time experimental data from the physical entity are continuously integrated into the virtual model to maintain synchronization and adaptive model updating. Within this framework, online dynamic response measurements from piezoelectric (PZT) transducers are processed using a Dynamic Piecewise Linear (DPL) algorithm, which facilitates automated and real-time detection of crack initiation during service. Experimental validation under various loading frequencies, stress ratios, amplitudes, and multi-block loading scenarios demonstrates reliable detection of sub-millimetre cracks (<2 mm). For fatigue crack growth prediction, a Physics-Informed Long Short-Term Memory (PI-LSTM) model is developed by embedding Walker’s law as a physical constraint, while Bayesian optimization ensures optimal hyperparameter selection. The continuously updated digital twin shows superior prediction accuracy and generalization across different specimens under different experimental conditions, highlighting its capability for adaptive, real-time structural health assessment. This study advances the development of intelligent digital twins for fatigue life prediction and condition-based maintenance in complex service environments.

The past decade has shown a trend to collect and mine large quantities of our data with the objective to assess the performance of materials and systems for various reasons. This rise of Artificial Intelligence has also largely affected the field of Structural Health Monitoring. A plethora of intelligent data science algorithms has been studied, developing the ability to recognise subtle changes in the (dynamic) behaviour of structures resulting from the presence of early stages of damage or deterioration. More recently, the direction has been altered again, towards inclusion of physics-based information. This is done on various levels and in different ways, but all with the shared objective to mitigate one of the most important drawbacks or limitations of pure data-centric methods: the limited performance in case of poor, incomplete, scares or non-representative data. This lack of data quality is not just solved by collecting more data and data with higher accuracy, but is also an inherent problem for monitoring systems of actual structures: maintenance interventions, securing the safe operation of the structure, imply hardly any data is collected of failed systems.

In addition, the blunt collection of data may work counterproductive if it comes to practical implementation of these methods. The collection and processing of this data is certainly not free of charge, while the additional information (hence the value) of more data is often limited. Hence, research is directed towards selection of relevant and informative features, reducing the data demand, while maximizing the information output.

Although these developments are doubtlessly of great value, a different perspective may provide new ideas on how to overcome some of the challenges of finding the right balance between knowledge and data to optimise the information output of a monitoring system. What is nature's perspective on how to extract information efficiently from large and diverse data sources?

Human sense can be considered as a sensor system with diverse set of sensors and a large variety and flexibility in how these are connected and are mutually communicating. The human sensory system collects a lot of data, but the processing of this data differs fundamentally from the way data is processed in data-centric methods. It is true that a neural networks mimics to some extent the way a brain works: general patterns of behaviour are trained at first, after which new observations are mapped on these patterns to identify the observed behaviour. In reality, human brains use a limited part of the data to guess what is happening. Despite clear evidence that human brains can be fooled, the general functionality of the approach is fairly good to say the least. This work shows the potential of applying this method, by introducing a layered approach and addressing how various sensor source can be combined and how decision can be made regarding the data be used and discarded.

Structural Health Monitoring (SHM) is becoming essential in civil engineering due to its ability to continuously assess the condition of infrastructures and detect potential damage. SHM techniques are generally categorized into data-driven (DD) and model-driven (MD) approaches. DD methods exploit long-term monitoring data to identify the standard structural behaviour and define relevant thresholds for anomaly detection. However, these methods typically lack labelled damage examples to train the algorithms. On the other hand, MD methods rely on updated finite element (FE) structural models to define the expected performance, paying the price of being computationally demanding, requiring substantial technical and economic resources. Recent research has therefore focused on transfer learning (TL) strategies that integrate the strengths of both paradigms. TL provides a powerful knowledge-generalization framework that allows information learned from an abundant, labelled source domain to be transferred to a related target domain where labelled data are scarce. Recent research has therefore focused on hybrid approaches (HAs) that integrate the strengths of both DD and MD paradigms. These strategies are designed to overcome the inherent limitations of each method, compensating for the lack of physical interpretability while mitigating the high computational demand. By merging physical consistency with the flexibility of statistical learning, HAs provide a more robust and reliable tool for SHM. In this work, a HA is applied to a span of a viaduct monitored with multi-year clinometer measurements. First, a FE model is developed and used to generate a large labelled dataset covering multiple damage scenarios. Subsequently, instrumental noise is stochastically injected into the synthetic data. This augmented dataset is then used to train and validate Physics-Aided Surrogate Models (SMs) based on neural networks, to ensure physical consistency while achieving high computational efficiency. Once trained, SMs are applied to real clinometer rotations to infer structural damage indicators. To further demonstrate the reliability of the system, the process is validated by integrating simulated damage scenarios with experimental monitoring data, thereby testing the framework’s robustness against complex, real-world conditions. Results, tested on a multi-set of real bridges, highlight the potential of this hybrid framework as a scalable and effective tool for next-generation SHM systems in bridges and large civil infrastructures.