This paper develops a dynamic Knowledge Graph (KG)-augmented Large Language Model (LLM) framework integrated with a Bidirectional Encoder Representations from Transformers-Cascade Relation Extraction (BERT-CasRel) architecture to address key challenges in aviation equipment fault diagnosis, including unstructured maintenance text processing, ambiguous domain semantics, static knowledge constraints, and limited explainable reasoning capabilities. The study first constructs a domain-specific aviation ontology and adopts a context-enhanced BERT-CasRel model to extract high-quality entity–relation triples from maintenance logs and technical documentation. These structured triples populate a dynamic aviation fault KG that supports hierarchical causal inference, subgraph refinement, and in-context learning for adaptive knowledge updating. Structured domain prompting enables bidirectional interaction between LLMs and the KG, facilitating traceable fault chain analysis and accurate root-cause diagnosis. Evaluated on CFM56-5 aero-engine turbine blade fault cases, the BERT-CasRel model achieves a triple extraction F1-score of 0.968, while the integrated LLM–KG framework attains fault diagnosis accuracy exceeding 95%. Benchmarking against conventional and state-of-the-art methods confirms the framework's superiority in extraction accuracy, diagnostic precision, interpretability, and scalability. It delivers strong cross-domain generalization and computational efficiency, mitigates LLM hallucinations, complies with aviation regulations, and provides an interpretable, scalable diagnostic solution while acknowledging limitations in large-scale knowledge iteration and full industrial deployment.

Wheeled robot fault diagnosis is indispensable for ensuring its reliable and safe operations. However, two challenges impede the application of prevalent intelligent diagnosis methods. 1) Multisensor fusion: The complexity of robot movements necessitates multisensor for comprehensive monitoring, generating strong-coupled, and high-dimensional data that complicate both intrinsic relationship mining and effective fusion; 2) Heterogeneous data silos: Dispersibility, heterogeneity and privacy constraints across different robots lead to non-independent and identically distributed (Non-IID) data silos, severely limiting the development of universal diagnostic models. To overcome these two problems, this article proposes a tailored federated physics-informed graph framework (FedMA-PIG). On the client side, the kinematics mathematical model is constructed for each robot, which explores the inter-sensor correlations and forms a physics-informed graph. It enables multisensor data fusion and assists the client in training a local graph neural network. On the federated framework side, a Non-IID federated framework based on a multianchor contrastive mechanism is devised. It employs multiple anchors to capture common knowledge from heterogeneous robot data, guiding feature representations toward corresponding anchors and away from others, thereby promoting consistency and mitigating inter-client data heterogeneity. Comprehensive experiments were conducted on three representative wheeled robots- Mecanum-wheeled, 4WD-wheeled, and Omni-wheeled- distributed across four federated clients. The results demonstrate that FedMA-PIG achieves generalized and superior diagnostic performance compared to state-of-the-art methods.

Spectral entropy (SE) is a promising nonlinear measure for detecting dynamic variations in vibration signals acquired from rolling element bearings (REB). However, in real world scenarios, characteristic spectral features relating to REB fault gets concealed by unwanted frequency components due to the association of heavy environmental noise. Consequently, original SE not only fails to detect incipient REB fault but also fails to monitor the progression of the fault along with predicting the remaining useful life of the faulty REB. Aiming to address aforementioned problems, in this paper, firstly, characteristic spectral features of REB fault is revealed by calculating the spectrum of the adaptively demodulated weighted squared envelope of the corresponding vibration signal. Subsequently, instead of using classical Shannon entropy theory corresponding to original SE, comprehensive prognosis of the analyzed REB health is achieved through the information quantification of the calculated spectrum by incorporating dispersion entropy (DE) theory. In this context, the proposed measure is named as adaptively demodulated dispersive spectral entropy (ADDSE). Two different run to failure REB data have been utilized to verify the effectiveness of the proposed ADDSE. Results show that the proposed ADDSE not only can overcome the limitations of the original SE in comprehensive dynamic prognosis of REB health but also demonstrate superior performance in compare to other conventional measures such as original DE and root mean square (RMS); advanced version of spectral entropy namely cumulative spectrum distribution entropy (CSDE) and three dimensional holo hilbert spectral entropy (MHHSE3D); alternative sparsity based measure namely Gini index (GI).

As a nonlinear measure, Lempel–Ziv complexity (LZC) can be considered as a suitable parameter for characterizing bearing health status by measuring the complexity of vibration signals. However, in continuous monitoring scenario under noisy condition, all components of a multicomponent bearing signal are not equally sensitive toward a change of LZC value. As a result, a direct application of LZC for bearing health monitoring not only suffers from its inefficient early fault warning but also fails to infer the fault progression. In this article, instead of direct utilization of a whole vibration signal, its fundamental component (FC) sensitive to LZC calculation is separated with the help of continuously adjustable parameterized tunable$Q$factor wavelet transform (TQWT). In this context, a study based on sparsity indices has been done for$Q$factor selection of TQWT. Since TQWT uses an oscillation-based bearing FC separation scheme for LZC calculation, the proposed measure is termed as oscillatory Lempel–Ziv complexity (OLZC). Two experimental cases are used for validation. Performance of OLZC is compared with original LZC, representative sparsity indices and recently proposed multiscale symbolic Lempel–Ziv complexity. Results demonstrate that the proposed OLZC can not only overcome the limitations of the original LZC but also performs better than other indices in comparison to continuous monitoring of bearing health.