The present paper deals with bond graph model-based for structural component fault detection and isolation. The structural conditions of fault detectability and isolability are obtained directly from the bond graph using the properties of the bicausality and the causal path. It is shown that the monitorability analysis using bond graph is automatically deduced using this unified tool, with respect to the detectability and isolability conditions. A real mechatronic system application of intelligent autonomous vehicle is given to show the efficiency and the simplicity analysis of the proposed approach. This paper was motivated by the problem of integrated design of a fault diagnosis system by considering both, system instrumentation and the set of specifications regarding faults. Existing methods dealing with such problems are based mainly on the existing system instrumentation. In this paper, a fault diagnosis system study and analysis is proposed. This is done by using a unified graphical tool such as Bond Graph tool which is used for system modeling, structural analysis and fault diagnosis conclusions. Therefore, system modeling, fault monitorability analysis, and fault indicator generation are all performed by using the same graphical tool. In addition, the proposed method may be exploited for monitorability analysis before industrial design, i.e., ability to detect and isolate faults with given instrumentation architecture and how to make faulty components monitorable by adding new sensors. To show the effectiveness of the proposed method, an application on real mechatronic system is considered.

Photovoltaic (PV) systems are integral to global renewable energy generation, yet their efficiency and reliability are frequently compromised by undetected faults, leading to significant energy losses, increased maintenance costs, and reduced operational lifespans. To address these challenges, this study proposes a novel hybrid deep learning framework that combines Stacked Sparse Auto-Encoders (SSAE) for autonomous feature extraction with an Optimized-Multi-Layer Perceptron (OMLP) for precise fault classification. The SSAE extracts high-dimensional fault features from raw operational data, while the OMLP leverages these features to classify faults with exceptional accuracy. The model was rigorously validated using real-world PV datasets, encompassing diverse fault types such as partial shading, open circuits, and module degradation under dynamic environmental conditions. Results demonstrate state-of-the-art performance, with the model achieving 99.82% accuracy, 99.7% precision, 99.4% sensitivity, and 100% specificity, outperforming traditional machine learning and deep learning approaches. These findings highlight the framework’s robustness and reliability in real-world applications. By significantly enhancing fault detection accuracy and computational efficiency, the proposed approach optimizes PV system performance, reduces operational costs, and supports sustainable energy production. This study concludes that the hybrid SSAE-Optimized MLP model represents a scalable and efficient solution for improving the reliability and longevity of renewable energy infrastructure, setting a new benchmark for intelligent maintenance strategies in the field.

This study deals with a hybrid method for the estimation of the remaining useful life (RUL) that does not require prior knowledge on the degradation phenomena, as the lack of information about the process of degradation is one of the main issues in the theory of fault prognosis of dynamic systems. The proposed method uses a dynamic model for the generation of fault indicators based on the principle of analytical redundancy, and the generation of a database by simulating the normal and faulty operations. Then, a data‐driven method is used to identify a cluster for each operating state. Taking into account uncertainties allows the generation of normal operation thresholds, which govern the launch of a kinematic model based on the calculation of the Euclidean distance for RUL estimation. The developed method is applied with three of the most used methods in literature, to a mechanical transmission system for the estimation of the remaining time before the breakage of the drive belt. The obtained results are discussed and evaluated with appropriate metrics in order to demonstrate the effectiveness of the proposed approach.