Accurate prediction of remaining useful life (RUL) plays a significant role in ensuring the safe flight of aircraft. With the recent rapid development of deep learning, there has been a growing trend towards more precise RUL prediction. However, while many current deep learning methods are capable of extracting spatial features—those along the sensor dimension—through convolutional kernels or fully connected layers, their extraction capacity is often limited due to the small scale of kernels and the high uncertainty associated with linear weights. Graph neural networks (GNNs), emerging as effective approaches for processing graph-structured data, explicitly consider the relationships between sensors. This is akin to imposing a constraint on the training process, thereby allowing the learned results to better approximate real-world situations. In order to address the challenge of GNNs in extracting temporal features, we augment our proposed framework for RUL prediction with a Transformer encoder, resulting in the adaptive graph convolutional transformer encoder (AGCTE). A case study using the C-MAPSS dataset is conducted to validate the effectiveness of our proposed model.

There are hundreds of various sensors used for online Prognosis and Health Management (PHM) of LREs. Inspired by the fact that a limited number of key sensors are selected for inflight control purposes in LRE, it is practical to optimal placement of redundant sensors for improving the diagnosability and economics of PHM systems. To strike a balance between sensor cost, real-time performance and diagnosability of the fault diagnosis algorithm in LRE, this paper proposes a novel Optimal Sensor Placement (OSP) method. Firstly, a Kernel Extreme Learning Machine-based (KELM) two-stage diagnosis algorithm is developed based on a system-level failure simulation model of LRE. Secondly, hierarchical diagnosability metrics are constructed to formulate the OSP problem in this paper. Thirdly, a Hierarchy Ranking Evolutionary Algorithm-based (HREA) two-stage OSP method is developed, achieving further optimization of Pareto solutions by the improved hypervolume indicator. Finally, the proposed method is validated using failure simulation datasets and hot-fire test-run experiment datasets. Additionally, four classical binary multi-objective optimization algorithms are introduced for comparison. The testing results demonstrate that the HREA-based OSP method outperforms other classical methods in effectively balancing the sensor cost, real-time performance and diagnosability of the diagnosis algorithm. The proposed method in this paper implements system-level OSP for LRE fault diagnosis and exhibits the potential for application in the development of reusable LREs.

The improvement of fault diagnosis for complex equipment is an important step towards intelligent systems. Unlike component-level fault detection, system-level fault diagnosis presents new challenges, such as the integration of multiple sensors, the handling of massive datasets, and the necessity for robust system-level decision-making. To address these challenges, this paper proposes a novel Physics-embedded Recurrent Graph Neural Network (Pe-RGNN). By embedding physical knowledge about the interdependencies among sensors, the graph constructed of sensors are built through physical knowledge obtaining from high-fidelity model. With the sensory graph, the parameters are modeled through recurrent graph neural network to extract the temporal information. The proposed model has the advantage of simultaneously considering both spatial and temporal information. Experimental results from both simulated and real-world tests demonstrate the superior performance of the Pe-RGNN in early fault detection and fault identification accuracy when compared to traditional graph neural networks. These findings highlight the importance of incorporating both physical knowledge and time-series information in fault diagnosis models.

This paper presents a study of aero-engine exhaust gas electrostatic sensor array to estimate the spatial position, charge amount and velocity of charged particle. Firstly, this study establishes a mathematical model to analyze the inducing characteristics and obtain the spatial sensitivity distribution of sensor array. Then, Tikhonov regularization and compressed sensing are used to estimate the spatial position and charge amount of particle based on the obtained sensitivity distribution; cross-correlation algorithm is used to determine particle’s velocity. An oil calibration test rig is established to verify the proposed methods. Thirteen spatial positions are selected as the test points. The estimation errors of spatial positions and charge amounts are both within 5% when the particles are locating at central area. The errors are higher when the particles are closer to the wall and may exceed 10%. The estimation errors of velocities by using cross-correlation are all within 2%. An air-gun test rig is further established to simulate the high velocity condition and distinguish different kinds of particles such as metal particles and non-metal particles.