Effective bearing fault diagnosis can ensure the safe operation of rotating machinery, which is important for the stable operation of rotating machinery. Feature fusion of multi-sensor data is a feasible method to improve fault diagnosis performance. To accurately detect, localize and identify bearing faults, we propose an attention-based two-stage multi-sensor feature fusion (ATS-MSFF) method for bearing fault diagnosis. The first stage, focuses on feature extraction from sensor itself, which utilizes the Channel-Attention to enhance key features in the sensor's signal. The second stage then focuses on feature fusion between sensors, such that the output of each sensor is endowed with additional critical information provided by other sensors. Experiments conducted on a publicly available PU dataset validate the effectiveness of our approach, achieving a high classification accuracy of 99.58% and maintaining stable performance in the presence of noise interference. In addition, the design of this framework takes into account the flexibility of practical applications and can adapt to different numbers of sensor configurations, providing a new solution for the accurate diagnosis of bearing faults.

Controlling three-phase current source rectifiers presents intricate challenges due to the input LC filters and DC bus capacitors. This paper proposes a multiple-loop-based control strategy for current source PWM rectifiers for low-voltage DC distribution systems to address these challenges. Firstly, the independent control of the active and reactive power is achieved based on the adjustment of the d-axis and q-axis currents on the grid side. Subsequently, a compensation scheme of capacitance current using the phasor method is proposed. Next, as an alternative damping method, the capacitor-voltage feedback is used to suppress the LC resonance, and the influence of the feedback gains on the system's stability is discussed. Furthermore, the dynamic performance is further improved by adopting DC-side inductor-current feedback and a damping factor. Ultimately, the feasibility and efficacy of the proposed approach are confirmed through experiments.

This article develops a new inductive displacement sensor with a segmented multi-group coil structure, which is suitable for the displacement measurement of control rods in nuclear reactors. Each group coil of the sensor consists of two excitation coils and one sensing coil. The excitation and sensing coils are segmented to extend the linearity range of the displacement sensor. It abandons the traditional sensor’s method of using nonlinear compensation to achieve large-stroke displacement measurement. Providing an alternating current (AC) signal to the excitation coil and processing the induced voltage generated by each sensing coil can directly achieve the high-precision measurement of core displacement. The mathematical model of the variations in the sensing coil voltage caused by the movement of the core is established. The impacts of the excitation coil structure, the number of turns of the excitation coil, and the excitation frequency on the output characteristics of the designed sensor are analyzed by finite element simulation. Based on the analysis and design, a sensor prototype is built and tested in the laboratory. The measurement results show that the linearity error is 0.35% and the maximum measuring error can be limited within 1.5 mm, which is sufficient to meet the practical requirements in a nuclear reactor environment.

With the advancement of artificial intelligence technology, fault diagnosis methods based on deep learning have been extensively studied due to their ability to automatically extract fault latent features and develop end-to-end diagnostic models. However, the existing methods focus on achieving high accuracy while neglecting model complexity. Therefore, this paper proposes an inter-turn short circuit fault diagnosis method of permanent magnet synchronous motor using data-level fusion and multi-objective tree-structured Parzen estimator (MOTPE) optimized Res-Net. In this method, the original three-phase current signals are fused into a new modal signal through Clarke transform at the data layer. Based on an improved Res-Net18, hyperparameters are optimized using MOTPE to achieve high-performance and lightweight model design. Experiments have validated the fault diagnosis model that integrates current signal fusion and MOTPE optimization, achieving an accuracy of 99.62%, with the best noise robustness and the lowest model complexity. Compared with single-objective tree-structured Parzen estimator, multi-objective Non-Dominated Sorting Genetic Algorithm-III, and Random algorithms, MOTPE not only maintains high accuracy but also achieves lower computational costs and a lightweight network structure.

Inductive position sensors have been widely employed in nuclear reactors for monitoring the control rod position. Accurate control rod position information plays a crucial role in regulating the power of nuclear reactors. However, the inductive sensors currently used for control rod position detection exhibit limited measurement accuracy, which may adversely affect safety and reliability. In response to these issues, a segmented coil structure inductive position sensor based on the resonance principle is proposed in this paper. To improve the sensor sensitivity, a resonant circuit is implemented by connecting capacitors in series with the excitation and sensing coils, and a theoretical model is developed. Furthermore, optimizing the resonant capacitance can reduce the nonlinearity error of the sensor. To enhance the load-driving capability of the excitation source, a single-phase inverter circuit is proposed as the excitation source. Only one excitation signal is used to excite multiple coils. ANSYS simulations and experimental measurements are performed to compare sensors with and without the resonant circuit. The results validate the theoretical analysis and demonstrate that the sensor employing the resonant circuit significantly improves sensitivity and enhances linearity range. The sensor achieves a sensitivity of 0.1909 V/mm and a nonlinearity error of less than 2% over the full-scale span.