Shanghai Electric (China)

Abstract Machine learning has demonstrated significant application potential in rolling bearing fault detection. However, traditional deep learning approaches typically depend on large-scale labeled datasets and are limited to identifying a single fault type, which poses challenges in meeting the complex fault diagnosis in real-world industrial scenarios, particularly under small sample conditions. To address these issues, this study proposes a multi-task attention fault diagnosis framework, convolutional attention neural network by neural discriminant dimension reduction (NDDR-CANN). The model employs multiple task-specific neural networks to extract distinct features such as fault type and severity, respectively, and achieves adaptive feature sharing across tasks by NDDR, thereby facilitating the transfer and complementarity of task-related knowledge and enhancing the model's generalization capability and predictive accuracy. Furthermore, by integrating a convolutional block attention module, the model can focus on salient features across channel and spatial dimensions, thereby improving feature discrimination. Experimental validation on the CWRU bearing dataset demonstrates that the proposed model exhibits robust effectiveness and adaptability under various fault conditions and small sample scenarios. Comparative experimental results further confirm that NDDR-CANN outperforms current mainstream few-shot fault diagnosis methodologies.

According to the duality of inductive wireless power transfer (IPT) and capacitive wireless power transfer (CPT) in coupler and tuning method, an integrated coil is designed in this paper to constructs a compact hybrid wireless power transfer (HPT). The magnetic field coupling and the electric field coupling constructed by coils and plates share one energy transmission channel, which enhances the transmission power and energy density. The inductance of coils and capacitances of plates compensate for each other. The resonant frequency of integrated coils is adjusted to kHz. Additional tuning inductances and capacitances are eliminated to build a lighter self-resonance coupler. The stacked structure design of coils and plates enables compact integration and miniaturization of integrated coils. A 300W experimental platform was built, and some comparative experiments were conducted. Compared to the coupler of IPT, the self-resonance coupler has improved performance in transmission power, anti-offset, and over-coupling suppression. HPT performs better than CPT in transmission distance, transmission power, and transmission efficiency. It has some advantages in the medium-power application scenarios such as inspection robots and drones, which expect couplers to have a smaller volume, lighter weight, higher power density and better anti-offset.

In a hybrid clock synchronization network, the convergence speed is low. Therefore, an optimization algorithm for a hybrid clock synchronization network under time and frequency measurement is proposed. Based on time and frequency measurement, an accurate formula for the transmission delay of a hybrid clock synchronization network is derived.The network is constructed into a star topology, and the optimal node is selected as the master clock source. At the same time, a formula for calculating the clock deviation of the slave node and a formula for calculating the mean and variance of the local time deviation adjusted by the reference node when multiple reference nodes work together are given to achieve precise joint optimization of multiple regions. Experimental results show that the synchronization stability of the Mesh topology optimized by the designed algorithm is 99.45%-100%, and the synchronization time is 5.80ms when there are 13 nodes. At the same time, the convergence time of the designed algorithm is 1.9s, and the synchronization standard deviation is 0.02μs, proving that it has a high convergence speed.

Wind-solar complementary dispatching distributes output tasks according to a fixed ratio during the dispatching process through historical data statistics, resulting in low dispatching efficiency. Therefore, a complementary dispatching method based on a multi-objective optimization model in wind-solar group collaborative control is studied. The objectives of maximizing economic benefits and minimizing power fluctuations are quantitatively integrated, and the constraints of related objective functions are designed. The hybrid particle swarm algorithm is used to dynamically solve the model and select the dispatching scheme that meets the conditions. For the schemes that meet the conditions, a feedback control error correction mechanism is designed to output the optimal dispatching scheme. The experimental results show that this method has the smallest total output and load deviation, a high 24-hour absorption rate and low power fluctuation, and has significant advantages in improving the economic efficiency of new energy absorption and grid stability.

A novel corrugated steel-reinforced concrete pipe that enhances electromagnetic shielding performance compared to the conventional reinforced concrete power pipes is developed and presented in this paper. In order to investigate the pipe’s behaviour under jacking and service conditions, the critical axial compression and flexural moment distributions were represented by two separate flat segments of a circular pipe cross-section, respectively. A total of six column specimens were designed for axial compression testing, while another four beam specimens were prepared for four-point bending tests to examine the bending behaviour. Prior to testing, all specimens were subjected to standard curing, and the material properties of steel and concrete were determined via standard tests. The load versus deformation curves of column specimens, the moment versus deflection curves of beam specimens, and the corresponding failure modes were obtained from the tested specimens. It was revealed that the load-carrying capacities of the corrugated steel-reinforced concrete cross-sections were comparable to those of the conventional reinforced concrete counterparts. Advanced finite element (FE) models incorporating the mechanical properties of encased corrugated steel plates (CSPs) and the damage development of concrete were developed and were validated against the experimental failure modes and load-carrying capacities. Based on both experimental and numerical results, the load-carrying capacity of corrugated steel-reinforced concrete cross-sections was evaluated by referring to Chinese standard GB/T 11836 and American standard ASTM C76. The experimental and numerical finding can pave the way for further research and applications of this novel type of corrugated steel-reinforced concrete pipe.