Search

Fault diagnosis methods based on deep learning have high requirements for the quality of the dataset, requiring a large amount of data for good model training to achieve accurate fault diagnosis. However, the fault signals that can be collected in practical applications are usually limited. A method for diagnosing roller faults based on small sample sound signals is proposed to address the problem of limited performance of intelligent fault diagnosis methods due to the difficulty in obtaining sound signals for roller faults and the small sample size. The feature transformation method is used to convert one-dimensional sound signals into two-dimensional time-frequency images, incorporating features from the frequency domain to improve the dataset's capability to express fault features. A dataset expansion method combining multiple types of time-frequency maps has been proposed. The method combines time-frequency maps drawn by three time-frequency analysis methods: short time fourier transform (STFT), continuous wavelet transform (CWT), and Hilbert Huang transform (HHT) to expand the dataset and increase data styles. The concept of deep transfer learning is introduced, using bearing datasets to pre-train the model, and then using roller data to fine-tune the pre-trained model to further improve the recognition accuracy of the model. The experimental results show that the dataset expansion method combining multiple types of time-frequency maps can effectively solve the problem of overfitting when training models with small sample data. After using transfer learning, the testing accuracy of the model reaches 98.81%, an improvement of 7% compared to not using transfer learning. There was no overfitting phenomenon, indicating that the model is well-trained. Compared to the method of generating adversarial networks to expand the STFT time-frequency map dataset and transfer learning, the method of dataset expansion by combining multiple types of time frequency maps and transfer learning has an accuracy improvement of 4%. It is easier to implement, and has stronger interpretability.

Dynamic coal quantity detection for conveyor belt is the foundation and key to achieving energy consumption optimization measures for multi-stage belt conveyors such as coal flow starting and automatic speed regulation. The existing coal quantity detection methods based on ultrasonic have low precision. Multiple ultrasonic sensors are susceptible to interference. In order to solve the above problems, a dynamic coal quantity detection system for conveyor belts based on ultrasonic array is designed. Using the principle of ultrasonic ranging, the coal material height corresponding to the detection points of each ultrasonic sensor array element is detected in real-time through an ultrasonic array. The cross-section slicing method is used to calculate the total volume of coal material passing through the conveyor belt per unit time. The real-time coal flow and total coal quantity of the conveyor belt are calculated based on the coal material stacking density. In order to reduce the crosstalk of the same frequency acoustic wave and the error caused by the attenuation of ultrasonic waves in harsh underground environments, 10 ultrasonic sensor arrays with different center frequencies are selected and arranged in a 2×5 linear array form. The collected coal height data is compensated through multiple rows of ultrasonic sensors to improve the accuracy of coal height data detection. The analysis results of real-time performance indicate that the ultrasonic array detection speed theoretically meets the coal quantity detection requirements of a belt conveyor with a belt speed of 5 m/s. The experimental results show that the average relative errors of regular material volume detection are 4.99% and 5.16% at belt speeds of 0.125 m/s and 0.170 m/s, respectively. Under simulated actual operating conditions, the average relative error of coal quantity detection is 5.56%. In the low belt speed state, the system has a measurement accuracy of over 94% for regular materials and coal. It basically achieves real-time and accurate detection of the dynamic coal quantity of the conveyor belt, meeting the coal quantity detection requirements of the belt conveyor.

Nuclear safety is a key component of the national security system, and it is the foundation and lifeline of the nuclear industry. Advanced and reliable nuclear safety technology is crucial for maintaining and improving intrinsic safety. Therefore, conducting strategic research on nuclear safety technologies is important for enhancing the nuclear industry in China. In this article, we conduct an in-depth research on China’s nuclear safety technology system using methods including academician interviews, field surveys, conference discussion, and literature review. The results show that, guiding by the overall national security and the nuclear safety concepts, China’s nuclear safety technology has made significant progress in recent years and its nuclear safety performance is good. However, China’s nuclear safety technology system still face several bottleneck problems. For example, the nuclear safety standards system needs improvement, the overall planning of nuclear safety software research and development is insufficient, and the precision and advanced nuclear safety equipment still depends on foreign countries. To continuously modernize the nuclear safety governance system and governance capacities and strengthen China’s nuclear industry, several suggestions are proposed. First, the nuclear safety standards system should be further improved. Second, independent nuclear safety software with high quality should be promoted by coordinating scientific research resources to tackle key problems. Third, government, industry, university, research, and application need to be coordinated to research and develop high-end nuclear safety equipment.