Search

Due to the lack of other component information in traditional magnetic leakage signal defects and the low accuracy of prediction methods, this paper proposes an improved residual network for magnetic leakage detection defect recognition method that predicts defect size and different detection speeds. A new defect diagnosis method based on ResNet18 on the Convolutional Neural Network (CNN) is proposed in this study. This method transfers the pre-trained ResNet18 network and replaces the activation function in the transferred network structure. It extracts features from transformed two-dimensional images obtained by converting the original experimental signals and signals with added noise, removing the influence of manual features. The results demonstrated that the improved ResNet18 network model, after transfer learning, achieved 100% prediction accuracy for all 10,000 grayscale images generated with defect lengths of 50 mm; width of 2 mm; and depths of 2 mm, 4 mm, 6 mm, and 8 mm. Moreover, the prediction accuracies for the quasi-static, slow, compensated fast, and fast scanning speeds were 99.20%, 98.50%, 93.30%, and 94.00%, respectively, for defect depths of 2 mm, 4 mm, 6 mm, and 8 mm. These accuracies surpass those of other models, demonstrating the significant improvement in prediction accuracy achieved by this method. [ABSTRACT FROM AUTHOR]

The study of tar behaviors in underground coal gasification (UCG) is essential for pollution control, system safety and conversion efficiency; however, existing studies have only focused on tar in products without revealing tar evolution in the reaction zone, and the experimental conditions in reported work are far from those in the real situation. In this work, tar behaviors were studied with a self-developed apparatus to simulate the UCG process. During the experiments, the sampling method along the gasification channel was used to collect tar at different positions; the gasification object was a large raw coal block 460 mm × 230 mm × 230 mm in size, and the flow rate of the inlet gas was adjusted according to the composition of products. The tar samples were not only taken from the outlet, but also from the reaction zone, and then analyzed using gas chromatography mass spectrometry. For all the tar samples, C15H13N and its isomer were the most abundant compounds, with a total percentage greater than 14%. Most of the top five chemicals contained more than nine carbon atoms in their molecular formulae, indicating that more heavy tar than light tar is formed by low-temperature pyrolysis. Compared with the upstream tar, the downstream tar had fewer PAHs and a lower boiling point, due to the decomposition of the heavy tar. The downstream tar contained more of the element fluorine (F) than upstream and outlet tars, indicating that tar pollution remaining in the reaction zone cannot be evaluated by monitoring the outlet tar. [ABSTRACT FROM AUTHOR]