Aiming at the problem that it is difficult to realize the traditional way of establishing mechanism model and then analyzing the safety of cryogenic loading system, this paper proposes to build a "system-subsystem-equipment" functional hierarchy model to simplify the difficulty of safety analysis. In order to solve the problems of scattered research results of system safety analysis and difficult knowledge sharing, this paper uses knowledge graph to integrate knowledge related to filling accidents, establishes ontology model of cryogenic loading system based on OWL, realizes the association among accident, evaluation index, monitoring data and other ontologies, and uses SWRL to establish the corresponding relationship among class, instance and attribute. The safety knowledge ontology and SWRL rules are put into the rule engine for reasoning and mining the hidden knowledge. Finally, an example of the actuator subsystem in cryogenic loading system is analyzed. The results show that the method can solve the problems of knowledge representation difficulty, lack of automatic semantic reasoning, poor reusability and sharing in cryogenic loading system.

Gastric cancer is one of the three major malignant tumors in the world. The earlier it is found, the higher cure rate get. However, the early clinical diagnosis is still not good enough. Thus, timely and accurate diagnosis of early gastric cancer lesions using computer-aided technology is urgently needed. This paper proposes an early gastric cancer detection (EGCD) model on the basis of CNN and attention mechanism. Due to the irregular shape of early gastric cancer under the endoscopy, the boundary is not obvious. First, improve YOLOv4, and propose E-YOLO with feature layer fusion. Combine CBAM to design EGCD to enhance the characteristic expression of cancer targets in channel and spatial, highlight cancer targets, and suppress background information. The results elaborate specificity of EGCD model is 77.39% and sensitivity is 82.27%. The average accuracy reaches 94.16%, which is 5% higher than YOLOv4. The method proposed has improved results in early gastric cancer detection and has high practical application value.

Sensors are of great importance in industrial facilities and automation systems. Its fault will bring deviation to the feedback loop and cause undesired results. The detection of sensor fault is a feasible scheme to promote system reliability and safety. The strong noise and disturbance in industrial situation will affect the fault detection procedure and results severely. In this paper, a fault detection scheme based on fine-tuned fractional-order chaotic system is developed for the situation that the system of interest is with strong noise and disturbance. The fault detection scheme is tested on two different kinds of sensor faults on different sensors of continuous stirred tank reactor with strong noise and disturbance. The simulation experiment results show that the fault detection scheme achieves relatively high performance.

With the improvement of technology, computer-aided technology is playing an increasingly important role in medical diagnosis. Early detection of esophageal cancer has a high probability of cure, so reducing the rate of missed diagnosis is very important for patients. Due to the particularity of medical images, the image data of early esophageal cancer is very limited. Therefore, we propose a new noise-filtered image augmentation method. After processing the original dataset using this method, a new image is generated to form the augmented the new dataset. The new dataset and the simple dataset are used in the early esophageal cancer recognition algorithm based on YOLOv3-Tiny for transfer learning to obtain training models. The experimental results show that the sensitivity of early esophageal cancer is improved after using the noise-filtered augmentation method. Therefore, the proposed noise-filtered image augmentation method can be applied in the recognition of early esophageal cancer.

A laboratory physical simulation and a CT scan test were conducted to analyze the anchoring mechanism of a system anchor bolt and a steel floral pipe in a layered rock mass. The following conclusions were drawn: (1) The anchoring effect of the system anchor bolt and steel floral pipe improves the strength parameter of the layered rock, and the system anchor bolt provides higher improvement. Neglecting the jointing effect, the improvement in the strength parameter of the layered rock due to the anchor bolt is primarily reflected by the following three aspects: the compressive zone effect of the preload, the repairing effect of the anchoring agent on defects in country rocks, and the reinforcement effect of the anchor bolt on the overall strength and density of the anchoring object. Considering the jointing effect, the improvement in the strength parameter of the layered rock due to the anchor bolt is reflected by improvement in the deformability and shear strength of the joint surface. (2) The stress-strain curves of the anchored specimens can be divided into different stages of damage evolution, based on the angles of bedding. (3) The crack-arresting effect of the anchor bolt is due to the weakening, cutting, and arresting of the cracks in the anchorage zone. The larger the anchorage zone, the better is the crack-arresting effect.

Uniaxial compression test was performed and CT scan images were obtained for the fractured specimens to study the mechanical property and fracture propagation of anchored rock with a hole under uniaxial compression. The results show that the supporting structure enhances greatly the strength of the specimens. However，different supporting structures have different effects of strength improving and the specimens with different beddings have the different effects of strength improving. Furthermore，it was found that the concrete and steel arch bridge can provide the supporting pressure for holes and prevent the damages such as stripping and spalling of holes effectively，which is necessary for maintaining the integrity of holes. Besides，from the analysis to CT scan images of the fracture surfaces of the destroyed specimens，it was found that the anchor effect of the anchor rod in the rock mass enables the rock mass to form an anchoring zone in a certain area. Thus，the anchoring zone can weaken and prevent the fracture development and change the path of fracture propagation in specimens.