In order to find out the susceptibility of debris flow after fire at different time points, the burned land where a serious fire occurred in Lushan Mountain, Changshou Township in March 2020 was selected as a demonstration research area. Based on the idea of “space for time”, the whole study was carried out. Through laboratory experiments, the root soil mechanical parameters of the study area at different time after fire were obtained. By using the experimental parameters obtained, the slope instability coefficient of different years after fire was obtained through the slope instability model. According to the slope stability division standard, the slope instability area of different years after fire was obtained. Finally, the source strength indexes of different years after fire were extracted. The dynamic evaluation index system of post-fire debris flow susceptibility at small watershed scale was established by taking source strength as static evaluation index and topographic and geomorphic index as static evaluation index. Using the entropy weight method to calculate the weight of index factors combined with the comprehensive index method, the dynamic susceptibility assessment of debris flow was carried out on the burned land of Lushan Mountain in Changshou Township. Based on the results, targeted remediation of watersheds that remain highly susceptible for many years after a fire and those that are highly susceptible within a short period of time can effectively prevent and reduce the probability of mud slides, while also saving economic costs and achieving truly effective disaster prevention and mitigation.

An open world object detection method based on shape perception and class balance optimization was proposed to address the issue of poor prediction performance of unknown class objects in open world object detection. Unknown classes referred to classes that were not labeled during the training phase. Due to the lack of guidance from labels, detecting unknown class objects was a challenging task. An unknown class enhanced detector has been constructed as an unknown class detection branch. During training, this detector was supervised using only known class labels, allowing it to learn the similarities in features of known class objects and generalize to unknown class objects. To improve the detector's sensitivity to unknown classes, the region proposal network (RPN) module's ability to distinguish between foreground and background was utilized. A specific filtering method was employed to select results with “unknown class potential” from the RPN output, which were then used as pseudo labels in the training process. Due to the absence of confidence scores, traditional non-maximum suppression (NMS) methods were difficult to apply for post-processing unknown objects. Therefore, a redundant unknown object suppression mechanism was designed, consisting of a center point-based grouping strategy and a redundancy score matrix based on shape perception. The center point-based grouping strategy included three methods based on the unknown class center points to determine the suppression range. Subsequently, a redundancy score matrix was constructed based on the redundancy scores of each prediction box within the group to suppress highly redundant predictions. Experimental results on open world object detection datasets demonstrated that the open world object detection based on shape perception and class balance optimization maintained high recall rates for unknown classes while achieving high prediction accuracy. This method effectively addressed the challenges of open world scenarios and avoided generating a large number of useless predictions.

Defect detection is regarded as an indispensable step in the industrial production process. At present, manual detection is faced with the problems of low efficiency and high cost. A ceramic small target defect detection algorithm based on deep learning was proposed. For small target defects, a slice pre-training layer was first added to reduce the loss of graphics memory resources by large-size images. Secondly, a small target detection layer was added for the detection of small target defects, and a large target detection layer was removed to reduce the number of parameters. In addition, a feature selection fusion module based on MLCA (mixed local channel attention) was proposed to improve the perception of small target defects. Finally, a detection head with shared parameters was designed to further reduce the number of learnable parameters of the algorithm. By comparing with the baseline model, taking the ceramic cup as an example, the detection accuracy of this algorithm has been improved by 20.9%. Combined with the developed detection software and experimental platform, the detection efficiency of the ceramic cup has been enhanced by about 46.9%.

To address the issue that traditional fault diagnosis methods struggle to accurately diagnose faults in the nuclear reactor coolant system (RCS) of nuclear power plants under uncertain conditions, a dynamic fuzzy radial basis function neural network (DFRBFNN) model was established for RCS fault diagnosis following these steps. First, based on the fault types and sample data of the RCS, the initial structure of the DFRBFNN model was determined. Then, using the radial basis function neural network method, the initial DFRBFNN model for RCS fault diagnosis was constructed, and a random initialization method was applied to initialize the connection weights from the defuzzification layer to the output layer of the initial DFRBFNN model. Finally, the error reduction rate method was used to adjust the structure and parameters of the initial DFRBFNN model, resulting in the final DFRBFNN model for RCS fault diagnosis. The established model was applied to diagnose loss of coolant, flow loss, and steam generator tube rupture accidents, and its performance was compared with traditional fault diagnosis models to verify its effectiveness. The research shows that the constructed DFRBFNN model can accurately diagnose RCS faults under uncertain conditions.

Conventional diagnostic methods that require a large amount of data support in practical engineering are difficult to effectively perform centrifugal pump fault diagnosis under small sample conditions. Therefore, the residual network (ResNet) in deep learning was combined with dilated convolution and extended into a siamese network to construct a dilated residual siamese network (DRSN). The dilated residual network was used as the feature extraction module of the siamese network, which enhanced the feature extraction ability of the model. Positive and negative sample pairs were constructed to extract more information from each sample, and make more effective use of limited data.The two sub-networks share parameters, the number of free parameters and lowering the risk of overfitting was reduced when the sample was insufficient. The proposed network model alleviated the problem of insufficient training samples, improved the efficiency of data utilization, and realized the fault classification of centrifugal pump under the condition of small samples. The research results show that even in the most sample-scarce situation, the accuracy of the model on the centrifugal pump test dataset can still reach 82.20%, which is at least 8.8 percentage points higher than other models.

Forest fires have the characteristics of strong suddenness, great destructiveness, many uncertain factors and high risk of fighting. In order to study effective strategies of firefighting, firstly, based on the theory of cellular automata, the forest fire system was analyzed, and a forest fire model considering external factors such as wind and flame retardant was established. Then, on this basis, the fire-fighting agent was modeled and correlated with the forest fire model, so as to build the fire-fighting model. Finally, the simulation algorithm based on cellular automata was designed to simulate the effect of fire-fighting strategies under the influence of different environmental factors, and the effect of different fire force allocation strategies. The results show that the method can combine fire-fighting simulation with actual decision making, and provide visualized and quantified strategy scheme for relevant departments to make fire-fighting decision, which is helpful to reduce forest fire loss and rescue cost.

To effectively address the contradiction between the uncertainty of the internal and external environment in the construction industry and the complexity and vulnerability of the construction supply chain, as well as to promote the overall security and stability of the construction supply chain network, and to prevent and mitigate the risk of disruption among node enterprises, a construction supply chain network invulnerability analysis method was proposed based on the complex network theory and cascade failure model. Firstly, from the perspectives of business, resources and information flow, the TOPSIS(technique for order preference by similarity to an ideal solution ) method was used to assess the importance of node enterprises based on multiple complex network centrality indicators. Secondly, combined with the operational characteristics of the construction supply chain, an improved load-capacity-elasticity cascade failure model was established to measure the impact of enterprise disruption from the perspective of network loss under intentional attack, and to analyze and explore the network invulnerability improvement strategy from the perspectives of node capacity, load, and resilience. After numerical simulation and analysis, the results show these as follows. When the upstream node enterprises of the construction supply chain network suffer from the impact of disruption risk, the supply chain network can show strong network invulnerability, but it should focus on the downstream supplier enterprises, so as to avoid the network as a whole suffering from more losses due to the shortage of the supply of construction materials or basic services. To keep the small difference in the business capacity among node enterprises, the large difference in the business load and an appropriate high level of risk remediation cost investment can effectively reduce the loss of the supply chain network when the node enterprises are interrupted, thus improving the level of network invulnerability. Among the multiple types of strategies, the node capacity strategy is better than the node resilience strategy and the node load strategy in order to improve the network invulnerability. The results of the study can provide scientific references for improving the security level of construction supply chain and proposing disruption risk management strategies.

Because Hainan Province is located in the tropics, it is often hit by extreme weather such as typhoons and rains, so it is very prone to geological disasters such as slope collapse and landslides, which eventually cause irreparable losses. In order to improve the stability of tropical soil slope in a green and environmentally friendly way, the slope was strengthened by microbial induced calcium carbonate precipitation (MICP) technology and carpet grass root slope consolidation. Suitable microbial strains were first screened out, and the preparation process of related microbial agents was optimized. Subsequently, carpet grass root system was implanted in the slope soil indoors, and MICP treatment was carried out after the formation of the root-soil complex. Subsequently, a series of laboratory tests and numerical simulation analyses were carried out to evaluate the reinforcement effect of this technology. The results showed that MICP technology and plant root treatment complemented each other in terms of mechanical brittleness and integrity of slope, and significantly improved the unconfined compressive strength and shear strength performance of soil, effectively enhanced the stability of slope, and reduced the risk of slope erosion. Finally, the numerical simulation verification was carried out by using Abaqus finite element software, which enhanced the reliability of the research results. It can be seen that the MICP combined with carpet grass root reinforcement method provides an effective reinforcement method for tropical soil slopes, which not only improves the mechanical properties of slopes, but also promotes ecological restoration and environmental sustainability. This result provides a new technical approach for the ecological reinforcement of slopes in tropical areas, and has certain theoretical and practical significance for protecting the ecological environment and reducing the risk of geological disasters.

The explosive popularity of the new generation of artificial intelligence technologies will profoundly impact the risk experience of perceptual subjects within risk societies. Factor analysis and multiple indicators and multiple causes (MIMIC) model were used to study 12 risk scenarios of generative AI, besides four indicators reflecting the public's risk perception and five dimensions affecting the public's risk perception were explored. The results show that the public's perception of the risks of generative AI can be reflected by expectations of safety, technology, user and corporate regulatory. The public's risk perception is affected by its subjective evaluation of technology risks, macro risks, equity risks, subject risks and application risks, among which both equity risks and macro risks have the most significant impact. It shows that the public's risk perception of generative artificial intelligence is mainly characterized by “self-oriented” and “precautionary”. On this basis, the public's risk perception of generative artificial intelligence from the perspectives of history and culture, risk communication and technology governance was analyzed further, and corresponding countermeasures was put forward.

To address the issue of land and capital waste caused by suboptimal site selection and construction models for urban drone landing and takeoff sites, the maximum coverage model is initially used for site selection. However, due to the uneven distribution of demand points and overly simplistic coverage determination criteria, the results show low coverage rates and overly concentrated site selection. To solve this problem, a method based on spatially continuous demand for the maximum coverage model of drone landing and takeoff site selection was proposed, considering factors such as no-fly zones and application scenarios. Demand objects were determined using a regular grid, and candidate sites were identified using the PIPS(polygon intersection point set) method. The feasibility of the improved model was validated through a case study of site selection for urban drone landing and takeoff sites in Binhai New Area, Tianjin. When the number of landing and takeoff sites was fixed at 14, the improved model increased the actual service area coverage rate from 62.03% to 88.61%. The results indicate that this method better meets the practical requirements for drone landing and takeoff site selection, resulting in more evenly distributed and rational site layouts, and significantly enhancing the service coverage rate of the drone landing and takeoff sites.