In response to the issues of parameter homogenization and insufficient rationality in the support design of the mining roadway in the Lingtai mining area, it is highly significant to conduct classification research on the mining roadway and propose differentiated support strategies. Firstly, based on a systematic analysis of the geological conditions of the surrounding rock and the matching relationship with the support system in the mining roadway of the Lingtai mining area, six classification indicators were determined. Secondly, a mining roadway classification method based on mixed data clustering was proposed, extracting the informational characteristics of numerical and categorical indicators and introducing a penalty competition mechanism to dynamically optimize the number of clusters during the classification process. Subsequently, the proposed method was applied to classify 37 segments of the Lingtai mining area's mining roadways into four categories, validating the feasibility of the method. Finally, a field test was conducted in the D segment of the 2502 transport roadway in Shaozhai Coal Mine. The results indicate that the maximum convergence of the roof and floor of the test roadway segment is 109 mm, and the maximum convergence of the two sides is 113 mm, with a small range of surrounding rock damage. The dynamic classification and differentiated support method for mining roadways can achieve stable control of the surrounding rock in the roadways.

The accurate detection of coke overflow in high-dust environments is a pivotal challenge in achieving intelligent coke loading. A method was proposed to address this issue for the intelligent detection of coke loading overflow, which was based on dark channel prior knowledge and the ResNet network. Firstly, a video collector was used to obtain video information of the coke loading scene, and the original time-series video image frames were processed to obtain the region of interest between the discharge port and the loader. Secondly, the prior knowledge method of dark channels was employed to process the regions of interest. Enhancing the contrast between the target areas and irrelevant areas within the regions of interest, thereby mitigating the effects of dust on subsequent detection models. Moreover, the problem of overflow detection was transformed into a binary classification task by labeling the regions of interest based on the actual loading of coke. Finally, the ResNet network was utilized for modeling, enabling the completion of model training and experimentation during the loading process of newly acquired coke. The experimental results demonstrate that the proposed method exhibits promising performance on new data, achieving an overall accuracy of 86.81%. Specifically, the accuracy, recall, and F1 score for the overflow class are 84. 12%, 90.74%, and 0.8730, respectively. Furthermore, the application of the dark channel prior algorithm in data processing results in a notable increase in the recall rate of the overflow class by 3.31%.

With the deepening research on federated learning, it has been observed that the privacy protection strategies employed within federated learning fall short of fully guaranteeing the security and confidentiality of user data. Moreover, the training process in federated learning encounters challenges regarding model convergence. In response to these aforementioned issues, an innovative solution termed adaptive differential privacy (DP-AdaMod) was proposed. Primarily, the model training process was fine-tuned by incorporating an adaptive learning rate algorithm to mitigate model fluctuations and the adverse effects of overfitting. Consequently, this enhancement led to improved training efficiency and optimal performance. Secondly, the application of differential privacy techniques ensured the privacy security in federated learning through the deliberate introduction of noise into the model gradients. Additionally, accurate quantification of privacy loss was achieved by implementing the moment accountant mechanism, facilitating a balanced trade-off between privacy preservation and analytical accuracy. This meticulous approach served to fortify system security. Lastly, the efficacy of the proposed solution was ascertained through comprehensive simulation experiments. The results substantiate the superior performance of the proposed method, evident by its exceptional accuracy, efficient utilization of privacy budget, and other notable facets.

Design cycle is one of the key factors that influence the development of customized mechanical products (CMP). Secondary development technology (SDT) is an effective method to enhance the efficiency of CMP. To better promote high-quality development of SDT in CMP applications, first, key technological advancements in SDT were briefly introduced based on the analyzes of the evolution of customized product production and the parameterized design methods. Then, research and application status of SDT in CMP design was analyzed from different perspectives, and existing problems are summarized. Finally, the future trends of SDT were explored in conjunction with modern information technologies, such as artificial intelligence (AI), collaborative design (CD), cloud technology (CT), and digital twins (DT), and the importance of domestic development platforms was emphasized. The results provides a reference for the application and further development of SDT in CMP.

To investigate the influence of Allee effect on population dynamics, the center theorem and bifurcation theory were used to study the bifurcation of a discrete predator-prey system with Allee effect on prey. The results indicate that an appropriate increase in the growth rate of the prey population will lead to the emergence of predators, the natural enemies of the prey by the transcritical bifurcation; Under the appropriate Allee effect, the system is stable, but if the Allee effect exceeds the critical value the system will undergo a double period flip bifurcation. The bifurcation parameter diagram also shows that as the Allee effect further strengthens, this flip bifurcation will lead to chaos in the system. From a biological perspective, in a certain ecological environment, a species that reproduces to a certain extent will encounter its natural enemies through “natural selection”. An appropriate Allee effect is beneficial for the stability of the predator-prey system, but if the Allee effect of the prey is too strong, it will cause a sharp decrease in predators and an effective restriction of no natural enemies. The prey will undergo a brief surge, exceed the environmental capacity, and then die out. To reduce the adverse effects of the bifurcations on the system, the state feedback is used to control the transcritical bifurcation and facilitate hybrid control to control the flip bifurcation. The numerical simulation results show complete consistency with the theoretical analyses.

Machine learning methods have been employed in the study area of Changyang Tujia Autonomous County for landslide hazard assessment, it could provide a scientific basis for geological disaster prevention and control efforts. Through the correlation analysis of 12 evaluation indicators (planar curvature, terrain undulation, surface roughness, slope, vegetation coverage, engineering lithology, distance to fault zone, distance to water system, rainfall, land use type, distance to buildings, and distance to roads) in the study area selected by historical landslide points, they were selected. And the evaluation model of the study area was constructed by calculating the information content of factors and integrate support vector machine (SVM) and gradient boosting decision tree (GBDT) models. The hazard of the study area was classified into four levels: extreme high, high, medium, and low, to generate hazard zoning. Subsequently, an assessment of the evaluation model was conducted. The results indicated that the very high hazard zone was mainly distributed in the southwest, central, and eastern parts of the research area. The distribution percentages of very high, high, medium, and low hazard zones predicted by the I-SVM and I-GBDT models were 15.86%, 21.29%, 33.51%, 28.68%, and 30.08%, 7.41%, 13.28%, 49.22%, respectively. The prediction of hazard zones by the I-SVM model aligned more closely with reality. The AUC values for the I-SVM and I-GBDT models were 0.859 and 0.829, respectively. The prediction of risk zones by the I-SVM model is deemed more reasonable and reliable.

According to the drawing load characteristics of horizontal rectangular anchor plate in red clay foundation, the vertical drawing model test of horizontal rectangular anchor plate in saturated red clay foundation is carried out by using a self-made visual drawing model test system combined with digital photographic measurement technology. The results show that the sliding surface of soil around the anchor shows different shapes with the change of buried depth ratio, but the initial Angle does not change with the change of buried depth ratio. The load displacement curve generally has obvious peak characteristics, but with the increase of the buried depth ratio, the characteristics gradually weaken. When the buried depth ratio is the same, the smaller the length-width ratio is, the more obvious the three-dimensional bearing characteristics of the anchor plate and the peak characteristics of the curve are. The bearing capacity coefficient increases with the increase of buried depth ratio, but the law is different under different aspect ratio. The bearing capacity coefficient decreases with the increase of length-width ratio, and the change law is consistent under different buried depth ratio. Conclusion: For test red clay, the anchor plate can be classified as shallow buried type at least within the range of buried depth ratio 4, and deep buried type if the buried depth ratio is greater than 8. Under the same conditions, under the influence of cohesion and dilatancy, the tensile strength coefficient of anchor plate in red clay foundation is between loose sand foundation and dense sand foundation.

With the continuous development of container cloud technology, it is of great significance to predict and analyze the overall trend and peak of cloud resource requests for efficient utilization and reasonable allocation of container cloud resources. Deep learning technology for load prediction has become a key technology to solve the unbalanced utilization of container cloud resources. Aiming at the problems of low prediction accuracy and insufficient capture sequence features existing in the current single model and combination model of load prediction, a cloud resource combination prediction model based on temporal convolutional network-long short-term memory(TCN-LSTM)was proposed. The hollow convolution in the combination model increased the sensitivity field without reducing the feature size to obtain longer time series features. The residual network could transfer information across layers to accelerate the convergence of the network, and the obtained time series features could effectively improve the prediction accuracy of LSTM. Useing Alibaba’s publicly available dataset to make predictions, the experiment shows that the proposed model is compared with the single prediction model and other combined models, and the error index-mean absolute error(MAE) is reduced by 8%~13.7% and root mean squared error (RMSE) by 9.8%~13.1%, which proves the effectiveness of the proposed model.

China’s offshore wind power sector is rapidly progressing with large-scale and efficient projects. To support large-capacity wind turbines, there is an increasing need for advanced tower structures that offer superior structural performance and economic efficiency. One such innovative solution is the hybrid fiber-reinforced composite materials (FRP)-concrete-steel prestressed double-skin wind turbine towers (PDSWTs) proposed by scholars at The Hong Kong Polytechnic University. PDSWTs boast excellent durability, high load-bearing capacity, and exceptional stability, resulting in the potential to reduce production and maintenance costs throughout the service life. A study on the analysis and design procedures of PDSWTs was presented using a tower that supports 12 MW offshore wind turbine as an example, based on the provisions in current design standards and finite element modelling. The results demonstrate that: the tower successfully meets the frequency requirements of wind turbine; the tower possesses high ultimate resistance under compression-bending, shear and torsional loadings of ultimate limit states; the strain, stress, crack width and tower deformation under serviceability limit states meet the requirements of current design standards; furthermore, a simplified method is proposed to check the fatigue resistance of steel and concrete in the tower sections. The result is expected to provide references for the design and application of PDSWTs.

In order to effectively utilize LNG(liquefied natural gas) cold energy and liquefy CO₂ in gas turbine exhaust gas, a new process of LNG cold energy to liquefy CO₂ and CO₂ power cycle was proposed. In this process, Reheat cycle and regenerative cycle were added on the basis of conventional Rankine cycle, and multi-flow strand heat exchanger was set up. Chemical process simulation software was used to simulate the process flow and sensitivity analysis of reflux temperature, interstage cooling temperature and maximum circulating temperature and pressure was carried out to obtain the best operating parameters. Exergy efficiency, specific work and CO₂ liquefaction rate of the system were analyzed and calculated by exergy analysis. Exergy efficiency of exergy was 54.16%, specific work was 335.9 kJ/kg LNG, and CO₂ liquefaction rate was 0.621 7 kg/kg LNG in a new process. The evaluation indexes of the new process were better than those of the existing process. As for the exergic efficiency, exergic efficiency of exergic was as the highest and the temperature of CO₂ after liquefied was as the constraint condition, exergic efficiency of exergic was 54.28% and specific power was 337.5 kJ /kg LNG, so the system performance was further improved.