In order to mitigate biomechanical injuries sustained by middle-sized children during frontal collisions of school buses, this study investigated the optimal design of intelligent airbags. Initially, a validated school bus simulation model was constructed based on sled test data. Subsequently, methods for determining head, thorax, and femur injury values for a 10-year-old THUMS dummy were established, and the model, including the school bus restraint system, THUMS dummy and intelligent airbag, was developed. Next, based on NSGA-III, the adaptive propagation factors were proposed, the Gaussian mutation operator and the evolutionary mechanism of the particle swarm optimization method were introduced, and the non-dominated sorting level was improved, thus putting forward the improved NSGA-III. Finally, the design optimization of the intelligent airbag was conducted using the improved NSGA-III to determine the optimal configuration of the airbag. The results demonstrate that the improved NSGA-III outperforms the other three state-of-the-art optimization algorithms. Under normal, 10° forward tilt, 20° forward tilt, 5° right tilt, 10° right tilt and lying sitting postures, the head injury criterion (HIC₁₅), intracranial pressure (IP), liver pressure (LP), left and right femur forces (F_L and F_R, respectively), and weighted injury criterion (WICC) for middle-sized children are significantly reduced when the proportional coefficient of the gas mass-flow rate of the inflation valve, the opening pressure and opening coefficient of the deflation valves, and the installation height of the air bag are set at 1.23, 1.37×10⁵ Pa, 2.10, and 478 mm, respectively.

In order to effectively monitor the potential performance degradation of core components in aircraft air conditioning systems, a baseline modeling method for aircraft air conditioning systems based on modified MSET was proposed in this paper. Firstly, QAR data of key components was selected as the feature parameter. Secondly, using healthy historical data as training sample, the density peak clustering method was used to screen state data and construct memory matrices for both non-low-temperature and low-temperature datasets. Then, the adaptive diagonal loading technique was applied to MSET process to reduce the abnormal fluctuations caused by the pathological state of the memory matrix. Finally, a performance baseline was established between the multivariate feature variables and the system operating state, and real flight data from Airbus A320 aircrafts was used to analysis. Results show that the proposed method can simultaneously establish performance baselines for multiple key components such as primary heat exchanger, main heat exchanger and air cycle machine. It can effectively detect flight cycles with performance degradation before component failure, and the detection results are relatively accurate, providing a reference standard for airlines' condition based on maintenance and health management.

To prevent coal spontaneous combustion disasters in the gob area and precisely delineate hazardous zones, the variation patterns of CO and O₂ volume fractions along the gob length were statistically analyzed based on field measurements. Critical warning threshold ranges for spontaneous combustion were established and a gob hazard zoning method based on gas volume fraction kernel density was proposed. The results demonstrate that O₂ volume fraction exhibits a linear variation with distinct stages, and the variation rate before the inflection point is lower than that after the point. The CO volume fraction follows a quadratic polynomial variation. The kernel densities of both CO and O₂ volume fractions decrease with increasing gob length. Their highest kernel density distributions occur within 0-10m and 0-30m, respectively. O₂ volume fraction is more stable before the inflection point but exhibits greater fluctuation afterward. The average upper and lower acceptable limits for O₂ volume fraction before and after the inflection point are 3.3% and 6.3%, respectively. The critical CO and O₂ volume fraction values for hazard zoning based on kernel density are 0.009% and 15.1%, which divides the gob area into four quadrants and three risk levels (safe, potentially hazardous, and hazardous).

In order to solve the safety constraint problem of uncertain space robot (USR) with time constraints and output constraints, a second-order sliding mode global fixed-time emergency control method was adopted. The dynamic model of the uncertain space robot was established by applying the momentum conservation theorem and Lagrangian method. To achieve the safety output goal, an obstacle Lyapunov function was introduced to limit the system error within the preset range. Based on this, the upper bound of the system's uncertain terms was estimated, and the system's time constraints were further considered. By using the power summation integration technique, a second-order sliding mode global fixed-time emergency controller was constructed. The results show that root-mean-square error (RMSE) of the proposed method is 55.71% and 56.45% lower than that of the terminal sliding mode fixed-time controller without output constraints in Case 1 and Case 2 respectively, and it can ensure that the outputs of the USR system are strictly limited within the preset safety boundary, indicating the superiority of the tracking performance of the designed control method.

In order to explore the theoretical connotation and practical path of the systematic causation model of oil and gas pipeline accidents, 241 domestic and international oil and gas long-distance pipeline accidents were counted first, and the types of domestic and foreign accident causes and the contents of investigation reports were compared and analyzed. Secondly, we integrated the factor system of HFACS model and the hierarchical structure of STAMP model, and put forward an accident causation model that covers 4 system levels and 17 influencing factors. Finally, the N-K model and social network analysis method were used to carry out risk coupling calculation, network model construction, centrality analysis and core-edge analysis. The results show that equipment failure and pipeline corrosion account for 35.3% and 22.8% of the causes of oil and gas long-distance pipeline accidents. The risk coupling value is directly proportional to the number of system-level couplings. Inadequate safety training, delayed emergency response, lack of safety oversight, lack of risk awareness, lack of risk communication, behavioral violations and equipment process issues are the core influences on the accidents. The full life cycle management of accidents includes the reduction phase, readiness phase, response phase, and recovery phase.

In order to prevent and control the spontaneous combustion of residual coal in alkaline coal mines, the effect of coal particle size on coal oxidation self-ignition in an alkaline environment had been explored. Through water quality tests and Fourier-transform infrared spectroscopy (FTIR) tests, the microstructural changes of coal with different particle sizes soaked in alkaline solution were analyzed and judged. The oxidation kinetics of coals with varying particle sizes soaked in alkaline solutions had been investigated through programmed heating experiments and calculations of apparent activation energy. The results show that during the process of soaking coal particles of varying sizes in an alkaline solution, the redox potential (ORP) and total dissolved solids (TDS) exhibit significant changes as soaking time increases, the content of active functional groups in the coal increases, the coal samples with a particle size of 0-1 mm exhibit significantly higher levels of active functional groups compared to the other test groups. After soaking in an alkaline solution, coal samples with smaller particle sizes exhibit lower apparent activation energy. As particle size decreases, the amount of oxygen consumed during oxidation increases, enhancing the oxidation capacity. Consequently, the quantities of CH₄, and C₂H₆ gases released in the later stages of oxidation also progressively rise. In an alkaline environment, smaller coal particle sizes experience more severe erosion. The lower the energy barrier that must be overcome for oxidation, the more intense the coal-oxygen composite reaction becomes.

To address the reliance on a single factor selection approach in traditional dam deformation prediction and the difficulty of comprehensively capturing complex inter-factor relationships among high-dimensional influencing factors, a factor screening method based on the fusion of multiple algorithms was proposed and applied to construct an LSTM model using optimally factors. Specifically, influencing factors were selected separately using the correlation coefficient method, neighborhood component analysis (NCA), and least absolute shrinkage and selection operator (LASSO) technique. The results from these individual methods were subsequently integrated. Since correlations exist among the factors, highly correlated ones were further eliminated using symmetrical uncertainty (SU), thereby an optimized factors set was obtained. The selected factor set was then used to develop a dam deformation prediction model via an LSTM network. A concrete-faced rockfill dam in Xinjiang was used as the case study. The model performance was evaluated using root mean square error (RMSE), mean absolute error (MAE), mean square error (MSE), and the coefficient of determination (R²). The results demonstrate that, compared with traditional factor selection methods, the proposed multi-algorithms are integrated and factors with significant influence on dam deformation are comprehensively and accurately identified. MSE is reduced by 20.11%-59.09%, RMSE by 10.61%-36.05%, and MAE by 9.95%-37.86%, and a superior predictive model was obtained, compared with models using conventional factor selection methods. For specific monitoring points, the maximum reductions in MSE, RMSE, and MAE reach 53.5%, 31.9%, and 34.7%, respectively, while the highest R² value attains 0.986 0.

To qucikly obtain the source strength and location information from hazardous gas leakage incidents, a multi-factor optimized Gaussian plume dispersion model was proposed, and the leakage source parameters were inverted by combining it with a CAFA. Key environmental factors such as wind speed distribution, surface resistance, and surface reflection were incorporated into the Gaussian plume model to enhance its fitting capability under complex conditions through multi-factor calibration. Furthermore, a chaotic mapping was introduced to improve the population diversity and global search ability of firefly algorithm(FA), thereby achieving an effective balance between global optimization and local refinement while reducing the risk of falling into local optima. The results indicate that, after optimization based on wind speed distribution, surface resistance, and surface reflection, the error of the Gaussian plume model is reduced by 16%. CAFA effectively can avoid falling into local optima, reducing the source strength inversion error from 63.56% to 0.22%, and the leak source coordinate inversion error from 1.5 m to 0.2 m.

A structural reliability evaluation method for gas pipeline segments with multiple defects, based on Copula functions, was proposed to address the characteristics of the interrelationship between defects and the difficulty in quantifying the highly nonlinear coupling effects of multiple defects. First, based on the limit state equations of pipeline segments corresponding to different defects, Monte Carlo (MC) method was used to calculate the failure probability and probability distribution of pipeline segments under the influence of a single defect. Second, the probability distribution of each defect were treated as random variables, and the Copula functions were employed to study the impact of the correlation between different defects on the pipeline failure probability. The joint failure probability of the pipeline segment under the coupling effect of two defects was obtained. Third, the reliability of the pipeline segment with multiple defects was determined based on the second-order narrow-bound theory. Finally, using this method, the impact of the correlation between corrosion and crack defects on the structural reliability of gas pipeline segments was explored, and the variation of pipeline reliability was analyzed under different defect depths and wall thickness ratios (0.2-0.8). The results show that, under the condition that the failure probabilities of the defects are at the same order of magnitude and do not differ significantly, the correlation of the same type of defects is stronger, and the coupling effect has a greater impact on the pipeline failure probability. Compared with traditional series methods that assume independence between defects, which underestimate the pipeline reliability, the results of this method are more consistent with practical needs. This approach provides theoretical support for the stable operation of gas pipeline segments with multiple defects.

In order to address the challenges of unclear spatial concentration distribution and uncertain future evolution in high-pressure large-diameter gas pipeline leakage scenarios, a predictive model for gas leakage dispersion was proposed by integrating machine learning-based dimensionality reduction and time series forecasting methods. Firstly, a multi-condition dataset of gas leakage concentration fields was generated using computational fluid dynamics simulations. Subsequently, the dimensionality reduction module and time series forecasting module of the predictive model were separately optimized and trained using this dataset. Finally, the model's predictive accuracy was evaluated on an independent test set, and the prediction errors under various forecast horizons were analyzed. The results show that the model achieves a mean absolute error of 0.000 5 and a mean absolute percentage error (mAPE) of 6.82% on the test set, with the mAPE remaining below 14% across different prediction time steps.