China Safety Science Journal

Identification and classification of Chinese and English safety related journals

LI Jie, LIU Jiao, QIAN Jinxin, TONG Ruipeng, LIANG Xushu

2022, 32(5): 1-7. doi:10.16265/j.cnki.issn1003-3033.2022.05.1995

Asbtract ( 586 )

HTML ( 21)

PDF (4515KB) ( 96 )

Figures and Tables | References | Related Articles | Metrics

In order to identify and classify safety related journals both in Chinese and English, based on Web of Science data platform, publications of China Safety Science Journal and Safety Science between 2018 and 2020 were searched from science citation index (SCI) and Chinese science citation database (CSCD). Then, VOSviewer software and journal co-citation analysis method were utilized to identify and analyze the lists and clusters of safety journals from references of selected publications. The results show that for these journals selected from the co-citation network of China Safety Science Journal, there are 10 journal clusters identified, including "safety management science", "safety science and human factor", "fire, combustion, engineering thermophysics and industrial process", "reliability" and etc. And citation sources of Safety Science are mainly from 7 journal clusters, namely "reliability engineering, expert system, ocean engineering and operational research", "safety management psychology and occupational health", "safety ergonomics", "safety science and building safety", "fire safety, evacuation and traffic", "traffic safety and occupational medicine" and "process safety, risk, disaster and engineering mechanics".

Impact of bus drivers' organizational identity on driving anger:mediating role of burnout

LI Qiong, WANG Yongjie, CHEN Wenqiang, BAI Pengxia, XUE Chenlei

2022, 32(5): 8-13. doi:10.16265/j.cnki.issn1003-3033.2022.05.0467

Asbtract ( 397 )

HTML ( 3)

PDF (1573KB) ( 55 )

Figures and Tables | References | Related Articles | Metrics

In order to explore antecedent variables that affect bus drivers' driving anger and their influencing mechanism, a mediation model among organizational identity, job burnout and driving anger was put forward with job burnout as an intermediary variable. Then, 583 bus drivers in Xi'an were surveyed by using organizational identity scale, burnout scale and driving anger scale. Finally, the model was verified by utilizing Process macro program. The results show that violations and egoism of other traffic subjects are more likely to induce driving anger of bus drivers. Organizational identity can significantly predict anger and burnout in a negative way, with the latter playing a completely mediating role between organizational identity and driving anger. In another words, the identity affects driving anger completely through burnout.

Crowd intervention strategies considering indoor respiratory exposure risks

LIANG Benbu, XIE Kefan, VAN DER WAL Natalie, LIU Jia

2022, 32(5): 14-20. doi:10.16265/j.cnki.issn1003-3033.2022.05.0842

Asbtract ( 287 )

HTML ( 2)

PDF (4982KB) ( 89 )

Figures and Tables | References | Related Articles | Metrics

In order to explore impacts of crowd intervention strategies on indoor respiratory exposure risks during major pandemics, a variety of crowd motion scenarios were established in general indoor conditions based on improved pedestrian dynamics model and respiratory infection probability model. Then, multi-agent simulation technology was utilized to simulate impacts of strategies, including protection optimization, pedestrian flow optimization and route optimization, on the exposure risks. The results show that indoor respiratory exposure risks are mainly determined by total pedestrian flow, individuals' stay length, movement route planning and duration of stay in contaminated areas. The carryover effect will be formed due to pedestrians' obedience behavior of social distancing, which will further increase exposure time to contaminated areas. The lower pathogen permeability of masks, and the greater space ventilation are, the lower infection probability the crowd will face.

Meta-analysis on impacts of role stress on employees' safety behavior:moderating effect of supervisor support

CHEN Qiao, LIU Suxia, MEI Qiang, ZHANG Jingjing

2022, 32(5): 21-27. doi:10.16265/j.cnki.issn1003-3033.2022.05.0079

Asbtract ( 293 )

HTML ( 3)

PDF (1403KB) ( 65 )

Figures and Tables | References | Related Articles | Metrics

In order to clarify impact mechanism of role stress on employees' safety behavior, meta-analysis was used to analyze relationship between role stress (role conflict, role ambiguity and role overload), supervisor support and employees' safety behavior (safety task behavior and safety contextual behavior) by statistically integrating 37 independent samples from 35 empirical studies. Then, moderating effects of economic development levels and industry types were investigated. The results show that role conflict, role ambiguity and role overload negatively affect employees' safety behavior, among which role ambiguity has the greatest impact on it, and also a greater impact on safety contextual behavior than on safety task behavior. However, supervisor support positively affects such behavior, and plays a moderating role between role conflict, role ambiguity, role overload and safety task behavior, as well as between role ambiguity and safety contextual behavior respectively. Moreover, employees' safety behavior is more vulnerable to role stress in developing economies and high-risk industries than in mature economies and general industries.

Relationship between error management climate and workers' safety performance in construction enterprises

CHEN Yanlin, HUANG Wenwei, LI Hongyang

2022, 32(5): 28-34. doi:10.16265/j.cnki.issn1003-3033.2022.05.1754

Asbtract ( 314 )

HTML ( 4)

PDF (1853KB) ( 187 )

Figures and Tables | References | Related Articles | Metrics

In order to clarify the internal influencing mechanism of error management climate on workers' safety performance in construction enterprises, workers' safety perception and self-efficacy were introduced as mediators and a SEM of relationship between the four was established on the basis of literature research and open interviews. 222 front-line construction workers from Guangzhou and Foshan were investigated by utilizing error management climate scale, safety perception scale, safety self-efficacy scale and safety performance scale. Finally, the hypothesis model was verified through model test and mediator effect test by using SPSS and AMOS. The results show that the direct relationship between error management climate and safety performance is not significant, but safety perception and safety self-efficacy indicate significant mediating effect independently as well as plays a significant chain intermediary role between the former two. Building a good error management climate and improving workers' safety perception and self-efficacy can effectively improve their safety performance.

Diagnosis of train wheelset tread damage based on EPSA-ResNet

HE Jing, HOU Na, ZHANG Changfan, HU Xinliang, LIU Jianhua

2022, 32(5): 35-40. doi:10.16265/j.cnki.issn1003-3033.2022.05.2166

Asbtract ( 219 )

HTML ( 2)

PDF (2718KB) ( 52 )

Figures and Tables | References | Related Articles | Metrics

In order to address problems of low accuracy, slow speed and few damaged samples in diagnosis of train wheelset tread damage, a diagnosis model based on EPSA-ResNet was proposed. Firstly, model parameters were obtained by pre-training ImageNet data set through transfer learning, and transferred to wheelset tread damage characteristics for fine-tuning, so that shared model structure and parameters were obtained. Secondly, 3×3 convolution in ResNet-50 residual blocks was replaced by a Pyramid Split Attention (PSA) Block before a new EPSA-ResNet was gained, which integrated multi-level features of spatial and channel attention, and feature re-calibration was carried out adaptively. Finally, damage diagnosis results were obtained by Softmax classifier. The results show that this method can identify damage state of train wheelset tread and local subtle differences between surrounding environment, with an diagnosis accuracy up to 99.79%, which is better than other deep neural network models.

Critical mechanical pressure rise for preventing buoyancy-dominated flow pattern in inclined tunnel fire with longitudinal ventilation

YANG Dong, LI Ping

2022, 32(5): 41-47. doi:10.16265/j.cnki.issn1003-3033.2022.05.0938

Asbtract ( 242 )

HTML ( 3)

PDF (3552KB) ( 48 )

Figures and Tables | References | Related Articles | Metrics

In order to prevent uncertainly of smoke flow in inclined tunnel fire with longitudinal ventilation, firstly, a theoretical model of critical fan-induced pressure corresponded to only existence of buoyancy-dominated flow pattern P_j_,min and only existence of fan-dominated flow pattern P_j_,max was proposed through potential function analysis. Then, small-scale experiments were carried out to validate the model's scientificity and effectiveness, and the magnitude of P_j_,c, which met critical velocity, were compared with the above P_j_,min and P_j_,max. The results show that P_j_,max is always larger than P_j_,c. If P_j_,c is adopted in a longitudinally ventilated inclined tunnel, there are two possible flow patterns caused by different fan activation time, one is the pattern that critical velocity is achieved and all of the smoke is exhausted downward, while the other is that buoyancy overwhelms fan-induced pressure rise, and thus part of smoke spreads upstream. If P_j_,max is adopted, effective longitudinal ventilation can be finally achieved under any fan activation time.

Suppression characteristics and mechanism of different hydrogen phosphates on ethylene explosion

WANG Yan, LI Zhong, ZHANG Yimin, WANG Yang, YANG Jingjing, JI Wentao

2022, 32(5): 48-54. doi:10.16265/j.cnki.issn1003-3033.2022.05.1564

Asbtract ( 282 )

HTML ( 3)

PDF (2060KB) ( 57 )

Figures and Tables | References | Related Articles | Metrics

In order to effectively reduce harms caused by ethylene explosion, three different hydrogen phosphates, namely ammonium dihydrogen phosphate, potassium dihydrogen phosphate, and calcium dihydrogen phosphate, were selected, and maximum explosion pressure as well as its rise rate and peak value reach time of ethylene under different concentration of hydrogen phosphates were tested on a 20 L spherical explosion test system. Then, inhibitory characteristics and mechanism of hydrogen phosphates on ethylene explosion were analyzed based on the change features of above mentioned parameters along with concentration variance. The results show that all three kinds of hydrogen phosphates have a certain inhibitory effect on ethylene explosion, and their optimal critical concentration for suppression is all 0.4 g/L, which means the suppression effect will be weakened either higher or lower than the value. When at same mass concentration, potassium dihydrogen phosphate has the best suppression effect, followed by calcium dihydrogen phosphate, and then ammonium dihydrogen phosphate.

Design and simulation of high-rise escape descent device based on orifice throttling principle

LU Yi, YANG Yuxuan, SHAO Shuzhen, HE Weilin, WANG Wei, YAN Xiaowen

2022, 32(5): 55-60. doi:10.16265/j.cnki.issn1003-3033.2022.05.1184

Asbtract ( 357 )

HTML ( 4)

PDF (1707KB) ( 49 )

Figures and Tables | References | Related Articles | Metrics

In order to address shortcomings of existing descent devices, such as large volume, complex structure and uncontrollable speed, a new high-rise escape device with controllable speed was designed. Firstly, relationship between its descent speed and orifice diameter was deduced. Then, a dynamic model of the device was established by using AMESim software, and its descent speed was simulated under different loads and different throttling apertures. Finally, safe adjustment range of orifice diameter was determined through simulation analysis, and purpose of variable speed and slow descent was achieved by adjusting orifice diameter of throttle holes connected in series between two hydraulic cylinders. The results show that the new descent device features simple structure and small volume, allowing users to freely adjust its descent speed within throttling aperture of 1-6 mm according to escape demand, with a maximum speed of no more than 1.5 m/s. For objects with different weights, the speed change range always tends to be stable, which greatly improves applicability and safety of the descent device.

Hesitant fuzzy method with variable weight for HAZOP risk sorting

WANG Tiedan, MU Yiqiang, PENG Dinghong

2022, 32(5): 61-67. doi:10.16265/j.cnki.issn1003-3033.2022.05.1335

Asbtract ( 306 )

HTML ( 1)

PDF (1767KB) ( 87 )

Figures and Tables | References | Related Articles | Metrics

In order to improve quality and efficiency of HAZOP, a hazard risk sorting method was proposed based on hesitant fuzzy information and incorporating continuous punishment-reward variable weight function. Firstly, HFS was used to describe group evaluation information of expert group for full convergence of HAZOP group's intelligence. Secondly, a continuous punishment-reward variable weight method was proposed to reflect influence of indicator state values on risk sorting. Then, HAZOP analysis process based on variable-weight hesitant fuzzy TOPSIS-Sort method was further constructed. Finally, the proposed method was used in a crude stabilization unit for risk level analysis, which verified its scientificity and effectiveness. The results show that five hazards, including H₁, are sorted into Negligible level, another five hazards, including H₄, are sorted into As Low As Reasonably Practicable level and the corresponding ranking is H₁₀<H₈<H₅<H₄<H₇.

Analysis on failure characteristics of coal seam floors and water inrush risks in mining under pressure

LI Gang, LIU Haizhen, YANG Qinghe, NIU Weifeng

2022, 32(5): 68-76. doi:10.16265/j.cnki.issn1003-3033.2022.05.1535

Asbtract ( 196 )

HTML ( 0)

PDF (11561KB) ( 27 )

Figures and Tables | References | Related Articles | Metrics

In order to study characteristics of floor failure under mining conditions, with working face 10409 of Xinzhi Coal Mine as a research object, distribution characteristics of floor stress field and damage evolution of floor strata were studied in detail by using a combination of theoretical analysis and numerical simulation methods. Then, based on verification results of on-site pressurized water test, distribution law of bearing pressure in coal seam floor strata and damage range of floor strata during mining process were obtained. Finally, mining risks of No. 10 coal seam under pressure was analyzed and evaluated. The results show that under influence of mining, the floor of coal seam presents periodic failure of tension failure-compaction-tension failure, and the critical value of working face length that affects failure depth of floors is about 160 m. The maximum failure depth of floor strata can reach 9.50 m, while maximum water inrush risk point lag behind working face is about 2.6 m.

Study on wake safety intervals of PA based on CFD

HE Xin, MA Yilong

2022, 32(5): 77-83. doi:10.16265/j.cnki.issn1003-3033.2022.05.2075

Asbtract ( 186 )

HTML ( 1)

PDF (3153KB) ( 39 )

Figures and Tables | References | Related Articles | Metrics

In order to address the lack of accuracy in wake prediction by simple mathematical modeling method under changeable environment, firstly, basic requirements of wake safety zones in PA mode for CSPRs were studied. Secondly, wingtip vortex under influence of crosswind was numerically simulated by adopting k-epsilon model in a vortex field created at the tail of transport wing, and spatial distribution of wake vortices before pairing was obtained. Then, a calculation method of PA wake safety intervals was put forward based on CFD after analyzing PA wake safety zones. Finally, with Shanghai Hongqiao Airport as an example, numerical simulation was conducted for a case where lead aircraft and trail aircraft were B747-400 and B737-800 respectively under crosswind of 3 m/s to verify the proposed method in terms of scientificity and effectiveness. The results show that crosswind can intensify transverse displacement of wakes, and trail aircraft can follow up flight in upwind direction of lead aircraft so as to avoid influence of wakes. When flight environment is constant, lead aircraft's wingspan is the decisive factor for maximum wake safety interval of PA. It is also found that the calculation results of safety intervals based on CFD can better fit experimental measured values, thus improving calculation accuracy. Moreover, CFD-based results are far less than current wake interval standard under the condition of same paired types.

Study on failure mode and stability of reserved rock in front of anti-slide piles

ZHU Dapeng, HE Lei, TIAN Li, QIN Liangkai

2022, 32(5): 84-89. doi:10.16265/j.cnki.issn1003-3033.2022.05.0935

Asbtract ( 135 )

HTML ( 0)

PDF (5347KB) ( 30 )

Figures and Tables | References | Related Articles | Metrics

In order to study stability of rock mass reserved in front of anti-slide piles under large excavation condition, deformation and failure modes of reserved rock mass at different widths were tested and analyzed based on similarity theory. Secondly, formulas for stability calculation of rock mass under different failure modes were deduced according to limit equilibrium theory. Finally, with landslide in Hunan province as research object, stability variance features of the mass under each failure mode was analyzed. The results show that along with increase of width, deformation trend of reserved rock is always in the form of deformation of the excavation surface - a few cracks - local uplift deformation in front of piles. The rock mass with larger width is prone to up-slide failure, while that with smaller width is prone to down-slide one. As width increases, stability coefficient increases in form of a quadratic function with the opening upward under down-slide failure.

Double parameters fire detection method based on machine learning

LIU Quanyi, ZHU Bo, DENG Li, SHI Hang, LIANG Guanghua

2022, 32(5): 90-96. doi:10.16265/j.cnki.issn1003-3033.2022.05.0874

Asbtract ( 206 )

HTML ( 4)

PDF (2796KB) ( 116 )

Figures and Tables | References | Related Articles | Metrics

In order to address false alarms or alarm failure caused by a single technology in fire detection, a combustion experimental platform was designed and built. PM10 and CO, concentration of which increased rapidly in the flue gas after fire, were selected as characteristic parameters.Then, data processing was conducted for them, and a fire detection model was established by adopting six machine learning algorithms, including logistic regression(LR), linear discriminant analysis (LDA), kNN algorithm, classification and regression tree(CART), naive Bayes, and support vector machine (SVM). Finally, the model's classification performance was assessed. The results show that among six algorithms, kNN features higher evaluation accuracy, recall rate, F₁ value and kappa value, with its accuracy of assessment reaching as high as 95.2%, making it possible to accurately identify combustion state. This method can accurately detect fire by classifying rapidly changing concentrations of PM10 and CO in combustion products.

Analysis on thermal characteristics of PRHRS of sodium-cooled reactors

WANG Yandong, CHEN Yongdong, DENG Jing

2022, 32(5): 97-103. doi:10.16265/j.cnki.issn1003-3033.2022.05.1048

Asbtract ( 213 )

HTML ( 2)

PDF (2041KB) ( 69 )

Figures and Tables | References | Related Articles | Metrics

In order to study thermal behavior characteristics of natural circulation of the system and related equipment's safety, a PRHRS thermal model of sodium-cooled reactor was established based on coupling of energy balance and flow balance. Then, influence mechanism of passive characteristics on system was explored, including potential difference and density difference. Finally, the model was verified with experimental results of PFBR, and influence laws of the number of tube passes, fin thickness, chimney diameter, and air door opening degree on thermodynamic cycle were clarified. The results show that along with increase of tube passes, heat transfer coefficient of air heat exchanger (AHX) decreases by 13.6% while pressure drop increases by 66.4%, and the system's heat transfer performance is better when the number of tube passes smaller. When inner diameter of exhaust chimney or fin thickness increases, the coefficient and pressure drop increase while heat transfer coefficient per unit pressure drop decreases, and there is an optimal solution in the design. Reducing area of inlet air door can effectively increase operating temperature of equipment and heat transfer effect of system in extreme low temperature conditions, which ensures safe operation of the system.

Study on catastrophe instability criterion of surrounding rock in tunnels blasting crossing water-rich fault zones

PENG Yaxiong, LIU Guangjin, HUANG Zhigang, LYU Hubo, WU Li

2022, 32(5): 104-111. doi:10.16265/j.cnki.issn1003-3033.2022.05.0887

Asbtract ( 191 )

HTML ( 0)

PDF (3827KB) ( 61 )

Figures and Tables | References | Related Articles | Metrics

In order to solve instability problems of tunnels blasting through rich water fault zones, based on instability mechanical model of tunnel-fault system under blasting, a cusp catastrophic model was established considering the combined effects of blasting and groundwater. Instability mechanism of surrounding rock under tunnel-fault system was revealed, and its instability criterion was developed. Then, with resource allocation project of Pingtan and Minjiangkou in Fujian as an example, rock stability of tunnels crossing water-rich fault zones was evaluated. The results show that the instability of surrounding rock is a result of internal and external factors. Due to long-term water-rock interaction and cumulative blasting, mechanical properties of surrounding rock in water-rich fault zones are constantly weakened, and catastrophe criterion, k value, is constantly reduced, which probably leads to instability and failure of the rock. The results of instability risk evaluation based on the criterion are consistent with field conditions.

Track circuit fault diagnosis method for massive imbalanced data

XING Yulong, WANG Jian, SHANGGUAN Wei, PENG Cong, ZHU Linfu

2022, 32(5): 112-118. doi:10.16265/j.cnki.issn1003-3033.2022.05.2389

Asbtract ( 209 )

HTML ( 2)

PDF (4619KB) ( 64 )

Figures and Tables | References | Related Articles | Metrics

In order to address deviation of decision-making boundary of track circuit diagnosis model due to imbalanced monitoring data and slow training speed caused by massive data, a fault diagnosis method based on data resampling and ensemble learning algorithm was proposed. Firstly, imbalanced data were processed by feature synthesis and resampling including random down-sampling and Synthetic Minority Oversampling Technique (SMOTE). Secondly, a fault diagnosis module for massive monitoring data was constructed based on LightGBM algorithm which could be trained efficiently, training and diagnosis flow was designed, and key parameters were selected by grid search and cross-validation. Finally, Macro-F₁, which was not affected by imbalanced data, was introduced as an evaluation indicator of the model. The results show that the comprehensive performance of each diagnosis model for imbalanced data can be improved by feature synthesis and data resampling. Compared with other algorithms, LightGBM is the best in terms of comprehensive performance and training time, ensuring superiority and rapidity when faced with massive data.

Test and simulation of influence of voids on tunnel structure safety

ZHENG Aichen, XIANG Jinyang, HUANG Feng, JIN Chenghao, LIU Xingchen

2022, 32(5): 119-126. doi:10.16265/j.cnki.issn1003-3033.2022.05.1777

Asbtract ( 238 )

HTML ( 2)

PDF (14094KB) ( 57 )

Figures and Tables | References | Related Articles | Metrics

In order to reduce damage of voids behind lining to tunnel structure, influence of voids on safety of secondary lining structure was studied through changing parameters like number, location and size of voids by using model test and numerical simulation. The results show that safety factor of lining structure is significantly reduced by voids, especially when they exist at the same time in vault, bottom and left and right waist of the arch. Changes in safety factor of arch foot is the most obvious, which decreases by 57%. For cases of two voids at the same time, that at left and right arch waist is most unfavourable to stress of lining structure, featuring the lowest safety factor. When vault void increases from 0° to 45°, the vault bending moment changes from positive to negative, and its safety factor firstly increases and then decreases, during which, it will be the smallest when circumferential angle is 45°, only 34% of that at 15°.

Preparation and flame retardant properties of silica gel foam

GONG Yawen, LIU Yingchun, JING Mingju, ZHANG Yingnan, ZHANG Bin

2022, 32(5): 127-133. doi:10.16265/j.cnki.issn1003-3033.2022.05.0114

Asbtract ( 325 )

HTML ( 1)

PDF (2907KB) ( 80 )

Figures and Tables | References | Related Articles | Metrics

In order to solve problems of low viscosity and poor stability of silica gel foam materials, a silica gel foam was prepared by using composite foaming agent, foam stabilizer, sodium silicate solution and acetic acid as raw materials. Firstly, sol-gel method and orthogonal test were used to study properties of sodium silicate solution's crosslinking system and compound foaming agent. Secondly, gelling time and viscosity of gels at different proportions were tested by Brookfield viscometer, and gel ratio suitable for solid combustibles (wood) was selected. Subsequently, on-site experiments were conducted by utilizing synchronous thermogravimetric (TG) analyzer to study flame retardant properties of gel foam on wood. The results show that silica gel foam has good flame retardant effect on wood combustion. The gel could be bonded to surface of solid combustibles (wood) during gelation process. In air atmosphere, carbon residue rate of wood is almost zero, while that for wood @gel foam system after experiment is about 48.63%, and according to the field wood burning tests, the wood's weight loss rate is 46% higher than gel-treated wood, indicating that the gel foam has better flame retardancy.

Determination method of comprehensive early-warning indexes for tailings dam based on improved POT model

ZENG Jihan, ZHANG Guang, WU Hao, HU Shaohua

2022, 32(5): 134-139. doi:10.16265/j.cnki.issn1003-3033.2022.05.1699

Asbtract ( 195 )

HTML ( 3)

PDF (2532KB) ( 62 )

Figures and Tables | References | Related Articles | Metrics

In order to improve online monitoring and early-warning system of tailings dam, a comprehensive early-warning model was proposed based on qualitative indexes and quantitative indexes. Firstly, warning indexes were divided into monitoring indexes and non-monitoring ones whose weights were evaluated respectively, and comprehensive risk values were obtained. Then, an improved POT model was introduced, grand Pareto distuibution(GPD) function of risk indicators was established and solved, and comprehensive early-warning values were determined according to 3σ criterion. Finally, case analysis was conducted with a tailings dam in Hubei Province as an example. The results show that the proposed comprehensive early-warning model can make up for limitations of traditional single-indicator models, and warning values of yellow level, orange level and red level are calculated to be 66.538, 71.387 and 80.585 respectively. Compared with the results of cloud model(CM), it is proved that the proposed model has rationality and effectiveness.

Detection method of high-altitude safety protective equipment for construction workers based on deep learning

ZHANG Meng, HAN Yu, LIU Zefeng

2022, 32(5): 140-146. doi:10.16265/j.cnki.issn1003-3033.2022.05.1141

Asbtract ( 435 )

HTML ( 16)

PDF (5600KB) ( 241 )

Figures and Tables | References | Related Articles | Metrics

In order to address the lack of dynamics and coordination in current detection technology of safety protection equipment, a new detection method for construction workers working at height was proposed based on deep learning. Comprehensive detection of helmet and seat belt in state of dynamic videos was realized by replacing backbone feature extraction network of YOLOv4 with lightweight network MobileNetV2. Then, tests were carried out to verify effectiveness of the method. The results show that detection speed of the proposed method is increased by 2.7 times in central processing unit (CPU) operating environment, and single-frame video detection speed for single target, multi-target and small target can be maintained between 25-27 milliseconds in graphics processing unit (GPU) operating environment. At the same time, mean average precision rate of 91.57%, 89.69% and 86.63% can be achieved.

Risk assessment method of risers based on two-dimensional cloud model and BN

LIU Jiahao, YU Yang, ZHANG Zhenxing, YU Jianxing, WANG Weiwei, FU Yiqin

2022, 32(5): 147-154. doi:10.16265/j.cnki.issn1003-3033.2022.05.0699

Asbtract ( 218 )

HTML ( 1)

PDF (2308KB) ( 56 )

Figures and Tables | References | Related Articles | Metrics

In order to explore key failure modes of non-bonded flexible risers, a risk assessment method combining quantitative assessment and probability reasoning was proposed. Firstly, risk matrix method was improved based on two-dimensional cloud model. Secondly, fault tree was transformed into BN, bottom event was quantitatively evaluated and its occurrence probability was calculated. Then, failure probability of top event was solved and fault source was investigated through BN bidirectional reasoning. Finally, preventive measures were proposed for key failure modes, and the assessment method was verified with an example. The results show that the failure probability of non-bonded flexible risers is 1.099×10^-2. Corrosion, erosion and ship movement are the three failure modes with higher risk levels, and should be paid attention to during maintenance.

Influence of access roads on accident risks of mountainous two-lane highway passing through villages and towns segments

JI Xiaofeng, GENG Zhaoshi, PU Yongming, LU Mengyuan, QIN Wenwen

2022, 32(5): 155-162. doi:10.16265/j.cnki.issn1003-3033.2022.05.0444

Asbtract ( 244 )

HTML ( 0)

PDF (1743KB) ( 63 )

Figures and Tables | References | Related Articles | Metrics

In order to reduce traffic accident risks of mountainous two-lane highways passing through villages and towns segments, SEM and N-K model were constructed based on road alignment, village and town environment as well as traffic accident data. Then, coupling effects of access roads and environment, vehicle and road alignment on traffic accident scale of such mountainous highways were explored, and influence of road-village distance, access road type and quantity, and intersection type on the accident risk and their coupling effects were quantitatively analyzed. The results show that access roads have significant impacts on the accident risk of the above-mentioned highways, and it has a strong coupling effect with horizontal curve factors and vertical curve ones. Road-village distance and intersection type have direct significant effect, while number and type of access roads have only indirect significant effect. The coupling risk of roads detouring villages for more than 100 meters and mechanical tillage road is maximum, with risk coupling degree at 0.359 6, while that of machine tillage road and three access roads is minimal, with coupling degree at 0.000 1.

Traffic accident severity prediction for secondary highways based on cluster analysis and SVM model

YANG Wenchen, ZHOU Yanning, TIAN Bijiang, GUO Fengxiang, HU Chengyu

2022, 32(5): 163-169. doi:10.16265/j.cnki.issn1003-3033.2022.05.1263

Asbtract ( 286 )

HTML ( 1)

PDF (3160KB) ( 106 )

Figures and Tables | References | Related Articles | Metrics

In order to identify influence of input features on machine-learning-based traffic accident severity (TAS) prediction model, 12 potential factors were firstly selected as input variables based on 1808 accidents of a secondary mountainous highway. Then, KM algorithm was used to discretize continuous feature variables of TAS, and RF algorithm was adopted to select key feature variables. Finally, by combining three kinds of input feature variables (potential features, KM features, and RF features) and SVM algorithm, three kinds of TAS prediction models were developed respectively (SVM^*, KM-SVM, and RF-SVM), and their prediction performance was systematically analyzed in terms of accuracy and applicability. The results show that severity prediction accuracy of RF-SVM model is significantly improved by discretizing continuous variables and identifying key feature parameters, with the accuracy for severe injuries or deaths being improved about 40%. Influence of feature selection on SVM model performance is less than that of discretization of continuous variables. And RF-SVM model, in spite of a better prediction performance than binary logistic regression model, has higher sensitivity to different input features.

Study on risk model of highway traffic accidents based on LSTM-BF

XIONG Xiaoxia, LIU Qingchao, SHEN Yujie, CAI Yingfeng, CHEN Long

2022, 32(5): 170-176. doi:10.16265/j.cnki.issn1003-3033.2022.05.1602

Asbtract ( 299 )

HTML ( 1)

PDF (3539KB) ( 106 )

Figures and Tables | References | Related Articles | Metrics

In order to reduce highway traffic accidents, a traffic risk state prediction model was studied by using LSTM neural network and BF. Firstly, time dependency existing in historical traffic flow risk data was studied through LSTM module. Then, real-time risk prediction performance was improved by BF module integrating LSTM prediction results. Finally, with accident data and traffic flow data of Ningbo roundabout in 2020 as an example, multi-step feature variables were constructed in the form of migration time window based on spatio-temporal data between upper and lower port stations within 20 min prior to accidents, and 5-fold cross-validation was carried out. The results show that the precision and recall rate of LSTM model are higher than that of random forest (RF) algorithm, and F₁ score of the final prediction result is close to 0.80 by adding BF module.

Scientific knowledge map analysis of coal mine noise research progress in China

YANG Ming, JIA Gaini, LIU Maomao, QUAN Xingyuan

2022, 32(5): 177-184. doi:10.16265/j.cnki.issn1003-3033.2022.05.1484

Asbtract ( 177 )

HTML ( 2)

PDF (6795KB) ( 296 )

Figures and Tables | References | Related Articles | Metrics

In order to understand the research progress of coal mine noise in China, based on CNKI, the relevant literature was statistically analyzed, and the scientific knowledge map of coal mine noise in China was drawn by visualization software. The results show that the annual number of research papers in the field of coal mine noise was more than 100 in the recent 10 years, Chongqing Research Institute Co., Ltd. of China Coal Science and Industry Group, Xi'an Research Institute Co., Ltd. of China Coal Science and Industry Group, Occupational Safety and Health Research Center of the State Administration of Work Safety, China University of Mining and Technology and other institutions have carried out a large number of research work on coal mine noise ; with the passage of time, the research content in the field of coal mine noise is dynamically changing and gradually widening, and coal mine noise and noise control have always been the research hotspot. There are many studies on the noise characteristics in long spaces such as subways and highway tunnels, while there are few studies on the noise characteristics of coal mine roadways, a special long space.

Research on change characteristics of environment in micro space of mine ventilation suits

YOU Bo, MAO Cong, SHI Shiliang, LIU Heqing, LU Yi, LI Min

2022, 32(5): 185-193. doi:10.16265/j.cnki.issn1003-3033.2022.05.0819

Asbtract ( 147 )

HTML ( 2)

PDF (10267KB) ( 26 )

Figures and Tables | References | Related Articles | Metrics

In order to reveal the heat transfer characteristics of mine ventilation suits in microspace, explore the local cooling effect of ventilation clothing on the human body, the heat source of ventilation clothing microspace was studied, and the local cooling mechanism of the ventilation clothing was analyzed through comparing numerical simulations with real person training test. The influences of ambient temperature, air intake pipe ventilation rate and intake pipe structure on human skin temperature and microspace climate environment were analyzed. The research results show that the working environment temperature has a significant impact on the human skin temperature and the microspace environment temperature, while the air intake duct ventilation is not sensitive to the human skin temperature and the microspace environment temperature, but it has a great influence on the wind speed of the microspace internal environment.When the inlet air volume increases by 5 m³/h, the skin temperature and microspace environment temperature only decrease by 0.1-0.2 ℃. With the increase of inlet air volume, the wind speed in the clothing microspace increase, the human skin blowing feeling strengthens, and the thermal comfort increases. It is found that the cooling effect is the best when the ventilation rate is 10 m³/h

Safety helmet wearing detection algorithm in complex construction environment based on improved YOLOv3

ZHAO Hongcheng, TIAN Xiuxia, YANG Zesen, BAI Wanrong

2022, 32(5): 194-200. doi:10.16265/j.cnki.issn1003-3033.2022.05.0714

Asbtract ( 270 )

HTML ( 9)

PDF (5008KB) ( 244 )

Figures and Tables | References | Related Articles | Metrics

In order to address problems of inaccurate or failed detection of safety helmet wearing under intelligent monitoring due to complex environment in construction sites, an improved YOLOv3 detection algorithm was proposed. Focal Loss was adopted to train difficult positive samples so as to improve the model's robustness in complex environment. Then, its multi-scale detection capabilities were improved by using spatial pyramid multi-level pooling based on initial network. Thirdly, attention mechanism was introduced, and channel and spatial attention modules were respectively integrated into YOLOv3's backbone and residual structure of detection layer network, so that it would focus on helmet feature learning. Finally, GIoU was utilized to improve positioning accuracy, and the algorithm's effectiveness was verified under different visual conditions in a complex construction environment. The results show that the improved model's mean accuracy reaches 88%, 13.3% higher than the original one, among which the precision of person and helmet are increased by 17.2% and 9.5%, while recall rate is increased by 15.3% and 7.6%.

Table of Content