With the development of the integration of aviation safety and artificial intelligence,research on the combination of risk assessment and artificial intelligence is particularly important in the field of risk manageme...With the development of the integration of aviation safety and artificial intelligence,research on the combination of risk assessment and artificial intelligence is particularly important in the field of risk management,but searching for an efficient and accurate risk assessment algorithm has become a challenge for the civil aviation industry.Therefore,an improved risk assessment algorithm(PS-AE-LSTM)based on long short-term memory network(LSTM)with autoencoder(AE)is proposed for the various supervised deep learning algorithms in flight safety that cannot adequately address the problem of the quality on risk level labels.Firstly,based on the normal distribution characteristics of flight data,a probability severity(PS)model is established to enhance the quality of risk assessment labels.Secondly,autoencoder is introduced to reconstruct the flight parameter data to improve the data quality.Finally,utilizing the time-series nature of flight data,a long and short-termmemory network is used to classify the risk level and improve the accuracy of risk assessment.Thus,a risk assessment experimentwas conducted to analyze a fleet landing phase dataset using the PS-AE-LSTMalgorithm to assess the risk level associated with aircraft hard landing events.The results show that the proposed algorithm achieves an accuracy of 86.45%compared with seven baseline models and has excellent risk assessment capability.展开更多
Knowledge graph technology play a more and more important role in various fields of industry and academia.This paper firstly introduces the general framework of the knowledge graph construction,which includes three st...Knowledge graph technology play a more and more important role in various fields of industry and academia.This paper firstly introduces the general framework of the knowledge graph construction,which includes three stages:information extraction,knowledge fusion and knowledge processing.In order to improve the efficiency of quality and safety supervision of transportation engineering construction,this paper constructs a knowledge graph by acquiring multi-sources heterogeneous data from supervision of transportation engineering quality and safety.It employs a bottom-up construction strategy and some natural language processing methods to solve the problems of the knowledge extraction for transportation engineering construction.We use the entity relation extraction method to extract the entity triples from the multi-sources heterogeneous data,and then employ knowledge inference to complete the edges in the constructed knowledge graph,finally perform quality evaluation to add the valid triples to the knowledge graph for updating.Subgraph matching technology is also exploited to retrieve the constructed knowledge graph for efficiently acquiring the useful knowledge about the quality and safety of transportation engineering projects.The results show that the constructed knowledge graph provides a practical and valuable tool for the quality and safety supervision of transportation engineering construction.展开更多
The paper presents the analysis of a human evacuation from a higher education building located in Iassy,Romania,by means of engineering techniques to approach fire safety.Because in Romania(as in other European countr...The paper presents the analysis of a human evacuation from a higher education building located in Iassy,Romania,by means of engineering techniques to approach fire safety.Because in Romania(as in other European countries)fire safety design of buildings is prescriptive and not performance-based,a fire safety engineering approach arouse great interest in many countries such as the U.S.A.,Australia,New Zealand,England,Sweden,Finland,etc.This paper is based on the assumption of starting a fire in the space of a hall for festivities,located on the ground floor of the building,near two human evacuation routes;We consider two building evacuation scenarios:two exits and,respectively,just one exit(assuming that the second would be accidentally blocked).展开更多
The key for dust control of coal mine is to clarify the dust concentration distribution and sedimentation in different areas. Both similarity experiment and numerical simulation method have certain restrictions and ar...The key for dust control of coal mine is to clarify the dust concentration distribution and sedimentation in different areas. Both similarity experiment and numerical simulation method have certain restrictions and are quite different from the actual situation on site. In order to study the dust sedimentation regularity of coal mine in large mining height, “filter membrane method” is adopted in this paper, i.e., to dry and weigh the filter membrane before and after sampling, collect the dust of respirable zone on mining face and calculate the dust concentration based on a main airway of 100 m. The result shows that: A large amount of dust will be produced during coal mining, wherein the maximum dust concentration from 6 m upstream to 100 m downstream of coal cutter is 121 mg/m3</sup>, while the minimum dust concentration is 61 mg/m3</sup>;The dust concentration in return airway is reduced with the distance increases, while the dust concentration at the entrance is 91 mg/m3</sup>;A large amount of dust may fall from roof during section advancing and improves the dust concentration of hydraulic support in walking area obviously;The dust granularity of mining face and return airway is 0 - 100 μm, but the amount of respirable dust is higher than 80%, the larger the dust particle size, the higher the dust concentration. Besides, dust in small particle size can be suspended in air flow for longer, but that in large particle size may subside under the action of gravity;To reduce dust exposure, the mining position shall be located in the windward direction of advancing or coal cutter. This research can provide guidance for taking dust prevention measures of working face in large mining height.展开更多
基金the National Natural Science Foundation of China(U2033213)the Fundamental Research Funds for the Central Universities(FZ2021ZZ01,FZ2022ZX50).
文摘With the development of the integration of aviation safety and artificial intelligence,research on the combination of risk assessment and artificial intelligence is particularly important in the field of risk management,but searching for an efficient and accurate risk assessment algorithm has become a challenge for the civil aviation industry.Therefore,an improved risk assessment algorithm(PS-AE-LSTM)based on long short-term memory network(LSTM)with autoencoder(AE)is proposed for the various supervised deep learning algorithms in flight safety that cannot adequately address the problem of the quality on risk level labels.Firstly,based on the normal distribution characteristics of flight data,a probability severity(PS)model is established to enhance the quality of risk assessment labels.Secondly,autoencoder is introduced to reconstruct the flight parameter data to improve the data quality.Finally,utilizing the time-series nature of flight data,a long and short-termmemory network is used to classify the risk level and improve the accuracy of risk assessment.Thus,a risk assessment experimentwas conducted to analyze a fleet landing phase dataset using the PS-AE-LSTMalgorithm to assess the risk level associated with aircraft hard landing events.The results show that the proposed algorithm achieves an accuracy of 86.45%compared with seven baseline models and has excellent risk assessment capability.
基金This work was supported by Scientific Research Project of Department of Transportation of Hunan Province under Grant No.201814.
文摘Knowledge graph technology play a more and more important role in various fields of industry and academia.This paper firstly introduces the general framework of the knowledge graph construction,which includes three stages:information extraction,knowledge fusion and knowledge processing.In order to improve the efficiency of quality and safety supervision of transportation engineering construction,this paper constructs a knowledge graph by acquiring multi-sources heterogeneous data from supervision of transportation engineering quality and safety.It employs a bottom-up construction strategy and some natural language processing methods to solve the problems of the knowledge extraction for transportation engineering construction.We use the entity relation extraction method to extract the entity triples from the multi-sources heterogeneous data,and then employ knowledge inference to complete the edges in the constructed knowledge graph,finally perform quality evaluation to add the valid triples to the knowledge graph for updating.Subgraph matching technology is also exploited to retrieve the constructed knowledge graph for efficiently acquiring the useful knowledge about the quality and safety of transportation engineering projects.The results show that the constructed knowledge graph provides a practical and valuable tool for the quality and safety supervision of transportation engineering construction.
文摘The paper presents the analysis of a human evacuation from a higher education building located in Iassy,Romania,by means of engineering techniques to approach fire safety.Because in Romania(as in other European countries)fire safety design of buildings is prescriptive and not performance-based,a fire safety engineering approach arouse great interest in many countries such as the U.S.A.,Australia,New Zealand,England,Sweden,Finland,etc.This paper is based on the assumption of starting a fire in the space of a hall for festivities,located on the ground floor of the building,near two human evacuation routes;We consider two building evacuation scenarios:two exits and,respectively,just one exit(assuming that the second would be accidentally blocked).
文摘The key for dust control of coal mine is to clarify the dust concentration distribution and sedimentation in different areas. Both similarity experiment and numerical simulation method have certain restrictions and are quite different from the actual situation on site. In order to study the dust sedimentation regularity of coal mine in large mining height, “filter membrane method” is adopted in this paper, i.e., to dry and weigh the filter membrane before and after sampling, collect the dust of respirable zone on mining face and calculate the dust concentration based on a main airway of 100 m. The result shows that: A large amount of dust will be produced during coal mining, wherein the maximum dust concentration from 6 m upstream to 100 m downstream of coal cutter is 121 mg/m3</sup>, while the minimum dust concentration is 61 mg/m3</sup>;The dust concentration in return airway is reduced with the distance increases, while the dust concentration at the entrance is 91 mg/m3</sup>;A large amount of dust may fall from roof during section advancing and improves the dust concentration of hydraulic support in walking area obviously;The dust granularity of mining face and return airway is 0 - 100 μm, but the amount of respirable dust is higher than 80%, the larger the dust particle size, the higher the dust concentration. Besides, dust in small particle size can be suspended in air flow for longer, but that in large particle size may subside under the action of gravity;To reduce dust exposure, the mining position shall be located in the windward direction of advancing or coal cutter. This research can provide guidance for taking dust prevention measures of working face in large mining height.
