Human factor analysis method and system for event that aircraft rushes out of runway

Abstract

The invention discloses a human factor analysis method and system for an event that an aircraft rushes out of a runway, and the method comprises the steps: constructing a human factor investigation decision support model HFIDSM through combining an accident tree analysis method and an improved HFACS model, and recognizing and obtaining dominant factors and hidden factors which cause the event that the aircraft rushes out of the runway; the cause factors of the event that the aircraft rushes out of the runway are excavated; constructing a human factor cause Bayesian network model of the event that the aircraft rushes out of the runway, and constructing a key cause chain causing the event that the aircraft rushes out of the runway; and analyzing the key cause chain, and optimizing the operation regulation of the aircraft according to the analysis result. According to the invention, by deeply analyzing the complex causes of the runway event, the key cause chain behind the event is accurately identified and quantitatively analyzed by using the Bayesian network, so that the understanding of the interaction relationship among multi-dimensional cause factors such as human errors, environmental factors and equipment faults is enhanced; and systematization and intellectualization of the investigation process are promoted.

Description

Technical Field

[0001] The present invention relates to the field of civil aviation safety management and accident investigation technology, and in particular to a method and system for analyzing human factors in an aircraft running off a runway. Background Art

[0002] Aircraft runway deviation incidents are a major type of accident that seriously threatens flight safety. There are many analysis methods for aircraft runway deviation incidents in the existing technology. For example, comprehensive research on runway deviation incidents at special airports, risk coupling assessment and prediction and early warning can reduce the probability of runway deviation incidents and mitigate the impact of accident consequences to a certain extent; there are also methods for real-time prediction of landing runway deviation risks, which can reduce the probability of runway deviation during landing to a certain extent. However, the above accident investigation methods focus on technical failures and operational errors, ignoring the in-depth analysis of human factors, and it is difficult to systematically reveal the complex cause and effect relationship behind the accident; and lack of systematicity and accuracy, it is difficult to effectively support accident prevention and decision-making improvement, especially in the analysis of human errors and organizational factors intertwined with multiple factors. Summary of the invention

[0003] The present invention aims to solve the technical problems that the investigation and analysis method of aircraft runway deviation events in the prior art ignores the hierarchical analysis of human factors and the accuracy and efficiency of event investigation are low, and provides a human factor analysis method and system for aircraft runway deviation events.

[0004] In view of the above technical problems, an embodiment of the present invention provides a method for analyzing human factors in an aircraft running off a runway, comprising:

[0005] Combining the accident tree analysis method with the improved HFACS model, a human factor investigation decision support model HFIDSM was constructed to identify and obtain the explicit and implicit factors that lead to aircraft running off the runway.

[0006] The causal factors of aircraft runway deviation events were explored through the human factors investigation decision support model HFIDSM;

[0007] Construct a Bayesian network model of the human factors that cause aircraft runway deviation events, quantify the contribution of the causal factors to the occurrence of the event, and construct the key causal chain that leads to aircraft runway deviation events;

[0008] Analyze the critical cause chain and optimize aircraft operating procedures based on the analysis results.

[0009] The present invention also provides a human factor analysis system for an aircraft running off a runway, comprising:

[0010] A construction module is used to combine the accident tree analysis method and the improved HFACS model to build a human factor investigation decision support model HFIDSM to identify and obtain the explicit and implicit factors that lead to aircraft running off the runway;

[0011] The causal factors mining module is used to mine the causal factors of aircraft runway excursions through the human factors investigation decision support model HFIDSM;

[0012] The key cause chain construction module is used to construct a Bayesian network model of the human factors that cause the aircraft to runway deviation, quantify the contribution of the causal factors to the event, and construct the key cause chain that leads to the aircraft running off the runway.

[0013] The optimization module is used to analyze the critical cause chain and optimize the aircraft operating procedures based on the analysis results.

[0014] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0015] In the present invention, the human factor investigation decision support model HFIDSM realizes a comprehensive analysis of the causal chain of unsafe incidents of aircraft running off the runway by combining the accident tree and the improved HFACS model, covering all links from direct operational errors to deep-level management defects; airlines and regulatory agencies can obtain the scientific basis required to formulate targeted improvement measures based on the model, thereby effectively improving aviation safety.

[0016] In the present invention, the HFIDSM model not only deeply explores the causes of runway deviation incidents, but also comprehensively reveals the complex relationship between explicit and implicit factors, providing a solid theoretical basis and decision-making support for accident prevention, investigation and subsequent improvement measures. In addition, the model also introduces the Bayesian network BN, which has strong scalability and flexibility, can efficiently identify the key causal chain in the accident, and reveal the dynamic interaction process of complex accident causes; it not only improves the comprehensiveness and depth of the accident investigation, but also ensures the accuracy and efficiency of the investigation results, and provides a strong scientific basis for aviation safety management and the formulation of preventive measures. The present invention provides a new systematic and scientific tool for the investigation and analysis of aircraft runway deviation incidents, which has important application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a method for analyzing human factors of an aircraft running off the runway in one embodiment of the present invention;

[0018] Figure 2 It is a structural schematic diagram of an improved HFACS model in one embodiment of the present invention;

[0019] Figure 3 It is a schematic diagram of analyzing the dominant factors in the human factors of an aircraft running off the runway in one embodiment of the present invention;

[0020] Figure 4 It is a schematic diagram of analyzing the hidden factors in the human factors of an aircraft running off the runway in one embodiment of the present invention;

[0021] Figure 5 A schematic diagram of the structure of a Bayesian network model in one embodiment of the present invention;

[0022] Figure 6 It is a structural diagram of a human factor analysis system for an aircraft running off the runway in one embodiment of the present invention;

[0023] Figure 7 A schematic diagram of causal reasoning of a Bayesian model for a runway deviation event in one embodiment of the present invention;

[0024] Figure 8 It is a schematic diagram of reverse diagnosis reasoning of a runway deviation event in one embodiment of the present invention;

[0025] Fig. 9 A schematic diagram of the Bayesian network sensitivity analysis results in one embodiment of the present invention;

[0026] Fig.10 A schematic diagram of the most critical causal chain of human factors causing runway deviation in one embodiment of the present invention;

[0027] Fig.11 Schematic diagram of the effectiveness verification of the human factor analysis model of the Bayesian network in one embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] In one embodiment, if Figure 1 As shown, the present invention provides a method for analyzing human factors in an aircraft running off a runway, comprising the following steps:

[0030] S10. Combining the accident tree analysis method and the improved HFACS model, a human factor investigation decision support model HFIDSM is constructed to identify and obtain the explicit factors and implicit factors that cause the aircraft to deviate from the runway; understandably, the accident tree analysis method is a tool for system safety analysis, which uses logic gates (such as AND gates and OR gates) to refine the fault causes of the system layer by layer to form an inverted tree diagram, and uses the accident tree analysis method to perform a structured decomposition of the aircraft deviating from the runway event to identify the various explicit factors (direct causes) that cause the event. The improved HFACS model uses its improved HFACS framework to deeply analyze the implicit factors (root causes) that trigger the explicit factors.

[0031] The improved HFACS model is an improvement made on the existing HFACS model. Specifically, the sub-indicators of each level are modified accordingly according to the aircraft running off the runway event. There are no more complicated connections between the levels, and the accident cause chain tends to be a simple relationship.

[0032] The improved HFACS model divides human factors into multiple levels, namely unsafe behaviors, preconditions for unsafe behaviors, unsafe supervision, and poor organizational management; among them, unsafe behaviors are usually explicit factors, and preconditions for safe behaviors, unsafe supervision, and poor organizational management are usually implicit factors. Implicit factors are often the deep-seated causes of unsafe behaviors. Through the analysis of implicit factors, the improved HFACS model can systematically reveal problems at the management and operational levels, providing deeper insights for accident prevention.

[0033] In one embodiment, if Figure 2 As shown, in order to improve the phenomenon that the indicators at each level in the existing HFACS model are out of touch with the factual information and need to be optimized, the HFACS model is improved from four aspects: poor organizational management, unsafe supervision, preconditions for unsafe behavior, and unsafe behavior, in order to analyze the deep-seated causes of human factors and explore the causal chain of unsafe incidents, so as to avoid similar incidents from happening again. Specifically:

[0034] Poor organizational management includes improper leadership decision-making, improper resource management and improper policy formulation. Among them, improper leadership decision-making includes weak awareness of risk classification and control, untimely risk hazard investigation, inaccurate and imperfect decision-making, etc.; improper resource management includes indicators such as incomplete organizational structure, imperfect rules and regulations, and insufficient human resource investment; improper policy formulation includes imperfect standard operating procedures and inadequate safety risk control.

[0035] Unsafe supervision includes indicators such as improper scheduling management, insufficient supervision, untimely risk warnings, and inadequate safety training. Among them, improper scheduling management includes indicators such as unreasonable team resource matching, weak load management capabilities, fatigue / overtime work, etc.; insufficient supervision includes indicators such as lack of real-time supervision, failure to perform supervisory duties, lack of feedback and reward mechanisms, and formal safety inspections; untimely risk warnings include indicators such as untimely control intelligence suggestions and weak dispatch risk management capabilities; inadequate safety training includes indicators such as untimely policy dissemination and incomplete skills training.

[0036] The prerequisites for unsafe behavior include indicators such as poor personnel status, complex operating environment, and mechanical / equipment failure. Among them, poor personnel status includes indicators such as personnel qualifications, physical and mental health, business capabilities, and work style. Personnel qualifications include expired licenses and expired medical certificates; physical and mental health mainly includes abnormal physiological factors and abnormal psychological factors; business capabilities include poor theoretical / practical ability, poor coordination and cooperation, and poor situational awareness; work style includes poor initiative and poor sense of responsibility. Complex operating environments include marginal weather and complex terrain. Marginal weather includes low visibility / night operation, wind shear / backwind and tailwind, and thunderstorms; complex terrain includes plateau / high plateau airports and special airports; machinery and equipment failures include hardware design / manufacturing defects, hardware not maintained according to procedures, imperfect software system warning functions, and equipment failures caused by personnel misoperation.

[0037] Combined with the definition of situational awareness, unsafe behaviors are divided into improper information preprocessing, improper problem solving / decision-making, and improper action execution; this makes it easier for investigators to combine factual information and identify problems with unsafe behaviors from different aspects in order to make targeted improvements.

[0038] In one embodiment, the step S10 further includes the following steps:

[0039] S101. Use the accident tree analysis method to reversely deduce the runway deviation event as the top-level event, analyze it layer by layer, and identify and obtain all the explicit factors at each level of the aircraft runway deviation event; understandably, through layer-by-layer analysis, the direct cause (explicit factor) leading to the runway deviation event can be clearly identified and obtained, such as Figure 3 As shown in the figure, the accident tree analysis reveals the logical relationship between various explicit factors, which helps to understand the specific path and causes of the incident.

[0040] S102. Use the improved HFACS model to analyze and obtain the implicit factors corresponding to all the explicit factors; understandably, combining explicit factors with implicit factors can form a complete causal chain, which helps to fully understand the deep-seated reasons for the incident. By analyzing the implicit factors, potential systemic defects and weak links in safety management can be revealed, providing targeted suggestions for improvement measures. Figure 4 As shown, this is a schematic diagram of the analysis results of implicit factors, which can clearly show the specific content of these implicit factors and clearly express the logical relationship between them.

[0041] S103. Combining explicit factors and implicit factors, construct a human factor investigation decision support model HFIDSM. Understandably, the human factor investigation decision support model HFIDSM combines explicit factors and implicit factors to form a comprehensive evaluation method that can systematically and scientifically analyze the human factors of aircraft runway deviation events. Specifically, the steps for constructing the human factor investigation decision support model HFIDSM include: data collection and preprocessing, identification and analysis of explicit factors, identification and analysis of implicit factors, and construction of the HFIDSM model based on the correlation between explicit factors and implicit factors. The specific construction procedure of the HFIDSM model can be set as needed.

[0042] S20. The causal factors of aircraft runway deviation incidents are explored through the human factor investigation decision support model HFIDSM. Understandably, by deeply analyzing the human factors in unsafe incidents from macro and micro perspectives, as well as the logical relationship between the various causes, the causal factors are finally obtained.

[0043] In one embodiment, the step S20 includes the following steps:

[0044] S201. Based on the human factor investigation decision support model HFIDSM, various relevant data of aircraft runway deviation events are collected, and various relevant data are summarized into associated data. The associated data are pre-processed to eliminate invalid data and irrelevant data to obtain associated data. It can be understood that the associated data of the runway deviation event include accident investigation reports, pilot operation records, flight data records, cockpit voice records, meteorological data and airport runway conditions.

[0045] S202. Using the accident analysis method, the top-level event is determined as the aircraft running off the runway event. Starting from the top-level event, the aircraft running off the runway event is structurally decomposed layer by layer, and the direct cause of the aircraft running off the runway event is analyzed to construct an accident tree. Understandably, this accident analysis method can show in detail how unsafe behaviors gradually lead to the occurrence of events through a series of logical steps, and reveal the position and role of explicit behaviors (such as pilots' decision-making errors, improper instructions from controllers, etc.) in the causal chain.

[0046] In one embodiment, constructing the accident tree in step S202 further includes the following steps:

[0047] Starting from the top-level event, the aircraft runway deviation event is structurally decomposed layer by layer.

[0048] Identify and obtain the direct causes of the top-level events, i.e., the intermediate events;

[0049] Continue to decompose each intermediate event until the basic event is found;

[0050] Use logic gates to connect top-level events, intermediate events, and basic events to form a graphical accident tree.

[0051] Understandably, the accident tree starts from the top-level event (i.e., the aircraft veers off the runway event) and performs structured decomposition layer by layer, gradually breaking down complex events into manageable components. This method helps to clarify the logical relationship of events and avoid missing important influencing factors. Through the construction of the accident tree, the direct cause of the top-level event (i.e., the intermediate event) can be systematically identified. These intermediate events provide clear directions and clues for further analysis. The intermediate events are further decomposed until the basic events (i.e., indivisible events or faults) are found. Basic events are the smallest units that constitute the accident tree, and they directly reflect the specific causes of the events. Use logic gates to connect the top-level events, intermediate events, and basic events to form a graphical accident tree. This graphical expression intuitively shows the causal and logical relationships between events, which is easy to understand and analyze. Accident tree analysis not only considers the role of a single factor, but also the interaction and influence between multiple factors, which helps to comprehensively and systematically reveal the causal factors of the event.

[0052] S203. Analyze the accident tree. According to the results of the accident tree analysis, take the direct cause of the aircraft running off the runway as the starting point for the analysis of the improved HFACS model. Apply the multi-level classification framework of the improved HFACS model to analyze and obtain the hidden factors behind the unsafe behavior. It is understandable that by analyzing these hidden factors, potential systemic defects can be revealed and weak links in safety management can be identified.

[0053] The multi-level classification framework of the improved HFACS model includes organizational influence, unsafe supervision, preconditions of unsafe behavior and unsafe behavior.

[0054] S204. Combining the explicit factors of the accident tree analysis and the implicit factors of the analysis of the improved HFACS model, the causal factors of the aircraft running off the runway are obtained. The causal factors can provide a basis for the prevention and management improvement of similar events in the future.

[0055] In one embodiment, the dominant factors include pilot operational errors, mechanical equipment failures, operating environment, human management omissions, improper scheduling management, poor personnel status, and operating environment;

[0056] The implicit factors include the pilot's psychological state, flight mission allocation, training management, implementation of operating procedures, organizational culture and resource allocation.

[0057] S30. Construct a Bayesian network model of the human factors that cause aircraft runway deviations, quantify the contribution of the causal factors to the occurrence of the event, and construct the key causal chain that leads to aircraft runway deviations; understandably, based on explicit and implicit factors, a systematic and scientific model is constructed through the Bayesian network to analyze the human factors causal chain in aircraft runway deviations. This model can help investigators more accurately identify and analyze the key factors in the event, and provide data support for preventing similar events.

[0058] In one embodiment, the step S30 includes the following steps:

[0059] S301, based on the explicit factors and the implicit factors, define variable nodes, express the causal relationship between variable nodes through directed edges, and establish a human factor causal Bayesian network structure according to variable nodes, directed edges and conditional probability tables, such as Figure 5 As shown in the figure, the Bayesian network model of human factors causation and the aircraft running off the runway event.

[0060] In one embodiment, the step S301 further includes the following steps:

[0061] S3011. Identify all relevant human factors that may cause an aircraft to deviate from the runway, including explicit factors and implicit factors. Understandably, ensure that no human factors that may affect an aircraft to deviate from the runway are missed, including explicit factors and implicit factors.

[0062] S3012. Define variable nodes based on human factors, express the causal relationship between variable nodes through directed edges, connect variable nodes with directed edges, and obtain directed edges to reflect the logical order and dependency of events; understandably, by defining variable nodes and directed edges, the causal relationship and logical order between human factors are clearly expressed. Directed edges not only express the connection between factors, but also reflect the logical order and dependency of events.

[0063] S3013. Provide a conditional probability table for each variable node and its parent node, quantify the specific numerical value of the correlation strength between the variable nodes, set the conditional probability value, and construct the conditional probability table based on the conditional probability value; understandably, each variable node represents a human factor, such as pilot operating error, fatigue work, etc.; a directed edge represents the influence of one variable node on another variable node, such as fatigue work leading to pilot operating error.

[0064] S3014. Establish the structure of the Bayesian network model of human factors causation based on variable nodes, directed edges and conditional probability tables, verify the Bayesian network model of human factors causation using actual cases or simulated data, and adjust the model structure and parameters based on the verification results. Understandably, the correlation strength between each variable node is quantified through the conditional probability table, making the model analysis more scientific and accurate. When there is not enough data, the probability value is directly assigned using expert knowledge to improve the practicality of the model.

[0065] S302. Use the Bayesian reasoning method to reason on the conditional probability table, identify the key factors among the causal factors, determine the key factors as key nodes, and perform weight evaluation on the key nodes; understandably, the conditional probability table lists all possible state combinations and their corresponding probability values ​​for each non-root node.

[0066] In one embodiment, the step S302 further includes the following steps:

[0067] S3021. After the Bayesian network model of human-caused factors is constructed, the Bayesian reasoning method is used to reversely infer the probability changes of other unobserved factors based on the human factors. It can be understood that through Bayesian reasoning, the interaction between various potential factors and their combined impact on the occurrence of events can be comprehensively and systematically considered; reverse inference can help identify which factors have the greatest impact on the occurrence of events in unobserved situations, and provide support for subsequent decision-making.

[0068] S3022. Through the results of Bayesian reasoning, analyze the influence of each node in the causal chain and determine the weight of the key nodes; understandably, determining the weight of the key nodes helps to clarify priorities in prevention and management, and give priority to solving those factors that have the greatest impact on the incident; through weight analysis, targeted improvement measures can be proposed to minimize the probability of the incident.

[0069] S3023. Use actual accident data to verify the inference results. If the inference results are found to be inconsistent with the actual situation, recheck the data input, Bayesian network model structure and conditional probability table, and make corresponding adjustments and optimizations. Understandably, through verification of actual data, the accuracy and reliability of the inference results can be ensured; through continuous adjustment and optimization of the model, the adaptability and predictive ability of the model can be improved to make it more in line with the actual situation.

[0070] S303. Apply the Bayesian network model of human factor causation to analyze the impact of different combinations of causal factors on aircraft runway excursion events, conduct diagnostic reasoning, identify the key causal chain of aircraft runway excursion events based on posterior probability, and provide improvement directions for the management department.

[0071] In one embodiment, the step S303 further includes the following steps:

[0072] S3031. Apply forward reasoning, starting from known nodes, using the topological structure and conditional probability table of the Bayesian network to predict the probability of an aircraft running off the runway; understandably, using the constructed Bayesian network model and known node states (such as equipment status, personnel qualifications, operating environment, etc.), forward reasoning (predictive reasoning) is used to calculate the probability of an aircraft running off the runway. This step can help the safety management department predict risks based on the current state and conditions before the event actually occurs, and take preventive measures in advance.

[0073] S3032, apply backward reasoning, set the aircraft running off the runway event as a known state, reversely infer the state probability changes of other nodes, and identify the factors that have the greatest impact on the aircraft running off the runway event, that is, the key causal factors; understandably, set the aircraft running off the runway event as a known state (that is, set its probability of occurrence to 100%), and through reverse reasoning (diagnostic reasoning), calculate and identify the factors that have the greatest impact on the aircraft running off the runway event, that is, the key causal factors. This step can help investigators quickly locate the cause of the accident and find out the links that need to be focused on.

[0074] S3033. Combine the results of forward and backward reasoning, and use sensitivity analysis to identify key parameters and obtain the key causal chain that led to the aircraft running off the runway. Understandably, by combining the results of forward and backward reasoning, we can more comprehensively understand the causal chain of the aircraft running off the runway. Further, use sensitivity analysis to evaluate the impact of different parameter changes on the probability of the final event, and identify the key parameters and causal chains that have the greatest impact on the results. This helps to determine the focus of prevention and improvement measures.

[0075] S3034. Display the key cause chain in the form of a chart or list, showing the cause-effect relationship and influencing factors in each chain, and providing intuitive analysis results for the aviation safety management department. Understandably, displaying the key cause chain in the form of a chart or list can clearly show the cause-effect relationship and influencing factors in each chain. This not only makes it easier for safety management departments and investigators to understand the complex causes of accidents, but also provides an intuitive reference for formulating preventive measures and improvement plans.

[0076] S40. Analyze the critical cause chain and optimize the aircraft's operating procedures based on the analysis results. Understandably, the model analysis results can also be used to optimize existing operating procedures, training plans and risk warning systems to improve the overall safety of aviation operations. By constructing and analyzing the critical cause chain, the deep-seated causes of aircraft runway deviation incidents can be systematically revealed, including potential defects in organizational management, supervision, personnel status, operational behavior and other aspects. These defects are often difficult to discover through individual analysis, and the construction of the critical cause chain allows these potential problems to be exposed.

[0077] Based on the analysis results of the key cause chain, the aircraft operation procedures can be optimized in a targeted manner. For example, if it is found that imperfect rules and regulations are an important cause of the accident, then the relevant rules and regulations can be optimized to ensure their integrity and effectiveness. If it is found that improper scheduling management affects the fatigue level of pilots, the scheduling system can be adjusted to ensure that pilots get enough rest.

[0078] The analysis results can also be used to improve pilot training programs. By identifying pilots’ inappropriate behaviors in specific situations, relevant training can be strengthened in a targeted manner to improve pilots’ skill levels and their ability to cope with complex situations. For example, if it is found that pilots have poor situational awareness in complex environments, situational awareness training can be strengthened to improve pilots’ ability to cope with them.

[0079] By optimizing operating procedures, improving training programs and risk early warning systems, the overall safety of aviation operations can be significantly improved, which will not only help reduce accidents such as aircraft running off the runway, but also improve the operational efficiency and market competitiveness of airlines.

[0080] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0081] In one embodiment, a human factor analysis system for an aircraft running off the runway is provided, and the human factor analysis system for an aircraft running off the runway corresponds to the human factor analysis method for an aircraft running off the runway in the above embodiment. As shown in Figure 6, the human factor analysis system for an aircraft running off the runway includes:

[0082] Building module 100, for combining the accident tree analysis method and the improved HFACS model, building a human factor investigation decision support model HFIDSM, identifying and obtaining explicit factors and implicit factors that lead to the aircraft running off the runway;

[0083] The causal factor mining module 200 is used to mine the causal factors of the aircraft running off the runway event through the human factor investigation decision support model HFIDSM;

[0084] The key cause chain construction module 300 is used to construct a Bayesian network model of the human factor cause of the aircraft running off the runway event, quantify the contribution of the cause factors to the event, and construct the key cause chain leading to the aircraft running off the runway event;

[0085] The optimization module 400 is used to analyze the critical cause chain and optimize the aircraft operation procedures according to the analysis results.

[0086] In one embodiment, the runway deviation events, which accounted for a relatively large proportion of serious symptom events in the past 10 years, were used for verification, and the causal chain of the runway deviation events was obtained as shown in Table 1:

[0087] Table 1 Human factors causal chain of runway deviation events

[0088]

[0089]

[0090] Based on the causal chain of HFIDSM analysis, a Bayesian network model of runway deviation events is constructed. The model contains four continuous causal paths that affect the probability of runway deviation events, namely poor organizational management, unsafe supervision, preconditions for unsafe behavior, and unsafe behavior. At the same time, each causal path also contains multiple human factor sub-nodes. The investigation reports of runway deviation accidents in recent years are mined and analyzed, and keywords related to causal factors are extracted. Their frequencies are calculated and converted into the prior probability of the factor as shown in Table 2:

[0091] Table 2 Prior probability analysis of human factors for runway deviation

[0092]

[0093] Direct assignment is an effective method that relies on expert knowledge to define the conditional probability of non-root nodes in Bayesian networks, especially when data is insufficient or domain knowledge is complex. This method is implemented through four core steps: first, clarify the topological structure of the Bayesian network, that is, the causal relationship between nodes, and identify the parent node of each non-root node; second, list all possible state combinations of non-root nodes and their parent nodes to ensure that all possible situations are covered; then, based on expert knowledge, directly assign probability values ​​to each non-root node under different state combinations of its parent nodes. This process relies on the rich experience and expertise of experts; finally, construct a conditional probability table (CPT) and fill in the above probability values ​​based on expert knowledge for subsequent analysis. This direct assignment method not only improves the accuracy of the model in the case of scarce data, but also enhances the application ability of the model in complex fields, providing reliable data support for the optimization of aircraft operating procedures.

[0094] In one embodiment, if Figure 7 As shown in the figure, by combining the prior probability obtained based on the statistical frequency of accident investigation reports and the conditional probability obtained based on expert experience, these probability values ​​are input into each node of the constructed Bayesian network model. This process adopts a top-down reasoning method, namely causal reasoning. Through the complex structure of the Bayesian network, the causal relationship between human factors can be systematically analyzed and quantified, thereby obtaining the causal reasoning results of human factors in runway deviation events. This method provides a scientific basis and comprehensive evaluation for the improvement of aviation safety management measures.

[0095] like Figure 8 As shown in the figure, the reasoning mechanism of the Bayesian network relies on the probabilistic relationship defined in the network, and calculates the posterior probability of the observed variable through the propagation and updating of information. This reasoning process can be divided into forward reasoning (or predictive reasoning) and backward reasoning (also known as diagnostic reasoning). Forward reasoning focuses on inferring potential results from known causes, while backward reasoning reverses this process, starting from the known results and inferring their possible causes and their impact.

[0096] In Bayesian networks, diagnostic reasoning specifically refers to reverse analysis and updating of the posterior probability distribution of all potential influencing factors (i.e., parent nodes) by setting the probability of a certain node (usually a result node) (e.g., "runway deviation" is set to 100%). This method helps to assess the specific impact of each risk factor when a specific event occurs, thereby revealing the key factors behind the event. By using Bayesian network software such as GeNIe, researchers can easily perform this bottom-up reasoning process. By setting the probability value of the target node to a specific value (such as 100%) and triggering the probability refresh function of the software, the software will automatically calculate and display the posterior probability values ​​of all relevant parent nodes, providing intuitive data support for understanding the causal relationship of complex events. This diagnostic reasoning method has a wide range of application value in aviation safety analysis, risk assessment, and decision-making.

[0097] The higher the posterior probability value of the Bayesian network node, the greater the risk contribution of this factor to the runway deviation event. It can be seen that the risk contribution of the second-level nodes of the human factors of the runway deviation event is as follows: unsafe behavior, preconditions for unsafe behavior, poor organizational management, and unsafe supervision. By comparing the diagnostic reasoning results with the prior probability and conditional probability results of each node, it can be seen that the probability of each root node has not changed, and the posterior probabilities of most non-root nodes have changed to varying degrees, as shown in Table 3:

[0098] Table 3 Probability change results of reverse reasoning of node “off track”

[0099]

[0100] like Fig. 9 As shown in the figure, sensitivity analysis is a key method to evaluate the sensitivity of Bayesian network models to changes in input parameters. It aims to deeply explore how small changes in various parameters (especially probability distribution) in the network significantly affect the output (posterior probability) of the model. Through sensitivity analysis, model designers can clearly identify which parameters or network structures occupy a core position in the model, and their changes will directly lead to significant fluctuations in the model output. This process not only deepens the understanding of the internal mechanism of the model, but also provides important clues for subsequent model optimization, ensuring that the model has sufficient stability and accuracy in practical applications.

[0101] In Bayesian networks, sensitivity analysis has become an important technical means to verify and optimize probability parameters. By making small adjustments to the model's numerical parameters (including prior probabilities and conditional probabilities) and observing how these adjustments affect changes in output parameters (posterior probabilities), a quantitative assessment of the model's sensitivity can be achieved. Those parameters that have the most significant impact on the output, i.e., highly sensitive parameters, are considered key factors in the model's reasoning process. Identifying and focusing on these parameters helps model designers estimate their values ​​more accurately, thereby significantly improving the accuracy and reliability of the model. In addition, sensitivity analysis can also help reveal potential weaknesses in the model and provide directional guidance for the improvement and optimization of the model structure.

[0102] like Fig.10 As shown, it is assumed that the state of the target node "off the runway" is Yes, that is, p("off the runway" state = Yes) = 100%. Then, the posterior probability of the network node is updated through reverse reasoning diagnosis. At this time, "unsafe behavior" is the parent node with the largest posterior probability of off the runway state, and the key causal chain reasoning is {"unsafe behavior" → "off the runway"}. Then, continue to look for the parent node with the largest posterior probability of the state from the node "unsafe behavior" until the root node of the network. Finally, it can be found that the most critical causal chain among the human factors of off the runway is {"improper action execution / improper information preprocessing" → "unsafe behavior" → "off the runway"}, which is crucial for the investigation and prevention of unsafe incidents. Similarly, three other key chains of human-caused causes of runway deviation can be found: {“poor theoretical and practical ability / poor coordination and cooperation ability” → “work skills” → “poor personnel status” → “preconditions for unsafe behavior” → “runway deviation”}, {“inadequate safety risk control” → “improper policy formulation” → “poor organizational management” → “runway deviation”}, {“incomplete skills training” → “inadequate safety training” → “unsafe supervision” → “runway deviation”}. The red nodes in the figure contain parameters that are important for the posterior probability distribution of the target node (runway deviation), and the gray nodes do not contain any parameters used to calculate the target posterior probability distribution. The darker the color, the greater the importance in the network; the thicker the arrow, the greater the weight value.

[0103] In a specific embodiment, the effectiveness of the Bayesian network model of human factors was verified. On July 7, 2018, Shenzhen Airlines flight ZH9127 deviated from the runway during landing in Hohhot, resulting in a tire blowout on the left outer main wheel and the left front wheel of the aircraft, but no one was injured. The incident was determined to be a serious sign of transport aviation due to human responsibility. After the incident, the North China Bureau responded quickly, established an investigation team and deployed emergency disposal and investigation work, and conducted a comprehensive investigation through on-site investigation, data decoding, evidence collection and other means. It was finally determined that the direct causes were the lack of flying skills of the operating pilot, the lack of crew resource management and coordination problems. At the same time, the crew's weak safety awareness, lack of emergency experience and complex thunderstorm weather also aggravated the severity of the incident.

[0104] In order to deeply analyze the human factors of the runway deviation incident, we compared the survey results with the Bayesian network model of the human factors of the runway deviation. Fig.11 As shown in the figure, in the model, we set the status of multiple key nodes such as "thunderstorm weather", "coordination", and "improper information preprocessing" to "Yes" according to the event situation to simulate the actual environment and human factors. Subsequently, through the top-down causal reasoning method, the network node status is refreshed, and the probability of the runway deviation event is calculated to be 78%. This result is highly consistent with the actual situation, which not only reveals the complex human-computer interaction process of the event, but also verifies the effectiveness and practicality of the human factor analysis model based on the Bayesian network in the investigation of aviation safety events.

[0105] The successful analysis of this runway deviation incident not only provides a scientific basis for the incident investigation, but also further verifies the potential of the Bayesian network model in identifying and managing aviation human factor risks. Through the application of the model, we can more accurately assess the impact of various human factors on flight safety and provide strong support for the formulation of targeted preventive measures. In the future, as the aviation industry continues to pay attention to safety management and technical means continue to advance, the human factor analysis model based on the Bayesian network is expected to play a more important role in the field of aviation safety and contribute to improving the level of global aviation safety.

[0106] The above are merely embodiments of the human factor analysis method and system for aircraft runway deviation events of the present invention, and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for analyzing human factors in aircraft runway deviation events, characterized in that: include: S10. Combine the accident tree analysis method and the improved HFACS model to build the human factor investigation decision support model HFIDSM to identify and obtain the explicit and implicit factors that lead to aircraft running off the runway; S20. Exploring the causal factors of aircraft runway deviation events through the human factors investigation decision support model HFIDSM; S30. Construct a Bayesian network model of the human factors that cause the aircraft to deviate from the runway, quantify the contribution of the causal factors to the occurrence of the event, and construct the key causal chain that leads to the aircraft to deviate from the runway; S40. Analyze the critical cause chain and optimize the aircraft operating procedures based on the analysis results.

2. The human factor analysis method for an aircraft runway deviation event according to claim 1, characterized in that: The step S10 further comprises the following steps: S101. Use the accident tree analysis method to reversely deduce the runway deviation event as the top-level event, analyze it layer by layer, and identify and obtain all the explicit factors of each level of the aircraft runway deviation event; S102, using the improved HFACS model to analyze and obtain the implicit factors corresponding to all the explicit factors; S103. Combining explicit factors and implicit factors, a human factor investigation decision support model HFIDSM is constructed.

3. The human factor analysis method for an aircraft runway deviation event according to claim 1, characterized in that: The step S20 comprises the following steps: S201, based on the human factor investigation decision support model HFIDSM, various relevant data of aircraft runway deviation events are collected, various relevant data are aggregated into associated data, and the associated data are pre-processed to eliminate invalid data and irrelevant data to obtain associated data; S202, using the accident analysis method, determining the top-level event as the aircraft running off the runway event, starting from the top-level event, structurally decomposing the aircraft running off the runway event layer by layer, analyzing the direct cause of the aircraft running off the runway event, and constructing an accident tree; S203. Analyze the accident tree. According to the results of the accident tree analysis, take the direct cause of the aircraft running off the runway as the starting point for analyzing the improved HFACS model. Apply the multi-level classification framework of the improved HFACS model to analyze and obtain the hidden factors behind the unsafe behavior. The multi-level classification framework of the improved HFACS model includes organizational influence, unsafe supervision, preconditions of unsafe behavior and unsafe behavior; S204. Combine the explicit factors of the accident tree analysis with the implicit factors of the analysis of the improved HFACS model to obtain the causal factors of the aircraft running off the runway.

4. The method for analyzing human factors of an aircraft running off the runway according to claim 3, characterized in that: The associated data of the runway overrun event includes accident investigation report, pilot operation record, flight data record, cockpit voice record, meteorological data and airport runway conditions.

5. The method for analyzing human factors of an aircraft running off the runway according to claim 3, characterized in that: The explicit factors include pilot operational errors, mechanical equipment failures, operating environment, human management omissions, improper scheduling management, poor personnel status and operating environment; The hidden factors include the pilot's psychological state, flight mission allocation, training management, implementation of operating procedures, organizational culture and resource allocation; The construction of the accident tree in step S202 further includes the following steps: Starting from the top-level event, the aircraft runway deviation event is structurally decomposed layer by layer. Identify and obtain the direct causes of the top-level events, i.e., the intermediate events; Continue to decompose each intermediate event until the basic event is found; Use logic gates to connect top-level events, intermediate events, and basic events to form a graphical accident tree.

6. The method for analyzing human factors of an aircraft running off the runway according to claim 1, characterized in that: The step S30 comprises the following steps: S301, based on the explicit factors and the implicit factors, define variable nodes, express the causal relationship between the variable nodes through directed edges, establish a human factor causal Bayesian network structure according to the variable nodes, directed edges and conditional probability tables, and form a human factor causal Bayesian network model, and an aircraft runway deviation event; S302, using the Bayesian reasoning method to reason on the conditional probability table, identifying the key factors among the causal factors, determining the key factors as key nodes, and performing weight evaluation on the key nodes; S303. Apply the Bayesian network model of human factor causation to analyze the impact of different combinations of causal factors on aircraft runway excursion events, conduct diagnostic reasoning, identify the key causal chain of aircraft runway excursion events based on posterior probability, and provide improvement directions for the management department.

7. The human factor analysis method for an aircraft runway deviation event according to claim 6, characterized in that: The step S301 further includes the following steps: S3011. Identify all relevant human factors of the aircraft running off the runway, including explicit factors and implicit factors; S3012, define variable nodes based on human factors, express the causal relationship between variable nodes through directed edges, connect variable nodes with directed edges, and obtain directed edges to reflect the logical order and dependency of events; S3013, equip each variable node and its parent node with a conditional probability table, quantify the specific value of the correlation strength between the variable nodes, set the conditional probability value, and construct the conditional probability table according to the conditional probability value; S3014. Establish the structure of the Bayesian network model of human factor causation based on variable nodes, directed edges and conditional probability tables, verify the Bayesian network model of human factor causation using actual cases or simulated data, and adjust the model structure and parameters based on the verification results.

8. The method for analyzing human factors of an aircraft running off the runway according to claim 7, characterized in that: The step S302 further includes the following steps: S3021. After constructing the Bayesian network model of human factors, use the Bayesian reasoning method to reversely infer the probability changes of other unobserved factors based on human factors; S3022. Analyze the influence of each node in the causal chain through the results of Bayesian reasoning and determine the weight of the key node; S3023. Use actual accident data to verify the inference results. If it is found that the inference results do not match the actual situation, recheck the data input, Bayesian network model structure and conditional probability table, and make corresponding adjustments and optimizations.

9. The method for analyzing human factors of an aircraft running off the runway according to claim 7, characterized in that: The step S303 further includes the following steps: S3031. Apply forward reasoning, starting from known nodes, using the topological structure and conditional probability table of the Bayesian network to predict the probability of an aircraft running off the runway; S3032. Apply backward reasoning, set the aircraft running off the runway event as a known state, reversely infer the state probability changes of other nodes, and identify the factors that have the greatest impact on the aircraft running off the runway event, namely, the key causal factors; S3033. Combine the results of forward and backward reasoning and use sensitivity analysis to identify key parameters and obtain the key cause chain that led to the aircraft running off the runway; S3034. Display the key cause chain in the form of a chart or list, showing the cause-effect relationship and influencing factors in each chain, and providing intuitive analysis results for the aviation safety management department.

10. A human factor analysis system for aircraft runway deviation events, characterized in that: include: A construction module is used to combine the accident tree analysis method and the improved HFACS model to build a human factor investigation decision support model HFIDSM to identify and obtain the explicit and implicit factors that lead to aircraft running off the runway; The causal factors mining module is used to mine the causal factors of aircraft runway excursions through the human factors investigation decision support model HFIDSM; The key cause chain construction module is used to construct a Bayesian network model of the human factors that cause the aircraft to runway deviation, quantify the contribution of the causal factors to the event, and construct the key cause chain that leads to the aircraft running off the runway. The optimization module is used to analyze the critical cause chain and optimize the aircraft operating procedures based on the analysis results.

Images