A flight whole-chain operation situation risk assessment method

Abstract

The present application relates to the technical field of civil aviation operation risk management and control, and particularly relates to a flight full-chain operation situation risk assessment method, comprising the following steps: constructing a full-chain risk event library, determining a trigger time, calculating a peak time, a standard deviation and an amplitude of a Gaussian pulse function, and constructing a Gaussian pulse energy representation model of the full-chain risk event; calculating a link coupling weight and a risk amplification coefficient of a full-chain function resonance analysis-cascading model coupling network; calculating a probability of a risk module entering an abnormal state at the next moment according to the risk amplification coefficient, selecting parameters through Monte Carlo simulation, calculating a total risk energy according to the parameters and the amplified amplitude, identifying a link control priority, and outputting a risk level of each operation stage. The present application realizes accurate representation of full-chain risk events, quantitative calculation of cross-stage coupling effects and accurate identification of high-risk links, and improves the pertinence and effectiveness of flight full-process risk management and control.

Description

Technical Field

[0001] The present invention relates to the technical field of civil aviation operation risk control, and particularly to a method for risk assessment of the full-chain operation situation of flights. Background Art

[0002] Flight operation is a typical multi-subsystem dynamic coupling complex system. Its safe and efficient operation is not only restricted by the inherent complexity of the system, but is also vulnerable to the superposition of multiple factors such as human operation deviation, equipment mechanical state degradation, extreme meteorological condition disturbance, and management process omission. A minor deviation in any link may be gradually amplified through the strong correlation between systems and finally evolve into an operation risk or even a safety incident. Therefore, the identification and assessment of flight operation risks mainly rely on the professional level of operation control responsible personnel and the fuzzy evaluation given by experts. The staff still stays in two fixed mindsets of "relying on regulations and not crossing the red line" and "relying on experience to handle and judge", lacking the ability to quantitatively describe the risk evolution process and dynamically predict, resulting in difficulty in timely capturing potential risks, accurately assessing the risk impact degree, and taking targeted improvement measures.

[0003] Although relevant scholars in the industry have carried out a large number of studies, the existing results generally have the following limitations:

[0004] Insufficient root cause analysis: Mostly statistically counting the occurrence frequency of risk events from a macroscopic level, without deeply exploring the underlying mechanism of risk formation and the non-linear resonance mechanism between functions;

[0005] Lack of dynamic description: Lack of refined modeling of the time distribution characteristics and energy transfer process of risk events, and it is difficult to reflect the dynamic evolution law of risks with the flight operation time sequence;

[0006] Inaccurate quantification of micro-coupling: No functional cascade network covering the entire operation chain has been established, and there is a lack of a quantitative calculation method for the risk coupling transfer effect across stages and subsystems;

[0007] Narrow coverage: Mostly focused on a single operation stage or local subsystems, and the full-chain risk assessment from ground preparation, takeoff, cruise, approach to landing guarantee has not been achieved.

[0008] The above problems together lead to the fact that the existing technology is difficult to meet the actual needs of flight operation for full-chain, dynamic, and quantitative risk assessment, and cannot provide operation control personnel with accurate basis for identifying high-risk links and hierarchical control, restricting the further improvement of flight operation safety guarantee level. Summary of the Invention

[0009] This invention aims to address at least one of the technical problems existing in related technologies. To this end, this invention provides a method for risk assessment of the entire flight operation chain, enabling accurate characterization of risk events across the entire chain, quantitative calculation of cross-stage coupling effects, and precise identification of high-risk links, thereby improving the pertinence and effectiveness of risk management throughout the entire flight process.

[0010] This invention provides a method for risk assessment of the entire flight operation chain, including:

[0011] S1: Construct a risk event database covering the entire flight supply chain;

[0012] S2: Determine the trigger time based on each risk event in the full-chain risk event database, construct a kernel density estimation function based on the trigger time, calculate the peak time based on the kernel density estimation function, calculate the standard deviation of the Gaussian impulse function using the numerical integration method, quantify the variability of the functional output using the fuzzy comprehensive evaluation method, obtain the amplitude, and construct a Gaussian impulse energy characterization model for the full-chain risk events based on the trigger time, peak time, standard deviation of the Gaussian impulse function, and amplitude.

[0013] S3: Constructing a full-chain functional resonance analysis-cascade model coupling network;

[0014] S4: Calculate the link coupling weights of the cascaded model coupled network in the full-chain functional resonance analysis, and calculate the risk amplification factor based on the link coupling weights;

[0015] S5: Calculate the amplitude after coupling amplification based on the risk amplification factor, and calculate the probability of the risk module entering an abnormal state at the next moment based on the amplitude after coupling amplification.

[0016] S6: Select the parameters of the Gaussian pulse energy characterization model for the whole chain of risk events through Monte Carlo simulation, and calculate the total risk energy based on the parameters and the amplitude after coupling and amplification.

[0017] S7: Identify links with different control priorities based on the probability of the risk module entering an abnormal state at the next moment, the link coupling weight, and the total risk energy, and output the risk level of each operation stage.

[0018] Furthermore, step S1 includes:

[0019] S11: Based on the actual flight operation process, the entire flight chain is divided into the takeoff preparation phase, takeoff phase, cruise phase, descent phase, approach phase, landing phase, and ground support phase;

[0020] S12: Extract all potential risk events from the entire flight incident report chain through a full scan of event elements;

[0021] S13: Compare and categorize the extracted potential risk events by reorganizing the event logic structure;

[0022] S14: By focusing on the impact intensity of the categorized risk events and eliminating events with low frequency of occurrence, weak impact, and overlapping concepts, multiple types of key risk events are selected to form a full-chain risk event database.

[0023] Furthermore, step S2 includes:

[0024] S21: Determine the trigger time based on each risk event in the full-chain risk event database;

[0025] S22: Construct a kernel density estimation function based on the trigger time, and calculate the peak time based on the kernel density estimation function;

[0026] S23: Calculate the standard deviation of the Gaussian impulse function using numerical integration;

[0027] S24: The fuzzy comprehensive evaluation method is used to quantify the variability of the functional output and obtain the amplitude of the Gaussian pulse function;

[0028] S25: Construct a Gaussian pulse energy characterization model for the entire chain of risk events based on the peak time, the standard deviation of the Gaussian pulse function, and the amplitude of the Gaussian pulse function.

[0029] Furthermore, the full-chain functional resonance analysis-cascade model coupled network includes the lower layer FRAM The model and the upper-level cascaded-recovery model;

[0030] lower level FRAM The model uses each type of key risk event in the full-chain risk event database as... FRAM Functional nodes establish connections between them based on the temporal logic and physical relationships of flight operations.

[0031] The upper-level cascade-recovery model constructs a serial cascade model based on the flight operation phases.

[0032] Furthermore, the calculation of link coupling weights includes:

[0033] S411: Construct time-series weight vectors, physical weight vectors, and information weight vectors using the analytic hierarchy process;

[0034] S412: Set a sample time point, calculate the overlap rate of the time distribution of upstream and downstream risk events based on the probability density of the upstream probability distribution and the probability density of the downstream probability distribution corresponding to the sample time point; calculate the peak time difference based on the peak time of the upstream risk event and the peak time of the downstream risk event.

[0035] S413: Multiply the peak time difference and the temporal distribution overlap rate to obtain the temporal conduction intensity;

[0036] S414: Calculate the buffer capacity based on the ratio of the peak time of the upstream risk event to the peak time of the downstream risk event, and multiply the buffer capacity by the connection stiffness between the upstream and downstream to obtain the physical transmission strength;

[0037] S415: Multiply the accuracy and criticality of information transmitted between upstream and downstream to obtain the information transmission strength;

[0038] S416: Determine the link's overall coupling weight based on the temporal transmission strength, physical transmission strength, information transmission strength, temporal weight vector, physical weight vector, and information weight vector.

[0039] Furthermore, the risk amplification factor is calculated based on the link coupling weight, including:

[0040] S421: Calculate the importance of the upstream node's functional modules based on the link's in-degree and out-degree;

[0041] S422: Calculate the risk amplification coefficient based on the importance of upstream node functional modules, the variability of functional output, and the comprehensive coupling weight of the link.

[0042] Furthermore, step S5 includes:

[0043] S51: Calculate the basic anomaly probability of the module based on the variability of the functional output;

[0044] S52: Set feature judgment indicators based on the risk items within the risk module, and calculate the degree of risk occurrence based on the feature judgment indicators;

[0045] S53: Based on expert scoring, conduct a comprehensive assessment of the environment, personnel, and equipment to obtain the coupling sensitivity coefficient;

[0046] S54: Determine the steepness based on historical data and coupling sensitivity coefficient, and determine the offset based on the risk energy that the risk module can withstand;

[0047] S55: Based on the module's basic anomaly probability, risk occurrence level, coupling sensitivity coefficient, steepness, and offset, based on state transition... Logistic The probabilistic decision function calculates the probability that the risk module will enter an abnormal state in the next moment.

[0048] Furthermore, step S6 includes:

[0049] S61: Select various parameters of the Gaussian pulse energy characterization model for the entire chain of risk events through Monte Carlo simulation;

[0050] S62: Calculate the total energy after amplification at the end of any stage at any time.

[0051] S63: Calculate the background energy transferred to the next stage based on the total energy after amplification at the end of the stage, the transfer efficiency, and the system's active recovery efficiency;

[0052] S64: Define the integral of the square of the Gaussian impulse function over time as the risk impulse energy. Quantify the change process of the system's risk energy by accumulating the energy to obtain the energy of each risk event after upstream amplification.

[0053] S65: Add the energy of each risk event amplified upstream to the background energy to obtain the total risk energy.

[0054] Furthermore, step S7 includes:

[0055] S71: Standardize the link coupling weights by using percentile-box plots to obtain standardized coupling weights;

[0056] S72: Standardize the total risk energy by using a percentile-box plot to obtain the standardized risk energy;

[0057] S73: Multiply the standardized coupling weights and the standardized risk energy to obtain the fusion strength;

[0058] S74: Multiply the fusion strength by the probability that the risk module will enter an abnormal state in the next moment to obtain a comprehensive risk assessment index;

[0059] S75: Set a first risk level determination threshold and a second risk level determination threshold; if the comprehensive risk assessment index is less than the first risk level determination threshold, it is determined to be a low-risk link.

[0060] If the comprehensive risk assessment index is greater than or equal to the first risk level threshold and less than the second threshold, it is determined to be a medium-risk link.

[0061] If the comprehensive risk assessment index is greater than or equal to the second threshold, it is determined to be a high-risk link.

[0062] Furthermore, Gaussian impulse functions include normal distribution functions, uniform distribution functions, bimodal distribution functions, gamma distribution functions, and Poisson distribution functions.

[0063] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects:

[0064] This invention expands the assessment scope to the seven core stages of the entire flight chain, constructing a full-chain risk event library containing multiple types of key risk events. This effectively avoids the limitations of single-stage assessments, comprehensively capturing the cross-stage risk transmission effects and achieving risk coverage across the entire flight operation process. Based on the Gaussian impulse model, it quantifies the temporal evolution characteristics of risk energy, combined with Monte Carlo simulation to capture the dynamic accumulation process of risks throughout the entire chain, clearly presenting the complete dynamic process of risk formation, evolution, and accumulation, achieving a breakthrough from static qualitative to dynamic quantitative analysis. Furthermore, it integrates temporal, physical, and informational three-dimensional coupling weights with the analytic hierarchy process (AHP) to systematically quantify cross-stage and cross-module risks. The resonance amplification effect and the three-dimensional coupling weights, after consistency verification, meet the reliability requirements, ensuring the scientific and rational nature of the coupling quantification results and solving the problem of the single dimension of coupling analysis in existing technologies. Through standardized processing and threshold determination, the invention accurately identifies high-risk links and risk levels at each stage of the entire chain, clarifies the control priorities of different links and the risk status at each stage, and provides clear targeting basis for risk control. This significantly improves the efficiency and accuracy of risk control. The invention achieves accurate characterization of risk events across the entire chain, quantitative calculation of cross-stage coupling effects, and accurate identification of high-risk links, thereby enhancing the pertinence and effectiveness of risk control throughout the entire flight process.

[0065] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0067] Figure 1 This is a flowchart illustrating a method for assessing the operational status and risks of a flight across the entire supply chain, as provided by this invention. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but cannot be used to limit the scope of this invention.

[0069] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0070] The following is combined Figure 1 This invention describes a method for assessing the operational status and risks of a flight across the entire supply chain.

[0071] like Figure 1 As shown, a method for assessing the operational status and risks of a flight across its entire value chain includes:

[0072] S1: Construct a risk event database covering the entire flight supply chain;

[0073] S11: Based on the actual flight operation process, the entire flight chain is divided into the takeoff preparation phase, takeoff phase, cruise phase, descent phase, approach phase, landing phase, and ground support phase;

[0074] Based on the actual flight operation process, the entire chain is divided into seven core stages: takeoff preparation, takeoff, cruise, descent, approach, landing, and ground support, to achieve complete coverage of the flight from the ground to the air and then to landing, ensuring that no operational link is missed.

[0075] S12: Extract all potential risk events from accident reports through a full scan of event elements; construct a database of key risk events for each stage using systemic risk cognition theory;

[0076] In some specific embodiments of the present invention, 200 aviation accident investigation reports covering the entire chain of operations are collected to ensure that the data comprehensively covers the three-dimensional risk sources of "human-machine-domain" and provides sufficient data support for the screening of key risk events. Based on 120 aviation operation event process record reports and other materials, the system risk cognition theory is adopted to first extract all potential risk events from the accident reports through a full scan of event elements, clarifying the key factors such as the time, premise, and environment of their occurrence, and ensuring the authenticity and completeness of the event description. Potential risk events include the crew not obtaining the latest real-time weather conditions and errors in fuel calculation.

[0077] S13: Compare and categorize the extracted potential risk events by reorganizing the event logic structure;

[0078] Cluster the dispersed initial risk scenarios according to their attributes and internal logic;

[0079] Based on the coding paradigm of "condition, phenomenon, action, result", concepts are integrated into a more explanatory framework, such as classifying crew stress level, crew fatigue and sleep management, crew mental health, and crew stress disorder into the crew status category.

[0080] S14: By focusing on the impact intensity of the categorized risk events and eliminating events with low frequency of occurrence, weak impact, and overlapping concepts, multiple types of key risk events are selected to form a full-chain risk event database.

[0081] S2: Determine the trigger time based on each risk event in the full-chain risk event database, construct a kernel density estimation function based on the trigger time, calculate the peak time based on the kernel density estimation function, calculate the standard deviation of the Gaussian impulse function using the numerical integration method, quantify the variability of the functional output using the fuzzy comprehensive evaluation method, obtain the amplitude, and construct a Gaussian impulse energy characterization model for the full-chain risk events based on the trigger time, peak time, standard deviation of the Gaussian impulse function, and amplitude.

[0082] S21: Determine the trigger time based on each risk event in the full-chain risk event database;

[0083] The trigger time of risk events can be divided into the following two cases: Category events: For risk events with a clearly defined time interval, they are defined according to the length of that time interval. Class Standard Deviation This reflects the dispersion of the occurrence time of risk events; Type B events: For risk events that only record the instantaneous trigger moment, the standard deviation is preset to a very small fixed value based on the actual data in the incident report. For example, in the approach phase, 0.5 min is selected as the Type B standard deviation based on historical data. This is to effectively capture the characteristics of multiple risk events overlapping in a short period of time.

[0084] for For risk events of this type, because the events have a duration, the timing of the events will fluctuate appropriately around the mean based on the superposition of risk energies. Therefore, the peak time may be... The trigger time is determined in the middle. This represents the midpoint of the duration of the risk event.

[0085] For Class B risk events, because the risk energy accumulates rapidly within a very short time once a momentary triggering event occurs, making this situation irreversible, the peak time is only... The trigger time is determined in the middle. This instant is chosen to reflect the irreversible nature of real-world events.

[0086] S22: Construct a kernel density estimation function based on the trigger time, and calculate the peak time based on the kernel density estimation function;

[0087] S221: Calculate the bandwidth based on the sample standard deviation and sample size. The calculation expression is as follows:

[0088]

[0089] in, For bandwidth, For the sample size, The standard deviation is the sample standard deviation.

[0090] S222: Calculate the Gaussian kernel function value for each time sample based on the bandwidth and time samples. Summate the Gaussian kernel function values ​​for each time sample and normalize them to obtain the probability density function. The calculation expression is as follows:

[0091]

[0092]

[0093] in, For the first One time sample, for Gaussian kernel function value, for The probability density function at time t;

[0094] To standardize the units and perform more accurate calculations of the probability density function, relative distance is used. To standardize the distance, Set it as a normalization constant to ensure that the final area of ​​the desired image is 1.

[0095] S223: The peak time is defined as the moment corresponding to the maximum value of the probability density function. The calculation expression is as follows:

[0096]

[0097] in, Peak time, To determine the time point at which the function reaches its maximum value, For the minimum moment, The maximum moment.

[0098] This method can accurately pinpoint the moment when a risk event is most likely to occur.

[0099] The time data of risk events are accurately calculated using the kernel density estimation method described above, and each distribution function is accurately fitted according to the required probability density to obtain smooth probability density waveforms of five types of Gaussian impulse functions.

[0100] Gaussian impulse functions include the normal distribution function, uniform distribution function, bimodal distribution function, gamma distribution function, and Poisson distribution function.

[0101] Among them, risk events with concentrated occurrence probability and small fluctuation range are suitable for normal distribution; risk events with clear occurrence time intervals and uniform probability distribution are suitable for uniform distribution; risk events with two obvious high-incidence periods are suitable for bimodal distribution; risk events with obvious long-tail characteristics of occurrence time (low probability long-term distribution) are suitable for gamma distribution; and within a fixed time and space, Poisson distribution is suitable for studying the occurrence frequency of rare risk events.

[0102] Risk events that satisfy the properties of a normal distribution:

[0103] Normal distribution model As shown below:

[0104]

[0105]

[0106]

[0107] in, Let be the probability density function. This is the average time.

[0108] Risk events that satisfy the uniform distribution characteristic:

[0109] Uniform distribution The model is shown below:

[0110]

[0111]

[0112]

[0113] in, At the start time, This is the end time.

[0114] Risk events that satisfy the bimodal distribution characteristic:

[0115] Bimodal distribution model As shown below:

[0116]

[0117]

[0118]

[0119]

[0120]

[0121]

[0122]

[0123] in, Let be the probability density function of the first peak. Let be the probability density function of the second peak. The time of the highest probability peak for the first peak. This represents the time when the second peak is most likely to trigger. The first peak trigger time is around The degree of dispersion, The trigger time for the second peak is around The degree of dispersion, The average time of the first peak. The average time for the second peak, The time sample taken for the first peak (can be any value within the range). The time sample taken for the first peak (can be any value within the range). The total number of samples in the first peak. This represents the total number of samples for the second peak, since both peaks are random variables. The first peak weight, As the second peak weight, set .

[0124] Risk events that satisfy the gamma distribution property:

[0125] Gamma distribution model As shown below:

[0126]

[0127]

[0128]

[0129]

[0130]

[0131]

[0132] in, For shape parameters, is a scaling parameter that represents the variability of the response time. This is a gamma function.

[0133] Risk events that satisfy the Poisson distribution property

[0134] The Poisson distribution model is shown below:

[0135]

[0136]

[0137] in, This represents the average number of events occurring within a given interval. This represents the number of risk events that occurred during the flight's operation at this stage. This represents the total number of flights observed during this phase.

[0138] S23: Calculate the standard deviation of the Gaussian impulse function using numerical integration. The calculation expression is as follows:

[0139]

[0140] in, Let be the standard deviation of the Gaussian impulse function. Step size, Values ​​should be taken based on different time scales of risk. For example, when a risk occurs instantaneously, values ​​should generally be taken on a second-by-second basis.

[0141] S24: The fuzzy comprehensive evaluation method is used to quantify the variability of the functional output and obtain the amplitude of the Gaussian pulse function;

[0142] S241: Determine the time dimension Precision dimension Stage Dimension Functional variability metrics, and based on the time dimension Precision dimension Stage Dimension The function variability index is used to calculate the function output variability, and the calculation expression is:

[0143]

[0144] in, The quantification components of the functional output variability index are shown in Table 1.

[0145] Table 1 Functional variability indicators and corresponding quantitative components

[0146]

[0147] S242: The fuzzy comprehensive evaluation method is used to calculate the comprehensive quantitative value of three-dimensional functional variability;

[0148] Experts from multiple fields were invited to score the data. To ensure a suitable range and interval for the scores, a probability evaluation value was set for the occurrence of different functional changes within a specific module. , These five evaluation values ​​correspond to very high, relatively high, average, relatively low, and very low, respectively. A fuzzy evaluation matrix is ​​constructed. For ease of calculation, the number of experts selected for each type of review role is the same.

[0149]

[0150] in, The number of expert review roles is categorized, with each role consisting of multiple experts in a particular field. To evaluate the number of levels, For the first The character is rated as number one in the category. The proportion of people at each level By the The character is rated as number one in the category. The number of people in each grade divided by the number of... The total number of people in each role category.

[0151] Set the decision vector for each expert review as follows: , ,in For the first The decision of the experts This reflects the varying degrees of importance of different review roles.

[0152] The expert fuzzy comprehensive evaluation result is obtained based on the fuzzy evaluation matrix and the decision vectors of each expert reviewer. The calculation expression is as follows:

[0153]

[0154] in, The result is the expert fuzzy comprehensive evaluation.

[0155] The expression for calculating the probability of functional variability is:

[0156]

[0157]

[0158] in, for Functional variability probability, The definition of functional variability is optional. This is the transpose of the matrix. For the first The results of the fuzzy comprehensive evaluation by experts at the provincial level For the first Level probability evaluation value, m This represents the number of functional variability.

[0159] Normalized functional variability probability The calculation expression is:

[0160]

[0161]

[0162] in, For the first Functional variability, for The probability of functional variability;

[0163] available The normalized functional variability probability.

[0164] Calculated comprehensive quantization value of three-dimensional functional variability for:

[0165]

[0166] in, for The normalized functional variability probability, for The normalized functional variability probability, for The normalized functional variability probability.

[0167] S243: Calculate the Gaussian impulse function amplitude based on the combined quantization value of functional output variability and three-dimensional functional variability. The calculation expression is as follows:

[0168]

[0169] in, The amplitude of the Gaussian impulse function is given.

[0170] To achieve accurate characterization of the impact intensity of risk events.

[0171] S25: Construct a Gaussian pulse energy characterization model for the entire chain of risk events based on the peak time, the standard deviation of the Gaussian pulse function, and the amplitude of the Gaussian pulse function.

[0172] S3: Constructing a full-chain functional resonance analysis-cascade model coupling network;

[0173] Full-chain functional resonance analysis - cascaded model coupling network including lower-level functional resonance analysis ( FRAM ) model and upper-level cascaded-recovery model;

[0174] lower level FRAM The model uses each type of key risk event in the full-chain risk event database as... FRAM Functional nodes, such as real-time weather information, crew coordination, airspace intrusion, etc., are input based on the six nodes included in each risk event. I ): The thing that starts or triggers this function; output ( O ): The result or state change produced after this function is executed; prerequisite ( P ): Conditions that must be met before the function can begin execution; Resources (R): Physical or intangible assets consumed or used during the execution of the function; Control (C): Things that monitor, regulate or guide the function during its execution; Time (T): Time constraints, timing, rhythm or sequence requirements related to the execution of the function; Label the interface characteristics of each node, clarify the functional attributes, data requirements and output results of the node, and lay the foundation for building the relationship between nodes;

[0175] Next, based on the temporal logic and physical correlation of flight operations, connections between nodes are established. Among them, risk events within the same operational phase are organized into intra-phase links according to their occurrence order and logical dependencies; ultimately forming a functional resonance coupling network within a single phase, clearly presenting the transmission path and correlation logic of risks within the phase.

[0176] The upper-level cascade-recovery model constructs a serial cascade model based on the relevant stages of flight operation, and arranges the preparation stage, takeoff stage, cruise stage, descent stage, approach stage, landing stage and ground support stage in sequence as key events according to the flight operation process.

[0177] S4: Calculate the link coupling weights of the cascaded model coupled network in the full-chain functional resonance analysis, and calculate the risk amplification factor based on the link coupling weights;

[0178] The calculation of link coupling weight includes:

[0179] S411: Construct time-series weight vectors, physical weight vectors, and information weight vectors using the analytic hierarchy process;

[0180] Establish the highest-level decision-making objectives to accurately assess the degree of coupling amplification between upstream and downstream modules. The evaluation criteria for the middle layer include temporal coupling, physical coupling, and information coupling. The bottom-level scheme assigns coupling strength in three dimensions—temporal, physical, and information—to each pair of upstream and downstream modules.

[0181] An importance judgment matrix was constructed by expert scoring, and the importance of the three dimensions was compared pairwise using the 1-9 scale.

[0182] The importance judgment matrix is ​​normalized column-wise by the sum-product method, and the average value is calculated row-wise to obtain the time-series weight vector, physical weight vector, and information weight vector.

[0183] The values ​​of the time-series weight vector, physical weight vector, and information weight vector meet the normalization requirements to ensure that the influence of each dimension on the coupling effect is reasonably reflected.

[0184] The temporal weight vector, physical weight vector, and information weight vector are tested (whether the sum of the weight vectors is approximately equal to 1).

[0185] The weight vectors for the three dimensions are calculated based on the weight vectors. The calculation expression is as follows:

[0186]

[0187] in, This is an importance judgment matrix. For the weight vector, This is the new weight vector;

[0188] pass Obtain the temporal components of the new weight vector Physical components of the new weight vector New weight vector information components ;

[0189] Calculate the maximum eigenvalues ​​of the new weight vector in the temporal, physical, and information dimensions, respectively. ;

[0190]

[0191] in, hour, for , hour, for , hour for .

[0192] To more accurately estimate, the average of the three largest eigenvalues ​​is calculated;

[0193] Calculate the consistency index :

[0194]

[0195] The random consistency index was obtained by looking up the table. RI;

[0196] Calculate the consistency ratio :

[0197]

[0198] Perform a consistency check on the obtained data. If... CR If the value is ≤0.1, then the selected data is reasonable.

[0199] S412: Set a sample time point, calculate the overlap rate of the time distribution of upstream and downstream risk events based on the probability density of the upstream probability distribution and the probability density of the downstream probability distribution corresponding to the sample time point; calculate the peak time difference based on the peak time of the upstream risk event and the peak time of the downstream risk event.

[0200] The formula for calculating the overlap rate of the time distribution of upstream and downstream risk events is as follows:

[0201]

[0202] in, The overlap rate of the time distribution of upstream and downstream risk events. Let be the probability density of the upstream probability distribution. Let be the probability density of the downstream probability distribution. To find the minimum value function;

[0203] The formula for calculating the peak time difference is:

[0204]

[0205] in, The peak time difference, This represents the peak time for upstream risk events. This represents the peak time for downstream risk events;

[0206] S413: Multiply the peak time difference and the temporal distribution overlap rate to obtain the temporal conduction intensity. The calculation expression is as follows:

[0207]

[0208] in, This represents the temporal conduction strength.

[0209] S414: Calculate the buffer capacity based on the ratio of the peak time of upstream risk events to the peak time of downstream risk events. Multiply the buffer capacity by the connection stiffness between upstream and downstream to obtain the physical conduction strength; the calculation expression is:

[0210]

[0211]

[0212] in, For buffering capacity, For the rigidity of the connection between upstream and downstream, The physical conductivity is denoted as σ0; the connection rigidity is shown in Table 2.

[0213] Table 2 Connectivity Stiffness

[0214]

[0215] S415: Accuracy of transmission between upstream and downstream and information criticality Multiply to obtain the information transmission strength; the calculation expression is:

[0216]

[0217] in, For information transmission strength; accuracy of transmission between upstream and downstream. As shown in Table 3, the criticality of information As shown in Table 4.

[0218] Table 3. Accuracy of transmission between upstream and downstream

[0219]

[0220] Table 4 Information Criticality

[0221]

[0222] S416: The link-wide coupling weight is determined based on the temporal conduction strength, physical conduction strength, information conduction strength, temporal weight vector, physical weight vector, and information weight vector. The calculation expression is as follows:

[0223]

[0224] in, For the comprehensive coupling weight of upstream and downstream links, For time-series weight vectors, For physical weight vectors, This is the information weight vector.

[0225] , , Each level includes three grades: weak, medium, and strong, with each grade corresponding to a preset quantitative value for the intensity of the influence.

[0226] The risk amplification factor is calculated based on the link coupling weight, including:

[0227] S421: Calculate the importance of the upstream node's functional modules based on the link's in-degree and out-degree;

[0228] The expression for calculating the importance of downstream node functional modules is as follows:

[0229]

[0230] in, The importance of downstream node functional modules, For downstream node link in-degree, For the downstream node link out-degree, For the upstream node link in-degree, For the out-degree of the upstream node link, The total degree of all nodes in the network is the maximum value, which reflects the importance of each risk event in the entire chain.

[0231] The expression for calculating the importance of upstream node functional modules is:

[0232]

[0233] in, The importance of upstream node functional modules.

[0234] S422: Calculate the risk amplification coefficient based on the importance of upstream node functional modules, the variability of functional output, and the comprehensive coupling weight of the link. The calculation expression is as follows:

[0235]

[0236] in, This is a risk amplification factor for downstream nodes. For downstream nodes The set of all upstream nodes, upstream node Importance of functional modules upstream node Functional output variability, The weight is used for the comprehensive coupling of upstream and downstream links.

[0237] The risk amplification coefficient of downstream nodes can quantify the amplification effect of upstream risk events on downstream risk events through the coupling link.

[0238] S5: Calculate the amplitude after coupling amplification based on the risk amplification factor, and calculate the probability of the risk module entering an abnormal state at the next moment based on the amplitude after coupling amplification.

[0239] To avoid excessive amplification of risk amplitude leading to assessment distortion, an upper limit coefficient for amplitude amplification is set. Amplified amplitude after coupling The calculation expression is:

[0240]

[0241] in, The amplitude of the Gaussian impulse function. For downstream nodes and upstream nodes The risk amplification factor, To find the minimum value function;

[0242] Using an upper limit constraint mechanism, the system's elastic boundary is typically set between 10% and 20%. Therefore, the maximum amplified amplitude is set to 20%, and... k =1.2 is the preset upper limit coefficient for amplitude amplification to ensure the rationality of the quantization results.

[0243] S51: Calculate the basic anomaly probability of the module based on the variability of functional output. The calculation expression is as follows:

[0244]

[0245] in, This represents the basic anomaly probability of the module. These are probability coefficients. The larger the module, the more unstable it is, and the greater the risk of an accident. Based on data on accidents caused by various risk modules over the years, and from the perspective of aviation reliability, we take... .

[0246] S52: Set feature judgment indicators based on the risk items within the risk module, and calculate the degree of risk occurrence based on the feature judgment indicators;

[0247] By setting internal risk items for each module Corresponding characteristic judgment indicators, such as thunderstorm weather in the meteorological module, include: lightning location data, atmospheric instability index, water vapor content, view obstruction, weather radar imagery, and risk assessment when a thunderstorm occurs. Perform calculations. The initial setting is 0. If the actual situation meets any of the feature judgment indicators, then... Increase As the number of satisfied items increases, the level of risk increases accordingly. ;

[0248] S53: Based on expert scoring, a comprehensive assessment of the environment, personnel, and equipment is conducted to obtain the coupling sensitivity coefficient, calculated as follows:

[0249]

[0250] in, The coupling sensitivity coefficient, R 1 represents the environment. R 2 refers to personnel. R 3 refers to equipment;

[0251] This reflects the system's overall tendency to amplify risk propagation, dividing the overall system into... Three subsystems, based on scores from different experts (Favorable), 1.0 (Normal) (Unfavorable) Three types of interval range values.

[0252] S54: Determine the steepness based on historical data and coupling sensitivity coefficient, and determine the offset based on the risk energy that the risk module can withstand;

[0253] Based on historical data and coupling sensitivity coefficients, the steepness is set as follows: Class A = 2.5, Class B = 2.0, Class C = 1.8, Class D = 1.5, Class E = 1.0;

[0254] The offset is determined based on the risk energy that the risk module can withstand. The offset is set as follows: Class A = 2.0, Class B = 1.5, Class C = 1.0, Class D = 0.5, Class E = 0.0.

[0255] S55: Based on the module's basic anomaly probability, risk occurrence level, coupling sensitivity coefficient, steepness, and offset, based on state transition... Logistic The probabilistic decision function calculates the probability that the risk module will enter an abnormal state in the next moment. The calculation expression is as follows:

[0256]

[0257]

[0258] in, for The probability that this element will enter an abnormal state at any given time. For module The set of modules whose upstream nodes affect it. upstream node exist The risk amplitude amplified by time coupling. η S-shaped function steepness, S-shaped function The offset, These are the input variables for the sigmoid function.

[0259] S6: Select the parameters of the Gaussian pulse energy characterization model for the whole chain of risk events through Monte Carlo simulation, and calculate the total risk energy based on the parameters and the amplitude after coupling and amplification.

[0260] S61: Select various parameters of the Gaussian pulse energy characterization model for the entire chain of risk events through Monte Carlo simulation;

[0261] The parameters of the Gaussian pulse energy characterization model for full-chain risk events include peak time, amplitude, standard deviation, and link coupling weight.

[0262] S62: Calculate the total energy after amplification at the end of any stage at any time. The calculation expression is:

[0263]

[0264] in, This represents the total energy after amplification at the end of the stage. Peak time, For the first The amplitude of each risk event.

[0265] S63: Calculate the background energy transferred to the next stage based on the total energy after amplification at the end of the stage, the transfer efficiency, and the system's active recovery efficiency;

[0266] Transmission efficiency can be obtained by combining historical data. With system active recovery efficiency , as well as The values ​​of are shown in Table 5.

[0267] Table 5 Transmission Efficiency With system active recovery efficiency

[0268]

[0269] Background energy transferred to the next stage The calculation expression is:

[0270]

[0271] in, The inherent risks of the current stage The values ​​of are shown in Table 6.

[0272] Table 6 The value of

[0273]

[0274] S64: The integral of the square of the Gaussian impulse function over time is defined as the risk impulse energy. The change process of the system's risk energy is quantified by energy accumulation to obtain the energy of each risk event after upstream amplification. The calculation expression is as follows:

[0275]

[0276] in, The energy amplified upstream for each risk event.

[0277] S65: Add the energy of each risk event after upstream amplification to the background energy to obtain the total risk energy. The calculation expression is as follows:

[0278]

[0279] in, Total energy of risk.

[0280] S7: Identify links with different control priorities based on the probability of the risk module entering an abnormal state in the next moment and the total risk energy, and output the risk level of each operation stage;

[0281] S71: The link coupling weight is standardized using a percentile-box plot to obtain the standardized coupling weight. The calculation expression is as follows:

[0282]

[0283] in, To standardize the coupling weights, For all W The 25th percentile of the sample, For all W The 75th percentile of the sample, W This represents the link coupling weight.

[0284] This represents the relative strength of the link's coupling strength within the entire sample. This indicates strong coupling at the tail.

[0285] S72: Standardize the total risk energy using a percentile-box plot to obtain the standardized risk energy. The calculation expression is as follows:

[0286]

[0287] in, Standardize the risk energy for upstream nodes. P 50 is all The median of P 90 is all The 90th percentile;

[0288] Based on the skewness of the energy distribution (the asymmetric distribution of the Gaussian impulse function), an appropriate quantile interval is selected for calculation, clearly presenting the impact of high-energy risk events.

[0289] Standardization using percentile-box plots eliminates the impact of dimensional differences. This means that there is high energy at the tail.

[0290] S73: Multiply the standardized coupling weights and the standardized risk energy to obtain the fusion strength. The calculation expression is as follows:

[0291]

[0292] in, For fusion strength.

[0293] S74: Multiply the fusion strength by the probability that the risk module will enter an abnormal state in the next moment to obtain the comprehensive risk assessment index. The calculation expression is as follows:

[0294]

[0295] in, For comprehensive risk assessment indicators;

[0296] S75: Set the threshold for determining the first risk level and the threshold for determining the second risk level If the comprehensive risk assessment index is less than the threshold for the first risk level, it is determined to be a low-risk link; this indicates that the link coupling strength and upstream energy are at a low level, and only routine monitoring is required without additional intervention.

[0297] If the comprehensive risk assessment index is greater than or equal to the first risk level threshold and less than the second threshold, it is determined to be a medium-risk link; at this point, it needs to be further subdivided into:

[0298] If the coupling strength is significant but the energy is low, then it is coupling monitoring, and the focus is on monitoring the link connection status;

[0299] If the upstream energy is significant but the coupling strength is low, then it is energy monitoring, with a focus on controlling the upstream risk events themselves;

[0300] If neither of these factors significantly exceeds the standard, it is classified as a normal medium-risk area, and regular periodic inspections can be strengthened.

[0301] If the comprehensive risk assessment index is greater than or equal to the second threshold, it is determined to be a high-risk link, indicating that the link is simultaneously in the strong coupling and high-energy tail region, with extremely strong risk transmission ability, and intervention measures need to be taken immediately.

[0302] Clearly define the urgency of risk control for each link; combine the cumulative risk energy value of each operation stage with the distribution of high-risk links, and output the risk level of each stage, such as low risk, medium risk, and high risk, to provide a clear priority basis for risk control across the entire chain.

[0303] This invention expands the assessment scope to the seven core stages of the entire flight chain, constructing a full-chain risk event library containing multiple types of key risk events. This effectively avoids the limitations of single-stage assessments, comprehensively capturing the cross-stage risk transmission effects and achieving risk coverage across the entire flight operation process. Based on the Gaussian impulse model, it quantifies the temporal evolution characteristics of risk energy, combined with Monte Carlo simulation to capture the dynamic accumulation process of risks throughout the entire chain, clearly presenting the complete dynamic process of risk formation, evolution, and accumulation, achieving a breakthrough from static qualitative to dynamic quantitative analysis. Furthermore, it integrates temporal, physical, and informational three-dimensional coupling weights with the analytic hierarchy process (AHP) to systematically quantify cross-stage and cross-module risks. The resonance amplification effect and the three-dimensional coupling weights, after consistency verification, meet the reliability requirements, ensuring the scientific and rational nature of the coupling quantification results and solving the problem of the single dimension of coupling analysis in existing technologies. Through standardized processing and threshold determination, the invention accurately identifies high-risk links and risk levels at each stage of the entire chain, clarifies the control priorities of different links and the risk status at each stage, and provides clear targeting basis for risk control. This significantly improves the efficiency and accuracy of risk control. The invention achieves accurate characterization of risk events across the entire chain, quantitative calculation of cross-stage coupling effects, and accurate identification of high-risk links, thereby enhancing the pertinence and effectiveness of risk control throughout the entire flight process.

[0304] Example:

[0305] S1: Construct a risk event database covering the entire flight supply chain;

[0306] Taking a certain airline's civil aircraft operating domestic cross-regional flights as the research object, this study covers the seven major operational stages of the entire chain, including takeoff preparation, takeoff, cruise, descent, approach, landing, and ground support. It takes two typical risk events as examples: "crew resource management capability" in the approach stage and "air traffic conflict" in the cruise stage.

[0307] S2: Determine the trigger time based on each risk event in the full-chain risk event database, construct a kernel density estimation function based on the trigger time, calculate the peak time based on the kernel density estimation function, calculate the standard deviation of the Gaussian impulse function using the numerical integration method, quantify the variability of the functional output using the fuzzy comprehensive evaluation method, obtain the amplitude, and construct a Gaussian impulse energy characterization model for the full-chain risk events based on the trigger time, peak time, standard deviation of the Gaussian impulse function, and amplitude.

[0308] Unit resource management capabilities:

[0309] Based on the clearly defined time interval of occurrence, the uniform probability distribution, and the setting of the start time as the initial approach phase to 0, a uniform distribution is established. min The end time is 15 minutes before the departure approach phase. min .

[0310] Five time points were selected as time samples. ( k =1, 2, 3, 4, 5), let , 7, , , ;

[0311] The solution obtained from the uniform distribution is... , .

[0312] This risk event is a risk event with a clearly defined time frame, and meets the criteria for an event... a The standard deviation is defined by its time interval length. The peak time calculated using kernel density estimation can be found in It has been determined.

[0313] Bandwidth is:

[0314]

[0315] Solving 3.433.

[0316] Calculate the probability density function:

[0317]

[0318] in ( k Given a sample of 1, 2, 3, 4, 5, a large number of samples need to be taken. Perform substitution calculations (take) (8 minutes for demonstration), relative distance. To standardize the distance,

[0319] when hour, , ;

[0320] when hour, 7, ;

[0321] when hour, , ;

[0322] when hour, , ;

[0323] when hour, , ;

[0324]

[0325]

[0326] The peak time is defined as the moment corresponding to the maximum value of the probability density function.

[0327]

[0328] right Take the derivative and set it to 0 to facilitate finding the extreme points:

[0329]

[0330] Finally obtained min.

[0331] Calculate the Gaussian impulse function :

[0332]

[0333] Solving .

[0334]

[0335] Solving .

[0336] Set the probability evaluation value for different functional changes of a certain module. These five evaluation values ​​correspond to very high, relatively high, average, relatively low, and very low, respectively, to construct a fuzzy evaluation matrix. There are three types of review roles: 1. Pilots, 2. Air Traffic Control, and 3. Aviation Safety Experts. Each role includes ten experts in the relevant field. The scores given by these experts are statistically analyzed.

[0337]

[0338] Set the decision vector for each expert review as follows: The expert fuzzy comprehensive evaluation results were obtained as follows:

[0339]

[0340] Solving for:

[0341] The quantified value of the probability of functional variability was calculated as follows:

[0342]

[0343] Solving for: .

[0344] Assume there are m types of variability (due to the time dimension) The options are: on time, too early, too late, or not at all; the precision dimension can be: precise, acceptable, or imprecise; the stage dimension can be: small, average, large, or huge; therefore, m can yield 48 possible scenarios.

[0345] The normalized probability is:

[0346]

[0347]

[0348] The calculation yielded the following results:

[0349]

[0350] After normalization, we get:

[0351]

[0352]

[0353]

[0354] The calculated comprehensive quantization value of the three-dimensional functional variability is:

[0355]

[0356] Solving .

[0357] Finally obtained This enables precise characterization of the impact intensity of risk events;

[0358]

[0359] Solve for amplitude .

[0360] Effectiveness of air traffic conflict early warning: Based on the long-tail characteristic of the occurrence time, it is adapted to gamma-ray observation; the cruise has multiple time ranges depending on the flight distance, taking the total time of the shortest route. .

[0361] Five time points were selected as time samples. ( k =1, 2, 3, 4, 5), let , , , , ;

[0362] According to the gamma distribution, we can solve for: , , , .

[0363] This risk event is a risk event with a clearly defined time frame, and meets the criteria for an event... a The standard deviation is defined by its time interval length. The peak time calculated using kernel density estimation can be found in It has been determined.

[0364] Bandwidth is:

[0365]

[0366] Solving 9.8 min.

[0367] Calculate the probability density function:

[0368]

[0369] in ( k Given a sample of 1, 2, 3, 4, 5, a large number of samples need to be taken. Perform substitution calculations (take) t 36 min (As a demonstration), take the relative distance To standardize the distance,

[0370] when hour, ,

[0371] when hour, ,

[0372] when hour, ,

[0373] when hour, ,

[0374] when hour, ,

[0375]

[0376]

[0377] The peak time is defined as the moment corresponding to the maximum value of the probability density function.

[0378]

[0379] right Take the derivative and set it to 0 to facilitate finding the extreme points:

[0380]

[0381] Finally obtained min.

[0382] Calculate the Gaussian impulse function :

[0383]

[0384] Solving .

[0385]

[0386] Solving .

[0387] Set the probability evaluation value for different functional changes of a certain module. These five evaluation values ​​correspond to very high, relatively high, average, relatively low, and very low, respectively, to construct a fuzzy evaluation matrix. There are three types of review roles: 1. Pilots, 2. Air Traffic Control, and 3. Aviation Safety Experts. Each role includes ten experts in the relevant field. The scores given by these experts are statistically analyzed.

[0388]

[0389] Set the decision vector for each expert review as follows: The expert fuzzy comprehensive evaluation results were obtained as follows:

[0390]

[0391] Solving for: .

[0392] The quantified value of the probability of functional variability was calculated as follows:

[0393]

[0394] Solving for: .

[0395] Assume there are m types of variability (due to the time dimension) The options are: on time, too early, too late, or not at all; the precision dimension can be: precise, acceptable, or imprecise; the stage dimension can be: small, average, large, or huge; therefore, m can yield 48 possible scenarios.

[0396] The normalized probability is:

[0397]

[0398]

[0399] The calculation yielded the following results:

[0400]

[0401] After normalization, we get:

[0402]

[0403]

[0404]

[0405]

[0406] The calculated comprehensive quantization value of the three-dimensional functional variability is:

[0407]

[0408] Solving .

[0409] Finally obtained A =4.1508.

[0410] S3: Constructing a full-chain functional resonance analysis-cascade model coupling network;

[0411] S4: Calculate the link coupling weights of the cascaded model coupled network in the full-chain functional resonance analysis, and calculate the risk amplification factor based on the link coupling weights;

[0412] All data are calculated based on the effectiveness of air traffic conflict early warning.

[0413] Using the Analytic Hierarchy Process (AHP), the first step is to establish the top-level decision-making objective (accurately assessing the degree of coupling amplification between upstream and downstream modules), the middle-level evaluation criteria (temporal coupling, physical coupling, and information coupling), and the bottom-level scheme (assigning coupling strength in three dimensions—temporal, physical, and information—to each pair of upstream and downstream modules). The second step involves constructing an importance judgment matrix using scores from 15 experts, and employing a 1-9 scale as the comparison metric to perform pairwise importance comparisons across the three dimensions, resulting in the importance judgment matrix shown in Table 7. For the temporal transmission strength dimension, For the physical conduction intensity dimension, This is the dimension of information transmission intensity.

[0414] Table 7 Importance Judgment Matrix

[0415]

[0416] The third step is to normalize the importance judgment matrix column by column using the sum-product method;

[0417] Find the sum of the columns: column = 1.833 Column = 3.5 Column = 6;

[0418] Normalize the data:

[0419] Column 1:

[0420] Line: 1 / 1.8333≈0.5455;

[0421] Line: 0.5 / 1.8333≈0.2727;

[0422] Line: 0.3333 / 1.8333 ≈ 0.1818;

[0423] Column 2:

[0424] Row: 2 / 3.5 ≈ 0.5714;

[0425] Line: 1 / 3.5 ≈ 0.2857;

[0426] Line: 0.5 / 3.5 ≈ 0.1429;

[0427] Column 3:

[0428] Line: 3 / 6 = 0.5000;

[0429] Line: 2 / 6 ≈ 0.3333;

[0430] Line: 1 / 6 ≈ 0.1667;

[0431] The normalized matrix is ​​shown in Table 8.

[0432] Table 8 Normalized Matrix

[0433]

[0434] Calculate the average by row:

[0435] The approximate value is 0.5390;

[0436] The approximate value is 0.2972;

[0437] The approximate value is 0.1638;

[0438] Right now:

[0439] ;

[0440] Conduct the inspection:

[0441] 0.5390 + 0.2972 + 0.1638 = 1

[0442] The weight vector is then:

[0443]

[0444]

[0445] = 1.6248;

[0446] = 0.8943;

[0447] = 0.4921;

[0448] The resulting three-dimensional vector , , Calculate the maximum eigenvalue of each of the three dimensions.

[0449] ≈ 3.0145

[0450] ≈ 3.0094

[0451] ≈ 3.0037

[0452] Averaging the three largest eigenvalues:

[0453]

[0454] Calculate the consistency index ;

[0455] The random consistency index was obtained by looking up the table. ;

[0456] Then calculate the consistency index ratio.

[0457] determination CR ≤0.1 is a reasonable value.

[0458] In terms of time, calculate the overlap rate of the time distribution of upstream and downstream risk events:

[0459] First, 20 sample time points were selected. The upstream probability distribution corresponding to the sample time points is shown in Table 9, and the downstream probability distribution is shown in Table 10.

[0460] Table 9. Upstream probability distribution corresponding to sample time points

[0461]

[0462] Table 10 Downstream probability distribution corresponding to sample time points

[0463]

[0464] Solving for:

[0465]

[0466] Take the peak time difference .

[0467]

[0468]

[0469] In terms of physical conduction strength, the connection rigidity between the upstream and downstream sides is selected, with a strength of strong. R =0.16;

[0470] Buffering capacity:

[0471]

[0472]

[0473]

[0474] Regarding information transmission strength, the accuracy M of information transmission between upstream and downstream is selected, with the strength being weak. Select key information N The intensity is medium. ;

[0475]

[0476]

[0477]

[0478] Calculated 0.0144.

[0479] With an in-degree of 2, an out-degree of 3, and a total degree of 5, and the maximum total degree of nodes in the network being 5, the calculated importance of the functional module is 1.0.

[0480] Risk amplification factor of downstream nodes:

[0481]

[0482] Solving It is 0.0646.

[0483] S5: Calculate the amplitude after coupling amplification based on the risk amplification factor, and calculate the probability of the risk module entering an abnormal state at the next moment based on the amplitude after coupling amplification.

[0484] set up k =1.2, calculate .

[0485] set up ,but =3.375× .

[0486] Flight operating system malfunction scenario: Cockpit display anomaly, aircraft controllability changes, other three characteristics normal, risk level. It is 0.4.

[0487] make R1 =0.8, R2 =1, R3 =

[0488] The coupling sensitivity coefficient was calculated. It is 0.5.

[0489] Take the steepness of the sigmoid function Ф η This is case A, version 2.0.S offset of type function Ф This is case 1.5, which falls under category B.

[0490] Based on state transition Logistic The probabilistic decision function calculates the probability that the risk module will enter an abnormal state in the next moment, and the result is:

[0491]

[0492] .

[0493] S6: Select the parameters of the Gaussian pulse energy characterization model for the whole chain of risk events through Monte Carlo simulation, and calculate the total risk energy based on the parameters and the amplitude after coupling and amplification.

[0494] Taking the descent phase to the approach phase, the calculation is as follows: =2.749.

[0495] Take the cruise phase to the descent phase and obtain , , δ =0.06.

[0496] Calculated =1.29705, =11.459.

[0497] S7: Identify links with different control priorities based on the probability of the risk module entering an abnormal state in the next moment and the total risk energy, and output the risk level of each operation stage;

[0498] W =0.0144, taken based on historical experience. P 25 = 3.2 P 75 = 4.6 P 50 = 3.5 P 90 = 4.5

[0499] Solving , ;

[0500]

[0501] Solving ;

[0502]

[0503] Solving for: .

[0504] Will Compared with preset thresholds, several priority links for "energy monitoring" are identified (such as "accuracy of takeoff performance data calculation → effectiveness of air traffic conflict early warning" and "crew resource management capability → accuracy of flight operation"). Combined with the cumulative risk energy value of each operational phase, the risk level of each phase is output, providing a clear basis for targeted risk management.

[0505] Differentiated control measures are proposed for the identified "immediate control" priority links and high-risk phases. For high-risk links that cross phases, a cross-departmental collaborative control mechanism is established, such as real-time data sharing between flight crews and air traffic control departments. For high-risk links within a phase, the operation process and procedures for that phase are optimized, such as strengthening special training on crew resource management during the approach phase. Based on the risk level of each phase, control resources are rationally allocated to achieve precise control of risks across the entire chain.

[0506] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assessing the operational status and risks of a flight across its entire supply chain, characterized in that, include: S1: Construct a risk event database covering the entire flight supply chain; S2: Determine the trigger time for each risk event in the full-chain risk event database, construct a kernel density estimation function based on the trigger time, calculate the peak time based on the kernel density estimation function, calculate the standard deviation of the Gaussian impulse function using the numerical integration method, quantify the variability of the functional output using the fuzzy comprehensive evaluation method, obtain the amplitude of the Gaussian impulse function, and construct a Gaussian impulse energy characterization model for the full-chain risk events based on the trigger time, peak time, standard deviation of the Gaussian impulse function, and amplitude of the Gaussian impulse function. S3: Constructing a full-chain functional resonance analysis-cascade model coupling network; Full-chain functional resonance analysis - cascade model coupling network including lower layer FRAM The model and the upper-level cascaded-recovery model; lower level FRAM The model uses each type of key risk event in the full-chain risk event database as... FRAM Functional nodes establish connections between them based on the temporal logic and physical relationships of flight operations. The upper-level cascade-recovery model constructs a serial cascade model based on the flight operation phases; S4: Calculate the link coupling weights of the cascaded model coupled network in the full-chain functional resonance analysis, and calculate the risk amplification factor based on the link coupling weights; The calculation of link coupling weight includes: S411: Construct time-series weight vectors, physical weight vectors, and information weight vectors using the analytic hierarchy process; S412: Set a sample time point, calculate the overlap rate of the time distribution of upstream and downstream risk events based on the probability density of the upstream probability distribution and the probability density of the downstream probability distribution corresponding to the sample time point; calculate the peak time difference based on the peak time of the upstream risk event and the peak time of the downstream risk event. S413: Multiply the peak time difference and the temporal distribution overlap rate to obtain the temporal conduction intensity; S414: Calculate the buffer capacity based on the ratio of the peak time of the upstream risk event to the peak time of the downstream risk event, and multiply the buffer capacity by the connection stiffness between the upstream and downstream to obtain the physical transmission strength; S415: Multiply the accuracy and criticality of information transmitted between upstream and downstream to obtain the information transmission strength; S416: Determine the overall coupling weight of the link based on the temporal transmission strength, physical transmission strength, information transmission strength, temporal weight vector, physical weight vector, and information weight vector; The risk amplification factor is calculated based on the link coupling weight, including: S421: Calculate the importance of the upstream node's functional modules based on the link's in-degree and out-degree; S422: Calculate the risk amplification coefficient based on the importance of upstream node functional modules, the variability of functional output, and the comprehensive coupling weight of the link; S5: Calculate the amplitude after coupling amplification based on the risk amplification factor, and calculate the probability of the risk module entering an abnormal state at the next moment based on the amplitude after coupling amplification. S6: Select the parameters of the Gaussian pulse energy characterization model for the whole chain of risk events through Monte Carlo simulation, and calculate the total risk energy based on the parameters and the amplitude after coupling and amplification. S7: Identify links with different control priorities based on the probability of the risk module entering an abnormal state at the next moment, the link coupling weight, and the total risk energy, and output the risk level of each operation stage.

2. The method for risk assessment of the entire flight operation status according to claim 1, characterized in that, Step S1 includes: S11: Based on the actual flight operation process, the entire flight chain is divided into the takeoff preparation phase, takeoff phase, cruise phase, descent phase, approach phase, landing phase, and ground support phase; S12: Extract all potential risk events from the entire flight incident report chain through a full scan of event elements; S13: Compare and categorize the extracted potential risk events by reorganizing the event logic structure; S14: By focusing on the impact intensity of the categorized risk events and eliminating events with low frequency of occurrence, weak impact, and overlapping concepts, multiple types of key risk events are selected to form a full-chain risk event database.

3. The method for risk assessment of the entire flight operation status according to claim 1, characterized in that, Step S2 includes: S21: Determine the trigger time based on each risk event in the full-chain risk event database; S22: Construct a kernel density estimation function based on the trigger time, and calculate the peak time based on the kernel density estimation function; S23: Calculate the standard deviation of the Gaussian impulse function using numerical integration; S24: The fuzzy comprehensive evaluation method is used to quantify the variability of the functional output and obtain the amplitude of the Gaussian pulse function; S25: Construct a Gaussian pulse energy characterization model for the entire chain of risk events based on the peak time, the standard deviation of the Gaussian pulse function, and the amplitude of the Gaussian pulse function.

4. The method for risk assessment of the entire flight operation status according to claim 1, characterized in that, The S5 steps include: S51: Calculate the basic anomaly probability of the module based on the variability of the functional output; S52: Set feature judgment indicators based on the risk items within the risk module, and calculate the degree of risk occurrence based on the feature judgment indicators; S53: Based on expert scoring, conduct a comprehensive assessment of the environment, personnel, and equipment to obtain the coupling sensitivity coefficient; S54: Determine the steepness based on historical data and coupling sensitivity coefficient, and determine the offset based on the risk energy that the risk module can withstand; S55: Based on the module's basic anomaly probability, risk occurrence level, coupling sensitivity coefficient, steepness, and offset, based on state transition... Logistic The probabilistic decision function calculates the probability that the risk module will enter an abnormal state in the next moment.

5. The method for risk assessment of the entire flight operation status according to claim 1, characterized in that, Step S6 includes: S61: Select various parameters of the Gaussian pulse energy characterization model for the entire chain of risk events through Monte Carlo simulation; S62: Calculate the total energy after amplification at the end of any stage at any time. S63: Calculate the background energy transferred to the next stage based on the total energy after amplification at the end of the stage, the transfer efficiency, and the system's active recovery efficiency; S64: Define the integral of the square of the Gaussian impulse function over time as the risk impulse energy. Quantify the change process of the system's risk energy by accumulating the energy to obtain the energy of each risk event after upstream amplification. S65: Add the energy of each risk event amplified upstream to the background energy to obtain the total risk energy.

6. The method for risk assessment of the entire flight operation status according to claim 1, characterized in that, Step S7 includes: S71: Standardize the link coupling weights by using percentile-box plots to obtain standardized coupling weights; S72: Standardize the total risk energy by using a percentile-box plot to obtain the standardized risk energy; S73: Multiply the standardized coupling weights and the standardized risk energy to obtain the fusion strength; S74: Multiply the fusion strength by the probability that the risk module will enter an abnormal state in the next moment to obtain a comprehensive risk assessment index; S75: Set a first risk level determination threshold and a second risk level determination threshold; if the comprehensive risk assessment index is less than the first risk level determination threshold, it is determined to be a low-risk link. If the comprehensive risk assessment index is greater than or equal to the first risk level threshold and less than the second threshold, it is determined to be a medium-risk link. If the comprehensive risk assessment index is greater than or equal to the second threshold, it is determined to be a high-risk link.

7. The method for risk assessment of the entire flight operation status according to claim 1, characterized in that, Gaussian impulse functions include the normal distribution function, uniform distribution function, bimodal distribution function, gamma distribution function, and Poisson distribution function.

Images