CN-122022492-A - Flight full-chain operation situation risk assessment method

CN122022492ACN 122022492 ACN122022492 ACN 122022492ACN-122022492-A

Abstract

The invention relates to the technical field of civil aviation operation risk management and control, in particular to a flight full-chain operation situation risk assessment method which comprises the steps of constructing a full-chain risk event library, determining trigger time, calculating peak time, standard deviation and amplitude of a Gaussian pulse function, constructing a Gaussian pulse energy characterization model of a full-chain risk event, calculating link coupling weight and risk amplification coefficient of a full-chain functional resonance analysis-cascade model coupling network, calculating probability of a risk module entering an abnormal state at the next moment according to the risk amplification coefficient, selecting various parameters through Monte Carlo simulation, calculating total risk energy according to the various parameters and the amplitude after coupling amplification, identifying link management and control priority, and outputting risk grade of each operation stage. The invention realizes the accurate representation of the full-chain risk event, the quantitative calculation of the cross-stage coupling effect and the accurate identification of the high-risk link, and improves the pertinence and the effectiveness of the full-flow risk management and control of the flight.

Inventors

SHI TONGYU
WANG YANTAO
YUAN YUJIE
WANG SHOUBING
WANG RUIQIAN
OUYANG BOWEN
WANG YUKUN
WEI SHIYAO

Assignees

中国民航大学

Dates

Publication Date: 20260512
Application Date: 20260408

Claims (10)

1. The method for evaluating the risk of the full-chain operation situation of the flight is characterized by comprising the following steps of: s1, constructing a flight full-chain risk event library; S2, determining triggering time according to each risk event in a full-chain risk event library, constructing a kernel density estimation function according to the triggering time, calculating peak time according to the kernel density estimation function, calculating standard deviation of a Gaussian pulse function by using a numerical integration method, quantifying function output variability by using a fuzzy comprehensive evaluation method to obtain Gaussian pulse function amplitude, and constructing a Gaussian pulse energy characterization model of the full-chain risk event according to the triggering time, the peak time, the standard deviation of the Gaussian pulse function and the Gaussian pulse function amplitude; S3, constructing a full-chain functional resonance analysis-cascade model coupling network; S4, calculating link coupling weight of the full-chain functional resonance analysis-cascade model coupling network, and calculating a risk amplification coefficient according to the link coupling weight; s5, calculating the amplitude after coupling amplification according to the risk amplification coefficient, and calculating the probability of the risk module entering an abnormal state at the next moment according to the amplitude after coupling amplification; S6, selecting various parameters of a Gaussian pulse energy characterization model of a full-chain risk event through Monte Carlo simulation, and calculating total risk energy according to the various parameters and the amplitude after coupling amplification; And S7, identifying links with different management and control priorities according to the probability of the risk module entering the abnormal state, the link coupling weight and the total risk energy at the next moment, and outputting the risk level of each operation stage.
2. The method for evaluating the risk of a full chain operation situation of a flight according to claim 1, wherein the step S1 comprises: s11, dividing the whole chain of the flight into a take-off preparation stage, a take-off stage, a cruising stage, a descending stage, an approaching stage, a landing stage and a ground guarantee stage based on the actual flow of the flight operation; S12, extracting all potential risk events from a flight full-chain accident report through event element full-scan; s13, comparing, classifying and sorting the extracted potential risk event through event logic structure recombination; And S14, carrying out event influence intensity focusing on the classified risk events, removing the events with low occurrence frequency, weak influence degree and overlapped concepts, and screening to obtain multiple types of key risk events to form a full-chain risk event library.
3. A method for risk assessment of a full chain operational scenario for a flight as claimed in claim 1, wherein step S2 comprises: S21, determining triggering time according to each risk event in a full-chain risk event library; S22, constructing a kernel density estimation function according to the triggering time, and calculating peak time according to the kernel density estimation function; S23, calculating the standard deviation of the Gaussian pulse function by using a numerical integration method; s24, quantifying the function output variability by adopting a fuzzy comprehensive evaluation method to obtain Gaussian pulse function amplitude; s25, constructing a Gaussian pulse energy characterization model of the full-chain risk event according to the peak time, the standard deviation of the Gaussian pulse function and the amplitude of the Gaussian pulse function.
4. The method for evaluating the risk of a full chain operation situation of a flight according to claim 1, wherein the full chain functional resonance analysis-cascade model coupling network comprises a lower FRAM model and an upper cascade-recovery model; The lower FRAM model takes each type of key risk event in the full-chain risk event library as an FRAM functional node, and establishes a connection relation between the functional nodes based on the time sequence logic and physical association of flight operation; the upper cascade-recovery model builds a serial cascade model according to the flight operation stage.
5. The method for evaluating the risk of a full chain operation situation of a flight according to claim 1, wherein the calculating of the link coupling weight comprises: s411, constructing a time sequence weight vector, a physical weight vector and an information weight vector by a analytic hierarchy process; s412, setting a sample time point, and calculating the time distribution overlapping rate of the upstream and downstream risk events according to the probability density of the upstream probability distribution and the probability density of the downstream probability distribution corresponding to the sample time point; s413, multiplying the peak time difference by the time distribution overlapping rate to obtain time sequence conduction intensity; s414, calculating the buffer capacity according to the ratio of the peak time of the upstream risk event to the peak time of the downstream risk event, and multiplying the buffer capacity by the connection rigidity between the upstream and the downstream to obtain the physical conduction strength; s415, multiplying the accuracy of the upstream and downstream transmission by the information criticality to obtain the information transmission intensity; And S416, determining the comprehensive coupling weight of the link according to the time sequence conduction intensity, the physical conduction intensity, the information conduction intensity, the time sequence weight vector, the physical weight vector and the information weight vector.
6. A method for risk assessment of a full chain operational scenario for a flight as claimed in claim 1, wherein calculating risk factors based on link coupling weights comprises: S421, calculating the importance of the upstream node functional module according to the link input degree and the link output degree; And S422, calculating a risk amplification coefficient based on the importance degree of the upstream node functional module, the variability of the functional output and the comprehensive coupling weight of the link.
7. A method for risk assessment of a full chain operational scenario for a flight as claimed in claim 1, wherein step S5 comprises: s51, calculating basic abnormality probability of a module according to the function output variability; S52, setting a characteristic judgment index according to the risk item in the risk module, and calculating the risk occurrence degree according to the characteristic judgment index; s53, comprehensively evaluating the environment, personnel and equipment according to expert scoring conditions to obtain a coupling sensitivity coefficient; S54, determining abruptness according to historical data and coupling sensitivity coefficients, and determining offset according to the sustainable risk energy of the risk module; S55, calculating the probability of the risk module entering the abnormal state at the next moment based on the Logistic probability decision function of the state transition according to the module basic abnormal probability, the risk occurrence degree, the coupling sensitivity coefficient, the abruptness and the offset.
8. A method for risk assessment of a full chain operational scenario for a flight as claimed in claim 1, wherein step S6 comprises: s61, selecting various parameters of a Gaussian pulse energy characterization model of a full-chain risk event through Monte Carlo simulation; S62, calculating total energy after amplification at any time of any stage; S63, calculating the background energy transferred to the next stage according to the total energy amplified at the end of the stage, the transfer efficiency and the active recovery efficiency of the system; s64, defining the integral of the square of the Gaussian pulse function in time as risk pulse energy, and quantitatively representing the change process of the system risk energy through energy accumulation to obtain the energy of each risk event amplified upstream; and S65, adding the energy of each risk event amplified by the upstream and the background energy to obtain total risk energy.
9. The method for evaluating the risk of a full chain operation situation of a flight according to claim 1, wherein the step S7 comprises: s71, normalizing the link coupling weight through a percentile-box diagram to obtain a normalized coupling weight; s72, normalizing the total risk energy through a percentile-box diagram to obtain normalized risk energy; s73, multiplying the standardized coupling weight and the standardized risk energy to obtain fusion strength; S74, multiplying the fusion strength by the probability of the risk module entering an abnormal state at the next moment to obtain a comprehensive risk evaluation index; S75, setting a first risk level judgment threshold value and a second risk level judgment threshold value, and judging a low risk link if the comprehensive risk evaluation index is smaller than the first risk level judgment threshold value; If the comprehensive risk evaluation index is greater than or equal to the first risk level judgment threshold value and smaller than the second threshold value, judging that the risk link is in the risk link; And if the comprehensive risk evaluation index is larger than or equal to the second threshold value, judging that the link is a high risk link.
10. The flight full chain operational situation risk assessment method according to claim 1, wherein the gaussian pulse function comprises a normal distribution function, a uniform distribution function, a bimodal distribution function, a gamma distribution function and a poisson distribution function.

Description

Flight full-chain operation situation risk assessment method Technical Field The invention relates to the technical field of civil aviation operation risk management and control, in particular to a flight full-chain operation situation risk assessment method. Background The flight operation is a typical multi-subsystem dynamic coupling complex system, the safe and efficient operation of the system is limited by inherent complexity of the system, and the system is also easily influenced by superposition of multiple factors such as manual operation deviation, equipment mechanical state degradation, extreme weather condition disturbance, management flow omission and the like, the tiny deviation of any link can be amplified step by step through strong correlation among the systems, and finally the operation risk and even the safety event are evolved. Therefore, the identification and assessment of flight operation risk is mainly dependent on the business level of the operation control responsibility personnel and fuzzy evaluation given by the expert. Workers still stay in two solidification thinking of 'relying on regulation not to go beyond the red line' and 'relying on experience to treat and judge', and lack of quantitative characterization and dynamic prediction capabilities on the risk evolution process leads to difficulty in timely capturing potential risks, accurately evaluating the influence degree of risks and taking targeted improvement measures. Although a great deal of research has been conducted by related scholars in the industry, the following limitations are common to the existing achievements: the root cause analysis is insufficient, namely, the occurrence frequency of risk events is counted from a macroscopic level, and a nonlinear resonance mechanism between a bottom layer mechanism and functions formed by risks is not deeply excavated; The lack of dynamic depiction is that the time distribution characteristic of the risk event and the energy transfer process lack of fine modeling, so that the dynamic evolution rule of the risk along with the flight operation time sequence is difficult to reflect; microcosmic coupling quantification is inaccurate, namely a functional cascade network covering a full-running chain is not established, and a quantification calculation method is lacked for risk coupling transfer effects of cross-stage and cross-subsystem; The coverage range is narrow, and the multi-focus is in a single operation stage or a local subsystem, so that full-chain risk assessment from ground preparation, take-off, cruising and approach to landing guarantee cannot be realized. The problems commonly lead to the prior art that the actual requirements of flight operation on full-chain, dynamic and quantitative risk assessment are difficult to meet, and accurate high-risk link identification and grading management and control basis cannot be provided for operation control personnel, so that further improvement of the flight operation safety guarantee level is restricted. Disclosure of Invention The present invention is directed to solving at least one of the technical problems existing in the related art. Therefore, the invention provides a flight full-chain running situation risk assessment method, which realizes the accurate representation of full-chain risk events, the quantitative calculation of cross-stage coupling effect and the accurate identification of high-risk links, and improves the pertinence and the effectiveness of flight full-flow risk management and control. The invention provides a flight full-chain operation situation risk assessment method, which comprises the following steps: s1, constructing a flight full-chain risk event library; S2, determining triggering time according to each risk event in a full-chain risk event library, constructing a kernel density estimation function according to the triggering time, calculating peak time according to the kernel density estimation function, calculating standard deviation of a Gaussian pulse function by using a numerical integration method, quantifying function output variability by using a fuzzy comprehensive evaluation method to obtain amplitude, and constructing a Gaussian pulse energy characterization model of the full-chain risk event according to the triggering time, the peak time, the standard deviation of the Gaussian pulse function and the amplitude; S3, constructing a full-chain functional resonance analysis-cascade model coupling network; S4, calculating link coupling weight of the full-chain functional resonance analysis-cascade model coupling network, and calculating a risk amplification coefficient according to the link coupling weight; s5, calculating the amplitude after coupling amplification according to the risk amplification coefficient, and calculating the probability of the risk module entering an abnormal state at the next moment according to the amplitude after coupling amplification; S6, selecting various