Method and system for optimizing adjustment of an aircraft ice accretion SFIP index

Abstract

The application provides an aircraft icing SFIP index optimization adjustment method and system, the index optimization adjustment method comprises the following steps: obtaining airline flight record data; extracting icing cases and non-icing cases based on prior probability; using ERA5 reanalysis data or model meteorological product data to calculate SFIP index values and values of each membership function; matching the observation data and the reanalysis data; calculating and selecting evaluation indexes; adjusting the SFIP index weight coefficient and optimizing the SFIP index algorithm in combination with regional influence factors. Different evaluation indexes are used as constraint conditions to adjust the weight coefficient of each membership function. For different regions, corresponding constraint conditions are selected according to the evaluation indexes concerned, so that the weight optimization scheme suitable for different regions is obtained.

Description

Technical Field

[0001] This invention relates to the field of flight safety, and in particular to a method and system for optimizing and adjusting the SFIP index of aircraft icing. Background Technology

[0002] With the development of China's aviation industry, the number of military and civilian air transport has increased significantly. Aircraft icing, a phenomenon that seriously threatens flight safety, has received considerable attention from both civil aviation and military departments. During flight, when an aircraft collides with supercooled liquid droplets, icing can form on components such as the propeller, wings, antennas, and windshield. Icing in different locations increases the aircraft's gravity and drag, reduces lift and thrust, and in severe cases, can lead to a stall. Therefore, it is crucial to determine and predict aircraft icing based on relevant icing index algorithms.

[0003] There is currently no weighting adjustment scheme for the SFIP index in China. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a method and system for optimizing and adjusting the SFIP index of aircraft icing to overcome or at least partially solve the above problems.

[0005] According to one aspect of the present invention, a method for optimizing and adjusting the SFIP index of aircraft icing is provided, the method comprising:

[0006] Obtain airline flight record data;

[0007] Extracting icing cases and non-icing cases based on prior probabilities;

[0008] Calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model weather product data;

[0009] Match the observation data with the reanalysis data;

[0010] Calculate and select evaluation indicators;

[0011] By incorporating regional influence factors, the weighting coefficients of the SFIP index are adjusted to optimize the SFIP index algorithm.

[0012] Optionally, obtaining airline flight record data specifically includes:

[0013] Based on the airline's flight record data, information is extracted from icing warnings;

[0014] By combining the location and time information corresponding to the warnings, abnormal warnings in the error dataset are filtered, and icing cases and non-icing cases are extracted and screened.

[0015] Summarize and analyze the characteristics of various events, such as spatial distribution, icing intensity, and icing type.

[0016] Optionally, the extraction of icing cases and non-icing cases based on prior probability specifically includes:

[0017] We used QAR data provided by the airline to remove outliers and collected information such as whether icing occurred, duration, and spatiotemporal location to obtain a full sample of icing events. To match the spatiotemporal resolution of the model data, we extracted icing cases from the full sample and extracted non-icing cases proportionally.

[0018] Optionally, the calculation of the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model meteorological product data specifically includes:

[0019] Calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or other types of model meteorological product data.

[0020] ERA5 reanalysis data is the fifth-generation atmospheric analysis data from the European Centre for Geosciences. It is necessary to extract four factors—temperature, humidity, vertical velocity, and liquid water content—from it to calculate the SFIP index and membership function.

[0021] Optionally, matching the observation data with the reanalysis data specifically includes:

[0022] The meteorological reanalysis data used are hourly data, and the QAR data are second-level data. In order to balance the objectivity and representativeness of the evaluation and avoid wasting computing power, an hourly matching scheme was designed.

[0023] By defining ice accumulation cases occurring within ten minutes before or after the hour as hourly ice accumulation cases, we first extract hourly ice accumulation cases.

[0024] To ensure that the occurrence of ice accumulation cases at the top of the hour and non-ice accumulation cases at the top of the hour have no bias on the results, the proportion of ice accumulation cases at the top of the hour to all ice accumulation cases is calculated, and an equal proportion of non-ice accumulation cases at the top of the hour are randomly selected based on this proportion.

[0025] For the hourly case, its center point is calculated, the coverage area of ​​the hourly case is calculated considering the aircraft flight speed, and the average SFIP index within a certain range of grid points around the center point is determined as the index value matching the case.

[0026] Optionally, the calculation and selection of evaluation indicators specifically includes:

[0027] Based on the original SFIP icing index, the index results of the corresponding points of the extracted cases are obtained, and the false alarm rate, false negative rate, and accuracy of the original SFIP index are calculated. The distribution characteristics are analyzed and summarized.

[0028] Optionally, the adjustment of the SFIP index weight coefficients by incorporating regional influence factors to optimize the SFIP index algorithm specifically includes:

[0029] The ice accumulation observation cases were binary quantized according to whether ice accumulation occurred, and a regression model was established based on the SFIP values ​​calculated at the case locations.

[0030] By introducing constraints such as false alarm rate, false alarm rate, and accuracy, and adjusting the weighting coefficients, an optimization and adjustment scheme for the aircraft icing SFIP index suitable for various regions of China is constructed.

[0031] This invention also provides an aircraft icing SFIP index optimization and adjustment system, comprising:

[0032] The icing case extraction module is used to acquire airline flight record data; it extracts icing cases and non-icing cases based on prior probability.

[0033] The SFIP icing index and membership function calculation module is used to calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model meteorological product data.

[0034] The case matching module is used to match observation data with reanalysis data;

[0035] The evaluation index calculation and selection module calculates and selects evaluation indicators.

[0036] The weight coefficient adjustment module adjusts the SFIP index weight coefficients by taking into account regional influence factors, thereby optimizing the SFIP index algorithm.

[0037] This invention provides a method for optimizing and adjusting the SFIP (Skimming Freedom of Inflation) index for aircraft icing. The method includes: acquiring airline flight record data; extracting icing and non-icing cases based on prior probabilities; calculating the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model meteorological product data; matching the observed data with the reanalysis data; calculating and selecting evaluation indicators; and adjusting the SFIP index weight coefficients based on regional influence factors to optimize the SFIP index algorithm. Different evaluation indicators are used as constraints to adjust the weight coefficients of each membership function. For different regions, corresponding constraints are selected based on the evaluation indicators of interest, thereby obtaining a weight optimization scheme suitable for different regions.

[0038] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 A flowchart of an aircraft icing SFIP index optimization and adjustment method provided in this embodiment of the invention;

[0041] Figure 2 A block diagram of an aircraft SFIP index optimization and adjustment system provided in this embodiment of the invention;

[0042] Figure 3 Flight record statistics chart provided for embodiments of the present invention;

[0043] Figure 4 This is a schematic diagram illustrating the duration of icing in a single icing event provided in an embodiment of the present invention.

[0044] Figure 5 This is a schematic diagram of the variation curve of the SFIP membership function relative to meteorological elements provided in an embodiment of the present invention;

[0045] Figure 6 This is a schematic diagram of the matching scheme between QAR observation cases and meteorological data calculation indices provided in the embodiments of the present invention;

[0046] Figure 7 This is a schematic diagram of the TSS distribution example relative to the values ​​of a and b provided in an embodiment of the present invention. Detailed Implementation

[0047] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0048] The terms "comprising" and "having," and any variations thereof, in the specification, embodiments, claims, and drawings of this invention are intended to cover non-exclusive inclusion, such as including a series of steps or units.

[0049] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0050] like Figure 1 As shown, a method for optimizing and adjusting the SFIP index of aircraft icing is provided, the method comprising:

[0051] Obtain airline flight record data;

[0052] Extracting icing cases and non-icing cases based on prior probabilities;

[0053] Calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model weather product data;

[0054] Match the observation data with the reanalysis data;

[0055] Calculate and select evaluation indicators; adjust the weight coefficients of the SFIP index in conjunction with regional impact factors, and optimize the SFIP index algorithm.

[0056] Extraction of Icing and Non-Icing Events in China: Based on the flight record data (QAR) of Chinese airlines, information on icing warnings is extracted. Combined with the location and time information corresponding to the warnings, abnormal warnings in the error dataset are filtered out. Icing and non-icing cases are extracted and screened. The characteristics of various events, such as spatial distribution, icing intensity, and icing type, are summarized and analyzed to provide a basis for the applicability analysis of the index in China.

[0057] Matching scheme between icing cases and SFIP index: The meteorological reanalysis data used are hourly data, and the QAR data are second-level data. To balance the objectivity and representativeness of the evaluation and avoid wasting computing power, an hourly matching scheme was designed. Icing cases occurring within ten minutes before or after the hour are defined as hourly icing cases. First, hourly icing cases are extracted. To ensure that hourly icing cases and non-iceing cases do not affect the results, the proportion of hourly icing cases among all icing cases is calculated. Based on this proportion, an equal proportion of hourly non-iceing cases are randomly selected. For hourly cases, the center point is calculated, and the coverage area of ​​the hourly case is calculated considering the aircraft flight speed. Combining the meteorological data resolution, the average SFIP index within a certain range of grid points around the center point is defined as the index value to be matched with the case.

[0058] Applicability assessment of the original SFIP icing index in China: Based on the original SFIP icing index, the index results of the corresponding points of the extracted cases are obtained, the false alarm rate, false negative rate, and accuracy of the original SFIP index are calculated, and the distribution characteristics are analyzed and summarized.

[0059] SFIP Index Membership Function Weight Optimization and Adjustment: Icing observation cases are binary quantized according to whether icing occurs. Based on the SFIP values ​​calculated at the case locations, a regression model is established. By introducing constraints such as false alarm rate, false miss rate, and accuracy, the weight coefficients are adjusted, and finally, an optimization and adjustment scheme for the aircraft icing SFIP index suitable for various regions of China is constructed.

[0060] like Figure 2 As shown, the ice accumulation case extraction module

[0061] We used QAR data provided by the airline to remove outliers and collected information such as whether icing occurred, duration, and spatiotemporal location to obtain a full sample of icing events. To match the spatiotemporal resolution of the model data, we extracted icing cases from the full sample and extracted non-icing cases proportionally.

[0062] SFIP Ice Accumulation Index and Membership Function Calculation Module

[0063] Calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or other types of model meteorological product data.

[0064] ERA5 reanalysis data is the fifth-generation atmospheric analysis data from the European Centre for Geographic Information Systems (ECDS). It is necessary to extract four factors—temperature, humidity, vertical velocity, and liquid water content—to calculate the SFIP index and membership function. Detailed formulas are described in the steps below.

[0065] Case matching module

[0066] There is a significant difference in spatiotemporal resolution between the observation case sites and the reanalysis data. Therefore, it is necessary to use an appropriate spatiotemporal interpolation scheme to interpolate the forecast results of the reanalysis data to the corresponding observation cases.

[0067] Commonly used spatiotemporal interpolation methods include linear interpolation and nearest neighbor interpolation. It is necessary to consider the characteristics of aircraft flight, the relationship between flight record resolution and reanalysis data resolution, and select the most representative SFIP index around the case point.

[0068] In the vertical direction, the SFIP values ​​calculated from the grid data are interpolated to the corresponding heights of the QAR observation cases using the bilinear interpolation method.

[0069] In the horizontal direction, assuming the aircraft's speed is 800 km / h during flight, the longest distance of a grid point during the level flight phase is approximately the spatial resolution of the model data * 110 km * 1.41. Taking a resolution of 0.25 degrees as an example, the longest distance is approximately 39 km. The flight time is 2.85 minutes. To correspond to the hourly case, the grid points within the corresponding range are expanded outward based on the aircraft's flight direction. The average value of the index calculation results of the surrounding 5*5 grid points is used to match the icing case to ensure reasonable coverage of the aircraft's flight range and that the meteorological information within the corresponding range can represent the meteorological characteristics during the aircraft's flight.

[0070] Commonly used spatiotemporal interpolation methods include linear interpolation and nearest neighbor interpolation. It is necessary to consider the characteristics of aircraft flight, the relationship between flight record resolution and reanalysis data resolution, and select the most representative SFIP index around the case point.

[0071] Evaluation index calculation and selection module

[0072] The mainstream international evaluation metrics include: PODY (probability of detection), representing the percentage of correct predictions; POFD (probability of false detection), representing the percentage of incorrect predictions; and TSS (true skill statistic), representing the overall evaluation result. The calculation formulas for each metric are as follows:

[0073] PODy = YY / (YY+NY)

[0074] POFD = YN / (YN+NN)

[0075] TSS = PODy - POFD

[0076] Where YY represents the number of cases where the forecast was correct when there was actual icing, NY represents the number of cases where the forecast was for no icing when there was actual icing, YN represents the number of cases where the forecast was for icing when there was no actual icing, and NN represents the number of cases where the forecast was for no icing when there was no actual icing.

[0077] For different regions, one or more indices are selected as constraints for adjusting the weighting coefficients, based on their respective priorities.

[0078] Weighting coefficient adjustment module

[0079] A binary regression relationship between SFIP values ​​and whether or not ice has accumulated was established. The correlation between each membership function and whether or not ice has accumulated was analyzed. Constraints were introduced, and the three weight coefficients were adjusted to obtain a weight coefficient scheme suitable for different regions and different priorities.

[0080] The specific steps for optimizing and adjusting the SFIP index of aircraft icing applicable to the Chinese region are as follows:

[0081] Step 1: Extracting Typical Cases Based on QAR Data

[0082] QAR stands for Quick Access Recorder. It is located at the nose of the aircraft and can collect hundreds of data points during the flight, providing an objective and comprehensive reflection of the aircraft's status and the pilot's actions throughout the entire flight.

[0083] The data contains four main categories of information: flight identification, time and location, meteorological elements, and icing information. Time and location include time, longitude, latitude, and altitude; meteorological elements include air pressure, wind speed, wind direction, total temperature, and calm water; and icing information includes two warnings: left-wing icing warning and right-wing icing warning.

[0084] like Figure 3 As shown, statistical analysis of three years of flight records shows that the flight records in 2019 were distributed relatively evenly across the months, which ensures that the selection of subsequent cases and the comparison with the SFIP index are not affected by seasonal differences. Therefore, QAR records from the corresponding year were selected for processing.

[0085] In terms of data processing, the first step is to remove outlier data, including data from outlier time points and data from outlier ranges. Then, statistics are collected on whether icing occurred at each recording point during the flight period, including the time and location, to obtain the duration, intensity, and other characteristic information of all icing events, thus obtaining the full sample of icing events. To match the temporal resolution of the model data, icing cases are extracted from the full sample. Considering the aircraft's flight speed and the spatial resolution of the data, icing cases within ten minutes before and after the hour are defined as hourly icing cases. To ensure that the impact of hourly icing cases and hourly non-icing cases on the results is unbiased, the proportion of hourly icing cases to all icing cases is calculated, and an equal proportion of hourly non-icing cases are randomly selected based on this proportion.

[0086] like Figure 4 The duration of icing in a single icing event is shown.

[0087] Step 2: SFIP Ice Accumulation Index and Membership Function Calculation Module

[0088] Using ERA5 reanalysis data from the same period, the SFIP index and membership function values ​​at each grid point were calculated. The formulas for calculating the SFIP index and membership functions are shown below:

[0089] SFIP=M T (aM RH +bMω +cM LWC )

[0090]

[0091] Where T1 = -28℃, T2 = -12℃, T3 = -1℃, and T4 = +1℃.

[0092]

[0093] Where RH1 = 60, RH2 = 95.

[0094]

[0095] Among them, ω1=-0.1m / s, ω2=0m / s, ω3=0.5m / s.

[0096]

[0097] like Figure 5 As shown, this is a schematic diagram of the variation curves of SFIP membership functions relative to meteorological elements.

[0098] Step 3: Matching Observational Cases with Model Calculation Results

[0099] Based on the spatial location information of the observed cases, and considering the aircraft's flight speed, the average value of the calculated results within a certain range around the location is selected as the diagnostic result of the model data. The specific approach is as follows:

[0100] In the vertical direction, the SFIP values ​​calculated from the grid data are interpolated to the corresponding heights of the QAR observation cases using the bilinear interpolation method.

[0101] In the horizontal direction, assuming the aircraft's speed is 800 km / h, the longest distance per grid point during level flight is approximately the spatial resolution of the model data * 110 km * 1.41. Taking a 0.25-degree resolution as an example, the longest distance is approximately 39 km, and the flight time is 2.85 minutes. Figure 6 As shown, in order to correspond to the hourly case, the grid points within the corresponding range are expanded outward based on the aircraft's flight direction. The average value of the index calculation results of the surrounding 5*5 grid points is used to match the icing case to ensure reasonable coverage of the aircraft's flight range and that the meteorological information within the corresponding range can represent the meteorological characteristics of the aircraft during flight.

[0102] The average value of these 25 grid point forecasts and the corresponding membership function values ​​will serve as the basis for subsequent evaluation index calculations and weight coefficient adjustments.

[0103] Step 4: Evaluation Indicator Calculation and Selection Module

[0104] Evaluation indicators are calculated based on the results of matching model calculations and observed icing conditions. The mainstream international evaluation indicators are:

[0105] PODy = YY / (YY+NY)

[0106] POFD = YN / (YN+NN)

[0107] TSS = PODy - POFD

[0108] Where YY represents the number of cases where the forecast was correct when there was actual icing, NY represents the number of cases where the forecast was for no icing when there was actual icing, YN represents the number of cases where the forecast was for icing when there was no actual icing, and NN represents the number of cases where the forecast was for no icing when there was no actual icing.

[0109] Taking the national average as an example, the calculation results of the three indices are as follows:

[0110] Table 1. Overall Evaluation Results of SFIP Index Based on ERA5 in 2019 Nationwide

[0111] PODy(%) POFD (%) TSS (%) SFIP 58.88 45.74 13.14

[0112] For the national region, focusing on overall accuracy, the TSS score is chosen as a constraint for subsequent weighting coefficient adjustments.

[0113] Step 5: Weighting Coefficient Adjustment Module

[0114] Based on the matched membership function values, an SFIP function based on weight coefficients a, b, and c is constructed.

[0115] The formula for calculating the SFIP value that matches the case is:

[0116]

[0117] The criteria for judging ice accumulation are as follows:

[0118] cond_icing:SFIP case >0

[0119] cond_no icing: SFIP case ≤0

[0120] Because a discriminant function exists, the optimal weight coefficients cannot be directly determined by differentiation. Initially, an interval of 0.1 is defined for adjusting each weight coefficient, with an adjustment range of 0 to 1. The weight coefficients are summed to 1. Based on the set values ​​of a and b, the corresponding c value can be obtained. After calculating the corresponding condition value, it is compared with the observed results to calculate the YY, NY, YN, and NN values. The TSS formula is then used to calculate the score. The TSS distribution sample relative to the values ​​of a and b is shown below. Figure 7 As shown.

[0121] The algorithm first identifies the highest-scoring point. Then, it calculates the gradient between this point and its surrounding grid points. A threshold is set; if all surrounding gradients exceed this threshold, the calculation range is narrowed to nine surrounding grid points. Within this range, the weight intervals of a and b are reduced, and the TSS value is further calculated. This process is repeated until the adjacent gradient values ​​of the point corresponding to the highest calculated TSS value are less than the threshold. The corresponding a and b values ​​at this point are the optimal thresholds. The adjusted weight coefficients are a = 0.395, b = 0.18, and c = 0.425.

[0122] Beneficial Effects: A method for extracting and sampling icing cases based on QAR data. QAR data is second-level data, while reanalysis data has an hourly time resolution, resulting in significant differences from QAR data. For comparison, icing cases within ten minutes before and after the hour were defined as hourly cases and extracted from them. Simultaneously, to ensure accuracy and unbiasedness in the analysis, other non-icing cases were sampled proportionally.

[0123] Matching method between icing cases and model calculation results. Considering the aircraft's flight speed and model data resolution, the average value of the calculation results within the flight range around the location and the corresponding membership function value were selected as the model data diagnostic results for matching with icing cases.

[0124] Weight coefficient adjustment schemes based on different evaluation index constraints. Different evaluation indices are used as constraints to adjust the weight coefficients of each membership function. For different regions, corresponding constraints are selected based on the evaluation indices they focus on, thus obtaining weight optimization schemes applicable to different regions.

[0125] To reduce the waste of computing resources while ensuring accuracy, the optimal weights are calculated by gradually narrowing the range and interval of weight adjustment by setting a gradient threshold.

[0126] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing and adjusting the SFIP index of aircraft icing, characterized in that, The index optimization and adjustment method includes: Obtain airline flight record data; Extracting icing cases and non-icing cases based on prior probabilities; Calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model weather product data; Matching observational data with reanalysis data includes: The meteorological reanalysis data used are hourly data, and the QAR data are second-level data. In order to balance the objectivity and representativeness of the evaluation and avoid wasting computing power, an hourly matching scheme was designed. By defining ice accumulation cases occurring within ten minutes before or after the hour as hourly ice accumulation cases, we first extract hourly ice accumulation cases. To ensure that the occurrence of ice accumulation cases at the top of the hour and non-ice accumulation cases at the top of the hour have no bias on the results, the proportion of ice accumulation cases at the top of the hour to all ice accumulation cases is calculated, and an equal proportion of non-ice accumulation cases at the top of the hour are randomly selected based on this proportion. For the hourly case, calculate its center point, consider the aircraft flight speed to calculate the coverage area of ​​the hourly case, and combine the meteorological data resolution to determine the average SFIP index within a certain range of grid points around the center point as the index value matching the case. Calculate and select evaluation indicators, including: Based on the original SFIP icing index, the index results of the corresponding points of the extracted cases are obtained, the false alarm rate, false negative rate and accuracy of the original SFIP index are calculated, and the distribution characteristics are analyzed and summarized. The SFIP index weighting coefficients are adjusted based on regional impact factors to optimize the SFIP index algorithm, including: The ice accumulation observation cases were binary quantized according to whether ice accumulation occurred, and a regression model was established based on the SFIP values ​​calculated at the case locations. By introducing constraints on false alarm rate, false alarm rate, and accuracy, and adjusting the weighting coefficients, an optimization and adjustment scheme for the aircraft icing SFIP index suitable for various regions of China is constructed.

2. The method of claim 1, wherein, The acquisition of airline flight record data specifically includes: Based on the airline's flight record data, information is extracted from icing warnings; By combining the location and time information corresponding to the warnings, abnormal warnings in the error dataset are filtered, and icing cases and non-icing cases are extracted and screened. The characteristics of various events were summarized and analyzed, including spatial distribution, icing intensity, and icing type.

3. The method of claim 1, wherein, The extraction of icing cases and non-icing cases based on prior probability specifically includes: We used QAR data provided by the airline to remove outliers and collected information on whether icing occurred, its duration, and its spatiotemporal location to obtain a full sample of icing events. To match the spatiotemporal resolution of the model data, we extracted icing cases from the full sample and extracted non-icing cases proportionally.

4. The method for optimizing and adjusting the SFIP index of aircraft icing according to claim 1, characterized in that, The calculation of the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model meteorological product data specifically includes: Calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or other types of model meteorological product data; ERA5 reanalysis data are fifth-generation atmospheric analysis data from the European Centre for Research on Atmospheric Sciences. Four factors need to be extracted from the data to calculate the SFIP index and membership function.

5. An aircraft icing SFIP index optimization and adjustment system, employing the aircraft icing SFIP index optimization and adjustment method according to any one of claims 1-4, characterized in that, The adjustment system includes: The icing case extraction module is used to acquire airline flight record data; it extracts icing cases and non-icing cases based on prior probability. The SFIP icing index and membership function calculation module is used to calculate the SFIP index value and the values ​​of each membership function using ERA5 reanalysis data or model meteorological product data. The case matching module is used to match observation data with reanalysis data; The evaluation index calculation and selection module calculates and selects evaluation indicators. The weight coefficient adjustment module adjusts the SFIP index weight coefficients by taking into account regional influence factors, thereby optimizing the SFIP index algorithm.

Images