Ice accumulation forecasting method and system, readable storage medium and program product

Abstract

The invention discloses an ice accumulation forecasting method and system, a readable storage medium and a program product, and the method comprises the steps: S1, selecting a specific region, obtaining meteorological data, and processing the meteorological data into a unified format, S2, constructing an initial ice accumulation potential algorithm of the specific region based on the meteorological data, and S3, constructing an initial ice accumulation potential algorithm of the specific region based on the meteorological data. S4, identifying a specific weather system and a key meteorological factor of the specific area, and establishing a membership function of the key meteorological factor under the specific weather system, and optimizing the initial icing potential algorithm in combination with the membership function to obtain a localized icing potential algorithm. And S5, outputting the icing potential based on the localized icing potential algorithm. Therefore, a localized icing potential algorithm can be established for a specific region, so that the icing risk can be predicted more accurately.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of civil aviation weather forecasting, and relates to an icing prediction method and system, a readable storage medium and a program product. BACKGROUND

[0002] It is known that aircraft icing (icing) is a serious flight safety hazard that can affect wing lift, increase drag, and deteriorate the aerodynamic performance of the aircraft, affecting the stability and maneuverability of the aircraft, and even causing engine failure. With the advancement of the domestic large aircraft project, the study of natural icing of aircraft has attracted more and more attention. However, the meteorological elements affecting aircraft icing are complex, each element has a direct impact on icing, and in addition, different elements also have mutual influence and indirectly affect icing.

[0003] The icing potential algorithm is to study the icing cloud microphysical process combined with macro meteorological data statistical analysis, to link the temperature, humidity, cloud cover, cloud water and other icing-related factors with the icing possibility, to make a membership curve, to find out the icing possibility and icing intensity under various conditions, and to give a calculation model of the aircraft icing possibility. At present, the optimal aircraft icing prediction algorithm usually uses fuzzy logic and decision tree method, and combines the icing formula obtained by statistical analysis of historical data to improve the accuracy and reliability of the prediction.

[0004] However, research shows that the climate characteristics and influencing factors are different in different regions, and the probability and regularity of icing occurrence are also different. If foreign algorithms are copied, the icing distribution in China cannot be fully reflected, and there are problems such as low hit rate and high false alarm rate. SUMMARY

[0005] Technical problems to be solved by the application

[0006] The application is formed to solve the above technical problems, and its purpose is to provide an icing prediction method and system, a readable storage medium and a program product, which can establish a localized icing potential algorithm for a specific region, thereby more accurately predicting icing risks.

[0007] Technical solutions adopted to solve the technical problems

[0008] The present application provides an ice accretion forecasting method, comprising: step S1: selecting a specific area, obtaining meteorological data, and processing the data into a unified format; step S2: constructing an initial ice accretion potential algorithm for the specific area based on the meteorological data; step S3: identifying a specific weather system and key meteorological factors for the specific area based on the meteorological data, and establishing a membership function for the key meteorological factors under the specific weather system; step S4: optimizing the initial ice accretion potential algorithm in combination with the membership function to obtain a localized ice accretion potential algorithm; and step S5: outputting ice accretion potential based on the localized ice accretion potential algorithm.

[0009] Preferably, in step S1, the meteorological data related to the specific area is obtained from pilot reports, reanalysis data, satellite data, aircraft meteorological data, civil aviation meteorological messages, radar data, and sounding data.

[0010] Preferably, in the step S1, processing the meteorological data into a unified format includes unifying the temporal resolution and unifying the spatial resolution.

[0011] Preferably, when unifying the time resolution, meteorological data of one hour before and after the icing time of a certain icing event in the specific area is selected and interpolated as shown in the following formula:

[0012]

[0013] Among them, X t is the meteorological data at the ice accumulation time t,

[0014] X t0 is the meteorological data of t0 one hour before the ice accumulation time t,

[0015] X t1 It is the meteorological data of the hour t1 after the ice accumulation time t.

[0016] Preferably, when unifying the spatial resolution, the meteorological data of the grid where the icing location of a certain icing event in the specific area is located is selected and interpolated as shown in the following formula:

[0017]

[0018] Where X is the meteorological data of the ice accumulation location,

[0019] d1, d2, d3, d4 are the distances between the ice accumulation location and the four vertices of the grid,

[0020] d5, d6, d7, d8 are the vertical distances between the ice accumulation location and the four sides of the grid,

[0021] is a weight.

[0022] Preferably, in the step S2, it comprises: selecting meteorological factors from the meteorological data processed into a unified format and constructing membership functions of the meteorological factors, and constructing the initial ice accretion potential algorithm based on the membership functions of the meteorological factors and corresponding weights.

[0023] Preferably, in the step S2, it comprises: classifying icing situations according to different meteorological conditions based on a decision tree, selecting different meteorological factors for each icing situation from the meteorological data processed into a unified format and constructing membership functions of the meteorological factors, and constructing different initial ice accretion potential algorithms based on the membership functions of the meteorological factors.

[0024] Preferably, the icing situations include a cloudless situation, a single-layer cloud situation, a convective cloud situation, a weak precipitation situation, and a multi-layer cloud situation.

[0025] Preferably, in the step S3, it comprises: determining the specific weather system that has a greater impact on ice accretion in a specific region and the key meteorological factor that has a greater impact on ice accretion under the specific weather system based on ice accretion events in the specific region, identifying the specific weather system based on the meteorological data processed into a unified format, and establishing membership functions of the key meteorological factor based on the meteorological data processed into a unified format and characteristic meteorological data of the specific weather system.

[0026] Preferably, when the specific region is the Northeast region, the specific weather system is a cold vortex and an upper trough, the key meteorological factor is vertical velocity, and the characteristic meteorological data is the height of the inversion layer and the wettest layer.

[0027] Preferably, in the step S4, the localized ice accretion potential algorithm is represented by the following formula:

[0028] CIP final = CIP0+ (1-CIP0) x m 关键气象因子

[0029] wherein CIP final is the localized ice accretion potential algorithm,

[0030] CIP0is the initial ice accretion potential algorithm,

[0031] m 关键气象因子 is a membership function of the key meteorological factor.

[0032] Therefore, since the aircraft icing is a nonlinear process, a small change in any meteorological element can lead to a sudden change in icing or not, and therefore the aircraft in-flight icing diagnosis method based on fuzzy logic proposed in the present application introduces the climate characteristics of a specific area on the basis of the CIP algorithm to optimize the icing potential algorithm and improve its local adaptability and prediction accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is a flowchart of the icing prediction method of an embodiment of the present application.

[0034] Figure 2 is a graph showing a cold vortex.

[0035] Figure 3 is a graph showing a high-level trough.

[0036] Figure 4 is a graph showing a membership function of vertical velocity.

[0037] Figure 5 is a graph showing the prediction results of the Ic index, CIP0, CIP final DETAILED DESCRIPTION

[0038] The present application will be further described below in conjunction with the following embodiments, which should be understood as merely illustrative of the present application, rather than limiting the present application. The same or corresponding reference numerals in the drawings represent the same components, and repeated descriptions are omitted. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without making creative efforts fall within the scope of protection of the present application.

[0039] An embodiment of the present application will be described below in conjunction with the drawings.

[0040] In step S1, a specific area is selected, meteorological data is obtained and processed into a unified format.

[0041] ​In this embodiment, after a specific region is selected, meteorological data related to the specific region can be obtained from, for example, pilot reports, reanalysis data, satellite data, aircraft meteorological data, civil aviation meteorological messages, radar data, sounding data, etc. Specifically, as pilot reports (PIREPs), for example, domestic civil aviation pilots can provide subjective reports of ice accumulation observations along the flight path. As reanalysis data (ERA5, NCEP-FNL, etc.), for example, it can come from meteorological agencies such as ECMWF, NCEP, etc., providing global meteorological background fields such as temperature and humidity, wind field, etc. As satellite data (MODIS, Himawari-8, etc.), for example, it can come from NASA, EUMETSAT, Chinese Fengyun series, etc., for monitoring cloud thickness, temperature, liquid water content, etc. As aircraft meteorological data (AMDAR), for example, it can come from airlines / meteorological agencies, providing high-altitude temperature and humidity, wind speed, etc. As civil aviation meteorological messages (METAR, TAF, SPECI), for example, it can come from airport automatic observation systems, for monitoring ground weather, predicting icing probability, etc. As radar data (weather radar, airborne radar), for example, it can come from meteorological departments, aircraft themselves, for monitoring precipitation type, cloud structure, etc. As sounding data (Radiosonde), for example, it can come from station data from meteorological stations, providing vertical temperature and humidity profiles, icing layer height, etc.

[0042] Based on the different data sources, different meteorological data related to a specific area can be obtained for ice accretion prediction. Specifically, for example, based on the (domestic civil aviation) pilot report, the weather time, location, ice accretion severity (light, medium, heavy), and ice accretion type (transparent ice, rime ice, mixed ice) during ice accretion can be identified, so that the temperature, humidity, cloud top temperature, etc. meteorological data of the ice accretion weather can be obtained as needed. Based on the reanalysis data, the temperature, relative humidity, atmospheric pressure, wind speed and direction, vertical speed, liquid water path (LWP), ice water content (IWC), liquid water content (LWC), and precipitation type can be identified, so that the temperature, humidity, wind speed, and liquid water content, etc. meteorological data can be obtained as needed. Based on satellite data, the cloud top temperature, cloud phase (ice cloud / water cloud), cloud optical thickness, liquid water path, ice water path (IWP), and cloud coverage can be identified, so that the cloud phase, liquid water content, cloud top temperature, etc. meteorological data can be obtained as needed. Based on aircraft meteorological data, the flight altitude, temperature, wind speed and direction, humidity, and air pressure can be identified, so that the upper air temperature and humidity, wind speed and direction, etc. meteorological data can be obtained as needed. Based on the civil aviation meteorological message, the ground temperature, dew point temperature, humidity, precipitation type (rain / snow / freezing rain), visibility, and wind speed and direction can be identified, so that the precipitation type, humidity, etc. meteorological data can be obtained as needed. Based on radar data, the radar reflectivity (precipitation intensity), cloud top height, echo intensity, precipitation type, and vertical wind field can be identified, so that the radar reflectivity (determination of liquid water in the cloud), precipitation type, etc. meteorological data can be obtained as needed. Based on the sounding data, the temperature profile, humidity profile, wind speed and direction, atmospheric stratification stability (CAPE / CIN), ice water content, and liquid water content can be identified, so that the upper air temperature and humidity profile, liquid water content, and stratification stability, etc. meteorological data can be obtained as needed.

[0043] In addition, it should be understood that depending on the different data sources, global data can be obtained first and then appropriate meteorological data related to a specific area can be selected, or data sources related to a specific area can be filtered out first and then meteorological data can be obtained. Essentially, there is no difference between these methods, only the difference in data processing amount, which does not affect the acquisition of meteorological data. It should also be understood that the above-mentioned means of obtaining meteorological data can use known technologies, such as code recognition based on a database or information extraction based on text, etc. In addition, the data sources for obtaining meteorological data and the means for obtaining meteorological data are not limited to the above.

[0044] Next, the obtained meteorological data is processed to have a unified format. The "unified format" referred to here provides consistent and accurate data basis for subsequent statistical analysis, at least including the unification of time interval and spatial grid (i.e. unified time and spatial resolution).

[0045] In this embodiment, for example, the icing time t and the icing location x can be determined based on icing events of an aircraft in a certain region, and interpolation is used to make data from different sources have uniform spatiotemporal resolution. The specific process is as follows.

[0046] First, the time resolution is unified. As an example of time interpolation, for example, data in one hour t0, t1 around the icing time t selected from the meteorological data obtained from reanalysis data (such as ERA5) can be interpolated.

[0047] Specifically, the data in one hour before and after the icing time t is selected, and the meteorological data is time-interpolated as shown in the following formula.

[0048]

[0049] wherein X t is the meteorological data (variable) at the icing time t, X t0 and X t1 are the meteorological data at one hour before t0 and one hour after t1 of the icing time t.

[0050] If the target time t is close to one hour before t0 of the icing time t (i.e., |t-t0|<10 minutes), the nearest known value X t is directly used: X t0 .

[0051] If the target time t is close to one hour after t1 of the icing time t (i.e., |t-t1|<10 minutes), the nearest known value X t is directly used: X t1 .

[0052] If the target time t is far from t0 and t1 (i.e., none of the above conditions is met), the inverse time difference weighted interpolation is used:

[0053]

[0054] wherein the weight indicates that the closer the time, the greater the weight and the stronger the influence.

[0055] In this way, the meteorological data from different data sources can be fused in the time scale. In addition, linear interpolation or cubic spline interpolation can also be used for time interpolation, as long as the time resolution is improved, without specific limitation.

[0056] Next, spatial resolution unification is performed. As an example of spatial interpolation, the following is specified. First, the grid in which the icing site x is located is determined. Here, different grids of different sizes are used according to different data sources. For example, if ERA5 is used as the data source, the grid resolution is 0.25°x0.25° (about 31 km x 31 km), and when the icing site x falls within a grid with a resolution of 0.25°x0.25°, the grid is selected as the basic grid for spatial interpolation. Of course, there can be special cases in which the icing site x falls exactly on the boundary of a grid with a resolution of 0.25°x0.25°, and two grids (with a resolution of 0.25°x0.5°) sharing the boundary can be selected as the basic grid for spatial interpolation.

[0057] After selecting the basic grid, the four vertices of the grid are set as vertices x1, x2, x3, and x4, the four edges of the grid are set as boundaries x1-x2, x2-x3, x3-x4, and x4-x1, the distances of the icing site x from the four vertices x1, x2, x3, and x4 are set as d1, d2, d3, and d4, and the vertical distances of the icing site x from the boundaries x1-x2, x2-x3, x3-x4, and x4-x1 are set as d5, d6, d7, and d8.

[0058] The required meteorological data is obtained from different data sources (such as ERA5, radar, satellite, ground civil aviation meteorological station, AMDAR, etc.) in the basic grid, and the spatial interpolation of the meteorological data is performed as shown in the following formula.

[0059]

[0060] where X is the meteorological data (variable) of the icing site x.

[0061] If at least one of the distances d1, d2, d3, and d4 of the icing site x to the four vertices x1, x2, x3, and x4 is less than 5 km (i.e., the minimum value of d1, d2, d3, and d4 < 5), the meteorological data X of the vertex corresponding to the minimum distance (i.e., the nearest vertex) is selected. i Direct assignment:

[0062] X = X i i = arg min(d1, d2, d3, d4)

[0063] where i = arg min(d1, d2, d3, d4) indicates the index of the minimum distance.

[0064] If d1, d2, d3, d4 are all greater than 5km, but at least one of the vertical distances d5, d6, d7, d8 from the icing site x to the four boundaries x1-x2, x2-x3, x3-x4, x4-x1 is less than 5km (i.e. the minimum of d5, d6, d7, d8 < 5), then the meteorological data X of the two vertices of the boundary corresponding to the minimum distance (i.e. the nearest boundary) is selected a and X b Inverse distance weighted interpolation is performed:

[0065]

[0066] wherein, is the weight defined according to the distance,

[0067] d a and d b is the distance from the icing site x to the two vertices of the nearest boundary.

[0068] If none of the above conditions are met, i.e. d1, d2, d3, d4, d5, d6, d7, d8 are all greater than 5km, then inverse distance weighted interpolation is performed using the meteorological data X1, X2, X3, X4 of the four vertices x1, x2, x3, x4:

[0069]

[0070] wherein, is the weight defined according to the distance,

[0071] d i is the distance from the icing site x to the four vertices, i = 1, 2, 3, 4.

[0072] By the spatial interpolation method as described above, the data can be smoothed while avoiding too large errors of the distant data points, so that the meteorological data of different sources are fused in the spatial scale. In addition, as described above, the spatial interpolation method is not specifically limited.

[0073] Therefore, the meteorological data with unified spatio-temporal resolution can be used as the meteorological data of a certain aircraft icing event in the specific region for subsequent statistical analysis. In addition, it should be understood that the judgment conditions listed in the process of unifying the spatio-temporal resolution (such as 10 minutes, 5km) are all examples, which can be set according to specific circumstances. In addition, it should be understood that the above-mentioned fusion of meteorological data of different sources by interpolation method is not limited to the above, and different methods can be used according to the characteristics of the data.

[0074] In step S2, an initial icing potential algorithm CIP0 of the specific region is constructed.

[0075] First, meteorological factors are selected from the unified meteorological data. Examples of these meteorological factors include temperature (T), relative humidity (RH), cloud top temperature (CTT), cloud liquid water content (CLW), vertical velocity (ω), cloud base height (CBH), radar echo intensity (dBZ), and cloud phase (Phase).

[0076] In this embodiment, a statistical analysis of icing events in a specific region can be performed to determine meteorological factors that have a significant impact on the probability of icing. Meteorological factors can also be appropriately selected based on the climatic characteristics of the specific region. For example, radar echo intensity and vertical velocity have a high similarity, and the processing speed of radar echo data is slow, so only vertical velocity can be selected. For example, cloud phase and liquid water in the cloud are both satellite inversion data with high similarity, and the liquid water content in the cloud and cloud top temperature can better represent the cloud condition. Therefore, cloud phase is not selected, and cloud top temperature and liquid water in the cloud are selected instead. For example, cloud base height cannot accurately represent the cloud condition in some scenarios and is therefore not selected. It should be understood that the selection criteria for meteorological factors are not limited and can be selected based on specific circumstances.

[0077] Then, the selected meteorological factors are statistically analyzed to establish the membership function of meteorological factors in a specific area.

[0078] For example, as an example, atmospheric temperature, relative humidity, cloud top temperature, vertical velocity of airflow, and liquid water content in the atmosphere are selected as meteorological factors, and fuzzy logic membership functions of atmospheric temperature, relative humidity, cloud top temperature, vertical velocity of airflow, and liquid water content in the atmosphere are established respectively. The details are as follows:

[0079] Atmospheric temperature fuzzy logic membership function M T :

[0080]

[0081] Relative humidity fuzzy logic membership function M RH :

[0082]

[0083] Fuzzy logic membership function M of airflow vertical velocity ω :

[0084]

[0085] Fuzzy logic membership function M of liquid water content in clouds CLW :

[0086]

[0087] Cloud top temperature fuzzy logic membership function MCTT :

[0088]

[0089] Then, based on the above membership functions and corresponding weight values, an initial ice accretion potential algorithm CIP0 is constructed.

[0090] CIP0 = w1 x M T + w2 x M RH + w3 x M ω + w4 x M CLW + w5 x M CTT

[0091] wherein wi represents the weight coefficient given by each membership function, which can be adjusted according to the weather characteristics of a specific region. In addition, it should be understood that the above membership functions can also be adjusted according to the climate characteristics of a specific region. It should also be understood that the above is to calculate the ice accretion potential by using the linear weighting method, but is not limited thereto.

[0092] For example, the present application can also classify the icing scenarios by combining the decision tree method, divide the atmosphere into five main icing scenarios of no cloud, single-layer cloud, convective cloud, weak precipitation, and multi-layer cloud according to different meteorological conditions, select different meteorological factors under each scenario, and use different ice accretion potential algorithms, thereby further improving the prediction accuracy.

[0093] Specifically, in the present embodiment, for example, the air column in the vertical direction of any location in a specific region can be obtained, and the following judgments can be made in sequence by the decision tree:

[0094] 1) When there is no height of RH≥85%, it is determined to be a no cloud scenario, the weather situation is weakly affected, only T and RH are selected as meteorological factors, and the initial ice accretion potential algorithm is:

[0095] CIP0 = M T x M RH

[0096] 2) When there is a height of RH≥85%, the height is taken as the cloud bottom, there is a continuous interval of RH>70% above the height and only one, it is determined to be a single-layer cloud scenario, the highest temperature at RH>70% is selected as CTT to indicate the influence of cloud top temperature on ice accretion in the cloud, and T, RH and CTT are selected as meteorological factors, and the initial ice accretion potential algorithm is:

[0097] CIP0 = M T x M RH x M CTT

[0098] 3) When there is a value of RH≥85% and more than 1 continuous RH>70% height layer exists upwardly from the value, it is determined to be a multi-layer cloud scenario, and the influence of cloud top temperature needs to be refined, so the temperature at the highest RH>70% in the layer is taken as CTT in the different cloud layers, T, RH and CTT are selected as the meteorological factors, and the initial ice accretion potential algorithm is:

[0099] CIP0=M T ×M RH ×M CTT

[0100] 4) When there is a height of RH≥85% and the minimum W is less than -0.2 Pa / s vertically upward, it is determined to be a convective cloud scenario, in order to describe the influence of cloud ice in the ascending motion on the aircraft ice accretion, the sum of cloud water and ice water content is taken as LWC, T, RH, CTT and LWC are selected as the meteorological factors, and the initial ice accretion potential algorithm is:

[0101] CIP0=M T ×M RH ×M CTT ×M CLW

[0102] 5) When there is a height of RH≥85% and the ground cumulative precipitation P is greater than 0 mm and less than 5 mm, it is determined to be a weak precipitation scenario, in order to describe the influence of large water droplets on the aircraft ice accretion, the sum of cloud water and rainwater content is taken as LWC, T, RH, CTT and LWC are selected as the meteorological factors, and the initial ice accretion potential algorithm is:

[0103] CIP0=M T ×M RH ×M CTT ×M CLW

[0104] Therefore, the embodiment is an optimization of the ice accretion potential algorithm (localization), and is not an optimization of a specific ice accretion potential algorithm. Therefore, a specific ice accretion potential algorithm can be selected according to actual needs and habits, and is not limited to the above enumeration.

[0105] In step S3, a specific weather system and key meteorological factors in a specific region are identified, and a membership function of the key meteorological factors under the specific weather system is established.

[0106] At present, it is generally recognized that there are five typical aircraft icing weather systems in China, which are high-level trough, cold vortex, shear line, stationary front and cold / warm front. According to the different geographical positions (specific regions) of the ice accretion, the main weather systems affecting the formation of the aircraft ice accretion are also different, so the identification of the weather system most affecting the icing in the specific region is needed.

[0107] In this embodiment, statistical analysis can be performed according to icing event data of a specific region, etc., to determine in advance which types of weather systems have a greater impact on icing weather in the specific region, and then set the weather systems with greater impact as specific weather systems. For example, the North China, Northeast China and Northwest China regions are greatly affected by upper troughs and cold vortices, and extensive icing is prone to occur in winter and spring. The East China and South China coastal regions are more affected by stationary fronts, cold fronts and warm fronts, and icing mainly occurs in frontal cloud systems in winter and spring. The Sichuan-Chongqing and Southwest mountainous regions are greatly affected by shear lines, with high humidity, and icing mainly occurs in local precipitation processes.

[0108] Next, the characteristics of different weather systems can be identified based on the obtained meteorological data. For example, the following is described by taking the ERA5 data as an example. For the upper trough weather system, in the 500 hPa potential height low value area, the isobars are wavy, the wind field has obvious convergence in front of the trough, the temperature gradient is large, and the humidity condition is suitable for icing. For the cold vortex weather system, a closed low pressure center is formed at 500 hPa potential height, the wind field rotates counterclockwise (in the Northern Hemisphere), and the conditions of low temperature, high humidity and strong upward motion are suitable for icing. For the shear line weather system, the wind direction shear is significant at 850 hPa or 700 hPa, the water vapor is sufficient, and the relative humidity is high. For the stationary front weather system, the wind speed is small and the wind direction difference is large at 850 hPa or 700 hPa, and the temperature gradient is strong, but the front moves slowly. For the cold front / warm front weather system, the temperature and wind direction change sharply at 850 hPa or 700 hPa, strong precipitation and icing usually occur before the cold front, and the temperature rises after the warm front, which may result in a wider icing range.

[0109] Further, in this embodiment, statistical analysis can also be performed according to icing event data of a specific region, etc., to determine in advance which meteorological data (elements) in the specific weather system have a greater impact on icing weather in the specific region, and then set the meteorological data (elements) with greater impact as key meteorological factors.

[0110] Next, the key meteorological factors can be determined for different weather systems based on the acquired meteorological data. For example, the following is described based on ERA5 data. Based on the icing event in a specific region and the corresponding ERA5 data, statistical analysis is performed by known methods to find the key meteorological factors most sensitive to icing. For example, in the North China region, under the cold vortex weather system, the key meteorological factors can be 500 hPa geopotential height, 700 hPa temperature, and 850 hPa relative humidity. In the Southwest region, under the shear line weather system, the key meteorological factors can be 700 hPa vertical velocity, 850 hPa wind speed, and 500 hPa relative humidity. In the East China region, under the stationary front weather system, the key meteorological factors can be 850 hPa temperature gradient, wind direction shear, and 700 hPa cloud water content. In the Northeast region, under the upper trough weather system, the key meteorological factors can be 500 hPa temperature, wind speed, and vertical velocity.

[0111] Based on the meteorological data processed into a unified format and the characteristic meteorological data of the specific weather system, the membership function of the key meteorological factor is established by statistical analysis or the like.

[0112] The membership obtained by fitting or probability density of a large amount of existing icing event statistical data, or using membership disclosed by others.

[0113] As the characteristic meteorological data of the specific weather system, for example, it can be data highly correlated with the key meteorological factor in the icing influence range.

[0114] In step S4, the initial icing potential algorithm is optimized in combination with the membership function of the key meteorological factor to obtain a localized icing potential algorithm CIP final . Specifically, in the present embodiment, the membership function m 关键气象因子 is combined with the initial icing potential algorithm CIP0 to perform localization correction to obtain the corrected icing potential algorithm CIP final :

[0115] CIP final = CIP0 + (1 - CIP0) x m 关键气象因子

[0116] In step S5, the icing potential is output based on the corrected icing potential algorithm CIP final .

[0117] Therefore, the aircraft in-flight icing diagnosis method based on fuzzy logic provided in the application determines the weather types that significantly affect aircraft icing in a specific region in combination with the climate characteristics of the specific region, optimizes the algorithm, improves the CIP algorithm by combining the key meteorological factors and their membership functions under the influence of the weather types, and obtains a modified icing potential algorithm, which is used to improve the locality and accuracy of the algorithm.

[0118] In addition, the application identifies and quantifies the weather types that significantly affect aircraft icing for a specific region, and introduces the key meteorological factors and their membership functions under the influence of these weather systems into the CIP algorithm, thereby optimizing the icing potential calculation. Through this optimization method, the local adaptability of the icing potential algorithm is improved, the applicability in different climate backgrounds is enhanced, and the regional adaptability is improved. In addition, by combining the icing weather characteristics of a specific region, the icing risk area can be more accurately identified, the icing potential calculation is more in line with the actual atmospheric conditions, and the reliability and local prediction ability of icing prediction are improved.

[0119] In addition, the application further combines fuzzy logic and decision tree methods to classify and identify different icing weather systems (such as no cloud, single-layer cloud, convective cloud, weak precipitation, and multi-layer cloud) according to meteorological principles. For different icing scenarios, icing potential calculation models are established respectively, and are optimized in combination with key meteorological factors to more finely calculate the icing probability caused by atmospheric conditions such as temperature, humidity, and airflow. Through this method, the scenario calculation ability of the icing potential algorithm is improved, and the accuracy of icing prediction is further improved.

[0120] In summary, the application optimizes the CIP algorithm by introducing local parameters in combination with weather types, realizes accurate optimization of icing potential calculation, improves regional adaptability, calculation accuracy, and weather system identification ability, thereby effectively improving the prediction accuracy of aircraft icing and reducing the risk of false positives and false negatives.

[0121] In the following, specific embodiments are exemplified to further illustrate and verify the accuracy of the modified icing potential algorithm CIP final of the application.

[0122]

Embodiment - Northeast Region

[0123] In this embodiment, the Northeast region is set as a specific region, and the local (optimized) icing potential algorithm CIP final of the Northeast region is calculated. However, it should be understood that the local optimization algorithm is not limited to the Northeast region exemplified in the embodiment, and can also be applied to regions such as the South China region, the Northwest region, and other regions with weather characteristics that can be classified and summarized.

[0124] Firstly, data of Northeast China is obtained and processed into a uniform format, and then an initial ice accretion potential algorithm CIP0 is constructed for Northeast China.

[0125] In this embodiment, CIP0 = mRH x mT x mCTT x mCLW, wherein,

[0126]

[0127] It should be understood that the ice accretion potential algorithm and the membership function in the above embodiment are a calculation example, and the ice accretion potential can not be calculated only by the above method.

[0128] Next, according to the statistical analysis of the ice accretion events in Northeast China, it is known that the upper trough and the cold vortex have a greater impact on the ice accretion weather, and it is also known that the vertical velocity is the key meteorological factor in these two types of weather systems. Therefore, it is necessary to identify these two types of weather systems, which are as follows.

[0129] The following will be described in combination with Figure 2 , 3 Figure 2 and Figure 3 In the above figures, the horizontal and vertical axes are longitude and latitude (based on the figures, the horizontal and vertical axes are 117.5°E-132.5°E and 37.5°N-55°N, respectively, which belongs to Northeast China), the ICE point represents the icing site (selecting a certain ice accretion event in Northeast China for analysis), and the three color bars below represent the geopotential height anomaly, the minimum 2% interval of pressure, and the curvature of the UV wind field, which are all represented by the general standards in the field.

[0130] In this embodiment, the geopotential height and wind field at 500 hPa are used as the identification factors of the two types of weather systems to automatically identify the height of the inversion layer and the inversion layer, so as to improve the accuracy of the ice accretion prediction. The specific method is as follows.

[0131] (1) Calculate the geopotential height anomaly (Zano) at 500 hPa.

[0132]

[0133] wherein, is the average geopotential height of a specific region.

[0134] The geopotential height anomaly (Zano) is used to measure the pressure anomaly of the current region, and Zano<0 represents the cold air activity area in the low pressure system, and Zano>0 represents the high pressure system (warm air dominant area).

[0135] (2) Identify the cold nest center.

[0136] ​For example, the Zano minimum 2% area is selected as the cold pool center. Figure 2 The gray gradient part in the figure indicates the low pressure system core, i.e., the cold pool center (shown in a red dashed box in the figure). Figure 2 The gray gradient part in the figure indicates the low pressure system core, i.e., the cold pool center (shown in a red dashed box in the figure).

[0137] (3) Identify the upper trough center.

[0138] For example, the curvature of Zano is calculated as follows:

[0139] The area with a curvature greater than or equal to 10 is selected as the upper trough center. Figure 3 The green gradient part in the figure indicates the upper trough center (shown in a red dashed box in the figure). Figure 3 The green gradient part in the figure indicates the upper trough center (shown in a red dashed box in the figure).

[0140] (4) Determine the characteristic meteorological data of the weather system.

[0141] When the centers of the two types of weather systems are determined, the icing-affected area of the weather system is set as an area extending a certain range (e.g., about 10° of latitude and longitude) west and south of the cold pool center, and an area extending a certain range (e.g., about 5° of latitude and longitude) around the upper trough center.

[0142] As shown in FIGS. Figure 2 ,The low pressure center is the cold pool center, and the area with a large wind field curvature is the upper trough center. In this way, the cold pool center and the upper trough can be identified by the geopotential height (Z500) and the wind field curvature, and the icing-affected area can be determined. 3 (5) Identify the key meteorological factors.

[0143] By identifying the heights of the inversion layer and the wettest layer within the above-mentioned icing-affected area, the vertical velocity ω and its membership function m inv as the key meteorological factors of icing weather in the Northeast region are calculated between the inversion layer H moist and the wettest layer H ω . The inversion layer H inv is a layer where the temperature increases with height, which is usually found by calculating the temperature profile and identifying the layer where the height increases but the temperature also increases. Below the inversion layer, the height layer with the maximum humidity is found, which is the wettest layer H moist . Specifically,

[0144] represents the membership of the vertical velocity to aircraft icing, where the horizontal axis is the vertical velocity (Pa / s), and the vertical velocity ω is the barometric vertical velocity, which is the rate of change of air particles with time and pressure in an isobaric coordinate system, and is expressed in units of Pa / s. Figure 4 Figure 4In the embodiment, the vertical velocity ω is mapped to a membership in the range of [0, 1] by a fuzzy function (which can be obtained by analysis or calculation, etc.). Based on the membership function m Figure 4 It can be seen that the membership function m ω is as follows:

[0145] m ω = 1, ω < -1;

[0146] m ω = 1 + 0.1ω, 0 > ω > -1 Pa / s;

[0147] m ω = -0.02ω, ω > 0 Pa / s.

[0148] Finally, based on the initial ice accumulation potential algorithm CIP0, the membership function m ω of the vertical velocity ω is combined to obtain the modified ice accumulation potential algorithm CIP final = CIP0 + (1 - CIP0) x m ω . Thus, based on the CIP final , the input related numerical values are calculated, and the ice accumulation prediction result can be output.

[0149] In summary, in the embodiment, in the northeast region, the cold vortex and high-level trough and other weather systems can be identified to more accurately determine the region where ice accumulation occurs. By identifying the inversion layer and the height of the wettest layer in the three-dimensional region, the membership function of the key meteorological factor can be established, the ice accumulation potential algorithm can be effectively optimized, and the accuracy of the ice accumulation prediction can be improved.

[0150] In the following, 44 northeast ice accumulation events in recent years are randomly selected to compare and measure the performance of each prediction algorithm. As a prediction algorithm, for example, the IC index commonly used in the art, the CIP0 before optimization, and the CIP final after optimization are selected.

[0151] First, the ice accumulation events are divided into "ice accumulation" or "non-ice accumulation" by the ice accumulation potential values of different ice accumulation potential algorithms (IC index, CIP0, CIP final ), and the hit number of different algorithms at different thresholds is calculated. Here, the threshold is set to 0.5, 0.6, 0.7, 0.8, and 0.9, that is, when the ice accumulation potential of a method exceeds the value, it is considered to be a successful prediction. The hit number refers to the number of times a certain algorithm correctly predicts the real ice accumulation event at a given threshold. If in 44 events, a certain algorithm predicts 30 ice accumulations, of which 13 are consistent with the actual ice accumulation event, then the hit number is 13. The hit rate = hit number / 44, which evaluates the overall prediction effect of different algorithms.

[0152] In this embodiment, the IC index is calculated as follows:

[0153]

[0154] where RH is the relative humidity and T is the temperature.

[0155] The calculation results are shown in Table 1 below. Figure 5

[0156]

[0157]

[0158] Figure 5 The horizontal axis represents the number of different events, and the vertical axis represents the ice accretion potential. The overall graph represents the ice accretion potential index of different events. The blue line segment represents the calculation results of IC, the red line segment represents the calculation results of CIP0, and the green line segment represents the calculation results of CIP final The calculation results are compared with the preset threshold value. If the calculation result exceeds the threshold value, it means that the calculation result hits an ice accretion event (i.e., is predicted to be an ice accretion event). As shown in Table 1, the number of hits of IC at the threshold value of 0.5 is 13, which means that it successfully predicts 13 ice accretion events in 44 events, and the hit rate is about 13 / 44≈29.5%. The number of hits of CIP final at the threshold value of 0.5 is 27, which means that it successfully predicts 27 events, and the hit rate is about 27 / 44≈61.4%, which is significantly better than IC. As the threshold value increases (from 0.5 to 0.9), the number of hits of all algorithms decreases, because a more stringent standard will reduce the number of events predicted to be ice accretion events. However, overall, CIP final has a significantly higher number of hits than IC and CIP0 at all threshold values, indicating that this algorithm can more accurately predict ice accretion events.

[0159] In this specification, the functions of the elements disclosed can be performed by electronic circuits or processing circuits that include a general-purpose processor, a special-purpose processor, an integrated circuit, an ASIC (Application Specific Integrated Circuit), an existing circuit, and / or a combination thereof configured to perform the disclosed functions. A processor is considered a processing circuit or circuit because it includes transistors or other circuits. In this disclosure, a circuit, unit, or means is hardware that performs the recited function, or is hardware programmed to perform the recited function. The hardware can be hardware disclosed in this specification, or other known hardware programmed or configured to perform the recited function. When the hardware is considered to be a processor, the circuit, means, or unit is a combination of hardware and software used for the configuration of the hardware and / or processor.

[0160] ​As another aspect, the present application also provides a computer readable storage medium, which can be included in the computer device described in the above embodiments and each modification, or can exist independently without being assembled into the computer device. The above computer readable storage medium carries one or more programs, when the one or more programs are executed by the computer device, the computer device implements the method of the above embodiments and modifications. For example, the computer device can implement each step shown in each figure.

[0161] According to an aspect of the present application, there is provided a computer program product comprising computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the method provided in various optional implementations of the above embodiments.

[0162] The above detailed description serves to further explain the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above is only one specific embodiment of the present application, and is not limited to the protection scope of the present application. Without departing from the basic characteristics of the present application, the present application can be embodied in various forms. Therefore, the embodiments in the present application are used for illustration and not limitation. Since the scope of the present application is defined by the claims and not by the specification, and all changes falling within the scope defined by the claims, or within the equivalent scope of the defined scope, should be understood as included in the claims. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for predicting ice accretion, characterized in that: include: Step S1: Select a specific area, obtain meteorological data and process it into a unified format. Step S2: constructing an initial ice accumulation potential algorithm for the specific area based on the meteorological data. Step S3: Based on the meteorological data, identifying the specific weather system and key meteorological factors of the specific area, and establishing a membership function of the key meteorological factors under the specific weather system, Step S4: Optimizing the initial ice accretion potential algorithm in combination with the membership function to obtain a localized ice accretion potential algorithm. Step S5: Outputting ice accretion potential based on the localized ice accretion potential algorithm.

2. The ice accretion prediction method according to claim 1, characterized in that: In step S1, the meteorological data related to the specific area is obtained from pilot reports, reanalysis data, satellite data, aircraft meteorological data, civil aviation meteorological messages, radar data, and sounding data.

3. The ice accretion prediction method according to claim 1, characterized in that: In the step S1, when the meteorological data is processed into a unified format, time resolution unification and spatial resolution unification are performed.

4. The ice accretion prediction method according to claim 3, characterized in that: When unifying the time resolution, the meteorological data of one hour before and after the ice accumulation time of a certain ice accumulation event in the specific area is selected and interpolated as shown in the following formula: Among them, X t is the meteorological data at the ice accumulation time t, X t0 is the meteorological data of t0 one hour before the ice accumulation time t, X t1 It is the meteorological data of the hour t1 after the ice accumulation time t.

5. The ice accretion prediction method according to claim 3, characterized in that: When unifying the spatial resolution, the meteorological data of the grid where the icing location of a certain icing event in the specific area is located is selected and interpolated as shown in the following formula: Where X is the meteorological data of the ice accumulation location, d1, d2, d3, d4 are the distances between the ice accumulation location and the four vertices of the grid, d5, d6, d7, d8 are the vertical distances between the ice accumulation location and the four sides of the grid, is the weight.

6. The ice accretion prediction method according to claim 1, characterized in that: In the step S2, it includes: selecting meteorological factors from the meteorological data processed into a unified format and constructing a membership function of the meteorological factors, and constructing the initial ice accretion potential algorithm based on the membership function of the meteorological factors and the corresponding weights.

7. The ice accretion prediction method according to claim 1, characterized in that: In the step S2, it includes: classifying icing scenarios according to different meteorological conditions based on a decision tree, selecting different meteorological factors for each icing scenario from the meteorological data processed into a unified format and constructing a membership function of the meteorological factors, and constructing different initial icing potential algorithms based on the membership functions of the meteorological factors.

8. The ice accretion prediction method according to claim 7, characterized in that: The icing scenarios include cloudless scenarios, single-layer cloud scenarios, convective cloud scenarios, weak precipitation scenarios, and multi-layer cloud scenarios.

9. The ice accretion prediction method according to claim 1, characterized in that: In step S3, it includes: Based on ice accumulation events in a specific area, the specific weather system that has a greater impact on ice accumulation in the specific area and the key meteorological factors that have a greater impact on ice accumulation under the specific weather system are pre-determined. identifying the specific weather system based on the meteorological data processed into a unified format, The membership function of the key meteorological factors is established based on the meteorological data processed into a unified format and the characteristic meteorological data of the specific weather system.

10. The ice accretion prediction method according to claim 9, characterized in that: When the specific area is the Northeast region, the specific weather system is a cold vortex and a high-altitude trough, the key meteorological factor is the vertical speed, and the characteristic meteorological data is the height of the inversion layer and the wettest layer.

11. The ice accretion prediction method according to claim 1, characterized in that: In step S4, the localized ice accretion potential algorithm is expressed by the following formula: Among them, CIP final =CIP0+(1-CIP0)×m 关键气象因子 CIP final is the localized ice accretion potential algorithm, CIP0 is the initial ice accretion potential algorithm, m 关键气象因子 is the membership function of the key meteorological factors.

12. An ice accretion forecasting system comprising one or more memories and one or more processors coupled to the one or more memories, characterized in that: The processor is configured to perform the method of any one of claims 1 to 11.

13. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 11 is implemented.

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 11 is implemented.

Images