A method and system for hierarchical icing prediction

Abstract

The present application relates to the technical field of aircraft icing prediction, and particularly relates to a hierarchical prediction method and system for aircraft icing, comprising the steps of: obtaining hierarchical indexes in a flight process; performing hierarchical processing on the hierarchical indexes to obtain at least one hierarchical membership degree; obtaining cloud top temperature in the flight process; determining a cloud top temperature membership degree according to the cloud top temperature and a corresponding cloud top temperature membership degree function; forming at least two groups of membership degrees according to the at least two hierarchical membership degrees and the cloud top temperature membership degree; calculating at least two initial icing potentials according to the at least two groups of membership degrees and an initial icing potential model; calculating at least two final icing potentials according to the at least two initial icing potentials; and outputting corresponding hierarchical prediction results according to the at least two final icing potentials. The present application can greatly improve the accuracy and efficiency of aircraft icing prediction with the aid of fuzzy logic.

Description

Technical Field

[0001] This invention relates to the field of aircraft icing forecasting technology, specifically to a graded forecasting method and system for icing. Background Technology

[0002] Aircraft icing can disrupt the aerodynamic shape of the fuselage and wings, significantly reducing lift, increasing drag, and increasing overall weight, directly affecting the aircraft's handling and flight stability. If ice covers or blocks critical components such as pitot tubes, engine air intakes, and control surfaces, it can also cause distorted avionics data, reduced engine power, and even control surface jamming and wing stall, seriously threatening flight safety.

[0003] In response, some predictive methods for aircraft icing have been proposed in traditional technologies.

[0004] For example, patent application CN116788512A proposes a method, device, and medium for predicting aircraft air icing potential based on fuzzy logic. The method includes: establishing fuzzy logic membership functions for atmospheric temperature, relative humidity, and cloud top temperature based on atmospheric temperature, relative humidity, cloud top temperature, vertical airflow velocity, and atmospheric liquid water content; establishing fuzzy logic membership functions for vertical airflow velocity and atmospheric liquid water content and obtaining the SCIP algorithm equation; acquiring parameters from future periods or historical forecasts, substituting them into the SCIP algorithm equation, and predicting the aircraft icing index; determining the degree of icing for the aircraft in the future period based on the icing index and icing degree judgment threshold, and planning and determining the aircraft flight route. This invention improves the generalization performance and anti-fitting performance of the diagnostic algorithm by adding two factors affecting air icing—vertical airflow velocity and atmospheric liquid water content—to the CIP algorithm.

[0005] For example, patent application CN115292656A discloses an invention providing a fuzzy logic-based method and apparatus for predicting aircraft icing, relating to the technical field of aircraft icing prediction. The invention includes: acquiring target data for the region to be predicted and constructing a membership function corresponding to the target data based on fuzzy logic; constructing a set of prediction index equations based on the target data, the membership function, and the weighting factors corresponding to the membership function; calculating the false alarm rate and accuracy of each prediction index equation based on the set of prediction index equations and the actual aircraft icing dataset corresponding to the region to be predicted; constructing an ROC curve based on the false alarm rate and accuracy of each prediction index equation, and determining the target prediction index equation based on the ROC curve; and determining the threshold for judging the severity of aircraft icing in the region to be predicted based on the false alarm rate and accuracy of the target prediction index equation and the actual aircraft icing dataset. This invention solves the technical problem of low prediction accuracy in existing aircraft icing prediction methods.

[0006] For example, patent application CN121051126A proposes an update method and system for aviation icing identification based on real-time aircraft detection. The method includes acquiring real-time aircraft data to determine icing occurrence information; constructing an icing report membership function based on the icing occurrence information to characterize the spatial and temporal distribution of icing occurrence probability; and calculating an icing potential index by combining the icing report membership function with a basic icing index to identify the three-dimensional distribution of icing areas. This method achieves precise icing location, dynamic prediction, and real-time early warning, improving aviation safety and flight efficiency.

[0007] However, these traditional technologies are insufficient to meet the needs for accurate and timely icing assessment and early warning during flight. Summary of the Invention

[0008] The purpose of this invention is to provide a graded forecasting method and system for aircraft icing, which partially solves or alleviates the above-mentioned shortcomings in the prior art and can greatly improve the accuracy and efficiency of aircraft icing forecasting by means of fuzzy logic.

[0009] To solve the aforementioned technical problems, the present invention specifically adopts the following technical solution:

[0010] A first aspect of the present invention is to provide a method for graded forecasting of ice accumulation, comprising the steps of:

[0011] S101, acquire the classification indicators during the flight process, the classification indicators including: temperature and relative humidity;

[0012] S102, perform hierarchical processing on the hierarchical index to obtain at least one hierarchical membership degree; wherein, S102 includes the following steps:

[0013] Obtain at least two pre-constructed hierarchical membership functions for the hierarchical index, and at least two of the hierarchical membership functions are used to define the hierarchical membership for different degrees of ice accumulation;

[0014] At least two hierarchical membership degrees are obtained by using the hierarchical index and the hierarchical membership function respectively;

[0015] S103, Obtain the cloud top temperature during the flight process;

[0016] S104, determine the cloud top temperature membership degree based on the cloud top temperature and the corresponding cloud top temperature membership function;

[0017] S105, at least two sets of membership degrees are formed based on at least two hierarchical membership degrees and the cloud top temperature membership degree;

[0018] S106, calculate at least two initial ice accumulation potentials based on at least two sets of membership degrees and initial ice accumulation potential models respectively;

[0019] S107, at least two final ice accumulation potentials are calculated based on the at least two initial ice accumulation potentials respectively;

[0020] S108, output the corresponding graded forecast results based on at least two of the final ice accumulation potentials.

[0021] In some embodiments, the steps further include:

[0022] S109, Construct the hierarchical membership function based on historical flight data, wherein the historical flight data includes: temperature and / or relative humidity.

[0023] In some embodiments, S109 includes the step of:

[0024] S1091, the historical flight data is divided into at least two datasets according to the degree of historical icing;

[0025] S1092, a hierarchical membership function is obtained by fitting the dataset. The hierarchical membership function is a function used to describe the changes in hierarchical indicators and the probability of icing, and the probability of icing = the number of icing events that occur at the temperature / the number of all historical flight data at the temperature.

[0026] In some embodiments, when the grading index is the temperature, S109 further includes the step: S1093, determining whether the grading membership function needs to be corrected;

[0027] S1093 includes:

[0028] S10931, the key verification area is set by the user;

[0029] S10932, identify whether an abnormal peak appears in the key verification area; wherein, when the abnormal probability corresponding to the peak is greater than the first probability value and less than the second probability value, the peak is identified as an abnormal peak.

[0030] S10933 is a signal for suggesting corrections to the abnormal peak.

[0031] In some embodiments, prior to S10933, the method further includes:

[0032] Obtain the number of historical flight data corresponding to the temperature of the abnormal peak;

[0033] When the quantity is less than or equal to a preset quantity threshold, then S10933 is allowed.

[0034] In some embodiments, prior to S10933, the method further includes:

[0035] Identify the number of abnormal peaks;

[0036] When the number of anomalies exceeds a preset anomaly threshold, then proceeding to step S10933 is permitted.

[0037] A second aspect of the present invention is to provide a graded forecasting system for ice accumulation, comprising:

[0038] The grading index acquisition module is used to acquire grading indexes during flight, including temperature and relative humidity.

[0039] A hierarchical processing module is used to perform hierarchical processing on the hierarchical indicators to obtain at least one hierarchical membership degree; wherein, the hierarchical processing module includes:

[0040] The function construction unit is used to obtain at least two pre-constructed hierarchical membership functions for the hierarchical index, and the at least two hierarchical membership functions are used to define the hierarchical membership of different degrees of ice accumulation.

[0041] The membership degree acquisition unit is used to obtain at least two hierarchical membership degrees by using the hierarchical index and the hierarchical membership degree function.

[0042] A cloud top temperature acquisition module is used to acquire the cloud top temperature during the flight process;

[0043] The cloud top membership determination module is used to determine the cloud top temperature membership degree based on the cloud top temperature and the corresponding cloud top temperature membership function.

[0044] A membership degree combination module is used to form at least two sets of membership degrees based on at least two hierarchical membership degrees and cloud top temperature membership degrees;

[0045] The first prediction module is used to calculate at least two initial ice accumulation potentials based on at least two sets of membership degrees and initial ice accumulation potential models, respectively.

[0046] The second prediction module is used to calculate at least two final ice accumulation potentials based on the at least two initial ice accumulation potentials respectively;

[0047] The graded forecast module is used to output corresponding graded forecast results based on at least two final ice accumulation potentials.

[0048] In some embodiments, it also includes:

[0049] The membership function construction module is used to construct the hierarchical membership function based on historical flight data, which includes temperature and / or relative humidity.

[0050] In some embodiments, the membership function construction module includes:

[0051] A dataset partitioning unit is used to divide the historical flight data into at least two datasets based on the degree of historical icing.

[0052] The hierarchical membership function fitting unit is used to fit a hierarchical membership function to the dataset. The membership function is a function used to describe the changes in hierarchical indicators and the probability of icing, and the probability of icing is equal to the number of icing events that occur at the temperature / the number of all historical flight data at the temperature.

[0053] In some embodiments, when the grading index is the temperature, the membership function construction module further includes: a correction judgment unit, used to determine whether the grading membership function needs to be corrected;

[0054] The correction judgment unit includes:

[0055] The verification area setting sub-unit is used for users to set key verification areas;

[0056] An anomaly identification subunit is used to identify whether an abnormal peak appears in the key verification area; wherein, when the anomaly probability corresponding to the peak is greater than a first probability value and less than a second probability value, the peak is identified as an abnormal peak.

[0057] A signal generation subunit is used to generate a suggested correction signal for the abnormal peak.

[0058] Beneficial technical effects:

[0059] Firstly, this invention proposes a restrictive icing classification prediction mechanism that classifies classification indicators while leaving unclassified indicators unclassified. This significantly simplifies calculations, allowing for more targeted calculation of icing potential from large amounts of flight data. Furthermore, it reduces the computational complexity of the computer system.

[0060] Alternatively, the present invention may choose to treat graded and non-graded indicators differently. From this perspective, the present invention can better control the calculation cost while calculating the ice accumulation potential through restrictive grade settings.

[0061] Secondly, by correcting the key verification region of the temperature membership function, this invention can avoid or reduce the large fluctuations in icing probability and low reliability caused by sample inhomogeneity or local sample insufficiency (such as the scarcity of flight detection data in certain temperature ranges), which can easily lead to abnormal peaks. Alternatively, by correcting the key verification region, it can also avoid or reduce abnormally high or low icing probability caused by detection instrument errors, reporting and recording deviations, etc.

[0062] In other words, this invention proposes a restrictive hierarchical membership function correction mechanism, which corrects specific temperature ranges and specific anomalous peaks of the hierarchical membership function, thereby improving the reliability of the membership function while avoiding membership distortion caused by over-correction and reducing the difficulty of correcting the hierarchical membership function. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale. Obviously, the drawings described below are some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0064] Figure 1 Example diagram of membership function for liquid water content in clouds provided by the present invention;

[0065] Figure 2 This is an example diagram illustrating the relationship between ice accumulation exposure time and ice accumulation amount under typical conditions provided by the present invention;

[0066] Figure 3 A flowchart illustrating a graded forecasting method for ice accumulation provided by the present invention;

[0067] Figure 4 Example diagram of temperature membership function provided for this invention;

[0068] Figure 5 Example graph of temperature normalization curve and membership function under weak icing conditions provided by the present invention;

[0069] Figure 6 Example graph of temperature normalization curve and membership function under light icing conditions provided by the present invention;

[0070] Figure 7 Example graph of temperature normalization curve and membership function under moderate icing conditions provided by the present invention;

[0071] Figure 8 Example graph of temperature normalization curve and membership function under heavy icing conditions provided by the present invention;

[0072] Figure 9 Example diagram of the original relative humidity membership function provided by the present invention;

[0073] Figure 10Example graphs of relative humidity normalization curves and membership functions under different icing conditions provided by this invention;

[0074] Figure 11 Example diagram of cloud top temperature membership function provided by the present invention;

[0075] Figure 12 Example diagram of the vertical velocity membership function provided by this invention;

[0076] Figure 13 A schematic diagram of the structure of a graded forecasting system for ice accumulation provided by the present invention;

[0077] Figure 14 This is a schematic block diagram of the structure of a computer device provided by the present invention. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0079] In this document, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0080] In this document, the terms "upper," "lower," "inner," "outer," "front," "rear," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the present invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0081] In this document, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0082] In this document, "and / or" includes any and all combinations of one or more of the listed related items.

[0083] In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0084] As used in this specification, the term "about" typically means + / -5% of the value, more typically + / -4% of the value, more typically + / -3% of the value, more typically + / -2% of the value, even more typically + / -1% of the value, and even more typically + / -0.5% of the value.

[0085] In this specification, certain embodiments may be disclosed in a range-bound format. It should be understood that this "range-bound" description is merely for convenience and brevity and should not be construed as a rigid limitation on the disclosed range. Therefore, the description of a range should be considered as having specifically disclosed all possible subranges and the individual numerical values ​​within those ranges. For example, a description of the range 1-6 should be considered as having specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., and the individual numbers within those ranges, such as 1, 2, 3, 4, 5, and 6. This rule applies regardless of the breadth of the range.

[0086] Example 1:

[0087] Please see Figure 3 In some embodiments, the present invention proposes a fuzzy logic-based method for predicting aircraft icing levels, comprising the following steps:

[0088] S101, acquire the classification indicators during the flight process, the classification indicators including: temperature and relative humidity;

[0089] S102, perform hierarchical processing on the hierarchical index to obtain at least one hierarchical membership degree; wherein, S102 includes the following steps:

[0090] Obtain at least two pre-constructed hierarchical membership functions for the hierarchical index, and at least two of the hierarchical membership functions are used to define the hierarchical membership for different degrees of ice accumulation;

[0091] At least two hierarchical membership degrees are obtained by using the hierarchical index and the hierarchical membership function respectively;

[0092] S103, Obtain the cloud top temperature during the flight process;

[0093] S104, determine the cloud top temperature membership degree based on the cloud top temperature and the corresponding cloud top temperature membership function;

[0094] S105, at least two sets of membership degrees are formed based on at least two hierarchical membership degrees and the cloud top temperature membership degree;

[0095] S106, calculate at least two initial ice accumulation potentials based on at least two sets of membership degrees and initial ice accumulation potential models respectively;

[0096] S107, at least two final ice accumulation potentials are calculated based on the at least two initial ice accumulation potentials respectively;

[0097] S108, output the corresponding graded forecast results based on at least two of the final ice accumulation potentials.

[0098] In some embodiments, the area where the ice accumulation potential needs to be predicted (or the target area) can be divided into grids (or pixels), and the ice accumulation potential in the grids can be calculated sequentially.

[0099] In some embodiments, relative humidity refers to the percentage of the actual amount of water vapor (vapor pressure) in the air at a given temperature to the maximum amount of water vapor that the air can hold at that temperature (saturated vapor pressure).

[0100] In some embodiments, cloud top temperature refers to the temperature of the top layer of clouds as observed / calculated by satellites or meteorological models.

[0101] In some embodiments, see Figures 5-8 and Figure 10 It is possible to construct hierarchical membership functions for different icing conditions (or degrees of icing).

[0102] In some embodiments, grading refers to assessing the likelihood of at least one degree of icing (e.g., light icing) occurring, corresponding to a grading index (e.g., temperature).

[0103] For example, based on the temperature value and the corresponding membership function for light and moderate icing, the membership degree for light icing at -8℃ is 'a', and the membership degree for moderate icing is 'b'.

[0104] In some embodiments, the hierarchical membership degree refers to the probability of different degrees of icing occurring corresponding to the hierarchical index.

[0105] The intensity of icing (or degree of icing) can be classified according to the icing rate (such as the thickness of icing on the wing per hour). For example, slight icing <0.6cm / h, light icing 0.6~2.5cm / h, moderate icing 2.5~7.5cm / h, and severe icing >7.5cm / h.

[0106] In some embodiments, cloud top temperature membership refers to the probability of icing occurring at different cloud top temperatures.

[0107] In this embodiment, the present invention proposes a graded prediction mechanism for aircraft icing, which can select different processing methods for different data types. Specifically, taking the temperature grading index as an example, the graded membership degrees of trace icing, light icing, moderate icing, and heavy icing corresponding to different temperatures can be obtained based on the grading index and the graded membership function. For cloud top temperature, the cloud top temperature membership degree is directly determined based on the cloud top temperature membership function.

[0108] In other words, this invention proposes a restrictive icing classification prediction mechanism that classifies classification indicators while leaving unclassified indicators unclassified. This significantly simplifies calculations, allowing for more targeted calculation of icing potential from large amounts of flight data. Furthermore, it reduces the computational complexity of the computer system.

[0109] Alternatively, the present invention may choose to treat graded and non-graded indicators differently. From this perspective, the present invention can better control the calculation cost while calculating the ice accumulation potential through restrictive grade settings.

[0110] In some embodiments, at least two hierarchical membership degrees can be obtained by using hierarchical indices and pre-built hierarchical membership functions.

[0111] For example, based on a certain temperature value, four graded membership functions (slight icing / mild icing / moderate icing / heavy icing) can be determined.

[0112] Table 1

[0113]

[0114] For example, please refer to Table 1. A set of membership degrees can be formed based on the temperature T1, relative humidity RH1 (i.e., two levels of membership degrees) and cloud top temperature CTT under the degree of slight icing.

[0115] For example, a set of membership degrees can be formed based on the temperature T2, relative humidity RH2 (i.e., two levels of membership degrees) and cloud top temperature CTT under the condition of light icing.

[0116] Similarly, four sets of membership degrees can be formed based on the hierarchical membership degrees under the four degrees of icing and the membership degrees of cloud top temperature.

[0117] In some embodiments, the initial icing potential model can be a decision tree model (as shown in Table 2). For example, the input data of the initial icing potential model can include grading indicators, cloud top temperature, cloud structure classification (e.g., no cloud, single-layer cloud, multi-layer cloud), precipitation type classification (e.g., no precipitation, rain, snowfall, etc.); the output data of the initial icing potential model can be the initial icing potential for different degrees of icing, with a larger value indicating a higher probability of icing.

[0118] In some embodiments, different vertical layers can be applied to calculate the initial icing potential value under different weather scenarios. For example, different vertical layers can be selected to calculate the initial icing potential value under single-layer cloud precipitation, multi-layer cloud precipitation, and no precipitation. For example, the existing weather scenario can be determined first, and different vertical layers can be applied to calculate the initial icing potential (InitialCIP) under different scenarios. For example, four different scenarios can be set up: clear sky, snowfall, single-layer cloud, and multi-layer cloud, and the aircraft icing situation under single-layer cloud and multi-layer cloud precipitation and no precipitation under specific scenarios is discussed. The equations used are shown in Table 2.

[0119] Table 2

[0120]

[0121] a. Clear sky

[0122] When no clouds are detected in a grid area (i.e., clear sky), the initial aircraft icing potential for that area is determined to be 0. This is because in clear sky conditions, there is a lack of supercooled water droplets and cloud droplets in the atmosphere, and there are no material carriers that can collide with the aircraft surface and cause icing. Furthermore, it is difficult to reach the saturated vapor pressure threshold that allows gaseous water vapor to condense directly into liquid water on the aircraft surface. Therefore, the initial icing potential for that area is determined to be 0.

[0123] b. Single-layer cloud

[0124] When the relative humidity is continuous throughout the entire vertical height of the cloud layer, it is considered that a single-layer cloud exists. First, the cloud top temperature membership function (M...) is used... CTT A preliminary assessment of the cloud phase is made, followed by consideration of the temperature membership function (M) at each altitude level. T ) and relative humidity membership function (M RHCalculate the initial icing potential. If liquid precipitation is observed below the cloud layer, the aircraft icing environment becomes more complex. In this case, when liquid precipitation is superimposed with a warm cloud top temperature (CTT > -15℃), it indicates that the collision-coalescence process of water droplets within the cloud is active. Therefore, if the ambient temperature is within the suitable icing range (-20-0℃), the risk of aircraft icing exists not only within the lowest cloud layer but also in the area below the cloud bottom. From the ground upwards, the icing potential extends at least to the cloud base and can further penetrate into the cloud interior; its initial icing potential is defined as:

[0125]

[0126] When ground meteorological observations record snowfall, it can be determined that all liquid water droplets in the clouds and the path of falling snow particles have been completely transformed into solid ice and snow particles. At this time, the aircraft icing potential in the area is determined to be 0.

[0127] c. Multi-layered cloud

[0128] When multiple cloud layers exist in the vertical structure of the atmosphere, the icing potential of each cloud layer needs to be determined independently. Specifically, the cloud top position of the upper cloud layer can be directly identified through satellite observation, so its icing potential can be calculated based on satellite remote sensing data. However, due to the obstruction of the upper cloud layer, direct observation of the lower cloud layer is limited, and numerical model data is needed to help determine its spatial location and cloud top height. In this model, if a dry layer exists, the simulated relative humidity threshold for this layer is below 84%, which is used to distinguish the vertical boundaries of the upper and lower cloud layers. Based on this, the lower cloud layer has independent cloud top temperature characteristics, and the model will calculate the icing layer thickness and icing potential separately for this layer to avoid interference between the meteorological parameters of the upper and lower cloud layers. In addition, when precipitation is observed at the ground, it is only applied to the lowest cloud layer and the elevation range below it, and does not extend upward to the middle and upper cloud layers. This is because the formation of ground precipitation is directly related to the collision-coalescence and melting of water droplets in the lowest cloud layer, and has a weaker correlation with the icing environment of the middle and upper cloud layers. Its initial icing potential is defined as:

[0129]

[0130] When ground meteorological observations record snowfall, it can be determined that all liquid water droplets in the clouds and the path of falling snow particles have been completely transformed into solid ice and snow particles. At this time, the icing potential of aircraft at all vertical heights in the area is determined to be 0.

[0131] In some embodiments, the final ice accumulation potential can be calculated from the initial ice accumulation potential as follows: (1).

[0132] Wherein, CLW refers to the liquid water content in the cloud, and W refers to the vertical velocity. is the membership function of the liquid water content in the cloud; is the membership function of the vertical velocity (e.g., when W is greater than zero, it corresponds to an upward vertical velocity). Here, a and b are the influence coefficients of vertical velocity and liquid water content on ice accumulation potential, respectively.

[0133] In some embodiments, the above formula (1) can be combined with at least two initial ice accumulation potentials. 1 and 2. At least two final icing potentials, CCIP1 and CCIP2, were calculated.

[0134] In some embodiments, different forecast thresholds can be preset for different degrees of icing. For example, the forecast threshold for trace icing can be 0.25, the forecast threshold for light icing can be 0.3, the forecast threshold for moderate icing can be 0.35, and the forecast threshold for heavy icing can be 0.55.

[0135] It is worth noting that different levels of icing require different response costs (e.g., light icing may only require enhanced monitoring, while heavy icing necessitates immediate changes to flight routes or altitudes). For example, responding to high-level alerts (corresponding to heavy icing) is costly. To address more severe icing levels, this invention sets a higher evidence threshold (e.g., a forecast threshold of 0.55) to prevent easily triggered forecasts that could cause unnecessary flight delays and operational losses, thereby improving the reliability of forecast results.

[0136] In other words, within the framework of the graded calculation of final icing potential of this invention, the same pixel may simultaneously output multiple high potential values ​​(i.e., the final icing potential corresponding to multiple icing levels all exceed their respective prediction thresholds, 0.3 for light icing and 0.6 for heavy icing). In this case, the highest icing level can be output (e.g., outputting a heavy icing warning). This approach simplifies the decision-making logic and reduces decision conflicts; in other words, it helps to achieve a better balance between computational difficulty and prediction accuracy.

[0137] In some embodiments, the steps further include:

[0138] S109, Construct a hierarchical membership function based on historical flight data, wherein the historical flight data includes: temperature and / or relative humidity.

[0139] In some embodiments, S109 includes the step of:

[0140] S1091, the historical flight data is divided into at least two datasets according to the degree of historical icing;

[0141] S1092, a hierarchical membership function is obtained by fitting the dataset. The membership function is a function used to describe the changes in hierarchical indicators and the probability of icing, and the probability of icing = the number of icing events that occur at the temperature / the number of all historical flight data at the temperature.

[0142] In some embodiments, to construct a hierarchical membership function for temperature under weak icing conditions, this invention first preprocesses the atmospheric temperature data from historical flight probes. The temperature range is discretized into several continuous intervals at 1°C intervals. The frequency of weak icing events within each temperature interval is statistically analyzed. The ratio of the frequency of weak icing events in each interval to the total number of flight probe records in that interval is calculated and normalized, ultimately yielding a temperature normalization curve under weak icing conditions (e.g., ...). Figure 5 (As shown in the figure). The normalized curve exhibits a significant single-peak characteristic, with the peak value located at -1℃, and a rapid decreasing trend from the peak towards both ends. This is because near -1℃, the atmospheric temperature is close to the freezing point but still below 0℃. Due to dynamic warming, supercooled water droplets are not easily frozen, thus making weak icing more likely. When the temperature is above -1℃, the thermal stability of supercooled water droplets increases, leading to a decrease in supercooled water droplet concentration and a lower probability of weak icing. When the temperature is below -1℃, the atmospheric temperature is far from the freezing point, and the frequency of weak icing events decreases accordingly. Referring to the normalized value, the closer the temperature is to -1℃, the higher the membership value, indicating a greater likelihood that the conditions for weak icing are met at that temperature.

[0143] In some embodiments, to construct a temperature membership function under light icing conditions, this invention first preprocesses the atmospheric temperature data from historical flight probes. The temperature range is discretized into several continuous intervals at 1°C intervals. The frequency of light icing events within each temperature interval is statistically analyzed. The ratio of the frequency of light icing events in each interval to the total number of flight probe records in that interval is calculated and normalized, ultimately yielding a temperature normalization curve under light icing conditions (e.g., ...). Figure 6 (As shown in the figure). The normalized curve, besides showing a significant peak near 0℃, exhibits slight fluctuations in the -14℃ to -6℃ range, generally decreasing rapidly from the peak towards both ends. The peak near 0℃ reflects that atmospheric conditions in this temperature range easily induce mild icing; when the temperature exceeds this peak range, the probability of mild icing decreases rapidly. Referring to the above normalization results, the membership function for mild icing (shown by the yellow curve in the figure) exhibits piecewise characteristics. Within the -28℃ to -14℃ range, the closer the temperature is to -14℃, the higher the membership value; in the -14℃ to -1℃ range, the membership function value is stable at 1, representing the highest probability of meeting the conditions for mild icing in this temperature range; thereafter, as the temperature increases, the membership function shows a rapid decreasing trend.

[0144] In some embodiments, to construct a temperature membership function under moderate icing conditions, the atmospheric temperature data from historical flight probes can first be preprocessed. The temperature range is discretized into several continuous intervals at 1°C intervals. The frequency of moderate icing events within each temperature interval is statistically analyzed. The ratio of the frequency of moderate icing events in each interval to the total number of flight probe records in that interval is calculated and normalized, ultimately yielding a temperature normalization curve under moderate icing conditions (e.g., ...). Figure 7 (As shown in the figure). The normalized curve shows obvious peaks at -2℃ and -13℃, with slight fluctuations in the range of -13℃ to -2℃, and a rapid decreasing trend from the peaks towards both ends. Referring to the above normalization results, the membership function of moderate icing (as shown by the yellow curve in the figure) exhibits piecewise characteristics. In the range of -28℃ to -13℃, the closer the temperature is to -13℃, the higher the membership value; in the range of -13℃ to -2℃, the membership function value is stable at 1, indicating that the probability of meeting the conditions for moderate icing is highest in this temperature range; thereafter, as the temperature increases, the membership function shows a rapid decreasing trend.

[0145] In some embodiments, to construct a temperature membership function under severe icing conditions, the atmospheric temperature data from historical flight probes can first be preprocessed. The temperature range is discretized into several continuous intervals at 1°C intervals. The frequency of severe icing events within each temperature interval is statistically analyzed. The ratio of the frequency of severe icing events in each interval to the total number of flight probe records in that interval is calculated and normalized, ultimately yielding a temperature normalization curve under severe icing conditions (e.g., ...). Figure 8 (As shown in the figure). The normalized curve shows obvious peaks at -3℃ and -14℃, with slight fluctuations in the range of -14℃ to -3℃, and a rapid decreasing trend from the peaks towards both ends. Referring to the above normalization results, the membership function of heavy icing (as shown by the yellow curve in the figure) exhibits piecewise characteristics. In the range of -28℃ to -14℃, the closer the temperature is to -14℃, the higher the membership value; in the range of -14℃ to -3℃, the membership function value is stable at 1, indicating that this temperature range has the highest probability of meeting the conditions for heavy icing; thereafter, as the temperature increases, the membership function shows a rapid decreasing trend.

[0146] It should be understood that the membership function of relative humidity (such as...) Figure 9As shown in the figure, its core purpose is to assess the presence of clouds within a flight area. The results show that the probability of icing increases positively with relative humidity, most frequently occurring in high relative humidity environments (i.e., areas with a higher probability of cloud formation), and gradually decreases as relative humidity decreases. It should be noted that some pilots reported relatively low relative humidity values. This phenomenon is mainly due to insufficient accuracy in relative humidity prediction, and may also be caused by positioning errors during pilot reporting (Brown et al. 1999). Overall, 74.9% of aircraft icing occurred when relative humidity exceeded 70%, and only 1.7% occurred in environments with relative humidity below 25%. The probability of aircraft icing gradually increases in the 70%-90% range. However, it should be noted that the relationship between relative humidity and icing depends more on model accuracy than on the physical mechanism of icing.

[0147] In some embodiments, to construct the relative humidity membership function under icing conditions, this invention first preprocesses the atmospheric relative humidity data from historical flight probes. The relative humidity range is discretized into several continuous intervals at 5% intervals. The frequency of icing events within each relative humidity interval is statistically analyzed. The ratio of the frequency of icing events in each interval to the total number of flight probe records in that interval is calculated and normalized, ultimately yielding the normalized relative humidity curve under icing conditions (e.g., ...). Figure 10 (As shown). The relative humidity normalized curves for different icing intensities show similar conclusions: the probability of icing increases positively with relative humidity, most commonly occurring in high relative humidity environments, and gradually decreases as relative humidity decreases. This is consistent with previous conclusions, indicating that icing intensity is not directly related to the distribution of relative humidity. Referring to the above normalization results, the membership function of icing (shown by the yellow curve in the figure) exhibits piecewise characteristics. Within the relative humidity range of 60% to 90%, the closer the relative humidity is to 90%, the higher the membership value; when the relative humidity is greater than 90%, the membership function value stabilizes at 1, representing the highest probability of meeting the icing conditions within this relative humidity range.

[0148] This invention can select historical flight detection data and divide the grading indicators into several equidistant intervals based on the physical value range of the grading indicators and the distribution characteristics of the observation data. Then, for each parameter interval, the number of samples of trace, light, moderate and heavy icing observed by the aircraft in that interval is counted, and the ratio of the sample to the total number of all observed samples in that interval (i.e., the icing probability) is calculated. Then, the ratio is normalized. Finally, the membership function value of a certain icing intensity corresponding to a specific parameter interval is formed based on the normalized value.

[0149] In some embodiments, when the grading index is the temperature, S109 further includes the step: S1093, determining whether the grading membership function needs to be corrected;

[0150] S1093 includes:

[0151] S10931, the key verification area is set by the user;

[0152] S10932, identify whether an abnormal peak appears in the key verification area; wherein, when the abnormal probability corresponding to the peak is greater than the first probability value and less than the second probability value, the peak is identified as an abnormal peak.

[0153] S10933 is a signal for suggesting corrections to the abnormal peak.

[0154] For example, please see Figure 6 Users can preset the critical verification zone to -15℃. The critical verification zone is used to determine whether the trend of the ice accumulation probability changing with the grading index is abnormal.

[0155] This invention, by correcting the key verification region of the temperature membership function, can avoid or reduce large fluctuations in icing probability and low reliability caused by sample inhomogeneity or local sample insufficiency (such as sparse flight detection data in certain temperature ranges), which can easily lead to abnormal peaks. Alternatively, by correcting the key verification region, it can also avoid or reduce abnormally high or low icing probability caused by instrument errors, reporting deviations, etc.

[0156] In other words, this invention proposes a restrictive hierarchical membership function correction mechanism, which corrects specific temperature ranges and specific anomalous peaks of the hierarchical membership function, thereby improving the reliability of the membership function while reducing the difficulty of correcting the hierarchical membership function.

[0157] In some embodiments, prior to S10933, the method further includes:

[0158] Obtain the number of historical flight data corresponding to the temperature of the abnormal peak;

[0159] When the quantity is less than or equal to a preset quantity threshold, then S10933 is allowed.

[0160] In some embodiments, when the number of historical flight data samples corresponding to an abnormal peak is small, it may indicate that there is a high probability of a deviation in the icing probability. In this case, a suggested correction signal can be generated for the corresponding abnormal peak.

[0161] In some embodiments, prior to S10933, the method further includes:

[0162] Identify the number of abnormal peaks;

[0163] When the number of anomalies exceeds a preset anomaly threshold, then proceeding to step S10933 is permitted.

[0164] In some embodiments, when only a few abnormal peaks appear in the critical verification area (e.g., the number of abnormal peaks is less than a preset abnormal threshold), the anomaly may be caused by random noise. Preferably, the present invention generates suggested correction signals for a large number of abnormal peaks, aiming to identify abnormal fluctuations that require intervention and can be resolved through correction, rather than responding to all anomalies.

[0165] Please see Figure 13 In some embodiments, the present invention proposes a graded forecasting system for ice accumulation, comprising:

[0166] The grading index acquisition module is used to acquire grading indexes during flight, including temperature and relative humidity.

[0167] A hierarchical processing module is used to perform hierarchical processing on the hierarchical indicators to obtain at least one hierarchical membership degree; wherein, the hierarchical processing module includes:

[0168] The function construction unit is used to obtain at least two pre-constructed hierarchical membership functions for the hierarchical index, and the at least two hierarchical membership functions are used to define the hierarchical membership of different degrees of ice accumulation.

[0169] The membership degree acquisition unit is used to obtain at least two hierarchical membership degrees by using the hierarchical index and the hierarchical membership degree function.

[0170] A cloud top temperature acquisition module is used to acquire the cloud top temperature during the flight process;

[0171] The cloud top membership determination module is used to determine the cloud top temperature membership degree based on the cloud top temperature and the corresponding cloud top temperature membership function.

[0172] A membership degree combination module is used to form at least two sets of membership degrees based on at least two hierarchical membership degrees and cloud top temperature membership degrees;

[0173] The first prediction module is used to calculate at least two initial ice accumulation potentials based on at least two sets of membership degrees and initial ice accumulation potential models, respectively.

[0174] The second prediction module is used to calculate at least two final ice accumulation potentials based on the at least two initial ice accumulation potentials respectively;

[0175] The graded forecast module is used to output corresponding graded forecast results based on at least two final ice accumulation potentials.

[0176] It should be understood that the aforementioned grading forecasting system for ice accumulation can be used to implement any of the method steps described in the embodiments of the present invention.

[0177] Example 2:

[0178] The following section will illustrate the graded forecasting method for ice accumulation proposed in this invention with specific examples:

[0179] In meteorological research, meteorological reanalysis data is used to dynamically fuse multi-source observational data with the physical processes of the model through numerical forecasting models and data assimilation techniques. This eliminates observational errors and spatiotemporal discontinuities. Historical multi-source observational data, including satellite remote sensing inversion products, ground meteorological station observations, upper-air meteorological sounding data, and meteorological radar observations, are systematically reprocessed, quality-controlled, and spatiotemporally interpolated to ultimately generate standardized meteorological datasets with spatiotemporal consistency and long-term continuity. The core value of such datasets lies in their ability to quantitatively reproduce the three-dimensional state of the atmosphere at different historical periods. This provides crucial support for climate change detection and attribution analysis, parameter optimization of numerical forecasting models, and improvement of forecast accuracy. Furthermore, it offers high-quality foundational data for regional meteorological simulation, extreme weather event replay, and environmental impact assessment.

[0180] In some embodiments, to meet the requirements of the present invention for the spatiotemporal resolution, coverage and accuracy of meteorological data, the fifth generation ECMWF atmospheric reanalysis of the global climate (ECMWF-ERA5) developed by the European Centre for Medium-Range Weather Forecasts (ECMWF) can be selected. This dataset boasts significant technical advantages. Its time coverage extends from 1940 to the present, meeting the needs of long-term meteorological analysis. With a horizontal resolution of 0.25°×0.25°, it is one of the highest horizontal resolution products currently available in global reanalysis data, enabling more detailed characterization of regional-scale atmospheric circulation features. Its temporal resolution is 1 hour, capturing short-term atmospheric state changes. Vertically, it includes 27 standard pressure layers, covering key atmospheric layers from near-surface 1000 hPa to upper atmosphere 100 hPa. Specifically, the pressure layers are: 1000 hPa, 975 hPa, 950 hPa, 925 hPa, 900 hPa, 875 hPa, 850 hPa, 825 hPa, 800 hPa, 775 hPa, 750 hPa, 700 hPa, 650 hPa, 600 hPa, 550 hPa, 500 hPa, 450 hPa, 400 hPa, 350 hPa, 300 hPa... hPa, 250 hPa, 225 hPa, 200 hPa, 175 hPa, 150 hPa, 125 hPa, and 100 hPa can fully reflect the distribution characteristics of atmospheric elements in the vertical direction.

[0181] From a data reliability perspective, the accuracy of ECMWF-ERA5 has been verified across multiple regions and meteorological elements globally. In comparisons of observed and reanalyzed values ​​of core meteorological parameters such as temperature, humidity, and wind speed, its errors remain at a low level. In this invention, ECMWF-ERA5 will serve as key driving data for the Weather Research and Forecasting (WRF) numerical model, providing standardized input for the initial meteorological field construction and boundary condition setting of the WRF model, thereby ensuring the rationality and accuracy of regional-scale meteorological simulations.

[0182] In some embodiments, to construct an icing meteorological database and establish an autonomous icing meteorological model, flight detection data from 53 cold cloud weather modification operations conducted between 2018 and 2024 can be obtained. This data was acquired using various cloud and fog environment detection instruments carried on aircraft, with the core being the microphysical characteristic parameters of icing clouds.

[0183] In some embodiments, the flight-related data (or historical flight data) used in this invention are all from the AIMMS atmospheric sounder, specifically including key parameters such as flight attitude, flight speed, ambient temperature, atmospheric pressure, and wind speed. Cloud microphysical parameters (including particle number concentration, median volume diameter (MVD), and liquid water content (LWC)) are acquired by the Cloud Droplet Probe 2 (CDP-2) system developed by DMT (Droplet Measurement Technologies) onboard the aircraft. This device can detect particles with diameters ranging from 2 to 50 μm, and its performance has been widely recognized in numerous previous flight detection tests. It is currently widely used in scenarios such as aircraft airworthiness certification testing.

[0184] In some embodiments, since there is a lack of direct records of aircraft icing intensity in historical flight detection data, numerical simulation methods can be used to obtain aircraft icing intensity. Specifically, the simulation is carried out using LOWICE, a computational fluid dynamics (CFD) software for aircraft icing developed by NASA based on Lewis wind tunnel tests. This software can output key results such as flow field distribution, drop impact characteristics, thermodynamic properties, and icing morphology, and has been widely verified and applied in the field of aircraft icing simulation (Cao et al., 2018; Han and Palacios, 2017; Sotomayor-Zakharov et al., 2024). By inputting the wing model, aircraft status parameters, cloud microphysical parameters, and exposure time into LEWICE, the wing icing thickness and maximum icing rate can be obtained. At the same time, according to the icing classification standards of the Federal Aviation Administration (FAA), the maximum icing rate is divided into: slight icing <0.6 cm / h, light icing 0.6~2.5 cm / h, moderate icing 2.5~7.5 cm / h, and heavy icing >7.5 cm / h.

[0185] Alternatively, one can use a method for predicting aircraft icing by constructing an icing numerical simulation lookup table, as described in patent application CN 118069954 A, to obtain data on the degree of icing.

[0186] Based on this, the present invention can further utilize LEWICE to construct a lookup table (LUT) for maximum icing rate of aircraft, aiming to quickly extract icing intensity information from massive historical flight detection data. Furthermore, the standard numerical model NACA0012 airfoil from Lewis wind tunnel testing is selected as the wing model, which has been widely used in icing simulation research (Olsen et al., 1984; Fukudomee et al., 2021; Mathieu et al., 2024). To ensure simulation accuracy, parameter settings were all based on existing research (Bernstein et al., 2005; Korolev and Isaac, 2006; Drage and Howe, 2008; Merino et al., 2019). Meteorological elements covering 95% of historical icing events were extracted from historical icing event studies as parameter boundaries (see the LOWICE parameter setting example provided in Table 3). The angle of attack was set to 4°, the temperature range was -28 to 4 ℃, the relative humidity was 70% to 100%, and the flight speed was 60 to 120 m / s. In accordance with Appendix C of Part 25 of the FAA Federal Aviation Regulations (FAR), the liquid water content (LWC) and the maximum droplet diameter (MVD) were set to 0 to 1 g / m³ and 5 to 100 μm, respectively, and the icing simulation duration was uniformly 10 min.

[0187] Table 3

[0188]

[0189] In some embodiments, a series of wing-type simulation experiments were conducted under fixed environmental conditions to study the correlation between icing exposure time and icing amount. Figure 2 A correlation graph showing the relationship between icing exposure time and icing amount under typical conditions (icing test temperature -8℃, flight speed 80m / s, relative humidity 100%, MVD 20 µm, LWC 0.8 g / cm³) is presented, illustrating the relationship between time and maximum icing thickness. Exposure times ranged from 1 to 20 minutes, with 1-minute intervals between tests. The results indicate that the maximum icing thickness increases with increasing exposure time. The uniform increase suggests that the rate of ice accumulation is unaffected by the duration of exposure, indicating a linear relationship between the two variables. Furthermore, Figure 2 The changes in ice type over time with ice accumulation exposure are shown. The location of maximum ice thickness was found to remain constant over time.

[0190] In some embodiments, this invention characterizes the growth conditions of icing clouds based on macro-meteorological features of the atmospheric environment, studies the significant intrinsic correlation between the formation and development of aircraft icing clouds and key atmospheric elements such as surface weather phenomena, air temperature stratification, cloud top temperature (CTT), and relative humidity (RH), and constructs a method for calculating icing potential. Its core approach lies in integrating data from numerical weather prediction, satellite, radiosonde, and ground observations, and using decision trees and fuzzy logic to analyze meteorological observation data, microphysical process correlations, and risk assessment logic. This research investigates the qualitative correlation and quantitative laws between various meteorological elements and aircraft icing risk, enabling accurate identification of aircraft icing risk. Specific steps include three-dimensional cloud structure judgment, initial icing potential calculation, and the construction of a final icing potential algorithm.

[0191] The three-dimensional structure of clouds is determined using relative humidity data output from the WRF numerical model and cloud top phase, cloud detection, and cloud top temperature data from geostationary meteorological satellites. First, cloud cover, cloud top height, cloud thickness, and precipitation type are assessed. Then, a threshold method is used to categorize aircraft icing into precipitation, non-precipitation, single-layer clouds, and multi-layer clouds. The specific steps are as follows:

[0192] (1) Cloud cover judgment

[0193] In some embodiments, satellite data can first be mapped onto a grid output by the WRF model, and matching data can be checked to determine whether clouds exist within each model grid cell. A satellite cloud detection product result of 3 for a grid cell indicates clear skies, 2 indicates clear skies, 1 indicates clouds, and 0 indicates clouds. When the cloud detection product is 1 or 0, the cloud base height, cloud top height, and the presence and type of precipitation are further assessed. Grid cells with cloud detection products of 3 or 2 are considered ice-free zones (the icing potential in the entire columnar structure is 0).

[0194] (2) Determining cloud top height and cloud thickness

[0195] In some embodiments, the determination of multi-layered clouds is based on the relative humidity threshold method. Specifically, when the relative humidity of a certain altitude layer on the vertical atmospheric profile is >84%, it is determined to be a cloud layer. On this basis, the cloud base and cloud top are further defined. If the relative humidity of a certain altitude layer is >84% and the relative humidity of the adjacent altitude layer directly below it is <84%, then the layer is the cloud base, and its height is the cloud base height. If the relative humidity of a certain altitude layer is >84% and the relative humidity of the adjacent altitude layer directly above it is <84%, then the layer is the cloud top, its height is the cloud top height, and the temperature of the layer is the initial cloud top temperature. When the cloud top temperature value of the corresponding pixel in the satellite data is greater than the initial cloud top temperature obtained by the relative humidity threshold method, the cloud top temperature in the satellite data is used as the actual cloud top temperature. At the same time, single-layered clouds and multi-layered clouds are distinguished according to the number of continuous cloud layers in the vertical whole-layer structure of the model grid unit. When there is only one continuous cloud layer, it is a single-layered cloud; when there are two or more continuous cloud layers, it is a multi-layered cloud. The height difference of continuous cloud layers is defined as the cloud layer thickness.

[0196] (3) Cloud precipitation type

[0197] In some embodiments, cloud precipitation type is obtained from the precipitation type numerical forecast product output by WRF. When the precipitation type product is 0, it is defined as non-precipitation. When the precipitation type product is 1 to 5, it is defined as precipitation. Among them, precipitation type product 1 represents rainfall, precipitation type product 2 represents snowfall, precipitation type product 3 represents freezing rain, precipitation type product 4 represents ice pellets, and precipitation type product 5 represents wet snow.

[0198] In some embodiments, the initial icing potential is calculated mainly by using fuzzy logic to construct the membership function of parameter fields such as temperature (T), relative humidity (RH), cloud top temperature (CTT) and aircraft icing, and then using a decision tree model to calculate the icing potential value for different weather scenarios.

[0199] In some embodiments, the membership function is constructed using fuzzy logic as the core technical approach. The system studies the statistical distribution relationship between three key meteorological parameters—temperature, relative humidity, and cloud top temperature—and different aircraft icing levels. This allows for the simulation of the continuous dynamic process of the transition of each parameter field from an icing environment to a non-icing environment, which is more in line with the actual evolution of atmospheric physics.

[0200] The applicant noted that, based on flight exploration data and fundamental research in cloud microphysics, supercooled liquid water is more likely to exist in clouds and precipitation under specific temperature conditions and cloud vertical structures. Specifically, such as... Figure 4 As shown, aircraft icing occurs most frequently at temperatures near 0°C, with the frequency decreasing as the temperature decreases. When the temperature is below -28°C, supercooled water is relatively rare (except for deep convective clouds and isolated "clean" clouds, which can still maintain supercooled water due to their special vertical motion intensity and water vapor supply conditions), thus reducing the likelihood of aircraft icing. Conversely, ice crystals have a lower probability of formation at temperatures near 0°C, but their formation probability increases significantly when the temperature drops below -10°C (Rogers & Yau, 1989; Rauber et al., 2000; Cober et al., 2001; Korolev et al., 2003). Furthermore, data from multiple field observation experiments (such as aviation meteorological special detection and cloud physics comprehensive observation experiments) show that aircraft icing most often occurs in the temperature range of -15 to -3℃ (Sand et al., 1984; Schultz & Politovich, 1992; Cober et al., 1995).

[0201] The applicant noted that the phase of cloud top condensate affects the composition of the underlying cloud layer. Relatively warm cloud top temperatures mean that the cloud layer is likely predominantly composed of liquid water. Conversely, if the cloud top temperature is low enough to produce ice crystals, these ice crystals will grow and fall into the underlying cloud layer, sometimes causing the cloud to become completely ice-crystallized. Geresdi et al. (2005) observed that as cloud top temperatures decrease, precipitation undergoes a gradual transition from liquid-dominated to ice-crystallized. To assess the probability of a cloud layer containing liquid water rather than being completely ice-crystallized at a given cloud top temperature, a cloud top temperature membership function can be plotted (…). Figure 11 ). Figure 11 The value peaks at CTT ≥ -12°C, as liquid water dominates in these warmer clouds; as CTT decreases, the value gradually decays but never reaches zero. Although low-temperature cloud tops typically indicate ice crystal dominance, liquid water may still be present in these clouds if the rate of liquid water formation exceeds the rate of consumption. The membership function gradually decays with decreasing cloud top temperature from -12°C to approximately -30°C, then levels off.

[0202] It should be understood that since cloud top temperature is primarily used to determine the phase state of particles in clouds, this invention does not consider it as a key factor in distinguishing icing intensity. Unlike the two classification indicators of temperature and relative humidity, this embodiment does not classify the membership degree of cloud top temperature.

[0203] In some embodiments, the final ice accumulation potential is calculated as follows:

[0204] (1)

[0205] The initial icing potential comprises three components: temperature, relative humidity, and cloud top temperature, comprehensively indicating the aircraft icing potential at a specific pixel within the target cloud. However, it is insufficient to provide guidance for aircraft icing forecasting; therefore, CLW (liquid water content in the cloud) and W (vertical velocity) can be used to adjust the initial icing potential. The membership function for the liquid water content in clouds; The vertical velocity membership function ( Figure 12 , Figure 1 ). Figure 12 In the diagram, the horizontal axis corresponds to different cloud liquid water contents, and the vertical axis corresponds to different cloud liquid water content membership degrees. Figure 1 In the diagram, the horizontal axis corresponds to different vertical velocities, and the vertical axis corresponds to different membership degrees of those velocities. When there is an upward vertical velocity, the adjustment factor is defined as... When there is a downward vertical velocity, the adjustment factor is defined as Where a and b are the influence coefficients of vertical velocity and liquid water content on ice accumulation potential, respectively. Preferably, a = 0.6 and b = 0.4.

[0206] In some embodiments, according to the above-described ice accumulation discrimination process, four CCIP (final ice accumulation potential) values ​​corresponding to trace ice accumulation, light ice accumulation, moderate ice accumulation and severe ice accumulation at the same pixel point can be calculated respectively.

[0207] In some embodiments, a preset forecast threshold can be used to determine the magnitude of the CCIP (final ice accumulation potential) value.

[0208] For example, if the CCIP value corresponding to a trace amount of icing is >0.25, it is determined that the probability of icing occurrence at this level is significantly higher; if the CCIP value corresponding to a light amount of icing is >0.3, it is determined that the probability of icing occurrence at this level is significantly higher; if the CCIP value corresponding to a moderate amount of icing is >0.35, it is determined that the probability of icing occurrence at this level is significantly higher; and if the CCIP value corresponding to a severe amount of icing is >0.55, it is determined that the probability of icing occurrence at this level is significantly higher.

[0209] In some embodiments, to ensure flight safety, if two or more different levels of icing all meet the criteria for a significantly higher probability of occurrence, the result with the most severe icing will be output first.

[0210] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0211] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a computer terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0212] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

[0213] In some embodiments, this application also provides a schematic block diagram of the structure of a computer device, please see... Figure 14 Computer programs can be used in situations such as Figure 14 It runs on the computer device shown. Figure 14As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The memory may include non-volatile storage media and internal memory. The non-volatile storage media may store an operating system and computer programs. The computer programs include program instructions that, when executed, cause the processor to perform arbitrary methods. The processor provides computational and control capabilities to support the operation of the entire computer device. The internal memory provides an environment for the execution of the computer programs in the non-volatile storage media; when executed by the processor, these programs cause the processor to perform arbitrary methods. The network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 14 The structures shown are merely block diagrams of a portion of the structure related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. It should be understood that the processor may be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

Claims

1. A method for graded forecasting of ice accumulation, characterized in that, Including the following steps: S101, acquire the classification indicators during the flight process, the classification indicators including: temperature and relative humidity; S102, perform hierarchical processing on the hierarchical index to obtain at least one hierarchical membership degree; wherein, S102 includes the following steps: Obtain at least two pre-constructed hierarchical membership functions for the hierarchical index, and at least two of the hierarchical membership functions are used to define the hierarchical membership for different degrees of ice accumulation; At least two hierarchical membership degrees are obtained by using the hierarchical index and the hierarchical membership function respectively; S103, Obtain the cloud top temperature during the flight process; S104, determine the cloud top temperature membership degree based on the cloud top temperature and the corresponding cloud top temperature membership function; S105, at least two sets of membership degrees are formed based on at least two hierarchical membership degrees and the cloud top temperature membership degree; S106, calculate at least two initial ice accumulation potentials based on at least two sets of membership degrees and initial ice accumulation potential models respectively; S107, at least two final ice accumulation potentials are calculated based on the at least two initial ice accumulation potentials respectively; S108, output the corresponding graded forecast results based on at least two of the final ice accumulation potentials.

2. The method according to claim 1, characterized in that, It also includes the following steps: S109, Construct the hierarchical membership function based on historical flight data, wherein the historical flight data includes: temperature and / or relative humidity.

3. The method according to claim 2, characterized in that, S109 includes the following steps: S1091, the historical flight data is divided into at least two datasets according to the degree of historical icing; S1092, a hierarchical membership function is obtained by fitting the dataset. The hierarchical membership function is a function used to describe the changes in hierarchical indicators and the probability of icing, and the probability of icing = the number of icing events that occur at the temperature / the number of all historical flight data at the temperature.

4. The method according to claim 3, characterized in that, When the grading index is the temperature, S109 further includes the step: S1093, determining whether the grading membership function needs to be corrected; S1093 includes: S10931, the key verification area is set by the user; S10932, identify whether an abnormal peak appears in the key verification area; wherein, when the abnormal probability corresponding to the peak is greater than the first probability value and less than the second probability value, the peak is identified as an abnormal peak. S10933 is a signal for suggesting corrections to the abnormal peak.

5. The method according to claim 4, characterized in that, Prior to S10933, it also includes: Obtain the number of historical flight data corresponding to the temperature of the abnormal peak; When the quantity is less than or equal to a preset quantity threshold, then proceeding to S10933 is permitted.

6. The method according to claim 4, characterized in that, Prior to S10933, it also includes: Identify the number of abnormal peaks; When the number of anomalies exceeds a preset anomaly threshold, then proceeding to step S10933 is permitted.

7. A graded forecasting system for ice accumulation, characterized in that, include: The grading index acquisition module is used to acquire grading indexes during flight, including temperature and relative humidity. A hierarchical processing module is used to perform hierarchical processing on the hierarchical indicators to obtain at least one hierarchical membership degree; wherein, the hierarchical processing module includes: The function construction unit is used to obtain at least two pre-constructed hierarchical membership functions for the hierarchical index, and the at least two hierarchical membership functions are used to define the hierarchical membership of different degrees of ice accumulation. The membership degree acquisition unit is used to obtain at least two hierarchical membership degrees by using the hierarchical index and the hierarchical membership degree function. A cloud top temperature acquisition module is used to acquire the cloud top temperature during the flight process; The cloud top membership determination module is used to determine the cloud top temperature membership degree based on the cloud top temperature and the corresponding cloud top temperature membership function. A membership degree combination module is used to form at least two sets of membership degrees based on at least two hierarchical membership degrees and cloud top temperature membership degrees; The first prediction module is used to calculate at least two initial ice accumulation potentials based on at least two sets of membership degrees and initial ice accumulation potential models, respectively. The second prediction module is used to calculate at least two final ice accumulation potentials based on the at least two initial ice accumulation potentials respectively; The graded forecast module is used to output corresponding graded forecast results based on at least two final ice accumulation potentials.

8. The system according to claim 7, characterized in that, Also includes: The membership function construction module is used to construct the hierarchical membership function based on historical flight data, which includes temperature and / or relative humidity.

9. The system according to claim 8, characterized in that, The membership function construction module includes: A dataset partitioning unit is used to divide the historical flight data into at least two datasets based on the degree of historical icing. The hierarchical membership function fitting unit is used to fit a hierarchical membership function to the dataset. The membership function is a function used to describe the changes in hierarchical indicators and the probability of icing, and the probability of icing is equal to the number of icing events that occur at the temperature / the number of all historical flight data at the temperature.

10. The system according to claim 9, characterized in that, When the grading index is the temperature, the membership function construction module further includes: a correction judgment unit, used to determine whether the grading membership function needs to be corrected; The correction judgment unit includes: The verification area setting sub-unit is used for users to set key verification areas; An anomaly identification subunit is used to identify whether an abnormal peak appears in the key verification area; wherein, when the anomaly probability corresponding to the peak is greater than a first probability value and less than a second probability value, the peak is identified as an abnormal peak. A signal generation subunit is used to generate a suggested correction signal for the abnormal peak.

Images