Method and system for evaluating icing risk of fan blade

Abstract

According to the fan blade icing risk assessment method and system provided by the invention, the to-be-processed fan blade icing condition description data is clustered based on the second fan blade icing condition description data to obtain the second clustering result; the reflection characteristic variable quantity of the second fan blade icing condition description data is smaller than the reflection characteristic variable quantity of the first fan blade icing condition description data, and determining the quality arrangement description data of the first fan blade icing quality information based on the first clustering result, and carrying out fusion processing on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result, and evaluating the target clustering result to obtain a fan blade icing risk evaluation result which can guarantee the accuracy and reliability of the fan blade icing risk evaluation result.

Description

Technical Field

[0001] The present application relates to the technical field of risk assessment, and more specifically, to a method and system for assessing the icing risk of wind turbine blades. Background Art

[0002] Icing on the blades will increase the weight and shape change of the blades, forming unbalanced loads, resulting in a reduction in the power generation of the wind turbine, and may even damage the blades. In addition, the ice cubes thrown out when the blades rotate may pose safety hazards to nearby personnel and property. Therefore, it is necessary to evaluate the icing risk, prevent and control the icing of wind turbine blades in advance, and improve the power generation efficiency of the wind turbine. Therefore, there is an urgent need for a technical solution for assessing the icing risk of wind turbine blades to improve the above technical problems. Summary of the Invention

[0003] To improve the technical problems existing in the related art, the present application provides a method and system for assessing the icing risk of wind turbine blades.

[0004] In a first aspect, there is provided a method for assessing the icing risk of wind turbine blades, including: Obtaining the description data of the icing situation of the wind turbine blades to be processed, where the description data of the icing situation of the wind turbine blades to be processed includes the first icing mass information of the wind turbine blades and the second icing mass information of the wind turbine blades, and the icing risk hazard level of the first icing mass information of the wind turbine blades is greater than that of the second icing mass information of the wind turbine blades; Determining the description data of the icing situation of the first wind turbine blades, and performing clustering processing on the description data of the icing situation of the wind turbine blades to be processed based on the description data of the icing situation of the first wind turbine blades to obtain a first clustering result; Determining the description data of the icing situation of the second wind turbine blades, and performing clustering processing on the description data of the icing situation of the wind turbine blades to be processed based on the description data of the icing situation of the second wind turbine blades to obtain a second clustering result, where the change amount of the reflection characteristics of the description data of the icing situation of the second wind turbine blades is less than that of the description data of the icing situation of the first wind turbine blades; Determining the quality arrangement description data of the first icing mass information of the wind turbine blades based on the first clustering result, and performing fusion processing on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result; Evaluating the target clustering result to obtain the assessment result of the icing risk of the wind turbine blades.

[0005] In the present application, the determining the description data of the icing situation of the first wind turbine blades includes: When the data describing the ice accretion situation of the to-be-processed wind turbine blade is the current data describing the ice accretion situation of the wind turbine blade, the data describing the ice accretion situation of the to-be-processed wind turbine blade is parsed to obtain a parsing result; Based on the parsing result, determine the maximum unit dynamic load value, the minimum unit dynamic load value, and the number of influencing factors of the wind turbine blade ice accretion corresponding to each unit dynamic load value; Based on the maximum unit dynamic load value and the minimum unit dynamic load value, determine a plurality of target parsing regions; Based on the number of influencing factors of the wind turbine blade ice accretion corresponding to each unit dynamic load value, determine the global number of influencing factor segments of the wind turbine blade ice accretion corresponding to each target parsing region; Based on the global number of influencing factor segments of the wind turbine blade ice accretion corresponding to each target parsing region, determine the first data describing the ice accretion situation of the wind turbine blade.

[0006] It should be understood that by accurately determining a plurality of target parsing regions, the first data describing the ice accretion situation of the wind turbine blade can be reliably determined.

[0007] In this application, the determining of a plurality of target parsing regions based on the maximum unit dynamic load value and the minimum unit dynamic load value includes: Obtain a constraint boundary for dividing the parsing region; Divide a plurality of parsing regions from the maximum unit dynamic load value to the minimum unit dynamic load value according to the constraint boundary; Take the parsing regions as the target parsing regions; Alternatively, select a plurality of target parsing regions from the plurality of parsing regions according to a preset region interval.

[0008] It should be understood that through the constraint boundary of the maximum unit dynamic load value and the minimum unit dynamic load value, a plurality of target parsing regions can be accurately determined.

[0009] In this application, the determining of the first data describing the ice accretion situation of the wind turbine blade based on the global number of influencing factor segments of the wind turbine blade ice accretion corresponding to each target parsing region includes: Arrange the target parsing regions in descending order according to the maximum unit dynamic load value of the region to obtain the arranged target parsing regions; Based on the arranged target parsing regions, determine the x-th comparison value of the global number of influencing factor segments of the wind turbine blade ice accretion corresponding to the (x + 1)-th target parsing region and the global number of influencing factor segments of the wind turbine blade ice accretion corresponding to the x-th target parsing region; When the x-th comparison value is greater than a preset comparison value setting, the regional minimum unit dynamic load value corresponding to the x-th target analysis region is used as the description data of the icing condition of the first wind turbine blade, where x is an integer between 1 and (M - 1), and M is the global number of target analysis regions.

[0010] It should be understood that when the global number of influencing factor segments of the icing of the wind turbine blades corresponding to each target analysis region is passed, the problem of uncertainty in each target analysis region is improved, so that the description data of the icing condition of the first wind turbine blade can be accurately determined.

[0011] In this application, the determination of the description data of the icing condition of the second wind turbine blade includes: Taking each unit dynamic load value as a candidate setting value in sequence from the minimum unit dynamic load value to the maximum unit dynamic load value; Determining a plurality of important influencing factor segments of the icing of the wind turbine blades and a plurality of potential influencing factor segments of the icing of the wind turbine blades corresponding to each candidate setting value; Determining the function processing result corresponding to each candidate setting value based on the unit dynamic load value corresponding to each important influencing factor segment of the icing of the wind turbine blades and the unit dynamic load value corresponding to each potential influencing factor segment of the icing of the wind turbine blades; Taking the candidate setting value corresponding to the maximum function processing result as the description data of the icing condition of the second wind turbine blade.

[0012] It should be understood that by accurately determining a plurality of important influencing factor segments of the icing of the wind turbine blades and a plurality of potential influencing factor segments of the icing of the wind turbine blades, the description data of the icing condition of the second wind turbine blade can be reliably determined.

[0013] In this application, the determination of the function processing result corresponding to each candidate setting value based on the unit dynamic load value corresponding to each important influencing factor segment of the icing of the wind turbine blades and the unit dynamic load value corresponding to each potential influencing factor segment of the icing of the wind turbine blades includes: Determining an important feature ratio and a potential feature ratio based on the global number of important influencing factor segments of the icing of the wind turbine blades and the global number of potential influencing factor segments of the icing of the wind turbine blades, and determining an important analysis average value based on the unit dynamic load value corresponding to each important influencing factor segment of the icing of the wind turbine blades and the global number of important influencing factor segments of the icing of the wind turbine blades; Determining a potential analysis average value based on the unit dynamic load value corresponding to each potential influencing factor segment of the icing of the wind turbine blades and the global number of potential influencing factor segments of the icing of the wind turbine blades; Determine the average value of the data analysis of the icing condition description of the wind turbine blade for the parsing result; Based on the average value of the icing condition description data of the wind turbine blade, the important feature ratio, the potential feature ratio, the important parsing average value, and the potential parsing average value, determine the function processing result corresponding to the candidate setting value.

[0014] It should be understood that when based on the unit dynamic load values corresponding to the influencing factor segments of the icing of the blades of each important wind turbine and the unit dynamic load values corresponding to the influencing factor segments of the icing of the blades of each potential wind turbine, the problem of inaccurate important parsing average value is improved, so that the function processing results corresponding to each candidate setting value can be accurately determined.

[0015] In this application, the quality arrangement description data for determining the first wind turbine blade icing quality information based on the first clustering result includes: Perform a filtering process on the first clustering result to obtain a filtered clustering result; Obtain the risk clusters in the filtered clustering result; Based on the risk clusters, combine the regional coefficients corresponding to the first wind turbine blade icing quality information; Based on the regional coefficients and the risk clusters, determine the quality arrangement description data of the first wind turbine blade icing quality information.

[0016] It should be understood that when based on the first clustering result, the problem of inaccurate clustering result filtering is improved, so that the quality arrangement description data of the first wind turbine blade icing quality information can be accurately determined.

[0017] In this application, the quality arrangement description data for determining the first wind turbine blade icing quality information based on the regional coefficients and the risk clusters includes: Obtain the blade load data corresponding to the first wind turbine blade icing quality information; Based on the regional coefficients, the risk clusters, and the blade load data, generate adaptive threshold reference wind turbine blade icing condition description data, where the influencing factor segments of the icing of the wind turbine blades within the first wind turbine blade icing quality information are of the first attribute, and the influencing factor segments of the icing of the wind turbine blades outside the first wind turbine blade icing quality information are of the second attribute; Based on the adaptive threshold reference wind turbine blade icing condition description data, determine the significant difference data of the influencing factor segments of the icing of each wind turbine blade within the first wind turbine blade icing quality information; Use the significant difference data of the influencing factor segments of the icing of each wind turbine blade as the quality arrangement description data of the icing quality information of the first wind turbine blade.

[0018] It should be understood that when based on the regional coefficient and the risk cluster, the problem of inaccurate second attributes of the influencing factor segments of the icing of the wind turbine blade is improved, so that the quality arrangement description data of the icing quality information of the first wind turbine blade can be accurately determined.

[0019] In this application, the fusion processing of the first clustering result and the second clustering result based on the quality arrangement description data to obtain the target clustering result includes: Obtain the significant difference data of the influencing factor segments of the icing of the y-th wind turbine blade in the second clustering result, where y = 1, 2,..., N, and N is the global number of the influencing factor segments of the icing of the wind turbine blade in the second clustering result; When it is determined that the influencing factor segment of the icing of the y-th wind turbine blade is located within the icing quality information of the first wind turbine blade based on the significant difference data of the influencing factor segment of the icing of the y-th wind turbine blade and the description data of the icing situation of the adaptive threshold reference wind turbine blade, obtain the first influencing factor parameter of the influencing factor segment of the icing of the y-th wind turbine blade in the first clustering result; Convert the influencing factor parameter of the influencing factor segment of the icing of the y-th wind turbine blade in the second clustering result into the first influencing factor parameter; When it is determined that the influencing factor segment of the icing of the y-th wind turbine blade is located outside the icing quality information of the first wind turbine blade, keep the influencing factor parameter of the influencing factor segment of the icing of the y-th wind turbine blade unchanged, and use the processed second clustering result as the target clustering result.

[0020] It should be understood that when the first clustering result and the second clustering result are fused based on the quality arrangement description data, the problem of inaccurate influencing factor parameters is improved, so that the target clustering result can be accurately obtained.

[0021] In this application, when the description data of the icing situation of the wind turbine blade to be processed includes a directory to be processed, the evaluation of the target clustering result to obtain the risk assessment result of the icing of the wind turbine blade includes: Perform directory position processing on the target clustering result to obtain the quality arrangement description data of the directory; Obtain the description data of the icing situation of the directory wind turbine blade from the target clustering result based on the quality arrangement description data of the directory; Use a preset artificial intelligence blade risk assessment thread to evaluate the data describing the icing condition of the directory fan blade, and obtain the risk assessment result of the fan blade icing.

[0022] It should be understood that when the data describing the icing condition of the fan blade to be processed includes the directory to be processed, when evaluating the target clustering result, the problem of inaccurate quality arrangement description data is improved, so that the risk assessment result of the fan blade icing can be accurately obtained.

[0023] In a second aspect, an evaluation system for the risk of fan blade icing is provided, including a processor and a memory that communicate with each other. The processor is configured to read and execute a computer program from the memory to implement the above method.

[0024] An evaluation method and system for the icing risk of a wind turbine blade. After obtaining the to-be-processed wind turbine blade icing condition description data including the icing mass information of the first wind turbine blade, first determine the icing condition description data of the first wind turbine blade, and perform clustering processing on the to-be-processed wind turbine blade icing condition description data based on the icing condition description data of the first wind turbine blade to obtain a first clustering result. The icing mass information of the first wind turbine blade is distinguished from the icing mass information of the second wind turbine blade through the icing condition description data of the first wind turbine blade. Then, determine the icing condition description data of the second wind turbine blade, and perform clustering processing on the to-be-processed wind turbine blade icing condition description data based on the icing condition description data of the second wind turbine blade to obtain a second clustering result. The change amount of the reflection feature of the icing condition description data of the second wind turbine blade is less than that of the icing condition description data of the first wind turbine blade. Then, based on the first clustering result, determine the quality arrangement description data of the icing mass information of the first wind turbine blade, and perform fusion processing on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result. Then, evaluate the target clustering result to obtain the wind turbine blade icing risk evaluation result. Since the to-be-processed wind turbine blade icing condition description data includes the icing mass information of the first wind turbine blade, if only the lower icing condition description data of the second wind turbine blade is used to perform clustering processing on the to-be-processed wind turbine blade icing condition description data, then the icing mass information of the first wind turbine blade will all adaptively threshold to the first attribute, resulting in the loss of data in the icing mass information of the first wind turbine blade. Therefore, in the embodiments of the present disclosure, the higher icing condition description data of the first wind turbine blade is preferentially used to perform clustering processing on the to-be-processed wind turbine blade icing condition description data, so as to distinguish the icing mass information of the first wind turbine blade in the to-be-processed wind turbine blade icing condition description data, and the attributes in the icing mass information of the first wind turbine blade can also be retained. Then, the lower icing condition description data of the second wind turbine blade is used to perform clustering processing on the to-be-processed wind turbine blade icing condition description data to perform clustering processing on other regions except the icing mass information of the first wind turbine blade (at this time, the icing mass information of the first wind turbine blade will all adaptively threshold to the first attribute). Finally, the first clustering result and the second clustering result are fused based on the quality arrangement description data of the icing mass information of the first wind turbine blade to obtain a target clustering result, so that the target clustering result includes the attribute information of the icing mass information of the first wind turbine blade and the icing mass information of the second wind turbine blade, ensuring the integrity of the target clustering result. Furthermore, using the target clustering result to evaluate the wind turbine blade icing condition description data can improve the evaluation accuracy and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] To more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings required for the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0026] Figure 1 It is a flowchart of an evaluation method for the icing risk of a wind turbine blade provided by an embodiment of the present application. Detailed implementation manners

[0027] To better understand the above technical solutions, the technical solutions of the present application will be described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solutions of the present application, rather than limitations on the technical solutions of the present application. Without conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0028] Please refer to Figure 1 , which shows an evaluation method for the icing risk of a wind turbine blade. The method may include the technical solutions described in the following steps S201 - step S205.

[0029] Step S201, obtain the description data of the icing situation of the wind turbine blade to be processed.

[0030] Among them, the description data of the icing situation of the wind turbine blade to be processed can be understood as the icing situation (such as the icing position and icing thickness, etc.) and the situation of the wind turbine blade (such as the surface roughness, etc.). After the blade is iced, its aerodynamic characteristics will change, resulting in an increase in the dynamic load of the unit, which will in turn have an adverse impact on the safe operation and economic benefits of the wind farm. Different climate environments and terrain conditions will cause the wind turbine blades to form ice with different textures and shapes during operation, and there are also significant differences in the icing processes and characteristics in different regions. In the embodiments of the present disclosure, the description data of the icing situation of the wind turbine blade to be processed can be the description data of the current icing situation of the wind turbine blade, or the analysis result.

[0031] For example, analyze the basic meteorological conditions for the formation of blade icing. Understanding the meteorological prediction and operation status of the wind farm is crucial for preventing blade icing. Generally, the following conditions need to be met for blade icing: the environmental temperature is lower than 0 °C, the surface temperature of the blade is lower than -5 °C, and the air humidity is higher than 85%. Although the range of these conditions is relatively narrow, wind farms are often located in areas where cold and hot air currents in the atmosphere meet, and the influence of micro-topography and micro-meteorology makes the blade icing phenomenon of some wind farm units relatively common.

[0032] Step S202: Determine the description data of the icing condition of the first wind turbine blade, and perform clustering processing on the description data of the icing condition of the to-be-processed wind turbine blade based on the description data of the icing condition of the first wind turbine blade to obtain a first clustering result.

[0033] For example, the description data of the icing condition of the wind turbine blade can be understood as clustering the description data of the icing condition of the to-be-processed wind turbine blade. The specific classifications can be glaze ice, rime ice, and mixed ice.

[0034] In this step, the determined description data of the icing condition of the first wind turbine blade is the analytical set value that can separate the icing quality information of the first wind turbine blade from the icing quality information of the second wind turbine blade in the description data of the icing condition of the to-be-processed wind turbine blade. When determining the description data of the icing condition of the first wind turbine blade, the description data of the icing condition of the to-be-processed wind turbine blade can be first analyzed to obtain an analysis result, then the number of influencing factors of the ice-covered wind turbine blade corresponding to each unit dynamic load value is counted, and then multiple analysis regions are divided from the highest unit dynamic load value to the lowest unit dynamic load value among each unit dynamic load value, and multiple target analysis regions are selected from them. The description data of the icing condition of the first wind turbine blade is determined by using the comparison value of the global quantity target of adjacent target analysis regions.

[0035] When performing clustering processing on the description data of the icing condition of the to-be-processed wind turbine blade based on the description data of the icing condition of the first wind turbine blade, the unit dynamic load value of the influencing factor segment of the ice-covered wind turbine blade with a unit dynamic load value greater than the description data of the icing condition of the first wind turbine blade in the analysis result corresponding to the description data of the icing condition of the to-be-processed wind turbine blade is set to a first preset value, and the unit dynamic load value of the influencing factor segment of the ice-covered wind turbine blade with a unit dynamic load value lower than or equal to the description data of the icing condition of the first wind turbine blade is set to a second preset value. The first preset value and the second preset value are different, and the comparison value of the first preset value and the second preset value is greater than a certain analytical set value.

[0036] Step S203: Determine the description data of the icing condition of the second wind turbine blade, and perform clustering processing on the description data of the icing condition of the to-be-processed wind turbine blade based on the description data of the icing condition of the second wind turbine blade to obtain a second clustering result. The change amount of the reflection characteristic of the description data of the icing condition of the second wind turbine blade is less than the change amount of the reflection characteristic of the description data of the icing condition of the first wind turbine blade.

[0037] Among them, the accuracy of the change amount of the reflection characteristic of the description data of the icing condition of the second wind turbine blade is greater than the accuracy of the change amount of the reflection characteristic of the description data of the icing condition of the first wind turbine blade.

[0038] This step is being implemented. Each unit dynamic load value between the minimum unit dynamic load value and the maximum unit dynamic load value in the parsing result corresponding to the data describing the icing condition of the fan blade to be processed can be used as a candidate setting value. Then, the important and potential function processing results are determined, and the candidate setting value corresponding to the maximum function processing result is used as the second data describing the icing condition of the fan blade.

[0039] When clustering the data describing the icing condition of the fan blade to be processed based on the second data describing the icing condition of the fan blade, the process is similar to that of step S202 for clustering based on the first data describing the icing condition of the fan blade. In the parsing result corresponding to the data describing the icing condition of the fan blade to be processed, the unit dynamic load value of the influencing factor segment of the wind turbine blade icing with a unit dynamic load value greater than that of the second data describing the icing condition of the fan blade is set to the first preset value, and the unit dynamic load value of the influencing factor segment of the wind turbine blade icing with a unit dynamic load value lower than or equal to that of the second data describing the icing condition of the fan blade is set to the second preset value.

[0040] Step S204, determining the quality arrangement description data of the first fan blade icing quality information based on the first clustering result, and performing a fusion process on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result.

[0041] For example, fusion can be understood as fusion and combination, etc.

[0042] When this step is implemented, first, the significant difference data of the risk cluster is determined based on the second clustering result. Then, the risk cluster is used for graphic combination to obtain an area coefficient that can characterize the shape of the first fan blade icing quality information. Using this area coefficient and the risk cluster, an adaptive threshold reference data describing the icing condition of the fan blade can be generated. In the adaptive threshold reference data describing the icing condition of the fan blade, the influencing factor segment of the wind turbine blade icing within the first fan blade icing quality information is the first attribute, and the influencing factor segment of the wind turbine blade icing outside the first fan blade icing quality information is the second attribute. Then, taking the adaptive threshold reference data describing the icing condition of the fan blade as a reference, when it is determined that the influencing factor segment of a certain wind turbine blade icing is within the first fan blade icing quality information, the influencing factor parameter of the influencing factor segment of the wind turbine blade icing in the second clustering result is converted into the influencing factor parameter of the influencing factor segment of the wind turbine blade icing in the first clustering result. After traversing all the influencing factor segments of the wind turbine blade icing in the second clustering result, the target clustering result is obtained.

[0043] Among them, the influencing factor segment of the wind turbine blade icing can be understood as the influencing factor attribute of the wind turbine blade icing.

[0044] Step S205: Evaluate the target clustering result to obtain the icing risk assessment result of the wind turbine blade.

[0045] When the data for describing the icing condition of the wind turbine blade to be processed includes a directory to be processed, in implementing this step, the directory location may first be determined based on the target clustering result, so as to determine the quality arrangement description data of the directory. Then, based on the quality arrangement description data of the directory, the data for describing the icing condition of the wind turbine blade of the directory is determined. Finally, the data for describing the icing condition of the wind turbine blade of the directory is decoded by using a preset artificial intelligence blade risk assessment thread, so as to obtain the icing risk assessment result of the wind turbine blade. The icing risk assessment result of the wind turbine blade may include the user identification information corresponding to the evaluation code. Among them, the artificial intelligence blade risk assessment thread in this application covers the functions of neural network, genetic algorithm, deep learning, reinforcement learning, machine learning, and ensemble learning, and can analyze various landslide factors in the data for describing the icing condition of the wind turbine blade of the directory. Therefore, the artificial intelligence blade risk assessment thread in this application has the functions of the following threads: Feedforward Neural Networks (FNN), Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Support Vector Machines (SVM), K-Means Clustering, and Hierarchical Clustering, etc. The applicant of this application has absorbed the advantages of the above models and improved them to obtain an artificial intelligence thread applicable to this field.

[0046] When the data for describing the icing condition of the wind turbine blade to be processed is other data for describing the icing condition of the wind turbine blade that does not include a directory to be processed, in implementing this step, the constraint data of the matters to be processed included in the data for describing the icing condition of the wind turbine blade to be processed may be determined based on the target clustering result, and then an evaluation is performed based on the constraint data to obtain the icing risk assessment result of the wind turbine blade.

[0047] In the evaluation method provided by the embodiments of the present disclosure, after obtaining the to-be-processed fan blade icing condition description data including the first fan blade icing mass information, first determine the first fan blade icing condition description data, and perform clustering processing on the to-be-processed fan blade icing condition description data based on the first fan blade icing condition description data to obtain a first clustering result. The first fan blade icing mass information is distinguished from the second fan blade icing mass information through the first fan blade icing condition description data. Then, determine the second fan blade icing condition description data, and perform clustering processing on the to-be-processed fan blade icing condition description data based on the second fan blade icing condition description data to obtain a second clustering result. The change amount of the reflection feature of the second fan blade icing condition description data is less than that of the first fan blade icing condition description data. Since the to-be-processed fan blade icing condition description data includes the first fan blade icing mass information with a relatively high icing risk hidden danger level, if only the lower second fan blade icing condition description data is used to perform clustering processing on the to-be-processed fan blade icing condition description data, then all the first fan blade icing mass information will adaptively threshold to the first attribute, resulting in data loss in the first fan blade icing mass information. Therefore, in the embodiments of the present disclosure, preferably, the higher first fan blade icing condition description data is used to perform clustering processing on the to-be-processed fan blade icing condition description data, so as to distinguish the first fan blade icing mass information in the to-be-processed fan blade icing condition description data, and the attributes in the first fan blade icing mass information can also be retained. Then, the lower second fan blade icing condition description data is used to perform clustering processing on the to-be-processed fan blade icing condition description data to perform clustering processing on other areas except the first fan blade icing mass information (at this time, the first fan blade icing mass information will all adaptively threshold to the first attribute). Finally, the first clustering result and the second clustering result are fused based on the quality arrangement description data of the first fan blade icing mass information to obtain a target clustering result, so that the target clustering result includes the attribute information of the first fan blade icing mass information and the second fan blade icing mass information, ensuring the integrity of the target clustering result. Furthermore, using the target clustering result to evaluate the fan blade icing condition description data can improve the evaluation efficiency.

[0048] For some possible embodiments, the above step S202 "determine the first fan blade icing condition description data" can be implemented through the following steps S2021 to S2025. The following explains each step.

[0049] Step S2021, when the data describing the ice accretion situation of the to-be-processed wind turbine blade is the current data describing the ice accretion situation of the wind turbine blade, perform parsing processing on the data describing the ice accretion situation of the to-be-processed wind turbine blade to obtain a parsing result.

[0050] Step S2022, based on the parsing result, determine the maximum unit dynamic load value, the minimum unit dynamic load value, and the number of influencing factors of ice accretion on the wind turbine blade corresponding to each unit dynamic load value.

[0051] After obtaining the parsing result of the data describing the ice accretion situation of the to-be-processed wind turbine blade, traverse each influencing factor segment of ice accretion on the wind turbine blade in the parsing result to determine the maximum unit dynamic load value, the minimum unit dynamic load value, and the number of influencing factors of ice accretion on the wind turbine blade corresponding to each unit dynamic load value.

[0052] For some possible embodiments, it is also possible to generate a parsing histogram based on the number of influencing factors of ice accretion on the wind turbine blade corresponding to each unit dynamic load value. Through the parsing histogram, the differences in the number of influencing factors of ice accretion on the wind turbine blade corresponding to different unit dynamic load values can be intuitively understood.

[0053] Step S2023, determine a plurality of target parsing regions based on the maximum unit dynamic load value and the minimum unit dynamic load value.

[0054] In the actual application process, the constraint boundaries for dividing the parsing regions can be obtained first, and then a plurality of parsing regions are divided from the maximum unit dynamic load value to the minimum unit dynamic load value according to the constraint boundaries. At this time, there are at least two ways to determine the target parsing regions adaptively: one is that all the directly divided parsing regions are used as the target parsing regions; the other is to select a plurality of target parsing regions from the plurality of parsing regions according to the preset region intervals.

[0055] Step S2024, determine the global number of influencing factor segments of ice accretion on the wind turbine blade corresponding to each target parsing region based on the number of influencing factors of ice accretion on the wind turbine blade corresponding to each unit dynamic load value.

[0056] When implementing this step, the sum of the number of influencing factors of ice accretion on the wind turbine blade corresponding to the unit dynamic load value included in each target parsing region is calculated to obtain the global number of influencing factor segments of ice accretion on the wind turbine blade corresponding to each target parsing region.

[0057] Step S2025, determine the first data describing the ice accretion situation of the wind turbine blade based on the global number of influencing factor segments of ice accretion on the wind turbine blade corresponding to each target parsing region.

[0058] For some possible embodiments, this step can be implemented through the following steps S251 to S253, and each step is described below.

[0059] Step S251: Arrange the respective target analysis regions in descending order according to the maximum unit dynamic load value of the region to obtain the arranged target analysis regions.

[0060] Here, each target analysis region corresponds to a maximum unit dynamic load value of the region and a minimum unit dynamic load value of the region. Arranging the respective target analysis regions in ascending order according to the maximum region analysis value, that is, arranging the respective target analysis regions in the order from bright to dark, to obtain the arranged target analysis regions.

[0061] Step S252: Based on the arranged respective target analysis regions, determine the x-th comparison value of the global quantity of the influencing factor segments of the ice accretion on the wind turbine blade corresponding to the (x + 1)-th target analysis region and the global quantity of the influencing factor segments of the ice accretion on the wind turbine blade corresponding to the x-th target analysis region.

[0062] Wherein, x is an integer between 1 and (M - 1), and M is the global quantity of the target analysis regions. In this step, the comparison value of the global quantity of the influencing factor segments of the ice accretion on the wind turbine blade corresponding to two adjacent target analysis regions is calculated. In implementation, the global quantity of the influencing factor segments of the ice accretion on the wind turbine blade of the (x + 1)-th target analysis region is subtracted from the global quantity of the influencing factor segments of the ice accretion on the wind turbine blade corresponding to the x-th target analysis region to obtain the x-th comparison value.

[0063] Step S253: Determine whether the x-th comparison value is greater than the preset comparison value setting.

[0064] Wherein, since the data describing the ice accretion situation of the wind turbine blade to be processed in the embodiments of the present disclosure includes the ice accretion quality information of the first wind turbine blade in most regions where the ice accretion risk hidden danger level is significantly greater than the data describing the ice accretion situation of the wind turbine blade to be processed, then when the comparison value of the global quantity of the influencing factor segments of the ice accretion on the wind turbine blade corresponding to two adjacent target analysis regions is greater than the preset comparison value setting, it indicates that the distinguishing unit dynamic load value between the ice accretion quality information of the first wind turbine blade and the ice accretion quality information of the second wind turbine blade is reached, and at this time, step S254 is entered; when the x-th comparison value is lower than or equal to the comparison value setting, it indicates that the distinguishing unit dynamic load value between the ice accretion quality information of the first wind turbine blade and the ice accretion quality information of the second wind turbine blade has not been reached, and at this time, step S255 is entered.

[0065] Step S254: Use the minimum unit dynamic load value of the region corresponding to the x-th target analysis region as the data describing the ice accretion situation of the first wind turbine blade.

[0066] Here, when it is determined that the comparison value of the global number of influence factor segments of the ice accretion on the wind turbine blades in two adjacent target analysis regions reaches the set comparison value, it indicates that the dynamic load value of the unit for distinguishing the ice accretion quality information of the first wind turbine blade and the ice accretion quality information of the second wind turbine blade is reached. Then, when describing the ice accretion situation data of the first wind turbine blade, the minimum unit dynamic load value corresponding to the x-th target analysis region is used as the ice accretion situation description data of the first wind turbine blade. That is to say, when the minimum unit dynamic load value corresponding to the previous target analysis region in two adjacent target analysis regions is used as the ice accretion situation description data of the first wind turbine blade.

[0067] Step S255, set x to x + 1, and enter step S252.

[0068] When determining the ice accretion situation description data of the first wind turbine blade for distinguishing the ice accretion quality information of the first wind turbine blade and the ice accretion quality information of the second wind turbine blade, through the above steps S2021 to S2025, first, the ice accretion situation description data to be processed is parsed to obtain a parsing result, and then the number of influence factors of the ice accretion on the wind turbine blades corresponding to each unit dynamic load value in the parsing result is counted, and multiple target analysis spaces are determined by using the maximum unit dynamic load value and the minimum unit dynamic load value of the parsing result. Furthermore, based on the comparison value of the global number of influence factor segments of the ice accretion on the wind turbine blades in adjacent target analysis regions, the ice accretion situation description data of the first wind turbine blade is determined. Since the depolarization ice accretion risk hidden danger level of the ice accretion quality information of the first wind turbine blade is significantly greater than that of the ice accretion quality information of the second wind turbine blade, in the embodiments of the present disclosure, the global number of influence factor segments of the ice accretion on the wind turbine blades is counted starting from the maximum unit dynamic load value forward. When it is determined that the comparison value of the global number of influence factor segments of the ice accretion on the wind turbine blades in adjacent target analysis regions is greater than the preset set comparison value, it indicates that a mutation has occurred in the global number of influence factor segments of the ice accretion on the wind turbine blades in these two adjacent target analysis regions. At this time, it is considered that the dynamic load value of the unit for distinguishing the ice accretion quality information of the first wind turbine blade and the ice accretion quality information of the second wind turbine blade is reached, so as to ensure that the determined ice accretion situation description data of the first wind turbine blade can accurately achieve the adaptive threshold distinction between the ice accretion quality information of the first wind turbine blade and the ice accretion quality information of the second wind turbine blade.

[0069] Regarding some possible embodiments, the "determining the ice accretion situation description data of the second wind turbine blade" in the above step S203 can be implemented through the following steps S2031 to S2034. The following explains each step.

[0070] Step S2031: Take each unit dynamic load value as a candidate setting value in sequence from the minimum unit dynamic load value to the maximum unit dynamic load value.

[0071] Step S2032: Determine multiple influencing factor segments of important wind turbine blade icing and multiple influencing factor segments of potential wind turbine blade icing corresponding to each candidate setting value.

[0072] In the embodiments of the present disclosure, the influencing factor segments of wind turbine blade icing with a unit dynamic load value lower than the candidate setting value are used as the influencing factor segments of potential wind turbine blade icing, and the influencing factor segments of wind turbine blade icing with a unit dynamic load value greater than or equal to the candidate setting value are used as the influencing factor segments of important wind turbine blade icing.

[0073] Step S2033: Determine the function processing results corresponding to each candidate setting value based on the unit dynamic load values corresponding to the influencing factor segments of important wind turbine blade icing and the unit dynamic load values corresponding to the influencing factor segments of potential wind turbine blade icing.

[0074] When implementing this step, first calculate the potential feature ratio and important feature ratio under each candidate setting value, and calculate the potential analysis average value and important analysis average value, and then use the variance calculation formula to calculate the important and potential function processing results corresponding to each candidate setting value.

[0075] Step S2034: Take the candidate setting value corresponding to the maximum function processing result as the description data of the icing condition of the second wind turbine blade.

[0076] Since variance is a measure of the uniformity of the analytical arrangement, the larger the function processing result between potential and important, the greater the difference between the two parts that make up the description data of the wind turbine blade icing condition. Therefore, in this step, taking the candidate setting value corresponding to the maximum function processing result as the description data of the second wind turbine blade icing condition can ensure that the important part and the potential part can be distinguished to the greatest extent using this description data of the second wind turbine blade icing condition.

[0077] Regarding some possible embodiments, the above-mentioned step S2033 "Determine the function processing results corresponding to each candidate setting value based on the unit dynamic load values corresponding to the influencing factor segments of important wind turbine blade icing and the unit dynamic load values corresponding to the influencing factor segments of potential wind turbine blade icing" can be implemented through the following steps S331 to S335. The following explains each step.

[0078] Step S331: Determine the important feature ratio and the potential feature ratio based on the global quantity of influence factor segments of important wind turbine blade icing and the global quantity of influence factor segments of potential wind turbine blade icing.

[0079] When implementing this step, first obtain the description data of the icing condition of the wind turbine blade in the parsing result, which is the global quantity of influence factor segments of wind turbine blade icing. Then divide the global quantity of influence factor segments of important wind turbine blade icing by the global quantity of influence factor segments of wind turbine blade icing in the description data of the icing condition of the wind turbine blade to obtain the important feature ratio, and divide the global quantity of influence factor segments of potential wind turbine blade icing by the global quantity of influence factor segments of wind turbine blade icing in the description data of the icing condition of the wind turbine blade to obtain the potential feature ratio. Therefore, both the important feature ratio and the potential feature ratio are real numbers greater than 0 and less than 1, and the sum of the important feature ratio and the potential feature ratio is 1.

[0080] Step S332: Determine the important parsing average value based on the corresponding unit dynamic load values of each influence factor segment of important wind turbine blade icing and the global quantity of influence factor segments of important wind turbine blade icing.

[0081] Here, sum up the corresponding unit dynamic load values of each influence factor segment of important wind turbine blade icing to obtain the important total unit dynamic load value, and then divide the important total unit dynamic load value by the global quantity of influence factor segments of important wind turbine blade icing to obtain the important parsing average value.

[0082] Step S333: Determine the potential parsing average value based on the corresponding unit dynamic load values of each influence factor segment of potential wind turbine blade icing and the global quantity of influence factor segments of potential wind turbine blade icing.

[0083] Here, sum up the corresponding unit dynamic load values of each influence factor segment of potential wind turbine blade icing to obtain the potential total unit dynamic load value, and then divide the potential total unit dynamic load value by the global quantity of influence factor segments of potential wind turbine blade icing to obtain the potential parsing average value.

[0084] Step S334: Determine the parsing average value of the description data of the icing condition of the wind turbine blade in the parsing result.

[0085] When implementing this step, sum up the unit dynamic load values of each influence factor segment of the wind turbine blade icing to obtain the total unit dynamic load value of the description data of the icing condition of the wind turbine blade, and then divide the total unit dynamic load value of the description data of the icing condition of the wind turbine blade by the global quantity of influence factor segments of the wind turbine blade icing in the description data of the icing condition of the wind turbine blade to obtain the parsing average value of the description data of the icing condition of the wind turbine blade.

[0086] Step S335: Determine the function processing result corresponding to the candidate set value based on the average value parsed from the icing condition description data of the fan blade, the important feature ratio, the potential feature ratio, the important parsing average value, and the potential parsing average value.

[0087] In the embodiment where steps S331 to S335 are located, first, the important feature ratio and the potential feature ratio are determined, then the important parsing average value and the potential parsing average value are determined. Furthermore, the function processing result corresponding to each candidate set value is determined by using the variance formula, thereby providing a data basis for determining the icing condition description data of the second fan blade.

[0088] Regarding some possible embodiments, the above step S204, "Based on the first clustering result, determine the quality arrangement description data of the icing quality information of the first fan blade, and based on the quality arrangement description data, perform a fusion process on the first clustering result and the second clustering result to obtain a target clustering result" can be implemented through steps S2041 to S2049.

[0089] Step S2041: Perform a filtering process on the first clustering result to obtain a filtered clustering result.

[0090] Although the icing condition description data of the first fan blade can distinguish the icing quality information of the first fan blade and the icing quality information of the second fan blade in the icing condition description data to be processed, since the icing condition description data of the second fan blade is determined by dividing the target parsing area, if the constraint boundary of the target parsing area is large, it will reduce the accuracy of the icing condition description data of the first fan blade, resulting in interference factors in the first clustering result obtained by performing clustering processing on the icing condition description data of the first fan blade. Therefore, it is necessary to perform a filtering process on the first clustering result to filter out the interference factors in the first clustering result. In the implementation of this step, median filtering can be performed on the first clustering result to obtain a filtered clustering result, or mean filtering can be performed on the first clustering result to obtain a filtered clustering result.

[0091] Step S2042: Obtain the risk clusters in the filtered clustering result.

[0092] In the actual implementation process, in order to improve the traversal efficiency, the risk clusters in the filtered clustering result can also be obtained by presetting the number of rows or columns in each interval.

[0093] Step S2043: Combine the area coefficients corresponding to the icing quality information of the first fan blade based on the risk clusters.

[0094] Step S2044: Determine the quality arrangement description data of the first wind turbine blade icing mass information based on the regional coefficient and the risk cluster.

[0095] Since the edges of the first wind turbine blade icing mass information are generally incomplete in the first clustering result obtained by clustering the to-be-processed wind turbine blade icing condition description data with the first wind turbine blade icing condition description data, after determining the regional coefficient, an adaptive threshold reference wind turbine blade icing condition description data with a complete edge can be drawn based on the regional coefficient. In the embodiments of the present disclosure, the quality arrangement description data of the first wind turbine blade icing mass information may be the significant difference data of the influencing factor segments of the icing of each wind turbine blade in the first wind turbine blade icing mass information. Therefore, after determining the significant difference data of each first attribute in the adaptive threshold reference wind turbine blade icing condition description data, the quality arrangement description data of the first wind turbine blade icing mass information is obtained.

[0096] Step S2045: Obtain the significant difference data of the influencing factor segment of the icing of the y-th wind turbine blade in the second clustering result.

[0097] Where y = 1, 2, …, N, and N is the global quantity of the influencing factor segments of the icing of the wind turbine blades in the second clustering result.

[0098] Step S2046: Determine whether the influencing factor segment of the icing of the y-th wind turbine blade is within the first wind turbine blade icing mass information based on the significant difference data of the influencing factor segment of the icing of the y-th wind turbine blade and the adaptive threshold reference wind turbine blade icing condition description data.

[0099] When implementing this step, based on the significant difference data of the influencing factor segment of the icing of the y-th wind turbine blade, the dynamic load value of the unit of the influencing factor segment of the icing of the y-th wind turbine blade in the adaptive threshold reference wind turbine blade icing condition description data can be obtained. If the dynamic load value of the unit of the influencing factor segment of the icing of the y-th wind turbine blade in the adaptive threshold reference wind turbine blade icing condition description data indicates that the influencing factor segment of the icing of the y-th wind turbine blade is the first attribute in the adaptive threshold reference wind turbine blade icing condition description data, it means that the influencing factor segment of the icing of the y-th wind turbine blade is within the first wind turbine blade icing mass information, and at this time, step S2047 is entered; if the dynamic load value of the unit of the influencing factor segment of the icing of the y-th wind turbine blade in the adaptive threshold reference wind turbine blade icing condition description data indicates that the influencing factor segment of the icing of the y-th wind turbine blade is the second attribute in the adaptive threshold reference wind turbine blade icing condition description data, it means that the influencing factor segment of the icing of the y-th wind turbine blade is outside the first wind turbine blade icing mass information, and at this time, step S2049 is entered.

[0100] Step S2047, obtain the first influencing factor parameter of the influencing factor segment of the y-th wind turbine blade icing in the first clustering result.

[0101] When implemented, based on the significant difference data of the influencing factor segment of the y-th wind turbine blade icing, the first influencing factor parameter of the influencing factor segment of the y-th wind turbine blade icing in the first clustering result can be obtained.

[0102] Step S2048, convert the influencing factor parameter of the influencing factor segment of the y-th wind turbine blade icing in the second clustering result into the first influencing factor parameter.

[0103] Step S2049, keep the influencing factor parameter of the influencing factor segment of the y-th wind turbine blade icing unchanged.

[0104] Step S20410, use the processed second clustering result as the target clustering result.

[0105] In the embodiments of the present disclosure, since the second clustering result is obtained by clustering the to-be-processed wind turbine blade icing condition description data based on the second wind turbine blade icing condition description data, for the first wind turbine blade icing quality information with a relatively high risk of icing hazards, in the second clustering result, the adaptive threshold is the first attribute, which may cause data loss in the first wind turbine blade icing quality information. The first clustering result is obtained by clustering the to-be-processed wind turbine blade icing condition description data based on the first wind turbine blade icing condition description data that is greater than the second wind turbine blade icing condition description data. Therefore, the attribute part of the first wind turbine blade icing quality information is retained in the first clustering result. However, most of the second wind turbine blade icing quality information will be adaptively thresholded as the second attribute. Therefore, when fusing the first clustering result and the second clustering result, convert the influencing factor parameter of the influencing factor segment of the wind turbine blade icing at the position corresponding to the first wind turbine blade icing quality information in the second clustering result into the influencing factor segment of the corresponding wind turbine blade icing in the first clustering result. In this way, the target clustering result can include both the attribute information in the highlighted first wind turbine blade icing quality information and the attribute information in the non-highlighted area.

[0106] Regarding some possible embodiments, the above step S2044 "determine the quality arrangement description data of the first wind turbine blade icing quality information based on the region coefficient and the risk cluster" can be implemented through the following steps S441 to S444. The following explains each step.

[0107] Step S441: Obtain the blade load data corresponding to the icing mass information of the first wind turbine blade.

[0108] In the embodiment of the present disclosure, the blade load data corresponding to the icing mass information of the first wind turbine blade can be pre-set, or can be determined according to the length and / or width of the icing mass information of the first wind turbine blade. For example, the blade load data corresponding to the icing mass information of the first wind turbine blade can be obtained by multiplying the length of the icing mass information of the first wind turbine blade by a preset coefficient.

[0109] Step S442: Generate adaptive threshold reference wind turbine blade icing condition description data based on the region coefficient, the risk cluster, and the blade load data.

[0110] Take the region coefficient, the risk cluster, and the blade load data as the output of this function, and the output is the adaptive threshold reference wind turbine blade icing condition description data. In this adaptive threshold reference wind turbine blade icing condition description data, the influencing factor segment of the wind turbine blade icing within the icing mass information of the first wind turbine blade is the first attribute, and the influencing factor segment of the wind turbine blade icing outside the icing mass information of the first wind turbine blade is the second attribute.

[0111] Step S443: Determine the significant difference data of each influencing factor segment of the wind turbine blade icing within the icing mass information of the first wind turbine blade based on the adaptive threshold reference wind turbine blade icing condition description data.

[0112] Here, by traversing each influencing factor segment of the wind turbine blade icing in the adaptive threshold reference wind turbine blade icing condition description data, the influencing factor segment of the white wind turbine blade icing in the adaptive threshold reference wind turbine blade icing condition description data can be obtained. Since the influencing factor segment of the white wind turbine blade icing in the adaptive threshold reference wind turbine blade icing condition description data is the influencing factor segment of the wind turbine blade icing within the icing mass information of the first wind turbine blade, the significant difference data of the influencing factor segment of the white wind turbine blade icing in the adaptive threshold reference wind turbine blade icing condition description data is used as the significant difference data of each influencing factor segment of the wind turbine blade icing within the icing mass information of the first wind turbine blade.

[0113] Step S444: Take the significant difference data of each influencing factor segment of the wind turbine blade icing as the quality arrangement description data of the icing mass information of the first wind turbine blade.

[0114] Through the above steps S441 to S444, it is possible to generate, based on the risk clusters in the first clustering result, the regional coefficient of the first wind turbine blade icing quality information, and the blade load data, the adaptive threshold reference wind turbine blade icing condition description data including the complete first wind turbine blade icing quality information corresponding to the regional coefficient. Furthermore, it is possible to determine the quality arrangement description data of the first wind turbine blade icing quality information based on the adaptive threshold reference wind turbine blade icing condition description data, providing reference data for the subsequent fusion processing of the first clustering result and the second clustering result.

[0115] On this basis, an evaluation device for the icing risk of wind turbine blades is provided. The device includes: A data acquisition module, configured to acquire the to-be-processed wind turbine blade icing condition description data, where the to-be-processed wind turbine blade icing condition description data includes the first wind turbine blade icing quality information and the second wind turbine blade icing quality information, and the icing risk potential level of the first wind turbine blade icing quality information is greater than that of the second wind turbine blade icing quality information; A first data determination module, configured to determine the first wind turbine blade icing condition description data, and perform clustering processing on the to-be-processed wind turbine blade icing condition description data based on the first wind turbine blade icing condition description data to obtain a first clustering result; A second data determination module, configured to determine the second wind turbine blade icing condition description data, and perform clustering processing on the to-be-processed wind turbine blade icing condition description data based on the second wind turbine blade icing condition description data to obtain a second clustering result, where the change amount of the reflection feature of the second wind turbine blade icing condition description data is less than that of the first wind turbine blade icing condition description data; A result clustering module, configured to determine the quality arrangement description data of the first wind turbine blade icing quality information based on the first clustering result, and perform fusion processing on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result; A result evaluation module, configured to evaluate the target clustering result to obtain the wind turbine blade icing risk evaluation result.

[0116] On this basis, an evaluation system for the icing risk of wind turbine blades is shown, including a processor and a memory that communicate with each other. The processor is configured to read and execute a computer program from the memory to implement the above method.

[0117] On this basis, a computer-readable storage medium is further provided, and the computer program stored thereon implements the above method when running.

[0118] In summary, based on the above solution, after obtaining the to-be-processed wind turbine blade icing condition description data including the icing mass information of the first wind turbine blade, first determine the icing condition description data of the first wind turbine blade, and perform clustering processing on the to-be-processed wind turbine blade icing condition description data based on the icing condition description data of the first wind turbine blade to obtain a first clustering result. The icing mass information of the first wind turbine blade is distinguished from the icing mass information of the second wind turbine blade through the icing condition description data of the first wind turbine blade. Then, determine the icing condition description data of the second wind turbine blade, and perform clustering processing on the to-be-processed wind turbine blade icing condition description data based on the icing condition description data of the second wind turbine blade to obtain a second clustering result. The change amount of the reflection feature of the icing condition description data of the second wind turbine blade is less than that of the icing condition description data of the first wind turbine blade. After that, determine the quality arrangement description data of the icing mass information of the first wind turbine blade based on the first clustering result, and perform fusion processing on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result. Then, evaluate the target clustering result to obtain a wind turbine blade icing risk assessment result. Since the to-be-processed wind turbine blade icing condition description data includes the icing mass information of the first wind turbine blade, if only the lower icing condition description data of the second wind turbine blade is used to perform clustering processing on the to-be-processed wind turbine blade icing condition description data, then the icing mass information of the first wind turbine blade will all adaptively threshold to the first attribute, resulting in data loss in the icing mass information of the first wind turbine blade. Therefore, in the embodiments of the present disclosure, the higher icing condition description data of the first wind turbine blade is preferentially used to perform clustering processing on the to-be-processed wind turbine blade icing condition description data, so as to distinguish the icing mass information of the first wind turbine blade in the to-be-processed wind turbine blade icing condition description data, and the attributes in the icing mass information of the first wind turbine blade can also be retained. Then, the lower icing condition description data of the second wind turbine blade is used to perform clustering processing on the to-be-processed wind turbine blade icing condition description data to perform clustering processing on other areas except the icing mass information of the first wind turbine blade (at this time, the icing mass information of the first wind turbine blade will all adaptively threshold to the first attribute). Finally, the first clustering result and the second clustering result are fused based on the quality arrangement description data of the icing mass information of the first wind turbine blade to obtain a target clustering result, so that the target clustering result includes the attribute information of the icing mass information of the first wind turbine blade and the second wind turbine blade, ensuring the integrity of the target clustering result. Furthermore, using the target clustering result to evaluate the wind turbine blade icing condition description data can improve the evaluation accuracy and reliability.

[0119] It should be understood that the systems and their modules shown above can be implemented in various ways. For example, in some embodiments, the systems and their modules can be implemented based on hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art can understand that the above methods and systems can be implemented using computer-executable instructions and / or included in processor control code. For example, such code is provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The systems and their modules of the present application can be implemented not only by hardware circuits such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips and transistors, or programmable hardware devices such as field programmable gate arrays and programmable logic devices, but also by software implemented by various types of processors, or by a combination of the above hardware circuits and software (e.g., firmware).

[0120] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the possible beneficial effects can be any one or several combinations of the above, or any other possible beneficial effects that can be obtained.

Claims

1. An assessment method for the icing risk of a fan blade, characterized in that, The method includes: Obtaining the icing condition description data of the wind turbine blade to be processed, where the icing condition description data of the wind turbine blade to be processed includes the icing mass information of the first wind turbine blade and the icing mass information of the second wind turbine blade, and the icing risk and hidden danger level of the icing mass information of the first wind turbine blade is greater than that of the icing mass information of the second wind turbine blade; Determining the icing condition description data of the first wind turbine blade, and performing clustering processing on the icing condition description data of the wind turbine blade to be processed based on the icing condition description data of the first wind turbine blade to obtain a first clustering result; Determining the icing condition description data of the second wind turbine blade, and performing clustering processing on the icing condition description data of the wind turbine blade to be processed based on the icing condition description data of the second wind turbine blade to obtain a second clustering result, where the change amount of the reflection feature of the icing condition description data of the second wind turbine blade is less than that of the icing condition description data of the first wind turbine blade; Determining the quality arrangement description data of the icing mass information of the first wind turbine blade based on the first clustering result, and performing fusion processing on the first clustering result and the second clustering result based on the quality arrangement description data to obtain a target clustering result; Evaluating the target clustering result to obtain the icing risk assessment result of the wind turbine blade.

2. The method according to claim 1, wherein The determining the icing condition description data of the first wind turbine blade includes: When the icing condition description data of the wind turbine blade to be processed is the current icing condition description data of the wind turbine blade, performing parsing processing on the icing condition description data of the wind turbine blade to be processed to obtain a parsing result; Based on the parsing result, determining the maximum unit dynamic load value, the minimum unit dynamic load value, and the number of influencing factors of the wind turbine blade icing corresponding to each unit dynamic load value; Determining a plurality of target parsing regions based on the maximum unit dynamic load value and the minimum unit dynamic load value; Determining the global number of influencing factor segments of the wind turbine blade icing corresponding to each target parsing region based on the number of influencing factors of the wind turbine blade icing corresponding to each unit dynamic load value; Determining the icing condition description data of the first wind turbine blade based on the global number of influencing factor segments of the wind turbine blade icing corresponding to each target parsing region.

3. The method according to claim 2, wherein The determining a plurality of target parsing regions based on the maximum unit dynamic load value and the minimum unit dynamic load value includes: Obtaining the constraint boundary for dividing the parsing region; Dividing a plurality of parsing regions from the maximum unit dynamic load value to the minimum unit dynamic load value according to the constraint boundary; Taking the parsing regions as the target parsing regions; Alternatively, selecting a plurality of target parsing regions from the plurality of parsing regions according to a preset region interval.

4. The method according to claim 2, wherein The determining the icing condition description data of the first wind turbine blade based on the global number of influencing factor segments of the wind turbine blade icing corresponding to each target parsing region includes: Arranging the target parsing regions in descending order according to the maximum unit dynamic load value of the region to obtain the arranged target parsing regions; Based on each of the arranged target analysis regions, determine the x-th comparison value of the global number of influence factor segments of ice accretion on the wind turbine blade corresponding to the (x + 1)-th target analysis region and the global number of influence factor segments of ice accretion on the wind turbine blade corresponding to the x-th target analysis region; When the x-th comparison value is greater than the preset comparison value setting, use the regional minimum unit dynamic load value corresponding to the x-th target analysis region as the first description data of the ice accretion situation of the wind turbine blade, where x is an integer between 1 and (M - 1), and M is the global number of target analysis regions.

5. The method according to claim 2, wherein The determination of the second description data of the ice accretion situation of the wind turbine blade includes: Successively use each unit dynamic load value from the minimum unit dynamic load value to the maximum unit dynamic load value as a candidate setting value; Determine multiple important influence factor segments of ice accretion on the wind turbine blade and multiple potential influence factor segments of ice accretion on the wind turbine blade corresponding to each candidate setting value; Based on the unit dynamic load value corresponding to each important influence factor segment of ice accretion on the wind turbine blade and the unit dynamic load value corresponding to each potential influence factor segment of ice accretion on the wind turbine blade, determine the function processing result corresponding to each candidate setting value; Use the candidate setting value corresponding to the maximum function processing result as the second description data of the ice accretion situation of the wind turbine blade.

6. The method according to claim 5, wherein The determination of the function processing result corresponding to each candidate setting value based on the unit dynamic load value corresponding to each important influence factor segment of ice accretion on the wind turbine blade and the unit dynamic load value corresponding to each potential influence factor segment of ice accretion on the wind turbine blade includes: Based on the global number of important influence factor segments of ice accretion on the wind turbine blade and the global number of potential influence factor segments of ice accretion on the wind turbine blade, determine the important feature ratio and the potential feature ratio, and based on the unit dynamic load value corresponding to each important influence factor segment of ice accretion on the wind turbine blade and the global number of important influence factor segments of ice accretion on the wind turbine blade, determine the important analysis average value; Based on the unit dynamic load value corresponding to each potential influence factor segment of ice accretion on the wind turbine blade and the global number of potential influence factor segments of ice accretion on the wind turbine blade, determine the potential analysis average value; Determine the average value of the description data of the ice accretion situation of the wind turbine blade for the analysis result; Based on the average value of the description data of the ice accretion situation of the wind turbine blade, the important feature ratio, the potential feature ratio, the important analysis average value, and the potential analysis average value, determine the function processing result corresponding to the candidate setting value.

7. The method according to claim 1, wherein The determination of the quality arrangement description data of the first ice accretion quality information of the wind turbine blade based on the first clustering result includes: Perform a filtering process on the first clustering result to obtain a filtered clustering result; Obtain the risk clusters in the filtered clustering result; Based on the risk clusters, combine the regional coefficients corresponding to the first ice accretion quality information of the wind turbine blade; Based on the regional coefficients and the risk clusters, determine the quality arrangement description data of the first ice accretion quality information of the wind turbine blade.

8. The method according to claim 7, wherein The quality arrangement description data for determining the icing mass information of the first wind turbine blade based on the region coefficient and the risk cluster includes: Obtaining blade load data corresponding to the icing mass information of the first wind turbine blade; Generating adaptive threshold reference wind turbine blade icing condition description data based on the region coefficient, the risk cluster, and the blade load data, where the influencing factor segments of the wind turbine blade icing within the icing mass information of the first wind turbine blade are the first attribute in the adaptive threshold reference wind turbine blade icing condition description data, and the influencing factor segments of the wind turbine blade icing outside the icing mass information of the first wind turbine blade are the second attribute; Determining the significant difference data of the influencing factor segments of the wind turbine blade icing within the icing mass information of the first wind turbine blade based on the adaptive threshold reference wind turbine blade icing condition description data; Taking the significant difference data of the influencing factor segments of each wind turbine blade icing as the quality arrangement description data of the icing mass information of the first wind turbine blade.

9. The method according to claim 8, wherein The fusion processing of the first clustering result and the second clustering result based on the quality arrangement description data to obtain the target clustering result includes: Obtaining the significant difference data of the influencing factor segments of the wind turbine blade icing of the y-th wind turbine in the second clustering result, where y = 1, 2,..., N, and N is the global quantity of the influencing factor segments of the wind turbine blade icing in the second clustering result; When it is determined that the influencing factor segment of the wind turbine blade icing of the y-th wind turbine is within the icing mass information of the first wind turbine blade based on the significant difference data of the influencing factor segment of the wind turbine blade icing of the y-th wind turbine and the adaptive threshold reference wind turbine blade icing condition description data, obtaining the first influencing factor parameter of the influencing factor segment of the wind turbine blade icing of the y-th wind turbine in the first clustering result; Converting the influencing factor parameter of the influencing factor segment of the wind turbine blade icing of the y-th wind turbine in the second clustering result into the first influencing factor parameter; When it is determined that the influencing factor segment of the wind turbine blade icing of the y-th wind turbine is outside the icing mass information of the first wind turbine blade, keeping the influencing factor parameter of the influencing factor segment of the wind turbine blade icing of the y-th wind turbine unchanged, and taking the processed second clustering result as the target clustering result; Among them, when the to-be-processed wind turbine blade icing condition description data includes a to-be-processed directory, the evaluation of the target clustering result to obtain the wind turbine blade icing risk assessment result includes: Performing directory 0 position processing on the target clustering result to obtain the quality arrangement description data of the directory; Obtaining the directory wind turbine blade icing condition description data from the target clustering result based on the quality arrangement description data of the directory; Evaluating the directory wind turbine blade icing condition description data by using a preset artificial intelligence blade risk assessment thread to obtain the wind turbine blade icing risk assessment result.

10. An assessment system for the icing risk of a fan blade, characterized in that, Including a processor and a memory that communicate with each other, the processor is configured to read and execute a computer program from the memory to implement the method according to any one of claims 1-9.

Images