Wind power blade icing state monitoring method and system based on fuzzy reasoning

Abstract

The invention belongs to the technical field of wind power generation, and particularly relates to a wind power blade icing state monitoring method and system based on fuzzy reasoning. The method comprises the following steps: acquiring non-icing historical operation data of a unit, averaging according to a wind speed interval by referring to an IEC standard, and fitting performance characteristics of the unit at different wind speeds; sCADA data of a unit in nearly four hours are acquired and cleaned, and whether the data are enough for monitoring is judged; averaging the cleaned data according to a fixed time period, dividing the data into two groups above a rated wind speed and below the rated wind speed, and respectively calculating performance average deviations; and in combination with a corresponding icing grade fuzzy inference rule, calculating two groups of icing grades by adopting a Mamdani inference algorithm, and weighting according to the number of the two groups of corresponding time periods to obtain an average blade icing grade of nearly 4 hours. The system comprises a performance feature fitting module, a data processing module, a performance average deviation calculation module and a blade icing grade calculation module.

Description

Technical Field

[0001] This invention belongs to the field of wind power generation technology, specifically relating to a method and system for monitoring the icing status of wind turbine blades based on fuzzy reasoning. Background Technology

[0002] Wind turbine blades are highly susceptible to icing under conditions of low temperature and high humidity, snowfall, or freezing rain. Icing alters the blade airfoil shape and surface roughness, increases blade weight, leading to a decrease in lift coefficient, an increase in drag coefficient, and premature stall. This, in turn, causes a significant reduction in turbine output power at the same wind speed, an increase in the wind speed required to reach rated power, a deviation from the optimal rotor speed-to-tip speed ratio, an increase in rotor unbalanced load, and increased turbine vibration and cyclic loads. In severe cases, it can lead to a series of problems such as difficulty starting and connecting to the grid, protective shutdown, and blade breakage, resulting in huge economic losses and significant safety hazards. Therefore, monitoring the icing status of wind turbine blades has significant engineering value.

[0003] Existing technologies often rely on external blade sensors for icing monitoring, including vibration sensors, dielectric constant patch sensors, ultrasonic guided waves, optical scattering sensors, fiber optic sensors, and visual monitoring. These technologies are significantly affected by installation location, environmental visibility, and fog / frost conditions, resulting in insufficient reliability and versatility. Some sensors require large-area deployment on the blade surface, making installation, debugging, and maintenance extremely inconvenient, increasing application costs, and raising the risk of lightning strikes. Many power plants, for cost and deployment convenience, tend to rely on software-based identification based on SCADA data. Common practices include using historical power curves as a benchmark, employing fixed wind speed compartments and thresholds to determine negative power deviations at the same wind speed, or using deviations in blade pitch angle from wind speed and power as supporting evidence. However, methods based on single indicators and fixed thresholds are easily affected by seasonal changes in air density, wake effects and yaw mismatch, grid-friendly power / AGC power limits, and adjustments to unit control strategies, leading to both false alarms and missed alarms. Furthermore, relying solely on power curve deviations is insufficient for timely detection of early or mild icing, and the ability to identify pitch control anomalies above rated operating conditions is limited. Overall, existing methods generally lack synergistic utilization and uncertainty handling mechanisms for multi-source evidence (such as power loss, pitch angle residuals, tip speed ratio drift, aerodynamic efficiency decline, and changes in correlation / coherence), making it difficult to reliably reuse them across different aircraft types, different sites, and complex operating conditions. Therefore, there is an urgent need for an icing status monitoring technology that can fully leverage existing SCADA and conventional sensor signals without increasing or minimizing hardware investment, and that robustly handles uncertainties in operating conditions and data quality through multi-feature fusion. This would improve the early detection rate of light to moderate icing, reduce false alarms, and support icing severity grading and subsequent maintenance. The deterministic threshold is unreasonable. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for monitoring the icing status of wind turbine blades based on fuzzy reasoning. This addresses the shortcomings of existing technologies, such as the significant impact of external blade sensors on icing monitoring due to factors like installation location, environmental visibility, fog, and frost, resulting in insufficient reliability and universality. Furthermore, the lack of multi-source evidence makes these sensors susceptible to interference from various factors, including seasonal changes in air density, wake effects and yaw mismatch, grid-friendly power / AGC power limits, and adjustments to turbine control strategies, leading to both false alarms and missed alarms. Additionally, relying solely on power curve deviations is insufficient for timely detection of early or mild icing, and the system has limited ability to identify pitch control anomalies above rated operating conditions, making it difficult to reliably reuse the technology across different turbine models, different sites, and complex operating conditions.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for monitoring the icing status of wind turbine blades based on fuzzy reasoning includes the following steps: Historical operating data of the unit under non-icing conditions were obtained, and the average was calculated according to the wind speed range with reference to IEC standards. The unit performance characteristics under different wind speeds were then obtained by fitting the data. Acquire SCADA data during the unit's operation period and clean the SCADA data during the unit's operation period; The SCADA data of the cleaning unit during the operation period are averaged according to a fixed time period. The averaged data is divided into two groups: data above the rated wind speed and data below the rated wind speed. The average performance deviation is calculated for the two groups of data respectively. Based on the average performance deviation of the two sets of data, and combined with the fuzzy inference rules for blade icing levels above and below the rated wind speed, the Mamdani inference algorithm is used to calculate the blade icing levels above and below the rated wind speed in real time within a certain period. Based on the blade icing levels above and below the rated wind speed within a certain period, and the number of time periods corresponding to the rated wind speed and below the rated wind speed respectively, the average blade icing level within a certain period is calculated by a weighted method.

[0006] The process involves acquiring historical operating data of the unit under non-icing conditions, averaging the data across wind speed ranges according to IEC standards, and fitting the data to obtain the unit's performance characteristics at different wind speeds. Specifically, this includes cleaning the historical operating data under non-icing conditions, averaging the cleaned data across wind speed ranges over 10 minutes according to IEC standards, and fitting the data to obtain the unit's performance characteristics at different wind speeds. The unit's performance characteristics under different wind speeds The unit performance includes characteristic power characteristics, pitch angle characteristics at different wind speeds, tip speed ratio characteristics at different wind speeds, and 1P-3P power spectrum amplitude ratio at different wind speeds.

[0007] Acquire SCADA data during the unit's operation period and clean the SCADA data during the unit's operation period. Specifically, acquire SCADA data during the unit's operation period using a sliding time window method, set an icing threshold, and compare the average ambient temperature in the SCADA data during the unit's operation period with the icing threshold to determine whether the blades are iced. If no icing occurs, the current judgment ends. If icing occurs, clean the SCADA data during the unit's operation period. Based on the cleaned SCADA data during the unit's operation period, determine whether the data is sufficient to determine the blade icing status. If the data is insufficient to determine whether the blades are iced, the current judgment ends, and SCADA data for the unit's operation period is acquired again.

[0008] The system compares the average ambient temperature in the SCADA data during the unit's operation with the icing threshold to determine whether the blades are icing. Specifically, if the average ambient temperature in the SCADA data during the unit's operation is higher than the icing threshold, the blades are considered not to be icing. If the average ambient temperature in the SCADA data during the unit's operation is less than or equal to the icing threshold, the blades are considered to be icing, and the process of cleaning the SCADA data during the unit's operation is then initiated.

[0009] The SCADA data during the unit's operation period is cleaned according to the unit's status, removing data from abnormal power generation points such as shutdown, idling, maintenance, repair, startup, power limitation, and pre-retraction of the propeller. If more than 50% of the cleaned SCADA data during the unit's operation period is removed, it is directly determined that there is insufficient data to determine whether the blades are icing; otherwise, the data is sufficient to determine whether the blades are icing.

[0010] The SCADA data collected during the operation of the cleaning unit was averaged over a fixed 2-minute interval. The averaged data was then divided into two groups based on the average wind speed: below the rated wind speed and above the rated wind speed. The data below the rated wind speed included... Within a 2-minute time interval, the data above the rated wind speed include Each segment lasts 2 minutes.

[0011] The average performance deviation is calculated for the two sets of data. The average performance deviation includes the average power loss ratio below the rated wind speed, the average tip speed ratio deviation below the rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation below the rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation above the rated wind speed, and the average pitch angle deviation above the rated wind speed.

[0012] Define the blade icing level The fuzzy universe of discourse is in the range [0, 100], and the leaf icing level is... The severity levels are categorized as "severe," "moderate," and "mild," based on average performance deviation and blade icing level. Fuzzy inference rules for blade icing levels above and below rated wind speed are defined separately. In the fuzzy inference rules for blade icing levels below rated wind speed, the main characteristics of blade icing are reduced power, reduced tip speed ratio, and increased average 1P-3P power spectrum amplitude ratio, based on the average power loss ratio below rated wind speed. Average tip speed ratio deviation ratio below rated wind speed and the deviation of the average 1P-3P power spectrum amplitude below the rated wind speed As input, blade icing level As output, in the fuzzy inference rule for blade icing levels above rated wind speed, the main characteristics of blade icing are a decrease in pitch angle and an increase in the average 1P-3P power spectrum amplitude ratio. The fuzzy inference rule for blade icing levels above rated wind speed uses the deviation of the average 1P-3P power spectrum amplitude ratio above rated wind speed as the key indicator. Average pitch angle deviation above rated wind speed As an output, blade icing level As output.

[0013] Based on the average power loss ratio below rated wind speed, the average tip speed ratio deviation ratio below rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation below rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation above rated wind speed, and the average pitch angle deviation above rated wind speed, as well as the fuzzy inference rules for blade icing levels above and below rated wind speed, the Mamdani inference algorithm is used to calculate the blade icing level above rated wind speed in real time over a given period. Blade icing levels below rated wind speed Based on the number of time periods corresponding to wind speeds above and below the rated wind speed, the average blade icing level within each time period is calculated using a weighted method.

[0014] In the formula, This represents the average leaf icing level over a given period of time. The number of time periods in the data below the rated wind speed. This refers to the number of time periods within the data above the rated wind speed.

[0015] A wind turbine blade icing status monitoring system based on fuzzy reasoning includes a performance characteristic fitting module, a data processing module, a performance average deviation calculation module, and a blade icing level calculation module. The performance characteristic fitting module is used to acquire historical operating data of the unit under non-icing conditions, calculate the average according to the wind speed range with reference to IEC standards, and fit the unit's performance characteristics under different wind speeds. The data processing module is used to acquire SCADA data during the unit's operation period and to clean the SCADA data during the unit's operation period. The average performance deviation calculation module is used to average the SCADA data of the cleaning unit during the operation period according to a fixed time period, divide the averaged data into two groups: above the rated wind speed and below the rated wind speed, and calculate the average performance deviation for each group of data. The blade icing level calculation module is used to calculate the blade icing level above and below the rated wind speed in real time based on the average performance deviation of two sets of data, combined with the fuzzy inference rules for blade icing levels above and below the rated wind speed, using the Mamdani inference algorithm. Based on the blade icing levels above and below the rated wind speed within a period of time, and the number of time periods corresponding to the rated wind speed above and below the rated wind speed, the average blade icing level within a period of time is calculated using a weighted method.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a fuzzy inference-based method for monitoring the icing status of wind turbine blades. By fusing multiple pieces of evidence, such as power loss, pitch angle residual, tip speed ratio drift, and the increase in the 1P-3P spectral amplitude ratio of power, it significantly improves the detection rate of light to moderate icing under complex operating conditions and reduces false alarms caused by a single icing criterion. The method uses fuzzy membership and rule bases to replace hard thresholds above and below rated wind speeds, respectively, to robustly cope with air density, wake, seasonal fluctuations, and other random uncertainties, reducing false alarms and missed alarms. Using the unit's recent month-long non-icing data to build self-constructed power characteristics, pitch angle characteristics, tip speed ratio characteristics, and the 1P-3P spectral amplitude ratio of power as reference benchmarks, it has adaptability and portability, reducing manual threshold parameter tuning across turbine models and stations. The method outputs a continuous and accurate "icing index" rather than a simple icing level, which is beneficial for linking de-icing control and maintenance scheduling. The method only requires software improvements to the SCADA system and does not require the addition of new sensors. Attached Figure Description

[0017] Figure 1 This is a flowchart of a wind turbine blade icing status monitoring method based on fuzzy reasoning in an embodiment of the present invention; Figure 2 This is a schematic diagram of the membership function constructed based on the average performance deviation of two sets of data in an embodiment of the present invention; Figure 3This is a schematic diagram of the membership function for constructing the leaf icing level in an embodiment of the present invention. Detailed Implementation

[0018] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0019] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] Example 1 This embodiment provides a method for monitoring the icing status of wind turbine blades based on fuzzy reasoning, which includes the following steps: Historical operating data of the unit under non-icing conditions were obtained, and the average was calculated according to the wind speed range with reference to IEC standards. The unit performance characteristics under different wind speeds were then obtained by fitting the data. Acquire SCADA data during the unit's operation period and clean the SCADA data during the unit's operation period; The SCADA data of the cleaning unit during the operation period are averaged according to a fixed time period. The averaged data is divided into two groups: data above the rated wind speed and data below the rated wind speed. The average performance deviation is calculated for the two groups of data respectively. Based on the average performance deviation of the two sets of data, and combined with the fuzzy inference rules for blade icing levels above and below the rated wind speed, the Mamdani inference algorithm is used to calculate the blade icing levels above and below the rated wind speed in real time within a certain period. Based on the blade icing levels above and below the rated wind speed within a certain period, and the number of time periods corresponding to the rated wind speed and below the rated wind speed respectively, the average blade icing level within a certain period is calculated by a weighted method.

[0021] Based on the above steps, combined with, for example Figure 1 The flowchart shown illustrates the method used in this embodiment to monitor and calculate the average leaf icing level over the past 4 hours. The specific implementation method is as follows: S1: Offline statistics of historical operating data of the unit under non-icing conditions. The historical operating data of the unit under non-icing conditions includes unit status, wind speed, active power, wind turbine speed, blade pitch angle, and ambient temperature. In order to reduce random errors, the amount of historical operating data collected should cover at least three months.

[0022] The historical operating data of the unit under non-icing conditions obtained above were preprocessed. The historical operating data under non-icing conditions were cleaned according to the unit status, and data of abnormal power generation operation points such as shutdown, idling, maintenance, repair, startup, power limitation, and pre-retraction of the propeller were removed.

[0023] Using historical operating data under non-icing conditions after cleaning, and referring to IEC standards, the average was calculated for 10 minutes according to wind speed ranges, and the data was fitted to obtain the unit's operating data under different wind speeds. The unit's performance characteristics under different wind speeds The unit performance includes characteristic power characteristics. Pitch angle characteristics under different wind speeds Blade tip speed ratio characteristics under different wind speeds and the 1P-3P spectral amplitude ratio of power at different wind speeds The formula for calculating the tip speed ratio is as follows: In the formula, The wind turbine rotation speed, Where is the radius of the wind turbine. Wind speed measured by a nacelle anemometer, corrected according to IEC standards; power 1P-3P spectral amplitude ratio This refers to performing a spectral analysis of the power across different wind speed ranges, where the spectral amplitude at the wind turbine rotation frequency, i.e., the 1P frequency, is... Spectral amplitude at 1P frequency The ratio is .

[0024] S2: Acquire SCADA data for the past 4 hours of unit operation using a sliding time window, including unit status, wind speed, active power, rotor speed, blade pitch angle, and ambient temperature. Set an icing threshold; if the average ambient temperature in the past 4 hours of SCADA data exceeds the icing threshold... If the average ambient temperature in the SCADA data over the past 4 hours is less than or equal to the icing threshold, then the blades are determined to be "not icy," and the current judgment ends.

[0025] S3: Clean the SCADA data obtained from S2 for the past 4 hours according to the unit status, removing data from abnormal power generation points such as shutdown, idling, maintenance, repair, startup, power limitation, and propeller retraction. If more than 50% of the data is removed, it is directly determined that "data is insufficient to determine whether the blades are icing" and the subsequent steps are skipped; otherwise, the data is sufficient to determine whether the blades are icing and the subsequent steps continue.

[0026] S4: Using nearly 4 hours of SCADA data after S3 cleaning, the average was calculated at fixed 2-minute intervals. The averaged data was then divided into two groups based on average wind speed: below the rated wind speed and above the rated wind speed. The data below the rated wind speed includes... Within a 2-minute time interval, the data above the rated wind speed include A 2-minute time interval. If all data is valid within 4 hours, then the following conditions are met: .

[0027] S5: Based on the two sets of data after cleaning in S4 and averaging over 2 minutes, calculate the average deviation of key performance for the two sets of data below the rated wind speed and above the rated wind speed, respectively.

[0028] S5-1: Calculate the average power loss ratio below the rated wind speed The calculation formula is defined as follows:

[0029] in, The first one below the rated wind speed Average wind speed over 2 minutes, The first one below the rated wind speed Average power over 2 minutes. From S1-3 The power characteristics were calculated. The larger the value, the higher the probability and degree of icing.

[0030] S5-2: Calculate the average tip speed ratio deviation ratio below the rated wind speed. The calculation formula is defined as follows:

[0031] in, The first one below the rated wind speed The average leaf tip speed ratio over 2 minutes. From S1-3 The tip speed ratio was calculated based on the blade tip speed characteristics. The larger the value, the higher the probability and degree of icing.

[0032] S5-3: Calculate the average 1P-3P power spectrum amplitude ratio deviation below the rated wind speed. The calculation formula is defined as follows:

[0033] in, The first one below the rated wind speed The average leaf tip speed ratio over 2 minutes. From S1-3 The tip speed ratio was calculated based on the blade tip speed characteristics. The larger the value, the higher the probability and degree of icing.

[0034] S5-4: Calculate the average 1P-3P power spectrum amplitude ratio deviation above the rated wind speed. The calculation formula is defined as follows:

[0035] in, The first one above the rated wind speed The average leaf tip speed ratio over 2 minutes. From S1-3 The tip speed ratio was calculated based on the blade tip speed characteristics. The larger the value, the higher the probability and degree of icing.

[0036] S5-5: Calculate the average pitch angle deviation above the rated wind speed. The calculation formula is defined as follows:

[0037] in, The first one above the rated wind speed The average pitch angle over 2 minutes. From S1-3 The pitch angle characteristics are calculated. The larger the value, the higher the probability and degree of icing.

[0038] S6: Construct the system defined in S5 , , , and The fuzzy domain and membership function.

[0039] S6-1: Average power loss ratio The reference range of the fuzzy domain is [0,1], which can be flexibly adjusted according to the actual situation.

[0040] S6-3: Average tip speed ratio deviation The reference range of the fuzzy domain is [0,1], which can be flexibly adjusted according to the actual situation.

[0041] S6-2: Average pitch angle deviation The fuzzy universe of discourse (in degrees, or radians) has a reference range of [0,6], which can be adjusted flexibly according to the actual situation.

[0042] S6-4: Average 1P-3P power spectrum amplitude ratio deviation and The reference range of the fuzzy universe of discourse is [0,10], which can be flexibly adjusted according to the actual situation.

[0043] S6-5: In specific embodiments, the membership function is selected as follows, and can be flexibly adjusted according to the actual situation, such as... Figure 2 As shown, where Figure 2 (a) represents the average power loss ratio below the rated wind speed. The fuzzy domain is [0,1], and the membership function is divided into three triangular fuzzy sets: "small", "medium", and "large". The "small" set has a membership degree of 1 at 0 and 0 at 0.5; the "medium" set has a membership degree of 1 at 0.5 and 0 at 0 and 1; the "large" set has a membership degree of 1 at 1 and 0 at 0.5. Figure 2 (b) indicates the proportion of average tip speed ratio deviation below the rated wind speed. The fuzzy universe of discourse is [0,1], and the membership function is also divided into three triangular fuzzy sets: "small", "medium", and "large", with distributions similar to... Figure 2 (a) Consistent. Figure 2 (c) indicates the average pitch angle deviation above the rated wind speed. The fuzzy domain is [0,6], and the membership function is divided into three triangular fuzzy sets: "small", "medium", and "large". The "small" set has a membership degree of 1 at 0 and 0 at 3; the "medium" set has a membership degree of 1 at 3 and 0 at 0 and 6; the "large" set has a membership degree of 1 at 6 and 0 at 3. Figure 2 (d) indicates the deviation of the average 1P-3P power spectrum amplitude below the rated wind speed. Deviation of average 1P-3P power spectrum amplitude above rated wind speed The fuzzy domain is [0,10]. The membership function is divided into three triangular fuzzy sets: "small", "medium", and "large". The "small" set has a membership degree of 1 at 0 and 0 at 5; the "medium" set has a membership degree of 1 at 5 and 0 at 0 and 10; the "large" set has a membership degree of 1 at 10 and 0 at 5.

[0044] S7: Constructing Leaf Icing Levels The fuzzy universe of discourse and membership function are determined by the following steps: S7-1: Indicates the level of leaf icing. The fuzzy domain is defined as the range [0, 100].

[0045] S7-2: Blade Icing Level The severity levels are categorized as "severe," "moderate," and "mild," with membership functions defined as follows (these can be adjusted flexibly according to actual circumstances). Figure 3 As shown, the blade icing level The fuzzy domain is [0, 100]. The membership function is divided into three triangular fuzzy sets: "small", "medium", and "large". Small corresponds to light icing, medium corresponds to moderate icing, and large corresponds to severe icing. Therefore, the light set has a membership of 1 at 0 and 0 at 50; the medium set has a membership of 1 at 50 and 0 at 0 and 100; and the severe set has a membership of 1 at 100 and 0 at 50.

[0046] S8: Calculated based on S5 , , , , and the blade icing levels defined by S7 Define fuzzy inference rules for blade icing levels above and below rated wind speed, respectively: S8-1: When the wind speed is below the rated wind speed, the main characteristics of blade icing are reduced power, reduced tip speed ratio, and increased average 1P-3P power spectrum amplitude ratio. Therefore, the average power loss ratio below the rated wind speed is... Average tip speed ratio deviation ratio below rated wind speed and the deviation of the average 1P-3P power spectrum amplitude below the rated wind speed As input to the fuzzy inference rule for blade icing levels below rated wind speed, blade icing level As the output of the fuzzy inference rule, the fuzzy inference rule for the blade icing level below the rated wind speed in this embodiment is as follows: (1): If Great, and Large or medium, and If it is large or medium, then It is "serious".

[0047] (2): If For the middle, and Large or medium, and If it is great, then It is "serious".

[0048] (3): If For the middle, and Great, and In the middle, then It is "serious".

[0049] (4): If Great, and It can be large, medium, or small, and If it is small, then It is rated as "medium".

[0050] (5): If Great, and Small, and If it is large or medium, then It is rated as "medium".

[0051] (6): If For the middle, and For the middle, and If it is in the middle, then It is rated as "medium".

[0052] (7): If Medium or small, and Small, and If it is great, then It is rated as "medium".

[0053] (8): If Medium or small, and Great, and If it is small, then It is rated as "medium".

[0054] (9): If Small, and For the middle, and If it is great, then It is rated as "medium".

[0055] (10): If Small, and Great, and If it is medium or large, then It is rated as "medium".

[0056] (11): If For the middle, and Small, and If it is in the middle, then It is classified as "mild".

[0057] (12): If For the middle, and Medium or small, and If it is small, then It is classified as "mild".

[0058] (13): If Small, and Medium or small, and If it is medium or small, then It is classified as "mild".

[0059] S8-2: When the wind speed is higher than the rated wind speed, the main characteristics of blade icing are a decrease in the pitch angle and an increase in the average 1P-3P power spectrum amplitude ratio. Therefore, the deviation of the average 1P-3P power spectrum amplitude ratio above the rated wind speed is... Average pitch angle deviation above rated wind speed As input to the fuzzy inference rules, the leaf icing level As the output of the fuzzy inference rule, the specific fuzzy inference rule for the blade icing level above the rated wind speed in this embodiment is as follows: (1): If Great, and If it is large or medium, then It is "serious".

[0060] (2): If For the middle, and If it is great, then It is "serious".

[0061] (3): If Small, and If it is great, then It is rated as "medium".

[0062] (4): If Great, and If it is small, then It is rated as "medium".

[0063] (5): If For the middle, and If it is in the middle, then It is rated as "medium".

[0064] (6): If For the middle, and If it is small, then It is classified as "mild".

[0065] (7): If Small, and If it is small or medium, then It is classified as "mild". S9: Calculated based on S5 , , , , Based on the fuzzy inference rules defined in S8, and using the Mamdani inference algorithm, the blade icing level above the rated wind speed in the past 4 hours is calculated in real time. Blade icing levels below rated wind speed .

[0066] S10: Blade icing rating at and below rated wind speed based on S9 calculations. , The average blade icing level for the most recent four hours is calculated using a weighted method, taking the number of time periods obtained from S4.

[0067] Finally, the SCADA system outputs a continuous icing index from 0 to 100; the higher the number, the more severe the icing. The number of time periods in the data below the rated wind speed. This refers to the number of time periods within the data above the rated wind speed.

[0068] Example 2 Based on the embodiments, a method for monitoring the icing status of wind turbine blades based on fuzzy reasoning is proposed. This embodiment proposes a system for monitoring the icing status of wind turbine blades based on fuzzy reasoning, including a performance feature fitting module, a data processing module, a performance average deviation calculation module, and a blade icing level calculation module. The performance characteristic fitting module is used to acquire historical operating data of the unit under non-icing conditions, calculate the average according to the wind speed range with reference to IEC standards, and fit the unit's performance characteristics under different wind speeds. The data processing module is used to acquire SCADA data during the unit's operation period and to clean the SCADA data during the unit's operation period. The average performance deviation calculation module is used to average the SCADA data of the cleaning unit during the operation period according to a fixed time period, divide the averaged data into two groups: above the rated wind speed and below the rated wind speed, and calculate the average performance deviation for each group of data. The blade icing level calculation module is used to calculate the blade icing level above and below the rated wind speed in real time based on the average performance deviation of two sets of data, combined with the fuzzy inference rules for blade icing levels above and below the rated wind speed, using the Mamdani inference algorithm. Based on the blade icing levels above and below the rated wind speed within a period of time, and the number of time periods corresponding to the rated wind speed above and below the rated wind speed, the average blade icing level within a period of time is calculated using a weighted method.

[0069] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A method for monitoring the icing status of wind turbine blades based on fuzzy reasoning, characterized in that, Includes the following steps: Historical operating data of the unit under non-icing conditions were obtained, and the average was calculated according to the wind speed range with reference to IEC standards. The unit performance characteristics under different wind speeds were then obtained by fitting the data. Acquire SCADA data during the unit's operation period and clean the SCADA data during the unit's operation period; The SCADA data of the cleaning unit during the operation period are averaged according to a fixed time period. The averaged data is divided into two groups: data above the rated wind speed and data below the rated wind speed. The average performance deviation is calculated for the two groups of data respectively. Based on the average performance deviation of the two sets of data, and combined with the fuzzy inference rules for blade icing levels above and below the rated wind speed, the Mamdani inference algorithm is used to calculate the blade icing levels above and below the rated wind speed in real time over a recent period. Based on the blade icing levels above and below the rated wind speed over a period of time, and the number of time periods corresponding to the rated wind speed above and below the rated wind speed, the average blade icing level over a recent period is calculated using a weighted method.

2. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 1, characterized in that, The process involves acquiring historical operating data of the unit under non-icing conditions, averaging the data across wind speed ranges according to IEC standards, and fitting the data to obtain the unit's performance characteristics at different wind speeds. Specifically, this includes cleaning the historical operating data under non-icing conditions, averaging the cleaned data across wind speed ranges over 10 minutes according to IEC standards, and fitting the data to obtain the unit's performance characteristics at different wind speeds. The unit's performance characteristics under different wind speeds The unit performance includes characteristic power characteristics, pitch angle characteristics at different wind speeds, tip speed ratio characteristics at different wind speeds, and 1P-3P power spectrum amplitude ratio at different wind speeds.

3. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 1, characterized in that, Acquire SCADA data during the unit's operation period and clean the SCADA data for the past 4 hours of unit operation. Specifically: acquire SCADA data during the unit's operation period using a sliding time window method, set an icing threshold, and compare the average ambient temperature in the SCADA data during the unit's operation period with the icing threshold to determine whether the blades are iced; if not iced, the current judgment ends; if iced, clean the SCADA data during the unit's operation period, and based on the cleaned SCADA data, determine whether the data is sufficient to determine the blade icing status; if the data is insufficient to determine whether the blades are iced, the current judgment ends, and SCADA data for the unit's operation period is acquired again.

4. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 3, characterized in that, The system compares the average ambient temperature in the SCADA data during the unit's operation with the icing threshold to determine whether the blades are icing. Specifically, if the average ambient temperature in the SCADA data during the unit's operation is higher than the icing threshold, the blades are considered not to be icing. If the average ambient temperature in the SCADA data during the unit's operation is less than or equal to the icing threshold, the blades are considered to be icing, and the process of cleaning the SCADA data during the unit's operation is then initiated.

5. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 4, characterized in that, The SCADA data during the unit's operation period is cleaned according to the unit's status, removing data from abnormal power generation points such as shutdown, idling, maintenance, repair, startup, power limitation, and pre-retraction of the propeller. If more than 50% of the cleaned SCADA data during the unit's operation period is removed, it is directly determined that there is insufficient data to determine whether the blades are icing; otherwise, the data is sufficient to determine whether the blades are icing.

6. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 1, characterized in that, The SCADA data collected during the operation of the cleaning unit was averaged over a fixed 2-minute interval. The averaged data was then divided into two groups based on the average wind speed: below the rated wind speed and above the rated wind speed. The data below the rated wind speed included... Within a 2-minute time interval, the data above the rated wind speed include Each segment lasts 2 minutes.

7. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 6, characterized in that, The average performance deviation is calculated for the two sets of data. The average performance deviation includes the average power loss ratio below the rated wind speed, the average tip speed ratio deviation below the rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation below the rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation above the rated wind speed, and the average pitch angle deviation above the rated wind speed.

8. The method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 7, characterized in that, Define the blade icing level The fuzzy universe of discourse is in the range [0, 100], and the leaf icing level is... The severity is categorized into "severe," "moderate," and "mild," based on average performance deviation and blade icing level. Fuzzy inference rules for blade icing levels above and below rated wind speed are defined separately. In the fuzzy inference rules for blade icing levels below rated wind speed, the main characteristics of blade icing are reduced power, reduced tip speed ratio, and increased average 1P-3P power spectrum amplitude ratio, based on the average power loss ratio below rated wind speed. Average tip speed ratio deviation ratio below rated wind speed and the deviation of the average 1P-3P power spectrum amplitude below the rated wind speed As input, blade icing level As output, in the fuzzy inference rule for blade icing levels above rated wind speed, the main characteristics of blade icing are a decrease in pitch angle and an increase in the average 1P-3P power spectrum amplitude ratio. The fuzzy inference rule for blade icing levels above rated wind speed uses the deviation of the average 1P-3P power spectrum amplitude ratio above rated wind speed as the key indicator. Average pitch angle deviation above rated wind speed As an output, blade icing level As output.

9. A method for monitoring the icing status of wind turbine blades based on fuzzy reasoning according to claim 8, characterized in that, Based on the average power loss ratio below rated wind speed, the average tip speed ratio deviation ratio below rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation below rated wind speed, the average 1P-3P power spectrum amplitude ratio deviation above rated wind speed, and the average pitch angle deviation above rated wind speed, as well as the fuzzy inference rules for blade icing levels above and below rated wind speed, the Mamdani inference algorithm is used to calculate the blade icing level above rated wind speed in real time over a given period. Blade icing levels below rated wind speed Based on the number of time periods corresponding to wind speeds above and below the rated wind speed, the average blade icing level within each time period is calculated using a weighted method. In the formula, This represents the average leaf icing level over a given period of time. The number of time periods in the data below the rated wind speed. This refers to the number of time periods within the data above the rated wind speed.

10. A wind turbine blade icing status monitoring system based on fuzzy reasoning, comprising a wind turbine blade icing status monitoring method based on fuzzy reasoning according to any one of claims 1 to 9, characterized in that, It includes a performance characteristic fitting module, a data processing module, a performance average deviation calculation module, and a blade icing level calculation module; The performance characteristic fitting module is used to acquire historical operating data of the unit under non-icing conditions, calculate the average according to the wind speed range with reference to IEC standards, and fit the unit's performance characteristics under different wind speeds. The data processing module is used to acquire SCADA data during the unit's operation period and to clean the SCADA data during the unit's operation period. The average performance deviation calculation module is used to average the SCADA data of the cleaning unit during the operation period according to a fixed time period, divide the averaged data into two groups: above the rated wind speed and below the rated wind speed, and calculate the average performance deviation for each group of data. The blade icing level calculation module is used to calculate the blade icing level above and below the rated wind speed in real time based on the average performance deviation of two sets of data, combined with the fuzzy inference rules for blade icing levels above and below the rated wind speed, using the Mamdani inference algorithm. Based on the blade icing levels above and below the rated wind speed within a period of time, and the number of time periods corresponding to the rated wind speed above and below the rated wind speed, the average blade icing level within a period of time is calculated using a weighted method.

Images