Failure judgment method for cooling system of fan frequency converter

Abstract

The invention discloses a failure judgment method for a cooling system of a fan frequency converter, which can more accurately identify faults of the cooling system, reduce false alarms and improve the reliability and operation efficiency of the system by introducing a model attenuation coefficient and a plurality of prediction deviation models. According to the method, system degradation caused by increase of the operating age limit of the fan is considered, the attenuation coefficient is dynamically adjusted, the fault judgment method can adapt to characteristic changes of the fan in different operating stages, the robustness of the system is enhanced, and even if the operating age limit of the fan is long and the performance of the cooling system is reduced, the fault can be effectively judged. And the running state of the cooling system can still be accurately judged, and normal power generation of the fan is prevented from being influenced by misinformation. The non-planned shutdown frequency and duration are effectively reduced, the operation efficiency of a wind power generation system is improved, and remarkable economic benefits are brought to energy companies.

Description

Technical Field

[0001] This invention relates to the field of wind turbine technology, and more specifically to a method for determining the failure of the cooling system of a wind turbine frequency converter. Background Technology

[0002] Currently, mainstream wind turbine manufacturers primarily rely on the pressure difference between the inlet and outlet of the cooling pump to determine whether the cooling system has failed. While this method can reflect the operating status of the cooling system to some extent, it has significant limitations. As wind turbines age, the cooling system and related components deteriorate to varying degrees, causing operating parameters to deviate from design values. This renders existing methods for determining cooling system failures inaccurate and unreliable over long-term operation, easily leading to false alarms, increasing the frequency and duration of unplanned wind turbine shutdowns, and severely impacting the operational efficiency and economic benefits of wind farms.

[0003] In an effort to reduce unplanned outages, some stations have resorted to tampering with cooling system sensor signals to mask alarms. This practice not only violates relevant management regulations but also poses serious safety hazards and increases the risk of production accidents.

[0004] Therefore, the purpose of this invention is to provide a more accurate, reliable and robust method for determining the failure status of the cooling system of a wind turbine frequency converter. This method can accurately determine the operating status of the cooling system even as the operating years of the wind turbine increase, reduce false alarms, reduce the number of unplanned shutdowns, improve the operating efficiency and safety of the wind power generation system, and avoid safety hazards caused by tampering with sensor signals, thus ensuring the stable and efficient operation of the wind power generation system. Summary of the Invention

[0005] The purpose of this invention is to provide a method for determining the failure of the cooling system of a fan inverter, effectively solving the technical problems existing in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution.

[0007] A method for determining the failure of a cooling system in a fan frequency converter, characterized by comprising the following steps: S1. Obtain environmental parameters and operating parameters of all wind turbines of the same model within a specific area over the past 6-12 months. The environmental parameters include at least ambient temperature, wind speed, and power generation. Construct a current-period prediction model ΔP_M_C for the wind turbine model based on the pressure difference of the inverter cooling system of all wind turbines of the same model. S2. Based on S1, obtain historical data on environmental parameters and wind turbine operating status within the area for 6-12 months after they have been put into operation, and build a model for all wind turbines of the same model to construct a basic prediction model ΔP_M_B based on the pressure difference of the inverter cooling system of all wind turbines of the same model. S3. Based on S2, test and calculate the current prediction model ΔP_M_C and the basic prediction model ΔP_M_B of the aircraft type to obtain the attenuation coefficient η. S4. Obtain environmental parameters and operating parameters of all fans of the same model within the same area as in step S1 within the most recent 6-12 months. Model each fan of the same model and construct a current fan prediction model ΔP_M_Ci for the pressure difference of the inverter cooling system of all fans of the same model. S5. Based on S4, compare the current differential pressure value of the cooling system of the fan inverter with the preset safety threshold T, and construct a safety threshold judgment model. If the current differential pressure value of the cooling system of the fan inverter is within the safe range, the result of the threshold judgment model is 0. If the current differential pressure value of the cooling system of the fan inverter is not within the safe range, the result is 1. The result of the threshold judgment model is represented by a. S6. Based on S5, calculate and construct an individual deviation judgment model according to the current cooling system differential pressure value of the fan inverter and the predicted value of the current fan prediction model ΔP_M_Ci. S7. Based on S5, calculate and construct a model deviation judgment model according to the current cooling system differential pressure value of the fan inverter and the predicted value of the current fan prediction model ΔP_M_Ci. S8. Based on S6 and S7, according to the attenuation coefficient η, the threshold judgment model result, the individual deviation judgment model result, and the model deviation judgment model result, a comprehensive judgment model f is constructed to obtain the comprehensive judgment coefficient Z. If the calculated Z value is greater than the set threshold T, a cooling system failure alarm is triggered; otherwise, it is considered that the fan can still operate normally.

[0008] Preferably, in step S1, the specific construction process of the current prediction model ΔP_M_C for the aircraft type is as follows: For all wind turbines of the same model in a specific area, the number of wind turbines is set to n, and a unified model is performed. The ambient temperature, wind speed and power generation of the most recent 6-12 months are selected as parameters, and the pressure difference ΔP of the corresponding wind turbine inverter cooling system is used as the target value. The least squares method is used to construct the current prediction model ΔP_M_C for this model.

[0009] Preferably, in step S2, the specific construction process of the aircraft model basic prediction model ΔP_M_B is as follows: For all wind turbines of the same model in a specific area, the number of wind turbines is set to n, and a unified model is performed. The ambient temperature, wind speed and power generation of the wind turbines are selected as parameters after 6-12 months of operation, and the pressure difference ΔP of the corresponding wind turbine inverter cooling system is used as the target value. The least squares method is used to construct the basic prediction model ΔP_M_B of the model.

[0010] Preferably, in step S3, the specific calculation process of the attenuation coefficient η is as follows: For the current model of wind turbines in a specific area, at least the typical ambient temperature (θ1, θ2 ... ..., θm), wind speed (v1, v2 ... ..., vm), and power generation (p1, p2 ... ..., pm) parameters should be selected. The pressure difference ΔP_C(ΔP_C1, ΔP_C2 ... ..., ΔP_Cm) and the pressure difference ΔP_B(ΔP_B1, ΔP_B2 ... ..., ΔP_Bm) of the cooling system under the typical parameters of this model should be calculated using the ΔP_M_C model and the ΔP_M_B model, respectively. η=α*(ΔP_C1 / ΔP_B1)+α*(ΔP_C2 / ΔP_B2) + ... ... + α*(ΔP_Cm / ΔP_Bm); Where α is the weighting coefficient, and its value is 1 / m.

[0011] Preferably, in step S4, the specific construction process of the current wind turbine prediction model ΔP_M_Ci is as follows: For all wind turbines of the same model within a specific area, the number of wind turbines is set to n, and each turbine is modeled separately. The ambient temperature, wind speed, and power generation over the past 6-12 months are selected as parameters, and the corresponding pressure difference ΔP of the wind turbine inverter cooling system is used as the target value. The least squares method is used to construct wind turbine prediction models ΔP_M_C1, ΔP_M_C2, ΔP_M_C3... ... ΔP_M_Cn for the n wind turbines in the specific area.

[0012] Preferably, in step S6, the specific content of the individual deviation determination model is as follows: Based on the pressure difference prediction result ΔPi_Pred obtained from the specific calculation of the current wind turbine prediction model ΔP_M_Ci, if the actual pressure difference value ΔPi of this wind turbine is within the confidence interval of the pressure difference prediction model result, the individual deviation judgment model judgment result is 0, otherwise the result is 1. The individual deviation judgment model result is represented by b. The confidence interval is: ΔPi_Pred ± 2SE, where SE is the standard deviation of the residuals, and the formula for calculating SE is: .

[0013] Preferably, in step S7, the specific content of the model deviation determination model is as follows: Using the current wind turbine prediction model ΔP_M_Ci, the predicted pressure difference for all wind turbines of this model is calculated as ΔP_Pred(ΔP1_Pred, ΔP2_Pred... ..., ΔPn_Pred); If the actual differential pressure value ΔPi of the current wind turbine is within the range of ΔP_Pred in the differential pressure prediction model (i.e., ΔPi is in ΔP1_Pred, ΔP2_Pred, ..., ΔPn_Pred), then it is not an outlier, and the model deviation judgment result is 0. Otherwise, the result is 1. The model deviation judgment result is represented by c.

[0014] Preferably, in step S8, the specific content of the comprehensive judgment model f is as follows: Based on the attenuation coefficient η, the results of the threshold judgment model, the results of the individual deviation judgment model, and the results of the model deviation judgment model, a comprehensive judgment model f(η, a, b, c) is constructed, and the comprehensive judgment coefficient Z is obtained, i.e., Z = f(η, a, b, c); where the calculation formula for Z is as follows: .

[0015] Preferably, the specific calculations using the least squares method in steps S1, S2, and S4 are as follows: Let (x, ΔP) be a set of observations, x = [θ, v, p]T, and the target pressure difference ΔP satisfy the following theoretical function: ΔP = f(x,ω); Where ω = [ω1,ω2,...,ωn]T are the parameters to be fitted; To find the optimal estimate of the parameter ω of the function f(x,ω), given m sets of observation data (xi,yi) (i=1,2,...,m), we solve for the objective function. ;; To achieve the best fit, calculate the objective function. The parameter with the smallest value i (i = 1, 2, ..., n).

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) By introducing a model attenuation coefficient and multiple prediction deviation models, this invention can more accurately identify cooling system faults, reduce false alarms, and improve system reliability and operating efficiency. This invention considers the system degradation caused by the increasing service life of wind turbines. By dynamically adjusting the attenuation coefficient, the fault judgment method can adapt to the characteristic changes of wind turbines at different operating stages, enhancing the robustness of the system. Even when the wind turbine has been in operation for a long time and the cooling system performance has declined, it can still accurately judge the operating status of the cooling system, avoiding the impact of false alarms on the normal power generation of the wind turbine. This effectively reduces the number and duration of unplanned shutdowns, improves the operating efficiency of the wind power generation system, and brings significant economic benefits to energy companies. Attached Figure Description

[0017] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0018] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0019] In the description of this invention, unless otherwise stated, "a plurality of" means two or more; the terms "upper," "lower," "left," "right," "inner," "outer," "front end," "rear end," "head," "tail," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0020] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] A method for determining the failure of the cooling system of a fan frequency converter includes the following steps: S1. Obtain environmental parameters and operating parameters of all wind turbines of the same model within a specific area over the past 6-12 months. The environmental parameters include at least ambient temperature, wind speed, and power generation. Construct a current-period prediction model ΔP_M_C for the wind turbine model based on the pressure difference of the inverter cooling system of all wind turbines of the same model. Specifically, the construction process of the current prediction model ΔP_M_C for this model is as follows: For all wind turbines of the same model in a specific area, set the number of wind turbines to n and perform unified modeling; select the ambient temperature, wind speed and power generation of the most recent 6-12 months as parameters, and the corresponding wind turbine inverter cooling system pressure difference ΔP as the target value, and use the least squares method to construct the current prediction model ΔP_M_C for this model.

[0022] S2. Based on S1, obtain historical data on environmental parameters and wind turbine operating status within the area for 6-12 months after they have been put into operation, and build a model for all wind turbines of the same model to construct a basic prediction model ΔP_M_B based on the pressure difference of the inverter cooling system of all wind turbines of the same model. Specifically, the construction process of the basic prediction model ΔP_M_B for the turbine model is as follows: For all wind turbines of the same model in a specific area, the number of wind turbines is set to n, and a unified model is performed; the ambient temperature, wind speed and power generation of the wind turbines for 6-12 months after they are put into operation are selected as parameters, and the pressure difference ΔP of the corresponding wind turbine inverter cooling system is used as the target value. The basic prediction model ΔP_M_B for the turbine model is constructed using the least squares method.

[0023] S3. Based on S2, test and calculate the current prediction model ΔP_M_C and the basic prediction model ΔP_M_B for the aircraft model to obtain the attenuation coefficient η. The specific calculation process of the attenuation coefficient η is as follows: For the current model of wind turbines in a specific area, at least the typical ambient temperature (θ1, θ2 ... ..., θm), wind speed (v1, v2 ... ..., vm), and power generation (p1, p2 ... ..., pm) parameters should be selected. The pressure difference ΔP_C(ΔP_C1, ΔP_C2 ... ..., ΔP_Cm) and the pressure difference ΔP_B(ΔP_B1, ΔP_B2 ... ..., ΔP_Bm) of the cooling system under the typical parameters of this model should be calculated using the ΔP_M_C model and the ΔP_M_B model, respectively. η=α*(ΔP_C1 / ΔP_B1)+α*(ΔP_C2 / ΔP_B2) + ... ... + α*(ΔP_Cm / ΔP_Bm); Where α is the weighting coefficient, with a value of 1 / m; in other embodiments, α can also be determined based on big data statistics, according to the probability / frequency of occurrence of typical working conditions, and the weighting coefficient is positively correlated with the probability of occurrence of typical working conditions.

[0024] S4. Obtain environmental parameters and operating parameters of all fans of the same model within the same area as in step S1 within the most recent 6-12 months. Model each fan of the same model and construct a current fan prediction model ΔP_M_Ci for the pressure difference of the inverter cooling system of all fans of the same model. The construction process of the current wind turbine prediction model ΔP_M_Ci is as follows: For all wind turbines of the same model within a specific area, the number of wind turbines is set to n, and each turbine is modeled separately. The ambient temperature, wind speed, and power generation over the past 6-12 months are selected as parameters, and the corresponding pressure difference ΔP of the wind turbine inverter cooling system is used as the target value. The least squares method is used to construct wind turbine prediction models ΔP_M_C1, ΔP_M_C2, ΔP_M_C3... ... ΔP_M_Cn for the n wind turbines in the specific area.

[0025] S5. Based on S4, compare the current differential pressure value of the cooling system of the fan inverter with the preset safety threshold T, and construct a safety threshold judgment model. If the current differential pressure value of the cooling system of the fan inverter is within the safe range, the result of the threshold judgment model is 0. If the current differential pressure value of the cooling system of the fan inverter is not within the safe range, the result is 1. The result of the threshold judgment model is represented by a. S6. Based on S5, calculate and construct an individual deviation judgment model according to the current cooling system differential pressure value of the fan inverter and the predicted value of the current fan prediction model ΔP_M_Ci. The specific details of the individual deviation judgment model are as follows: Based on the pressure difference prediction result ΔPi_Pred obtained from the specific calculation of the current wind turbine prediction model ΔP_M_Ci, if the actual pressure difference value ΔPi of this wind turbine is within the confidence interval of the pressure difference prediction model result, the individual deviation judgment model judgment result is 0, otherwise the result is 1. The individual deviation judgment model result is represented by b. The confidence interval is: ΔPi_Pred ± 2SE, where SE is the standard deviation of the residuals, and the formula for calculating SE is: .

[0026] S7. Based on S5, calculate and construct a model deviation judgment model according to the current cooling system differential pressure value of the fan inverter and the predicted value of the current fan prediction model ΔP_M_Ci. The specific details of the model deviation determination model are as follows: Using the current wind turbine prediction model ΔP_M_Ci, the predicted pressure difference for all wind turbines of this model is calculated as ΔP_Pred(ΔP1_Pred, ΔP2_Pred... ..., ΔPn_Pred); If the actual differential pressure value ΔPi of the current wind turbine is within the range of ΔP_Pred in the differential pressure prediction model (i.e., ΔPi is in ΔP1_Pred, ΔP2_Pred, ..., ΔPn_Pred), then it is not an outlier, and the model deviation judgment result is 0. Otherwise, the result is 1. The model deviation judgment result is represented by c.

[0027] Outliers can be identified using existing z-score statistical methods, with the specific process as follows: Step 1: Calculate the mean μ and standard deviation σ of the prediction set: μ = (ΔP1_Pred + ΔP2_Pred + ... + ΔPn_Pred) / n; σ = sqrt{[(ΔP1_Pred-μ)² + (ΔP2_Pred-μ)² + ... + (ΔPn_Pred-μ)²] / n}; Step 2: Calculate the Z-score value of the current wind turbine: z_i = (ΔP_i - μ) / σ; Step 3: Anomaly detection, where the threshold value is 3: c = { 0, if |z_i| ≤ 3; 1, if |z_i| > 3}; Individual deviation judgment models and model deviation judgment models are constructed separately. By calculating the confidence interval of the predicted differential pressure value and detecting outliers, it is determined whether the actual differential pressure value of the current wind turbine deviates from the normal range. This invention not only considers the uncertainty of the predicted value, but also can identify obviously abnormal data points, further improving the accuracy of fault diagnosis.

[0028] S8. Based on S6 and S7, according to the attenuation coefficient η, the threshold judgment model result, the individual deviation judgment model result, and the model deviation judgment model result, a comprehensive judgment model f is constructed to obtain the comprehensive judgment coefficient Z. If the calculated Z value is greater than the set threshold T, a cooling system failure alarm is triggered; otherwise, it is considered that the fan can still operate normally.

[0029] In step S8, the specific content of the comprehensive judgment model f is as follows: Based on the attenuation coefficient η, the results of the threshold judgment model, the results of the individual deviation judgment model, and the results of the model deviation judgment model, a comprehensive judgment model f(η, a, b, c) is constructed, and the comprehensive judgment coefficient Z is obtained, i.e., Z = f(η, a, b, c); where the calculation formula for Z is as follows: .

[0030] The value of η is dynamically adjusted according to the operating status of the wind turbine. As the operating years of the wind farm increase, the pressure difference of the wind turbine cooling system will tend to decrease. That is, η is 1 during the first 6-12 months of operation of the wind turbine, and then this value is corrected every year as the wind turbine operates, and η will gradually decrease.

[0031] The method for determining the T-value is as follows: Step 1, collect historical cooling system fault sample data of the fan; Step 2: Calculate the model decision value t(t1,t2,t3,...,ti) using the sample data and operating condition data as input. Step 3: Take the 95th percentile (from smallest to largest) of the positive sample judgment value as the threshold T to ensure that the fault can be identified.

[0032] The specific calculations using the least squares method in steps S1, S2, and S4 are as follows: Let (x, ΔP) be a set of observations, x = [θ, v, p]T, and the target pressure difference ΔP satisfy the following theoretical function: ΔP = f(x,ω); Where ω = [ω1,ω2,...,ωn]T are the parameters to be fitted; To find the optimal estimate of the parameter ω of the function f(x,ω), given m sets of observation data (xi,yi) (i=1,2,...,m), we solve for the objective function. ; To achieve the best fit, calculate the objective function. The parameter with the smallest value i (i = 1, 2, ..., n).

[0033] This invention utilizes multiple predictive models, combined with historical data on environmental parameters and wind turbine operating status, to dynamically monitor and diagnose the differential pressure in the cooling system of a wind turbine inverter. This invention considers not only the current operating status but also long-term historical data to more accurately assess the health of the cooling system. By introducing model attenuation coefficients and multiple predictive deviation models, this invention can more accurately identify cooling system faults, reduce false alarms, and improve system reliability and operating efficiency. The method of this invention enhances the system's robustness, enabling accurate assessment of the cooling system's operating status even after the wind turbine has been in operation for a long time and the cooling system's performance has deteriorated, avoiding disruptions to normal wind turbine power generation due to false alarms. This effectively reduces the number and duration of unplanned outages, improves the operating efficiency of wind power systems, and brings significant economic benefits to energy companies.

[0034] The above embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any modifications, equivalent changes, improvements, etc., made in accordance with the claims of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for determining the failure of the cooling system of a fan frequency converter, characterized in that, Includes the following steps: S1. Obtain environmental parameters and operating parameters of all wind turbines of the same model within a specific area over the past 6-12 months. The environmental parameters include at least ambient temperature, wind speed, and power generation. Construct a current-period prediction model ΔP_M_C for the wind turbine model based on the pressure difference of the inverter cooling system of all wind turbines of the same model. S2. Based on S1, obtain historical data on environmental parameters and wind turbine operating status within the area for 6-12 months after they have been put into operation, and build a model for all wind turbines of the same model to construct a basic prediction model ΔP_M_B based on the pressure difference of the inverter cooling system of all wind turbines of the same model. S3. Based on S2, test and calculate the current prediction model ΔP_M_C and the basic prediction model ΔP_M_B of the aircraft type to obtain the attenuation coefficient η. S4. Obtain environmental parameters and operating parameters of all fans of the same model within the same area as in step S1 within the most recent 6-12 months. Model each fan of the same model and construct a current fan prediction model ΔP_M_Ci for the pressure difference of the inverter cooling system of all fans of the same model. S5. Based on S4, compare the current differential pressure value of the cooling system of the fan inverter with the preset safety threshold T, and construct a safety threshold judgment model. If the current differential pressure value of the cooling system of the fan inverter is within the safe range, the result of the threshold judgment model is 0. If the current differential pressure value of the cooling system of the fan inverter is not within the safe range, the result is 1. The result of the threshold judgment model is represented by a. S6. Based on S5, calculate and construct an individual deviation judgment model according to the current cooling system differential pressure value of the fan inverter and the predicted value of the current fan prediction model ΔP_M_Ci. S7. Based on S5, calculate and construct a model deviation judgment model according to the current cooling system differential pressure value of the fan inverter and the predicted value of the current fan prediction model ΔP_M_Ci. S8. Based on S6 and S7, according to the attenuation coefficient η, the threshold judgment model result, the individual deviation judgment model result, and the model deviation judgment model result, a comprehensive judgment model f is constructed to obtain the comprehensive judgment coefficient Z. If the calculated Z value is greater than the set threshold T, a cooling system failure alarm is triggered; otherwise, it is considered that the fan can still operate normally.

2. The method for determining the failure of the cooling system of a fan frequency converter according to claim 1, characterized in that, In step S1, the specific construction process of the current prediction model ΔP_M_C for the aircraft type is as follows: For all wind turbines of the same model in a specific area, the number of wind turbines is set to n, and a unified model is performed. The ambient temperature, wind speed and power generation of the most recent 6-12 months are selected as parameters, and the pressure difference ΔP of the corresponding wind turbine inverter cooling system is used as the target value. The least squares method is used to construct the current prediction model ΔP_M_C for this model.

3. The method for determining the failure of the cooling system of a fan frequency converter according to claim 2, characterized in that, In step S2, the specific construction process of the aircraft basic prediction model ΔP_M_B is as follows: For all wind turbines of the same model in a specific area, the number of wind turbines is set to n, and a unified model is performed. The ambient temperature, wind speed and power generation of the wind turbines are selected as parameters after 6-12 months of operation, and the pressure difference ΔP of the corresponding wind turbine inverter cooling system is used as the target value. The least squares method is used to construct the basic prediction model ΔP_M_B of the model.

4. The method for determining the failure of the cooling system of a fan frequency converter according to claim 3, characterized in that, In step S3, the specific calculation process for the attenuation coefficient η is as follows: For the current model of wind turbines in a specific area, at least the typical ambient temperature (θ1, θ2 ... ..., θm), wind speed (v1, v2 ... ..., vm), and power generation (p1, p2 ... ..., pm) parameters should be selected. The pressure difference ΔP_C(ΔP_C1, ΔP_C2 ... ..., ΔP_Cm) and the pressure difference ΔP_B(ΔP_B1, ΔP_B2 ... ..., ΔP_Bm) of the cooling system under the typical parameters of this model should be calculated using the ΔP_M_C model and the ΔP_M_B model, respectively. η=α*(ΔP_C1 / ΔP_B1)+α*(ΔP_C2 / ΔP_B2) + ... ... + α*(ΔP_Cm / ΔP_Bm); Where α is the weighting coefficient, and its value is 1 / m.

5. The method for determining the failure of the cooling system of a fan frequency converter according to claim 4, characterized in that, In step S4, the specific construction process of the current wind turbine prediction model ΔP_M_Ci is as follows: For all wind turbines of the same model within a specific area, the number of wind turbines is set to n, and each turbine is modeled separately. The ambient temperature, wind speed, and power generation over the past 6-12 months are selected as parameters, and the corresponding pressure difference ΔP of the wind turbine inverter cooling system is used as the target value. The least squares method is used to construct wind turbine prediction models ΔP_M_C1, ΔP_M_C2, ΔP_M_C3... ... ΔP_M_Cn for the n wind turbines in the specific area.

6. The method for determining the failure of the cooling system of a fan frequency converter according to claim 5, characterized in that, In step S6, the specific content of the individual deviation determination model is as follows: Based on the pressure difference prediction result ΔPi_Pred obtained from the specific calculation of the current wind turbine prediction model ΔP_M_Ci, if the actual pressure difference value ΔPi of this wind turbine is within the confidence interval of the pressure difference prediction model result, the individual deviation judgment model judgment result is 0, otherwise the result is 1. The individual deviation judgment model result is represented by b. The confidence interval is: ΔPi_Pred ± 2SE, where SE is the standard deviation of the residuals, and the formula for calculating SE is: 。 7. The method for determining the failure of the cooling system of a fan frequency converter according to claim 6, characterized in that, In step S7, the specific content of the aircraft deviation determination model is as follows: Using the current wind turbine prediction model ΔP_M_Ci, the predicted pressure difference for all wind turbines of this model is calculated as ΔP_Pred(ΔP1_Pred, ΔP2_Pred... ..., ΔPn_Pred); If the actual differential pressure value ΔPi of the current wind turbine is within the range of ΔP_Pred in the differential pressure prediction model (i.e., ΔPi is in ΔP1_Pred, ΔP2_Pred, ..., ΔPn_Pred), then it is not an outlier, and the model deviation judgment result is 0. Otherwise, the result is 1. The model deviation judgment result is represented by c.

8. The method for determining the failure of the cooling system of a fan frequency converter according to claim 7, characterized in that, In step S8, the specific content of the comprehensive judgment model f is as follows: Based on the attenuation coefficient η, the results of the threshold judgment model, the results of the individual deviation judgment model, and the results of the model deviation judgment model, a comprehensive judgment model f(η, a, b, c) is constructed, and the comprehensive judgment coefficient Z is obtained, i.e., Z = f(η, a, b, c); where the calculation formula for Z is as follows: 。 9. A method for determining the failure of a cooling system in a fan frequency converter according to claim 6 or 7, characterized in that, The specific calculations using the least squares method in steps S1, S2, and S4 are as follows: Let (x, ΔP) be a set of observations, x = [θ, v, p]T, and the target pressure difference ΔP satisfy the following theoretical function: ΔP = f(x,ω) Where ω = [ω1,ω2,...,ωn]T are the parameters to be fitted; To find the optimal estimate of the parameter ω of the function f(x,ω), given m sets of observation data (xi,yi) (i=1,2,...,m), we solve for the objective function. ； To achieve the best fit, calculate the objective function. The parameter with the smallest value i (i = 1,2,...,n).

Images