High-power variable pitch system adaptive to SL1500 wind turbine generator and temperature control method

Abstract

The invention provides a high-power variable pitch system adaptive to an SL1500 wind turbine generator and a temperature control method, and is applied to the technical field of data processing. According to the method, the problem that the temperature of the variable-pitch motor of the SL1500 wind turbine generator is high is solved, unit operation data, variable-pitch system influence factors and abnormal state information of the motor are obtained firstly, and rated torque and operation temperature basic data of the variable-pitch motor are obtained through preprocessing; the temperature control improvement probability of the high-power variable-pitch system is calculated by dynamically fusing engine reinforced blade bearing wear and lubrication state weights and combining space-time correlation analysis and a regression model, and temperature anomaly feature information is generated by comparing data before and after technical improvement. The method comprises the following steps of: obtaining a multi-modal feature, processing the multi-modal feature to obtain an overload heating identification result, generating a risk influence factor in combination with a multi-task decision matrix and a technical improvement key index, and finally outputting a dynamic early warning result of over-high motor temperature by means of a time-space correlation early warning engine and a multi-target decision, thereby realizing problem solving and risk management and control.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a high-power pitch system and temperature control method adapted to SL1500 wind turbine units. Background Technology

[0002] As a wind turbine model, the SL1500 wind turbine's pitch system is a core component ensuring the safe and efficient operation of the unit. The temperature control of the pitch motor directly affects the unit's operational stability and service life. In existing technology, SL1500 wind turbines mostly use 23NMKEB pitch motors. In actual operation, this motor is prone to overload problems due to insufficient torque, which in turn causes abnormal temperature rises, with the highest temperature approaching 90℃. In severe cases, this can lead to the unit operating with limited power or shutting down due to malfunction.

[0003] Current technical solutions for temperature control of pitch motors rely on a limited data acquisition dimension, focusing primarily on the motor's own temperature and torque data. They fail to adequately incorporate key influencing factors such as blade bearing wear, lubrication status, and inverter operating conditions. Furthermore, the lack of a systematic data preprocessing mechanism results in insufficient data accuracy and completeness. In terms of influencing factor weighting, existing solutions do not emphasize the dominant role of core factors like blade bearing wear and lubrication status, making it difficult to accurately pinpoint the root cause of temperature anomalies. Simultaneously, existing technologies lack scientific spatiotemporal correlation analysis and regression models, making it impossible to quantify the improvement effect of technical upgrades on temperature control. Moreover, the absence of a multi-dimensional, collaborative overload heating identification and risk assessment mechanism hinders the early prediction of excessively high temperature risks and the development of targeted control strategies.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0006] According to one aspect of this application, a high-power pitch and temperature control method adapted to SL1500 wind turbine units is provided, comprising: acquiring data information on SL1500 unit operation data, pitch system influencing factor data, and motor abnormal state information; performing data preprocessing on the data information to generate basic data on the rated torque and operating temperature of the pitch motor; processing the influencing factor data and abnormal state information based on a dynamic fusion engine, strengthening the core influencing weights of blade bearing wear and lubrication status through a dynamic weight allocator, combining spatiotemporal correlation analysis, using an equipment operation state regression model to calculate the probability of improvement in temperature control by the high-power pitch system, and comparing and verifying the improvement scheme with the original system by combining data before and after the technical upgrade. Based on operational differences, abnormal temperature characteristics of the pitch system are identified; multimodal dynamic characteristic information is processed to generate pitch motor overload heating identification results; the identification results and risk levels in the influencing factor data are combined with a multi-task decision matrix, and the overload heating identification results corresponding to the high-power pitch system and the unit fault risk level are combined with the multi-task decision matrix, and the technical indicators independently related to the stable operation of the unit selected in the technical transformation plan are integrated to generate the risk impact factor of pitch system overload heating on unit operation; based on the spatiotemporal correlation early warning engine combined with multi-objective dynamic decision strategy, the identification results, risk impact factors and abnormal characteristic identification information are processed to generate dynamic early warning results of excessive pitch motor temperature.

[0007] Another aspect of this application discloses a high-power pitch and temperature control device adapted to SL1500 wind turbine units, comprising: an acquisition module for acquiring data type information of SL1500 unit operation, pitch system influencing factor data, and motor abnormal state information; a processing module for preprocessing the data type information to generate basic data on the rated torque and operating temperature of the pitch motor; processing the influencing factor data and abnormal state information based on a dynamic fusion engine, strengthening the core influencing weights of blade bearing wear and lubrication status through a dynamic weight allocator, combining spatiotemporal correlation analysis, using an equipment operating state regression model to calculate the probability of improvement in temperature control by the high-power pitch system, and verifying the improvement by comparing data before and after the technical upgrade. The proposed solution differs from the original system in its operation, generating information on abnormal temperature characteristics of the pitch system. Multimodal dynamic characteristic information is processed to generate pitch motor overload heating identification results. The identification results and risk levels in the influencing factor data are combined with a multi-task decision matrix. The overload heating identification results corresponding to the high-power pitch system and the unit fault risk level are combined with the multi-task decision matrix, integrating technical indicators independently related to stable unit operation selected from the technical upgrade plan, to generate risk impact factors of pitch system overload heating on unit operation. Based on a spatiotemporal correlation early warning engine combined with a multi-objective dynamic decision-making strategy, the identification results, risk impact factors, and abnormal characteristic identification information are processed to generate dynamic early warning results for excessively high pitch motor temperature.

[0008] According to another aspect of this application, an electronic device includes: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the above-described high-power pitch and temperature control method adapted for SL1500 wind turbines by executing the executable instructions.

[0009] According to another aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a second processor, implements the above-described high-power pitch and temperature control method adapted to SL1500 wind turbine units.

[0010] This application provides a high-power pitch control system and temperature control method adapted to SL1500 wind turbine units. The purpose of this application is to comprehensively collect unit operation data, pitch system influencing factors and abnormal motor status information, optimize the data preprocessing process, scientifically allocate the weights of influencing factors, and construct a spatiotemporal correlation analysis and multi-objective decision-making model to achieve accurate identification of overload heating, quantitative risk assessment and dynamic early warning of excessive temperature, and ultimately solve the problem of excessive pitch motor temperature and ensure stable unit operation.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0012] Figure 1 This document shows a flowchart illustrating a high-power pitch and temperature control method adapted to an SL1500 wind turbine, provided in an embodiment of this application. Figure 2 A schematic diagram of a high-power pitch and temperature control device adapted to an SL1500 wind turbine is shown in one embodiment of this application. Detailed Implementation

[0013] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0014] The following is combined with Figure 1 This application describes a high-power pitch and temperature control method adapted to SL1500 wind turbines according to exemplary embodiments. It should be noted that the following application scenarios are shown only to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way. Rather, the embodiments of this application are applicable to any suitable scenario.

[0015] In one implementation, Figure 1A schematic flowchart illustrating a high-power pitch and temperature control method adapted to an SL1500 wind turbine according to an embodiment of this application is shown.

[0016] S101, obtain data on the operating data of the SL1500 unit, data on the pitch system influencing factors, and information on abnormal motor status.

[0017] In one implementation, the data type information for the SL1500 unit focuses on the core operating status parameters of the unit, comprehensively covering key data dimensions associated with the pitch system to ensure that the data can directly support subsequent torque and temperature-related analyses. Pitch motor operating parameters include the real-time torque value, operating temperature, rated current (13.5A), and rated frequency (133.3Hz) of the original 23NMKEB pitch motor, while also recording the corresponding operating parameters of the new 32NM pitch motor to achieve data comparison between the old and new systems. Unit power generation data is recorded at 1-minute sampling intervals (0-1500KW range), synchronously correlated with the pitch operation frequency under different power levels to match the needs of temperature change trend analysis. Operating condition data covers real-time wind speed, wind direction, unit operating time, and number of start-stop cycles, with wind speed divided into intervals of 0.5m / s to clearly define the load impact of different operating conditions on the pitch system.

[0018] The pitch system's influencing factor data revolves around the core cause of increased torque, accurately capturing key factors leading to motor overload and overheating, ensuring each factor has a clear monitoring logic and data source. Blade bearing condition data is collected via vibration sensors, recording bearing operating vibration frequency (Hz) and wear (mm) three times daily to quantify the impact of bearing aging on pitch torque. Pitch drive system lubrication data monitors lubricant viscosity (mm). 2 The system records the following parameters: / s, impurity content (mg / L), replacement cycle (days), and lubrication system pressure (MPa) to determine if the lubrication status meets the standards. It also records the input and output voltage (305-500V range), current (including apparent current and maximum short-time current), and overload frequency of the pitch inverter, with a focus on recording the voltage stability and fault trigger count of the braking circuit.

[0019] The abnormal status information of the motor focuses on the abnormal behavior of the pitch motor during operation, clearly defining the abnormal judgment criteria and data collection methods to provide a basis for subsequent fault risk assessment. Temperature anomaly data: When the pitch motor temperature exceeds 70℃ (the warning threshold after technical upgrade) or approaches 90℃ (the highest abnormal temperature before technical upgrade), the system automatically records the start time, duration, and peak temperature of the anomaly, synchronously linking it to the power generation and torque values ​​at that time. Torque overload anomaly data: When the torque exceeds 23NM (the mild overload threshold) or 50NM (the severe overload threshold), the system records the overload occurrence time, peak torque (up to 70NM), and overload duration, indicating whether it leads to power-limited operation or a fault shutdown of the unit. Example 3: Mechanical motion anomaly data: This records the number of times blade jamming occurs, the pitch angle when jamming occurs, and the torque required to unlock. It also captures abnormal situations such as motor start-up failure and brake failure, recording detailed fault codes and on-site operating condition descriptions.

[0020] S102 performs data preprocessing on data type information to generate basic data on the rated torque and operating temperature of the pitch motor.

[0021] In one implementation, based on the reliability requirements of torque and temperature data during wind turbine retrofitting, as well as equipment aging interference and differences in operating conditions, the monitoring equipment for the pitch system in the data type information is selected and its parameters are determined. Real-time monitoring parameters from the pitch motor torque sensor and continuous acquisition parameters from the temperature sensor are obtained. Based on the core requirement of torque and temperature data reliability during the SL1500 unit retrofitting, and while avoiding measurement deviations caused by equipment aging and parameter fluctuations caused by different operating conditions (such as wind speed 0.5-25m / s and power generation 0-1500KW), suitable monitoring equipment is selected and parameter standards are clearly defined. A high-precision magnetoelectric torque sensor is selected, with a measurement range of 0-100NM (covering a maximum torque of 70NM during faults), an accuracy class of 0.2, and a sampling frequency of 10Hz. This ensures real-time capture of torque changes in both the original 23NMKEB motor and the new 32NM motor, adapting to the pitch torque fluctuation scenarios of the unit. The temperature sensor is a PT100 platinum resistance sensor with a measurement range of -40℃ to 150℃, an error of ≤ ±0.5℃, and a sampling frequency of 5Hz. It is installed at key temperature measurement points on the pitch motor windings and housing to continuously collect motor operating temperature data and meet the monitoring needs of temperature changes from nearly 90℃ to below 70℃ before and after technical upgrades.

[0022] Based on the priority of data usage, the data collected by the two types of sensors were initially screened, prioritizing the accuracy of torque data required for rated torque calculation and the continuity of operating temperature data. According to the priority of data usage, data supporting rated torque calculation and operating temperature trend analysis were listed as core priority. Invalid data was filtered out to ensure the quality of key data. During torque data screening, jump values ​​caused by sensor malfunctions (such as data that instantly jumps from 30NM to 100NM without supporting operating conditions) and zero-value data under shutdown conditions were removed. Real-time torque data during normal unit operation (wind speed ≥3m / s, power generation ≥100KW) was retained to ensure the accuracy of rated torque calculation. For example, the continuous effective torque value of a 23NM motor under an overload condition of around 40NM was selected.

[0023] When filtering temperature data, missing values ​​caused by poor sensor contact and abnormal peak values ​​caused by extreme environmental interference (such as instantaneous high temperature data generated during a lightning strike) are removed. Temperature data with a continuous sampling interval of no more than 10 seconds are retained to ensure the continuity of operating temperature statistics. For example, continuous monitoring data of motor temperature stabilizing below 70℃ after technical modification are fully retained.

[0024] Constraints were applied to ensure data consistency and integrity, as well as to adapt the data to the SL1500 unit's operating scenario. The selected torque data underwent rated torque calculation, and the temperature data underwent operational temperature statistical processing. These were then substituted into the torque-temperature correlation verification formula: Pre-upgrade temperature warning threshold = Base temperature + (Actual torque - 23 NM) × Temperature influence coefficient. Multi-dimensional constraints were applied to ensure data validity. The selected data underwent specific processing, and the correlation verification formula was used to verify logical consistency. Regarding data constraints, the consistency constraint required that the torque and temperature change trends match (torque increases synchronously with temperature, and vice versa), such as removing contradictory data where the torque was stable at 23 NM but the temperature suddenly increased by 20°C. The integrity constraint required that the data loss rate in a single batch not exceed 5%, otherwise, data was re-collected. The compatibility constraint required that the data conform to the SL1500 unit's operating scenario, such as selecting torque and temperature data corresponding to wind speeds of 5-15 m / s (normal operating wind speed).

[0025] In terms of data processing and formula calculation, when calculating torque data, the rated torque is calculated by averaging multiple samples. For example, the effective torque data of a 23NM motor over 10 consecutive minutes is averaged to obtain the actual rated torque. When calculating temperature data, the average temperature, maximum temperature, and temperature rise rate per hour are calculated. When substituting into the torque-temperature correlation verification formula, the base temperature is set to 40℃ and the temperature influence coefficient is 0.8℃ / NM. If the actual torque is 40NM, then the temperature warning threshold before the technical upgrade = 40℃ + (40-23) × 0.8℃ / NM = 53.6℃. This is used to verify whether the temperature data is within a reasonable warning range.

[0026] The integrated processing of rated torque and operating temperature data of the pitch motor generates pre-processed basic data for rated torque and operating temperature, along with quality verification information on torque monitoring data effectiveness and temperature sensor detection error range. The processed torque and temperature data are then structured and integrated, supplemented with data quality verification information to form a complete set of basic data. During data integration, the rated torque and operating temperature data of the pitch motor are aligned by timestamp to generate a three-dimensional basic dataset of "time-torque-temperature," such as: at 10:00 AM on January 23, 2024, the torque of a 23NM motor is 42NM, corresponding to a temperature of 55℃; at 10:00 AM on January 24, 2024, the torque of a 32NM motor is 45NM, corresponding to a temperature of 62℃, clearly showing the relationship between the data before and after the technical upgrade.

[0027] Regarding quality verification information, the validity rate of torque monitoring data is calculated as "valid data volume / total collected data volume × 100%". For example, if a batch collects 1,000 torque data points and 985 are valid, the validity rate is 98.5%. The detection error range of temperature sensors is determined through calibration experiments. For example, the detection error of this batch of PT100 sensors in the 50℃-80℃ range is ±0.3℃. The above verification information is attached to the basic data to ensure data reliability.

[0028] S103 processes influencing factor data and abnormal state information based on a dynamic fusion engine. It strengthens the core influence weights of blade bearing wear and lubrication status through a dynamic weight allocator. Combined with spatiotemporal correlation analysis, it uses a regression model of equipment operation status to calculate the probability of improvement in temperature control of the high-power pitch system. It also compares and analyzes the differences in operation between the improved scheme and the original system by combining data before and after the technical upgrade, and generates pitch system temperature anomaly characteristic identification information.

[0029] In one implementation, the influencing factor data and abnormal state information are processed based on the preset processing rules of the dynamic fusion engine. A dynamic weight allocator strengthens the core influencing weights of blade bearing wear and lubrication status, while simultaneously incorporating secondary influencing factors such as pitch motor brake circuit failure and configuring appropriate weights. According to the preset influencing factor classification processing rules of the dynamic fusion engine, the pitch system influencing factor data (blade bearing wear, lubrication status, brake circuit failure, etc.) and motor abnormal state information (excessive temperature, torque overload, etc.) are classified and integrated. The dynamic weight allocator clarifies the weight ratio of each factor, highlighting the dominant role of the core influencing factors.

[0030] The core influencing factors are weighted as follows: blade bearing wear is weighted at 0.45, and lubrication status at 0.4, totaling 85%, to strengthen its core impact on pitch torque and temperature. Simultaneously, pitch motor brake circuit failure is classified as a secondary influencing factor with an appropriate weight of 0.15, ensuring that secondary factors are not ignored while maintaining the influence of core factors. Data preprocessing includes smoothing abnormal fluctuations in blade bearing wear data (such as sudden increases in wear without changes in operating conditions), using linear interpolation to fill in missing values ​​in lubrication status data, and aligning brake circuit failure records with timestamps to ensure data consistency for subsequent analysis.

[0031] Based on the operating conditions of the SL1500 unit and the temperature control requirements of the pitch system, spatiotemporal correlation analysis rules were designed. These rules focus on the dynamic correlation between torque and temperature under different operating periods and wind speed scenarios, generating a set of correlation features and quantifiable indicators for operating conditions, torque, and temperature. Furthermore, by combining the actual operating conditions of the SL1500 unit (wind speed, power generation, operating period, etc.) with the temperature control objectives of the pitch system, targeted spatiotemporal correlation analysis rules were designed, focusing on the dynamic correlation between torque and temperature under different scenarios, and extracting a quantifiable set of correlation features.

[0032] The spatiotemporal correlation analysis rules are designed to divide the operating time into three periods: early morning (00:00-06:00), daytime (06:00-18:00), and nighttime (18:00-24:00). Based on wind speed, the rules are divided into three scenarios: low wind speed (3-8 m / s), medium wind speed (8-15 m / s), and high wind speed (15-25 m / s). The correlation analysis logic between torque and temperature under different time periods and wind speed combinations is clearly defined. For example, during high wind speed periods, the focus is on monitoring the synchronicity of sudden torque changes and rapid temperature increases. A set of correlation features and quantitative indicators are generated by extracting quantitative indicators such as the average torque, average temperature, and torque-temperature correlation coefficient under the "wind speed 15-25 m / s + daytime period" conditions. This forms a set of operating condition-torque-temperature correlation features. For example, a set of feature data might be: wind speed 18 m / s, torque 45 NM, temperature 68℃, and correlation coefficient 0.92, visually presenting the correlation pattern among the three.

[0033] The system utilizes a regression model of equipment operating status to process the associated feature set. It integrates linear regression and operating condition correction algorithms to quantify the probability of improvement in temperature control by the high-power pitch system, outputting the improvement probability value and confidence interval. Alternatively, the system uses the generated operating condition-torque-temperature associated feature set as input data, integrating linear regression (quantifying the linear correlation between torque and temperature) and operating condition correction algorithms (correcting for interference from different wind speeds and time periods) to accurately calculate the probability of improvement in temperature control and the confidence interval for the high-power pitch system (32NM motor).

[0034] The model adopts a "dual-algorithm collaboration + multi-feature input" architecture. Its core logic is to first establish a basic correlation between torque and temperature through linear regression, and then dynamically adjust deviations using an operating condition correction algorithm to ensure that the calculation results closely match the actual operating scenarios of the SL1500 unit. The model's input data is a set of operating condition-torque-temperature correlation features, including five core features: wind speed range (3-25 m / s), operating period (early morning / daytime / nighttime), mean torque (0-100 NM), mean temperature (-40℃-150℃), and torque-temperature correlation coefficient. Each feature category is accompanied by a data quality verification label (valid / invalid).

[0035] The linear regression algorithm serves as the main model, quantifying the core linear relationship of "torque change → temperature change." The operating condition correction algorithm, as the auxiliary model, dynamically corrects for the interference of different wind speeds and time periods on temperature, with the correction coefficient strongly correlated with the operating condition characteristics. The outputs of both algorithms are weighted and fused (main model weight 0.7, auxiliary model weight 0.3) to obtain the final improvement probability. The output includes the improvement probability value (0-100%), a 95% confidence interval (reflecting the statistical reliability of the results), and an operating condition adaptation label (indicating the wind speed-time combination corresponding to that probability).

[0036] The linear regression algorithm focuses on "torque" as the core influencing factor, constructing a linear correlation equation between torque and temperature to clarify the impact of torque changes on temperature before and after the technical upgrade. Under stable operating conditions, the temperature and torque changes of the SL1500 unit's pitch motor exhibit an approximately linear relationship (excluding abnormal interference such as extreme wind speeds and mechanical failures). The model parameters are configured as follows: Dependent variable Y: Pitch motor operating temperature (°C), selected as the average temperature under the same operating conditions before and after the technical upgrade to avoid the influence of instantaneous fluctuations. Independent variable X: Pitch motor torque value (NM), covering the overload range (23-70NM) of the original 23NM motor and the rated range (0-32NM) of the new 32NM motor. Regularization parameter: L2 regularization (ridge regression) is used, with a regularization strength λ=0.01 to avoid model overfitting (due to a small number of abnormal torque-temperature correlation samples in historical data).

[0037] The data sample consisted of 5000 valid data sets before and after the technical upgrade (a total of 10000 sets), covering the core operating scenarios of medium wind speed (8-15 m / s) and normal operating hours (6:00-18:00 during the day) to ensure sample representativeness. The basic correlation equation was obtained by minimizing the sum of squared residuals between the predicted and actual temperature values ​​using the least squares method: Before the technical upgrade (23NM motor): =35.2+1.2X(R 2 =0.89, indicating that torque can explain 89% of the temperature change); After technical modification (32NM motor): =33.6+0.7X(R) 2=0.91, indicating improved fit (due to the reduced overload amplitude caused by the torque redundancy of the 32NM motor). Before the technical upgrade, for every 1NM increase in torque, the motor temperature rose by an average of 1.2℃; after the upgrade, for every 1NM increase in torque, the temperature rose by an average of 0.7℃, clearly demonstrating the temperature control advantage of the 32NM motor.

[0038] By integrating the set of related features and improvement probability data, the module for comparing and verifying data before and after the technical upgrade is activated. Through difference analysis and significance testing, the differences between the improved scheme and the original system in terms of temperature control and operational stability are quantified, generating information on abnormal temperature characteristics of the pitch system and an inspection report.

[0039] Difference analysis and significance test were performed to compare the temperature data before and after the technical upgrade under the same operating conditions (wind speed 12m / s, power generation 800KW). The average temperature before the upgrade was 78℃, and after the upgrade it was 62℃, with a temperature difference of 16℃. The t-test showed that the P value was <0.05, indicating that the temperature difference before and after the upgrade was statistically significant and the improvement plan was effective.

[0040] Temperature anomaly identification and inspection report generation identifies abnormal characteristics of the original system, such as "temperature rises by 5°C every 10 minutes after torque exceeds 40 NM" and "continuous operation after temperature exceeds 85°C is prone to triggering shutdown failure". The inspection report records in detail key data such as the temperature peak before and after the technical modification (90°C→68°C), the overload temperature control effect (temperature decrease of 18% for slight overload), and the percentage increase in stable running time (30%), fully presenting the implementation effect of the solution and the improvement of temperature anomalies.

[0041] S104 processes the multimodal dynamic feature information to generate the overload heating identification result of the pitch motor.

[0042] In one implementation, the acquired multimodal dynamic feature information (covering dimensions such as torque, temperature, inverter operation, and bearing status) is systematically analyzed. A multi-dimensional feature extraction method is employed to accurately capture the core feature change patterns directly related to motor overload heating. Quantitative analysis clarifies the correlation strength between each feature and overload heating, forming a core feature set and feature association weight information. Core feature extraction and pattern capture include: extracting the difference between the actual torque and rated torque of the original 23NM motor and its duration, such as capturing the change patterns of "torque exceeding 23NM for more than 15 minutes" and "peak value exceeding 50NM"; calculating the temperature rise rate as "temperature increase every 5 minutes," such as the rapid temperature rise pattern of "temperature rising from 60℃ to 75℃ within 5 minutes"; statistically analyzing the daily overload trigger frequency of the inverter; and calculating the bearing friction coefficient using vibration sensor data to capture the correlation pattern of "temperature rising synchronously when friction coefficient > 0.3."

[0043] The correlation metrics and weights are configured by calculating the correlation between each feature and overload heating using the Pearson correlation coefficient. The correlation between torque overload amplitude and temperature rise rate is 0.91, the correlation between frequency converter overload frequency and bearing friction coefficient is 0.75, and the correlation between torque overload amplitude and temperature rise rate and bearing friction coefficient ...

[0044] Based on the operating scenarios of the SL1500 unit and the characteristics of the pitch system, pre-stored overload heating feature screening algorithms and data quality verification rules are invoked to denoise and complete the core feature set of motor overload heating, generating a standardized overload heating feature dataset. Feature denoising involves using a moving average filtering algorithm to remove instantaneous jump values ​​in torque exceeding the standard (such as instantaneous 100NM data caused by sensor malfunction), and using an outlier detection algorithm to remove extreme outliers in temperature rise rates (such as instantaneous temperature rise of 30℃ caused by lightning strikes), retaining valid feature data that conforms to the unit's operating logic.

[0045] Specifically, the core objective of feature denoising is to eliminate abnormal interference data in torque exceedance magnitude and temperature rise rate, while retaining effective features that conform to the physical logic of SL1500 unit operation. This provides high-quality data support for subsequent overload heating identification. A moving average filtering algorithm is used to process instantaneous jumps in torque exceedance magnitude. The moving average filtering algorithm constructs a fixed-length time window, calculates the mean value of the torque data within the window, smooths instantaneous fluctuations, and accurately identifies and eliminates jump values ​​caused by sensor malfunctions. Considering the torque variation characteristics of the SL1500 unit's pitch motor, the time window length is set to 5 seconds (corresponding to a torque sensor sampling frequency of 10Hz, with 50 data points within the window), and a rectangular window type is used (with equal weights for mean calculation) to avoid weighting bias. The real-time collected torque data is sequentially slid into the window in chronological order. The arithmetic mean of the data within the window is calculated, and this mean is used as the filtered torque value at the current moment. At the same time, the jump threshold is set to 30NM (based on historical data where the maximum torque fluctuation during normal operation of the unit is ≤20NM). When the absolute value of the difference between the original torque value and the filtered torque value exceeds 30NM, it is determined to be an instantaneous jump value.

[0046] If the sensor is falsely triggered by electromagnetic interference, resulting in a transient jump sequence of "30NM→100NM→32NM", when the window slides over this sequence, the average data within 5 seconds is calculated to be 35NM. The difference between the original 100NM and the average value reaches 65NM, exceeding the 30NM threshold. Therefore, the 100NM jump data is directly discarded, and the filtered 35NM is retained as valid data to avoid misidentification due to false overload signals.

[0047] Data completion and standardization were performed using the K-nearest neighbor algorithm to complete missing data in the bearing friction coefficient (e.g., 10 minutes of missing data due to brief sensor downtime). All core features were standardized using the normalization formula "eigenvalue / maximum feature value," mapping features such as torque overshoot and temperature rise rate to the [0,1] interval, generating a standardized overload heating feature dataset. This ensured a consistent data format suitable for subsequent analysis. Data completion aimed to fill in missing data caused by brief sensor downtime, while standardization mapped core features of different dimensions and magnitudes to the [0,1] interval, eliminating dimensional differences and ensuring direct fusion analysis by subsequent algorithm models. This was achieved through the following two steps: The K-Nearest Neighbor (KNN) algorithm was used to complete the missing bearing friction coefficient data. Based on the logic that bearing friction coefficients are correlated under similar operating conditions, the KNN algorithm accurately completes the missing values ​​by finding the most similar historical data for the period of missing data. Combining the operating conditions of the SL1500 unit, a K value of 5 was set (selecting 5 most similar operating condition samples), and Euclidean distance was used as the distance metric (quantifying the similarity of multi-dimensional operating conditions). The operating condition feature dimensions included wind speed, power generation, pitch angle, and operating time (all parameters strongly correlated with the bearing friction coefficient).

[0048] First, a historical operating condition-friction coefficient database is constructed, containing operating condition data (wind speed 3-25m / s, power generation 0-1500KW, pitch angle 0-90°, operating time 0-8760 hours) and corresponding bearing friction coefficients during normal operation periods over the past year. When the sensor is briefly offline (e.g., 10 minutes) resulting in missing friction coefficients, the operating condition data for the missing period is extracted (e.g., wind speed 15m / s, power generation 800KW, pitch angle 30°, operating time 5000 hours). The Euclidean distance between this operating condition and all historical operating conditions is calculated in the database. The five historical samples with the smallest distances are selected, and the average friction coefficient of these five samples is used as the fill-in value for the missing period. The sensor was offline for 10 minutes. The missing time period was characterized by "wind speed 14.8 m / s, power generation 790 KW, pitch angle 31°, and operating time 5020 hours". The database was searched and found that the friction coefficients of the five most similar conditions were 0.28, 0.29, 0.27, 0.28, and 0.29, respectively. The calculated average value was 0.282. This value was used as the supplementary data to fill in the missing time period to ensure the continuity of the friction coefficient data.

[0049] The standardized overload heating feature dataset and feature association weight information are integrated, and the real-time motor overload heating analysis module and anomaly pattern mining engine are activated to generate overload heating identification results for the pitch motor. The real-time analysis module calculates the standardized feature data (such as torque exceedance of 0.8, temperature rise rate of 0.75, bearing friction coefficient of 0.6, and inverter overload frequency of 0.3) and feature weights, and substitutes them into the overload heating judgment model to calculate a comprehensive score of 0.72. A scoring threshold of 0.6 is set, and when the comprehensive score is ≥0.6, it is judged that there is an overload heating risk.

[0050] Anomaly pattern mining and result generation: The anomaly pattern mining engine discovers through association rule mining that when "torque exceeds the standard by ≥0.7 and temperature rise rate is ≥0.65", the probability of motor overload heating reaches 92%. Combined with the comprehensive score of the real-time analysis module, the final result of pitch motor overload heating identification is generated, clearly marked as "mild overload heating exists", "severe overload heating exists" or "no overload heating", and also includes core triggering features (such as "torque exceeding the standard by 0.8 is the main triggering factor").

[0051] S105 processes the identification results and risk levels in the impact factor data in conjunction with the multi-task decision matrix. It combines the overload heating identification results of the high-power pitch system and the unit failure risk level with the multi-task decision matrix, and integrates the technical indicators that are independently related to the stable operation of the unit selected in the technical transformation plan to generate the risk impact factor of pitch system overload heating on unit operation.

[0052] In one implementation, the overload heating identification results of the pitch motor and the unit failure risk level are retrieved from the impact factor data to establish the input dimensions of the multi-task decision matrix. Simultaneously, the matrix's row dimension is defined as the severity level of overload heating, and the column dimension as the failure risk level gradient, forming the basic framework for decision analysis. First, two types of core information are extracted from the impact factor data: the overload heating identification results of the pitch motor and the unit failure risk level. These two types of information are then determined as the input dimensions of the multi-task decision matrix. The matrix's row dimension represents the severity of overload heating, categorized into three levels based on the actual operating conditions of the SL1500 unit: mild overload heating (torque 23NM-40NM, temperature not exceeding 70℃), moderate overload heating (torque 40NM-50NM, temperature 70℃-85℃), and severe overload heating (torque ≥50NM, temperature ≥85℃). The matrix's column dimension represents the fault risk level gradient, also categorized into three levels: low risk (only slight temperature increase, not affecting unit operation), medium risk (temperature continues to rise, unit operates with limited power), and high risk (temperature exceeds the limit, unit fails and shuts down). This forms the basis for a complete decision analysis framework.

[0053] If the overload heating identification result of a certain SL1500 unit is "moderate overload heating" (torque 45NM, temperature 78℃) and the unit's fault risk level is "medium risk" (power limitation operation has occurred), then the data is positioned in the decision matrix as "row dimension: moderate overload heating, column dimension: medium risk", which clarifies its analytical position in the framework.

[0054] Based on preset overload-risk correlation weight standards and unit stable operation thresholds, the overload heating identification results corresponding to the high-power pitch system and the unit fault risk level are substituted into a multi-task decision matrix. Preliminary risk assessment results are output through matrix operations. The matrix weights are dynamically adjusted based on past unit fault cases and technical upgrade effect data. The preliminary assessment results must cover the impact coefficients of overload on power generation efficiency and equipment lifespan. First, preset overload-risk correlation weight standards are set, with weight values ​​determined based on historical operating data of the SL1500 unit. The correlation weights for mild overload heating with low, medium, and high risk are 0.1, 0.3, and 0.6, respectively; the correlation weights for moderate overload heating are 0.2, 0.5, and 0.8; and the correlation weights for severe overload heating are 0.3, 0.7, and 1.0. Simultaneously, the unit stable operation thresholds are defined: power generation efficiency is maintained above 90% of the design value, and the equipment lifespan loss rate does not exceed 0.005% per day.

[0055] The overload heating identification results of the high-power pitch system (32NM motor) and the unit failure risk level were substituted into a matrix. Combined with dynamically adjusted weights (based on data from 100 similar past failure cases and 50 unit upgrades, the correlation weight between moderate overload heating and moderate risk was adjusted from 0.5 to 0.45), a preliminary risk assessment result was obtained through matrix multiplication. This result must include the impact coefficients of overload on power generation efficiency and equipment lifespan; the higher the impact coefficient, the greater the negative impact on unit operation.

[0056] The overload heating identification result of a certain generating unit is "mild overload heating", the fault risk level is "medium risk", and the corresponding correlation weight is 0.3. Through matrix operation, the preliminary risk assessment results are: power generation efficiency impact coefficient 0.2 (i.e., power generation efficiency decreases by 20%, from the design value of 95% to 76%), and equipment lifespan impact coefficient 0.15 (i.e., the equipment lifespan loss rate increases to 0.00575% per day), clarifying the specific impact of overload on the core operating indicators of the unit.

[0057] Technical indicators independently related to the stable operation of the unit in the technical upgrade plan were selected and incorporated into the multi-task decision matrix as correction parameters. The preliminary risk assessment results were corrected through indicator weight allocation to generate the risk impact factor of pitch system overload heating on unit operation. Technical indicators independently related to the stable operation of the unit were selected from the technical upgrade plan. These indicators excluded factors directly related to overload heating and fault risk, focusing only on the newly added stability assurance characteristics after the technical upgrade. Specifically, these included: the torque redundancy coefficient of the new 32NM motor (1.39, i.e., the ratio of the rated torque of 32NM to the original 23NM), the inverter overload protection response speed (≤0.1 seconds), and the successful unlocking rate of the pitch jammed motor after the pitch program optimization (≥99%).

[0058] These technical indicators are assigned appropriate weights: torque redundancy coefficient (0.5), inverter overload protection response speed (0.3), and propeller jamming unlocking success rate (0.2). These are then incorporated into the multi-task decision matrix as correction parameters. The preliminary risk assessment results are corrected through weighted calculations. The corrected comprehensive result is the risk impact factor of pitch system overload heating on unit operation. This factor comprehensively reflects the actual risk level of overload heating on unit operation after the technical upgrade.

[0059] Based on the preliminary assessment results above, calculations were performed after incorporating corrected parameters. The torque redundancy coefficient is 0.5 × 0.3 (preliminary risk weight) = 0.15; the inverter overload protection response speed is 0.3 × 0.3 = 0.09; the propeller jamming unlocking success rate is 0.2 × 0.3 = 0.06; and the corrected risk impact factor is 0.15 + 0.09 + 0.06 = 0.3. Combining these impact coefficients, the final risk impact factors are defined as "power generation efficiency impact coefficient 0.12 (a 12% decrease after correction), equipment lifespan impact coefficient 0.09 (a loss rate of 0.00545% per day after correction)," accurately quantifying the actual risk of overload heating after the technical upgrade.

[0060] S106, based on the spatiotemporal correlation early warning engine and combined with the multi-objective dynamic decision-making strategy, processes the identification results, risk impact factors and abnormal feature identification information to generate dynamic early warning results for excessively high pitch motor temperature.

[0061] In one implementation, based on the real-time requirements of pitch motor temperature warning and the needs of safe unit operation, a correlation mapping rule is established between pitch motor overload heating identification results, overload heating risk influencing factors, and temperature anomaly characteristic identification information. This binds the overload heating identification results with the risk influencing factors and temperature anomaly characteristic identification information, generating correlation data between identification, risk, and anomaly. The correlation mapping rule is established around the real-time requirements of pitch motor temperature warning (requiring timely response in the early stages of temperature anomalies) and the needs of safe unit operation (avoiding power limiting or shutdown due to excessively high temperatures). The rule must clearly define the correspondence between different overload heating identification results, risk influencing factors, and temperature anomaly characteristics to ensure a logical closed loop.

[0062] The specific mapping rules are as follows: If the overload heating identification result is "mild overload heating" (torque 23NM-40NM, temperature not exceeding 70℃), and the power generation efficiency influence coefficient and equipment life influence coefficient in the risk impact factors are ≤0.1 and ≤0.08, the corresponding temperature anomaly characteristic is "temperature rises slowly after torque exceeds 23NM, increasing by 3℃ every 15 minutes"; if the identification result is "moderate overload heating" (torque 40NM-50NM, temperature 70℃-85℃), and the two coefficients in the risk impact factors are 0.1-0.2 and 0.08-0.15 respectively, the corresponding temperature anomaly characteristic is "temperature rises rapidly after torque exceeds 40NM, increasing by 5℃ every 10 minutes"; if the identification result is "severe overload heating" (torque ≥50NM, temperature ≥85℃), and the two coefficients in the risk impact factors are both ≥0.2, the corresponding temperature anomaly characteristic is "temperature approaches 90℃, accompanied by torque continuously exceeding 50NM, the unit is about to trigger a fault shutdown".

[0063] The three elements are bound together according to the above rules to generate correlation data for identification, risk, and anomaly. Example: The overload heating identification result of a certain SL1500 unit is "moderate overload heating" (torque 46NM, temperature 79℃), the risk impact factors are "power generation efficiency impact coefficient 0.16, equipment life impact coefficient 0.12", and the temperature anomaly characteristic is "temperature rises by 5.2℃ every 10 minutes under torque 46NM". The correlation data formed after binding is "moderate overload heating - power generation efficiency impact 0.16 + equipment life impact 0.12 - temperature rise of 5.2℃ every 10 minutes".

[0064] Real-time analysis of the correlation data between identification, risk, and anomaly extracts core early warning indicators. Missing data and outliers are filtered and supplemented, and redundant data that does not conform to the unit's operating logic is removed, generating standardized basic early warning data. The generated correlation data is then analyzed in real-time to extract core early warning indicators. These core indicators include six items: overload heating severity level, power generation efficiency impact coefficient, equipment lifespan impact coefficient, temperature rise rate, current temperature value, and peak torque. These indicators directly reflect the severity and potential risks of temperature anomalies.

[0065] During the analysis process, missing data and outliers are handled as follows: if the "temperature rise rate" is missing in a certain related data, it is filled in by the average temperature rise rate within 10 minutes before and after; if an outlier such as "torque peak 120NM" appears, which exceeds the maximum torque of 70NM when the unit is in fault, it is determined to be caused by sensor failure and the data is directly removed; at the same time, redundant data that does not conform to the unit's operating logic is removed, such as contradictory data such as "torque 20NM (not overloaded) but temperature 80℃", as well as invalid data in the shutdown state.

[0066] The processed valid data is formatted according to a unified standard to generate standardized early warning base data. Example: The parsed and processed data is "Overload heating level: moderate overload; power generation efficiency influence coefficient: 0.16; equipment life influence coefficient: 0.12; temperature rise rate: 5.2℃ / 10 minutes; current temperature: 79℃; peak torque: 46NM". All indicators are in a unified format and the data is valid, and can be directly used for subsequent early warning model construction.

[0067] Based on standardized early warning data and a spatiotemporal correlation early warning engine, a multi-dimensional collaborative early warning model is constructed. Combined with the operating timeline of the SL1500 unit, the model achieves spatiotemporal synchronous updates of identification results, risk impact factors, and temperature anomaly characteristic information, generating collaborative early warning data. Using standardized early warning data as input, and combined with the spatiotemporal correlation early warning engine, a multi-dimensional collaborative early warning model is constructed. The model covers three core dimensions: overload heating, risk impact, and temperature anomaly. Each dimension corresponds to a core early warning indicator. A weighted algorithm is used to achieve collaborative analysis of multi-dimensional data, with the following weight allocation: temperature-related indicators (current temperature, rate of temperature rise) account for 0.4, overload-related indicators (overload level, peak torque) account for 0.3, and risk impact indicators (power generation efficiency, equipment lifespan impact coefficient) account for 0.3.

[0068] By combining the operating timeline of the SL1500 unit, data is updated at 1-minute intervals to achieve spatiotemporal synchronization of identification results, risk influencing factors, and temperature anomaly characteristics. For example, if the standardized data is acquired at 10:00, it is updated at 10:01 to "Overload heating level: moderate overload; power generation efficiency impact coefficient: 0.17; equipment lifespan impact coefficient: 0.13; temperature rise rate: 5.3℃ / 10 minutes; current temperature: 80.2℃; peak torque: 47NM". The model calculates the synchronized multi-dimensional data in real time to generate collaborative early warning data.

[0069] Example collaborative early warning data is "Time: 10:01; Overload heating level: moderate overload; Comprehensive risk coefficient: 0.14 (0.16×0.3+0.13×0.3); Temperature risk coefficient: 0.81 (80.2℃ / 90℃ fault temperature×0.4+5.3℃ / 10 minutes÷8℃ / 10 minutes dangerous temperature rise rate×0.4); Collaborative early warning coefficient: 0.95 (Comprehensive risk coefficient + Temperature risk coefficient)", which intuitively presents the collaborative risk level of the current temperature anomaly.

[0070] Combining wind farm operation and maintenance scenarios with multi-objective dynamic decision-making strategies, this paper designs and visualizes collaborative early warning data, and generates dynamic early warning results for excessive pitch motor temperature. Based on the wind farm operation and maintenance scenario (where maintenance personnel need to quickly identify risks and take corresponding measures), a multi-objective dynamic decision-making strategy is adopted to visualize and design collaborative early warning data. The early warning is divided into three levels: Level 1 (low risk): collaborative early warning coefficient 0.3-0.6, no shutdown required, only close monitoring; Level 2 (medium risk): collaborative early warning coefficient 0.6-0.9, unit operating parameters need to be adjusted (e.g., reducing power generation), maintenance personnel need to inspect on-site; Level 3 (high risk): collaborative early warning coefficient ≥0.9, immediate shutdown is required to avoid equipment damage.

[0071] The visualization design uses a real-time data panel to display core early warning indicators, collaborative early warning coefficients, and early warning levels, facilitating quick viewing by maintenance personnel. The final result generates a dynamic early warning for overheating of the pitch motor, which is updated in real-time along with the collaborative early warning data, ensuring the timeliness and accuracy of the warning. Based on the aforementioned collaborative early warning coefficient of 0.95, the generated dynamic early warning result is: "Warning Level: Level 3; Warning Information: The pitch motor is in a moderate overload heating state, with a current temperature of 80.2℃ and a rapid temperature rise rate, which has already significantly impacted power generation efficiency and equipment lifespan. Continued operation will trigger a fault shutdown; Recommended Measures: Immediately shut down the machine and check the blade bearing wear and lubrication status."

[0072] In one implementation, such as Figure 2 As shown, this application also provides a high-power pitch and temperature control device adapted to the SL1500 wind turbine, comprising: The acquisition module 201 is used to acquire data on the operating data of the SL1500 unit, data on the pitch system influencing factors, and information on abnormal motor status. Processing module 202 is used to preprocess data of data types to generate basic data on the rated torque and operating temperature of the pitch motor; it processes the influencing factor data and abnormal state information based on the dynamic fusion engine, strengthens the core influencing weights of blade bearing wear and lubrication status through a dynamic weight allocator, and calculates the probability of the high-power pitch system improving temperature control using a regression model of equipment operation status combined with spatiotemporal correlation analysis. It also compares and analyzes the differences in operation between the improved scheme and the original system by combining data before and after the technical upgrade, and generates pitch system temperature abnormality feature identification information; it processes multimodal dynamic feature information to generate pitch motor overload heating identification results; it processes the identification results and risk levels in the influencing factor data in combination with a multi-task decision matrix, combines the overload heating identification results of the high-power pitch system and the unit fault risk level in combination with the multi-task decision matrix, and integrates the technical indicators that are independently related to the stable operation of the unit selected in the technical upgrade scheme to generate the risk impact factor of pitch system overload heating on unit operation; and it processes the identification results, risk impact factors and abnormal feature identification information based on the spatiotemporal correlation early warning engine and multi-objective dynamic decision strategy to generate dynamic early warning results for excessive pitch motor temperature.

[0073] The computer-readable storage medium provided in the above embodiments of this application and the high-power pitch and temperature control method for SL1500 wind turbines provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0074] The various embodiments in this application are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for evaluating the high-power pitch and temperature control method, electronic device, electronic equipment, and readable storage medium adapted to the SL1500 wind turbine are basically similar to the embodiments of the high-power pitch and temperature control method adapted to the SL1500 wind turbine described above, and are therefore described relatively simply. Relevant parts can be referred to in the descriptions of the embodiments of the high-power pitch and temperature control method adapted to the SL1500 wind turbine described above.

Claims

1. A method for adapting high-power variable pitch and temperature control for SL1500 wind turbines, characterized by, Comprise: Obtain SL1500 unit operation data type information, variable pitch system influence factor data and motor abnormal state information; Data preprocessing is carried out on the data type information to generate variable pitch motor rated torque and running temperature basic data; Based on the dynamic fusion engine, the influence factor data and abnormal state information are processed, the core influence weight of blade bearing wear and lubrication state is strengthened through the dynamic weight distributor, the improvement probability of temperature control of high-power variable pitch system is calculated by using equipment operation state regression model combined with space-time correlation analysis, the operation difference between the improved scheme and the original system is verified and analyzed by comparing the data before and after the technical improvement, and variable pitch system temperature abnormal feature recognition information is generated; The multi-modal dynamic feature information is processed to generate variable pitch motor overload heating identification result; The identification result and risk level in the influence factor data are processed combined with multi-task decision matrix, the overload heating identification result of high-power variable pitch system corresponding to the unit fault risk level is combined with multi-task decision matrix, the technical indexes independently related to stable operation of the unit are screened out in the technical improvement scheme, and the risk influence factor of variable pitch system overload heating on unit operation is generated; Based on the space-time correlation early warning engine combined with multi-objective dynamic decision strategy, the identification result, risk influence factor and abnormal feature recognition information are processed to generate variable pitch motor temperature too high dynamic early warning result.

2. The method of claim 1, wherein, Data preprocessing is carried out on the data type information to generate variable pitch motor rated torque and running temperature basic data, comprising: Based on the demand of wind turbine technical improvement for the reliability of torque and temperature data, and the interference of equipment aging and operation condition difference, the variable pitch system monitoring equipment in the data type information is selected and the parameters are determined, the real-time monitoring parameters of variable pitch motor torque sensor and the continuous acquisition parameters of temperature sensor are obtained; Based on the data use priority, the data collected by the two types of sensors are preliminarily screened, and the accuracy of torque data required for rated torque calculation and the continuity of running temperature data are preferentially guaranteed; The constraint information of data logic uniformity and data integrity, and the constraint information suitable for SL1500 unit operation scene are applied to the screened torque data for rated torque calculation processing and the running temperature statistical processing of temperature data, and the torque-temperature correlation verification formula is substituted, which is technical improvement before temperature early warning threshold = basic temperature + (actual torque-23NM) x temperature influence coefficient); The integrated variable pitch motor rated torque data and running temperature data are processed to generate preprocessed variable pitch motor rated torque and running temperature basic data, and the quality verification information of torque monitoring data efficiency and temperature sensor detection error range is also attached.

3. The method of claim 1, wherein, Based on the dynamic fusion engine, the influence factor data and abnormal state information are processed, the core influence weight of blade bearing wear and lubrication state is strengthened through the dynamic weight distributor, the improvement probability of temperature control of high-power variable pitch system is calculated by using equipment operation state regression model combined with space-time correlation analysis, the operation difference between the improved scheme and the original system is verified and analyzed by comparing the data before and after the technical improvement, and variable pitch system temperature abnormal feature recognition information is generated, comprising: The influence factor data and abnormal state information are processed based on preset processing rules of a dynamic fusion engine, core influence weights of blade bearing wear and lubrication state are strengthened through a dynamic weight distributor, and secondary influence factors of pitch motor brake loop failure are synchronously incorporated and configured with adaptive weights; Time-space correlation analysis rules are designed in combination with operating conditions of the SL1500 unit and temperature control requirements of the pitch system, dynamic correlation dimensions of torque and temperature in different operating periods and wind speed scenarios are focused on, and a working condition-torque-temperature correlation feature set and a quantitative index are generated; A device operating state regression model is called to process the correlation feature set, a linear regression and a working condition correction algorithm are fused to quantify an improvement probability of the temperature control of the high-power pitch system, and an improvement probability value and a confidence interval are output; The correlation feature set and the improvement probability data are integrated, a data comparison and verification module before and after technical improvement is started, differences between the improvement scheme and the original system in temperature control and operating stability are quantified through difference analysis and significance test, and pitch system temperature abnormal feature recognition information and a verification report are generated.

4. The method of claim 1, wherein, Multi-modal dynamic feature information is processed to generate pitch motor overload heating recognition results, including: The acquired multi-modal dynamic feature information is processed, a multi-dimensional feature extraction method is adopted to capture the change rules of torque over-standard amplitude, temperature rise rate, frequency of inverter overload and bearing friction coefficient in motor operation, the correlation degrees of the features and the overload heating are quantified, a motor overload heating core feature set and feature correlation weight information are generated; In combination with operating scenarios of the SL1500 unit and characteristics of the pitch system, a pre-stored overload heating feature screening algorithm and data quality verification rules are called to denoise and complete the motor overload heating core feature set, and a standardized overload heating feature data set is generated; The standardized overload heating feature data set and the feature correlation weight information are integrated, a motor overload heating real-time analysis module and an abnormal rule mining engine are started, and pitch motor overload heating recognition results are generated.

5. The method of claim 4, wherein, The recognition results and risk levels in the influence factor data are processed in combination with a multi-task decision matrix, the pitch motor overload heating recognition results and the unit fault risk levels corresponding to the high-power pitch system are combined with the multi-task decision matrix, the technical indexes independently related to stable operation of the unit that are screened in the technical improvement scheme are fused, and pitch system overload heating risk influence factors on operation of the unit are generated, including: The pitch motor overload heating recognition results and the unit fault risk levels in the influence factor data are called, input dimensions of the multi-task decision matrix are established, the row dimensions of the matrix are classified according to the severity of the overload heating, and the column dimensions are graded according to the fault risk levels, and a decision analysis basic framework is formed; Based on preset overload-risk correlation weight standards and stable operation thresholds of the unit, the pitch motor overload heating recognition results and the unit fault risk levels corresponding to the high-power pitch system are substituted into the multi-task decision matrix, preliminary risk assessment results are output through matrix operation, the matrix weights are dynamically adjusted according to past unit fault cases and technical improvement effect data, and the preliminary assessment results need to cover influence coefficients of the overload on power generation efficiency and equipment life. The technical index independent of the stable operation of the unit in the screening technical improvement scheme is taken as a correction parameter and integrated into the multi-task decision matrix, the preliminary risk assessment result is corrected through index weight distribution, and a risk influence factor of the overheat of the variable pitch system on the operation of the unit is generated.

6. The method of claim 1, wherein, The recognition result, the risk influence factor and the abnormal feature recognition information are processed based on the time-space correlation early warning engine and the multi-target dynamic decision strategy to generate a high temperature of the variable pitch motor dynamic early warning result, including: Based on the real-time of the variable pitch motor temperature early warning and the safety operation demand of the unit, a correlation mapping rule of the overheat recognition result of the variable pitch motor, the overheat risk influence factor and the temperature abnormal feature recognition information is established, the overheat recognition result is bound with the risk influence factor and the temperature abnormal feature recognition information, and correlation data of recognition-risk-abnormality is generated; The correlation data of recognition-risk-abnormality is analyzed in real time, core early warning indexes in the data are extracted, data missing and abnormal values are filtered and completed, redundant data not meeting the unit operation logic is eliminated, and standardized early warning basic data is generated; Based on the standardized early warning basic data and the time-space correlation early warning engine, a multi-dimensional collaborative early warning model is constructed, the time-space synchronous update of the recognition result, the risk influence factor and the temperature abnormal feature information is realized in combination with the SL1500 unit operation working condition timeline, and collaborative early warning data is generated; In combination with the wind farm operation and maintenance scene and the multi-target dynamic decision strategy, the collaborative early warning data is visualized and early warning grading is designed to generate the high temperature of the variable pitch motor dynamic early warning result.

7. A high power pitch and temperature control device for adapting SL1500 wind turbines, characterized by, The device comprises: an acquisition module configured to acquire SL1500 unit operation data type information, variable pitch system influence factor data and motor abnormal state information; a processing module configured to perform data preprocessing on the data type information to generate variable pitch motor rated torque and operation temperature basic data, perform processing on the influence factor data and the abnormal state information based on a dynamic fusion engine, strengthen core influence weights of blade bearing wear and lubrication state through a dynamic weight distributor, calculate an improvement probability of temperature control of a high-power variable pitch system by using an equipment operation state regression model in combination with time-space correlation analysis, generate variable pitch motor overheat recognition results by processing multi-modal dynamic feature information, and process the recognition results and risk levels in the influence factor data in combination with a multi-task decision matrix to generate a risk influence factor of the overheat of the variable pitch system on the operation of the unit by combining the overheat recognition results of the high-power variable pitch system, the unit fault risk levels and the multi-task decision matrix, and integrating the technical index independent of the stable operation of the unit screened in the technical improvement scheme.

8. An electronic device, comprising: It comprises: a first processor; and a memory for storing executable instructions of the first processor; The first processor is configured to execute the executable instructions to perform the method for adapting large-power variable pitch and temperature control of SL1500 wind turbine units according to any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the second processor to implement the method for adapting large-power variable pitch and temperature control of SL1500 wind turbine units according to any one of claims 1-6.

Images