A wind turbine generator efficiency intelligent analysis method and system

Abstract

The present application relates to the field of wind power generation technology, and discloses a kind of wind turbine efficiency intelligent analysis method and system.The method of the present application includes constructing wind energy sensitivity matrix, dividing wind speed and wind direction change rate into two-dimensional grid, calculating the power generation efficiency of each grid;Real-time calculation of wind speed standard deviation and wind direction change rate, generate dynamic efficiency deviation index DEVI to evaluate efficiency;Based on weather forecast and matrix data, predict the power loss in the next 24 hours;Trigger level one early warning and visualize the efficiency attenuation area, while executing yaw compensation, power curve verification, cluster collaborative optimization and model self-verification optimization mechanism, the system includes data acquisition and preprocessing, matrix management, dynamic evaluation, loss prediction, early warning visualization, optimization control and self-verification module;The present application improves the accuracy and real-time performance of efficiency evaluation, reduces power loss, and provides support for fine operation of wind power.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, and in particular to a method and system for intelligent analysis of wind turbine performance. Background Technology

[0002] Wind turbine efficiency analysis is a core technical step in improving wind power generation efficiency and reducing operation and maintenance costs. Current technologies primarily use SCADA systems to collect real-time operational data and combine it with theoretical power curves for efficiency assessment. However, these methods suffer from the following bottlenecks: traditional methods insufficiently consider the dynamic characteristics of wind conditions, treating wind speed as a single variable and failing to effectively quantify the combined impact of wind direction change rate and wind speed stability on power generation efficiency; existing assessment models lack the ability to dynamically respond to the turbine's operating status and cannot reflect in real-time the efficiency degradation caused by changes in environmental factors such as turbulence intensity. These problems lead to a significant discrepancy between the actual operating efficiency of wind turbines and theoretical values, resulting in substantial power loss. Summary of the Invention

[0003] The technical problem to be solved by this invention is that the existing technology does not adequately consider the dynamic characteristics of wind energy conditions and lacks dynamic response to the operating status of the unit. To address this, we propose an intelligent analysis method and system for wind turbine performance.

[0004] To achieve the above objectives, this application adopts the following technical solution: a wind turbine efficiency intelligent analysis method, comprising the following steps: S1. Wind energy sensitivity matrix construction: wind speed range according to Divide the wind direction change rate into 100 intervals. according to Divide into 100 intervals and construct a two-dimensional grid. Calculate the power generation efficiency for each grid. , For grid The average power generation efficiency, To fall into the grid Number of data points The actual power generation of the k-th data point. According to wind speed The power obtained from the theoretical power curve The angle between the wind direction and the direction the fan is facing. S2. Dynamic performance evaluation: Real-time calculation of wind speed standard deviation. (This refers to the wind deviation correction factor.) and wind direction change rate Generate dynamic performance deviation index The value ranges from 0 to 100; the higher the value, the better the performance. ,in The current grid retrieved from the wind energy sensitivity matrix Corresponding efficiency , The standard deviation of real-time wind speed. For the maximum permissible turbulence intensity, This represents the real-time rate of change of wind direction. For the maximum permissible rate of change of wind direction, S3. Power Loss Prediction: Based on weather forecasts and matrix data, predict the power loss in the next h-period. ,in This represents the theoretical power for the h-th time period. To retrieve the efficiency of the grid corresponding to the h-th time period from the wind energy sensitivity matrix, This is the fluctuation sensitivity coefficient. To predict the difference between the standard deviation of wind speed and the current standard deviation of wind speed, Let h be the duration of the h-th time period. To predict the total duration, a fixed value of 24 hours is used. Prediction efficiency... S4. Intelligent Early Warning: (The following text appears to be a list of parameters and their corresponding functions, and is not translated.) Or total loss in 24 hours A level-one early warning is triggered, and the area of ​​performance degradation is visualized through a GIS heat map; and an optimization verification mechanism is executed.

[0005] Preferably, the method for updating the wind energy sensitivity matrix is ​​as follows: Where M represents the amount of new data added this week, and the update cycle is 7 days ± 12 hours. The mesh after the last update efficiency, The actual power of the m-th newly added data. To adjust wind speed based on newly added data The theoretical power of the query.

[0006] Preferably, the maximum permissible turbulence intensity The methods for determining the value include: For direct-drive units: For direct-drive units, the wind speed increases slightly when it exceeds 10 m / s; for doubly-fed units: The maximum speed for a doubly-fed generator unit is no more than 4.0 m / s; among which... The current real-time wind speed is expressed in m / s.

[0007] Preferably, the prediction efficiency Fluctuation sensitivity coefficient The dynamic adjustment method is ,when Time increases Values ​​are used to improve sensitivity to fluctuations. Dynamic performance deviation index for current time.

[0008] Preferably, the visualization of the performance attenuation area comprises: superimposing the GIS map with the unit location coordinates; marking a circular area with a radius of 50 meters, wherein the real-time power generation efficiency of the unit is taken from the wind energy sensitivity matrix ; the performance classification is: green transparent layer when ; yellow translucent layer when ; and red flashing layer when

[0009] Preferably, the first-level early warning triggers the automatic execution of: S61. Generating a blade contamination diagnosis report: when and , calculate the contamination index: , 0.85 is the baseline efficiency of clean blades, and 0.75 is the efficiency threshold when the contamination is severe; S62. Push the cleaning scheme: recommend operation in period, and estimate the power generation recovery value , 0.92 is the expected efficiency after cleaning.

[0010] Preferably, the optimization verification mechanism comprises yaw system error compensation, theoretical power curve verification, wind power cluster collaborative optimization, and model self-verification, wherein the yaw system error compensation step is: S71. Calculate the optimal wind-against angle correction amount: , is the correction angle variable to be optimized, is the predicted wind direction deviation angle of the h period; S72. Calculate the yaw adjustment angle: , wherein is the corrected yaw angle, is the current yaw angle, and k is the corresponding coefficient.

[0011] Preferably, the theoretical power curve verification method: when the actual power is continuously lower than the theoretical value, , automatically fit a new power curve , the coefficient is calculated by the least squares method.

[0012] Preferably, the wind power cluster collaborative optimization: by constructing a field-level sensitivity matrix: , generate a scheduling instruction: when the local , reduce the output of the unit by 20%, and allocate the incremental value to the unit, is the overall average power generation efficiency of the wind farm, N is the total number of units in the wind farm, is the efficiency of the i-th unit, is the theoretical power of the i-th unit.

[0013] Preferably, the model self-verification mechanism: calculate the key indicators every month, when the accuracy index is less than 0.85, execute the start 3-year history data training, and reconfigure the matrix to send the verification report to the management center.

[0014] The technical effects and advantages of the present application: in the present application, the present application quantifies the influence of wind speed and wind direction change on power generation efficiency by constructing a wind energy sensitivity matrix, combines dynamic performance evaluation and power loss prediction, and realizes accurate analysis of wind turbine performance; Introducing yaw compensation, power curve adaptive updating and optimization mechanism, can dynamically correct system error and improve power generation efficiency; Through cluster collaborative optimization to realize reasonable allocation of resources, combined with model self-verification to ensure long-term stability; not only improve the accuracy and real-time performance of performance evaluation, but also reduce power loss through closed-loop optimization, and provide strong support for fine operation of wind power. BRIEF DESCRIPTION OF DRAWINGS

[0015] The disclosure of the present application will be described with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present application. In the drawings, the same reference numerals are used to refer to the same parts:

[0016] Figure 1 The method flowchart of the present application. DETAILED DESCRIPTION

[0017] It is easy to understand that according to the technical scheme of the present application, those skilled in the art can propose a variety of structures and implementation ways which can be replaced with each other without changing the essential spirit of the present application. Therefore, the following specific embodiments and drawings are only exemplary description of the technical scheme of the present application, and should not be regarded as the whole or as the limitation or restriction of the technical scheme of the present application.

[0018] Referring to Figure 1 , the present application provides a technical scheme: a wind turbine performance intelligent analysis method, characterized in that it comprises the following steps: S1. Wind energy sensitivity matrix construction: dividing the wind speed interval into 100 intervals, dividing the wind direction change rate interval into 100 intervals, constructing a two-dimensional grid , i represents the index of the wind speed interval, j represents the index of the wind direction change rate interval, and through combination, a specific cell in the two-dimensional grid can be accurately located, and the power generation efficiency of each grid is calculated , , is the average power generation efficiency of the grid , falling into the grid Number of data points The actual power generation of the k-th data point. To be based on wind speed The power obtained from the theoretical power curve The angle between the wind direction and the direction the fan is facing. S2. Dynamic performance evaluation: Real-time calculation of wind speed standard deviation. (This refers to the wind deviation correction factor.) and wind direction change rate Generate dynamic performance deviation index The value ranges from 0 to 100; the higher the value, the better the performance. ,in The current grid retrieved from the wind energy sensitivity matrix Corresponding efficiency , The standard deviation of real-time wind speed. For the maximum permissible turbulence intensity, This represents the real-time rate of change of wind direction. For the maximum permissible rate of change of wind direction, S3. Power Loss Prediction: Based on weather forecasts and matrix data, predict the power loss in the next h-period. ,in This represents the theoretical power for the h-th time period. To retrieve the efficiency of the grid corresponding to the h-th time period from the wind energy sensitivity matrix, This is the fluctuation sensitivity coefficient. To predict the difference between the standard deviation of wind speed and the current standard deviation of wind speed, Let h be the duration of the h-th time period. To predict the total duration, a fixed value of 24 hours is used. Prediction efficiency... S4. Intelligent Early Warning: When... Or total loss in 24 hours A Level 1 warning is triggered, and the area of ​​performance degradation is visualized through a GIS heat map; and an optimization verification mechanism is executed.

[0019] By dividing the wind speed and wind direction change rate into a two-dimensional grid, the ratio of actual power generation to theoretical power in each grid is statistically analyzed. Combined with the cosine correction of the wind deviation angle, a wind energy sensitivity matrix is ​​established. Real-time data on wind speed standard deviation and wind direction change rate are collected, and the current efficiency status is calculated using the dynamic efficiency deviation index formula. Combined with meteorological forecast data, the efficiency value of the corresponding grid is queried in the matrix, and the power generation efficiency for each future period is predicted through the exponential decay function, thereby calculating the power loss. When the dynamic efficiency deviation index is lower than the preset threshold or the total power loss exceeds the safe range, an early warning is triggered and the efficiency decay area is visualized.

[0020] The application further proposes an updating method of the wind energy sensitivity matrix: Wherein M is the newly added data amount of this week, which is calculated by the actual power and the theoretical power ratio collected by the SCADA system, and is used to reflect the correction demand of the latest operation condition on the power generation efficiency, and the updating period is 7 days ± 12 hours, is the efficiency of the grid after the last update, is the actual power of the mth newly added data, is the theoretical power queried according to the newly added data wind speed

[0021] refers to the grid power generation efficiency after the current update, which is realized by the weighted average of the historical efficiency and the newly added data, the influence of data fluctuation on the stability of the matrix is balanced by retaining 70% of the historical efficiency value and superimposing 30% of the average of the newly added data; the updating period is set to 7 days ± 12 hours, so that the data collection period can be flexibly adjusted by ± 12 hours according to the weather abnormality, for example, when the typhoon passes through and causes data abnormality and loss, the update can be performed after the data is supplemented for 12 hours; through the dynamic updating mechanism, the latest running state can be reflected by the matrix data, and the interference of short-term data noise on the stability of the matrix can be avoided.

[0022] The application further proposes a method for determining the value of the maximum allowable turbulence intensity The method includes: for the direct drive unit: The maximum allowable turbulence intensity of the direct drive unit slightly increases when the wind speed is higher than 10 m / s, which refers to the upper limit of the wind speed fluctuation that the unit can withstand under the current operation condition, which is dynamically calculated by the real-time wind speed data combined with the unit type characteristics, the direct drive unit refers to the unit adopting a permanent magnet synchronous generator direct drive structure, and the wind speed compensation term is included in the calculation formula of the maximum allowable turbulence intensity of the direct drive unit, which is used to reflect the change of mechanical load bearing capacity under high wind speed condition; for the doubly-fed unit: The maximum allowable turbulence intensity of the doubly-fed unit is not more than 4.0 m / s, the doubly-fed unit adopts a doubly-fed asynchronous generator structure, and the upper limit constraint is set in the calculation formula of the maximum allowable turbulence intensity of the doubly-fed unit, which is used to limit the mechanical stress of the transmission system under turbulence condition; wherein is the current real-time wind speed, unit: m / s.

[0023] ​​In the calculation of the dynamic performance deviation index, the maximum allowable turbulence intensity parameter is dynamically generated by acquiring the unit type identifier and current wind speed data in real time and selecting the corresponding calculation formula. For direct-drive units, when the real-time wind speed exceeds 10 m / s, the allowable turbulence intensity will increase by 0.1 m / s for every 1 m / s increase in wind speed. This design considers the mechanical strength redundancy of the direct-drive structure under high wind speeds. For doubly-fed units, a segmented calculation method is used to ensure that the turbulence intensity does not exceed the engineering safety threshold of 4.0 m / s, while reflecting the impact of wind speed changes on the transmission system through a linear relationship. The differentiated calculation method can accurately match the structural characteristics of different types of units and avoid the evaluation deviation caused by using a uniform threshold.

[0024] This application further proposes prediction efficiency. Fluctuation sensitivity coefficient Dynamic adjustment method: ,when Time increases Values ​​are used to improve sensitivity to fluctuations. The current dynamic performance deviation index is used. During the power loss prediction process, the fluctuation sensitivity coefficient is adjusted in real time based on the current dynamic performance deviation index of the unit. When the unit is in a highly efficient and stable state, the dynamic performance deviation index is high. At this time, the fluctuation sensitivity coefficient is increased to make the prediction efficiency more sensitive to wind speed fluctuations and accurately capture the impact of wind speed changes. When the unit is in a state of performance decline, the dynamic performance deviation index is low. At this time, the fluctuation sensitivity coefficient is reduced to decrease the weight of wind speed fluctuations in the prediction model and avoid excessive fluctuations in the prediction results due to the instability of the unit's state. A smooth transition is achieved through an exponential function, where the difference between the dynamic performance deviation index and the benchmark value of 70 is used as an adjustment factor to ensure that the coefficient adjustment range matches the degree of change in the unit's state.

[0025] This application further proposes a visualization method for performance degradation areas: The method involves overlaying the turbine location coordinates onto a GIS map. This overlay matches the actual geographical location information of the wind turbines with the geographic information system coordinates, using a GPS positioning module or geographic coordinate transformation algorithm to accurately mark the turbine distribution on the digital map; using a radius... The circular marking area of ​​rice, in which The real-time power generation efficiency of this unit is taken from the wind energy sensitivity matrix. The performance classification is as follows: It is a green transparent layer at that time; It was a yellow, semi-transparent layer at that time; The time is a red flashing layer.

[0026] Preferably, after the first-level early warning is triggered, the following steps are automatically executed: S61. Generate a leaf pollution diagnosis report: when and , calculate pollution index: , 0.85 is the baseline efficiency of clean blades, 0.75 is the efficiency threshold when heavily polluted; S62. Push cleaning scheme: recommend operation in period, estimate power generation recovery value , 0.92 is the expected efficiency after cleaning.

[0027] Preferably, the optimization verification mechanism includes yaw system error compensation, theoretical power curve verification, wind power cluster collaborative optimization and model self verification, wherein the yaw system error compensation step is: S71. Calculate the optimal wind-against angle correction amount: , is the correction angle variable to be optimized, is the predicted wind direction deviation angle of the h period; S72. Calculate the yaw adjustment angle: , wherein is the corrected yaw angle, is the current yaw angle, and k is the corresponding coefficient.

[0028] Preferably, the theoretical power curve verification method: when the actual power is continuously lower than the theoretical value, , automatically fit a new power curve , the coefficient is calculated by least squares method.

[0029] Preferably, the wind power cluster collaborative optimization: by constructing a field-level sensitivity matrix: , generate scheduling instructions: when the local , reduce the output of the unit by 20%, allocate the incremental value unit, is the overall average power generation efficiency of the wind farm, N is the total number of wind turbine generators in the wind farm, efficiency of the i-th unit, is the theoretical power of the i-th unit.

[0030] The wind power cluster collaborative optimization evaluates the overall efficiency of the whole field in real time by constructing a field-level sensitivity matrix, which associates the efficiency of each unit with its theoretical power, so that the efficiency change of large-capacity units or units in high-wind-speed areas has a greater impact on the overall efficiency index; When detecting that the local unit efficiency is lower than the set threshold, automatically generate scheduling instructions to reduce its output, and at the same time, preferentially allocate the released load to high-efficiency units with efficiency higher than the set threshold, thereby realizing dynamic optimization and configuration of the internal power generation resources of the wind farm.

[0031] The present application further proposes a model self verification mechanism: calculate key indicators every month, when the accuracy index is less than 0.85, execute start 3-year historical data training, and reconstruct the matrix to send a verification report to the management center.

[0032] The model self-verification mechanism sets a fixed cycle of monthly execution, calls stored real-time running data and historical prediction records, calculates the correlation coefficient of the current model prediction power generation efficiency and the actual power generation efficiency as a key indicator; when the correlation coefficient is less than 0.85, the data preprocessing process is automatically started, the historical data containing complete seasonal characteristics in the past three years are screened, and the power generation efficiency prediction model is retrained; after training, the efficiency value in the wind energy sensitivity matrix is reconstructed based on the updated model parameters, and the prediction deviation before and after the model correction, the statistical characteristics of the reconstructed matrix and the recommended operation and maintenance measures are compared to generate a standardized report, which is transmitted to the wind farm central control platform through an encrypted communication protocol.

[0033] The present application further proposes another embodiment, a wind turbine efficiency intelligent analysis system, comprising: a data acquisition and preprocessing module: through the interface of sensors, SCADA systems, weather stations and the like, real-time acquisition of wind turbine running real power generation, equipment state and the like, wind speed, wind direction, wind direction change rate and the like, environmental data and historical data and the like; interface with the weather forecast platform to obtain 24-hour wind speed, wind direction prediction data; clean, complete, time alignment of the collected data, and discretization processing according to 0.25m / s wind speed interval, 0.5° / h wind direction change rate interval, to prepare for matrix construction; a wind energy sensitivity matrix management module: constructs and dynamically updates the mapping relationship between wind energy and power generation efficiency as the core benchmark for efficiency analysis, calculates the average power generation efficiency of each grid according to the wind speed-wind direction change rate two-dimensional grid (100x100) based on the preprocessed historical data; update the matrix every week using the exponential smoothing method to ensure that the matrix adapts to the long-term changes in the performance of the unit; a dynamic efficiency evaluation module: real-time calculation of wind speed standard deviation, wind direction change rate, querying of the corresponding efficiency of the current grid from the wind energy sensitivity matrix, calculation of the dynamic efficiency deviation index, and division of the efficiency level according to the DEVI value; an electric quantity loss prediction module: based on the weather forecast data and the matrix query result, calculating the loss electric quantity of the future period, accumulating the loss electric quantity of each period for 24 hours to obtain the total loss summary; an intelligent early warning and visualization module: used to trigger abnormal early warning and intuitively display the efficiency state, assist operation and maintenance decision, superimpose the unit position on the GIS map, and mark the efficiency area with different colors and radii of circular markers; an optimization control module: automatically generates optimization strategies for efficiency abnormal problems, reduces losses and improves efficiency; a model self-verification module: regularly evaluates the system analysis accuracy to ensure long-term reliability.

[0034] The technical scope of the present application is not limited to the content in the above description, and those skilled in the art can make various modifications and modifications to the above embodiments without departing from the technical idea of the present application, and these modifications and modifications should all belong to the protection scope of the present application.

Claims

1. A wind turbine generator efficiency intelligent analysis method, characterized in that, Comprising the following steps: S1. Wind Energy Sensitivity Matrix Construction: Constructing wind speed ranges according to Divide the wind direction change rate into 100 intervals. according to Divide into 100 intervals and construct a two-dimensional grid. Calculate the power generation efficiency for each grid. , For grid The average power generation efficiency, To fall into the grid Number of data points The actual power generation of the k-th data point. To be based on wind speed The power obtained from the theoretical power curve The angle between the wind direction and the direction the fan is facing. This is the wind deviation correction factor; S2. Dynamic performance evaluation: Real-time calculation of wind speed standard deviation and wind direction change rate , generating a dynamic performance deviation index DEVI, wherein is the current grid queried from the wind energy sensitivity matrix corresponding efficiency , is the real-time wind speed standard deviation, is the maximum allowed turbulence intensity, is the real-time wind direction change rate, is the maximum allowed wind direction change rate, ; S3. Power loss prediction: Predicting the loss of power in the future h period based on weather forecasts and matrix data: wherein is the theoretical power in the h period, is the efficiency of the corresponding grid in the h period queried from the wind energy sensitivity matrix, is the fluctuation sensitivity coefficient, is the difference between the predicted wind speed standard deviation and the current wind speed standard deviation, is the length of the h period, is the predicted total length, a fixed value of 24h, and the predicted efficiency is the predicted power generation efficiency in the h period; S4. Intelligent early warning: when or 24-hour total loss trigger a first-level warning, and the performance degradation area is visualized by GIS heat map. And perform optimization verification mechanism, optimization verification mechanism includes yaw system error compensation, theoretical power curve verification, wind power cluster collaborative optimization and model self verification.

2. The method of claim 1, wherein, The method for updating the wind energy sensitivity matrix is ​​as follows: Where M represents the amount of new data added this week, and the update cycle is 7 days ± 12 hours. The mesh after the last update efficiency, The actual power of the m-th newly added data. To adjust wind speed based on newly added data The theoretical power of the query.

3. The method of claim 1, wherein, The maximum allowable turbulence intensity The value method includes: for direct drive unit: , the wind speed of direct drive unit is slightly increased when it is higher than 10 m / s; for double-fed unit: , the maximum wind speed of double-fed unit is not more than 4.0 m / s; wherein is the current real-time wind speed, unit: m / s.

4. The method of claim 1, wherein, The predicted efficiency The fluctuation sensitivity coefficient The dynamic adjustment method is When DEVI > 70, increase The value to improve the sensitivity to fluctuations, DEVI is the dynamic efficiency deviation index of the current moment.

5. The method of claim 1, wherein, The visualization of the performance decay area includes: superimposing the GIS map with the unit location coordinates; using a circular marker area with a radius of 1 km, where The performance decay area is colored in green, yellow or red, depending on the real-time performance of the unit, taken from the wind energy sensitivity matrix The performance classification is: Green transparent layer when the performance is above 90%; Yellow semi-transparent layer when the performance is between 70% and 90%; Red flashing layer when the performance is below 70%.​ 6. The method of claim 1, wherein, The first early warning trigger automatically executes: S61. Generate a blade contamination diagnosis report: when and DEVI < 60, calculate the contamination index: , 0.85 is the baseline efficiency of clean blades, and 0.75 is the efficiency threshold when severely contaminated. S62. Push cleaning solution: recommend in Periodic operation, estimated power generation recovery value 0.92 is the expected efficiency after cleaning.

7. The method of claim 1, wherein, The yaw system error compensation step is: S71. calculating the optimal wind-aiming angle correction amount: , is a correction angle variable to be optimized, is a predicted wind direction deviation angle of an hth time period; S72. calculating a yaw adjustment angle: , wherein is a corrected yaw angle, is a current yaw angle, and k is a corresponding coefficient.

8. The method of claim 7, wherein, The theoretical power curve verification method: when the actual power is continuously lower than the theoretical value, , automatically fitting a new power curve , the coefficient calculation uses the least square method.

9. The method of claim 8, wherein, The wind power cluster collaborative optimization: by constructing field level sensitivity matrix: , generating scheduling instructions: when local , reduce the output of the unit 20%, distribution increment value unit, for the overall average power generation efficiency of the wind farm, N is the total number of wind farm units, for the efficiency of the i unit, for the theoretical power of the i unit.

10. The method of claim 9, wherein, The model self verification mechanism: calculate key indicators every month, when the accuracy index is less than 0.85, start 3-year historical data training and reconstruct the matrix to send verification report to the management center.

Images