An offshore wind farm power prediction evaluation method, device, equipment and medium

Abstract

This invention relates to the field of new energy power generation technology, and discloses a method, device, equipment, and medium for power prediction and evaluation of offshore wind farms. The evaluation method provided by this invention first uses preprocessing methods to ensure data quality based on real-time acquired operating data of each turbine in the target wind farm. Then, using a scenario-adaptive parameter generation method, a matching set of evaluation hyperparameters is dynamically output based on the preprocessed operating data of each turbine. Next, by employing a state-environment perception evaluation method and using the prediction error of each turbine and the comprehensive evaluation weight, the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results is significantly reduced. Simultaneously, effective interval constraints are introduced to eliminate invalid prediction periods under extreme sea conditions. Finally, the weighted calculation yields the power prediction accuracy evaluation result for the corresponding target wind farm, ensuring that the final result truly reflects the prediction performance of wind turbines under complex operating conditions.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation technology, specifically to a method, device, equipment, and medium for predicting and evaluating the power of offshore wind farms. Background Technology

[0002] The offshore wind power industry is developing rapidly, and wind turbine power forecasting plays a crucial role in grid dispatching, wind power grid connection and operation, and power generation planning. Power forecasting evaluation methods, as an important means to measure forecasting effectiveness and optimize forecasting models, are widely used in the field of wind power operation and management.

[0003] In the wind farm power prediction and evaluation methods disclosed in related technologies, the prediction effect is quantitatively analyzed by using conventional statistical indicators based on directly collected actual power and predicted power as the basic data.

[0004] However, offshore wind farms are susceptible to various factors during operation, such as wind speed fluctuations, sea state changes, and turbine operating status switching. Operating conditions vary significantly at different times and under different conditions. Therefore, existing power prediction and evaluation methods struggle to accurately reflect the power prediction performance of offshore wind farms under complex operating conditions. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for predicting and evaluating the power of offshore wind farms, in order to solve the problem that power prediction and evaluation methods in related technologies are difficult to accurately reflect the power prediction effect of offshore wind farms under complex operating conditions.

[0006] In a first aspect, the present invention provides a method for predicting and evaluating the power output of offshore wind farms, the method comprising: Based on the real-time acquired operating data of each unit in the target wind farm, a preprocessing method is used to obtain preprocessed operating data of multiple units; the operating data includes actual power, predicted power, operating conditions, unit health data, environmental data and wind speed data; Based on the preprocessed operating data of each unit, the evaluation hyperparameter set of each unit in the target wind farm is obtained by using the scenario adaptive parameter generation method. Based on the preprocessed operating data of each unit and the corresponding set of evaluation hyperparameters, the state environment perception evaluation method is used to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm. By combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, the power prediction accuracy evaluation result of the target wind farm is obtained using the weighted accuracy calculation method.

[0007] By adopting the above implementation method, firstly, based on the real-time acquired operating data of each unit in the target wind farm, the data quality is ensured through preprocessing methods; then, using a scenario-adaptive parameter generation method, a matching set of evaluation hyperparameters is dynamically output based on the preprocessed operating data of each unit; next, by adopting a state-environment perception evaluation method, and by using the prediction error of each unit and the comprehensive evaluation weight, the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results is significantly reduced; at the same time, effective interval constraints are introduced to eliminate invalid prediction periods under extreme sea conditions; finally, the power prediction accuracy evaluation result of the corresponding target wind farm is obtained through weighted calculation, ensuring that the final result truly reflects the prediction performance of wind turbine units under complex operating conditions.

[0008] In one optional implementation, the step of obtaining the prediction error, comprehensive evaluation weight, and effective evaluation interval constraint for each unit in the target wind farm based on the preprocessed operating data of each unit and the corresponding set of evaluation hyperparameters, using a state-environment perception evaluation method, includes: Based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, the prediction error of the corresponding unit at multiple times is obtained by using the wind condition sensing power error evaluation method. Based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, the comprehensive evaluation weight of the corresponding unit and the prediction error at each time moment is obtained by using the multi-dimensional state-aware weighting method. Based on the preprocessed environmental data of each unit, the effective evaluation interval constraints of the corresponding unit are obtained by using the extreme sea state identification method.

[0009] By adopting the above implementation method, firstly, based on the preprocessed wind speed, actual power, and predicted power of each unit, and combined with scenario adaptive hyperparameters, the prediction error of the corresponding unit at each moment is calculated using the wind condition-aware power error evaluation method, avoiding interference from extreme values. Secondly, based on the actual power, operating conditions, and unit health data of each unit, a multi-dimensional state-aware weighting method is adopted, integrating power level weight, operating condition weight, and health weight to obtain the comprehensive evaluation weight of the corresponding unit at each moment, effectively reducing the error contribution of power curtailment, low power, and non-steady-state operating conditions. At the same time, based on environmental data, an effective evaluation interval constraint is generated through an extreme sea state identification method, eliminating invalid prediction periods under extreme sea states such as typhoons and swells. This ensures that the final weighted power prediction accuracy evaluation result is continuous and stable, improving the accuracy of the final power prediction evaluation.

[0010] In one optional implementation, based on the preprocessed wind speed data, actual power, and predicted power of each unit, and combined with the evaluation hyperparameter set of the corresponding unit, the wind condition-aware power error evaluation method is used to obtain the prediction error of the corresponding unit at multiple times, including: Based on the evaluation hyperparameter set and wind speed data of each unit, the dynamic generalization of the corresponding unit at multiple moments is obtained by using the wind speed dynamic adjustment evaluation function. Based on the absolute error between the actual power and the predicted power of each unit at each moment, and combining the actual power and dynamic normalization term of the corresponding unit at the corresponding moment, a predefined error evaluation method is used to obtain the prediction error of the corresponding unit at multiple moments.

[0011] By adopting the above implementation method, based on the evaluation hyperparameter set and wind speed data of each unit, the dynamic normalization term of the corresponding unit at each time is calculated using the wind speed dynamic adjustment evaluation function; then, based on the absolute error between the actual power and the predicted power of each unit at each time, the prediction error is calculated by combining the actual power of the corresponding unit with the dynamic normalization term; when the power approaches zero, the normalization term is used to ensure that the error is bounded; when the power is high, the error is smoothly degraded into a standard relative error to maintain physical comparability; finally, by utilizing the adaptive adjustment characteristic of the dynamic normalization term with wind speed changes, greater fault tolerance is given in the low wind speed and high randomness range, and reasonable reduction is made in the high wind speed and power-limited range, thereby ensuring that the final prediction error truly reflects the power prediction performance of each unit under complex wind conditions.

[0012] In one optional implementation, the comprehensive evaluation weight matching the prediction error of the corresponding unit with each time moment is obtained by combining the preprocessed wind speed data, actual power, and predicted power of each unit with the evaluation hyperparameter set of the corresponding unit and using a multi-dimensional state-aware weighting method, including: Based on the actual power of each unit at each time and the evaluation hyperparameter set of the corresponding unit at the corresponding time, the power sensing weight of each unit at the corresponding time is obtained by using the power sensing weight calculation method. Based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the operating condition perception weight of each unit at the corresponding time is obtained by using the operating condition perception weight calculation method. Based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the health perception weight of each unit at the corresponding time is obtained by using the health perception weight calculation method. The power perception weight, operating condition perception weight, and health perception weight of each unit at each time are multiplied to obtain the comprehensive evaluation weight that matches the prediction error of the corresponding unit at the corresponding time.

[0013] By adopting the above implementation method, based on the actual power and evaluation hyperparameter set of each unit at each moment, the power perception weight at each moment is obtained using the power perception weight calculation method, so that the error contribution at high power moments is greater, reflecting the energy value; then, based on the operating conditions and evaluation hyperparameter set of each unit at each moment, corresponding weights are assigned to states such as power curtailment and grid connection start-up and shutdown, reducing the interference of unsteady-state conditions; then, based on the unit health data and hyperparameters of each unit at each moment, the power curve deviation caused by unit aging is corrected through health perception weight; finally, by multiplying the three to obtain a comprehensive evaluation weight that matches the prediction error, the joint perception of power level, operating conditions and equipment health is realized, ensuring that the final power prediction accuracy evaluation result of each unit is continuous and stable across the entire power range.

[0014] In one optional implementation, the step of combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, and using a weighted accuracy calculation method to obtain the power prediction accuracy evaluation result of the target wind farm, includes: Based on the effective evaluation interval constraint for each unit, the prediction error and comprehensive evaluation weight of the corresponding unit at each time point are screened to obtain the prediction error and comprehensive evaluation weight of multiple effective time points. Based on the prediction error and comprehensive evaluation weight of each unit at each effective moment, the weighted error of the corresponding unit at a single moment is obtained by using the weighted multiplication statistical method. Based on the single-time weighted error of multiple effective times for each unit, a comprehensive method is used to sum them to obtain the total weighted error of each unit. The total weight of each unit is obtained by summing the comprehensive evaluation weights for the entire effective time interval. Based on the total weighted error and total weight of each unit, the power prediction accuracy of each unit in the target wind farm is obtained using the accuracy evaluation method, and the power prediction evaluation result of the target wind farm is obtained.

[0015] By adopting the above implementation method, firstly, based on the effective evaluation interval constraint of each unit, the effective time points under normal sea conditions for each unit are screened out, eliminating invalid interference from extreme sea conditions such as typhoons and swells. For each effective time point, the prediction error is multiplied by the comprehensive evaluation weight to obtain the single-time weighted error, which reasonably reduces the error contribution under non-ideal operating conditions such as low power, power curtailment, grid connection start-up and shutdown, and aging, while highlighting the error contribution during high-power normal power generation. Finally, the single-time weighted errors within the effective interval of the entire time period are summed to obtain the total weighted error, and the total weight is obtained by summing the comprehensive evaluation weights. Finally, the power prediction accuracy is calculated through the accuracy evaluation method to obtain a power prediction evaluation result consistent with the actual operating value.

[0016] In one optional implementation, the step of obtaining a set of evaluation hyperparameters for each unit in the target wind farm by using a scenario adaptive parameter generation method based on the preprocessed operating data of each unit includes: Based on the preprocessed unit operation data of each unit, the operation scenario vector of the corresponding unit is obtained by using the scenario vector construction method; Based on the operating scenario vector of each unit, an evaluation hyperparameter set matching the corresponding unit is obtained using the hyperparameter adaptive mapping method.

[0017] By adopting the above implementation method, based on the preprocessed actual power, wind speed, operating conditions, unit health and environmental data of each unit, the operating scenario vector of each unit is first generated using the scenario vector construction method to achieve a unified characterization of the real-time operating status of the wind farm; then, based on the operating scenario vector of each unit, the hyperparameter adaptive mapping method is used to dynamically output the evaluation hyperparameter set that matches the current operating conditions of the corresponding unit, so that the evaluation parameters are automatically adjusted according to wind speed fluctuations, unit aging degree and power curtailment conditions, ensuring that the evaluation indicators of each unit are adaptable to different wind conditions, different life stages and different operating modes, and improving the robustness of the evaluation method in complex marine environments.

[0018] In one optional implementation, the preprocessing method is used to obtain preprocessed operating data for multiple units based on the real-time acquired operating data of each unit in the target wind farm, including: Based on the real-time acquired operating data of each unit in the target wind farm, a time synchronization method is used to perform unified time axis alignment to obtain the time-synchronized operating data of each unit. Based on the time-synchronized operating data of each unit, threshold judgment and statistical discrimination methods are used to filter the data to obtain the operating data of each unit after removing outliers. Based on the operational data of each unit after removing outliers, interpolation or neighboring values ​​are used to fill in the missing data for the corresponding time periods, resulting in operational data with missing values ​​filled in for each unit, which is then output as the preprocessed operational data for the corresponding unit.

[0019] By adopting the above implementation method, the operating data of each unit acquired in real time is first synchronized in time, unifying the time axis of data such as actual power, predicted power, operating conditions, unit health, environment and wind speed, and eliminating time deviation of multi-source data; then, threshold judgment and statistical discrimination methods are used to identify and remove outliers to avoid interference introduced by sensor failure or communication errors; then, missing time periods are filled in by interpolation or neighboring values ​​to ensure the continuity and integrity of the data sequence, providing reliable basic data for subsequent scene vector construction, hyperparameter adaptive mapping, prediction error calculation and comprehensive evaluation weight generation, significantly improving the stability and accuracy of the entire power prediction and evaluation method in complex marine environments.

[0020] Secondly, the present invention provides an offshore wind farm power prediction and evaluation device, the device comprising: The data processing module is used to obtain preprocessed operating data of multiple units based on the real-time acquired operating data of each unit in the target wind farm using preprocessing methods; the operating data includes actual power, predicted power, operating conditions, unit health data, environmental data and wind speed data; The parameter construction module is used to obtain the set of evaluation hyperparameters for each unit in the target wind farm by using the scenario adaptive parameter generation method based on the preprocessed operating data of each unit. The parameter evaluation module is used to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm based on the preprocessed operating data of each unit and the corresponding set of evaluation hyperparameters, using the state environment perception evaluation method. The results generation module is used to integrate the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, and use the weighted accuracy calculation method to obtain the power prediction accuracy evaluation result of the target wind farm.

[0021] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the offshore wind farm power prediction and evaluation method of the first aspect or any corresponding embodiment described above.

[0022] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the offshore wind farm power prediction and evaluation method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0023] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first process of the offshore wind farm power prediction and evaluation method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the second process of the offshore wind farm power prediction and evaluation method according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the third process of the offshore wind farm power prediction and evaluation method according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the fourth process of the offshore wind farm power prediction and evaluation method according to an embodiment of the present invention; Figure 6 This is a structural block diagram of an offshore wind farm power prediction and evaluation device according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0027] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0028] As an optional application scenario of this invention, such as Figure 1 As shown, the terminal device 110 is equipped with a power prediction and evaluation application 101. The user 130 can interact with the power prediction and evaluation application 101 through the terminal device 110 and / or the access device of the terminal device 110 to perform wind turbine power prediction and evaluation related operations.

[0029] For example, the power prediction and evaluation application 101 can provide wind turbine power prediction and evaluation related services, which may include functions such as inputting pre-processed operating data, viewing prediction errors, displaying evaluation results, and adjusting parameters. For example, it supports users to input the operating data of the target wind farm, view real-time prediction errors, retrieve valid evaluation intervals and final accuracy results, etc., which corresponds to the core requirements of the power prediction and evaluation method in this application.

[0030] exist Figure 1 In the application scenario shown, when the power prediction and evaluation application 101 is active, the terminal device 110 can display the operation interface 102 of the application 101. The interface 102 may include a running data input page, a prediction error display page, an evaluation result query page, a parameter setting page, etc., to meet the user's various operation needs for wind turbine power prediction and evaluation.

[0031] In some embodiments, the terminal device 110 is communicatively connected to the server 120 to achieve a stable supply of power prediction and evaluation services. The terminal device 110 may be a mobile terminal, fixed terminal, or portable terminal, including but not limited to mobile phones, desktop computers, laptops, industrial control terminals, etc., and may also include accessories and peripherals of the aforementioned terminal devices. The server 120 may be a computing system or server capable of providing computing power, including but not limited to mainframes, edge computing nodes, computing devices in cloud environments, etc., used to store wind farm operation data, process prediction error calculations, and generate power prediction and evaluation results.

[0032] It should be noted that, Figure 1 This is merely one application scenario example of the power prediction and evaluation method of this application, and does not limit the scope of protection of this invention. The power prediction and evaluation method of this application can be adapted to various terminal devices and server communication modes, meeting the actual operation and maintenance needs of different wind farms.

[0033] The embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the operation interface shown in the drawings is merely an example. In practice, different interface layouts and functional modules can be designed according to the operation and maintenance needs of wind farms. The various operation elements on the page can be arranged, omitted, or replaced according to actual needs, and no limitations are made in the embodiments of the present invention. Furthermore, the embodiments described below mainly refer to the terminal device 110. It should be understood that the operations described relative to the terminal device 110 can be executed by the power prediction and evaluation application 101 on the terminal device 110, or jointly executed by the application 101 and its server (e.g., server 120), corresponding to the offshore wind farm power prediction and evaluation method in this application.

[0034] The wind power prediction and evaluation methods disclosed in related technologies mostly adopt traditional statistical indicators such as mean absolute error, root mean square error, and mean absolute percentage error, which have the following limitations in practical engineering applications: On the one hand, the publicly available power prediction and evaluation methods in related technologies do not adaptively adjust to the actual operating scenarios of wind turbines. The evaluation parameters are fixed and singular, and cannot be dynamically adapted according to information such as wind speed characteristics, power levels, and unit operating status. This leads to distorted evaluation results under complex operating conditions such as low wind speed, high fluctuations, power curtailment, and unit start-up and shutdown, making it difficult to objectively reflect the true performance of the prediction model. On the other hand, the publicly available wind power prediction and evaluation methods in related technologies do not fully consider the impact of complex marine environments and do not identify and eliminate periods without effective power generation significance, such as extreme sea states. This easily incorporates invalid operating condition errors into the evaluation system, reducing the reliability and engineering applicability of the evaluation results.

[0035] To overcome the shortcomings of wind power prediction and evaluation methods in the aforementioned related technologies, this application provides a method for predicting and evaluating offshore wind farm power. First, based on real-time acquired operational data of the target wind farm, a preprocessing method is used to ensure data quality. Then, a scenario-adaptive parameter generation method is used to dynamically output a matching set of evaluation hyperparameters based on the preprocessed operational data. Next, a state-environment awareness evaluation method is adopted to ensure that the error is bounded when the power approaches zero and degenerates to a standard relative error when the power is high. A comprehensive evaluation weight is calculated by jointly using power level weight, operating condition weight, and turbine health weight, significantly reducing the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results. Simultaneously, an effective interval constraint is introduced to eliminate invalid prediction periods under extreme sea conditions. Finally, the weighted calculation of the power prediction accuracy evaluation result ensures that the final result truly reflects the prediction performance of the wind turbine under complex operating conditions.

[0036] According to an embodiment of the present invention, an embodiment of a method for predicting and evaluating the power of offshore wind farms is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0037] This embodiment provides a method for predicting and evaluating the power output of offshore wind farms, which can be used in the aforementioned wind farm terminal equipment. Figure 2 This is a flowchart of an offshore wind farm power prediction and evaluation method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: S201, based on the real-time acquired operating data of each unit in the target wind farm, uses a preprocessing method to obtain preprocessed operating data of multiple units; the operating data includes actual power, predicted power, operating conditions, unit health data, environmental data and wind speed data.

[0038] The preprocessing method involves performing time-series alignment, outlier removal, and missing value completion on the raw operating data of each unit acquired in real time, in order to ensure the quality and consistency of the preprocessed operating data.

[0039] Actual power is the active power actually generated by the wind turbine or wind farm.

[0040] Predicted power is an estimated value of active power given in advance based on future meteorological data using a power prediction model.

[0041] Operating conditions refer to the operating status of wind turbine units, including normal power generation, power-limited operation, grid connection, or shutdown.

[0042] Unit health data are indicators used to reflect the aging degree, health status, or available capacity of a unit. Unit health data includes: health index and years of service.

[0043] Environmental data represents information about the external environmental characteristics that affect the operation of wind turbines, including data on wave height, ocean current speed, typhoons, swells, and other extreme sea conditions.

[0044] Wind speed data refers to the real-time or predicted wind speed value at the hub height of the wind turbine, which is used to evaluate the predicted power of the wind turbine.

[0045] By using preprocessing methods to preprocess the real-time acquired operational data, reliable basic data is provided for subsequent scenario vector construction, hyperparameter adaptive mapping, prediction error calculation, and comprehensive evaluation weight generation for each unit, significantly improving the stability and accuracy of the entire power prediction and evaluation method in complex marine environments.

[0046] S202, based on the preprocessed operating data of each unit, uses the scenario adaptive parameter generation method to obtain the set of evaluation hyperparameters for each unit in the target wind farm.

[0047] The scenario adaptive parameter generation method is a mapping method that dynamically outputs a matching set of evaluation hyperparameters based on the current wind farm operation scenario vector. This method is used to adjust for the different wind speeds and fluctuations caused by the different locations of each turbine in the target wind farm, thereby ensuring the accuracy of the power prediction results for each turbine in an objective evaluation.

[0048] The evaluation hyperparameter set is a set of adaptively adjustable parameters used to control subsequent error normalization intensity, power weight, operating condition weight, and health weight based on the unit's condition.

[0049] S203, based on the preprocessed operating data and evaluation hyperparameter set of each unit, uses the state-environment perception evaluation method to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm.

[0050] The environmental condition perception evaluation method is an evaluation system that integrates wind conditions, operating conditions, unit health, and environmental constraints, and calculates prediction errors, comprehensive evaluation weights, and effective evaluation interval constraints.

[0051] Prediction error is the deviation between the predicted power and the actual power at each moment, and is a bounded relative error value obtained after soft normalization.

[0052] The comprehensive evaluation weight is a composite weight obtained by multiplying the power perception weight, operating condition perception weight, and health perception weight. It is used to measure the contribution of the error at the corresponding time point to the final evaluation result.

[0053] The effective evaluation interval constraint is a binary identifier generated based on the extreme sea state identification method. It is set to 1 under normal sea state and 0 under extreme sea state, and is used to remove data during extreme sea state periods such as typhoons and swells.

[0054] S204. By combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, the power prediction accuracy evaluation result of the target wind farm is obtained using the weighted accuracy calculation method.

[0055] The weighted accuracy calculation method is to multiply the prediction errors of the effective times of multiple units in the target wind farm by the comprehensive evaluation weights, sum them, and then divide by the total weights to finally obtain the power prediction accuracy evaluation result.

[0056] The State-Environment-aware Power Accuracy Score (SEPAS) is used to reflect the prediction performance of wind farms under complex operating conditions; the higher the value, the more accurate the prediction.

[0057] Specifically, S204 above includes: a1. Based on the effective evaluation interval constraint of each unit, the prediction error and comprehensive evaluation weight of the corresponding unit at each time point are screened to obtain the prediction error and comprehensive evaluation weight of multiple effective time points. a2, based on the prediction error and comprehensive evaluation weight of each unit at each effective moment, the weighted error of the corresponding unit at a single moment is obtained by using the weighted multiplication statistical method; a3, based on the single-time weighted error of multiple effective times for each unit, the total weighted error of each unit is obtained by summing the results using a comprehensive method; a4, sum the comprehensive evaluation weights of each unit within the effective time interval to obtain the total weight of each unit; a5. Based on the total weighted error and total weight of each unit, the power prediction accuracy of each unit in the target wind farm is obtained using the accuracy evaluation method, and the power prediction evaluation result of the target wind farm is obtained.

[0058] For example, the power prediction accuracy evaluation results for each of the above units satisfy the following:

[0059] in, This represents the evaluation result of the power prediction accuracy of any given unit. This represents the prediction error at time t for the corresponding unit; This represents the effective interval evaluation constraint value for the corresponding unit at time t. The value represents the comprehensive evaluation weight of the corresponding unit at time t; T represents the total time period corresponding to the power prediction evaluation of the corresponding unit.

[0060] First, based on the effective evaluation interval constraints for each unit, effective time points under normal sea conditions are selected, eliminating invalid interference from extreme sea conditions such as typhoons and swells. For each effective time point, the prediction error is multiplied by the comprehensive evaluation weight to obtain the single-time weighted error, which reasonably reduces the error contribution under non-ideal operating conditions such as low power, power curtailment, grid connection start-up and shutdown, and aging, while highlighting the error contribution during high-power normal power generation. Finally, the single-time weighted errors within the effective time interval are summed to obtain the total weighted error, and the total weight is obtained by summing the comprehensive evaluation weights. Finally, the power prediction accuracy is calculated using the accuracy evaluation method to obtain a power prediction evaluation result consistent with the actual operating value.

[0061] The offshore wind farm power prediction and evaluation method provided in this embodiment firstly ensures data quality through preprocessing based on real-time acquired target wind farm operation data. Then, using a scenario-adaptive parameter generation method, a matching set of evaluation hyperparameters is dynamically output based on the preprocessed operation data. Next, by employing a state-environment awareness evaluation method, the error is bounded when power approaches zero and degrades to a standard relative error at high power. A comprehensive evaluation weight is calculated jointly using power level weight, operating condition weight, and turbine health weight, significantly reducing the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results. Simultaneously, an effective interval constraint is introduced to eliminate invalid prediction periods under extreme sea conditions. Finally, the weighted calculation of the power prediction accuracy evaluation result ensures that the final result truly reflects the prediction performance of the wind turbine under complex operating conditions.

[0062] This embodiment provides a method for predicting and evaluating the power output of offshore wind farms, which can be used in the aforementioned wind farm terminal equipment. Figure 3 This is a flowchart of an offshore wind farm power prediction and evaluation method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: S301 uses preprocessing methods to obtain preprocessed operating data of multiple units based on real-time acquired operating data of each unit in the target wind farm. The operating data includes actual power, predicted power, operating conditions, unit health data, environmental data, and wind speed data.

[0063] Specifically, S301 includes: S3011, based on the real-time acquired operating data of each unit in the target wind farm, uses a time synchronization method to perform unified time axis alignment, and obtains the time-synchronized operating data of each unit; S3012, based on the time-synchronized operating data of each unit, uses threshold judgment and statistical discrimination methods to filter and obtain the operating data of each unit after removing outliers; S3013, based on the operating data of each unit after removing outliers, uses interpolation or neighboring values ​​to fill in the missing data for the corresponding time period, to obtain the operating data of each unit after filling in the missing values, and outputs it as the preprocessed operating data of the corresponding unit.

[0064] First, the real-time wind farm operation data is synchronized to unify the timelines of actual power, predicted power, operating conditions, turbine health, environment, and wind speed, eliminating time deviations from multiple data sources. Then, threshold judgment and statistical discrimination methods are used to identify and remove outliers to avoid interference introduced by sensor failures or communication errors. Next, missing time periods are filled in using interpolation or neighboring values ​​to ensure the continuity and integrity of the data sequence. This provides reliable basic data for subsequent scene vector construction, hyperparameter adaptive mapping, prediction error calculation, and comprehensive evaluation weight generation, significantly improving the stability and accuracy of the entire power prediction and evaluation method in complex marine environments.

[0065] S302, based on the preprocessed operating data of each turbine, uses a scenario-adaptive parameter generation method to obtain the set of evaluation hyperparameters for each turbine in the target wind farm. For details, please refer to [link to relevant documentation]. Figure 2 S202 of the illustrated embodiment will not be described again here.

[0066] S303, based on the preprocessed operating data and evaluation hyperparameter set of each unit, utilizes the state-environment perception evaluation method to obtain the prediction error, comprehensive evaluation weight, and effective evaluation interval constraints for each unit in the target wind farm. For details, please refer to [link to relevant documentation]. Figure 2 S203 of the illustrated embodiment will not be described again here.

[0067] S304, by combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, a weighted accuracy calculation method is used to obtain the power prediction accuracy evaluation result of the target wind farm. For details, please refer to [link to relevant documentation]. Figure 2 S204 of the illustrated embodiment will not be described again here.

[0068] The offshore wind farm power prediction and evaluation method provided in this embodiment firstly ensures data quality through preprocessing based on real-time acquired target wind farm operation data. Then, using a scenario-adaptive parameter generation method, a matching set of evaluation hyperparameters is dynamically output based on the preprocessed operation data. Next, by employing a state-environment awareness evaluation method, the error is bounded when power approaches zero and degrades to a standard relative error at high power. A comprehensive evaluation weight is calculated jointly using power level weight, operating condition weight, and turbine health weight, significantly reducing the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results. Simultaneously, an effective interval constraint is introduced to eliminate invalid prediction periods under extreme sea conditions. Finally, the weighted calculation of the power prediction accuracy evaluation result ensures that the final result truly reflects the prediction performance of the wind turbine under complex operating conditions.

[0069] This embodiment provides a method for predicting and evaluating the power output of offshore wind farms, which can be used in the aforementioned wind farm terminal equipment. Figure 4This is a flowchart of an offshore wind farm power prediction and evaluation method according to an embodiment of the present invention, such as... Figure 4 As shown, the process includes the following steps: S401, based on real-time acquired operating data of each turbine in the target wind farm, uses preprocessing methods to obtain preprocessed operating data for multiple turbines. The operating data includes actual power, predicted power, operating conditions, turbine health data, environmental data, and wind speed data. For details, please refer to [link to relevant documentation]. Figure 2 S201 of the illustrated embodiment will not be described again here.

[0070] S402, based on the preprocessed operating data of each unit, uses the scenario adaptive parameter generation method to obtain the set of evaluation hyperparameters for each unit in the target wind farm.

[0071] Specifically, S402 above includes: S4021, based on the preprocessed unit operation data of each unit, the operation scenario vector of the corresponding unit is obtained by using the scenario vector construction method; S4022, based on the operating scenario vector of each unit, uses the hyperparameter adaptive mapping method to obtain the set of evaluation hyperparameters that match the corresponding unit.

[0072] For example, the operating scenario vector for each unit constructed in S4021 above includes:

[0073] in, This represents the operating scenario vector of any unit in the target wind farm; This represents the average wind speed within the time window corresponding to time t for the corresponding unit. This indicates the degree of wind speed fluctuation for the corresponding unit; Indicates the service life of the corresponding generator unit; This indicates the unit availability rate of the corresponding generator set; This indicates the power rationing or non-full-capacity status of the corresponding generating unit.

[0074] Specifically, S4022 includes: Vector the operating scenarios of each unit Input parameter mapping function Output the set of adaptive parameters required for the model or evaluation. The adaptive parameter set includes prediction model hyperparameters, loss function weights, or evaluation weight factors; while the parameter mapping function... Depending on the operation of the station and the unit, it can be implemented using rule-based, fuzzy inference, or piecewise function-based methods to meet the needs of different engineering applications.

[0075] Based on preprocessed data on actual power, wind speed, operating conditions, turbine health, and environment, an operating scenario vector is first generated using a scenario vector construction method to achieve a unified characterization of the real-time operating status of the wind farm. Then, based on this operating scenario vector, a hyperparameter adaptive mapping method is used to dynamically output a set of evaluation hyperparameters that match the current operating conditions. This facilitates the automatic adjustment of evaluation parameters with wind speed fluctuations, turbine aging, and power curtailment conditions, ensuring that evaluation indicators are adaptable to different wind conditions, different life stages, and different operating modes, thereby improving the robustness of the evaluation method in complex marine environments.

[0076] S403, based on the preprocessed operating data and evaluation hyperparameter set of each unit, utilizes the state-environment perception evaluation method to obtain the prediction error, comprehensive evaluation weight, and effective evaluation interval constraints for each unit in the target wind farm. For details, please refer to [link to relevant documentation]. Figure 2 S203 of the illustrated embodiment will not be described again here.

[0077] S404, by combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, and using a weighted accuracy calculation method, the power prediction accuracy evaluation result of the target wind farm is obtained. For details, please refer to [link to relevant documentation]. Figure 2 S204 of the illustrated embodiment will not be described again here.

[0078] The offshore wind farm power prediction and evaluation method provided in this embodiment firstly ensures data quality through preprocessing based on real-time acquired target wind farm operation data. Then, using a scenario-adaptive parameter generation method, a matching set of evaluation hyperparameters is dynamically output based on the preprocessed operation data. Next, by employing a state-environment awareness evaluation method, the error is bounded when power approaches zero and degrades to a standard relative error at high power. A comprehensive evaluation weight is calculated jointly using power level weight, operating condition weight, and turbine health weight, significantly reducing the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results. Simultaneously, an effective interval constraint is introduced to eliminate invalid prediction periods under extreme sea conditions. Finally, the weighted calculation of the power prediction accuracy evaluation result ensures that the final result truly reflects the prediction performance of the wind turbine under complex operating conditions.

[0079] This embodiment provides a method for predicting and evaluating the power output of offshore wind farms, which can be used in the aforementioned wind farm terminal equipment. Figure 5 This is a flowchart of an offshore wind farm power prediction and evaluation method according to an embodiment of the present invention, such as... Figure 5 As shown, the process includes the following steps: S501, based on real-time acquired operating data of each turbine in the target wind farm, uses preprocessing methods to obtain preprocessed operating data for multiple turbines. The operating data includes actual power, predicted power, operating conditions, turbine health data, environmental data, and wind speed data. For details, please refer to [link to relevant documentation]. Figure 2 S201 of the illustrated embodiment will not be described again here.

[0080] S502, based on the preprocessed operating data of each turbine, uses a scenario-adaptive parameter generation method to obtain the set of evaluation hyperparameters for each turbine in the target wind farm. For details, please refer to [link to relevant documentation]. Figure 2 S202 of the illustrated embodiment will not be described again here.

[0081] S503, based on the preprocessed operating data and evaluation hyperparameter set of each unit, uses the state-environment perception evaluation method to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm.

[0082] Specifically, the aforementioned S503 includes: S5031, based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, uses the wind condition sensing power error evaluation method to obtain the prediction error of the corresponding unit at multiple times.

[0083] Specifically, S5031 includes: b1, based on the evaluation hyperparameter set and wind speed data of each unit, the dynamic generalization of the corresponding unit at multiple moments is calculated using the wind speed dynamic adjustment evaluation function; b2, based on the absolute error between the actual power and the predicted power of each unit at each moment, combined with the actual power and dynamic normalization term of the corresponding unit at the corresponding moment, and using a predefined error evaluation method, the prediction error of the corresponding unit at multiple moments is obtained.

[0084] For example, the prediction error at each time step satisfies the following:

[0085]

[0086]

[0087] in, This represents the prediction error at time t; This represents a term constructed based on the wind speed dynamic adjustment evaluation function; This represents the predicted power at time t; This represents the actual power at time t; This represents the rated power of the wind turbine generator, used as a reference scaling factor. This represents the baseline normalization coefficient, with a value range of [0.01, 0.05], determined based on the evaluation hyperparameter set; This represents the wind condition sensitive modulation coefficient, which is used to characterize the influence of wind speed on the normalization term. When the value is 1, it corresponds to most offshore wind farms. When the value is 2, it corresponds to offshore wind farms with a high proportion of low wind speeds and violent wind speed fluctuations. This represents the wind speed modulation function, based on real-time wind speed. Dynamic adjustment; This indicates the cut-in wind speed of the wind turbine, used to indicate that wind turbines below this wind speed do not generate electricity; This represents the critical value of the high wind speed power limiting zone corresponding to the rated power, used to characterize the wind turbine unit starting to operate with limited power when the wind speed is higher than this value.

[0088] By using the evaluation hyperparameter set and wind speed data, the dynamic normalization term is calculated at each moment using the wind speed dynamic adjustment evaluation function. Then, based on the absolute error between the actual power and the predicted power, the prediction error is calculated by combining the actual power and the dynamic normalization term. When the power approaches zero, the normalization term is used to ensure that the error is bounded. When the power is high, the error is smoothly degraded into a standard relative error to maintain physical comparability. Finally, the dynamic normalization term is adaptively adjusted with wind speed changes, giving greater tolerance in the low wind speed and high randomness range, and reasonably reducing it in the high wind speed and power-limited range, thereby ensuring that the final prediction error truly reflects the power prediction performance of the wind turbine under complex wind conditions.

[0089] S5032, based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, uses a multi-dimensional state-aware weighting method to obtain the comprehensive evaluation weight matching the prediction error of the corresponding unit with each time moment.

[0090] Specifically, S5032 includes: c1, based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the power sensing weight of each unit at the corresponding time is obtained by using the power sensing weight calculation method; c2, based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the operating condition perception weight of each unit at the corresponding time is obtained by using the operating condition perception weight calculation method. c3, based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the health perception weight of each unit at the corresponding time is obtained by using the health perception weight calculation method. c4 multiplies the power perception weight, operating condition perception weight, and health perception weight corresponding to each unit at each time to obtain the comprehensive evaluation weight that matches the prediction error of the corresponding unit at the corresponding time.

[0091] For example, the comprehensive evaluation weights matched with the prediction error at each time step satisfy the following:

[0092]

[0093]

[0094]

[0095] in, This represents the overall evaluation weight at time t; The weighting function based on power level indicates the contribution of prediction error to the evaluation result, and is positively correlated with its contribution to the total power generation at that moment. This represents the actual power at time t; This indicates the rated power of the wind turbine generator set; The power sensitivity index is represented, with a value range of [0.5, 2], and is determined based on the set of evaluation hyperparameters. This represents the weighting function for sensing operating conditions, based on the operating status of the wind farm. Specifically, when the wind farm is operating in normal power generation mode, the value is 1; when the wind farm is operating in a power curtailment state, the value is [value missing]. The corresponding value range is [0.1, 0.4]. When the wind farm's operating status is grid-connected or start-up / shutdown, the value is [value]. The corresponding value range is [0, 0.2]; Represents the weighting function for the perception of unit health status; This represents the unit health index, with a value range of (0, 1]. The closer the unit health index is to 1, the healthier the corresponding unit is. The health modulation index ranges from [0.5, 2]. A larger value indicates that the health status of the equipment has a stronger attenuation effect on the weight, thereby reducing the deviation of the evaluation results caused by the power curve shift due to equipment aging.

[0096] By calculating the power perception weight for each moment based on the actual power and evaluation hyperparameters, the power perception weight is obtained using a power perception weight calculation method, which makes the error contribution at high power moments greater and reflects the energy value. Then, based on the operating conditions and hyperparameters, corresponding weights are assigned to states such as power curtailment and grid-connected start-up and shutdown to reduce the interference of unsteady-state conditions. Next, based on the unit health data and hyperparameters, the power curve deviation caused by unit aging is corrected through health perception weights. Finally, by multiplying the three to obtain a comprehensive evaluation weight that matches the prediction error, the joint perception of power level, operating conditions and equipment health is achieved, ensuring that the final power prediction accuracy evaluation result is continuous and stable across the entire power range.

[0097] S5033, based on the preprocessed environmental data of each unit, uses an extreme sea state identification method to obtain the effective evaluation interval constraints for the corresponding unit.

[0098] For example, S5033 above includes:

[0099] in, The effective interval evaluation constraint value at time t is represented. It takes the value of 1 when the external environment is normal sea state and takes the value of 0 when the external environment is extreme sea state. Extreme sea state includes wind speed greater than 25m / s, wave height greater than 5m, or current speed greater than 1.5m / s.

[0100] First, based on preprocessed wind speed, actual power, and predicted power, combined with scenario-adaptive hyperparameters, a wind-sensing power error evaluation method is used to calculate the prediction error at each moment, ensuring that low-power errors are bounded and high-power errors degenerate into standard relative errors, avoiding interference from extreme values. Second, based on actual power, operating conditions, and unit health data, a multi-dimensional state-sensing weighting method is adopted, fusing power level weights, operating condition weights, and health weights to obtain a comprehensive evaluation weight matching each moment, effectively reducing the error contribution from power curtailment, low power, and unsteady-state conditions. Simultaneously, based on environmental data, an effective evaluation interval constraint is generated through an extreme sea state identification method, eliminating invalid prediction periods under extreme sea states such as typhoons and swells. This ensures the continuous and stable evaluation result of the final weighted power prediction accuracy, improving the accuracy of the final power prediction evaluation.

[0101] S504, by combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, uses a weighted accuracy calculation method to obtain the power prediction accuracy evaluation result of the target wind farm. For details, please refer to [link to relevant documentation]. Figure 2 S204 of the illustrated embodiment will not be described again here.

[0102] The offshore wind farm power prediction and evaluation method provided in this embodiment firstly ensures data quality through preprocessing based on real-time acquired target wind farm operation data. Then, using a scenario-adaptive parameter generation method, a matching set of evaluation hyperparameters is dynamically output based on the preprocessed operation data. Next, by employing a state-environment awareness evaluation method, the error is bounded when power approaches zero and degrades to a standard relative error at high power. A comprehensive evaluation weight is calculated jointly using power level weight, operating condition weight, and turbine health weight, significantly reducing the interference of power curtailment, low power, and unsteady-state conditions on the evaluation results. Simultaneously, an effective interval constraint is introduced to eliminate invalid prediction periods under extreme sea conditions. Finally, the weighted calculation of the power prediction accuracy evaluation result ensures that the final result truly reflects the prediction performance of the wind turbine under complex operating conditions.

[0103] This embodiment also provides an offshore wind farm power prediction and evaluation device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0104] This embodiment provides a device for predicting and evaluating the power output of offshore wind farms, such as... Figure 6 As shown, it includes: The data processing module 610 is used to obtain preprocessed operating data of multiple units based on the real-time acquired operating data of each unit in the target wind farm using preprocessing methods; the operating data includes actual power, predicted power, operating conditions, unit health data, environmental data and wind speed data; The parameter construction module 620 is used to obtain the set of evaluation hyperparameters for each unit in the target wind farm by using the scenario adaptive parameter generation method based on the preprocessed operating data of each unit. The parameter evaluation module 630 is used to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm based on the preprocessed operating data of each unit and the corresponding evaluation hyperparameter set, using the state environment perception evaluation method. The result generation module 640 is used to integrate the prediction error of each unit, the comprehensive evaluation weight and the effective evaluation interval constraint, and use the weighted accuracy calculation method to obtain the power prediction accuracy evaluation result of the target wind farm.

[0105] In some alternative implementations, the data processing module 610 includes: The first processing unit is used to perform unified time axis alignment based on the real-time acquired operating data of each unit in the target wind farm using a time synchronization method to obtain the time-synchronized operating data of each unit. The second processing unit is used to filter the time-synchronized operating data of each unit using threshold judgment and statistical discrimination methods to obtain the operating data of each unit after removing outliers. The third processing unit is used to fill in the missing data for the time period corresponding to the missing data by interpolation or neighboring values ​​based on the operating data of each unit after removing outliers, so as to obtain the operating data of each unit after filling in the missing values, and output it as the preprocessed operating data of the corresponding unit.

[0106] In some alternative implementations, the parameter construction module 620 includes: The scenario construction unit is used to obtain the corresponding unit's operating scenario vector based on the preprocessed unit operation data of each unit using the scenario vector construction method; The parameter construction unit is used to obtain the set of evaluation hyperparameters matching the corresponding unit based on the operating scenario vector of each unit and using the hyperparameter adaptive mapping method.

[0107] In some alternative implementations, the parameter evaluation module 630 includes: The first evaluation unit is used to obtain the prediction error of the corresponding unit at multiple times based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, and using the wind condition sensing power error evaluation method. The second evaluation unit is used to obtain the comprehensive evaluation weight of the corresponding unit and the prediction error at each time point based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, and using the multi-dimensional state-aware weighting method. The third evaluation unit is used to obtain the effective evaluation interval constraints for each unit based on the preprocessed environmental data of each unit and by using extreme sea state identification methods.

[0108] In some alternative implementations, the first evaluation unit is specifically used for: Based on the evaluation hyperparameter set and wind speed data of each unit, the dynamic generalization of the corresponding unit at multiple moments is obtained by using the wind speed dynamic adjustment evaluation function. Based on the absolute error between the actual power and the predicted power of each unit at each moment, and combining the actual power and dynamic normalization term of the corresponding unit at the corresponding moment, a predefined error evaluation method is used to obtain the prediction error of the corresponding unit at multiple moments.

[0109] In some alternative implementations, the second evaluation unit is specifically used for: Based on the actual power of each unit at each time and the evaluation hyperparameter set of the corresponding unit at the corresponding time, the power sensing weight of each unit at the corresponding time is obtained by using the power sensing weight calculation method. Based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the operating condition perception weight of each unit at the corresponding time is obtained by using the operating condition perception weight calculation method. Based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the health perception weight of each unit at the corresponding time is obtained by using the health perception weight calculation method. The power perception weight, operating condition perception weight, and health perception weight of each unit at each time are multiplied to obtain the comprehensive evaluation weight that matches the prediction error of the corresponding unit at the corresponding time.

[0110] In some alternative implementations, the result generation module 640 is specifically used for: Based on the effective evaluation interval constraint for each unit, the prediction error and comprehensive evaluation weight of the corresponding unit at each time point are screened to obtain the prediction error and comprehensive evaluation weight of multiple effective time points. Based on the prediction error and comprehensive evaluation weight of each unit at each effective moment, the weighted error of the corresponding unit at a single moment is obtained by using the weighted multiplication statistical method. Based on the single-time weighted error of multiple effective times for each unit, a comprehensive method is used to sum them to obtain the total weighted error of each unit. The total weight of each unit is obtained by summing the comprehensive evaluation weights for the entire effective time interval. Based on the total weighted error and total weight of each unit, the power prediction accuracy of each unit in the target wind farm is obtained using the accuracy evaluation method, and the power prediction evaluation result of the target wind farm is obtained.

[0111] The offshore wind farm power prediction and evaluation device provided in this embodiment of the invention can execute the offshore wind farm power prediction and evaluation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0112] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0113] The following is a detailed reference. Figure 7 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 701, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 702 or a program loaded from memory 708 into random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the electronic device. The processor 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0114] Typically, the following devices can be connected to I / O interface 705: input devices 706 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 707 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 708 including, for example, magnetic tapes, hard disks, etc.; and communication devices 709. Communication device 709 allows electronic devices to exchange data via wireless or wired communication with other devices. Although Figure 7 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0115] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 709, or installed from a memory 708, or installed from a ROM 702. When the computer program is executed by the processor 701, it performs the functions defined in the offshore wind farm power prediction and evaluation method of the embodiments of the present invention.

[0116] Figure 7 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0117] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the offshore wind farm power prediction and evaluation method shown in the above embodiments is implemented.

[0118] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0119] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for predicting and evaluating the power output of offshore wind farms, characterized in that, The method includes: Based on the real-time acquired operating data of each unit in the target wind farm, a preprocessing method is used to obtain preprocessed operating data of multiple units; the operating data includes actual power, predicted power, operating conditions, unit health data, environmental data and wind speed data; Based on the preprocessed operating data of each unit, the evaluation hyperparameter set of each unit in the target wind farm is obtained by using the scenario adaptive parameter generation method. Based on the preprocessed operating data of each unit and the corresponding set of evaluation hyperparameters, the state environment perception evaluation method is used to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm. By combining the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, the power prediction accuracy evaluation result of the target wind farm is obtained using the weighted accuracy calculation method.

2. The method according to claim 1, characterized in that, Based on the preprocessed operating data of each unit and the corresponding set of evaluation hyperparameters, the state-environment perception evaluation method is used to obtain the prediction error, comprehensive evaluation weight, and effective evaluation interval constraints for each unit in the target wind farm, including: Based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, the prediction error of the corresponding unit at multiple times is obtained by using the wind condition sensing power error evaluation method. Based on the preprocessed wind speed data, actual power and predicted power of each unit, combined with the evaluation hyperparameter set of the corresponding unit, the comprehensive evaluation weight of the corresponding unit and the prediction error at each time moment is obtained by using the multi-dimensional state-aware weighting method. Based on the preprocessed environmental data of each unit, the effective evaluation interval constraints of the corresponding unit are obtained by using the extreme sea state identification method.

3. The method according to claim 2, characterized in that, Based on the preprocessed wind speed data, actual power, and predicted power of each generator unit, and combined with the evaluation hyperparameter set of the corresponding generator unit, the wind condition-aware power error evaluation method is used to obtain the prediction error of the corresponding generator unit at multiple times, including: Based on the evaluation hyperparameter set and wind speed data of each unit, the dynamic generalization of the corresponding unit at multiple moments is obtained by using the wind speed dynamic adjustment evaluation function. Based on the absolute error between the actual power and the predicted power of each unit at each moment, and combining the actual power and dynamic normalization term of the corresponding unit at the corresponding moment, a predefined error evaluation method is used to obtain the prediction error of the corresponding unit at multiple moments.

4. The method according to claim 2, characterized in that, Based on the preprocessed wind speed data, actual power, and predicted power of each generator unit, and combined with the evaluation hyperparameter set of the corresponding generator unit, a multi-dimensional state-aware weighting method is used to obtain the comprehensive evaluation weight matching the prediction error of the corresponding generator unit at each time moment, including: Based on the actual power of each unit at each time and the evaluation hyperparameter set of the corresponding unit at the corresponding time, the power sensing weight of each unit at the corresponding time is obtained by using the power sensing weight calculation method. Based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the operating condition perception weight of each unit at the corresponding time is obtained by using the operating condition perception weight calculation method. Based on the actual power of each unit at each time and the set of evaluation hyperparameters of the corresponding unit at the corresponding time, the health perception weight of each unit at the corresponding time is obtained by using the health perception weight calculation method. The power perception weight, operating condition perception weight, and health perception weight of each unit at each time are multiplied to obtain the comprehensive evaluation weight that matches the prediction error of the corresponding unit at the corresponding time.

5. The method according to claim 1, characterized in that, The prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint are combined, and a weighted accuracy calculation method is used to obtain the power prediction accuracy evaluation result of the target wind farm, including: Based on the effective evaluation interval constraint for each unit, the prediction error and comprehensive evaluation weight of the corresponding unit at each time point are screened to obtain the prediction error and comprehensive evaluation weight of multiple effective time points. Based on the prediction error and comprehensive evaluation weight of each unit at each effective moment, the weighted error of the corresponding unit at a single moment is obtained by using the weighted multiplication statistical method. Based on the single-time weighted error of multiple effective times for each unit, a comprehensive method is used to sum them to obtain the total weighted error of each unit. The total weight of each unit is obtained by summing the comprehensive evaluation weights for the entire effective time interval. Based on the total weighted error and total weight of each unit, the power prediction accuracy of each unit in the target wind farm is obtained using the accuracy evaluation method, and the power prediction evaluation result of the target wind farm is obtained.

6. The method according to claim 1, characterized in that, Based on the preprocessed operating data of each unit, the scenario adaptive parameter generation method is used to obtain the evaluation hyperparameter set for each unit in the target wind farm, including: Based on the preprocessed unit operation data of each unit, the operation scenario vector of the corresponding unit is obtained by using the scenario vector construction method. Based on the operating scenario vector of each unit, an evaluation hyperparameter set matching the corresponding unit is obtained using the hyperparameter adaptive mapping method.

7. The method according to any one of claims 1 to 6, characterized in that, The preprocessing method is used to obtain preprocessed operating data for multiple units based on the real-time acquired operating data of each unit in the target wind farm, including: Based on the real-time acquired operating data of each unit in the target wind farm, a time synchronization method is used to perform unified time axis alignment to obtain the time-synchronized operating data of each unit. Based on the time-synchronized operating data of each unit, threshold judgment and statistical discrimination methods are used to filter the data to obtain the operating data of each unit after removing outliers. Based on the operational data of each unit after removing outliers, interpolation or neighboring values ​​are used to fill in the missing data for the corresponding time period to obtain the operational data of each unit after missing values ​​are filled in, and the output is the preprocessed operational data of the corresponding unit.

8. A power prediction and evaluation device for offshore wind farms, characterized in that, The device includes: The data processing module is used to obtain preprocessed operating data of multiple units based on the real-time acquired operating data of each unit in the target wind farm using preprocessing methods; the operating data includes actual power, predicted power, operating conditions, unit health data, environmental data and wind speed data; The parameter construction module is used to obtain the set of evaluation hyperparameters for each unit in the target wind farm by using the scenario adaptive parameter generation method based on the preprocessed operating data of each unit. The parameter evaluation module is used to obtain the prediction error, comprehensive evaluation weight and effective evaluation interval constraint of each unit in the target wind farm based on the preprocessed operating data of each unit and the corresponding set of evaluation hyperparameters, using the state environment perception evaluation method. The results generation module is used to integrate the prediction error of each unit, the comprehensive evaluation weight, and the effective evaluation interval constraint, and use the weighted accuracy calculation method to obtain the power prediction accuracy evaluation result of the target wind farm.

9. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the offshore wind farm power prediction and evaluation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the offshore wind farm power prediction and evaluation method according to any one of claims 1 to 7.

Images