A dynamic optimization method and system for wind turbine operation and maintenance windows integrating multi-scale quantity and price forecasting

Abstract

This invention relates to the field of wind power generation technology, specifically to a dynamic optimization method and system for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasting. The method includes: collecting and preprocessing multi-source data; constructing a multi-scale quantity and price forecasting model; training the forecasting sub-model to obtain the theoretical power, contracted power output, and spot electricity price for a future period; obtaining the opportunity cost of downtime; constructing a dynamic optimization model with the objective of minimizing downtime opportunity cost and constrained by safe operating conditions; solving the dynamic optimization model to output the optimal operation and maintenance window; and executing the above steps on a rolling basis at set intervals. This invention integrates safety, energy efficiency, and economy, and introduces multi-scale quantity and price forecasting as a pre-driving factor for decision-making. It proactively predicts wind speed, power generation, and electricity price for a future period, thereby constructing a dynamic optimization model with the objective function of minimizing downtime loss per unit time. This model is used to find the optimal downtime within the boundaries of safety constraints, thus guiding operation and maintenance activities.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, specifically to a dynamic optimization method and system for wind turbine operation and maintenance windows that integrates multi-scale quantity and price prediction. Background Technology

[0002] To ensure the safe and stable operation of wind turbines, wind farm operation and maintenance mainly follow two modes: one is periodic maintenance based on time cycles (such as semi-annual or annual inspections), and the other is temporary troubleshooting for sudden faults. When deciding when to shut down the turbines for maintenance, existing technical solutions do indeed refer to meteorological conditions, such as simply determining whether the current wind speed is below the safe operating threshold (such as 10 m / s) and whether there is severe weather such as rain, snow, or lightning, to ensure the safety of personnel and equipment during climbing operations and turbine operation.

[0003] The closest existing technology to this invention is a wind turbine operation and maintenance window decision system based on meteorological threshold judgment and a scheduling method based on fixed maintenance plans. Its typical technical path is as follows: By accessing real-time meteorological data (including wind speed, wind direction, temperature, humidity and rainfall, etc.) from the local wind measurement tower or meteorological server of the wind farm, the system makes a judgment based on the static safety operation threshold combination preset by the operation and maintenance personnel (such as 10-minute average wind speed < 8m / s, no rainfall, temperature > -10℃, etc.). When all real-time meteorological conditions meet the threshold requirements, the system automatically marks it as an operable window and notifies the operation and maintenance team. The team then starts the planned shutdown maintenance work or the shutdown maintenance work of the faulty unit according to this window and its own personnel arrangements.

[0004] While the aforementioned technical solutions play a basic role in ensuring operational safety, their decision-making logic is overly simplistic and static, lacking consideration for economic benefits. This results in poor economic efficiency for shutdown timing, focusing only on physical constraints such as wind speed and weather while ignoring the real-time power generation capacity (power generation potential) of the wind turbine itself and the rapidly changing electricity market price signals. Under this solution, the turbine might be ignored during favorable periods of "light wind and low electricity prices," but shut down during periods of medium wind and high electricity prices, leading to significant losses in the product of power generation and electricity price (i.e., potential revenue). Furthermore, these wind farms typically sign medium- to long-term contracts and participate in spot market transactions. Existing technology cannot handle the interplay between the pressure of allocating electricity under medium- to long-term contracts and the high electricity price window in the spot market. For example, if a day is characterized by strong winds and high electricity prices, but a small wind window that meets safety conditions appears in between, shutting down during this window could directly result in the inability to complete the medium- to long-term daily electricity allocation, forcing the purchase of electricity from the high-priced spot market to fulfill the contract, causing secondary losses. Existing technologies lack the ability to integrate power forecasting information across multiple time scales (medium- to long-term / day-ahead / real-time) for comprehensive decision-making.

[0005] In summary, the use of static meteorological thresholds for triggering in existing technologies leads to technical defects such as the inability to comprehensively consider power generation capacity and price factors, and excessively high opportunity costs for operation and maintenance, thus failing to maximize the operational benefits of wind farms. Summary of the Invention

[0006] To address the problems in existing technologies, this invention provides a dynamic optimization method and system for wind turbine operation and maintenance windows that integrates multi-scale quantity and price prediction. The goal is to maximize the operational benefits of wind farms throughout their entire lifespan or in the future, and to dynamically determine the optimal timing for shutdown and maintenance.

[0007] This invention is achieved through the following technical solution: A dynamic optimization method for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasts includes: Collect and preprocess multi-source data; multi-source data includes meteorological data, unit status data, and market electricity price data; Construct multi-scale quantity and price forecasting models, including ultra-short-term / short-term power forecasting models, medium-term contract decomposition models, and spot electricity price forecasting models; The preprocessed data is used to train each prediction sub-model to obtain the theoretical power, contracted power output, and spot electricity price for a future time period. The opportunity cost of outages is determined based on the theoretical power output, contracted power output, and spot electricity price for a future period. Construct a dynamic optimization model with the objective of minimizing downtime opportunity cost and with safe operating conditions as constraints; Solve the dynamic optimization model and output the optimal operation and maintenance window; Set an interval to continuously execute the above steps.

[0008] Preferably, the meteorological data includes historical and forecast data on wind speed, wind direction, temperature, humidity, and air pressure; the unit status data includes the real-time operating status and power generation of the unit; and the market electricity price data includes the monthly / daily breakdown of the electricity volume under the medium- and long-term contracts signed by the power station, as well as the day-ahead / real-time spot market electricity price forecast data for the future period.

[0009] Preferably, the preprocessing process for multi-source data is as follows: the collected data is cleaned, normalized, and time-aligned to form a unified time series dataset.

[0010] Preferably, the ultra-short-term / short-term power prediction model is established based on NWP wind speed prediction and historical power curves, and uses a deep learning network to predict the theoretical power generation of wind turbines for a certain period of time within the next T hours. P _ forecast ( t ); The medium-term contract decomposition model dynamically decomposes the medium- and long-term contracted electricity volume of power plants into each forecast time point based on historical power generation patterns, holidays, and maintenance plans. t To obtain the contracted electricity volume that must be undertaken at each point in time. E _ contract ( t ) and the corresponding contract electricity price Price _ contract ; The spot electricity price forecasting model is built upon historical electricity price data, load forecasting, renewable energy output forecasting, and grid congestion information. It employs a hybrid model that includes time-series models and regression analysis of meteorological factors to predict the spot market clearing price for the next T hours. Price _ spot ( t ).

[0011] Preferably, the optimization variables in the dynamic optimization model are the decision variables. , =1 indicates that at time... Start maintenance shutdown. =0 indicates normal operation; The constraints of the dynamic optimization model are: Safe operating constraints: W _ safe ( ) = 1, if and only if the wind speed is within the predicted future time period. WS _ forecast ( )< WS _ threshold _ safe Furthermore, temperature, rainfall, and visibility are all within safe ranges; otherwise... W _ safe ( ) = 0; The maximum number of units that can be operated and maintained simultaneously within the same time period is: N _ max ; For each shutdown task, there is one and only one start time. =1.

[0012] Preferably, the objective function of the dynamic optimization model is:

[0013]

[0014]

[0015] In the formula: If at any moment The total power generation loss cost caused by a single shutdown maintenance event; For time resolution; To comprehensively consider electricity price weights and dynamically reflect default risk and market gains: when the actual power generation is less than the contracted amount, the difference is implicitly included in the loss based on the cost of purchasing electricity at the higher spot price; when the power generation exceeds the contracted amount, the revenue is settled based on the spot electricity price.

[0016] Preferably, a dynamic programming or mixed-integer linear programming solver is used, provided that the above constraints are met and W _ safe ( All time points where ) = 1 Above, solve for Total _ Loss Minimum optimal solution ; Output the optimal solution The optimal operation and maintenance window is recommended, and a second-best window and a third-best window are provided as alternatives.

[0017] A system for implementing the aforementioned dynamic optimization method for wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting includes: The data acquisition and processing module is used to acquire and preprocess multi-source data. The prediction sub-model building module is used to build a multi-scale quantity and price prediction model, and to train each prediction sub-model using preprocessed data to obtain the theoretical power, contract decomposed electricity and spot electricity price for a certain period in the future. The downtime opportunity cost acquisition module is used to obtain the downtime opportunity cost based on the theoretical power, contracted power breakdown, and spot electricity price for a certain future period. The dynamic optimization model building module is used to build a dynamic optimization model with the goal of minimizing downtime opportunity cost and with safe operating conditions as constraints. The solver module is used to solve the dynamic optimization model and output the optimal operation and maintenance window.

[0018] An electronic device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method.

[0019] A storage medium having a computer program stored thereon, the computer program being executed by a processor to perform the steps of the method.

[0020] Compared with the prior art, the present invention has the following beneficial effects: This invention presents a dynamic optimization method for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasting. It combines safety, energy efficiency, and economy, and introduces multi-scale quantity and price forecasting as a pre-driving factor for decision-making. This proactively predicts wind speed, power generation, and electricity price over a future period, thereby constructing a dynamic optimization model with the objective function of minimizing downtime losses per unit time. This model is used to find the optimal downtime within safety constraints, thus guiding operation and maintenance activities. Specifically, this invention integrates ultra-short-term power forecasting, medium- and long-term contract decomposition, and spot electricity price forecasting as inputs for operation and maintenance window optimization. Then, it constructs an optimization function with the objective of minimizing "downtime opportunity cost," and designs a comprehensive electricity price weighting coefficient λ(t) coupled with contracted electricity volume and spot electricity price to accurately quantify the comprehensive economic losses caused by downtime. Furthermore, it relies on a dynamic rolling optimization mechanism to periodically re-optimize and dynamically adjust the operation and maintenance plan using the latest data, forming a closed-loop feedback control. Attached Figure Description

[0021] Figure 1 This is an overall flowchart of a dynamic optimization method for wind turbine operation and maintenance windows according to the present invention; Figure 2 This is a structural block diagram of a system for implementing a dynamic optimization method for the operation and maintenance window of a wind turbine generator according to the present invention. Detailed Implementation

[0022] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification.

[0023] Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to fully and completely disclose the invention and to fully convey its scope to those skilled in the art. The terminology used in the exemplary embodiments illustrated in the drawings is not intended to limit the invention. In the drawings, the same units / elements are referred to by the same reference numerals.

[0024] Unless otherwise stated, the terms used herein (including technical terms) have their common meaning as understood by one of ordinary skill in the art. Furthermore, it is understood that terms defined in commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and not to be interpreted as having an idealized or overly formal meaning.

[0025] Multi-scale quantity and price forecasting: A comprehensive forecasting method that integrates power generation at multiple time scales, electricity volume allocated from medium- and long-term contracts, and spot market electricity price forecasts.

[0026] Operation and maintenance window: refers to the time period during which wind turbine units can be safely shut down for maintenance. In this invention, this window not only meets meteorological safety conditions, but is also dynamically determined as the economically optimal start time and duration through an optimization model.

[0027] Opportunity cost of downtime: refers to the potential revenue loss caused by the abandonment of power generation due to wind turbine downtime for maintenance. In this invention, this cost is calculated jointly from the theoretical power generation during the downtime period, the contracted electricity price, and the spot electricity price.

[0028] Medium- and long-term contract breakdown: The process of dynamically allocating the monthly / weekly medium- and long-term contract electricity volume signed by wind farms to each smaller time unit (such as hour) based on historical power generation patterns and expected maintenance plans.

[0029] Comprehensive Electricity Price Weight: A dynamic coefficient defined in this invention, used to calculate the unit electricity price at time t due to outage losses. It integrates medium- and long-term contract electricity prices with spot electricity prices, reflecting the comprehensive economic cost of outages.

[0030] Safe operation constraints: A set of meteorological condition thresholds to ensure the personal safety of maintenance personnel and the safety of equipment.

[0031] The present invention will be further described in detail below with reference to specific embodiments. These descriptions are for explanation purposes only and are not intended to limit the scope of the invention.

[0032] This invention discloses a dynamic optimization method for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasts, referring to... Figure 1 , 2 ,include: S100 collects and preprocesses multi-source data, including meteorological data, unit status data, and market electricity price data.

[0033] The multi-source data is acquired in real time or periodically from wind farm data acquisition and monitoring control systems (SCADA), wind measurement towers, numerical weather prediction (NWP) systems, and power trading center API interfaces. Specifically: Meteorological data includes historical and forecast data on wind speed, wind direction, temperature, humidity, and air pressure; unit status data includes real-time operating status and power generation; market electricity price data includes monthly / daily breakdowns of electricity volume under medium- and long-term contracts signed by the power station, and forecasts of day-ahead / real-time spot market electricity prices for the future period.

[0034] During preprocessing, the collected data is cleaned, normalized, and time-aligned to form a unified time series dataset. In actual operation, the time resolution can be set to 15 minutes or 1 hour.

[0035] In one embodiment, during cleaning, linear interpolation, spline interpolation, or moving average based on front and back windows is used for filling; for large areas of continuous missing data, discarding or marking is used; median filtering or wavelet transform is used for denoising and smoothing; at the same time, duplicate timestamp records are removed to solve the problem of data accumulation caused by fluctuations in data acquisition frequency.

[0036] In one embodiment, to eliminate the impact of differences in feature dimensions on model convergence, multidimensional data is mapped to a unified numerical range. Then, the timestamps of all data sources are uniformly converted to UTC standard time and parsed into a Unix timestamp format accurate to milliseconds. Finally, the time-series features of the variables are concatenated along the feature axis to form a multidimensional feature matrix, and a unified time series dataset with shape [number of samples, time step, feature dimension] is generated, which can be directly input into subsequent deep learning models.

[0037] S200 constructs a multi-scale quantity and price forecasting model, including an ultra-short-term / short-term power forecasting model, a medium-term contract decomposition model, and a spot electricity price forecasting model.

[0038] Each prediction sub-model is trained using the preprocessed data to obtain the theoretical power, contracted power output, and spot electricity price for a future time period. Details are as follows: The ultra-short-term / short-term power prediction model is based on NWP wind speed prediction and historical power curves. It uses deep learning networks (such as LSTM or Transformer) to predict the theoretical power generation of wind turbines for a certain time period (e.g., 15 minutes) within the next T hours. P _ forecast ( t In one embodiment, the power prediction model uses physical methods (based on computational fluid dynamics models) or statistical methods (such as support vector regression, SVR) instead of deep learning models. For example, the deep learning power prediction method based on multi-point NWP disclosed in CN106650982B uses multi-point NWP within a specified area as input to increase the model input information. It also uses the complex function learning and layer-by-layer feature extraction capabilities of deep learning to deeply mine the useful information provided by multi-point NWP data. The power of any wind turbine, wind farm, or wind farm group within the specified area is used as the output to achieve large-scale wind power prediction.

[0039] The medium-term contract decomposition model dynamically decomposes the medium- and long-term contract electricity volumes (such as monthly and weekly contracts) of power plants into each forecast time point based on historical power generation patterns, holidays, and maintenance plans. t To obtain the contracted electricity volume that must be undertaken at each point in time. E _ contract ( t) and the corresponding contract electricity price Price _ contract .

[0040] The spot electricity price forecasting model is built upon historical electricity price data, load forecasts, renewable energy output forecasts, and grid congestion information. It employs a hybrid model that includes time-series models and regression analysis of meteorological factors to predict the spot market clearing price for the next T hours. Price _ spot ( t Such as the spot electricity price forecasting method based on multidimensional matching and model fusion of similar days disclosed in CN121921051A. In one embodiment, the electricity price forecasting adopts an autoregressive integral moving average model (ARIMA) or a simulation model based on multi-agent reinforcement learning (such as CN121546550A).

[0041] S300 calculates the opportunity cost of outages based on theoretical power, contracted power allocation, and spot electricity prices for a future period.

[0042] S400 constructs a dynamic optimization model with the objective of minimizing downtime opportunity cost and constrained by safe operating conditions, as follows: The optimization variables in a dynamic optimization model are the decision variables. , =1 indicates that at time... Start maintenance shutdown. =0 indicates normal operation; The constraints of the dynamic optimization model are: Safe operating constraints: W _ safe ( ) = 1, if and only if the wind speed is within the predicted future time period. WS _ forecast ( )< WS _ threshold _ safe Furthermore, temperature, rainfall, and visibility are all within safe ranges; otherwise... W _ safe ( ) = 0; The maximum number of units that can be operated and maintained simultaneously within the same time period is: N _ max ; For each shutdown task, there is one and only one start time. =1.

[0043] The objective function of the dynamic optimization model is:

[0044]

[0045]

[0046] In the formula: If at any moment Perform maintenance shutdown; single shutdown event (duration) D For example, the total power generation loss cost caused by 4-8 hours; For time resolution (e.g., 0.25 hours); To comprehensively consider electricity price weights and dynamically reflect default risk and market gains: when the actual power generation is less than the contracted amount, the difference is implicitly included in the loss based on the cost of purchasing electricity at the higher spot price; when the power generation exceeds the contracted amount, the revenue is settled based on the spot electricity price.

[0047] In one embodiment, safety constraints may not be limited to wind speed, but may also include more refined parameters that affect the safety of high-altitude operations, such as turbulence intensity, wind shear, and cloud base height.

[0048] In one embodiment, resource constraints may include the skill level of the maintenance team, the status of spare parts inventory, etc.

[0049] In one embodiment, the dynamic optimization model is optimized with the goal of "maximizing potential benefits", that is, calculating the difference between the expected benefits under the condition of no shutdown and the expected benefits under the condition of shutdown.

[0050] S500 solves the dynamic optimization model and outputs the optimal operation and maintenance window. Specifically, it uses a dynamic programming or mixed-integer linear programming solver, under the condition that the above constraints are met, and W _ safe ( All time points where ) = 1 Above, solve for Total _ Loss Minimum optimal solution ; Output the optimal solution As an optimal maintenance window suggestion, for example, "It is recommended that Unit #3 be shut down for maintenance starting at 02:00 tomorrow (XXXX year X month X day) within the next 48 hours, and the maintenance is expected to last for 5 hours."

[0051] In one embodiment, in addition to dynamic programming / mixed integer linear programming, heuristic algorithms such as genetic algorithms and particle swarm optimization can be used to obtain approximate optimal solutions, especially when the number of units is large, in order to improve the solution speed.

[0052] In addition, the solver can provide suboptimal windows and third windows as alternatives.

[0053] S600 involves executing the above steps on a rolling basis at set intervals or when the prediction error exceeds a threshold (wind speed, power, electricity consumption, etc.). This is because new NWP data and electricity price data are constantly updated over time. For example, executing S100 to S500 every 1 to 4 hours allows for the re-optimization and dynamic adjustment of maintenance plans that have not yet been implemented, adapting to market and weather changes. By recalculating the entire process from S100 to S500 every 1 to 4 hours, the algorithm can identify deviations between the original plan and the actual environment, dynamically fine-tuning unit start-up and shutdown strategies, maintenance windows, and output plans. Near real-time adaptive optimization effectively mitigates the risk of wind and solar curtailment caused by sudden weather changes, and maximizes asset returns during peak market electricity price periods. This ensures that maintenance decisions are always at the optimal Pareto front in complex and ever-changing market and physical environments, thereby achieving a dynamic balance between safety and economy.

[0054] This invention also discloses a system for implementing the aforementioned dynamic optimization method for wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting, comprising: The data acquisition and processing module is used to collect and preprocess multi-source data; this module connects to SCADA, NWP server, and electricity market information API.

[0055] The prediction sub-model building module is used to build a multi-scale quantity and price prediction model, and to train each prediction sub-model using preprocessed data to obtain the theoretical power, contract decomposed electricity and spot electricity price for a certain period in the future. The downtime opportunity cost acquisition module is used to obtain the downtime opportunity cost based on the theoretical power, contracted power breakdown, and spot electricity price for a certain future period. The dynamic optimization model building module, including the objective function builder and constraint manager, is used to build a dynamic optimization model with the objective of minimizing downtime opportunity cost and with safe operating conditions as constraints.

[0056] The solution module includes a solver (which runs the aforementioned dynamic programming / mixed-integer linear programming algorithm) for solving the dynamic optimization model and outputting the optimal operation and maintenance window; The human-computer interaction interface is used to display optimal operation and maintenance window suggestions, input constraint parameters (such as security thresholds and resource limits), and receive confirmation instructions from operation and maintenance personnel.

[0057] This invention presents a dynamic optimization method for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasts. It merges power generation forecasts with multi-scale electricity prices (medium- to long-term + spot) into the objective function, identifying the time when the product of "power generation loss × electricity price" is minimized for shutdown. This avoids significant potential revenue losses caused by shutdowns during periods of high winds and high electricity prices, representing a fundamental improvement over existing "safety-only" decision-making methods. Simultaneously, a medium-term contract decomposition model is introduced to balance performance obligations with market speculative gains, ensuring that the operation and maintenance plan serves the overall operational strategy of the wind farm, rather than merely the equipment itself. Furthermore, a rolling optimization mechanism is employed to continuously revise the plan using the latest forecast data, overcoming the limitations of static plans in handling weather and electricity price uncertainties, ensuring that the recommended operation and maintenance window always closely approximates the actual optimal solution.

[0058] This invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, and is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions from a computer storage medium to implement the corresponding method flow or corresponding function. This invention also discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method. The computer-readable storage medium is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both built-in storage media in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space containing the terminal's operating system. Furthermore, this storage space also contains one or more instructions suitable for loading and execution by a processor; these instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium.

[0059] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0060] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0061] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0062] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0063] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the functions specified in one or more boxes. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the technical solution of the present invention in any way. Those skilled in the art should understand that, without departing from the spirit and principles of the present invention, the technical solution can be modified and replaced in several simple ways, and these modifications and replacements are all within the scope of protection covered by the claims.

Claims

1. A dynamic optimization method for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasts, characterized in that, include: Collect and preprocess multi-source data; multi-source data includes meteorological data, unit status data, and market electricity price data; Construct multi-scale quantity and price forecasting models, including ultra-short-term / short-term power forecasting models, medium-term contract decomposition models, and spot electricity price forecasting models; The preprocessed data is used to train each prediction sub-model to obtain the theoretical power, contracted power output, and spot electricity price for a future time period. The opportunity cost of outages is determined based on the theoretical power output, contracted power output, and spot electricity price for a future period. Construct a dynamic optimization model with the objective of minimizing downtime opportunity cost and with safe operating conditions as constraints; Solve the dynamic optimization model and output the optimal operation and maintenance window; Set an interval to continuously execute the above steps.

2. The method for dynamic optimization of wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting as described in claim 1, characterized in that, Meteorological data includes historical and forecast data on wind speed, wind direction, temperature, humidity, and air pressure; unit status data includes real-time operating status and power generation; market electricity price data includes monthly / daily breakdowns of electricity volume under medium- and long-term contracts signed by the power station, and forecasts of day-ahead / real-time spot market electricity prices for the future period.

3. The method for dynamic optimization of wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting as described in claim 1, characterized in that, The preprocessing of multi-source data involves cleaning, normalizing, and aligning the collected data over time to form a unified time series dataset.

4. The method for dynamic optimization of wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting as described in claim 1, characterized in that, The ultra-short-term / short-term power prediction model is based on NWP wind speed prediction and historical power curves, and uses a deep learning network to predict the theoretical power generation of wind turbines for a certain period of time within the next T hours. P _ forecast ( t ); The medium-term contract decomposition model dynamically decomposes the medium- and long-term contracted electricity volume of power plants into each forecast time point based on historical power generation patterns, holidays, and maintenance plans. t To obtain the contracted electricity volume that must be undertaken at each point in time. E _ contract ( t ) and the corresponding contract electricity price Price _ contract ; The spot electricity price forecasting model is built upon historical electricity price data, load forecasting, renewable energy output forecasting, and grid congestion information. It employs a hybrid model that includes time-series models and regression analysis of meteorological factors to predict the spot market clearing price for the next T hours. Price _ spot ( t ).

5. The method for dynamic optimization of wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting as described in claim 4, characterized in that, The optimization variables in a dynamic optimization model are the decision variables. , =1 indicates that at time... Start maintenance shutdown. =0 indicates normal operation; The constraints of the dynamic optimization model are: Safe operating constraints: W _ safe ( ) = 1, if and only if the wind speed is within the predicted future time period. WS _ forecast ( )< WS _ threshold _ safe Furthermore, temperature, rainfall, and visibility are all within safe ranges; otherwise... W _ safe ( ) = 0; The maximum number of units that can be operated and maintained simultaneously within the same time period is: N _ max ; For each shutdown task, there is one and only one start time. =1.

6. The method for dynamic optimization of wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting as described in claim 5, characterized in that, The objective function of the dynamic optimization model is: In the formula: If at any moment The total power generation loss cost caused by a single shutdown maintenance event; For time resolution; To comprehensively consider electricity price weights and dynamically reflect default risk and market gains: when the actual power generation is less than the contracted amount, the difference is implicitly included in the loss based on the cost of purchasing electricity at the higher spot price; when the power generation exceeds the contracted amount, the revenue is settled based on the spot electricity price.

7. The method for dynamic optimization of wind turbine operation and maintenance windows based on multi-scale quantity and price forecasting as described in claim 6, characterized in that, Using dynamic programming or mixed-integer linear programming solvers, under the above constraints and W _ safe ( All time points where ) = 1 Above, solve for Total _ Loss Minimum optimal solution ; Output the optimal solution The optimal operation and maintenance window is recommended, and a second-best window and a third-best window are provided as alternatives.

8. A system for implementing the dynamic optimization method for wind turbine operation and maintenance windows that integrates multi-scale quantity and price forecasts as described in any one of claims 1 to 7, characterized in that, include: The data acquisition and processing module is used to acquire and preprocess multi-source data. The prediction sub-model building module is used to build a multi-scale quantity and price prediction model, and to train each prediction sub-model using preprocessed data to obtain the theoretical power, contract decomposed electricity and spot electricity price for a certain period in the future. The downtime opportunity cost acquisition module is used to obtain the downtime opportunity cost based on the theoretical power, contracted power breakdown, and spot electricity price for a certain future period. The dynamic optimization model building module is used to build a dynamic optimization model with the goal of minimizing downtime opportunity cost and with safe operating conditions as constraints. The solver module is used to solve the dynamic optimization model and output the optimal operation and maintenance window.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images