New energy station operation optimization system and method based on big data

Abstract

The invention relates to the technical field of new energy power generation, and discloses a new energy station operation optimization system and method based on big data, and the system comprises a data collection and integration module which collects multi-source data, including equipment operation data, meteorological data and power grid data, inside and outside a new energy station through a multi-source data collection sub-module in the data collection and integration module, collected data is subjected to primary processing through the edge calculation preprocessing sub-module, format unification, denoising processing and feature extraction are included, standardized time series data are generated, and the processed standardized time series data are transmitted to the big data storage and calculation module. Through the data anomaly detection and restoration module, abnormal values in the data collected by the new energy station can be effectively identified and restored, a regularized restoration method is combined with statistical characteristics, the integrity and accuracy of input data are ensured, and subsequent prediction and optimization deviations caused by data noise or collection errors are reduced.

Description

Technical Field

[0001] The present invention relates to the field of renewable energy power generation technology, and specifically to a new energy station operation optimization system and method based on big data. Background Art

[0002] With the growth of global energy demand and the advancement of carbon reduction targets, new energy power generation technologies represented by wind power and photovoltaics are developing rapidly. However, there are significant technical challenges in the operation of new energy stations, including power generation uncertainty, data quality issues, low resource scheduling efficiency, and insufficient grid stability, which limit the operating efficiency of new energy stations.

[0003] In the operation of new energy stations, multi-source data collection involves information such as equipment operation, meteorological conditions, and power grid status. However, due to the accuracy of the collection equipment, the stability of network transmission, and the influence of environmental noise, there are outliers and noise data in the collected data. Traditional anomaly detection methods mostly rely on simple rules or statistical analysis, which cannot efficiently identify complex abnormal patterns, and the repair strategies lack accuracy; Renewable energy generation is significantly affected by weather changes and environmental factors, and is highly random and intermittent. Traditional forecasting methods are mostly based on linear regression or time series models, which cannot capture complex nonlinear relationships. At the same time, load demand forecasting faces similar problems. The grid load is affected by multiple factors such as electricity price fluctuations and electricity consumption behavior. The prediction accuracy of traditional models is limited, resulting in an imbalance in resource allocation, which in turn reduces the power generation utilization rate and economic benefits of the site. Traditional resource scheduling methods mostly use static allocation strategies, which cannot dynamically adapt to the actual needs of power generation fluctuations and grid load changes. At the same time, in the scheduling of energy storage equipment, real-time electricity prices and equipment operating status are not fully considered, resulting in low energy storage utilization and high operating costs. In addition, the lack of intelligent scheduling algorithms reduces the overall resource optimization capabilities of the station and cannot meet the efficient operation needs of new energy stations; The volatility and intermittent characteristics of renewable energy power generation pose a challenge to the stability of the power grid. During the grid-connected operation of new energy stations, the high volatility of power output can easily cause instability in the grid frequency and voltage. Traditional grid-connected optimization methods cannot dynamically adjust the energy storage charging and discharging and grid-connected power output, and cannot effectively reduce peak loads and smooth power fluctuations, limiting the large-scale grid-connected application of new energy stations.

[0004] Therefore, those skilled in the art provide a new energy station operation optimization system and method based on big data to solve the above-mentioned problems. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a new energy station operation optimization system and method based on big data to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A new energy station operation optimization system based on big data, comprising: Data collection and integration module: The multi-source data inside and outside the new energy station are collected through its internal multi-source data collection sub-module, including equipment operation data, meteorological data and power grid data. The collected data is preliminarily processed by the edge computing pre-processing sub-module, including format unification, denoising and feature extraction, to generate standardized time series data. After processing, the standardized time series data is transmitted to the big data storage and computing module; Big Data Storage and Computing Module: Receives standardized time series data from the Data Collection and Integration Module and processes it through the submodules in the Big Data Storage and Computing Module: Data anomaly detection and repair module: receives data from the big data storage and computing module, performs anomaly detection operations on the data, identifies outliers in the data, repairs outliers based on regular processing methods, generates high-quality data streams, and the repaired data is passed to the data prediction module; Data prediction module: receives high-quality data streams from the data anomaly detection and repair module, builds a deep learning model using multimodal data, and predicts future power generation and grid load. The prediction results are passed as input to the intelligent optimization and analysis module to optimize resource allocation. Intelligent optimization and analysis module: Based on the results of the data prediction module, the resource allocation of the new energy stations is optimized. At the same time, the operating status of the stations is analyzed. The optimized allocation plan and operating status information are transmitted to the application and interaction module for further operation and monitoring by users. Application and interaction module: Receives the results from the intelligent optimization and analysis module, displays the optimization allocation and operation status information, and allows users to remotely manage and monitor through the provided operation interface. It supports the adjustment of station operation parameters and the management of resource allocation plans.

[0007] Preferably, in the data acquisition and integration module, the denoising process adopts a sliding window algorithm: , in, is the data after smoothing. For the The original data value at the moment, n is the length of the sliding window.

[0008] Preferably, the big data storage and computing module includes: Real-time data storage submodule: stores pre-processed real-time data, supports efficient reading and streaming computing. Real-time data will be passed to the streaming computing submodule for real-time computing of operating indicators; Historical data storage submodule: stores long-term data in a distributed file system, supports data query and offline analysis, and historical data is passed to the batch computing submodule for further analysis; Stream computing submodule: performs stream processing on real-time data, calculates real-time operating indicators and status information, and passes the results to the data anomaly detection and repair module; Batch computing submodule: performs offline computing and analysis on historical data to provide long-term trend data for the prediction module.

[0009] Preferably, the big data storage and calculation module performs streaming calculation on real-time data to calculate the key operating indicators KPI of the new energy station, and the calculation formula is: , in, is the key performance indicator at time t, For the The power output value at the moment, n is the size of the time window.

[0010] Preferably, the data anomaly detection and repair module includes: The anomaly detection submodule is used to receive data from the streaming computing submodule and the batch computing submodule, analyze the degree of deviation of the data, and detect abnormal values ​​in the data; The data repair submodule repairs the detected abnormal data, uses a regularization method or refers to the data of adjacent time points to correct the abnormal values, generates clean time series data, and passes it to the data prediction module; The data anomaly detection and repair module identifies outliers in the data through an anomaly detection formula, and the anomaly detection formula is: , in, is the anomaly score of the ith data point, is the value of the ith data point, is the mean of the data, is the standard deviation of the data.

[0011] Preferably, the data prediction module includes: The power generation prediction submodule is used to receive data from the data repair submodule, analyze historical features using a deep learning model, and predict wind power and photovoltaic power generation at future times; The grid load forecasting submodule predicts future grid load demand based on historical load data and electricity price fluctuation trends, and provides load information for the optimization module; The data prediction module predicts future power generation through a deep learning model, and the prediction model is defined as: , in, For the predicted time The power generation at the moment, For deep learning models, is the input feature vector.

[0012] Preferably, the intelligent optimization and analysis module includes: The resource scheduling optimization submodule optimizes the resource allocation strategy of the new energy station based on the results of the power generation prediction and grid load prediction submodules, and determines the power allocation plan for the power generation and energy storage systems; The equipment operation status analysis submodule uses the clean equipment data generated by the data repair submodule to analyze the equipment's operation status to identify potential failure risks; The energy storage and grid connection optimization submodule combines the resource scheduling optimization results to calculate the charging and discharging strategy of the energy storage system and optimize the power interaction between the station and the power grid.

[0013] Preferably, the intelligent optimization and analysis module optimizes the allocation of new energy station resources based on the prediction results, the optimization goal is to maximize the benefits, and the objective function is: , in, is the total revenue, is the real-time electricity price at time t, is the total power generated at time t, is the operating cost corresponding to the power generated at time t.

[0014] Preferably, the application and interaction module includes: The data display submodule is used to obtain the calculation results of the resource scheduling optimization submodule and the operation status analysis submodule, and to display the power generation, load demand, resource allocation and equipment status in the form of a graphical interface; The user operation interface submodule provides an interactive interface for users and receives input instructions from users; The remote management submodule supports users to access and control the operating status of new energy stations through remote devices.

[0015] A new energy station operation optimization method based on big data, comprising: Step 1: Acquire multi-source data inside and outside the new energy station, including equipment operation data, meteorological data and power grid data, perform preliminary processing on the collected data, complete data format unification, denoising and feature extraction, generate standardized time series data, and output the processed standardized time series data to step 2; Step 2: Receive the standardized time series data from step 1 and pass it to step 2 to perform anomaly detection on the received data, identify outliers in the data, and repair the detected outliers using a regularized processing method to generate a high-quality data stream, and pass the repaired data as output to step 3; Step 3: Receive the high-quality data stream from step 2, build a deep learning model based on multimodal data, and predict the power generation and grid load at future times. The prediction result is used as the output of step 3 and passed to step 4 for further optimizing resource allocation. Step 4: Receive the prediction result outputted from step 3, optimize the resource allocation of the new energy station based on the predicted power generation and grid load, analyze the operation status of the station, generate an optimized resource allocation plan and operation status information, and pass the optimization result as output to step 5; Step 5: Receive the optimized allocation plan and operating status information from step 4, visualize the results, and support users to remotely manage and monitor station operations through the provided operation interface. Users can adjust the station's operating parameters and resource allocation plan in the operation interface to further improve the operation optimization process of the new energy station.

[0016] The present invention provides a new energy station operation optimization system and method based on big data, which has the following beneficial effects: 1. The present invention can effectively identify and repair outliers in the data collected by new energy stations through the data anomaly detection and repair module, and adopts a regularized repair method combined with statistical characteristics to ensure the integrity and accuracy of the input data and reduce subsequent prediction and optimization deviations caused by data noise or collection errors.

[0017] 2. The present invention integrates multimodal data such as equipment operating status, meteorological conditions and grid demand, and uses deep learning models to capture the complex nonlinear relationship between multiple variables, to achieve high-precision prediction of power generation and grid load, and can dynamically adapt to the complexity and uncertainty of new energy sites, providing more reliable input for resource optimization and scheduling.

[0018] 3. Based on the prediction results of power generation and grid load, the present invention comprehensively considers the real-time electricity price, equipment operating status and energy storage conditions, adopts intelligent algorithms to generate resource scheduling plans, dynamically optimizes scheduling station resources and energy storage strategies, significantly improves power generation utilization and economic benefits, enhances the stability of the grid, and significantly improves the operation and management level of new energy stations with the global optimization capability driven by intelligent algorithms.

[0019] 4. The present invention realizes the coordinated control between the energy storage system and the power grid through dynamic planning, effectively alleviating the volatility problem of renewable energy power generation, and can adjust the energy storage charging and discharging strategy and grid-connected power output in real time, smooth out power fluctuations and reduce peak loads, reduce the impact of renewable energy sites on the stability of grid operation, and realize large-scale renewable energy grid connection, which has important technical significance and innovative value. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a system framework diagram of the present invention; Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0021] In order to make the technical personnel in the technical field understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in combination with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a partial embodiment of the present invention, not a complete embodiment. Based on the embodiment of the present invention, other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present invention.

[0022] The present invention is described in detail below in conjunction with the accompanying drawings: Example: Please refer to the attached Figure 1 and attached Figure 2 The embodiment of the present invention provides a new energy station operation optimization system based on big data, including: Data collection and integration module: The multi-source data inside and outside the new energy station are collected through its internal multi-source data collection sub-module, including equipment operation data, meteorological data and power grid data. The collected data is preliminarily processed by the edge computing pre-processing sub-module, including format unification, denoising and feature extraction, to generate standardized time series data. After processing, the standardized time series data is transmitted to the big data storage and computing module; Big Data Storage and Computing Module: Receives standardized time series data from the Data Collection and Integration Module and processes it through the submodules in the Big Data Storage and Computing Module: Data anomaly detection and repair module: receives data from the big data storage and computing module, performs anomaly detection operations on the data, identifies outliers in the data, repairs outliers based on regular processing methods, generates high-quality data streams, and the repaired data is passed to the data prediction module; Data prediction module: receives high-quality data streams from the data anomaly detection and repair module, builds a deep learning model using multimodal data, and predicts future power generation and grid load. The prediction results are passed as input to the intelligent optimization and analysis module to optimize resource allocation. Intelligent optimization and analysis module: Based on the results of the data prediction module, the resource allocation of the new energy stations is optimized. At the same time, the operating status of the stations is analyzed. The optimized allocation plan and operating status information are transmitted to the application and interaction module for further operation and monitoring by users. Application and interaction module: Receives the results from the intelligent optimization and analysis module, displays the optimization allocation and operation status information, and allows users to remotely manage and monitor through the provided operation interface. It supports the adjustment of station operation parameters and the management of resource allocation plans.

[0023] The data collection and integration module can obtain multi-source data inside and outside the new energy station, including equipment operation data, meteorological data and power grid data, and pre-process and format the data through edge computing devices to ensure data consistency and integrity, providing basic data for subsequent analysis and optimization; The big data storage and computing module adopts a hot-cold tiered storage architecture, storing real-time data in an in-memory database to support fast computing, and storing historical data in a distributed file system to support long-term analysis. At the same time, it combines streaming computing and batch processing methods to efficiently process real-time and historical data, providing computing power for real-time monitoring and long-term optimization of station operations. The data anomaly detection and repair module detects and repairs the collected data through regularization methods, effectively identifies and corrects outliers, ensures the accuracy and completeness of input data, reduces noise interference caused by collection errors or emergencies, and provides input for subsequent prediction and optimization modules; The data prediction module achieves high-precision prediction of future power generation and grid load by integrating multimodal data and building a deep learning model. Compared with traditional linear prediction methods, it can capture the complex nonlinear relationship between multiple variables, significantly improve the accuracy of prediction, and provide a scientific basis for resource scheduling and operation optimization. Based on the prediction results, the intelligent optimization and analysis module uses intelligent optimization algorithms to dynamically dispatch station resources, generate power distribution for power generation equipment, charge and discharge plans for energy storage equipment, and grid-connected power distribution strategies. At the same time, it analyzes the station operation status, identifies potential problems, and provides support for efficient station management. The application and interaction module displays the optimization results and station operation status in a graphical manner, so that users can intuitively understand information such as power generation power distribution, load demand changes and equipment health status. In addition, the module provides an interactive interface, so users can adjust optimization parameters or initiate scheduling tasks in real time, making station operation management more flexible and convenient.

[0024] In the data acquisition and integration module, the denoising process uses a sliding window algorithm: , in, is the data after smoothing. For the The original data value at the moment, n is the length of the sliding window.

[0025] The multi-source data acquisition submodule obtains the equipment operation data, meteorological data and power grid data inside and outside the new energy station. By integrating multi-dimensional information, the system can fully reflect the operation status of the new energy station and provide a rich data foundation for subsequent optimization; The edge computing preprocessing submodule unifies the format of the collected data, eliminates processing obstacles caused by inconsistent data structures, and generates standardized time series data, so that subsequent modules can seamlessly receive and process multi-source data, ensuring smooth circulation and application of data in the system; De-noising time series data through a sliding window algorithm can smooth fluctuations in the data, reduce the interference of noise on subsequent analysis, enhance the continuity and stability of the time series, and generate more reliable input data; The edge computing preprocessing submodule performs preliminary feature extraction at the data acquisition end, extracts key indicators from the raw data, reduces the processing burden of subsequent modules, can quickly identify the core status of equipment operation, and provide efficient data support for subsequent prediction and optimization.

[0026] Big data storage and computing modules include: Real-time data storage submodule: stores pre-processed real-time data, supports efficient reading and streaming computing, and the real-time data will be passed to the streaming computing submodule for real-time computing of operating indicators; Historical data storage submodule: stores long-term data in a distributed file system, supports data query and offline analysis, and historical data is passed to the batch computing submodule for further analysis; Stream computing submodule: performs stream processing on real-time data, calculates real-time operating indicators and status information, and passes the results to the data anomaly detection and repair module; Batch computing submodule: performs offline computing and analysis on historical data to provide long-term trend data for the prediction module.

[0027] The real-time data storage submodule can quickly receive and store pre-processed data from the data acquisition and integration module. At the same time, it supports efficient data reading and real-time computing capabilities, ensuring the performance stability of the system when processing high-frequency data, and providing rapid response capabilities for real-time monitoring and optimization decisions; The historical data storage submodule stores long-term data in a distributed file system according to rules and categories, supports persistent management and fast query of large-scale data, provides reliable basic data support for historical trend analysis and data mining through an efficient retrieval mechanism, and accumulates long-term data resources for prediction and optimization modules; The streaming computing submodule can dynamically process real-time data, calculate key operating indicators and real-time status information of the station, and pass the results to the anomaly detection and repair module. It can obtain the operating status of the new energy station in a timely manner and provide rapid feedback for anomaly detection, prediction and scheduling optimization; The batch calculation submodule extracts long-term trends and patterns in the data through offline analysis of historical data, and generates characteristic data that is instructive for the prediction module. It can make up for the limitations of real-time calculations and help the system establish a more comprehensive new energy station operation model.

[0028] The big data storage and computing module performs streaming computing on real-time data and calculates the key operating indicators KPI of new energy stations. The calculation formula is: , in, is the key performance indicator at time t, For the The power output value at the moment, n is the size of the time window.

[0029] By calculating key operating indicators, the system can reflect the power output characteristics of the new energy station at time t in real time. Through the average power data in the window, this indicator can effectively describe the current power generation level and provide a core reference for real-time monitoring and operation optimization. The key operating indicators calculate the average power output through a sliding time window, which can eliminate the impact of sharp fluctuations in single-point data. Compared with the power value at a single moment, Providing a stable and representative description of the operating status helps improve the data reliability of subsequent modules; By continuous calculation The system can dynamically monitor the operating performance of power generation equipment, and when indicators change abnormally, it can assist in determining whether the equipment has experienced a decrease in efficiency, excessive power fluctuations, or other operating problems.

[0030] The data anomaly detection and repair module includes: The anomaly detection submodule is used to receive data from the streaming computing submodule and the batch computing submodule, analyze the degree of deviation of the data, and detect abnormal values ​​in the data; The data repair submodule repairs the detected abnormal data, uses a regularization method or refers to the data of adjacent time points to correct the abnormal values, generates clean time series data, and passes it to the data prediction module; The data anomaly detection and repair module identifies outliers in the data through the anomaly detection formula. The anomaly detection formula is: , in, is the anomaly score of the ith data point, is the value of the ith data point, is the mean of the data, is the standard deviation of the data.

[0031] The anomaly detection submodule uses anomaly detection formulas to accurately analyze the degree of deviation in data and quickly identify outliers in power output, meteorological conditions, or load demand by calculating the standardized difference between each data point and the mean; The data repair submodule corrects outliers through regularization methods or by referring to data at adjacent time points to generate continuous, non-anomaly time series data, avoiding the impact of data mutation or loss on subsequent processing. The repaired data can provide reliable input, laying a good data foundation for deep learning prediction models and intelligent optimization algorithms. Through the anomaly detection formula, the system can quickly respond to anomalies in the collected data, such as anomalies caused by sensor failure, network transmission problems or sudden environmental changes. The repaired data eliminates the interference of anomalies on operation analysis and enhances the stability and robustness of the system in complex environments.

[0032] The data prediction module includes: The power generation prediction submodule is used to receive data from the data repair submodule, analyze historical features using a deep learning model, and predict wind power and photovoltaic power generation at future times; The grid load forecasting submodule predicts future grid load demand based on historical load data and electricity price fluctuation trends, and provides load information for the optimization module; The data prediction module predicts future power generation through a deep learning model. The prediction model is defined as: , in, For the predicted time The power generation at the moment, For deep learning models, is the input feature vector.

[0033] The power generation prediction submodule uses a deep learning model to analyze the characteristics of historical data, which can capture the nonlinear changes in wind power and photovoltaic power generation. By integrating multiple influencing factors such as equipment operating status and meteorological conditions, the prediction results are more accurate, providing reliable power generation estimates for power allocation and scheduling optimization of new energy sites. The grid load forecasting submodule combines historical load data and real-time electricity price fluctuation trends to dynamically forecast future grid load demand. By capturing the time series characteristics of grid load changes, it provides detailed load information for the resource scheduling optimization module, helping stations better match power generation capacity with grid demand. The deep learning model of the data prediction module can integrate equipment operation data, meteorological data and historical load data to form multimodal feature input. Compared with the traditional single variable prediction method, the fusion of multimodal data significantly improves the accuracy and reliability of the prediction, and adapts to the needs of new energy station operation in complex scenarios.

[0034] Intelligent optimization and analysis modules include: The resource scheduling optimization submodule optimizes the resource allocation strategy of the new energy station based on the results of the power generation prediction and grid load prediction submodules, and determines the power allocation plan for the power generation and energy storage systems; The equipment operation status analysis submodule uses the clean equipment data generated by the data repair submodule to analyze the equipment's operation status to identify potential failure risks; The energy storage and grid connection optimization submodule combines the resource scheduling optimization results to calculate the charging and discharging strategy of the energy storage system and optimize the power interaction between the station and the power grid.

[0035] The resource scheduling optimization submodule dynamically optimizes the resource allocation strategy of new energy stations based on the results of the power generation prediction and grid load prediction submodules, and determines the power allocation plan of power generation equipment and energy storage system in real time through intelligent algorithms. This submodule can effectively balance power generation capacity and load demand, improve the utilization efficiency of station resources, and avoid energy waste or shortage. The equipment operation status analysis submodule uses the high-quality equipment operation data generated by the data repair submodule to monitor the equipment's operation status in real time and analyze its health status. By identifying potential failure risks in equipment operation, this submodule provides a basis for site managers to plan maintenance in advance, reduce unplanned downtime caused by failures, and improve the safety and reliability of equipment. The energy storage and grid-connection optimization submodule calculates the charging and discharging strategies of the energy storage equipment based on the resource scheduling optimization results, and rationally arranges the energy flow of the energy storage system. Through intelligent energy storage management, this submodule can achieve peak shaving and valley filling and smooth power output, improve the economy and utilization rate of the energy storage system, and provide an effective buffer for the volatility of renewable energy power generation.

[0036] The intelligent optimization and analysis module optimizes the allocation of new energy station resources based on the prediction results. The optimization goal is to maximize the benefits. The objective function is: , in, is the total revenue, is the real-time electricity price at time t, is the total power generated at time t, is the operating cost corresponding to the power generated at time t.

[0037] The intelligent optimization and analysis module optimizes the resource allocation strategy through the objective function, taking the benefit as the optimization target. By comprehensively considering the real-time electricity price, power generation and operating cost, the module can dynamically adjust the resource utilization to ensure that the new energy station obtains the maximum economic benefit during the operation cycle; The objective function includes the real-time electricity price. During the optimization process, the market opportunities during the peak electricity price period can be fully utilized to concentrate the output of power generation to obtain higher profits. During the low electricity price period, the power generation is reduced or the energy storage system is reasonably dispatched to avoid energy waste and improve profit elasticity. Through the power generation and operating costs in the objective function, the optimization algorithm can refine the resource allocation strategy of the new energy station, dynamically determine the power output of each device and the charging and discharging behavior of energy storage, reduce resource waste, and ensure that the station operation meets the goal of maximizing revenue; The operating cost in the objective function serves as a constraint condition, which can balance the relationship between power generation efficiency and operating cost. By controlling the power output level and optimizing the algorithm, it can avoid the erosion of revenue by excessive operating costs and achieve cost-effectiveness.

[0038] Application and interaction modules include: The data display submodule is used to obtain the calculation results of the resource scheduling optimization submodule and the operation status analysis submodule, and to display the power generation, load demand, resource allocation and equipment status in the form of a graphical interface; The user operation interface submodule provides an interactive interface for users and receives input instructions from users; The remote management submodule supports users to access and control the operating status of new energy stations through remote devices.

[0039] The data display submodule displays the calculation results of the resource scheduling optimization submodule and the operation status analysis submodule through a graphical interface, including power generation, load demand, resource allocation and equipment operation status. The graphical presentation makes complex data and optimization results easier to understand, helping users quickly understand the overall operation of new energy stations and reducing management difficulty. The user operation interface submodule provides an interactive interface for users and can receive user input commands, including adjusting scheduling parameters, setting power allocation strategies or triggering maintenance plans. Through flexible interactive design, users can adjust the station operation plan according to real-time needs, enhancing management flexibility and customization capabilities; The remote management submodule allows users to access the operating status of new energy stations through remote devices, providing real-time monitoring and remote control capabilities. Users can understand the power generation situation and equipment status of the station remotely and issue operating instructions, breaking the traditional management model’s reliance on on-site operations and improving management efficiency.

[0040] A new energy station operation optimization method based on big data, comprising: Step 1: Acquire multi-source data inside and outside the new energy station, including equipment operation data, meteorological data and power grid data, perform preliminary processing on the collected data, complete data format unification, denoising and feature extraction, generate standardized time series data, and output the processed standardized time series data to step 2; Step 2: Receive the standardized time series data from step 1 and pass it to step 2 to perform anomaly detection on the received data, identify outliers in the data, and repair the detected outliers using a regularized processing method to generate a high-quality data stream, and pass the repaired data as output to step 3; Step 3: Receive the high-quality data stream from step 2, build a deep learning model based on multimodal data, and predict the power generation and grid load at future times. The prediction result is used as the output of step 3 and passed to step 4 for further optimizing resource allocation. Step 4: Receive the prediction result outputted from step 3, optimize the resource allocation of the new energy station based on the predicted power generation and grid load, analyze the operation status of the station, generate an optimized resource allocation plan and operation status information, and pass the optimization result as output to step 5; Step 5: Receive the optimized allocation plan and operating status information from step 4, visualize the results, and support users to remotely manage and monitor station operations through the provided operation interface. Users can adjust the station's operating parameters and resource allocation plan in the operation interface to further improve the operation optimization process of the new energy station.

[0041] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A new energy station operation optimization system based on big data, characterized in that: include: Data collection and integration module: The multi-source data inside and outside the new energy station are collected through its internal multi-source data collection sub-module, including equipment operation data, meteorological data and power grid data. The collected data is preliminarily processed by the edge computing pre-processing sub-module, including format unification, denoising and feature extraction, to generate standardized time series data. After processing, the standardized time series data is transmitted to the big data storage and computing module; Big Data Storage and Computing Module: Receives standardized time series data from the Data Collection and Integration Module and processes it through the submodules in the Big Data Storage and Computing Module: Data anomaly detection and repair module: receives data from the big data storage and computing module, performs anomaly detection operations on the data, identifies outliers in the data, repairs outliers based on regular processing methods, generates high-quality data streams, and the repaired data is passed to the data prediction module; Data prediction module: receives high-quality data streams from the data anomaly detection and repair module, builds a deep learning model using multimodal data, and predicts future power generation and grid load. The prediction results are passed as input to the intelligent optimization and analysis module to optimize resource allocation. Intelligent optimization and analysis module: Based on the results of the data prediction module, the resource allocation of the new energy stations is optimized. At the same time, the operating status of the stations is analyzed. The optimized allocation plan and operating status information are transmitted to the application and interaction module for further operation and monitoring by users; Application and interaction module: Receives the results from the intelligent optimization and analysis module, displays the optimization allocation and operation status information, and allows users to remotely manage and monitor through the provided operation interface. It supports the adjustment of station operation parameters and the management of resource allocation plans.

2. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: In the data acquisition and integration module, the denoising process uses a sliding window algorithm: ， in, is the data after smoothing. For the The original data value at the moment, n is the length of the sliding window.

3. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: The big data storage and computing module includes: Real-time data storage submodule: stores pre-processed real-time data, supports efficient reading and streaming computing. Real-time data will be passed to the streaming computing submodule for real-time computing of operating indicators; Historical data storage submodule: stores long-term data in a distributed file system, supports data query and offline analysis, and historical data is passed to the batch computing submodule for further analysis; Stream computing submodule: performs stream processing on real-time data, calculates real-time operating indicators and status information, and passes the results to the data anomaly detection and repair module; Batch computing submodule: performs offline computing and analysis on historical data to provide long-term trend data for the prediction module.

4. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: The big data storage and calculation module performs streaming calculations on real-time data to calculate the key operating indicators KPI of the new energy station. The calculation formula is: ， in, is the key performance indicator at time t, For the The power output value at the moment, n is the size of the time window.

5. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: The data anomaly detection and repair module includes: The anomaly detection submodule is used to receive data from the streaming computing submodule and the batch computing submodule, analyze the degree of deviation of the data, and detect abnormal values ​​in the data; The data repair submodule repairs the detected abnormal data, uses a regularization method or refers to the data of adjacent time points to correct the abnormal values, generates clean time series data, and passes it to the data prediction module; The data anomaly detection and repair module identifies outliers in the data through an anomaly detection formula, and the anomaly detection formula is: ， in, is the anomaly score of the ith data point, is the value of the ith data point, is the mean of the data, is the standard deviation of the data.

6. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: The data prediction module comprises: The power generation prediction submodule is used to receive data from the data repair submodule, analyze historical features using a deep learning model, and predict wind power and photovoltaic power generation at future times; The grid load forecasting submodule predicts future grid load demand based on historical load data and electricity price fluctuation trends, and provides load information for the optimization module; The data prediction module predicts future power generation through a deep learning model, and the prediction model is defined as: , in, For the predicted time The power generation at the moment, For deep learning models, is the input feature vector.

7. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: The intelligent optimization and analysis module includes: The resource scheduling optimization submodule optimizes the resource allocation strategy of the new energy station based on the results of the power generation prediction and grid load prediction submodules, and determines the power allocation plan for the power generation and energy storage systems; The equipment operation status analysis submodule uses the clean equipment data generated by the data repair submodule to analyze the equipment's operation status to identify potential failure risks; The energy storage and grid connection optimization submodule combines the resource scheduling optimization results to calculate the charging and discharging strategy of the energy storage system and optimize the power interaction between the station and the power grid.

8. According to the big data-based new energy station operation optimization system of claim 1, it is characterized in that: The intelligent optimization and analysis module optimizes the allocation of new energy station resources based on the prediction results. The optimization goal is to maximize the benefits. The objective function is: ， in, is the total revenue, is the real-time electricity price at time t, is the total power generated at time t, is the operating cost corresponding to the power generated at time t.

9. The new energy station operation optimization system based on big data according to claim 1 is characterized in that: The application and interaction module includes: The data display submodule is used to obtain the calculation results of the resource scheduling optimization submodule and the operation status analysis submodule, and to display the power generation, load demand, resource allocation and equipment status in the form of a graphical interface; The user operation interface submodule provides an interactive interface for users and receives input instructions from users; The remote management submodule supports users to access and control the operating status of new energy stations through remote devices.

10. A method for optimizing the operation of a new energy station based on big data, according to a system for optimizing the operation of a new energy station based on big data according to any one of claims 1 to 9, characterized in that: include: Step 1: Acquire multi-source data inside and outside the new energy station, including equipment operation data, meteorological data and power grid data, perform preliminary processing on the collected data, complete data format unification, denoising and feature extraction, generate standardized time series data, and output the processed standardized time series data to step 2; Step 2: Receive the standardized time series data from step 1 and pass it to step 2 to perform anomaly detection on the received data, identify outliers in the data, and repair the detected outliers using a regularized processing method to generate a high-quality data stream, and pass the repaired data as output to step 3; Step 3: Receive the high-quality data stream from step 2, build a deep learning model based on multimodal data, and predict the power generation and grid load at future times. The prediction result is used as the output of step 3 and passed to step 4 for further optimizing resource allocation. Step 4: Receive the prediction result outputted from step 3, optimize the resource allocation of the new energy station based on the predicted power generation and grid load, analyze the operation status of the station, generate an optimized resource allocation plan and operation status information, and pass the optimization result as output to step 5; Step 5: Receive the optimized allocation plan and operating status information from step 4, visualize the results, and support users to remotely manage and monitor station operations through the provided operation interface. Users can adjust the station's operating parameters and resource allocation plan in the operation interface to further improve the operation optimization process of the new energy station.

Images