A new energy station operation intelligent management method and system

Abstract

The present invention provides a method and system for intelligent management of the operation of a new energy station, which relates to the technical field of operation of a new energy station. The method includes: collecting various types of data in real time through sensors and monitoring equipment in the new energy station; transmitting the collected various types of data to a central data center and performing iterative optimization processing to determine the final combination of data classification, cleaning and compression parameters to build a big data platform for a new energy station; and constructing a digital twin model of the new energy station based on the constructed big data platform of the new energy station to reflect the operation status, equipment health status and environmental changes of the station in real time. The present invention realizes the automatic generation and closed-loop management of the operation optimization plan through real-time data collection, digital twin model construction and simulation analysis, which effectively improves the operation efficiency, prediction accuracy and fault response capability of the new energy station.

Description

Technical Field

[0001] The present invention relates to the field of new energy station operation technology, and in particular to a new energy station operation intelligent management method and system. Background Art

[0002] The traditional operation and management method of new energy stations relies heavily on manual experience for decision-making and combines some basic automatic control systems to monitor the operation status of the stations. However, this method has certain limitations in terms of comprehensiveness and real-time performance, as shown in the following aspects:

[0003] The machine system may only be able to obtain limited key parameters and cannot fully capture all the detailed information of the station operation. The frequency and accuracy of data collection may be limited, resulting in weak real-time perception of the station status. Data processing and analysis often rely on preset rules and thresholds, lacking the ability of deep mining and intelligent analysis. The decision-making process is mainly based on manual experience and may be limited by personal knowledge, experience and judgment. Although the automated control system can perform some preset operations, it lacks the ability to make autonomous decisions and respond flexibly when faced with complex or unexpected situations. There may be a certain delay in decision-making because human intervention is required or waiting for the system to complete a fixed processing flow.

[0004] Due to the limitations of information acquisition and processing, as well as the delay in decision-making, the system may be slow to respond to changes in the station status. When equipment fails or energy output fluctuates, it may not be possible to take effective measures in time for adjustment and optimization. Due to the lack of comprehensive real-time monitoring and precise optimization strategies, energy output may fluctuate greatly due to multiple factors such as equipment performance and environmental factors. Equipment operation may not be in the optimal state, resulting in low energy conversion efficiency and low output efficiency. Fault warning and diagnosis mainly rely on manual inspections and simple automated alarm systems, which may not be able to detect potential fault hazards in time. Fault handling and maintenance are often carried out after the problem occurs, lacking foresight and initiative, resulting in frequent equipment failures and increased maintenance costs. Due to low energy output efficiency and frequent equipment failures, operating costs may rise, including energy consumption, equipment maintenance, labor costs, etc. The lack of effective optimization strategies and control measures makes it difficult to effectively control operating costs. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a smart management method for the operation of a new energy station, so as to realize comprehensive and real-time monitoring and management of the station.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0007] In a first aspect, a method for intelligent management of operation of a new energy station is provided, the method comprising:

[0008] Collect various data in real time through sensors and monitoring equipment in new energy stations;

[0009] The collected data is transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a new energy station big data platform;

[0010] Based on the big data platform of the new energy station, a digital twin model of the new energy station is constructed to reflect the operation status, equipment health status and environmental changes of the station in real time;

[0011] Analyze the data of the digital twin model and predict the energy output and demand trends and equipment failures in the future by simulating different operating scenarios to obtain analysis results;

[0012] Based on the analysis results and real-time feedback from the digital twin model, an operation optimization plan is automatically generated to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic arrays, and energy storage systems;

[0013] Apply the operation optimization plan to actual operation, monitor the actual operation effect in real time, and obtain operation data;

[0014] Compare and analyze the operating data with the predictions of the digital twin model, continuously adjust the digital twin model and optimize the operating plan to form a closed-loop management.

[0015] Furthermore, the collected data will be transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a new energy station big data platform, including:

[0016] Set an objective function to evaluate the data processing performance of different parameter combinations, including classification accuracy, cleaning effect, and compression rate;

[0017] A set of parameter combinations are randomly generated as the initial harmony library, each harmony represents a data classification, cleaning and compression scheme,

[0018] A group of harmonies is randomly determined from the harmony library as the current harmony, and a new harmony is generated by randomly adjusting some parameters in the current harmony;

[0019] Calculate the objective function value corresponding to the new harmony, evaluate its quality, and obtain the evaluation result;

[0020] Continuously perform harmony selection, parameter adjustment, new harmony evaluation and harmony library update operations until the preset number of iterations is reached to obtain the final data classification, cleaning and compression parameter combination. The data is processed according to the final data classification, cleaning and compression parameter combination and stored in the central data center to build a big data platform for new energy sites.

[0021] Furthermore, the calculation formula of the objective function value corresponding to the new harmony is:

[0022] ;

[0023] in, represents the objective function value corresponding to the new harmony; Represents the weight coefficient of classification accuracy; Indicates classification accuracy; The weight coefficient indicating the cleaning effect; Indicates the number of valid data after cleaning; Indicates the number of invalid data after cleaning; Indicates the number of erroneous data after cleaning; Indicates the number of original valid data; Indicates the number of original invalid data; Indicates the number of raw error data; Indicates the number of reference valid data; Indicates the number of reference invalid data; Indicates the number of reference error data; represents a positive number; represents the exponential parameter; The weight coefficient representing the compression ratio; Indicates the adjusted compression ratio.

[0024] Furthermore, based on the big data platform of the new energy station, a digital twin model of the new energy station is constructed to reflect the operation status, equipment health status and environmental changes of the station in real time, including:

[0025] Extract data from the big data platform of new energy stations, including equipment operation data, environmental monitoring data and equipment maintenance records;

[0026] Based on the extracted data, determine the structure and framework of the digital twin model, including determining the model's hierarchy, the relationships between components, and the direction of data flow;

[0027] A digital twin model of the new energy station is constructed based on the structure and framework of the digital twin model, and the real-time data in the big data platform is mapped to the digital twin model to reflect the operating status, equipment health status and environmental changes of the station in real time.

[0028] Furthermore, the data of the digital twin model is analyzed, and by simulating different operating scenarios, the energy output and demand trends and equipment failures in the future are predicted, and the analysis results are obtained, including:

[0029] Integrate the real-time equipment operation data, environmental monitoring data and equipment maintenance records obtained in the digital twin model to obtain a data set containing multi-dimensional information;

[0030] Analyze multi-dimensional information data sets to identify key characteristics related to energy output, demand trends, and equipment failures, including equipment operating parameters, extreme values ​​of environmental factors, and frequency of equipment maintenance;

[0031] The time series analysis model is used as a prediction model, and the key features and corresponding historical data are used as input data to train the prediction model to obtain a trained prediction model;

[0032] Define multiple operating scenarios, including different weather conditions and energy demand changes, and use the defined multiple operating scenarios as input conditions to import the trained prediction model, including parameters of weather conditions and energy demand change patterns;

[0033] The prediction function of the prediction model is used to simulate the energy output and demand trends in different scenarios, and predict equipment failures to obtain prediction results.

[0034] Furthermore, the calculation formula of the prediction function is:

[0035] ;

[0036] in, Represents the predicted output value; represents the predicted energy output; represents the predicted energy demand; represents the predicted frequency of equipment failure; Indicates the weight coefficient of equipment operation data; Indicates equipment operation data; Represents the weight coefficient of time series data; Time series data representing historical energy production and demand records; Represents the weight coefficient of environmental data; Represents environmental data, defined as: ; Indicates the ambient temperature; Indicates humidity; Indicates wind speed; Indicates ambient air pressure; Represents the bias term of the prediction model; represents the number of interaction terms; represents the coefficient of the interaction effect; A coefficient indicating the effect of compression rate; represents a positive number; Represents the index of the interaction term.

[0037] Furthermore, based on the analysis results and real-time feedback from the digital twin model, an operation optimization plan is automatically generated to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic arrays, and energy storage systems, including:

[0038] Encode the start and stop of wind turbines, photovoltaic panels, and energy storage systems and the adjustment of equipment parameters into chromosomes;

[0039] A set of initial solutions is randomly generated, each of which represents a running optimization plan;

[0040] Using the analysis results and real-time feedback from the digital twin model, the fitness of each solution is evaluated, and the population is iteratively updated through selection, crossover, and mutation operations to obtain the final solution;

[0041] Continue to iterate. When the preset number of iterations is reached, the initial operation optimization plan is determined from the final population according to the fitness of the solution.

[0042] Parsing the initial operation optimization plan into control instructions and parameter adjustment values;

[0043] According to the control instructions, the equipment is remotely started and stopped, and according to the parameter adjustment values, the operating parameters of the equipment are remotely adjusted.

[0044] Furthermore, the fitness calculation formula of each solution is:

[0045] ;

[0046] in, Indicates The fitness value of each solution; , , , represents the weight coefficient; Indicated in Understand the actual energy output of new energy stations; Indicates the maximum energy output; Indicated in Understand the predicted energy demand; Indicates the actual energy demand; Indicates the maximum demand difference; Indicated in Understand the equipment failure rate; Indicated in Let’s understand the running costs.

[0047] Furthermore, the operation data is compared and analyzed with the predictions of the digital twin model, and the digital twin model is continuously adjusted and the operation plan is optimized to form a closed-loop management, including:

[0048] Compare and analyze the actual operation data with the predicted data of the digital twin model to identify differences and deviations;

[0049] Analyze the differences and deviations and obtain analysis results;

[0050] According to the analysis results, the digital twin model is adjusted, including correcting the digital twin model parameters, optimizing the digital twin model structure, and optimizing the current operation plan to obtain the optimized and adjusted operation plan;

[0051] According to the optimized and adjusted operation plan, the wind turbines, photovoltaic panel arrays and energy storage system equipment are remotely controlled and parameter adjusted, and the operating status and performance parameters of the equipment are monitored in real time. The real-time operation data is fed back to the digital twin model, and the cycle is continuously iterated to form a closed-loop management.

[0052] Second, a new energy station operation intelligent management system includes:

[0053] The acquisition module is used to collect various data in real time through sensors and monitoring equipment in new energy stations;

[0054] The collected data is transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a big data platform for new energy stations. Based on the constructed big data platform for new energy stations, a digital twin model of the new energy station is constructed to reflect the operating status of the station, equipment health status and environmental changes in real time.

[0055] The processing module is used to analyze the data of the digital twin model, and by simulating different operating scenarios, predict the energy output and demand trends and equipment failures in the future to obtain analysis results; based on the analysis results and real-time feedback from the digital twin model, automatically generate an operation optimization plan to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic panel arrays, and energy storage systems; apply the operation optimization plan to actual operation, monitor the actual operation results in real time, and obtain operation data; compare and analyze the operation data with the predictions of the digital twin model, continuously adjust the digital twin model and optimize the operation plan to form a closed-loop management.

[0056] The above solution of the present invention includes at least the following beneficial effects:

[0057] By collecting and analyzing various data of new energy stations in real time, we can more accurately grasp the operating status and environmental changes of the stations, so as to formulate more optimized operating strategies. This helps wind turbines, photovoltaic arrays and other equipment to operate in the best state, improve energy conversion efficiency, and thus increase energy output. Building a digital twin model can reflect the health status of the equipment in real time. By analyzing the model data, potential fault hazards can be discovered in time and early warning can be issued. This helps to carry out equipment maintenance in advance and avoid failures, thereby reducing equipment failure rates, maintenance costs and downtime. By simulating different operating scenarios, it is possible to predict energy output and demand trends in the future, so as to formulate energy scheduling plans. This helps to avoid energy waste and shortages, optimize energy configuration, and reduce operating costs.

[0058] Relying on the big data platform and digital twin model, it can provide comprehensive and accurate data support, making the decision-making process more scientific and accurate. This helps managers make correct decisions in a complex and changing operating environment and improve management efficiency. Through real-time monitoring and feedback mechanisms, it can quickly respond to changes in the status of the station and adjust the operation optimization plan in time. This helps to shorten the system response time, improve the flexibility and adaptability of the system, and ensure that the station always remains in operation. By continuously optimizing energy output efficiency, reducing equipment failure rate and operating costs, it helps to improve the overall economic and environmental benefits of new energy stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a flow chart of a new energy station operation intelligent management method provided by an embodiment of the present invention.

[0060] Figure 2 It is a schematic diagram of a new energy station operation intelligent management system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0062] like Figure 1 As shown, an embodiment of the present invention provides a new energy station operation intelligent management method, the method comprising the following steps:

[0063] Step 11, collect various data in real time through sensors and monitoring equipment in the new energy station;

[0064] Step 12: Transmit the collected data to the central data center and perform iterative optimization processing to determine the final combination of data classification, cleaning and compression parameters to build a new energy station big data platform;

[0065] Step 13: Based on the constructed big data platform of the new energy station, a digital twin model of the new energy station is constructed to reflect the operation status, equipment health status and environmental changes of the station in real time;

[0066] Step 14: Analyze the data of the digital twin model, and predict the energy output and demand trends and equipment failures in the future by simulating different operation scenarios to obtain analysis results;

[0067] Step 15: Based on the analysis results and real-time feedback from the digital twin model, an operation optimization plan is automatically generated to guide the remote start and stop and parameter adjustment of the wind turbine, photovoltaic panel array, and energy storage system;

[0068] Step 16, applying the operation optimization plan to actual operation, and monitoring the actual operation effect in real time to obtain operation data;

[0069] Step 17: Compare and analyze the operating data with the predictions of the digital twin model, continuously adjust the digital twin model and optimize the operating plan to form a closed-loop management.

[0070] In an embodiment of the present invention, various types of data in the new energy station are collected in real time and transmitted to the central data center for iterative optimization processing, thereby ensuring the accuracy, completeness and timeliness of the data. This makes management decisions more data-driven, improves the level of refined management, and can more accurately grasp the operating status and performance of the station. By building a digital twin model of a new energy station, real-time monitoring of the station's operating status, equipment health, and environmental changes is achieved. At the same time, by simulating different operating scenarios, the digital twin model can predict energy output and demand trends and equipment failures in the future, providing strong support for management decisions. According to the analysis results and real-time feedback of the digital twin model, an operation optimization plan is automatically generated, and the remote start and stop and parameter adjustment of wind turbines, photovoltaic panel arrays, and energy storage systems are guided. This helps to optimize energy configuration, improve energy output efficiency, reduce energy waste, and thus improve the economic benefits of new energy stations.

[0071] Through the prediction function of the digital twin model, potential hidden faults can be discovered in time and early warning can be issued. This makes equipment maintenance more proactive and preventive, reduces equipment failure rate, reduces maintenance costs and downtime, and improves equipment reliability and service life. The operation optimization plan is applied to actual operation, and the actual operation effect is monitored in real time to obtain operation data. By comparing and analyzing the operation data with the prediction of the digital twin model, the digital twin model is continuously adjusted and the operation plan is optimized to form a closed-loop management. This helps to continuously improve and optimize the operation management of new energy stations, and continuously improve the management level and economic benefits. The application of advanced technical means such as real-time data collection, digital twin model construction, operation scenario simulation and optimization has enhanced the adaptability and flexibility of new energy stations to complex and changing environments. Regardless of weather conditions, equipment status or market demand, the operation strategy can be adjusted quickly to ensure that the station always remains in operation.

[0072] In a preferred embodiment of the present invention, the above step 11, collecting various data in real time through sensors and monitoring equipment in the new energy station, may include:

[0073] According to the layout, equipment type and operating characteristics of the new energy site, determine the key points and equipment that need to be monitored, such as the speed, power, and temperature of the wind turbine, the voltage, current, and irradiance of the photovoltaic panel array, and the power, temperature, and charging and discharging status of the energy storage system. For each monitoring point, select the appropriate sensor type, such as temperature sensor, humidity sensor, current sensor, voltage sensor, wind speed sensor, irradiance sensor, etc., to ensure that the required data can be accurately collected. Install the selected sensors in the relevant equipment and areas of the new energy site in the predetermined position and manner to ensure that the sensors can work stably and accurately. Install a data collector near each monitoring point or equipment to receive the data transmitted by the sensor and perform preliminary processing and storage. The data collector can select an intelligent device with data preprocessing, data compression and communication functions.

[0074] Establish a stable communication network to connect the data collector to the central data center. The communication network can be a wired network (such as optical fiber, Ethernet) or a wireless network (such as Wi-Fi, LoRa, NB-IoT, etc.), depending on the actual situation of the site and communication needs. The data collector transmits the collected data to the central data center in real time through the communication network.

[0075] In a preferred embodiment of the present invention, the above step 12, transmitting the collected data to the central data center and performing iterative optimization processing to determine the final combination of data classification, cleaning and compression parameters to build a new energy station big data platform, may include:

[0076] Step 122, setting an objective function to evaluate the data processing performance of different parameter combinations, including classification accuracy, cleaning effect, and compression rate;

[0077] Step 123, randomly generating a set of parameter combinations as an initial harmony library, each harmony representing a data classification, cleaning and compression scheme;

[0078] Step 124, randomly determining a group of harmonies from the harmony library as the current harmony, and generating a new harmony by randomly adjusting some parameters in the current harmony;

[0079] Step 125, calculating the objective function value corresponding to the new harmony, evaluating the quality, and obtaining the evaluation result;

[0080] Step 126, continuously perform harmony selection, parameter adjustment, new harmony evaluation and harmony library update operations until a preset number of iterations is reached, and the final data classification, cleaning and compression parameter combination is obtained, and the data is processed according to the final data classification, cleaning and compression parameter combination, and stored in the central data center to build a big data platform for new energy sites.

[0081] In an embodiment of the present invention, an objective function is set to evaluate the comprehensive index of data processing performance under different parameter combinations. A set of parameter combinations is randomly generated as the initial harmony library. Each harmony represents a data classification, cleaning and compression scheme, including the selection of classification algorithm, the setting of cleaning rules, compression algorithm and parameters, etc. A set of harmonies is randomly selected from the harmony library as the current harmony. A new harmony is generated by randomly adjusting some parameters in the current harmony. Part of the data is processed using the parameter combination corresponding to the new harmony, the objective function value is calculated, and its pros and cons are evaluated. If the objective function value of the new harmony is better than the worst harmony in the current harmony library, the new harmony is added to the harmony library and the worst harmony is replaced. The harmony selection, parameter adjustment, new harmony evaluation and harmony library update operations are repeated until the preset number of iterations is reached, and the final data classification, cleaning and compression parameter combination is determined from the harmony library by the objective function value of the harmony. All the collected data are processed using this set of final parameter combinations, including classification, cleaning and compression, and the processed data is stored in the central data center to build a big data platform for new energy stations.

[0082] By setting the objective function and using the harmony search algorithm, the final combination of data processing parameters can be automatically found, avoiding the tediousness and inefficiency of traditional manual debugging. In the iterative process, distributed computing or parallel processing technology can be used to simultaneously evaluate the performance of multiple harmonies, further accelerating the optimization process.

[0083] By continuously optimizing classification algorithms and parameters, the classification accuracy of data can be improved, ensuring that data is correctly classified, and providing a reliable basis for subsequent data analysis and application. The optimized cleaning rules can more effectively identify and handle problems such as outliers and missing values ​​in the data, improving the integrity and accuracy of the data.

[0084] Choosing appropriate compression algorithms and parameters can minimize the cost of data storage and transmission while ensuring data quality. The realization of the entire process can rely on an intelligent data management system to achieve automatic data collection, processing, storage and analysis, reduce manual intervention, and improve management efficiency. The constructed big data platform can continue to expand with the development of new energy stations to adapt to more types and scales of data processing needs. Based on the optimized data, more accurate data models and prediction algorithms can be constructed to provide intelligent decision-making support for the operation and management of new energy stations. Improve the accuracy and efficiency of data processing. Building an efficient and intelligent big data platform can improve the operational efficiency and management level of new energy stations.

[0085] When specifically applied, this includes:

[0086] Select a set of classification algorithms, such as decision trees. Determine a cleaning method, such as missing value handling. Select a compression algorithm, such as ZIP, and define its parameters, such as compression level. Randomly generate a set of parameter combinations to form an initial harmony library. Each harmony contains information such as the type of classification algorithm, data cleaning rules, compression algorithm and parameters (e.g., compression level). Set the size of the harmony library, such as 100 harmonies. Randomly select a set of harmonies from the harmony library as the current harmony. Make small random adjustments to some parameters in the current harmony to generate a new harmony. For example, you can randomly change the parameters of the classification algorithm, cleaning rules, or compression algorithm. Use the parameter combination of the new harmony to process a portion of the data, which includes data classification, cleaning, and compression. Calculate the objective function value from the processed data. The objective function can be a weighted combination of classification accuracy, processing time, resource consumption, etc. Compare the objective function value of the new harmony with the objective function value of the worst harmony in the current harmony library. If the value of the new harmony is better than the worst harmony, then add the new harmony to the harmony library and replace the worst harmony. Repeat until the preset number of iterations (e.g. 100) is reached. The final parameter combination is obtained by calculating the objective function values ​​of all harmonies. All collected data are classified, cleaned and compressed using the determined final parameter combination. The processed data is stored in the central data center to ensure that the data can support the construction of the big data platform of the new energy station.

[0087] In a preferred embodiment of the present invention, the calculation formula of the objective function value corresponding to the new harmony is:

[0088] ;

[0089] in, represents the objective function value corresponding to the new harmony; Represents the weight coefficient of classification accuracy; Indicates classification accuracy; The weight coefficient indicating the cleaning effect; Indicates the number of valid data after cleaning; Indicates the number of invalid data after cleaning; Indicates the number of erroneous data after cleaning; Indicates the number of original valid data; Indicates the number of original invalid data; Indicates the number of raw error data; Indicates the number of reference valid data; Indicates the number of reference invalid data; Indicates the number of reference error data; represents a positive number; represents the exponential parameter; The weight coefficient representing the compression ratio; Indicates the adjusted compression ratio.

[0090] In an embodiment of the present invention, is the objective function value corresponding to the new harmony to be calculated, It is the weight coefficient of classification accuracy, which determines the importance of classification accuracy in the objective function. It is the classification accuracy, which indicates the correctness of the classification, and its value range is between 0 and 1. is the weight coefficient of the cleaning effect, which determines the importance of the cleaning effect in the objective function. is the number of valid data after cleaning. is the number of invalid data after cleaning. is the number of erroneous data after cleaning. is the number of original valid data. is the number of original invalid data. is the amount of raw error data. It is the number of reference valid data, used to evaluate the cleaning effect. is the number of reference invalid data, used to evaluate the cleaning effect. is the number of reference error data, used to evaluate the cleaning effect. is a small positive number used to prevent the denominator from being zero. It is an index parameter used to adjust the calculation method of the cleaning effect to make it more in line with actual needs. is the weight coefficient of the compression rate, which determines the importance of the compression rate in the objective function. It is the adjusted compression ratio, which indicates the degree of data compression, and its value range is between 0 and 1.

[0091] First, we need to know the classification accuracy and its weight coefficient Then, and Multiply them together to get the contribution of classification accuracy to the objective function value. Calculate the proportion of valid data after cleaning, that is, Next, calculate the ratio of valid and invalid data in the original data, that is, . Calculate the proportion of valid data in the reference data, that is . Calculate the proportion of valid data in the original data (considering invalid data), that is Substitute the above ratio into the cleaning effect formula to get the cleaning effect value. Substitute the cleaning effect value with its weight coefficient and exponential parameters Combined with the above, the contribution of the cleaning effect to the objective function value is calculated. First, we need to know the adjusted compression rate and its weight coefficient .Will and Multiply them together to get the contribution of compression rate to the objective function value. Add the contribution of classification accuracy, cleaning effect and compression rate to get the total objective function value. .

[0092] By classification accuracy ( ), cleaning effect and compression rate ( ) three indicators to comprehensively evaluate the performance of the new harmony. The quality of data cleaning and compression efficiency are taken into account, thus providing a more comprehensive performance evaluation. Weight coefficient , and The introduction of allows users to flexibly adjust the importance of these three indicators in the objective function according to actual needs. For example, in some application scenarios, the classification accuracy may be more important, so you can increase In other scenarios, data cleaning effect or compression rate may be more critical, and you can adjust it accordingly. or The cleaning effect part reflects the quality of the cleaned data in detail through complex proportional calculations. By comparing the number of valid data, invalid data, and erroneous data after cleaning with the corresponding number of original data and reference data, the degree of improvement of the data by the cleaning process can be accurately evaluated. The introduction of exponential parameters further allows users to adjust the impact of cleaning effect in the objective function according to actual needs. The positive term effectively prevents the denominator from being zero, ensuring the stability and reliability of the calculation. This is especially important when dealing with real data, because there may be some special cases in the data set. As optimization goals, the algorithm can be guided to continuously optimize the classification accuracy, cleaning effect and compression rate during the training process, so as to find the final new harmony. The formula promotes the efficient use of data by comprehensively considering the three links of classification, cleaning and compression. Under limited resources, by optimizing these three aspects, the quality and value of data can be maximized.

[0093] In a preferred embodiment of the present invention, the above step 13, based on the constructed big data platform of the new energy station, constructs a digital twin model of the new energy station to reflect the operation status, equipment health status and environmental changes of the station in real time, which may include:

[0094] Step 133, extracting data from the big data platform of the new energy station, including equipment operation data, environmental monitoring data and equipment maintenance records;

[0095] Step 134, determining the structure and framework of the digital twin model based on the extracted data, including determining the hierarchical structure of the model, the relationship between components, and the direction of data flow;

[0096] Step 135, construct a digital twin model of the new energy station according to the structure and framework of the digital twin model, and map the real-time data in the big data platform to the digital twin model to reflect the operating status of the station, equipment health status and environmental changes in real time.

[0097] In an embodiment of the present invention, key data is extracted from the big data platform of the new energy site, including but not limited to equipment operation data (such as power output, temperature, vibration, etc.), environmental monitoring data (such as wind speed, wind direction, temperature, humidity, etc.) and equipment maintenance records (such as maintenance history, replacement parts records, etc.).

[0098] These data are obtained through the big data platform to ensure the real-time and accuracy of the data. The extracted data is cleaned to remove noise, outliers and duplicate data to ensure data quality. The data is formatted and standardized to meet the input requirements of the digital twin model. According to the actual situation of the new energy station, the hierarchical structure of the digital twin model is determined, including the equipment layer, system layer, station layer, etc. The equipment layer focuses on the operating status and performance of a single device; the system layer focuses on the collaborative work between devices and the overall performance of the system; the station layer focuses on the operating status and environmental impact of the entire station. Analyze the relationship between equipment, system and station, and determine the connection mode and data flow path between components in the digital twin model. Ensure that data can flow smoothly in the model, providing a basis for subsequent simulation, analysis and optimization. Construct a digital twin model of the new energy station according to the designed structure and framework. The digital twin model includes the three-dimensional model of the equipment, operation logic, control strategy, etc., to ensure that the model can accurately reflect the operating status of the actual station. Map the real-time data in the big data platform to the digital twin model to ensure that the model can reflect the operating status of the station, equipment health and environmental changes in real time. Through the data interface, data interaction and synchronous update between the big data platform and the digital twin model are realized. The accuracy and reliability of the digital twin model are evaluated by comparing and verifying with the data of the actual site. If the model is consistent with the behavior and performance of the actual site, the model is valid; otherwise, the model needs to be adjusted and optimized.

[0099] According to the verification results, the digital twin model is optimized by adjusting parameters, improving algorithms, and other optimization measures to improve the accuracy and prediction ability of the model. The structure and framework of the model are continuously optimized to make it more in line with the actual operation status and needs of the station.

[0100] The digital twin model can monitor the operating status and equipment health of new energy stations in real time, detect potential problems and failures in a timely manner, and issue early warning signals. This helps operators take timely measures to deal with them, avoid the occurrence and expansion of accidents, and ensure the safe and stable operation of the stations. Through the simulation and analysis of the digital twin model, the operation strategy and control parameters of the new energy station can be optimized. According to the results predicted by the model, the operation mode and power output of the equipment can be adjusted to improve the power generation efficiency and energy utilization of the station.

[0101] The digital twin model can predict the maintenance needs and replacement cycles of equipment, and formulate maintenance plans and spare parts procurement plans in advance. This helps reduce sudden failures and downtime, and reduces operation and maintenance costs and repair expenses. Through the simulation and analysis results of the model, we can more accurately understand the operating status and development trends of the station, providing more reliable data support for decision-making. The application of digital twin technology has promoted the innovative development of new energy stations.

[0102] When specifically applied, this includes:

[0103] Obtain key data from the big data platform of new energy stations, including equipment operation data (such as power output, temperature, etc.), environmental monitoring data (such as wind speed, humidity, etc.), and equipment maintenance records. Ensure the real-time and accuracy of the data, and provide a solid foundation for subsequent model construction. Clean the extracted data to remove noise, outliers, and duplicate data. Perform format conversion and standardization to make the data meet the input specifications of the digital twin model.

[0104] According to the actual situation of the station, the hierarchical structure of the design model includes equipment layer, system layer and station layer.

[0105] Equipment layer: Focus on the detailed operating status and performance of a single device. For each device, we build a 3D model and embed its unique operating logic and control strategy. For example, in a wind farm, a detailed 3D model is built for each wind turbine, including its blades, gearbox, generator and other key components, and its behavior in actual operation is simulated.

[0106] System layer: Consider the collaborative work between devices and the overall performance of the system. At this layer, individual device models are connected to form a system model to simulate the interaction between devices and the operation of the entire system. For example, in a wind farm, consider the arrangement of wind turbines, the impact of wind speed and direction on each generator, and the power generation efficiency and stability of the entire wind farm.

[0107] Site layer: Focus on the operation status and environmental impact of the entire site. At this layer, the system model is combined with the environmental model of the site to simulate the overall operation status of the site and the impact of the environment on the site performance. For example, consider the site's geographical location, climate conditions, topography and other factors, and how these factors affect the power generation efficiency and equipment life of the wind farm.

[0108] After the models at each level are built, they are integrated together and tested to ensure the accuracy and reliability of the models. This involves:

[0109] Ensure that real-time data in the big data platform can be correctly mapped to the digital twin model, and that real-time data transmission and synchronous updates are achieved through the data interface. Perform various functional tests on the model to ensure that it can correctly simulate the operating conditions of the actual site. This includes simulating scenarios such as normal operation, failure conditions, extreme weather conditions, and observing the response and behavior of the model. Based on the test results, make necessary adjustments and optimizations to the model to improve its accuracy and predictive ability. This may involve adjusting model parameters, improving algorithms, optimizing data structures, and other aspects.

[0110] Taking a wind farm as an example, the real-time power output, ambient temperature, and historical maintenance records of a wind turbine are extracted from the big data platform. A three-dimensional model is built for the wind turbine, and its operating logic and control strategy are embedded; at the same time, an overall layout model of the wind farm is built, considering the impact of environmental factors on power generation efficiency. Real-time wind speed, wind direction data, and cleaned power output data are mapped to the digital twin model to ensure that the model can reflect the operating status of the wind farm in real time. By comparing the wind turbine power output predicted by the model with the actual data, it is found that the model has low prediction accuracy in specific wind speed ranges. Subsequently, the model parameters are adjusted, such as increasing the consideration of specific wind speed ranges, to improve the prediction accuracy of the model in these ranges.

[0111] In a preferred embodiment of the present invention, the above step 14 analyzes the data of the digital twin model and predicts the energy output and demand trends and equipment failure conditions in the future by simulating different operation scenarios to obtain analysis results, which may include:

[0112] Step 144, integrating the real-time equipment operation data, environmental monitoring data, and equipment maintenance records obtained in the digital twin model to obtain a data set containing multi-dimensional information;

[0113] Step 145 , analyzing the multi-dimensional information data set to determine key features related to energy output, demand trends, and equipment failures, including equipment operating parameters, extreme values ​​of environmental factors, and frequency of equipment maintenance;

[0114] Step 146, using the time series analysis model as a prediction model, and using the key features and corresponding historical data as input data, training the prediction model to obtain a trained prediction model;

[0115] Step 147, defining multiple operating scenarios, including different weather conditions and energy demand changes, and using the defined multiple operating scenarios as input conditions to import the trained prediction model, including parameters of weather conditions and energy demand change patterns;

[0116] Step 148, simulating the energy output and demand trends under different scenarios through the prediction function of the prediction model, and predicting equipment failure conditions to obtain prediction results.

[0117] In an embodiment of the present invention, the equipment operation data (such as power output, temperature, vibration, etc.), environmental monitoring data (such as wind speed, wind direction, temperature, humidity, etc.) and equipment maintenance records (such as maintenance history, replacement parts records, etc.) acquired in real time in the digital twin model are integrated. A data set containing multi-dimensional information is formed, which comprehensively reflects the operating status, environmental factors and equipment health status of the new energy station. Using data analysis technology, the multi-dimensional information data set is deeply analyzed to determine the key features closely related to energy output, demand trends and equipment failures. These features may include specific operating parameters of the equipment (such as the fluctuation range of power output), extreme values ​​of environmental factors (such as maximum wind speed, maximum temperature, etc.), and the frequency and type of equipment maintenance. A set of key features are extracted, which will serve as important inputs to the prediction model. The time series analysis model is selected as the prediction model, and the extracted key features and corresponding historical data are used as input data to train the prediction model. By continuously adjusting the model parameters and optimizing the model structure, the prediction accuracy of the model is improved. Get a well-trained prediction model that can accurately predict energy output, demand trends, and equipment failures in the future based on historical data. Define multiple operation scenarios based on the actual operation of the new energy station and possible future changes. These scenarios should cover different weather conditions (such as sunny, cloudy, rainy, stormy, etc.), changes in energy demand (such as peak, trough, holidays, etc.), and possible equipment failure modes. Form a set of detailed operation scenario descriptions, including parameter settings for weather conditions, change patterns of energy demand, and probability distribution of equipment failures. Import the defined operation scenarios as input conditions into the trained prediction model. Use the model's prediction function to simulate energy output and demand trends in different scenarios and predict equipment failures. Get a series of prediction results, including energy output, demand, and the probability and type of equipment failures in different scenarios.

[0118] By integrating multi-dimensional information data sets and refining key features, the prediction model can more accurately reflect the actual operating conditions and future trends of new energy stations, thereby improving the accuracy and reliability of predictions. The prediction results provide a scientific basis for the energy management of new energy stations. Based on the prediction results, the stations can reasonably adjust the energy production plan, optimize energy distribution and scheduling, improve energy utilization efficiency, and reduce operating costs.

[0119] By predicting the conditions and types of equipment failures, the station can formulate maintenance plans and spare parts procurement plans in advance to avoid the adverse effects of equipment failures on production. At the same time, preventive maintenance can be performed on the equipment to extend the equipment life and improve equipment reliability. The prediction results can also provide support for the decision-making of new energy stations. Management can formulate more reasonable strategic planning, investment decisions and risk management measures based on the prediction results to promote the sustainable development of the station.

[0120] When specifically applied, this includes:

[0121] Integrate the equipment operation data (such as power output, temperature, vibration, etc.), environmental monitoring data (such as wind speed, wind direction, temperature, humidity, etc.) and equipment maintenance records (such as maintenance history, replacement parts records, etc.) obtained in real time in the digital twin model to form a data set containing multi-dimensional information. The Pearson correlation coefficient is selected as an indicator to measure the degree of linear correlation between features. The Pearson correlation coefficient ranges from -1 to 1. The closer the absolute value is to 1, the stronger the linear correlation between the two features is, and the closer it is to 0, the weaker the linear correlation is. For all features in the data set, calculate the Pearson correlation coefficient between them to form a correlation matrix. This matrix can help us intuitively see which features have strong correlations. By looking at the correlation matrix, we can identify features that are closely related to energy output, demand trends, and equipment failures. For example, if it is found that there is a high positive correlation between wind speed and energy output, and a high negative correlation between temperature and equipment failure rate, then these features may be key features that need attention.

[0122] Recursive Feature Elimination (RFE) was selected as the feature selection algorithm. RFE repeatedly builds the model, removing the least important features each time until the predetermined number of features is reached or the model performance is no longer significantly improved. In order to apply RFE, a base model needs to be built first. This model will be used to evaluate the importance of each feature. Using the RFE algorithm and the base model, the features in the dataset are recursively eliminated. In each iteration, the least important feature in the current model is removed, and the model is retrained to evaluate the importance of the remaining features. Based on the results of RFE, it can be determined which features have an impact on the forecasting target (energy output, demand trend, and equipment failure). These features are the key features that need to be extracted. The results of correlation analysis and RFE are comprehensively considered to extract a set of key features that are closely related to the forecasting target and are important in the model. These features should be able to fully reflect the relationship between the equipment's operating conditions (such as power output, speed, vibration, etc.), environmental factors (such as wind speed, wind direction, temperature, humidity, etc.), and equipment health (such as maintenance history, replacement parts record, etc.) and energy output, demand trend, and equipment failure.

[0123] For example, in wind power plants, data analysis shows that when the wind speed is within the range of 5-10m / s, the power output of wind turbines is most stable and the failure rate is low; when the wind speed exceeds 15m / s, the power output fluctuation increases and the equipment failure rate rises. At the same time, it is also found that after the equipment has been running continuously for more than 200 hours, the failure rate begins to gradually increase, especially under high temperature and large vibration conditions.

[0124] Select the time series analysis model as the prediction tool, use the extracted key features and corresponding historical data as input to train the model. Ensure the accuracy and stability of the prediction model by continuously adjusting the model parameters and optimizing the model structure. According to the actual operation of the new energy station and possible changes in the future.

[0125] For example, using the trained prediction model, it can be predicted that in the next week, the energy output of the wind farm will maintain a steady growth because the wind speed will remain in the range of 5-10m / s and the ambient temperature will be moderate. At the same time, because some equipment has been running continuously for more than 200 hours, the model predicts that these equipment may fail in the next few days and maintenance preparations need to be made in advance.

[0126] Define multiple operation scenarios based on the actual operation of new energy stations and possible future changes. For example, define different weather conditions such as sunny days, cloudy days, rainy days, and storms, as well as different energy demand change scenarios such as peak periods, trough periods, and holidays. Use the defined operation scenarios as input conditions and import them into the trained prediction model to simulate energy output and demand trends in different scenarios and predict equipment failures.

[0127] For example, in the defined operation scenario, it is assumed that there will be a storm in the next week, and the wind speed will increase sharply to more than 20m / s. Through simulation and prediction, the model predicts that during this storm, the energy output of the wind power station will drop significantly, and the equipment failure rate will increase significantly. Based on this prediction result, the station can prepare countermeasures in advance, such as reducing energy production plans, strengthening equipment inspections and maintenance preparations, etc.

[0128] In a preferred embodiment of the present invention, the calculation formula of the prediction function is:

[0129] ;

[0130] in, Represents the predicted output value; represents the predicted energy output; represents the predicted energy demand; represents the predicted frequency of equipment failure; Indicates the weight coefficient of equipment operation data; Indicates equipment operation data; Represents the weight coefficient of time series data; Time series data representing historical energy production and demand records; Represents the weight coefficient of environmental data; Represents environmental data, defined as: ; Indicates the ambient temperature; Indicates humidity; Indicates wind speed; Indicates ambient air pressure; Represents the bias term of the prediction model; represents the number of interaction terms; represents the coefficient of the interaction effect; A coefficient indicating the effect of compression rate; represents a positive number; Represents the index of the interaction term.

[0131] In an embodiment of the present invention, the device operation data is collected . Collect time series data of historical energy production and demand records . Collect environmental data , including ambient temperature ,humidity , wind speed , Ambient air pressure . Determine the forecasted energy output , predicted energy demand and predicted equipment failure frequency . Calculate the weighted value of equipment operation data: . Calculate weighted values ​​for time series data: . Calculate the weighted value of the environmental data: For each interaction term (from 1 to n) Calculate interaction effects × ,X× and × Calculates the weighted sum of the interaction effects . Add the weighted sums of all interaction terms. Calculate The value of the logarithmic result and the coefficient affecting the compression rate Multiply them together to get the contribution of the compressibility effect. Add the sum of the base term, the interaction term, and the compressibility effect, and then add the bias term , and get the final predicted output value .

[0132] The prediction function takes into account the equipment operation data , time series data , Environmental data and their interactions can more comprehensively reflect the actual situation of energy output, demand, and equipment failure frequency, thereby improving the accuracy of prediction. , can capture the mutual influence between different factors and further improve the precision of prediction. The weight coefficient is , , , interaction effect coefficient , compression rate influence coefficient , positive number These are all adjustable parameters and can be adjusted according to actual data and requirements. Environmental data , in order to more accurately reflect the impact of environmental factors on the prediction results. Accurate prediction results provide strong support for energy management, equipment scheduling, fault prevention and other decision-making, helping to reduce the uncertainty of decision-making. , can provide early warning for equipment maintenance and fault prevention, and help reduce the risks and losses caused by equipment failure.

[0133] In a preferred embodiment of the present invention, the above step 15 automatically generates an operation optimization plan based on the analysis results and real-time feedback of the digital twin model to guide the remote start and stop and parameter adjustment of the wind turbine, photovoltaic panel array, and energy storage system, which may include:

[0134] Step 151, encoding the start and stop of the wind turbine, photovoltaic panel, and energy storage system and the adjustment of equipment parameters into chromosomes;

[0135] Step 152, randomly generating a set of initial solutions, each solution representing an operation optimization scheme;

[0136] Step 153, using the analysis results and real-time feedback of the digital twin model, evaluate the fitness of each solution, and iteratively update the population through selection, crossover and mutation operations to obtain the final solution;

[0137] Step 154, continue iterating, and when the preset number of iterations is reached, determine the initial operation optimization plan from the final population according to the fitness of the solution;

[0138] Step 155, parsing the initial operation optimization plan into control instructions and parameter adjustment values;

[0139] Step 156, remotely start and stop the device according to the control instruction, and remotely adjust the operating parameters of the device according to the parameter adjustment value.

[0140] In an embodiment of the present invention, the start and stop states of wind turbines, photovoltaic panels, and energy storage systems and the adjustment of equipment parameters (such as wind speed, light intensity, energy storage charge and discharge rate, etc.) are encoded as chromosomes. Chromosomes can be encoded in binary or real numbers. A set of initial solutions is randomly generated, each of which represents an operation optimization scheme. These solutions constitute the initial population as the starting point of the genetic algorithm. Based on the analysis results of historical data, weather forecasts, equipment performance models, etc., the potential effect of each solution (i.e., each operation optimization scheme) is evaluated. The operation of the actual system is simulated by a digital twin model to obtain the performance of each solution in the actual system, including real-time feedback such as energy output, consumption, and equipment status. The fitness of each solution is calculated based on the evaluation results and feedback data. Fitness can be defined as the weighted sum of indicators such as energy efficiency, equipment utilization, and system stability. According to the fitness, an excellent solution is selected as the parent generation for generating the next generation. The selection operation adopts roulette selection. The selected parent generation solution is crossover operated to generate a new child generation solution. The crossover operation can adopt methods such as single-point crossover, multi-point crossover, or uniform crossover. The newly generated offspring solutions are mutated to increase the diversity of the population. The mutation operation can randomly change the values ​​of certain genes in the chromosome.

[0141] Repeat fitness evaluation, selection, and genetic operations to iteratively update the population. Each iteration generates a new solution and eliminates solutions with lower fitness. When the preset number of iterations is reached or other stop conditions are met, the final solution is selected from the final population according to fitness as the initial operation optimization plan. The initial operation optimization plan is parsed into specific control instructions and parameter adjustment values. For example, the start and stop state of the wind turbine is converted into a start or stop instruction, and the tilt angle of the photovoltaic panel is adjusted to the optimal value. According to the control instructions, the wind turbine, photovoltaic panel, and energy storage system are started and stopped through the remote communication system. At the same time, the operating parameters of the equipment are remotely adjusted according to the parameter adjustment values.

[0142] By optimizing the operation plan through genetic algorithms, the potential of wind turbines, photovoltaic panels and energy storage systems can be fully utilized to improve energy output and utilization. Real-time feedback and adjustment mechanisms can ensure that the system always operates in the optimal state and reduce energy waste. Genetic algorithms can consider the impact of various factors (such as weather changes, equipment performance, etc.) on system operation to generate optimization plans. Real-time feedback mechanisms can detect and deal with potential problems in a timely manner to prevent system failures or crashes. Through remote control and parameter adjustment, the number and frequency of on-site operation and maintenance personnel can be reduced, reducing operation and maintenance costs. The generated optimization plan can provide a scientific basis to help them make more informed decisions.

[0143] When specifically applied, this includes:

[0144] Collect historical data, including the operation data of wind turbines, photovoltaic panels, and energy storage systems, as well as the corresponding weather conditions, energy output, and consumption. Obtain future weather forecast data, including wind speed, light intensity, etc., to predict future energy output and consumption. According to the technical parameters and historical operation data of the equipment, establish an equipment performance model to evaluate the performance of the equipment under different operation schemes. Encode the start and stop status and equipment parameters (such as wind speed threshold, light intensity threshold, energy storage charge and discharge rate, etc.) of wind turbines, photovoltaic panels, and energy storage systems into chromosomes. For example, binary coding is used to represent the start and stop status (0 means stop, 1 means start), and real number coding is used to represent equipment parameters. Randomly generate a set of initial solutions, each of which represents an operation optimization scheme. These solutions constitute the initial population as the starting point of the genetic algorithm. Based on historical data, weather forecasts, and equipment performance models, evaluate the potential effect of each solution (i.e., each operation optimization scheme). This includes the weighted sum of indicators such as energy efficiency, equipment utilization, and system stability. The digital twin model is used to simulate the operation of the actual system and obtain the performance of each solution in the actual system, including real-time feedback on energy output, consumption, equipment status, etc. This helps to more accurately evaluate the actual effect of each solution. The roulette selection method is used to select excellent solutions as parents according to fitness to generate the next generation. The selected parent solutions are crossovered to generate new child solutions. The newly generated child solutions are mutated to increase the diversity of the population. Some gene values ​​in the chromosome can be randomly changed, such as adjusting equipment parameters or changing the start-stop state. Fitness evaluation, selection and genetic operations are repeated to iteratively update the population. New solutions are generated in each iteration, and solutions with lower fitness are eliminated. When the preset number of iterations is reached or other stop conditions are met, the final solution is selected from the final population according to fitness as the initial operation optimization plan. The initial operation optimization plan is parsed into specific control instructions and parameter adjustment values. For example, the start-stop state of the wind turbine is converted into a start or stop instruction, the tilt angle of the photovoltaic panel or the tracking system is adjusted to the optimal value, and the charge and discharge rate of the energy storage system is adjusted to the final strategy. According to the control instructions and parameter adjustment values, the wind turbines, photovoltaic panels and energy storage systems are remotely controlled and parameter adjustments are made through the remote communication system.

[0145] Suppose there is a hybrid energy system that includes wind turbines, photovoltaic panels, and energy storage systems. It is hoped that the operation of this system can be optimized to maximize energy efficiency and reduce equipment failures. Encode the wind turbines, photovoltaic panels, and energy storage systems separately. For example, the start and stop status of the wind turbine is represented by binary coding (0 or 1), and the wind speed threshold is represented by real number coding (such as a value between 5m / s and 15m / s). The start and stop status of the photovoltaic panel is also represented by binary coding, and the light intensity threshold is represented by real number coding (such as a value between 100W / m² and 1000W / m²). The charge and discharge rate of the energy storage system is also represented by real number coding. A set of initial solutions is randomly generated, each of which contains the start and stop status and equipment parameters of the wind turbine, photovoltaic panel, and energy storage system. Based on historical data and weather forecasts, the energy efficiency, equipment utilization, and system stability of each solution are evaluated. For example, for a certain solution, we simulate its operation in the next week and calculate the total energy output, consumption, and number of equipment failures. Use the digital twin model to more accurately simulate the performance of each solution in the actual system. This can include simulating the output power of wind turbines at different wind speeds, the power generation efficiency of photovoltaic panels at different light intensities, and the charging and discharging behavior of energy storage systems. The population is updated iteratively through selection, crossover, and mutation operations. For example, we select solutions with higher fitness as parents, generate new child solutions through single-point crossover, and randomly change some gene values ​​to increase diversity. After multiple iterations, the final operation optimization plan is determined from the final population based on the fitness of the solution.

[0146] In a preferred embodiment of the present invention, the calculation formula for the fitness of each solution is:

[0147] ;

[0148] in, Indicates The fitness value of each solution; , , , represents the weight coefficient; Indicated in Understand the actual energy output of new energy stations; Indicates the maximum energy output; Indicated in Understand the predicted energy demand; Indicates the actual energy demand; Indicates the maximum demand difference; Indicated in Understand the equipment failure rate; Indicated in Let’s understand the running costs.

[0149] In an embodiment of the present invention, It is in Let’s take a look at the actual energy output of new energy stations. is the maximum energy output (determined by equipment specifications). It is in The energy demand is predicted based on weather forecasts and historical demand patterns. is the actual energy demand (real-time data). is the maximum demand difference (a preset value used to normalize demand differences). It is in Understand the equipment failure rate (based on equipment performance and operating history). It is in Understand the operating costs (including maintenance, energy consumption, etc.). , , , These weight coefficients are used to balance the importance of different indicators in fitness calculation and need to be set according to actual needs.

[0150] A set of initial solutions is randomly generated, each of which contains the start / stop status and equipment parameters of wind turbines, photovoltaic panels and energy storage systems. The diversity of the initial population is ensured to cover the possible solution space. For each solution , calculate the fitness by following the steps below :

[0151] Calculate the energy output ratio , which represents the ratio of actual energy output to maximum energy output. Calculate the demand difference ratio , represents the difference between the predicted energy demand and the actual energy demand, normalized to the maximum demand difference. Equipment failure rate As part of the fitness calculation. Running cost As part of the fitness calculation, the energy output ratio, demand difference ratio, equipment failure rate, and operating cost are combined according to the weight coefficient to calculate the fitness. .

[0152] The optimization algorithm continuously searches for the energy that can maximize the energy output. near The solution can improve the energy conversion efficiency of new energy stations, thereby increasing the total energy output. The project aims to reduce the difference between predicted energy demand and actual energy demand. Through optimization, the system can more accurately predict and meet energy demand, reduce energy waste or shortage caused by mismatch between supply and demand, and improve the flexibility and reliability of energy utilization. The optimization algorithm will tend to select solutions that can reduce the equipment failure rate, thereby reducing the downtime and maintenance costs caused by equipment failure and improving the overall stability and availability of the system. The project focuses on the control of operating costs. Through the optimization algorithm, we can find a solution with low operating costs while ensuring energy output and supply and demand balance. This helps to reduce the operating costs of new energy stations and improve their economy and competitiveness. The formula combines multiple factors such as energy output, supply and demand balance, equipment failure rate and operating costs, and uses weight coefficients to calculate the cost of the equipment. , , , The adjustment can achieve the trade-offs and choices of different optimization objectives. This comprehensive optimization method can more comprehensively reflect the actual operation status of new energy stations and provide decision makers with more accurate and comprehensive optimization suggestions.

[0153] In a preferred embodiment of the present invention, the above step 16, applying the operation optimization scheme to the actual operation, and monitoring the actual operation effect in real time to obtain the operation data; the above step 17, comparing and analyzing the operation data with the prediction of the digital twin model, continuously adjusting the digital twin model and optimizing the operation scheme to form a closed-loop management, may include:

[0154] Step 171, comparing and analyzing the actual operation data with the predicted data of the digital twin model to determine the differences and deviations;

[0155] Step 172, analyzing the differences and deviations to obtain analysis results;

[0156] Step 173, according to the analysis results, the digital twin model is adjusted, including correcting the digital twin model parameters, optimizing the digital twin model structure, and optimizing the current operation plan to obtain an optimized and adjusted operation plan;

[0157] Step 174, remotely control and adjust the parameters of the wind turbine, photovoltaic panel array and energy storage system equipment according to the optimized and adjusted operation plan, monitor the operation status and performance parameters of the equipment in real time, and feed back the real-time operation data to the digital twin model, continuously iterate and form a closed-loop management.

[0158] In an embodiment of the present invention, sensors are deployed at key nodes of wind turbines, photovoltaic arrays, energy storage systems, and power grids to collect key operating data such as voltage, current, power, temperature, wind speed, and light intensity in real time. The collected data is transmitted to a data center or cloud storage platform by wired or wireless means to ensure the integrity and security of the data. The original data is preprocessed by cleaning, denoising, calibration, and other operations to improve the quality and reliability of the data. The initial version of the digital twin model is constructed based on historical operating data, equipment parameters, environmental factors, etc., including physical models, behavioral models, and state models. The model is run on the digital twin platform, and simulation calculations are performed according to the current environmental conditions (such as wind speed, light intensity) and the set operating strategy to obtain predicted data. Ensure that the actual operating data and the predicted data of the digital twin model are synchronized in time. Calculate the difference and deviation between the actual operating data and the predicted data, including indicators such as absolute value difference, relative error, and mean square error. The comparison between the actual data and the predicted data is intuitively displayed in the form of charts, curves, etc., which is convenient for analysts to quickly identify problems. Conduct in-depth analysis of differences and deviations to identify the root causes of the differences, such as equipment aging, environmental changes, model inaccuracies, etc. Evaluate the impact of differences and deviations on system performance, energy efficiency, safety, etc. Based on the analysis results, modify the parameters of the digital twin model, such as adjusting equipment efficiency, changing environmental parameters, etc.

[0159] Optimize the structure of the digital twin model, such as improving the algorithm, to improve the accuracy and reliability of the model. Verify and test the adjusted digital twin model to ensure the effectiveness of the model. Based on the prediction results and difference analysis of the digital twin model, optimize and adjust the current operation strategy, such as adjusting the start and stop time of wind turbines and photovoltaic panels, and the charging and discharging strategy of the energy storage system. Formulate an optimized operation plan, including specific control instructions, parameter adjustment values, etc. Send the optimized operation plan to the wind turbine, photovoltaic panel array and energy storage system equipment through the remote communication system. The equipment adjusts parameters and controls operations according to the received instructions. Monitor the operating status and performance parameters of the equipment in real time, such as power output, voltage, current, temperature, etc. Feedback real-time operation data to the digital twin model for further verification and optimization of the model. Continuously adjust the digital twin model and optimize the operation plan to form a closed-loop management and continuously optimize the operation plan and digital twin model.

[0160] By real-time monitoring and comparing actual operation data with the predicted data of the digital twin model, the operation strategy can be discovered and adjusted in time to improve the operation efficiency and energy utilization of the system. Optimizing the operation plan can reduce unnecessary energy consumption and loss, reduce operation and maintenance costs, and improve economic benefits. Through the prediction and real-time monitoring of the digital twin model, potential failures and safety hazards of equipment can be discovered in time, maintenance and repair can be carried out in advance, and the reliability and safety of the system can be enhanced.

[0161] When specifically applied, this includes:

[0162] Deploy wind speed sensors, temperature sensors, vibration sensors, etc. on wind turbines to monitor key parameters such as wind speed, unit temperature, and vibration in real time. Deploy light intensity sensors, temperature sensors, etc. in photovoltaic panel arrays to monitor light intensity, panel surface temperature, etc. Deploy voltage sensors, current sensors, temperature sensors, etc. in energy storage systems to monitor the charge and discharge status and temperature of energy storage equipment. Deploy voltage sensors, current sensors, etc. at key nodes of the power grid to monitor electrical parameters such as voltage and current of the power grid.

[0163] For example, in a wind farm, wind speed sensors measure wind speed in real time and transmit the data to a data center via wired or wireless means.

[0164] Use wired (such as optical fiber, Ethernet) or wireless (such as 4G / 5G, Wi-Fi) communication methods to transmit the collected data to the data center or cloud storage platform. On the data center or cloud storage platform, establish a database system to store and manage the collected data.

[0165] For example, the data acquisition system of a photovoltaic power station uploads data such as light intensity and temperature to a cloud storage platform in real time through a 4G network, ensuring the real-time and security of the data.

[0166] Based on historical operating data, equipment parameters, environmental factors, etc., the initial version of the digital twin model is constructed, including a physical model (describing the physical characteristics of the equipment), a behavioral model (describing the behavioral logic of the equipment), and a state model (describing the state changes of the equipment). On the digital twin platform, simulation scenarios are set according to current or predicted environmental conditions (such as wind speed and light intensity). These environmental conditions can be obtained through real-time data interfaces or generated through historical data or simulation data. According to the operating objectives and constraints of the equipment, the operating strategy in the simulation calculation is set. For example, wind turbines may need to adjust the start and stop and power output according to wind speed changes; photovoltaic panels may need to adjust the tilt angle and power regulation according to light intensity and temperature; energy storage systems may need to adjust the charging and discharging strategy according to grid demand and battery status. Run the constructed model on the digital twin platform and perform simulation calculations according to the set environmental conditions and operating strategies. During the simulation process, the operating status, performance parameters, and output data of the equipment can be monitored and recorded in real time. The predicted data of the equipment is obtained through simulation calculations, including but not limited to power output, temperature distribution, and failure probability. These data can be used for equipment operation optimization, failure prediction, and decision support.

[0167] For example, based on historical wind speed data and wind turbine parameters, a digital twin model of a wind turbine is constructed. Through simulation calculations, the output power of the wind turbine at different wind speeds is predicted. Ensure that the actual operating data and the predicted data of the digital twin model are synchronized in time. Calculate the difference and deviation between the actual operating data and the predicted data, including relative error indicators. Visually display the comparison between the actual data and the predicted data in the form of charts, curves, etc. The absolute value difference ;in, Indicates the value of actual running data; is the predicted output value.

[0168] For example, by comparing the actual wind turbine output power with the predicted power of the digital twin model, it is found that the actual power is lower than the predicted power in a certain period of time, and the difference is displayed in a chart. In-depth analysis of differences and deviations is carried out to find out the root causes of the differences, such as equipment aging, environmental changes, model inaccuracies, etc. Evaluate the impact of differences and deviations on system performance, energy efficiency, safety, etc.

[0169] For example, through analysis, it was found that the reason why the actual power of wind turbines is lower than the predicted power is that the wind energy capture efficiency decreases due to blade aging. Further evaluation found that this will lead to a decrease in the overall energy efficiency of the wind farm. According to the analysis results, the parameters of the digital twin model are corrected, such as adjusting the equipment efficiency and changing the environmental parameters. The corrected model is simulated again to verify the correction effect. Based on the analysis results of blade aging, the efficiency parameters of the wind turbines in the digital twin model are adjusted. Through re-simulation calculations, it is found that the predicted power of the corrected model is closer to the actual power.

[0170] like Figure 2 As shown, an embodiment of the present invention further provides a new energy station operation intelligent management system 20, comprising:

[0171] The acquisition module 21 is used to collect various data in real time through sensors and monitoring equipment in the new energy station;

[0172] The collected data is transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a big data platform for new energy stations. Based on the constructed big data platform for new energy stations, a digital twin model of the new energy station is constructed to reflect the operating status of the station, equipment health status and environmental changes in real time.

[0173] The processing module 22 is used to analyze the data of the digital twin model, and by simulating different operating scenarios, predict the energy output and demand trends and equipment failure conditions in the future to obtain analysis results; based on the analysis results and the real-time feedback of the digital twin model, automatically generate an operation optimization plan to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic panel arrays, and energy storage systems; apply the operation optimization plan to actual operation, and monitor the actual operation results in real time to obtain operation data; compare and analyze the operation data with the predictions of the digital twin model, continuously adjust the digital twin model and optimize the operation plan to form a closed-loop management.

[0174] It should be noted that the system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0175] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A new energy station operation intelligent management method, characterized in that: The method comprises: Collect various data in real time through sensors and monitoring equipment in new energy stations; The collected data is transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a new energy station big data platform; Based on the big data platform of the new energy station, a digital twin model of the new energy station is constructed to reflect the operation status, equipment health status and environmental changes of the station in real time; Analyze the data of the digital twin model and predict the energy output and demand trends and equipment failures in the future by simulating different operating scenarios to obtain analysis results; Based on the analysis results and real-time feedback from the digital twin model, an operation optimization plan is automatically generated to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic arrays, and energy storage systems; Apply the operation optimization plan to actual operation, monitor the actual operation effect in real time, and obtain operation data; Compare and analyze the operation data with the predictions of the digital twin model, continuously adjust the digital twin model and optimize the operation plan to form a closed-loop management; The collected data is transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a new energy station big data platform, including: Set an objective function to evaluate the data processing performance of different parameter combinations, including classification accuracy, cleaning effect, and compression rate; A set of parameter combinations are randomly generated as the initial harmony library, each harmony represents a data classification, cleaning and compression scheme, A group of harmonies is randomly determined from the harmony library as the current harmony, and a new harmony is generated by randomly adjusting some parameters in the current harmony; Calculate the objective function value corresponding to the new harmony, evaluate its quality, and obtain the evaluation result; Continuously perform harmony selection, parameter adjustment, new harmony evaluation and harmony library update operations until the preset number of iterations is reached to obtain the final data classification, cleaning and compression parameter combination, and process the data according to the final data classification, cleaning and compression parameter combination, and store it in the central data center to build a big data platform for new energy stations; The calculation formula of the objective function value corresponding to the new harmony is: ； in, represents the objective function value corresponding to the new harmony; Represents the weight coefficient of classification accuracy; Indicates classification accuracy; The weight coefficient indicating the cleaning effect; Indicates the number of valid data after cleaning; Indicates the number of invalid data after cleaning; Indicates the number of erroneous data after cleaning; Indicates the number of original valid data; Indicates the number of original invalid data; Indicates the number of raw error data; Indicates the number of reference valid data; Indicates the number of reference invalid data; Indicates the number of reference error data; represents a positive number; represents the exponential parameter; The weight coefficient representing the compression ratio; Indicates the adjusted compression ratio.

2. The intelligent management method for new energy station operation according to claim 1 is characterized in that: Based on the big data platform of the new energy station, a digital twin model of the new energy station is constructed to reflect the operation status, equipment health status and environmental changes of the station in real time, including: Extract data from the big data platform of new energy stations, including equipment operation data, environmental monitoring data and equipment maintenance records; Based on the extracted data, determine the structure and framework of the digital twin model, including determining the model's hierarchy, the relationships between components, and the direction of data flow; A digital twin model of the new energy station is constructed based on the structure and framework of the digital twin model, and the real-time data in the big data platform is mapped to the digital twin model to reflect the operating status, equipment health status and environmental changes of the station in real time.

3. The intelligent management method for new energy station operation according to claim 2 is characterized in that: Analyze the data of the digital twin model and predict the energy output and demand trends and equipment failures in the future by simulating different operating scenarios. The analysis results include: Integrate the real-time equipment operation data, environmental monitoring data and equipment maintenance records obtained in the digital twin model to obtain a data set containing multi-dimensional information; Analyze multi-dimensional information data sets to identify key characteristics related to energy output, demand trends, and equipment failures, including equipment operating parameters, extreme values ​​of environmental factors, and frequency of equipment maintenance; The time series analysis model is used as a prediction model, and the key features and corresponding historical data are used as input data to train the prediction model to obtain a trained prediction model; Define multiple operating scenarios, including different weather conditions and energy demand changes, and use the defined multiple operating scenarios as input conditions to import the trained prediction model, including parameters of weather conditions and energy demand change patterns; The prediction function of the prediction model is used to simulate the energy output and demand trends in different scenarios, and predict equipment failures to obtain prediction results.

4. The intelligent management method for new energy station operation according to claim 3 is characterized in that: The calculation formula of the prediction function is: ； in, Represents the predicted output value; represents the predicted energy output; represents the predicted energy demand; represents the predicted frequency of equipment failure; Indicates the weight coefficient of equipment operation data; Indicates equipment operation data; Represents the weight coefficient of time series data; Time series data representing historical energy production and demand records; Represents the weight coefficient of environmental data; Represents environmental data; defined as: ; Indicates the ambient temperature; Indicates humidity; Indicates wind speed; Indicates ambient air pressure; Represents the bias term of the prediction model; represents the number of interaction terms; represents the coefficient of the interaction effect; A coefficient indicating the effect of compression rate; represents a positive number; Represents the index of the interaction term.

5. The intelligent management method for new energy station operation according to claim 4 is characterized in that: Based on the analysis results and real-time feedback from the digital twin model, an operation optimization plan is automatically generated to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic arrays, and energy storage systems, including: Encode the start and stop of wind turbines, photovoltaic panels, and energy storage systems and the adjustment of equipment parameters into chromosomes; A set of initial solutions is randomly generated, each of which represents a running optimization plan; Using the analysis results and real-time feedback from the digital twin model, the fitness of each solution is evaluated, and the population is iteratively updated through selection, crossover, and mutation operations to obtain the final solution; Continue to iterate. When the preset number of iterations is reached, the initial operation optimization plan is determined from the final population according to the fitness of the solution. Parsing the initial operation optimization plan into control instructions and parameter adjustment values; According to the control instructions, the equipment is remotely started and stopped, and according to the parameter adjustment values, the operating parameters of the equipment are remotely adjusted.

6. The intelligent management method for new energy station operation according to claim 5 is characterized in that: The fitness of each solution is calculated as: ； in, Indicates The fitness value of each solution; represents the weight coefficient; Indicated in Understand the actual energy output of new energy stations; Indicates the maximum energy output; Indicated in Understand the predicted energy demand; Indicates the actual energy demand; Indicates the maximum demand difference; Indicated in Understand the equipment failure rate; Indicated in Let’s understand the running costs.

7. The intelligent management method for new energy station operation according to claim 6 is characterized in that: Compare and analyze the operation data with the predictions of the digital twin model, continuously adjust the digital twin model and optimize the operation plan to form a closed-loop management, including: Compare and analyze the actual operation data with the predicted data of the digital twin model to identify differences and deviations; Analyze the differences and deviations and obtain analysis results; According to the analysis results, the digital twin model is adjusted, including correcting the digital twin model parameters, optimizing the digital twin model structure, and optimizing the current operation plan to obtain the optimized and adjusted operation plan; According to the optimized and adjusted operation plan, the wind turbines, photovoltaic panel arrays and energy storage system equipment are remotely controlled and parameter adjusted, and the operating status and performance parameters of the equipment are monitored in real time. The real-time operation data is fed back to the digital twin model, and the cycle is continuously iterated to form a closed-loop management.

8. A new energy station operation intelligent management system, the system implements the method as described in any one of claims 1 to 7, characterized in that: include: The acquisition module is used to collect various data in real time through sensors and monitoring equipment in new energy stations; The collected data is transmitted to the central data center for iterative optimization and processing to determine the final combination of data classification, cleaning and compression parameters to build a big data platform for new energy stations. Based on the constructed big data platform for new energy stations, a digital twin model of the new energy station is constructed to reflect the operating status of the station, equipment health status and environmental changes in real time. The processing module is used to analyze the data of the digital twin model and predict the energy output and demand trends and equipment failures in the future by simulating different operating scenarios to obtain analysis results; based on the analysis results and real-time feedback from the digital twin model, an operation optimization plan is automatically generated to guide the remote start and stop and parameter adjustment of wind turbines, photovoltaic arrays, and energy storage systems; Apply the operation optimization plan to actual operation, monitor the actual operation effect in real time, and obtain operation data; compare and analyze the operation data with the prediction of the digital twin model, continuously adjust the digital twin model and optimize the operation plan to form a closed-loop management.

Images