Power load power demand data collection method based on flywheel energy storage system

Abstract

The application relates to the technical field of data collection and discloses a computing power load power demand data collection method based on a flywheel energy storage system. The method first acquires flywheel energy storage system related data, classifies the data, constructs a database, and performs dimension reduction processing; then extracts a load fluctuation mode, classifies computing power load, generates power demand parameter characteristics, and automatically matches and collects parameter configurations; then generates a collection instruction according to parameter range adjustment of power grid dispatching requirements; then updates the database, adjusts a load prediction model, judges the applicability of a new load mode, and expands a configuration list; finally, a visual monitoring interface is constructed, data is stored in an encrypted manner, and permission control is set. The method can accurately collect computing power load power demand data, adapt to dynamic changes of load, meet power grid dispatching requirements, improve the intelligence, safety and efficiency of data collection, and help optimize operation of a power system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data acquisition, in particular to a computing power load power demand data acquisition method based on a flywheel energy storage system. BACKGROUND

[0002] With the acceleration of digitalization, the computing power infrastructure plays an increasingly important role in modern society. The scale of computing power facilities such as data centers and cloud computing platforms continues to expand, and their power consumption continues to rise, becoming an important load in the power system. Accurate acquisition and mastery of computing power load power demand data is crucial for the rational planning, scheduling and optimized operation of the power system.

[0003] Traditional data acquisition methods have many limitations in dealing with computing power load power demand. Computing power load has significant dynamic variation characteristics. Its power demand is affected by factors such as business volume fluctuations and user usage habits, showing complex variation patterns. For example, during promotional activities on Internet e-commerce platforms, data processing volume increases explosively, computing power load increases instantaneously, and power demand also fluctuates significantly. Traditional data acquisition methods are difficult to capture this rapid and complex change in real time and accurately, resulting in data that cannot truly reflect the actual power demand of computing power load.

[0004] In terms of combination with flywheel energy storage systems, traditional methods have deficiencies. Flywheel energy storage systems, as a new type of energy storage technology, have the advantages of fast response speed and high energy density, and can be used in power systems to regulate load fluctuations and improve power quality. However, traditional data acquisition methods do not fully consider the characteristics of flywheel energy storage systems, and cannot effectively match the acquisition parameters with the operating state of the flywheel energy storage system. For example, during the charging and discharging process of the flywheel energy storage system, its speed, energy storage capacity and other parameters will change, and traditional acquisition methods cannot dynamically adjust the acquisition parameters according to these changes, affecting the accuracy and effectiveness of the acquired data.

[0005] Power grid scheduling has increasingly high requirements for computing power load power demand data. The power grid needs to accurately predict and schedule loads based on this data to ensure the safe and stable operation of the power system. However, traditional data acquisition methods cannot meet the strict requirements of power grid scheduling in terms of data accuracy, real-time performance and completeness in different scenarios. During peak electricity consumption, the power grid needs to quickly acquire accurate computing power load power demand data to reasonably allocate power resources and avoid power shortages or overloads. However, the data acquired by traditional methods often has delays and large errors, making it difficult to meet the actual needs of power grid scheduling.

[0006] With the continuous expansion of the scale and the increase of the complexity of the power system, the amount of data grows exponentially. Traditional data acquisition methods also face challenges in data processing and storage. They lack effective data dimensionality reduction, classification and feature extraction means, resulting in high data storage and transmission costs, and it is difficult to quickly extract valuable information from massive data, affecting the overall operation efficiency of the power system and the scientificity of decision-making. SUMMARY

[0007] The purpose of the present application is to provide a computing power load power demand data acquisition method based on a flywheel energy storage system to solve the problems raised in the background art.

[0008] To achieve the above-mentioned purpose, the present application provides the following technical solution: a computing power load power demand data acquisition method based on a flywheel energy storage system, the method comprising:

[0009] Obtaining real-time speed, energy storage capacity and power output data of the flywheel energy storage system, classifying power demand data, constructing a power feature database, and performing data dimensionality reduction processing;

[0010] According to the power feature database, extract the load fluctuation mode, classify the computing power load to be collected; according to the dynamic change characteristics of the computing power load, generate the power demand parameters under each load classification; according to the use preference and configuration condition of the system operator, automatically match the adaptive collection parameter configuration for the real-time demand of different loads;

[0011] According to different power grid dispatching requirements, dynamically adjust the parameter range in the configuration list to generate collection instructions that accurately describe the power demand;

[0012] Update the power feature database, adjust the corresponding load prediction model according to the database update result, and judge whether the new load mode exceeds the applicable range of the current collection parameters;

[0013] According to the new load mode, expand the collection parameter configuration list and apply it to real-time data collection.

[0014] Preferably, the method for obtaining real-time speed, energy storage capacity and power output data of the flywheel energy storage system, classifying power demand data, constructing a power feature database, and performing data dimensionality reduction processing comprises:

[0015] A plurality of types of sensor networks are deployed to collect flywheel rotor speed, temperature, and energy storage unit power change data; real-time power output power and load demand curves are obtained through a power grid scheduling platform; historical operation logs and fault records are integrated to verify the completeness and consistency of the data; power demand data is divided into base load, peak load, and fluctuating load according to load type, and a timestamp label is added to each type of data; a distributed storage system is used to establish a table structure of a power feature database to store speed, capacity, power, and load classification information; based on a principal component analysis method, the power demand data is converted into a low-dimensional feature vector; the reduced features include load fluctuation amplitude, duration, periodicity, and associated equipment state;

[0016] The principal component analysis method based on the power demand data is converted into a low-dimensional feature vector, specifically including:

[0017] The original power data is normalized to eliminate dimensional differences; local features of time series data are extracted through a sliding window mechanism to construct a multi-dimensional feature matrix; the covariance matrix is used to calculate the feature vector and the characteristic value, and the principal components with a cumulative contribution rate exceeding a preset threshold are screened; the screened principal components are input into the load prediction model as low-dimensional feature vectors.

[0018] Preferably, the load fluctuation pattern is extracted from the power feature database, and the required computing power load is classified, including:

[0019] The collection priority and response time constraints of the computing power load are defined; the key parameters of the load fluctuation pattern are input into the power feature database for matching and searching to screen historical data related to the target computing power load; for each historical data, wavelet transform is used to extract the frequency domain features and abrupt point information of the load waveform; the extracted features are clustered according to load type, fluctuation intensity, and duration to form a computing power load classification rule library.

[0020] Preferably, the power demand parameters under each load classification are generated according to the dynamic change characteristics of the computing power load, including:

[0021] Based on the load classification rule library, the dynamic change interval of the computing power load is determined; in combination with the power grid operation standards and the technical indicators of the flywheel energy storage system, the power demand parameter threshold values under each load classification are set; equipment operation state data and external environmental variables are collected to construct a multi-dimensional parameter correlation matrix; a long short-term memory network model is used to extract features and predict trends of time series parameters; a load feature coding table is generated according to the prediction results for parameter configuration of real-time collection instructions.

[0022] Preferably, the adaptive collection parameter configuration is automatically matched for real-time demand of different loads according to the use preferences and configuration conditions of system operators, including:

[0023] Analyzing the historical configuration records of the operator and the parameter adjustment frequency, extracting the commonly used parameter combination and adjustment strategy; limiting the range and step of the configurable parameters according to the technical level and equipment permission of the operator; dynamically generating a candidate set of parameter configurations combined with the current power grid load state and the remaining capacity of the flywheel energy storage system; using the analytic hierarchy process to prioritize the candidate set and recommend the optimal parameter configuration scheme;

[0024] Based on the Euclidean distance similarity calculation method, the acquisition parameter configuration list under the similar scene in the historical data is recommended, which specifically includes:

[0025] Compare the load feature vector under the historical scene with the current scene item by item, calculate the Euclidean distance and generate a similarity score; filter the historical scenes with a similarity higher than the threshold value according to the score, and extract their parameter configuration scheme; fuse the configuration schemes of multiple similar scenes to generate a weighted average recommended list.

[0026] Preferably, the parameter range in the configuration list is dynamically adjusted according to different power grid dispatching requirements to generate an acquisition instruction that accurately describes the power demand, including:

[0027] Parse the load regulation target and time window requirement in the power grid dispatching instruction; according to the output result of the load prediction model, divide multiple regulation stages and set stage parameter constraints; use the particle swarm optimization algorithm to iteratively adjust the parameter range to ensure that the dynamic response requirement of the power grid dispatching is met; encode the adjusted parameter range into a standardized acquisition instruction and issue it to the flywheel energy storage control system;

[0028] Group the power grid dispatching requirements based on the density clustering algorithm, and determine the parameter threshold value of different groups, including:

[0029] Extract the load regulation amplitude, response speed and duration characteristics in the dispatching requirement; calculate the data point density in the feature space, and identify the high-density area as the core clustering cluster; merge adjacent density-reachable clustering clusters and remove noise data; according to the clustering result, divide the dispatching requirement groups and set the upper and lower limits of the parameter threshold value for each group.

[0030] Preferably, the power feature database is updated, the corresponding load prediction model is adjusted according to the database update result, and it is judged whether the newly added load mode is beyond the applicable range of the current acquisition parameters, including:

[0031] Access real-time grid monitoring data and flywheel energy storage system operation logs, update load pattern samples in the power feature database; extract features from the new load pattern and compare them with existing patterns for similarity; based on a random forest regression model, predict the impact of the new pattern on grid frequency stability; build a gradient boosting decision tree model to determine whether the new pattern can be adapted by adjusting flywheel speed and energy storage capacity; if it exceeds the current parameter range, trigger the parameter expansion process;

[0032] Based on a random forest regression model, predict the impact of the new load pattern on grid stability, specifically including:

[0033] Collect grid frequency deviation, voltage fluctuation and load sudden event data as training samples; build multiple regression decision trees and output stability scores through a voting mechanism; input the features of the new load pattern into the model to get its predicted score for grid stability, and determine the risk level according to the score threshold.

[0034] Preferably, the method further comprises:

[0035] Extract the key features of the new load pattern, including fluctuation period, peak power and duration; calculate the feasible interval of parameter expansion based on the maximum charge and discharge rate and capacity limit of the flywheel energy storage system; use a genetic algorithm to perform multi-objective optimization on parameter combinations to select the optimal solution set that meets system constraints; add the optimized parameters to the configuration list and update it to the real-time collection system;

[0036] Based on the genetic algorithm optimization strategy, determine the constraint conditions of parameter configuration and dynamically match them with the system operating state, specifically including:

[0037] Define the objective function of parameter optimization, including minimizing load response time and balancing system loss; initialize the parameter population and generate the child population through crossover and mutation operations; calculate the fitness value of each individual and retain high fitness individuals for the next generation iteration; finally output the optimal parameter configuration scheme that meets multiple constraints.

[0038] Preferably, the method further comprises:

[0039] Build a visual monitoring interface for power demand data collection, integrating load curve, parameter configuration status and system alarm information; allow operators to manually adjust the collection frequency and data accuracy through interactive controls; use a time series database to store collection data, supporting millisecond-level data queries and backtracking analysis.

[0040] Preferably, the method further comprises:

[0041] The collected data is encrypted and transmitted and stored, sensitive information is protected by end-to-end protection using national encryption algorithm, and a multi-level access permission control mechanism is set to limit the operation permission of unauthorized users.

[0042] Compared with the prior art, the beneficial effects of the present application are:

[0043] At the data collection and processing level, the flywheel rotor speed, temperature and energy storage unit power change data are collected through the deployment of multiple types of sensor networks, and the real-time power output power and load demand curve are obtained combined with the power grid scheduling platform, and the historical operation log and fault record verification data are integrated to verify the data integrity and consistency, this multi-source data fusion method ensures the comprehensiveness and reliability of the collected data. The power demand data is divided according to the load type and a time stamp label is added, which facilitates subsequent classification management and analysis. Principal component analysis method is used for data dimension reduction, which reduces the complexity of data processing, improves the data processing efficiency, and reduces the storage and transmission cost while retaining the key information.

[0044] For load fluctuation mode and computing load classification, the present application defines collection priority and response time constraint, uses power feature database to match and search historical data, uses wavelet transform to extract frequency domain features and mutation point information and clusters to form a classification rule library, which makes the classification of computing load more scientific and accurate, and better adapts to the characteristics of different types of computing load, providing a solid foundation for subsequent power demand parameter feature generation and collection parameter configuration.

[0045] In terms of power demand parameter feature generation, based on the load classification rule library, determine the dynamic change interval, set the parameter threshold value combined with the power grid operation standard and flywheel energy storage system technical index, collect multi-dimensional data to construct the association matrix, use the long short-term memory network model for feature extraction and trend prediction and generate the load feature coding table, which can accurately grasp the dynamic changes of computing load power demand in real time, and provide strong support for accurate collection.

[0046] According to the system operator's use preference and configuration condition, the collection parameter configuration is automatically matched, the candidate set is generated by analyzing the historical configuration record and adjusting the frequency, combined with the current power grid and energy storage system state, and the analytic hierarchy process is used for sorting, and based on the Euclidean distance similarity, the historical similar scene configuration list is recommended, which improves the intelligence and individualization level of parameter configuration, reduces the complexity and error of manual configuration, and improves the usability and operation efficiency of the system.

[0047] In response to the grid dispatching demand, the dispatching instruction is analyzed, the adjustment stage is divided, the parameter constraint is set, the particle swarm optimization algorithm is used to adjust the parameter range and encode as the acquisition instruction, and the parameter threshold is determined based on the density clustering algorithm. It can quickly and accurately respond to the dynamic demand of grid dispatching, ensure the stable operation of the power system, improve the allocation efficiency of power resources, and reduce the risk of grid operation.

[0048] The power feature database is updated and the load prediction model is adjusted, the random forest regression model and the gradient boosting decision tree model are used to evaluate the influence of the new load mode and judge whether it is suitable, and the parameter expansion process is triggered in time, which ensures the timeliness of the data and the accuracy of the model, so that the system can continuously adapt to the changes of the new load mode.

[0049] Based on the expansion of the acquisition parameter configuration list of the new load mode, the genetic algorithm is used to optimize the parameters to ensure that the parameter configuration meets the system constraints, further improving the adaptability and accuracy of data acquisition. In addition, a visual monitoring interface is constructed to facilitate operator monitoring and manual adjustment, a time series database is used to store data to support efficient query and backtracking analysis; data encryption transmission and storage are used and a multi-level access permission control mechanism is set to ensure the security and privacy of the data. In summary, the present application improves the quality and efficiency of the calculation power load power demand data acquisition from multiple aspects, and provides strong technical support for the optimized operation, precise scheduling and safe management of the power system. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 The working principle diagram of the calculation power load power demand data acquisition method based on the flywheel energy storage system is described.

[0051] Figure 2 The flowchart for obtaining flywheel energy storage system data and data processing is described.

[0052] Figure 3 The flowchart for updating the power feature database and judging the new load mode is described.

[0053] Figure 4 The flowchart for expanding and applying the acquisition parameter configuration based on the new load mode is described. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described in detail below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0055] Please refer to Figures 1-4The application provides a computing power load power demand data acquisition method based on a flywheel energy storage system, and the specific implementation steps are as follows:

[0056] Real-time speed, energy storage capacity and power output data of the flywheel energy storage system are acquired, power demand data is classified, a power feature database is constructed, and data dimension reduction processing is performed.

[0057] According to the power feature database, load fluctuation patterns are extracted, and computing power loads to be acquired are classified; according to the dynamic change characteristics of the computing power load, power demand parameters under each load classification are generated; according to the use preferences and configuration conditions of the system operator, the acquisition parameter configuration suitable for the real-time demand of different loads is automatically matched;

[0058] According to different power grid dispatching requirements, the parameter range in the configuration list is dynamically adjusted to generate acquisition instructions accurately describing power demand;

[0059] The power feature database is updated, the corresponding load prediction model is adjusted according to the database update result, and it is judged whether the newly added load pattern exceeds the applicable range of the current acquisition parameter;

[0060] According to the newly added load pattern, the acquisition parameter configuration list is expanded and applied to real-time data acquisition.

[0061] Embodiment 1:

[0062] The system acquires real-time speed, energy storage capacity and power output data of the flywheel energy storage system, classifies power demand data, constructs a power feature database, and performs data dimension reduction processing, and the specific process is as follows.

[0063] In the data acquisition stage, in order to accurately acquire the flywheel energy storage system operation state data, a multi-type sensor network needs to be deployed. For the flywheel rotor, a high-precision speed sensor is installed, which monitors the rotation speed of the flywheel rotor in real time through photoelectric or magneto-electric induction principle, converts the mechanical rotation signal into an electrical signal or a digital signal, and collects data at a sampling frequency of milliseconds to ensure that subtle changes in rotation speed can be captured; at the same time, a temperature sensor is configured, which is installed close to the surface or key parts of the flywheel rotor, and real-time temperature data during rotor operation is collected to avoid affecting the performance of the flywheel or causing safety problems due to high temperature. For the energy storage unit, an electric quantity sensor is deployed, which measures battery voltage, current and other parameters, and calculates the real-time electric quantity change data of the energy storage unit by using specific algorithms such as ampere-hour integration method or coulomb counting method, to accurately reflect the charging and discharging state of the energy storage unit.

[0064] The power grid dispatching platform is the key source of obtaining real-time power output power and load demand curve. The system establishes stable and reliable communication connection with the power grid dispatching platform for data interaction, such as using industrial Ethernet, optical fiber communication and the like. Real-time power output power data sent by the power grid dispatching platform is received by using a standardized communication protocol, such as IEC61850 or Modbus protocol. These data contain the power output by the flywheel energy storage system to the power grid at different time nodes. At the same time, the load demand curve is obtained, which intuitively presents the demand trend of the power grid for power in a certain period of time, providing a basis for subsequent data processing.

[0065] In order to ensure data quality, the system integrates historical operation logs and fault records. The historical operation logs record the past operation parameters, operation records and other information of the flywheel energy storage system; the fault records contain various fault phenomena, occurrence time, treatment measures and other contents of the system in the running process. The system uses data checking algorithm to compare and analyze the newly collected data with the historical data, checks the integrity of the data to ensure that there is no data missing or omission, verifies the consistency of the data to prevent the occurrence of mutually contradictory data, and ensures the accuracy of subsequent data processing and analysis.

[0066] After completing data collection and verification, the power demand data is classified. According to the load type, it is divided into basic load, peak load and fluctuating load. The basic load refers to the relatively stable and small fluctuation power demand in the power grid under normal operation, which is usually generated by some continuously running and power consumption stable equipment, such as some continuous production industrial equipment, long-term lighting facilities, etc.; the peak load shows the power demand with a large increase in a short time, which is common in the peak period of electricity consumption, such as the period when air conditioners are concentratedly turned on in summer, the period when large-scale activity places are densely used, etc.; the fluctuating load is the power demand with irregular changes, which may be generated by some intermittent equipment or equipment greatly affected by external factors, such as wind power and photovoltaic power generation equipment, whose power output will fluctuate with the change of natural conditions. In the classification process, a time stamp label accurate to seconds is added to each type of data, which is convenient for subsequent analysis of the change and mutual relationship of different types of loads at different time points.

[0067] When constructing the power feature database, a distributed storage system is used to establish a table structure. The distributed storage system combines the advantages of distributed file system and distributed database, and has the characteristics of high scalability, high reliability and high performance. In the table structure design, fields are set to store the speed data of the flywheel rotor, the capacity data of the energy storage unit, the power output power data and the load classification information, etc. Through reasonable table structure design, efficient storage and rapid retrieval of data are realized, which is convenient for subsequent analysis and processing of data.

[0068] The power demand data is processed by dimension reduction based on principal component analysis. First, the original power data is normalized. Since the dimensions and value ranges of different parameters in the original data are different, such as the unit of rotational speed is revolutions per minute and the unit of power is kilowatts, the normalization process maps data of different dimensions to the same interval, eliminating the influence of dimension differences on data analysis. Normalization methods include minimum-maximum normalization or Z-score normalization. Next, the sliding window mechanism is used to extract local features of time series data at fixed time intervals to construct a multi-dimensional feature matrix. For example, taking 1 minute as the sliding window, the mean, variance, maximum, and minimum of the rotational speed, power, and other data in this time period are extracted as a row of data in the multi-dimensional feature matrix. Then, the covariance matrix is used to calculate the eigenvectors and eigenvalues, which reflect the correlation between the features. By performing eigenvalue decomposition on the covariance matrix, the eigenvectors and eigenvalues are obtained. Set a cumulative contribution rate threshold, such as 85% or 90%, and select the principal components whose cumulative contribution rate exceeds the threshold. These principal components contain the main information of the original data. The selected principal components are used as low-dimensional feature vectors, which contain key information such as load fluctuation amplitude, duration, periodicity, and associated equipment status, and are input into the load prediction model to provide data support for subsequent load analysis and prediction.

[0069] After completing the data dimension reduction, the system continues to extract the load fluctuation pattern based on the power feature database, classifies the computing power load to be collected, generates features for the power demand parameters under each load classification based on the dynamic characteristics of the computing power load, automatically matches and configures the collection parameters according to the usage preferences and configuration conditions of the system operator, dynamically adjusts the parameter range in the configuration list according to different grid dispatching requirements, generates collection instructions that accurately describe the power demand, updates the power feature database, adjusts the corresponding load prediction model based on the database update results, determines whether the newly added load pattern exceeds the applicable range of the current collection parameters, and extends the collection parameter configuration list based on the newly added load pattern and applies it to real-time data collection. In subsequent steps, the low-dimensional feature vectors obtained after data dimension reduction are used as the basis for accurate collection and processing of computing power load power demand data through a series of algorithms and models, ensuring that the system can efficiently and accurately meet the needs of grid operation and management.

[0070] Embodiment 2:

[0071] When classifying power demand data, constructing a power feature database, and performing data dimension reduction, the specific operations are as follows.

[0072] The deployment of multi-type sensor networks is the foundation for obtaining accurate data. Among them, the speed sensor adopts a high-precision Hall effect sensor, which converts mechanical rotation into electrical signals by detecting the pulse signals generated by the rotation of the magnet on the flywheel rotor. This sensor has high resolution and fast response characteristics, and can accurately capture the speed changes of the flywheel rotor under different working conditions. The sampling frequency is set to 1000 Hz to ensure accurate acquisition of speed data even at high rotation speeds. The temperature sensor uses a platinum resistance temperature sensor, which has high measurement accuracy and good stability. It uses a contact measurement method and is tightly attached to the key parts of the flywheel rotor surface to monitor the temperature changes in real time. To avoid environmental interference, the sensor is wrapped with thermal insulation material, and a signal conditioning circuit is provided to amplify and filter the collected temperature signals, improving signal quality. For the energy storage unit, an intelligent power monitoring module is deployed, which integrates voltage sensors, current sensors and microprocessors to accurately calculate the remaining power and charge-discharge state of the energy storage unit by measuring the voltage and current of the battery in real time and combining the battery's charge-discharge characteristic curve.

[0073] The grid dispatch platform interacts with the system through an industrial Ethernet interface. The IEC61850 communication protocol is used to ensure the reliability and real-time performance of data transmission. The system sends data requests to the grid dispatch platform at regular intervals to receive real-time power output data and load demand curves. The power output data includes active power, reactive power and power factor parameters, which are stored in time series format for subsequent analysis. The load demand curve reflects the power demand trend of the grid at different time periods, which is analyzed and processed by the system to extract key feature points such as peak demand, valley demand and demand change rate.

[0074] To ensure data quality, the system integrates historical operation logs and fault records. Historical operation logs are stored in a distributed file system, containing all running parameters and operation records since the system was put into use. Fault records are stored in a structured way in a relational database, recording the occurrence time, fault phenomenon, processing process and final result of each fault. The system develops a special data integration module, which first cleans the historical data to remove duplicate and invalid data, then performs format unification processing to convert data of different sources and formats into a unified data structure. Finally, the data verification algorithm is used to verify the integrity and consistency of the integrated data, ensuring the accuracy and reliability of the data.

[0075] In terms of load type classification, the system classifies according to the variation characteristics of power demand. The identification of base load is based on long-term data analysis, and by calculating the mean and standard deviation of the load, the load with fluctuation range within ±10% of the mean is defined as the base load. This kind of load mainly comes from continuously running industrial equipment, lighting systems and basic electrical equipment in residential life. The determination of peak load combines time factors and power change rate, when the load power rises more than 50% of the mean in a short time (such as 15 minutes), it is determined as peak load. The identification of fluctuating load is relatively complex, the system uses frequency domain analysis method, Fourier transform is carried out on the load data, and the high frequency component is extracted, the load with high frequency component ratio is defined as fluctuating load. Add time stamp label to each type of data to the millisecond level, time stamp uses coordinated universal time (UTC), to ensure the consistency of data time in distributed system.

[0076] The distributed storage system adopts the combination of Hadoop distributed file system (HDFS) and Apache HBase distributed database. HDFS is responsible for storing raw sensor data and historical operation logs, which has high fault tolerance and high throughput characteristics, and can handle massive data storage and read-write operations. HBase is used to store structured power feature data such as speed, capacity, power and load classification information, etc. In the table structure design, the combination of partition table and index table is adopted, and partition is carried out according to time and load type to improve data query efficiency. At the same time, multi-level index is established, including time index, load type index and parameter value index, which supports fast data retrieval and statistical analysis.

[0077] In data dimensionality reduction processing, the original power data is processed based on principal component analysis method. First, normalization operation is carried out, Z-score normalization method is adopted, each parameter value is converted to standard normal distribution, and the influence of different parameter dimensions is eliminated. Sliding window mechanism divides the time series data according to fixed time interval (such as 1 minute), extracts various statistical features in each window, including mean, variance, maximum, minimum, median and quartile. These features constitute a multi-dimensional feature matrix, each row of the matrix represents the power data features in a time window. By calculating the covariance matrix of the feature matrix, the correlation between features is analyzed. The covariance matrix is decomposed to get the eigenvector and eigenvalue. According to the size of eigenvalue, the principal components with cumulative contribution rate exceeding 90% are selected. These principal components contain the main information of the original data, and are used as low-dimensional feature vectors to obtain low-dimensional feature vectors containing load fluctuation amplitude, duration, periodicity and associated equipment state, etc. The dimension of the low-dimensional feature vector is greatly reduced compared with the original data, reducing the complexity of data storage and processing, while retaining the main features of the data, providing efficient data representation for subsequent load analysis and prediction.

[0078] After the data dimension reduction is completed, the low-dimensional feature vectors are input to the subsequent analysis module. The system uses these feature vectors for load pattern recognition and classification, quickly and accurately identifying the type and characteristics of the load by comparing historical data and current features. At the same time, these feature vectors also provide input for the load prediction model, which predicts future power demand trends based on historical data and current features, providing decision support for grid dispatching and energy storage system control. In addition, the system also uses low-dimensional feature vectors for anomaly detection, setting normal feature ranges to promptly detect abnormal situations in the power system, such as equipment failure and load mutation, and issue warning signals to ensure the safe and stable operation of the power system.

[0079] Embodiment 3:

[0080] In the classification and feature generation process of computing power load, the collection priority and response time constraints of computing power load must be determined first. The collection priority is set according to the importance of computing power equipment in the business process. For key computing power equipment that ensures the operation of core business, high priority is given, and the system is required to prioritize the collection of its power demand data; while for auxiliary or non-critical business computing power equipment, a relatively low priority is set. The response time constraint is determined according to the real-time data requirements of different business scenarios. For example, real-time transaction processing, online gaming and other businesses that require high data real-time performance require response times to be controlled within milliseconds; for data processing, background analysis and other businesses, the response time can be relaxed to seconds or minutes.

[0081] The key parameters of the load fluctuation pattern are input into the power feature database for matching and retrieval. These key parameters include load change frequency (unit: times / minute, indicating the number of load changes per unit time), amplitude (unit: kilowatts, indicating the difference in power before and after load change), duration (unit: minutes, indicating the length of time the load is in a certain state), etc. The system searches for historical data related to the target computing power load in the power feature database based on these parameters. During the retrieval process, an efficient data indexing mechanism is used to quickly locate the data records that meet the conditions by pre-classifying and indexing the historical data according to the key parameters.

[0082] For each piece of retrieved historical data, wavelet transform is applied to analyze the load waveform. Wavelet transform is a mathematical tool that converts time-domain signals into a joint time-frequency representation. Its basic principle is to perform an inner product operation with wavelet functions of different scales and positions to obtain the characteristics of the signal at different frequencies and time resolutions. Through wavelet transform, the variation law of the load waveform at different frequency components can be extracted, and the rapid change part corresponding to the high-frequency component and the slow change trend corresponding to the low-frequency component can be captured. At the same time, the mutation point information in the waveform can also be accurately determined, which often corresponds to the start, shutdown or running state switching of the computing device and other events.

[0083] The extracted features are clustered according to the load type, fluctuation intensity and duration, and a clustering algorithm (such as K-means clustering algorithm) is used to form a computing load classification rule library. The core idea of K-means clustering algorithm is to divide the data into clusters, so that the sum of squared distances from each data point to its cluster center is minimized. In the clustering process, the cluster centers are randomly initialized, then the distance from each feature data point to each cluster center is calculated, and the data point is assigned to the nearest cluster. Then, the center position of each cluster is recalculated, and the above assignment and center calculation process is repeated until the cluster center no longer changes significantly or the preset iteration number is reached. Through clustering, computing loads with similar features are classified into one category, forming different classification rules, which describe the feature range and typical mode of each computing load in detail.

[0084] In the feature generation link, based on the load classification rule library, the dynamic change interval of computing load is determined. Combined with the grid operation standard and the technical indicators of flywheel energy storage system, the threshold values of power demand parameters under each load classification are set. For example, for the power parameter, according to the maximum output power of the flywheel energy storage system and the safe operation power range of the grid, the upper limit and lower limit of power are set for different types of computing loads; for the current parameter, according to the rated current of the device and the carrying capacity of the line, the current threshold is determined. Collect device operating state data, including device temperature , voltage , current , etc., and external environmental variables such as environmental temperature , humidity , etc., to construct a multi-dimensional parameter correlation matrix. Each row in the matrix represents a sampling time, and each column corresponds to a parameter. The relationship between various parameters is integrated and represented in the form of a matrix.

[0085] The long short-term memory (LSTM) model is used for feature extraction and trend prediction of time series parameters. The LSTM model is a special recurrent neural network (RNN) that can effectively handle long-term dependencies in long sequence data by introducing a gating mechanism. The LSTM model includes an input gate, a forget gate, and an output gate. The input gate controls the degree of current input information entering the cell state, the forget gate determines which information in the previous cell state is retained, and the output gate determines the output based on the cell state and the current input. During training, the time series data in the multi-dimensional parameter correlation matrix is used as input, which is processed by multiple LSTM units to extract complex features and trends hidden in the data. The model adjusts internal parameters through the backpropagation algorithm to minimize the error between the predicted value and the actual value. According to the prediction results of the LSTM model, a load feature encoding table is generated, which uniformly encodes different load features for real-time acquisition instruction parameter configuration, enabling the system to automatically adjust data acquisition parameters based on predicted load features, improving the relevance and effectiveness of collected data.

[0086] Embodiment 4:

[0087] The automatic matching of adaptive acquisition parameter configuration and parameter adjustment process based on grid dispatching requirements is as follows.

[0088] The system first analyzes the historical configuration records and parameter adjustment frequency of the operator. Suppose there is an operator A who has set the acquisition frequency to 1 time per minute and the data precision to two decimal places for a certain type of common basic load computing power equipment in the past month. The system uses data mining algorithms to deeply analyze these historical operation records of operator A and extracts the commonly used parameter combination as (acquisition frequency: 1 time / minute, data precision: two decimal places). At the same time, it summarizes that operator A is used to prioritize data stability when facing basic load computing power equipment and rarely adjusts parameters frequently. Based on this, the system sets the range of configurable parameters for operator A based on his technical level and equipment permissions. If his technical level is intermediate, the acquisition frequency can be adjusted between 0.5-2 times / minute with a step size of 0.1 times / minute, and the data precision can be selected between one decimal place and three decimal places with a step size of one decimal place.

[0089] The system dynamically generates a candidate set of parameter configurations in combination with the current grid load state and the remaining capacity of the flywheel energy storage system. For example, when the grid is in a low electricity consumption period, the flywheel energy storage system has a remaining capacity of 80%, and the current data collection load is a non-core computing task of a data center (which belongs to fluctuating load), the system uses a state evaluation algorithm to generate a candidate set of parameter configurations by considering factors such as light grid load, sufficient energy storage capacity, and load fluctuation characteristics. The candidate set may include: Scheme One (collection frequency: 0.8 times / minute, data precision: one decimal place), Scheme Two (collection frequency: 1 time / minute, data precision: two decimal places), and Scheme Three (collection frequency: 1.2 times / minute, data precision: one decimal place).

[0090] Using the analytic hierarchy process, the system constructs a hierarchical structure model to prioritize the candidate set. In this model, parameters such as collection frequency and data precision are considered as the criterion layer, different parameter configuration schemes are considered as the scheme layer, and the target layer is to select the optimal parameter configuration to meet the current load demand. The system considers the importance of each parameter in different scenarios and assigns appropriate weights to each parameter. For the above fluctuating load and light grid load, sufficient energy storage capacity scenario, assume the weight of collection frequency is 0.6 and the weight of data precision is 0.4. Through the comprehensive evaluation of the parameters in each candidate scheme, it is calculated that the priority of Scheme Two is the highest, and Scheme Two is presented to the operator as the recommended optimal parameter configuration scheme.

[0091] Based on the Euclidean distance similarity calculation method, the system compares the load feature vectors of historical scenarios with the current scenario item by item. Assume that there is a scenario B in the historical scenarios, its load feature vector is (load change frequency: 0.5 times / minute, load change amplitude: 5 kW, duration: 30 minutes), and the load feature vector of the current scenario is (load change frequency: 0.6 times / minute, load change amplitude: 6 kW, duration: 25 minutes). The system compares the corresponding elements of the two vectors, calculates the Euclidean distance and generates a similarity score. Set the similarity threshold to 0.8, filter out historical scenarios with a similarity higher than the threshold, and extract their parameter configuration schemes. If the similarity score of scenario B reaches 0.85, it meets the conditions, and the parameter configuration scheme of scenario B (collection frequency: 0.7 times / minute, data precision: two decimal places) is extracted. If there are multiple similar scenarios, the system fuses the configuration schemes of these scenarios to generate a recommended list after weighted averaging.

[0092] In the adjustment of grid dispatching demand parameters, the system analyzes the load adjustment target and time window requirement in the grid dispatching instruction. For example, the grid dispatching instruction requires that the load in a certain area be adjusted from the current 500 kilowatts to 600 kilowatts in the next 2 hours. According to the output result of the load prediction model, the system divides multiple adjustment stages. Assuming that it is divided into three stages: the first stage (0-30 minutes), the load is adjusted to 530 kilowatts; the second stage (30-90 minutes), the load is adjusted to 570 kilowatts; the third stage (90-120 minutes), the load is adjusted to 600 kilowatts, and parameter constraints are set for each stage, such as the power adjustment rate allowed in the first stage is not more than 1 kilowatt / minute.

[0093] Using the particle swarm optimization algorithm, the system sets the position and speed of the particles. Each particle represents a set of parameter configurations, the position of the particle corresponds to the value of the parameter, and the speed represents the trend of the parameter change. By iteratively adjusting the parameter range, the parameter meets the dynamic response requirement of grid dispatching. In the iteration process, the particle constantly updates its position and speed according to its own optimal position and the optimal position of the group, and gradually finds a better parameter configuration. Finally, the adjusted parameter range is encoded into a standardized collection instruction, which is issued to the flywheel energy storage control system through the communication protocol.

[0094] Based on the density clustering algorithm, the system extracts the load adjustment amplitude, response speed and duration characteristics in the dispatching demand. For example, for a series of dispatching demands, some require a load adjustment amplitude of 100 kilowatts, a response speed of completing adjustment within 5 minutes, and a duration of 60 minutes; some require a load adjustment amplitude of 50 kilowatts, a response speed of completing adjustment within 3 minutes, and a duration of 30 minutes, etc. The system calculates the data point density of these characteristics in the feature space, identifies the high-density area as the core clustering cluster, merges adjacent density-reachable clustering clusters, removes noise data, and divides the dispatching demand groups according to the clustering result. For example, the dispatching demand is divided into high adjustment amplitude, long duration group, medium adjustment amplitude, medium duration group, etc., and parameter threshold upper and lower limits are set for each group. For the high adjustment amplitude, long duration group, the lower limit of the power adjustment amplitude is set to 80 kilowatts, the upper limit is unlimited, the response speed lower limit is 5 minutes, and the upper limit is unlimited, etc.

[0095] Example 5:

[0096] In terms of updating the database, judging the load mode, and expanding the parameter configuration list, the system will continuously access real-time power grid monitoring data and the operation log of the flywheel energy storage system. Real-time power grid monitoring data includes voltage, frequency, and other information of each node in the power grid. These data are collected at fixed time intervals by specialized monitoring equipment and transmitted to the system through a stable communication network. The operation log of the flywheel energy storage system records the operating status of each component within the system, such as the speed change of the flywheel, the charging and discharging process of the energy storage unit, and the operation instructions of related control modules. These log data are generated in real time during system operation and stored in local storage devices, and then transmitted to the system for integration, adding new data to the load mode samples in the power feature database.

[0097] When extracting features from the new load mode, the same method as the existing mode is used. For example, wavelet transform is used to analyze the waveform of the new load mode, capture its change rule at different frequency components, and determine the mutation points in the waveform. These mutation points may correspond to the connection of new computing equipment, the switching of equipment operating state, etc. Then, using the similarity measurement algorithm, the features of the new load mode are compared with the existing mode, and the similarity between them is calculated. If the similarity is high, it means that the new load mode has certain similarity with the existing mode, and the processing method of the existing mode can be referred to; if the similarity is low, further in-depth analysis is needed.

[0098] Based on the random forest regression model, the system collects power grid frequency deviation, voltage fluctuation, and load mutation event data as training samples. Power grid frequency deviation data reflects the difference between actual power grid operating frequency and standard frequency, voltage fluctuation data reflects the voltage change of each node in the power grid, and load mutation event data records the significant change of load in a short time. By constructing multiple regression decision trees, each decision tree is constructed based on different training sample subsets and features, and then a stable score is output through the voting mechanism. After inputting the features of the new load mode into the model, the model will integrate the results of multiple decision trees to get the predicted score of the grid stability, and judge the risk level according to the pre-set score threshold.

[0099] The system constructs a gradient boosting decision tree model, which continuously iterates to learn whether the new mode can be adapted by adjusting the flywheel speed and energy storage capacity. Based on historical data and the current load mode, the model gradually optimizes the decision-making process, and each iteration adjusts based on the previous prediction error to improve the accuracy of the prediction. If the judgment result shows that the new load mode exceeds the applicable range of the current collection parameters, the parameter expansion process will be triggered.

[0100] In the parameter expansion link, the system extracts the key features of the new load pattern, including fluctuation period, peak power and duration, etc. For example, if the new load pattern shows periodic power fluctuations, the fluctuation period is an important feature; the peak power reflects the maximum power value that the load pattern may reach during operation; the duration indicates the length of time that the load pattern remains in a certain state. According to the maximum charge and discharge rate and capacity limit of the flywheel energy storage system, the feasible interval of parameter expansion is determined through a series of calculations. The maximum charge and discharge rate determines the maximum energy that the system can release or absorb in a unit of time, and the capacity limit specifies the maximum amount of electricity that the system can store, which will constrain the range of parameter expansion.

[0101] The genetic algorithm is used to define the objective function of parameter optimization, including minimizing the load response time and balancing the system loss. The load response time refers to the time required for the system to respond to load changes and make adjustments, and shortening this time can improve the response speed and stability of the system; the system loss balancing aims to reasonably allocate the energy loss of the system during operation, avoid excessive loss of some components, and prolong the service life of the system. Initialize the parameter population, each individual represents a set of possible parameter configurations, generate the offspring population through crossover and mutation operations. Crossover operation is to exchange part of the genes of two individuals to generate new individuals; mutation operation is to randomly change the genes of individuals to increase the diversity of the population. Calculate the fitness value of each individual, the higher the fitness value, the more the parameter configuration corresponding to the individual meets the requirements of the objective function, and the high fitness individuals are retained for the next generation iteration. After multiple iterations, the optimal parameter configuration scheme that meets multiple constraints is finally output, and the optimized parameters are added to the configuration list and updated to the real-time acquisition system.

[0102] In addition, the system builds a visual monitoring interface for power demand data acquisition. This interface integrates load curve, parameter configuration status and system alarm information. The load curve shows the power load change in different time periods in a graphical way, making it easy for operators to intuitively understand the fluctuation trend of the load; the parameter configuration status displays the current acquisition parameters used by the system, as well as the basis for setting these parameters and the applicable scenarios; the system alarm information will prompt when abnormal conditions are detected, such as load exceeding the preset range, abnormal equipment operating state, etc. Through interactive controls, operators can manually adjust the acquisition frequency and data accuracy to meet individual needs in different scenarios. The system uses a time series database to store acquisition data, which is optimized specifically for time series data, can efficiently store and query data arranged in chronological order, supports millisecond-level data query and backtracking analysis, and facilitates in-depth research and analysis of historical data by operators.

[0103] Meanwhile, the system performs encrypted transmission and storage of the collected data, uses national secret algorithm to perform end-to-end protection on sensitive information. In the data transmission process, the data is encrypted to ensure that the data is not stolen or tampered with during network transmission; in the storage link, the stored data is also encrypted to prevent illegal access to the data in the storage device. A multi-level access permission control mechanism is set up, different access permissions are allocated according to the identity and responsibility of the user, the operation permission of unauthorized users is limited, only the user with the corresponding permission can view, modify and other operations on the data, to ensure the security and privacy of the data.

[0104] It should be noted that, in this paper, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "include" "contain" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0105] Although the embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for collecting data on power demand for computing load based on a flywheel energy storage system, characterized in that, The method comprises: Obtaining real-time rotating speed, energy storage capacity and power output data of a flywheel energy storage system, classifying power demand data, constructing a power feature database, and performing data dimension reduction processing; According to the power feature database, extracting load fluctuation patterns and classifying the computing power loads to be collected; according to the dynamic change characteristics of the computing power loads, generating features of the power demand parameters under each load classification; according to the use preferences and configuration conditions of the system operator, automatically matching the adaptive collection parameter configuration for the real-time demand of different loads; According to different power grid dispatching requirements, dynamically adjusting the parameter range in the configuration list to generate collection instructions that accurately describe the power demand; Updating the power feature database, adjusting the corresponding load prediction model according to the database update result, and judging whether the newly added load mode exceeds the applicable range of the current collection parameters; updating the power feature database, adjusting the corresponding load prediction model according to the database update, and judging whether the newly added load mode exceeds the applicable range of the current collection parameters; According to the newly added load mode, expanding the collection parameter configuration list and applying it to real-time data collection; The method comprises: Deploying a multi-type sensor network to collect the rotating speed, temperature and energy storage unit power change data of the flywheel rotor; obtaining real-time power output power and load demand curve through a power grid dispatching platform; integrating historical operation logs and fault records to verify the completeness and consistency of the data; dividing the power demand data into basic load, peak load and fluctuating load according to the load type, and adding a timestamp label to each type of data; using a distributed storage system to establish the table structure of the power feature database, storing the rotating speed, capacity, power and load classification information; based on the principal component analysis method, converting the power demand data into a low-dimensional feature vector; the reduced features include load fluctuation amplitude, duration, periodicity and associated equipment state; The method comprises: Normalizing the original power data to eliminate dimensional differences; extracting local features of time series data through a sliding window mechanism to construct a multi-dimensional feature matrix; calculating the feature vector and eigenvalue using the covariance matrix, and selecting the principal components with a cumulative contribution rate exceeding a preset threshold; inputting the selected principal components as low-dimensional feature vectors into the load prediction model; The method comprises: Access real-time grid monitoring data and flywheel energy storage system operation logs, update load pattern samples in the power feature database; extract features from new load patterns and compare them with existing patterns for similarity; based on a random forest regression model, predict the impact of new patterns on grid frequency stability; build a gradient boosting decision tree model to determine whether new patterns can be adapted by adjusting flywheel speed and energy storage capacity; if it exceeds the current parameter range, trigger the parameter expansion process; Based on the random forest regression model, the influence of the new load pattern on the stability of the power grid is predicted, which specifically includes: Collect grid frequency deviation, voltage fluctuation and load mutation event data as training samples; build multiple regression decision trees and output stability scores through a voting mechanism; input the features of the new load pattern into the model to get the predicted score of its impact on grid stability, and judge the risk level according to the score threshold.

2. The flywheel-based energy storage system-based computing power load power demand data collection method according to claim 1, characterized in that, The power feature database is used to extract load fluctuation patterns and classify the required computing power loads, including: Define the collection priority and response time constraints of computing power loads; input the key parameters of load fluctuation patterns into the power feature database for matching and searching; filter the historical data related to the target computing power load; for each historical data, use wavelet transform to extract the frequency domain features and mutation point information of the load waveform; cluster the extracted features by load type, fluctuation intensity and duration to form a computing power load classification rule base.

3. The flywheel-based energy storage system-based computing power load power demand data collection method according to claim 1, characterized in that, According to the dynamic change characteristics of the computing power load, the power demand parameters under each load classification are generated, including: Based on the load classification rule base, determine the dynamic change interval of the computing power load; set the power demand parameter threshold for each load classification based on the grid operation standards and technical indicators of the flywheel energy storage system; collect equipment operating state data and external environmental variables to build a multi-dimensional parameter correlation matrix; use a long short-term memory network model to extract features and predict trends of time series parameters; generate a load feature coding table based on the prediction results for real-time acquisition of parameter configuration.

4. The flywheel-based energy storage system-based computing power load power demand data collection method according to claim 1, characterized in that, According to the use preferences and configuration conditions of system operators, automatically match the adaptive acquisition parameter configuration for real-time demand of different loads, including: Analyze the operator's historical configuration records and parameter adjustment frequency to extract their commonly used parameter combinations and adjustment strategies; limit the range and step size of configurable parameters according to the operator's technical level and device permissions; dynamically generate a parameter configuration candidate set based on the current grid load state and the remaining capacity of the flywheel energy storage system; use the analytic hierarchy process to prioritize the candidate set and recommend the optimal parameter configuration scheme; Based on the Euclidean distance similarity calculation method, recommend the acquisition parameter configuration list in the similar scenarios in the historical data, including: Compare the load feature vectors of the historical scenarios with the current scenario item by item, calculate the Euclidean distance and generate a similarity score; filter historical scenarios with a similarity score higher than the threshold to extract their parameter configuration schemes; fuse multiple similar scenario configuration schemes to generate a weighted average recommended list.

5. The flywheel-based energy storage system-based computing power load power demand data collection method according to claim 1, characterized in that, The parameter range in the configuration list is dynamically adjusted according to different power grid scheduling requirements to generate collection instructions accurately describing power demand, including: Parsing the load regulation target and time window requirement in the power grid scheduling instruction; dividing multiple regulation stages and setting stage parameter constraints according to the output results of the load prediction model; iteratively adjusting the parameter range using the particle swarm optimization algorithm to ensure that the dynamic response requirements of the power grid scheduling are met; encoding the adjusted parameter range into standardized collection instructions and issuing them to the flywheel energy storage control system; Grouping the power grid scheduling requirements based on the density clustering algorithm to determine the parameter threshold of different groups, including: Extracting the load regulation amplitude, response speed and duration characteristics in the scheduling requirements; calculating the data point density in the feature space to identify high-density areas as core clustering clusters; merging adjacent density-reachable clustering clusters and removing noise data; dividing the scheduling requirement groups according to the clustering results and setting the upper and lower limits of the parameter threshold for each group.

6. The flywheel-based energy storage system-based computing power load electricity demand data collection method according to claim 1, characterized in that, The collection parameter configuration list is expanded according to the new load mode and applied to real-time data collection, including: Extracting the key features of the new load mode, including fluctuation period, peak power and duration; calculating the feasible interval of parameter expansion according to the maximum charge and discharge rate and capacity limit of the flywheel energy storage system; using genetic algorithm to perform multi-objective optimization on parameter combinations to select the optimal solution set that meets system constraints; adding the optimized parameters to the configuration list and synchronously updating to the real-time collection system; Based on the genetic algorithm optimization strategy, the constraint conditions of parameter configuration are determined and dynamically matched with the system running state, including: Defining the objective function of parameter optimization, including minimizing the load response time and balancing the system loss; initializing the parameter population to generate the child population through crossover and mutation operations; calculating the fitness value of each individual to retain high fitness individuals for the next generation iteration; finally outputting the optimal parameter configuration scheme that meets multiple constraints.

7. The flywheel-based energy storage system-based computing power load electricity demand data collection method according to claim 1, characterized in that, The method further includes: Building a visual monitoring interface for power demand data collection, integrating load curve, parameter configuration status and system alarm information; allowing operators to manually adjust the collection frequency and data accuracy through interactive controls; using a time series database to store collection data to support millisecond-level data query and backtracking analysis.

8. The flywheel-based energy storage system-based computing power load electricity demand data collection method according to claim 1, characterized in that, The method further includes: Encrypting the collection data for transmission and storage, using the national encryption algorithm to protect sensitive information end-to-end; setting a multi-level access permission control mechanism to limit the operation permissions of unauthorized users.

Images