A soc estimation method for a flywheel energy storage array system

Abstract

The application relates to the technical field of flywheel energy storage, and discloses an SOC estimation method for a flywheel energy storage array system. The method synchronously collects operation parameters such as flywheel rotating speed and bus current through a multi-source sensing unit, inputs a dynamic parameter observation model to estimate dynamic characteristic parameters, the model dynamically adjusts observation weights based on historical data, then inputs the dynamic characteristic parameters into a multi-model fusion algorithm to generate a fused state characteristic vector, the algorithm assigns fusion coefficients according to the spatial correlation of each flywheel unit, uses an error correction model to correct the deviation of the state characteristic vector to obtain an SOC estimation result and stores the result. The application also relates to steps such as model construction, data collection and processing, and algorithm optimization. The application can improve the SOC estimation precision, adapt to complex working conditions, guarantee data integrity and safety, and is of great significance to the development of flywheel energy storage technology.

Description

Technical Field

[0001] The present invention relates to the technical field of flywheel energy storage, and in particular to a SOC estimation method for a flywheel energy storage array system. Background Art

[0002] In today's energy landscape, with the rapid development of renewable energy, energy storage technology is becoming increasingly important. Flywheel energy storage, as a new physical energy storage method, boasts broad application prospects in numerous fields, including grid peak regulation, uninterruptible power supplies, and electric vehicles, thanks to its advantages such as high power density, long life, and rapid response. However, accurately estimating the SOC (state of charge) of flywheel energy storage array systems remains a critical issue.

[0003] Flywheel energy storage array systems operate under complex and variable operating conditions. Flywheel performance varies significantly under varying charge and discharge rates, ambient temperatures, and mechanical stress conditions. Traditional SOC estimation methods, such as the ampere-hour integration method, primarily calculate SOC based on the integral of current over time. However, in practice, this method's estimation accuracy degrades over time due to factors such as current measurement errors, variations in charge and discharge efficiency, and self-discharge, making it unable to meet the requirements for real-time, precise monitoring of flywheel energy storage array systems.

[0004] The open-circuit voltage method estimates SOC by measuring the open-circuit voltage of a flywheel energy storage system. However, the relationship between open-circuit voltage and SOC is not a simple linear one and is affected by various factors such as temperature and battery aging. This makes this method difficult to accurately implement in actual flywheel energy storage array systems. Furthermore, during operation, individual flywheel units interact with each other, and traditional methods often overlook this spatial correlation, resulting in an inability to fully and accurately reflect the SOC status of the entire array system.

[0005] Furthermore, in complex industrial environments, flywheel energy storage array system sensors are susceptible to electromagnetic interference, mechanical vibration, and other factors, resulting in measurement errors. If these errors are not promptly addressed and corrected, they can further reduce the accuracy of SOC estimation. Furthermore, as flywheel energy storage array systems continue to scale, the amount of data is growing exponentially. Efficiently processing and analyzing this data to achieve accurate SOC estimation presents another challenge in this field. Summary of the Invention

[0006] The object of the present invention is to provide a SOC estimation method for a flywheel energy storage array system to solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a SOC estimation method for a flywheel energy storage array system, the method comprising:

[0008] Synchronously collecting operating parameters of the flywheel energy storage array through a multi-source sensing unit, wherein the operating parameters include flywheel speed, bus current, winding temperature, and mechanical vibration intensity;

[0009] Inputting the operating parameters into a preset dynamic parameter observation model to estimate the dynamic characteristic parameters of the flywheel system in real time, wherein the dynamic parameter observation model dynamically adjusts the observation weight based on historical operating data of the flywheel unit;

[0010] Inputting the dynamic characteristic parameters into a preset multi-model fusion algorithm to generate a fused state feature vector, wherein the multi-model fusion algorithm allocates fusion coefficients according to the spatial correlation of each flywheel unit;

[0011] The state characteristic vector is corrected for deviation using a preset error correction model to obtain an SOC estimation result, and the corrected result is stored in a state database in a time series.

[0012] Preferably, the step of constructing the dynamic parameter observation model includes:

[0013] Acquire a historical operation data set, wherein each data in the historical operation data set is marked with an error type and an error amplitude of a dynamic characteristic parameter;

[0014] Divide the training subsets based on the error type and error amplitude, each training subset corresponding to a dynamic characteristic scenario;

[0015] The initial observation model is trained in parallel using the training subset until the parameter estimation error rate of the initial observation model for each scene is less than or equal to a preset first threshold, thereby stopping the training and obtaining an intermediate observation model;

[0016] Inputting the historical operation data set into the intermediate observation model, and verifying whether the parameter estimation result output by the intermediate observation model meets the preset accuracy range;

[0017] If so, the intermediate observation model is determined as the dynamic parameter observation model.

[0018] Preferably, the synchronous acquisition of the operating parameters of the flywheel energy storage array by the multi-source sensing unit includes:

[0019] establishing a communication link with a target flywheel unit, wherein the target flywheel unit is deployed at a preset node position of the energy storage array;

[0020] Continuously reading the real-time operation data stream of the target flywheel unit according to a preset sampling period, and marking a collection time stamp based on a timing feature of the real-time operation data stream;

[0021] According to the topological structure of the energy storage array, the real-time operation data streams of different nodes at the same timestamp are spatially synchronized to form a spatially associated operation parameter set.

[0022] Preferably, inputting the operating parameters into a preset dynamic parameter observation model includes:

[0023] Extracting abnormal fluctuation segments from the operating parameters, wherein the abnormal fluctuation segments are signal segments in which the parameter change rate exceeds a preset dynamic threshold within a continuous time window;

[0024] Generating a dynamic characteristic evaluation index based on the duration and change slope of the abnormal fluctuation segment;

[0025] The corresponding parameter estimation algorithm is dynamically selected according to the dynamic characteristic evaluation index, wherein the Kalman filtering algorithm is used for short-term high-frequency fluctuations and the least squares fitting algorithm is used for long-term low-frequency fluctuations.

[0026] Preferably, the method further comprises:

[0027] After completing the dynamic characteristic parameter estimation, performing data integrity check on the state characteristic vector;

[0028] If the check finds that the data missing rate exceeds the preset second threshold, the preset redundancy compensation model is triggered to perform priority compensation on the missing data, wherein the high-priority missing data is a parameter segment whose continuous missing duration exceeds the preset duration.

[0029] Preferably, the multi-model fusion algorithm includes the following fusion steps:

[0030] Constructing a distributed state observation model based on the node distribution of the energy storage array, wherein each node corresponds to the operating state of a flywheel unit;

[0031] Calculate the fusion weight coefficient based on the dynamic characteristics difference of adjacent nodes;

[0032] Combined with the time series correlation of the operating parameters, a joint fusion calculation is performed on the multi-node status.

[0033] Preferably, the method further comprises:

[0034] After the fusion calculation is completed, the fusion result is verified for consistency. The verification method includes comparing the deviation between the fused data and the actual data of the adjacent nodes.

[0035] If the deviation exceeds a preset third threshold, the fusion weight coefficient is readjusted and the fusion calculation is iterated until the deviation is less than the third threshold.

[0036] Preferably, using a preset error correction model to perform deviation correction on the state characteristic vector includes:

[0037] Classifying correction priority labels according to error types, wherein the correction priority labels include transient error categories and steady-state error categories;

[0038] Under each correction label, the correction sub-strategies are divided based on the error amplitude range;

[0039] The corrected SOC data is stored in an independent storage partition of the state database according to the priority tag.

[0040] Preferably, the method further comprises:

[0041] Modify the encryption verification rules of the priority tag according to the preset access rights configuration;

[0042] Upon receiving a data query request, verify whether the permission identifier provided by the requester matches the encryption rule of the target revision tag;

[0043] If there is a match, the data access channel corresponding to the correction tag is opened.

[0044] Preferably, the method further comprises optimizing the multi-model fusion algorithm in the following manner:

[0045] Calculating the fusion error distribution of the SOC estimation results under different working conditions;

[0046] Determining a weight adjustment amount of the fusion model according to the error distribution, wherein a high error working condition increases the fusion weight of adjacent nodes, and a low error working condition increases the fusion weight of historical data;

[0047] The multi-model fusion algorithm is iteratively optimized based on the weight adjustment amount until the fusion error rate is less than a preset fourth threshold.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The SOC estimation method for a flywheel energy storage array system provided by this invention exhibits significant advantages in multiple aspects, positively impacting the development and application of flywheel energy storage technology. To improve estimation accuracy, the present invention utilizes multi-source sensing units to simultaneously collect multiple operating parameters, such as flywheel speed, bus current, winding temperature, and mechanical vibration intensity, enabling comprehensive acquisition of system operating status information. Compared to traditional SOC estimation methods that rely solely on a single or a few parameters, multi-parameter acquisition significantly enriches the data dimension. Inputting these parameters into a dynamic parameter observation model that dynamically adjusts observation weights based on historical flywheel unit operating data enables more accurate estimation of the flywheel system's dynamic characteristic parameters. For example, under varying ambient temperatures, historical data can help the model automatically adjust the observation weights for winding temperature parameters, making the estimation of dynamic characteristic parameters more accurate. Utilizing a multi-model fusion algorithm, fusion coefficients are assigned based on the spatial correlation of each flywheel unit to generate a fused state feature vector, fully accounting for the interrelationships within the system. This avoids the estimation bias caused by traditional methods that ignore spatial correlations, significantly improving the accuracy of SOC estimation and bringing the estimated results closer to the true state of charge.

[0050] From the perspective of adapting to complex operating conditions, the present invention possesses strong robustness. When faced with complex operating conditions such as varying charge and discharge rates, ambient temperature fluctuations, and mechanical vibration, the dynamic parameter observation model can flexibly adjust its observation weights based on real-time collected operating parameters and historical data to adapt to the changing operating conditions. For operating parameters with abnormal fluctuations, the system dynamically selects appropriate parameter estimation algorithms (e.g., Kalman filtering for short-term, high-frequency fluctuations and least squares fitting for long-term, low-frequency fluctuations) by extracting abnormal fluctuation segments and generating dynamic characteristic evaluation indicators. This ensures accurate estimation of dynamic characteristic parameters in a variety of complex situations. In high-temperature environments, when short-term, high-frequency fluctuations in flywheel speed occur, the Kalman filtering algorithm can quickly and effectively process noise and interference, accurately estimating dynamic characteristic parameters and thus ensuring the reliability of SOC estimation.

[0051] In terms of data processing and system stability, the present invention has taken a series of effective measures. After completing the estimation of dynamic characteristic parameters, the state characteristic vector is checked for data integrity. When it is found that the data missing rate exceeds the preset threshold, the redundant compensation model is triggered to compensate for the missing data in priority. This ensures the integrity of the data, avoids estimation errors caused by missing data, and improves the stability and reliability of the system. After the fusion calculation is completed, the multi-model fusion algorithm verifies the consistency of the fusion result. If the deviation exceeds the preset threshold, the fusion weight coefficient is readjusted and the fusion calculation is iterated until the deviation is less than the threshold. This self-optimization and adjustment mechanism can effectively eliminate errors in the fusion process, ensure the accuracy of the fusion results, and further improve the reliability of SOC estimation.

[0052] In addition, the present invention also focuses on data management and security. A preset error correction model is used to correct deviations in the state feature vector, and the corrected results are stored in the state database in a time series. Correction priority tags are assigned based on error type, and correction sub-strategies are divided based on error amplitude ranges under different correction tags. Corrected SOC data is stored in independent storage partitions according to priority tags, making it convenient for users to quickly query and analyze different types of error data. By configuring encryption verification rules for correction priority tags, data access channels are only opened when the verification requester's permission identifier matches, effectively ensuring data security and preventing the leakage of sensitive information.

[0053] The multi-model fusion algorithm of the present invention also statistically analyzes the fusion error distribution of SOC estimation results under different operating conditions, determines weight adjustments based on the error distribution, and iteratively optimizes the algorithm until the fusion error rate falls below a preset threshold. This enables the algorithm to continuously adapt to system changes and continuously improve estimation accuracy, providing strong support for the long-term stable operation of flywheel energy storage array systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a working principle diagram of the SOC estimation method for a flywheel energy storage array system according to the present invention;

[0055] Figure 2 A diagram showing the steps for collecting operating parameters of a flywheel energy storage array;

[0056] Figure 3 Diagram of steps for inputting dynamic parameter observation model for operating parameters;

[0057] Figure 4 Figure 2 is a step diagram for processing the fusion results. DETAILED DESCRIPTION

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] See also Figures 1-4 The present invention provides a method for estimating the state of charge (SOC) of a flywheel energy storage array system, aiming to more accurately and efficiently estimate the SOC (State of Charge) of the flywheel energy storage array system. The following is a detailed implementation of this method.

[0060] The multi-source sensing units are used to synchronously collect the operating parameters of the flywheel energy storage array. In actual application scenarios, the flywheel energy storage array can be composed of multiple flywheel units distributed in different locations. Multi-source sensing units are deployed at key locations for real-time monitoring of operating parameters. The operating parameters include flywheel speed, bus current, winding temperature, and mechanical vibration intensity. For example, in a large energy storage power station, each flywheel unit is equipped with a high-precision speed sensor to measure the flywheel speed, a current sensor to obtain the bus current, a temperature sensor to monitor the winding temperature, and a vibration sensor to detect the mechanical vibration intensity. These sensors convert the collected analog signals into digital signals for subsequent processing.

[0061] The collected operating parameters are input into a preset dynamic parameter observation model to estimate the dynamic characteristic parameters of the flywheel system in real time. The dynamic parameter observation model is not fixed, but dynamically adjusts the observation weights based on the historical operating data of the flywheel unit. During the historical operating data accumulation phase, the system continuously records the operating parameters of the flywheel under different operating conditions and the corresponding true values of the dynamic characteristic parameters. Through in-depth analysis of these data, the model can automatically adjust the observation weights of each parameter according to the similarity between the current operating state and the historical data, thereby more accurately estimating the dynamic characteristic parameters. For example, when the system detects that the current operating condition is similar to that in a certain historical data segment, the model will increase the weight of the related parameters in that historical data segment, making the estimation result closer to the actual situation.

[0062] The estimated dynamic characteristic parameters are input into a preset multi-model fusion algorithm to generate a fused state feature vector. The multi-model fusion algorithm assigns fusion coefficients based on the spatial correlation of each flywheel unit. In the flywheel energy storage array, there is a certain spatial relationship between flywheel units at different locations, and units close to each other may have higher correlation in operating state. The algorithm takes into account this spatial correlation and assigns different fusion coefficients to different flywheel units. For example, for two adjacent flywheel units, since they are close in physical location and are similarly affected by environmental factors, the algorithm assigns them relatively high fusion coefficients, so that their information can be more fully fused when generating the state feature vector.

[0063] The state eigenvector is corrected for deviations using a preset error correction model to produce an SOC estimate. The error correction model identifies and corrects any potential deviations in the state eigenvector. For example, deviations in the state eigenvector caused by sensor errors or other factors are adjusted according to preset rules within the correction model. Once the corrected results are obtained, they are stored in a time series in the state database. In the state database, data is stored in chronological order to facilitate subsequent query and analysis. Each data entry includes an accurate time stamp and the corresponding SOC estimate, allowing users to easily monitor the energy storage system status at any given moment.

[0064] The specific implementation of the present invention is further described in detail below through 6 examples.

[0065] Example 1:

[0066] When constructing a dynamic parameter observation model, the first step is to obtain a historical operating data set. The source of the historical operating data set is very critical, and it is usually accumulated during the long-term operation of the flywheel energy storage array. These data record the operation of the flywheel under various working conditions, including operating parameters under different load conditions, ambient temperature, operating time and other factors, as well as the corresponding dynamic characteristic parameter true values. For example, in an actual energy storage system that has been operating for a year, the system continuously records operating parameters such as flywheel speed, bus current, winding temperature, mechanical vibration intensity, etc. every hour, and at the same time obtains the corresponding dynamic characteristic parameter true values ​​through high-precision measurement equipment. These data together constitute the historical operating data set.

[0067] After acquiring the historical operating data set, training subsets are divided based on the error type and error amplitude of the dynamic characteristic parameters annotated in the data. Different error types may represent different operating abnormalities, such as errors caused by sensor failure or errors caused by wear of mechanical components. The error amplitude reflects the severity of the error. For example, the error type is divided into several categories, such as sensor error and mechanical error. Under the sensor error category, data with an error amplitude in the range of 0-5% is divided into one training subset, data with an error amplitude in the range of 5%-10% is divided into another training subset, and so on. Each training subset corresponds to a dynamic characteristic scenario.

[0068] Next, the initial observation model is trained in parallel using the training subsets. During the training process, a large amount of computing resources will be used to process multiple training subsets at the same time. Taking the common neural network model as an example, an independent neural network structure is constructed for each training subset, and these neural networks are trained at the same time. The goal of training is to make the parameter estimation error rate of the initial observation model for each scenario less than or equal to the preset first threshold. The preset first threshold is set according to the actual application needs and system accuracy requirements. For example, in some energy storage systems with high accuracy requirements, the first threshold may be set to 3%. When, after multiple iterative training, the parameter estimation error rate of the model for the scenarios corresponding to all training subsets reaches or is lower than this threshold, the training is stopped to obtain the intermediate observation model.

[0069] After obtaining the intermediate observation model, the historical operation data set is input into the model again, and it is verified whether the parameter estimation results output by the intermediate observation model meet the preset accuracy range. The setting of the preset accuracy range is related to the first threshold, but it more comprehensively considers the error requirements in practical applications. For example, the preset accuracy range may require that under different working conditions, the error between the parameter estimation result and the true value must not only be within a certain percentage range, but also meet certain absolute value limits. If the output result of the intermediate observation model meets this preset accuracy range, then the intermediate observation model is determined as the final dynamic parameter observation model.

[0070] Example 2:

[0071] When synchronously collecting the operating parameters of a flywheel energy storage array through multi-source sensing units, a communication link with the target flywheel unit must first be established. In actual flywheel energy storage arrays, the target flywheel unit is deployed at a preset node position in the energy storage array. The selection of the preset node location is usually based on the overall layout of the system and performance optimization considerations, such as selecting a key node on the power transmission path or a location with good heat dissipation conditions. Taking a distributed energy storage system as an example, a communication connection with the target flywheel unit is established through a wireless communication module (such as ZigBee, Bluetooth, etc.) or a wired communication line (such as optical fiber, Ethernet cable, etc.). In the process of establishing the communication link, operations such as device address identification and communication protocol negotiation are required to ensure the stability and accuracy of communication.

[0072] After the communication link is established, the real-time operating data stream of the target flywheel unit is continuously read according to the preset sampling period. The determination of the preset sampling period should take into account multiple factors, including the system's real-time requirements, data processing capabilities, and the response speed of the sensor. For example, in an energy storage system with high real-time requirements, the preset sampling period may be set to 100 milliseconds. When reading the real-time operating data stream, the acquisition timestamp is marked based on the timing characteristics of the real-time operating data stream. The timestamp marking can accurately record the acquisition time of each set of data, providing a time reference for subsequent data processing and analysis. For example, a high-precision clock chip is used to add a timestamp accurate to the microsecond level to each acquired data.

[0073] Finally, based on the topological structure of the energy storage array, the real-time operating data streams of different nodes at the same timestamp are spatially aligned to form a spatially correlated set of operating parameters. The topological structure of the energy storage array can take many forms, such as series, parallel, star, and mesh. Taking the star topology as an example, the central node is connected to each edge node (i.e., the node where the target flywheel unit is located). When performing spatial synchronization alignment, the connection relationship between each node is determined based on the topological structure, and the real-time operating data streams collected from different nodes at the same timestamp are integrated to ensure that each data point corresponds to the operating parameters at different locations at the same time, thereby forming a spatially correlated set of operating parameters, providing a comprehensive data foundation for subsequent analysis and processing.

[0074] Example 3:

[0075] When the operating parameters are input into the preset dynamic parameter observation model, the abnormal fluctuation segments in the operating parameters are first extracted. The abnormal fluctuation segment refers to the signal segment in which the parameter change rate exceeds the preset dynamic threshold within a continuous time window. The setting of the preset dynamic threshold needs to be determined based on the parameter change range of the flywheel energy storage array under normal operating conditions. For example, for the parameter of flywheel speed, its speed changes relatively smoothly during normal operation. According to the statistical analysis of historical data, it is determined that the speed change of more than 100 revolutions per minute within 1 second is an abnormal fluctuation, then this 100 revolutions per minute is the preset dynamic threshold. In actual operation, the operating parameters are monitored by sliding the time window. Once it is found that the parameter change rate exceeds the preset dynamic threshold, the signal segment within this time period is marked as an abnormal fluctuation segment.

[0076] A dynamic characteristic evaluation index is generated based on the duration and slope of the abnormal fluctuation segment. The duration reflects the length of the abnormal fluctuation, while the slope reflects the severity of the parameter change. For example, if an abnormal fluctuation segment lasts for 5 seconds and has a slope of 200 revolutions per minute², these data will be used to calculate the dynamic characteristic evaluation index. The dynamic characteristic evaluation index can be calculated using a weighted summation method, for example, dynamic characteristic evaluation index = 0.6 × duration + 0.4 × slope. The weights of 0.6 and 0.4 here are determined based on practical experience and extensive experimental data to balance the impact of duration and slope on the evaluation index.

[0077] The corresponding parameter estimation algorithm is dynamically selected according to the dynamic characteristic evaluation index. For short-term high-frequency fluctuations, the Kalman filter algorithm is used. Short-term high-frequency fluctuations usually mean that the system is affected by sudden and rapidly changing interference factors. The Kalman filter algorithm can use the system's state equation and observation equation to make the best estimate of the system state through two steps of prediction and update. In this invention, it is assumed that the state equation of the system is ,in represents the state vector of the system at time k, It is the state transition matrix, which is used to describe the state transition relationship of the system from time k-1 to time k. is the control input matrix, is the control input vector, is the process noise vector, representing the uncertainty factor within the system. The observation equation is , is the observation vector at time k, is the observation matrix, which is used to map the system state to the observation space. is the observation noise vector, reflecting the error in the observation process. The Kalman filter algorithm can accurately estimate the system state in the presence of noise through continuous iterative calculations.

[0078] For long-term low-frequency fluctuations, the least squares fitting algorithm is used. Long-term low-frequency fluctuations generally indicate that the system has a relatively slow, long-term trend of change. The principle of the least squares fitting algorithm is to find the best function matching the data by minimizing the sum of squares of errors. Suppose we have a set of observation data , , the function to be fitted is , by minimizing To determine the parameters and The best fitting curve is obtained to estimate the dynamic characteristic parameters of the system.

[0079] Example 4:

[0080] After dynamic characteristic parameter estimation is complete, the state eigenvector needs to be checked for data integrity. This ensures that the data in the state eigenvector is not missing or corrupted, ensuring the accuracy of subsequent calculations and analysis. In practical applications, data loss may occur due to communication failures, sensor malfunctions, and other factors. For example, during data transmission, the flywheel speed data at a certain moment may not be successfully transmitted to the processing unit due to signal interference, resulting in missing data in the state eigenvector.

[0081] The data loss rate is calculated as follows: Data loss rate = Number of missing data / Total data × 100%. The preset second threshold is set based on the system's requirements for data integrity. For example, in some application scenarios with high data integrity requirements, the second threshold may be set to 5%. When verification finds that the data loss rate exceeds the preset second threshold, the preset redundancy compensation model is triggered to prioritize the missing data. In the redundancy compensation model, high-priority missing data is defined as a parameter segment with a continuous missing duration exceeding the preset duration. The setting of the preset duration should also be determined based on actual application requirements. For example, in an energy storage system with high real-time requirements, the preset duration may be set to 10 seconds.

[0082] For high-priority missing data, backup sensor data is used for compensation. If there is no backup sensor data, interpolation calculation is performed based on historical data and data from adjacent moments. For example, for continuously missing flywheel speed data, the speed data from the previous moment and the next moment can be used to estimate it using linear interpolation. Assume that the speed at the previous moment is , the speed at the next moment is , the missing time is , the previous moment is , the next moment is , then the estimated speed at the missing moment is For low-priority missing data, statistical averaging can be used to compensate for it, for example, calculating the average value of data under other similar working conditions in the same time period to fill in the missing values.

[0083] Example 5:

[0084] The specific fusion steps of the multi-model fusion algorithm are as follows: First, a distributed state observation model is constructed based on the node distribution of the energy storage array. In an actual flywheel energy storage array, each node corresponds to the operating status of a flywheel unit. For example, in an energy storage array consisting of 10 flywheel units, it is divided into 10 nodes, each equipped with corresponding sensors to monitor the operating parameters of the flywheel unit, such as speed, current, temperature, etc. Based on the operating parameters of each node, independent state observation models are constructed. These models can reflect the changes in the operating status of each flywheel unit in real time.

[0085] Next, the fusion weight coefficient is calculated based on the dynamic characteristic differences of adjacent nodes. The dynamic characteristic differences of adjacent nodes can be measured by calculating the similarity of their operating parameters. For example, the Euclidean distance of parameters such as flywheel speed and bus current between two adjacent nodes is calculated. The Euclidean distance calculation formula is ,in and The two nodes are The value of the operating parameter, is the number of operating parameters. The smaller the Euclidean distance, the more similar the dynamic characteristics of the two nodes are, and the higher the fusion weight coefficient is. Assume that the Euclidean distance between node A and node B is , set a reference distance , fusion weight coefficient , which ensures that the smaller the distance, the larger the weight coefficient, and the weight coefficient is between 0-1.

[0086] Finally, the time series correlation of the operating parameters is combined to perform a joint fusion calculation on the multi-node status. The time series correlation of the operating parameters can be determined by calculating the autocorrelation function. The calculation formula of the autocorrelation function is: ,in is time series data, is the data length, is the delay time. The correlation strength of data at different moments is determined based on the value of the autocorrelation function. During joint fusion calculations, time series data with stronger correlations are given higher weight. For example, when calculating the fused state feature vector at a given moment, the weights of data from the previous and next moments with strong correlations are increased, ensuring that the fusion result better reflects the actual state trends of the system.

[0087] After the fusion calculation is completed, the consistency of the fusion result is verified. The verification method includes comparing the deviation between the fused data and the actual data of the adjacent nodes. The actual data of the adjacent nodes can be obtained in real time through sensors. For example, the difference between the fused flywheel speed and the flywheel speed actually measured by the adjacent nodes is calculated. If the difference exceeds the preset third threshold, it is considered that the fusion result has a deviation. The preset third threshold is set according to the system's requirements for accuracy. For example, in an energy storage system with high accuracy requirements, the third threshold may be set to 5 rpm. If the deviation exceeds the preset third threshold, the fusion weight coefficient is readjusted and the fusion calculation is iterated until the deviation is less than the third threshold. When adjusting the fusion weight coefficient, optimization algorithms such as gradient descent can be used to gradually find the optimal combination of weight coefficients to make the fusion result more accurate.

[0088] Example 6:

[0089] When using a preset error correction model to correct the state eigenvector's deviation, the system first prioritizes the correction based on the error type. Error types are categorized as transient and steady-state. Transient errors are typically caused by sudden disturbances, such as momentary voltage fluctuations or external shocks. These errors are characterized by their short duration and sudden nature. Steady-state errors, which persist during the system's long-term stable operation, may be caused by inherent sensor bias, model inaccuracies, and other factors.

[0090] Within each correction label, correction sub-strategies are divided based on the error amplitude range. For example, in the transient error category, when the error amplitude is within the 0-10% range, a simple filtering algorithm is used for correction; when the error amplitude is within the 10%-20% range, a more complex model prediction correction is performed by combining historical data and current operating status. In the steady-state error category, small error amplitudes can be corrected through regular sensor calibration; large error amplitudes require readjustment of model parameters.

[0091] The corrected SOC data is stored in separate storage partitions of the state database according to priority tags. Within the state database, separate storage areas are allocated for transient error-corrected data and steady-state error-corrected data. This allows users to quickly query and analyze data based on their needs. For example, if a user wants to understand the system's SOC changes under sudden interference, they can directly retrieve the data from the transient error-corrected data storage partition. If a user is concerned about the long-term stability of the system, they can query the data from the steady-state error-corrected data storage partition.

[0092] In addition, encryption verification rules for correction priority tags are configured based on preset access rights. In a multi-user energy storage system management platform, different users may have different access rights. For example, the system administrator has the highest authority and can access data of all correction tags; ordinary operators may only be able to access steady-state error correction data. When a data query request is received, it is verified whether the permission identifier provided by the requester matches the encryption rules of the target correction tag. The permission identifier can be a user's account password, digital certificate, etc. If it matches, the data access channel for the corresponding correction tag is opened to ensure data security and confidentiality.

[0093] At the same time, the multi-model fusion algorithm was optimized as follows: the fusion error distribution of the SOC estimation results under different operating conditions was statistically analyzed. These different operating conditions include varying load conditions, ambient temperature, and operating time. For example, under high-load conditions, the error between the SOC estimation result and the actual value was statistically analyzed; under low-temperature ambient conditions, the corresponding error was also statistically analyzed. Through extensive experiments and the accumulation of actual operating data, the fusion error distribution under different operating conditions was determined.

[0094] The weight adjustment of the fusion model is determined based on the error distribution. For high error conditions, the fusion weight of adjacent nodes is increased. This is because under high error conditions, it may mean that the measurement data of a single node is greatly disturbed. In this case, increasing the fusion weight of adjacent nodes can improve the estimation accuracy with the help of relatively reliable data from adjacent nodes. Assuming that under high error conditions, the original adjacent node fusion weight is , set the weight adjustment ratio to ( ), the adjusted adjacent node fusion weight is .

[0095] For low error conditions, increase the fusion weight of historical data. Low error conditions indicate that the current system is relatively stable, and historical data can reflect the normal operation of the system. Increasing its fusion weight helps to improve the estimation accuracy by utilizing accumulated experience. Suppose the original historical data fusion weight is , the weight adjustment ratio is ( ), the adjusted historical data fusion weight is .

[0096] Based on the aforementioned weight adjustments, the multi-model fusion algorithm is iteratively optimized until the fusion error rate falls below a preset fourth threshold. During each iteration, the multi-model fusion calculation is re-performed based on the adjusted weights, and the fusion error rate is calculated again. If the fusion error rate is still greater than the preset fourth threshold, the weights are adjusted again according to the aforementioned rules and the next iteration is performed. If the fusion error rate is less than the fourth threshold, the optimization is considered complete. The preset fourth threshold is determined based on the estimation accuracy requirements of the actual application. For example, in energy storage systems with extremely high accuracy requirements, the fourth threshold may be set to 2%.

[0097] In actual operation, a large amount of operating data under different operating conditions is first collected, including operating parameters such as flywheel speed, bus current, winding temperature, mechanical vibration intensity, and the corresponding actual SOC values. This data is used to calculate the initial fusion error distribution and determine the initial weight adjustment amount. The process of weight adjustment, fusion calculation, and error rate statistics is then repeated to gradually optimize the multi-model fusion algorithm, making the SOC estimation results increasingly accurate. This optimization process can adapt to different operating conditions, improve the overall performance of SOC estimation for flywheel energy storage array systems, and provide strong support for the efficient management and reliable operation of energy storage systems.

[0098] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0099] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A SOC estimation method for a flywheel energy storage array system, characterized in that: include: Synchronously collecting operating parameters of the flywheel energy storage array through a multi-source sensing unit, wherein the operating parameters include flywheel speed, bus current, winding temperature, and mechanical vibration intensity; Inputting the operating parameters into a preset dynamic parameter observation model to estimate the dynamic characteristic parameters of the flywheel system in real time, wherein the dynamic parameter observation model dynamically adjusts the observation weight based on historical operating data of the flywheel unit; Inputting the dynamic characteristic parameters into a preset multi-model fusion algorithm to generate a fused state feature vector, wherein the multi-model fusion algorithm allocates fusion coefficients according to the spatial correlation of each flywheel unit; Using a preset error correction model to correct the deviation of the state characteristic vector to obtain an SOC estimation result, and storing the corrected result in a state database in a time series; The steps of constructing the dynamic parameter observation model include: Acquire a historical operation data set, wherein each data in the historical operation data set is marked with an error type and an error amplitude of a dynamic characteristic parameter; Divide the training subsets based on the error type and error amplitude, each training subset corresponding to a dynamic characteristic scenario; The initial observation model is trained in parallel using the training subset until the parameter estimation error rate of the initial observation model for each scene is less than or equal to a preset first threshold, thereby stopping the training and obtaining an intermediate observation model; Inputting the historical operation data set into the intermediate observation model, and verifying whether the parameter estimation result output by the intermediate observation model meets the preset accuracy range; If so, the intermediate observation model is determined as the dynamic parameter observation model; The multi-model fusion algorithm includes the following fusion steps: Constructing a distributed state observation model based on the node distribution of the energy storage array, wherein each node corresponds to the operating state of a flywheel unit; Calculate the fusion weight coefficient based on the dynamic characteristics difference of adjacent nodes; Combined with the time series correlation of the operating parameters, a joint fusion calculation is performed on the multi-node status.

2. The SOC estimation method for a flywheel energy storage array system according to claim 1, characterized in that: The synchronous acquisition of the operating parameters of the flywheel energy storage array by the multi-source sensing unit includes: establishing a communication link with a target flywheel unit, wherein the target flywheel unit is deployed at a preset node position of the energy storage array; Continuously reading the real-time operation data stream of the target flywheel unit according to a preset sampling period, and marking a collection time stamp based on a timing feature of the real-time operation data stream; According to the topological structure of the energy storage array, the real-time operation data streams of different nodes at the same timestamp are spatially synchronized to form a spatially associated operation parameter set.

3. The SOC estimation method for a flywheel energy storage array system according to claim 1, characterized in that: Inputting the operating parameters into a preset dynamic parameter observation model includes: Extracting abnormal fluctuation segments from the operating parameters, wherein the abnormal fluctuation segments are signal segments in which the parameter change rate exceeds a preset dynamic threshold within a continuous time window; Generating a dynamic characteristic evaluation index based on the duration and change slope of the abnormal fluctuation segment; The corresponding parameter estimation algorithm is dynamically selected according to the dynamic characteristic evaluation index, wherein the Kalman filtering algorithm is used for short-term high-frequency fluctuations and the least squares fitting algorithm is used for long-term low-frequency fluctuations.

4. The SOC estimation method for a flywheel energy storage array system according to claim 3, characterized in that: The method further comprises: After completing the dynamic characteristic parameter estimation, performing data integrity check on the state characteristic vector; If the check finds that the data missing rate exceeds the preset second threshold, the preset redundancy compensation model is triggered to perform priority compensation on the missing data, wherein the high-priority missing data is a parameter segment whose continuous missing duration exceeds the preset duration.

5. The SOC estimation method for a flywheel energy storage array system according to claim 1, characterized in that: The method further comprises: After the fusion calculation is completed, the fusion result is verified for consistency. The verification method includes comparing the deviation between the fused data and the actual data of the adjacent nodes. If the deviation exceeds a preset third threshold, the fusion weight coefficient is readjusted and the fusion calculation is iterated until the deviation is less than the third threshold.

6. The SOC estimation method for a flywheel energy storage array system according to claim 1, characterized in that: Correcting the deviation of the state characteristic vector using a preset error correction model includes: Classifying correction priority labels according to error types, wherein the correction priority labels include transient error categories and steady-state error categories; Under each correction label, the correction sub-strategies are divided based on the error amplitude range; The corrected SOC data is stored in an independent storage partition of the state database according to the priority tag.

7. The SOC estimation method for a flywheel energy storage array system according to claim 6, characterized in that: The method further comprises: Modify the encryption verification rules of the priority tag according to the preset access rights configuration; Upon receiving a data query request, verify whether the permission identifier provided by the requester matches the encryption rule of the target revision tag; If there is a match, the data access channel corresponding to the correction tag is opened.

8. The SOC estimation method for a flywheel energy storage array system according to claim 1, characterized in that: The method further includes optimizing the multi-model fusion algorithm in the following manner: Calculating the fusion error distribution of the SOC estimation results under different working conditions; Determining a weight adjustment amount of the fusion model according to the error distribution, wherein a high error working condition increases the fusion weight of adjacent nodes, and a low error working condition increases the fusion weight of historical data; The multi-model fusion algorithm is iteratively optimized based on the weight adjustment amount until the fusion error rate is less than a preset fourth threshold.

Images