Data center transient voltage treatment flywheel energy storage system

Abstract

The invention relates to the technical field of voltage governance, in particular to a data center transient voltage governance flywheel energy storage system which comprises a disturbance moment identification module, a load sudden change judgment module, a capacity correction distribution module, a scheduling sequence adjustment module and a state change tracking module. And analyzing the voltage curve of the flywheel energy storage monitoring node in each time period. According to the method, the coupling relation between voltage disturbance and load change is synchronously captured through multi-dimensional data, the influence of the server running dynamic state on the flywheel energy state is excavated in real time, the energy distribution structure is adjusted in combination with the friction loss trend, and the flywheel priority is dynamically optimized through load proportion fluctuation and the main loop state. According to the method, node energy fluctuation rapid screening and change tracking are realized, the response strategy of the flywheel energy storage system to short-time voltage disturbance is refined, the pertinence and elasticity of energy storage scheduling in a data center operation scene are improved, and the stability level and risk protection capability of the whole power supply system are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of voltage management technology, and in particular to a flywheel energy storage system for transient voltage management in a data center. Background Art

[0002] The field of voltage management involves monitoring, regulating and controlling the voltage level in the power system, aiming to ensure the voltage stability and quality of the power supply. This technical field mainly includes the detection and analysis of transient disturbances of voltage fluctuations in the power network, and the adoption of energy storage compensation and reactive power regulation and other means to ensure the reliable operation of the power supply system in various key power consumption scenarios. Among them, the flywheel energy storage system for transient voltage management of data centers refers to the short-term voltage changes caused by load fluctuations or grid disturbances during the operation of the data center. It uses a high-speed rotating flywheel to store electrical energy in the form of mechanical energy, and quickly releases energy to achieve instant compensation for voltage when voltage transients occur, realize rapid two-way energy conversion and power injection, and complete the management of transient voltage fluctuations in the data center.

[0003] In the process of voltage fluctuation judgment, existing technologies have difficulty distinguishing the linkage effects of multi-node load responses and energy status. It is difficult to adapt to variable working conditions using a single parameter or static strategy. During large-scale load switching or concurrent power supply, some energy storage units are not scheduled in time or energy regulation is delayed, which can easily lead to local energy shortages and equipment status deviations, affecting the reliability of short-term voltage management and limiting the dynamic adaptability of the data center power supply system. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a flywheel energy storage system for transient voltage management in data centers.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a flywheel energy storage system for transient voltage control in a data center, the system comprising: The disturbance moment identification module analyzes the voltage curves of the flywheel energy storage monitoring node in each time period based on the data center voltage change rate and amplitude, compares the synchronization of rate and amplitude changes, determines the correlation between the fluctuation moment and the operating status of the flywheel energy storage system, and obtains the timing characteristics of the fluctuation event; The load mutation determination module determines the flywheel energy storage operation changes in the server cluster log based on the timing characteristics of the fluctuation event, analyzes the instantaneous power changes, compares the load switching rate and the number of concurrent energy supply nodes before and after the disturbance, identifies the critical interval, and obtains the load disturbance critical interval; The capacity correction distribution module compares the current speed of the flywheel with the trend of the mechanical friction rate based on the load disturbance key interval, determines the speed and friction coordination between the electric bearing and the energy, adjusts the available energy distribution, calculates the capacity change, and obtains the energy storage distribution correction feature; The scheduling sequence adjustment module compares the total operating power of the main circuit with the flywheel data based on the energy storage allocation correction characteristics, analyzes the changes in the server area load ratio caused by voltage offset, determines the impact of parameter changes on the scheduling sequence, and obtains the energy storage priority ranking index.

[0006] The present invention has the following improvements: the fluctuation event timing characteristics include the event occurrence time, fluctuation distribution characteristics, and associated marks; the load disturbance key interval includes interval time information, load disturbance type, and disturbance event identifier; the energy storage allocation correction characteristics include corrected capacity distribution, energy storage dynamic coefficient, and energy distribution structure; and the energy storage priority ranking indicators include ranking order, priority number, and allocation weight.

[0007] The present invention is improved in that the disturbance moment identification module includes: The rate fluctuation identification submodule analyzes the voltage curves recorded by the monitoring nodes based on the voltage change rate and amplitude of the data center, determines the sudden changes in the voltage change rate within a continuous period, and combines the node distribution characteristics to screen the concentrated time periods where fluctuations occur. It summarizes the related change intervals and obtains the rate fluctuation interval distribution; The amplitude offset identification submodule calls the rate fluctuation interval distribution, analyzes the voltage data of each node in the corresponding time period, compares the difference between the data and the rated voltage, selects the time segment with continuous voltage fluctuation and critical amplitude change, and obtains the amplitude offset section parameter; The fluctuation feature extraction submodule analyzes the overlapping time segments according to the amplitude offset segment parameters, determines the correlation between the time segments and the start and stop moments of the flywheel energy storage system, identifies the associated fluctuation events, and obtains the time series characteristics of the fluctuation events.

[0008] The present invention is improved in that the load mutation determination module includes: The energy supply rate extraction submodule extracts the instantaneous energy supply power in each period before and after the disturbance based on the timing characteristics of the fluctuation event and the voltage change rate data collected by the server cluster log and the flywheel energy storage monitoring node, and compares the instantaneous energy supply power with the load switching rate in the same period to obtain the energy supply change rate characteristics; The node fluctuation quantification submodule monitors the changes in the number of concurrent energy supply nodes based on the energy supply change rate characteristics, calculates the fluctuation of the number of energy supply nodes in the differential disturbance stage, identifies the key time period with prominent parameter changes, and obtains the load disturbance key interval.

[0009] The present invention is improved in that the capacity correction allocation module includes: The coordination gradient analysis submodule analyzes the synchronization of the flywheel speed change and the mechanical friction rate based on the load disturbance key interval, compares the change gradients of the two in each period, determines whether the fluctuations of the two show a coordinated trend, and calculates the degree of fluctuation coordination in the difference section to obtain the coordination gradient amplitude; The energy distribution adjustment submodule analyzes the coordination of the flywheel energy storage unit in energy release according to the coordination gradient amplitude, compares the balance between energy sharing and dynamic power output among the energy supply nodes, adjusts the energy output sequence, and obtains energy distribution structure information; The energy storage distribution correction submodule calls the energy distribution structure information, analyzes the difference between the current speed of each flywheel and the upper limit speed, and obtains the energy storage distribution correction characteristics by combining the moment of inertia, friction coefficient and power release difference.

[0010] The present invention is improved in that the scheduling sequence adjustment module includes: The main circuit comparison submodule analyzes the synchronous changes of the total operating power curve of the distribution main circuit and the output power curve of the flywheel energy storage system based on the energy storage distribution correction characteristics, determines the synchronous offset of the power curve within the monitoring period, selects the cycle segments with key synchronization characteristics, and obtains the power offset cycle segments; The load ratio analysis submodule compares the energy distribution of the flywheel energy storage system with the current load status of each server area based on the power offset period, calculates the mean square error of the power change rate of each server area, identifies the number of server areas with critical change rate amplitudes, and obtains the load fluctuation normalization amount; The scheduling sequence generation submodule determines the energy output efficiency and the remaining energy ratio of the flywheel energy storage device according to the load fluctuation normalization amount, calculates the scheduling sequence difference in the target period, adjusts the scheduling priority, and obtains the energy storage priority ranking index.

[0011] The present invention is improved in that the system further comprises: The state change tracking module analyzes the change in the number of the first flywheel in the order according to the energy storage priority ranking index, compares the change amplitude of the residual energy in consecutive cycles, determines the stability of the energy change, identifies the key number, marks it as a state jump object, updates the data list, and obtains the state jump tracking data; The state transition tracking data includes a transition flag, number information, and a change history record.

[0012] The present invention is improved in that the state change tracking module includes: The energy fluctuation analysis submodule analyzes the change in the remaining available energy of the flywheel number in consecutive cycles based on the energy storage priority ranking index, compares the energy fluctuation amplitude between each monitoring cycle, determines whether there is an abnormal deviation in the change, and generates the periodic energy fluctuation amplitude; The stability judgment submodule calls the periodic energy fluctuation amplitude, judges its fluctuation persistence in the time series, analyzes whether there are regular fluctuations or sudden changes in the energy change process, identifies the coherence characteristics of the energy change, and obtains the energy response stability trend; The number jump screening submodule analyzes the fluctuation coherence reflected by the stable trend of the energy response, screens the data sequence with prominent changes in the flywheel number, determines whether the fluctuation characteristics meet the state jump characteristics, and synchronously updates the data list to obtain the state jump tracking data.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, the coupling relationship between voltage disturbances and load changes is captured synchronously through multi-dimensional data, the impact of server operation dynamics on the flywheel energy state is mined in real time, the energy distribution structure is adjusted in combination with the friction loss trend, and the flywheel priority is dynamically optimized through load ratio fluctuations and main circuit status. Rapid screening and change tracking of node energy fluctuations are achieved, and the response strategy of the flywheel energy storage system to short-term voltage disturbances is refined. The pertinence and flexibility of energy storage scheduling in data center operation scenarios are improved, and the stability level and risk protection capability of the overall power supply system are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a system flow chart of the present invention; Figure 2 This is a flow chart of the disturbance moment identification module in the present invention; Figure 3 This is a flow chart of the load mutation determination module in the present invention; Figure 4 This is a flow chart of the capacity correction allocation module in the present invention; Figure 5 This is a flow chart of the scheduling sequence adjustment module in the present invention; Figure 6 This is a flow chart of the state change tracking module in the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0016] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction, and therefore should not be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0017] Example See also Figure 1 The present invention provides a technical solution: a flywheel energy storage system for transient voltage control in a data center, comprising: The disturbance moment identification module analyzes the voltage curves of the flywheel energy storage monitoring nodes in different time periods based on the voltage change rate and amplitude of the data center. It compares whether the rate change time and amplitude offset time occur at the same time, determines the correlation between each fluctuation moment and the operating status of the flywheel energy storage system, and summarizes the fluctuation event information to obtain the fluctuation event time series characteristics. The load mutation determination module determines the operational changes of the flywheel energy storage in the server cluster logs during the corresponding period based on the timing characteristics of the fluctuation event. It analyzes the change process of the instantaneous energy supply power, compares the changes in the load switching rate and the number of concurrent energy supply nodes before and after the disturbance, identifies the key time periods with prominent changes in various parameters, and obtains the key interval of the load disturbance. The capacity correction allocation module compares the change trend of the flywheel's current speed and the mechanical friction rate based on the load disturbance key interval, determines the speed and friction coordination between the electric bearing and the energy, adjusts the flywheel's available energy distribution, and calculates the capacity change parameters to obtain the energy storage distribution correction characteristics; The scheduling sequence adjustment module, based on the energy storage allocation correction characteristics, compares the total operating power of the distribution main circuit with the flywheel data, analyzes the changes in the server area load ratio caused by the instantaneous voltage offset, determines the impact of various parameter changes in continuous cycles on the scheduling sequence of the flywheel energy storage system, adjusts the scheduling sequence, and obtains the energy storage priority ranking index; The state change tracking module analyzes the changes in the flywheel number corresponding to the first item in the sorting according to the energy storage priority ranking index, compares the change amplitude of the remaining available energy of the number in consecutive cycles, judges the stability of the energy parameter change process, identifies the key number of the change, marks it as the state jump object, and synchronously updates it to the data list to obtain the state jump tracking data.

[0018] The timing characteristics of fluctuation events include the time of event occurrence, fluctuation distribution characteristics, and associated marks. The key intervals of load disturbances include interval time information, load disturbance type, and disturbance event identification. The energy storage allocation correction characteristics include the corrected capacity distribution, energy storage dynamic coefficient, and energy distribution structure. The energy storage priority sorting indicators include sorting order, priority number, and allocation weight. The state jump tracking data includes jump flags, number information, and change history records.

[0019] In Module 1, the monitoring node refers to the voltage collection point installed on the flywheel energy storage system or its connected busbar. The node collects real-time voltage data to determine whether there are abnormal voltage changes in the system. The voltage curve refers to the record of voltage changes over time collected by each monitoring node over a period of time, which can be used to observe voltage fluctuations and trends. The rate change time refers to the specific moment in the voltage curve when the voltage change rate fluctuates significantly, such as the moment when the voltage suddenly rises or falls. The amplitude deviation time refers to the moment in the voltage curve when the voltage amplitude deviates significantly from the rated value, indicating abnormal voltage fluctuations. The fluctuation moment refers to the set of all moments corresponding to the above rate change or amplitude deviation, which is used to determine subsequent disturbance events.

[0020] In module 2, the change process of instantaneous energy supply power refers to the change in power output or compensation by each node (parallel power supply point) of the flywheel energy storage system per unit time, reflecting the response capability to instantaneous load changes; the load switching rate refers to the frequency with which the load (such as servers, air conditioners, etc.) in the data center switches between power-on and offline states, affecting the overall power fluctuation; the concurrent energy supply node refers to the number of multiple flywheel energy storage units or branches actually participating in power supply (energy storage discharge) at the same time, reflecting the system's concurrent energy supply capability.

[0021] In Module 3, the trend of change refers to the direction and magnitude of change in the flywheel's rotational speed or mechanical friction rate over a period of time, and is an important basis for determining the energy storage status. Electric bearings and energy: Electric bearings: Bearing systems that support high-speed flywheel rotation while minimizing friction losses. Energy: This refers specifically to the kinetic energy stored by the flywheel through rotation. The so-called "between electric bearings and energy" actually refers to the correspondence between the flywheel's operating conditions (bearing friction) and the kinetic energy it can store and release. Speed ​​and friction coordination refers to the match between the flywheel's current speed and the bearing friction conditions. If friction is excessive but the speed does not decrease, this is an anomaly. Under normal circumstances, increased friction will cause the speed to decrease, and the two should change synchronously.

[0022] In Module 4, the change in zone load ratio refers to the change in the proportion of power load in each distribution zone of the data center due to factors such as voltage disturbances and equipment switching. It is an important data that affects the scheduling and allocation of flywheel energy storage. The impact of scheduling order refers to the impact of multiple parameter changes (such as energy storage capacity, load ratio, voltage change, etc.) on the ranking weight of the flywheel energy storage system priority scheduling (that is, which flywheel has priority in compensation).

[0023] In Module 5, the change in flywheel number refers to the change in parameters such as the operating status and energy status of a specific flywheel number during the scheduling and energy release processes; the available energy change amplitude refers to the difference in the energy that can be released by a flywheel device between two monitoring cycles, which is used to determine the health of its energy storage or the severity of the change; the state jump object refers to the flywheel number whose operating status has suddenly changed, which is screened out through the above analysis, and its behavior and status data need to be tracked in detail.

[0024] See also Figure 2 , the disturbance moment identification module includes: The rate fluctuation identification submodule analyzes the voltage curves recorded by the monitoring nodes based on the voltage change rate and amplitude of the data center, determines the sudden changes in the voltage change rate within a continuous period, and combines the node distribution characteristics to screen the concentrated time periods where fluctuations occur. It summarizes the related change intervals and obtains the rate fluctuation interval distribution; The voltage time series is collected from the monitoring nodes installed at the access point of the flywheel energy storage system or distributed on the key power bus. The sampling period is usually set to 10 milliseconds. Then, by traversing the continuous voltage data of each node, the voltage values ​​at each two adjacent moments are divided by the time interval to obtain the instantaneous rate of change of the voltage. All the results are compared to see if there is a rate mutation greater than the set reference value. The reference value can be set to the top 5% of the rate data in the past month, for example, 5 volts per second. If the rate of the current data is greater than the reference value, it is determined to be a mutation point. After that, a set of mutation time points is independently formed at each node. Then, a sliding time window statistical analysis method is used. For example, every 0.2 seconds is a window. The number of mutation points in the window is counted to see if it exceeds 3. If so, it is marked as a concentrated fluctuation time period. Then, the fluctuation time periods in all nodes are compared horizontally. If the interval between adjacent fluctuation time periods is less than 0.1 seconds, they are merged into a longer fluctuation interval. Finally, the time periods with concentrated mutations are summarized to form a time period set of rate fluctuation intervals. For example, at 12:01:34 noon, 5 nodes frequently experience rate mutations in the same second, concentrated in the same time window. This window will be identified as a rate fluctuation interval distribution result.

[0025] The amplitude offset identification submodule uses the rate fluctuation interval distribution to analyze the voltage data of each node in the corresponding time period, compares the difference between the data and the rated voltage, and selects the time segments with continuous voltage fluctuations and critical amplitude changes to obtain the amplitude offset segment parameters; The previously obtained time range of the rate fluctuation interval is used to obtain the raw voltage data of all monitoring nodes within the period. The data is then subtracted from the standard voltage value set under normal operating conditions to obtain the offset amplitude at each moment. The average of all offset values ​​within each time period is then compared with a preset amplitude reference value. This amplitude reference value can be set to approximately 10 volts based on the offset distribution in historical data. If the average offset value for the current time period exceeds this reference value and the offset state persists for longer than 0.15 seconds—that is, multiple consecutive data points are in the offset state for at least 150 milliseconds—then it is determined to be a valid segment of abnormal voltage amplitude fluctuation. All segments that meet the conditions are then recorded based on information such as the start and end time, maximum offset value, and number of participating nodes to form a set of amplitude offset segment parameters. For example, between 12:01:34 and 12:01:34.29 noon, if the voltages of five nodes were more than 12 volts higher than the rated value, and this high state lasted for more than 0.17 seconds, this time period is selected as the amplitude offset segment.

[0026] The fluctuation feature extraction submodule analyzes the overlapping time segments based on the amplitude offset segment parameters, determines the correlation between the time segments and the start and stop moments of the flywheel energy storage system, identifies the associated fluctuation events, and obtains the time series characteristics of the fluctuation events; Extract the start and end time information of the segment, then access the operation records of the flywheel energy storage system, find all the recorded flywheel start and stop time points, and compare whether each amplitude offset segment is close to the flywheel start and stop time. The close standard is that the time difference does not exceed 0.1 seconds. If this condition is met, it is considered that this section of fluctuation is related to the operating status of the flywheel, and then continue to check the overlap of multiple related time periods. If the time overlap is greater than 0.05 seconds, they are merged into a longer event time period and the corresponding flywheel number is recorded. The start point, end point, duration, flywheel number involved, voltage offset degree and other information of the time period are output as the timing characteristics of the fluctuation event. For example, if flywheel number 12 is started at 12:01:34.20, and a voltage fluctuation with an offset value of 13 volts occurs from 12:01:34.12 to 12:01:34.29, then this fluctuation event is recorded as being associated with the start of flywheel number 12, and its time segment and characteristic parameters are used to construct the timing characteristic information of the fluctuation event.

[0027] See also Figure 3 , the load mutation determination module includes: The energy supply rate extraction submodule extracts the instantaneous energy supply power in each period before and after the disturbance based on the timing characteristics of the fluctuation event and the voltage change rate data collected by the server cluster log and the flywheel energy storage monitoring node. It also compares the instantaneous energy supply power with the load switching rate in the same period to obtain the energy supply change rate characteristics. It is necessary to clarify the start and end time of the fluctuation event, the associated flywheel number and its trigger type, and use this time segment as an analysis window to read the voltage change rate data recorded by the flywheel energy storage monitoring node. At the same time, read the server cluster operation log that matches the time period, extract its load status record and energy supply request record, and then set an observation time window before and after the fluctuation event. For example, the length of each time window is set to 1 second, and calculate the instantaneous power output by the flywheel energy storage system to the server cluster power supply node in the two windows. The power value is obtained by collecting the output voltage and current data in the energy storage discharge branch and multiplying them. For example, the output voltage of flywheel F-07 from 12:02:10.00 to 12:02:11.00 is 220 volts and the current is 8 amps, then the instantaneous energy supply power is 1760 watts. The power at each time point is collected in turn to form a time series, and then the data is time-aligned. The average power supply power and fluctuation amplitude are compared in the two time windows before and after the disturbance event to determine the power change trend. The load switching behavior marked in the server cluster log is then called to extract the server numbers and quantity of newly powered on or off servers at each time point to form a load switching rate time series. For example, if within one second after the fluctuation occurs, 6 servers are switched from the shutdown state to the running state, the load switching rate is recorded as 6 servers per second. The flywheel output power is then time-correlated with the load switching rate to determine whether there is a rapid change in energy output during the period of high load switching rate. The judgment reference value is set to a power mutation rate greater than 200 watts per second and a duration of more than 0.2 seconds. If this condition is met, the time slice is marked as a significant energy supply change segment. The energy supply difference before and after the disturbance, the duration of the mutation segment, and the time alignment result with the load switching behavior are extracted to form the energy supply change rate feature.

[0028] The node fluctuation quantification submodule monitors the changes in the number of concurrent energy supply nodes based on the energy supply change rate characteristics, and calculates the fluctuation of the number of energy supply nodes in the differential disturbance stage using the formula: ; Identify the key time periods where parameter changes are prominent and obtain the key intervals of load disturbance ,in, Represents the first The number of concurrent energy supply nodes in a time period, Represents the first The number of concurrent energy supply nodes in a time period, Represents the first Characteristics of energy supply change rate during time periods, Represents the first Characteristics of energy supply change rate during time periods, Represents the number of time periods for perturbation analysis.

[0029] The critical load disturbance interval refers to the continuous time interval during which various parameters (such as instantaneous energy supply power, load switching rate, number of concurrent energy supply nodes, etc.) change most significantly and fluctuate most concentratedly by analyzing the operating data of the flywheel energy storage system and server cluster during voltage fluctuations or sudden load changes in the data center. This interval can reflect the main time period when load disturbance events occur, thereby guiding the energy storage system scheduling and response of the data center under power fluctuations.

[0030] The operation records of each flywheel node in different stages before and after the disturbance are called in time sequence, and the number of nodes participating in the power supply in each sub-period before the disturbance starts and after the disturbance response is extracted, and paired with the corresponding energy supply change rate characteristics. In actual operation, the start time of the disturbance event is 14:00 on May 10, 2025, and the analysis period is set to 3 stages, each stage is 2 minutes long. Then to Corresponding to the period before the disturbance, to Corresponding to the post-disturbance period, the number of concurrent energy supply nodes recorded during the collection period was 5, 6, and 7 before the disturbance, and 11, 10, and 12 after the disturbance. The energy supply change rate characteristics were 3.2kW / node, 3.9kW / node, and 4.1kW / node before the disturbance, and 6.5kW / node, 5.8kW / node, and 6.2kW / node after the disturbance. To unify the dimensions and ensure the rationality of the calculation, the change in the number of nodes was normalized to the range of 0.00-1.00. The normalized result is: 6→0.60, 4→0.40, 5→0.50; At the same time, the energy supply change rate is normalized to the range of 0.00–1.00, and the normalized value is: 3.3→0.66, 1.9→0.38, 2.1→0.42, after normalization, the difference is input into the following formula: ; ; ; The obtained disturbance response intensity is , indicating that there is a medium-intensity synchronous response relationship between the change in the number of energy supply nodes and / or their corresponding rate characteristics during the disturbance, which is used to mark to This is the key period during which resources need to be dispatched first in response to the disturbance. This value can also be compared with the disturbance results of other periods. , then it is identified as a sensitive segment that needs to be adjusted in scheduling order. , it is excluded from the priority scheduling sequence. The formula forms a quantitative expression of the degree of synchronization between the node dynamic response and the energy supply adjustment in the overall disturbance event by normalizing the product of the node change and the energy supply rate change and averaging the values ​​in all time periods.

[0031] See also Figure 4 , the capacity correction allocation module includes: The coordination gradient analysis submodule analyzes the synchronization between the flywheel speed change and the mechanical friction rate based on the key interval of the load disturbance, compares the change gradients of the two in each period, determines whether the fluctuations of the two show a coordinated trend, and calculates the degree of fluctuation coordination in the difference section to obtain the coordination gradient amplitude; The real-time speed of each flywheel and the friction rate of the electric bearing in the disturbance interval are extracted from the sensor data of the flywheel energy storage system, and the data are sorted into two independent sequences in chronological order. Subsequently, the two sequences are differentially processed to obtain the speed change amplitude and the friction rate change amplitude at each time point. Then, a sliding scan is performed in units of a fixed-length time window, and the change directions of the two in each time period are compared section by section to see if they are consistent. If the directions are consistent and the amplitude difference is within the set limit, it is determined to be a synchronous change. In the specific implementation, the change direction consistency rate can be set to be greater than 80% and the change difference can be set to no more than 20% of the original value as the basis for coordination judgment. For example, from 12:12:00 to 12:12:05 on June 6, 2025, the speed of the flywheel No. F-09 dropped by 120 revolutions per minute in 1 second, and the friction rate increased by 6 units. The change direction is opposite, which does not constitute a Coordination, while the speed of number F-03 decreased by 80 revolutions per minute, and the friction rate increased by 6 units, with the same direction and similar amplitude, which constitutes a coordinated segment. Then, the proportion of time determined to be coordinated in all time slices in the entire disturbance interval is counted to form a synchronization rate evaluation result, and then the fluctuation distribution of speed and friction rate in each difference segment is extracted. The average value of the absolute difference of the gradient offset between the two is calculated as the fluctuation coordination difference value, and then the proportion of the maximum gradient difference accounted for by this value is subtracted from 1 to convert it into a coordination gradient amplitude index between 0 and 1. If the coordination gradient amplitude is above 0.75, it is considered highly coordinated, between 0.5 and 0.75, it is moderately coordinated, and below 0.5, it is uncoordinated. In this example, if the coordination gradient amplitude of flywheel F-03 in the disturbance segment is 0.82, its speed change and friction rate change are significantly synchronized, and the coordination gradient amplitude corresponding to each flywheel number is output.

[0032] The energy distribution adjustment submodule analyzes the coordination of the flywheel energy storage unit in energy release according to the coordinated gradient amplitude, compares the balance between energy sharing and dynamic power output among the energy supply nodes, adjusts the energy output sequence, and obtains energy distribution structure information; According to the flywheel number, the energy supply record and energy release of each flywheel unit in the disturbance section are read, and a corresponding relationship is established between the flywheel and the power supply node where it is located. The dynamic power output sequence of each node is extracted and time-aligned with the flywheel energy release time series. Then, the energy proportion released by each flywheel in the same time period is counted, and the proportion is compared with the ratio of the actual energy supply power of the node in the time period. If it is found that the release proportion is significantly high and the output power proportion is low, it means that the flywheel has poor coordination in the section. Otherwise, it means that its energy supply response is good and the coordination is high. Then, all flywheels are prioritized according to the coordination gradient amplitude. The higher the coordination amplitude, the higher the ranking, for example, F-0 3 coordination amplitude is 0.82, F-07 is 0.65, and F-12 is 0.49, then the order is F-03, F-07, and F-12. Then, according to the sorting order, the energy release plan is adjusted. During the energy supply period, the high-priority flywheel is set as the first discharge target, and the low-priority flywheel is postponed or left idle. The target flywheel number of each power supply task is reallocated. For example, the period from 12:12:04 to 12:12:05, which was originally discharged by F-12, is adjusted to F-03 after this round of sorting. Structured data of the flywheel number, start and end time, released energy value, and corresponding node number in each power supply period is generated to form new energy distribution structure information.

[0033] The energy storage distribution correction submodule calls the energy distribution structure information, analyzes the difference between the current speed of each flywheel and the upper limit speed, and combines the moment of inertia, friction coefficient and power release difference to adopt the formula: ; Get the energy storage distribution correction characteristics ,in, Indicates the total number of flywheels participating in concurrent scheduling in the current cycle, Indicates the The moment of inertia of a flywheel reflects the response ability of the flywheel unit to the change of speed during the energy storage process. Indicates the The difference between the current speed of each flywheel and the upper speed limit reflects the remaining dynamic release space of the unit. Indicates the The bearing friction coefficient of a flywheel is used to measure the energy loss level of the flywheel during rotation. Indicates the The power release difference of each flywheel in this cycle is used to identify the change in the flywheel power output per unit time in the current cycle.

[0034] The energy storage allocation correction feature is a composite data structure generated during the energy storage system scheduling process based on the flywheel operating status and scheduling structure optimization results. It is used to measure the changing trend of the dynamic energy supply capacity of each flywheel energy storage unit under the current scheduling configuration. It guides the ranking and participation level evaluation of flywheels in dynamic energy supply scheduling, and serves as the basis for priority sorting and dynamic scheduling adjustment.

[0035] For each flywheel unit involved in the scheduling in the current cycle, six operating parameters are called: number, moment of inertia, current speed, upper limit speed, bearing friction coefficient, and power release change. They are compared and calculated item by item. First, the difference between the current speed and the set upper limit speed of each flywheel unit is analyzed to obtain the remaining range of its release capacity. For example, the current speed of flywheel F1 is 4600 rad / s, and the set upper limit is 4800 rad / s, with a difference of 200 rad / s. Then, the moment of inertia of the flywheel is obtained as 0.85 kg·m 2 , the bearing friction coefficient is 0.018, and the power release change is 11kW. They are uniformly normalized and calculated as follows: inertia is 0.68, speed difference is 0.83, friction coefficient is 0.69, and power release is 0.71. Substituting them into the formula, when performing the calculation, taking the four flywheels F1-F4 as an example, the normalized values ​​of their parameters are as follows: F1: , , , ; F2: , , , ; F3: , , , ; F4: , , , ; Substitute into the formula and calculate as follows: ; ; ; ; Obtain: ; The results show that in the current cycle, under the adjusted energy structure scheduling, the energy storage and release coordination correction amplitude of the flywheel energy supply system is This value can be used to compare and analyze with the average correction amplitude of historical operation, so as to provide a calculation basis for subsequent scheduling priority sorting. The formula is introduced by 、 、 、 The four types of parameters are combined with composite ratio calculation and square root conversion, which not only considers the energy release capacity of the flywheel itself, but also integrates its friction consumption and dynamic power output characteristics into the correction calculation, thereby forming a comprehensive quantitative standard for the flywheel scheduling performance, so that the energy storage system scheduling has a comprehensive trade-off basis for adaptability and responsiveness.

[0036] See also Figure 5 ,The scheduling order adjustment module includes: The main circuit comparison submodule analyzes the synchronous changes of the total operating power curve of the distribution main circuit and the output power curve of the flywheel energy storage system based on the energy storage distribution correction characteristics, determines the synchronization offset of the power curve within the monitoring period, selects the cycle segments with key synchronization characteristics, and obtains the power offset cycle segments; Extract the total operating power time series of the distribution main circuit during the monitoring period. This series can be composed of the real-time voltage and current product collected by the total power monitoring point. At the same time, extract the output power time series from each discharge node of the flywheel energy storage system and summarize them to obtain the total output power curve of the flywheel. Then align the two power curves at the same time step and divide them into segments through the sliding time window. For example, a synchronous comparison unit is constructed with each 10 seconds as a segment. Then, the change trend analysis is performed segment by segment. By calculating the change direction of the two power sequences in each segment and recording their same direction or reverse mark, the time period corresponding to the synchronous direction ratio reaching more than 80% is marked as the same trend. Then calculate the instantaneous value of the two curves in each segment. The absolute value of the power difference is averaged at 12:15:00 on June 6, 2025, and the ratio of this value to the average power level of the main circuit is compared. If the difference ratio is less than 10%, it is determined to be synchronized with the fluctuation amplitude. The segments that meet both conditions are filtered as power synchronization offset segments. For example, if the average power of the main circuit is 18,000 watts and the flywheel output is 17,500 watts from 12:15:00 to 12:15:10 on June 6, 2025, and the change direction of the two is consistent and the average difference is 400 watts, accounting for about 2.2%, then this segment is included in the power offset period segment. Finally, the start and end time of each segment, the average power of the main circuit, the average power of the flywheel output, the difference percentage, the trend consistency mark and other data are extracted to form the power offset period segment.

[0037] The load ratio analysis submodule compares the flywheel energy storage system's energy distribution with the current load status of each server zone based on the power offset period. It calculates the mean square error of the power change rate of each server zone, identifies the number of server zones with critical change rate amplitudes, and obtains the load fluctuation normalization value. Call the load record data of the server area within the cycle period, extract the current power consumption time series of each physical area one by one, and simultaneously extract the flywheel energy supply record, establish the mapping relationship between the flywheel power supply node and the server area, re-project the flywheel energy supply distribution to each area, calculate the power change rate series of each server area by time segment in each power offset segment, and then perform mean square error calculation on the power change rate series of all areas to characterize the distribution of the degree of fluctuation in each area, and then set the benchmark threshold to judge whether the change rate amplitude is prominent, and set the mean square error with reference to the data change range of the past week. The difference threshold is 50 watts per second. The number of regions whose mean square error exceeds the threshold is counted as the identification result. For example, during the offset period from 12:15:00 to 12:15:10, there are 12 server regions, of which 5 regions have a mean square error of more than 50 watts per second. The number of key change regions is recorded as 5. This number is then divided by the total number of regions and rounded to two decimal places to obtain a normalized load fluctuation value of 0.42. This value is used as a normalized indicator representing the degree of load fluctuation imbalance in the server region during the current power offset period, and the normalized value is output.

[0038] The scheduling sequence generation submodule determines the energy output efficiency and residual energy ratio of the flywheel energy storage device based on the load fluctuation normalization amount, and calculates the scheduling sequence difference in the target cycle using the formula: ; Adjust the dispatch priority to obtain the energy storage priority ranking index, where: Indicates the number The energy storage priority ranking index corresponding to the flywheel energy storage device, Indicates the number The remaining available energy ratio of the flywheel energy storage device in the target cycle, Indicates the number The unit energy supply efficiency of the flywheel energy storage device, Indicates the normalized load fluctuation in the current cycle. Indicates the number The flywheel energy storage device is The current dispatch sequence number of the server region, Indicates the number The flywheel energy storage device is The dispatch number that each server region should be assigned in rotation order, Indicates the number of server regions.

[0039] The energy storage priority ranking index refers to a set of ranking data determined for each flywheel device in the flywheel energy storage system based on factors such as the comprehensive operating status of each flywheel energy storage device in the current scheduling cycle, the proportion of remaining available energy, the unit energy supply efficiency, the differences in scheduling order between regions, and load fluctuations. This index is used to reflect the priority order of each flywheel device in the energy storage compensation scheduling, thereby determining which flywheel devices should be given priority in energy supply and compensation when power grid fluctuations or load mutations occur, ensuring the dynamic response capability and power supply security of the overall system operation.

[0040] Determine the energy output efficiency and residual energy ratio of each numbered flywheel energy storage device in the current cycle, establish a scheduling performance index set for all numbered devices, calculate the product of each device's unit energy compensation capacity and the actual residual energy ratio, and obtain its basic output level. Taking the flywheel device numbered 03 as an example, its residual energy in the current monitoring cycle is , the rated energy storage of the equipment is , the remaining energy ratio is calculated as: ; Output efficiency =0.94, which is the output performance value normalized by the flywheel speed, friction loss, temperature rise, etc. The product of the two is: ; Load fluctuation normalization It has been determined in advance to be 0.60, and the square root expression is substituted into it to calculate: ; The first part of the indicators are: ; Then analyze the scheduling deviation of the flywheel device No. 03 in the four server areas, assuming that its current numbering order is , the rotation order is , then the deviation is , squared to , the sum is 3, and the number of server regions is , so the second part of the result is: ; Substituting the above results into the complete formula: ; In the formula, the first term expresses the comprehensive energy supply capacity of the flywheel in the current cycle and modulates its priority through the square root function of the fluctuation normalization quantity; the second term measures the coordination degree of scheduling by the mean square of the number deviation of the server area. The smaller the value, the more reasonable the number arrangement. The calculation result is This shows that although the flywheel device No. 03 has sufficient energy supply capacity, its scheduling consistency has a large deviation and the overall ranking is relatively backward. It is recommended to be placed in the rear standby queue or participate in non-main load compensation tasks.

[0041] See also Figure 6 , the state change tracking module includes: The energy fluctuation analysis submodule analyzes the changes in the remaining available energy of the flywheel number in consecutive cycles based on the energy storage priority ranking index, compares the energy fluctuation amplitude between each monitoring cycle, determines whether there is an abnormal deviation in the change, and generates the cycle energy fluctuation amplitude; The remaining available energy data of each flywheel number in multiple consecutive monitoring cycles are extracted in order of priority. Each cycle can be set to 5 minutes, and it is organized into a two-dimensional time series table according to the dual index of number and time. Then, the energy difference between adjacent cycles is calculated for the continuous cycle data of each flywheel number. The absolute change value of the energy remaining between two cycles is calculated by traversal. Then, the average of all cycle differences of each number is calculated to obtain the energy fluctuation amplitude of the flywheel during the observation time. After that, by comparing the difference between the energy change value of each cycle and the historical fluctuation mean of the number, it is judged whether there is an abnormal offset. The offset judgment The basis for diagnosis is set as the single-cycle fluctuation amplitude is more than twice the average fluctuation value of the number in the past 12 hours, and the current remaining energy is less than 30% of its capacity. For example, the remaining energy of number F-15 in cycle T1 is 6.8 kJ, which drops to 3.1 kJ in T2. ​​The average historical fluctuation is 1.5 kJ, and the current decline is 3.7 kJ, which is more than twice the historical average and the remaining ratio is less than 30%. The abnormal offset condition is met, and the cycle is recorded as an abnormal fluctuation segment. All cycle fluctuation differences corresponding to each flywheel number, whether it is abnormal, start and end time and current remaining energy are output to form a cycle energy fluctuation amplitude data set.

[0042] The stability judgment submodule uses the amplitude of periodic energy fluctuations to determine the persistence of fluctuations in the time series, analyzes whether there are regular fluctuations or sudden changes in the energy change process, identifies the coherence characteristics of energy changes, and obtains the stability trend of energy response; The energy fluctuation time series described above was expanded by flywheel number, and a complete time series fluctuation graph was constructed for each number. A sliding window method was used to check whether the fluctuation values ​​showed continuous high-amplitude deviations between adjacent windows to determine whether the fluctuation trend was continuous. If the abnormal deviation judgment criteria were met for three or more consecutive cycles, it was considered a persistent unstable state. At the same time, whether there were cases in the fluctuation segment where the fluctuation direction of the two cycles before and after was opposite and the value reversed significantly was identified. If the reversal amplitude exceeded twice the historical average fluctuation amplitude of the number, it was considered a sudden change. For example, number F-10 dropped 3.2 kJ at T1, rose 2.9 kJ at T2, and then dropped 3.5 kJ at T3. The repeated direction and sharp amplitude constituted a sudden change. The ratio of the number of abnormal segments to the total number of cycles for each flywheel in the fluctuation graph was then counted, and the stability rate of the number within the analysis segment was calculated. If the stability rate was less than 60%, the number was marked as unstable. The fluctuation direction sequence, reversal point location, continuous abnormal length, and stability rate were output by number to form a data table of energy response stability trends for each number.

[0043] The number jump screening submodule analyzes the fluctuation continuity reflected by the stable trend of the energy response, screens the data sequence with prominent changes in the flywheel number, determines whether the fluctuation characteristics meet the state jump characteristics, and synchronously updates the data list to obtain the state jump tracking data.

[0044] A list of numbers with stability rates below the set standard is screened, and key points are extracted from the energy fluctuation sequence of the numbers, identifying moments of direction reversal or sudden increase in amplitude. The key moments are then time-compared with flywheel start and stop events in historical scheduling records. If the key fluctuation moment occurs when the flywheel is not involved in scheduling, it is judged to be a non-task-driven change. The simultaneous records of environmental interference and power supply disturbances are further excluded. If no external driving factor can be identified and the fluctuation pattern meets the mutation characteristics, the flywheel number is marked as a state jump target. For example, number F-22 has not been scheduled in the past 12 hours, but the energy surplus value has experienced three high-frequency sudden drops, with the fluctuation form of a drop of more than 60% within 2 hours, and the stability rate is 38%. At the same time, there is no corresponding load record. In this case, the number is identified as a state jump target. The number, jump time, energy values ​​before and after the change, historical fluctuation trajectory, and the corresponding monitoring node number are synchronously written into the data list to form the state jump tracking data.

[0045] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A flywheel energy storage system for transient voltage control in a data center, characterized in that: The system comprises: The disturbance moment identification module analyzes the voltage curves of the flywheel energy storage monitoring node in each time period based on the data center voltage change rate and amplitude, compares the synchronization of rate and amplitude changes, determines the correlation between the fluctuation moment and the operating status of the flywheel energy storage system, and obtains the timing characteristics of the fluctuation event; The load mutation determination module determines the flywheel energy storage operation changes in the server cluster log based on the timing characteristics of the fluctuation event, analyzes the instantaneous power changes, compares the load switching rate and the number of concurrent energy supply nodes before and after the disturbance, identifies the critical interval, and obtains the load disturbance critical interval; The capacity correction distribution module compares the current speed of the flywheel with the trend of the mechanical friction rate based on the load disturbance key interval, determines the speed and friction coordination between the electric bearing and the energy, adjusts the available energy distribution, calculates the capacity change, and obtains the energy storage distribution correction feature; The scheduling sequence adjustment module compares the total operating power of the main circuit with the flywheel data based on the energy storage allocation correction characteristics, analyzes the changes in the server area load ratio caused by voltage offset, determines the impact of parameter changes on the scheduling sequence, and obtains the energy storage priority ranking index.

2. The flywheel energy storage system for transient voltage control in a data center according to claim 1 is characterized in that: The fluctuation event timing characteristics include the time of event occurrence, fluctuation distribution characteristics, and associated marks; the load disturbance key interval includes interval time information, load disturbance type, and disturbance event identifier; the energy storage allocation correction characteristics include corrected capacity distribution, energy storage dynamic coefficient, and energy distribution structure; the energy storage priority ranking indicators include ranking order, priority number, and allocation weight.

3. The flywheel energy storage system for transient voltage control in a data center according to claim 1, characterized in that: The disturbance moment identification module includes: The rate fluctuation identification submodule analyzes the voltage curves recorded by the monitoring nodes based on the voltage change rate and amplitude of the data center, determines the sudden changes in the voltage change rate within a continuous period, and combines the node distribution characteristics to screen the concentrated time periods where fluctuations occur. It summarizes the related change intervals and obtains the rate fluctuation interval distribution; The amplitude offset identification submodule calls the rate fluctuation interval distribution, analyzes the voltage data of each node in the corresponding time period, compares the difference between the data and the rated voltage, selects the time segment with continuous voltage fluctuation and critical amplitude change, and obtains the amplitude offset section parameter; The fluctuation feature extraction submodule analyzes the overlapping time segments according to the amplitude offset segment parameters, determines the correlation between the time segments and the start and stop moments of the flywheel energy storage system, identifies the associated fluctuation events, and obtains the time series characteristics of the fluctuation events.

4. The flywheel energy storage system for transient voltage control in a data center according to claim 1, characterized in that: The load mutation determination module includes: The energy supply rate extraction submodule extracts the instantaneous energy supply power in each period before and after the disturbance based on the timing characteristics of the fluctuation event and the voltage change rate data collected by the server cluster log and the flywheel energy storage monitoring node, and compares the instantaneous energy supply power with the load switching rate in the same period to obtain the energy supply change rate characteristics; The node fluctuation quantification submodule monitors the changes in the number of concurrent energy supply nodes based on the energy supply change rate characteristics, calculates the fluctuation of the number of energy supply nodes in the differential disturbance stage, identifies the key time period with prominent parameter changes, and obtains the load disturbance key interval.

5. The flywheel energy storage system for transient voltage control in a data center according to claim 1, characterized in that: The capacity correction allocation module includes: The coordination gradient analysis submodule analyzes the synchronization of the flywheel speed change and the mechanical friction rate based on the load disturbance key interval, compares the change gradients of the two in each period, determines whether the fluctuations of the two show a coordinated trend, and calculates the degree of fluctuation coordination in the difference section to obtain the coordination gradient amplitude; The energy distribution adjustment submodule analyzes the coordination of the flywheel energy storage unit in energy release according to the coordination gradient amplitude, compares the balance between energy sharing and dynamic power output among the energy supply nodes, adjusts the energy output sequence, and obtains energy distribution structure information; The energy storage distribution correction submodule calls the energy distribution structure information, analyzes the difference between the current speed of each flywheel and the upper limit speed, and obtains the energy storage distribution correction characteristics by combining the moment of inertia, friction coefficient and power release difference.

6. The flywheel energy storage system for transient voltage control in a data center according to claim 1, characterized in that: The scheduling order adjustment module includes: The main circuit comparison submodule analyzes the synchronous changes of the total operating power curve of the distribution main circuit and the output power curve of the flywheel energy storage system based on the energy storage distribution correction characteristics, determines the synchronous offset of the power curve within the monitoring period, selects the cycle segments with key synchronization characteristics, and obtains the power offset cycle segments; The load ratio analysis submodule compares the energy distribution of the flywheel energy storage system with the current load status of each server area based on the power offset period, calculates the mean square error of the power change rate of each server area, identifies the number of server areas with critical change rate amplitudes, and obtains the load fluctuation normalization amount; The scheduling sequence generation submodule determines the energy output efficiency and the remaining energy ratio of the flywheel energy storage device according to the load fluctuation normalization amount, calculates the scheduling sequence difference in the target period, adjusts the scheduling priority, and obtains the energy storage priority ranking index.

7. The flywheel energy storage system for transient voltage control in a data center according to claim 1, characterized in that: The system further comprises: The state change tracking module analyzes the change in the number of the first flywheel in the order according to the energy storage priority ranking index, compares the change amplitude of the residual energy in consecutive cycles, determines the stability of the energy change, identifies the key number, marks it as a state jump object, updates the data list, and obtains the state jump tracking data; The state transition tracking data includes a transition flag, number information, and a change history record.

8. The flywheel energy storage system for transient voltage control in a data center according to claim 7, characterized in that: The state change tracking module includes: The energy fluctuation analysis submodule analyzes the change in the remaining available energy of the flywheel number in consecutive cycles based on the energy storage priority ranking index, compares the energy fluctuation amplitude between each monitoring cycle, determines whether there is an abnormal deviation in the change, and generates the periodic energy fluctuation amplitude; The stability judgment submodule calls the periodic energy fluctuation amplitude, judges its fluctuation persistence in the time series, analyzes whether there are regular fluctuations or sudden changes in the energy change process, identifies the coherence characteristics of the energy change, and obtains the energy response stability trend; The number jump screening submodule analyzes the fluctuation coherence reflected by the stable trend of the energy response, screens the data sequence with prominent changes in the flywheel number, determines whether the fluctuation characteristics meet the state jump characteristics, and synchronously updates the data list to obtain the state jump tracking data.

Images