Distributed base station energy storage resource virtual power plant anomaly monitoring method and system

Abstract

The application provides a distributed base station energy storage resource virtual power plant anomaly monitoring method and system, wherein the method comprises the following steps: firstly, acquiring the grid frequency modulation demand, real-time electricity price, energy storage state, mechanical stress and radio frequency signal data of each distributed base station in the virtual power plant; then, aggregating the frequency modulation demand, electricity price and energy storage state data to construct a time series database of electricity price load response, and simultaneously fusing the mechanical stress and radio frequency signal to generate stress electromagnetic coupling fingerprints; then, based on the time series database and the physical fingerprints, the charging and discharging competitive behavior characteristics of each base station are generated by Nash equilibrium model analysis; finally, by comparing the deviation degree of the actual behavior characteristics and the expected equilibrium state of the model, the abnormal energy storage node is accurately positioned, and the credible evidence and traceability of the abnormal behavior are realized by using the blockchain smart contract. The application improves the accuracy and traceability reliability of the virtual power plant energy storage anomaly monitoring.

Description

Technical Field

[0001] This application relates to the fields of power systems and information communication technology, and in particular to a method and system for anomaly monitoring of distributed base station energy storage resource virtual power plants. Background Technology

[0002] As the scale of virtual power plants participating in electricity market frequency regulation expands, the charging and discharging behavior of distributed base station energy storage under the incentive of electricity prices is prone to overcharging / over-discharging anomalies. It is necessary to monitor the operating status of energy storage equipment in real time, and combine market frequency regulation requirements with the physical status of the equipment to achieve accurate identification and reliable traceability of abnormal behavior.

[0003] Existing solutions analyze the matching degree between historical charging and discharging data of energy storage devices and grid frequency regulation commands, combine preset thresholds to identify anomalies, and use data signature technology to preserve evidence of anomaly records.

[0004] This scheme relies on static thresholds to determine anomalies, which makes it difficult to adapt to dynamic scenarios such as electricity price fluctuations and changes in equipment status. Furthermore, it does not consider the physical characteristics of energy storage devices, such as mechanical stress and electromagnetic interference, which limits the accuracy and reliability of anomaly detection. Summary of the Invention

[0005] This application provides a method and system for monitoring anomalies in distributed base station energy storage virtual power plants, in order to solve the problems of low accuracy and poor traceability reliability in the existing technology for monitoring anomalies in virtual power plant energy storage.

[0006] Firstly, this application provides a method for anomaly monitoring of distributed base station energy storage resource virtual power plants, including:

[0007] Acquire grid frequency regulation demand information, electricity price trend data, historical energy storage status data, mechanical stress signals, and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant;

[0008] The power grid frequency regulation demand information, the electricity price trend data, and the historical energy storage status data are aggregated to form an electricity price load response time-series database.

[0009] By fusing the mechanical stress signal and the radio frequency disturbance signal, a stress electromagnetic coupling fingerprint is generated;

[0010] Based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model, charging and discharging competition behavior characteristics are generated;

[0011] Based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, abnormal energy storage nodes in the virtual power plant are located. Based on the abnormal energy storage nodes, an immutable abnormal monitoring report is generated through a blockchain smart contract.

[0012] Optionally, the step of generating charging and discharging competition behavior characteristics based on the electricity price load response time-series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model, includes:

[0013] The price value is parsed from the window-level response sequence of the electricity price load response time series database as the electricity price incentive parameter, and the rate of change of state of charge is parsed as the energy storage response parameter.

[0014] The zero-crossing interval sequence in the stress electromagnetic coupling fingerprint is converted into a frequency stability index, and the amplitude envelope in the stress electromagnetic coupling fingerprint is converted into a mechanical strength attenuation index.

[0015] The energy storage response parameters, the frequency stability index, and the mechanical strength attenuation index are input into the supply-side constraint sub-model of the Nash equilibrium model to generate a set of feasible behavioral strategies.

[0016] The electricity price incentive parameters are input into the demand-side game sub-model of the Nash equilibrium model to generate a set of expected behavioral strategies.

[0017] Based on the expected behavior strategy set and the feasible behavior strategy set, the charging and discharging competition behavior characteristics are output.

[0018] Optionally, the step of outputting charging and discharging competition behavior characteristics based on the expected behavior strategy set and the feasible behavior strategy set includes:

[0019] The expected behavior strategy set is sorted according to the time window of the scheduling cycle to construct a first strategy matrix, where each first row vector corresponds to a theoretical charging and discharging strategy for a time window.

[0020] The feasible behavior strategy set is sorted according to the time window with the same scheduling period to construct a second strategy matrix, where each second row vector corresponds to an executable charging and discharging strategy in the same time window;

[0021] For each time window, the theoretical charging / discharging strategy is matched with the executable charging / discharging strategy to generate a set of conflicting strategy units;

[0022] Based on the distribution density and temporal continuity of the conflict strategy unit set, charging and discharging competition behavior characteristics are generated.

[0023] Optionally, the step of matching the theoretical charge / discharge strategy with the executable charge / discharge strategy for each time window to generate a set of conflicting strategy units includes:

[0024] The power change rate and duration parameters are extracted from the theoretical charge-discharge strategy.

[0025] Extract the allowable power variation range and duration threshold from the executable charge / discharge strategy;

[0026] When the power change rate is less than the minimum value or greater than the maximum value of the allowable power change range, the current time window is marked as a power conflict strategy unit;

[0027] When the duration parameter is greater than the duration threshold, the current time window is marked as a duration conflict strategy unit;

[0028] If both power conflict and duration conflict exist simultaneously, the current time window will be remarked as a composite conflict strategy unit.

[0029] After traversing all time windows, the time windows that are used as power conflict strategy units, duration conflict strategy units, or composite conflict strategy units are summarized to form a conflict strategy unit set.

[0030] Optionally, the aggregation of the power grid frequency regulation demand information, the electricity price trend data, and the historical energy storage status data to form an electricity price load response time-series database includes:

[0031] The command value in the power grid frequency regulation demand information, the price value in the electricity price trend data, and the state of charge value in the energy storage status history data are bound together with three parameters according to the same time mark to generate a basic data unit.

[0032] Based on the minimum time granularity of virtual power plants participating in electricity market frequency regulation, the continuous time period of virtual power plants in the electricity market is divided into multiple analysis windows;

[0033] Within each analysis window, the basic data units are arranged in chronological order to form a window-level response sequence;

[0034] The window-level response sequences of all analysis windows are integrated according to the time dimension to construct a time-series database of electricity price load response.

[0035] Optionally, the step of fusing the mechanical stress signal and the radio frequency disturbance signal to generate a stress electromagnetic coupling fingerprint includes:

[0036] The mechanical stress signal and the radio frequency disturbance signal are sampled respectively to obtain stress time-domain waveform data and radio frequency time-domain waveform data;

[0037] The stress time-domain waveform data and the radio frequency time-domain waveform data marked at the same time are superimposed to generate hybrid waveform data;

[0038] Extract the zero-crossing interval sequence and amplitude envelope from the mixed waveform data;

[0039] The standard deviation of the zero-crossing interval sequence and the average value of the amplitude envelope are combined according to a preset weight ratio to generate a stress electromagnetic coupling fingerprint.

[0040] Optionally, the step of locating abnormal energy storage nodes in the virtual power plant based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, and generating an immutable anomaly monitoring report based on the abnormal energy storage nodes through a blockchain smart contract, includes:

[0041] Calculate the deviation quantization value between the charging and discharging competition behavior characteristics and the characteristics corresponding to the expected equilibrium state;

[0042] When the deviation quantization value of any distributed base station exceeds the preset deviation threshold N times in a row, the distributed base station is marked as an abnormal energy storage node, where N is greater than or equal to 2.

[0043] Extract the key window-level response sequence corresponding to the abnormal energy storage node from the electricity price load response time-series database;

[0044] The state of charge values, electricity price values, and frequency regulation command values ​​in the key window-level response sequence are reorganized into anomaly decision chain data in chronological order.

[0045] Through blockchain smart contracts, the abnormal decision chain data and the corresponding deviation quantification value are reliably stored and an anomaly monitoring report is generated.

[0046] Secondly, this application provides a distributed base station energy storage resource virtual power plant anomaly monitoring system, comprising:

[0047] The acquisition module is used to acquire grid frequency regulation demand information, electricity price trend data, energy storage status historical data, mechanical stress signals, and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant.

[0048] The aggregation module is used to aggregate the power grid frequency regulation demand information, the electricity price trend data, and the historical data of energy storage status to form an electricity price load response time-series database.

[0049] The fusion module is used to fuse the mechanical stress signal and the radio frequency disturbance signal to generate a stress electromagnetic coupling fingerprint;

[0050] The generation module is used to generate charging and discharging competition behavior features based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model;

[0051] The positioning module is used to locate abnormal energy storage nodes in the virtual power plant based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, and to generate an immutable abnormal monitoring report based on the abnormal energy storage nodes through a blockchain smart contract.

[0052] Thirdly, this application provides a computing device, including a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the method for abnormal monitoring of a distributed base station energy storage resource virtual power plant as described in any of the first aspects.

[0053] Fourthly, this application provides a computer storage medium storing computer program instructions thereon, which, when executed by a processor, implement the method for abnormal monitoring of a distributed base station energy storage resource virtual power plant as described in any one of the first aspects.

[0054] This application provides a method for anomaly monitoring of distributed base station energy storage resources in a virtual power plant. The method includes: acquiring grid frequency regulation demand information, electricity price trend data, historical energy storage status data, mechanical stress signals, and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant; aggregating the grid frequency regulation demand information, the electricity price trend data, and the historical energy storage status data to form an electricity price load response time-series database; fusing the mechanical stress signals and the radio frequency disturbance signals to generate a stress electromagnetic coupling fingerprint; generating charging and discharging competition behavior characteristics based on the electricity price load response time-series database and the stress electromagnetic coupling fingerprint, combined with a Nash equilibrium model; locating abnormal energy storage nodes in the virtual power plant according to the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model; and generating an immutable anomaly monitoring report based on the abnormal energy storage nodes through a blockchain smart contract.

[0055] The technical solution provided in this application has the following beneficial effects:

[0056] This application achieves multi-source data acquisition covering the electricity market and equipment physical status by acquiring grid frequency regulation demand information, electricity price trend data, historical energy storage status data, mechanical stress signals, and radio frequency disturbance signals; it aggregates data to form an electricity price load response time-series database, establishing a dynamic correlation between market incentives and energy storage response; it fuses and generates stress electromagnetic coupling fingerprints to provide unique characteristics of equipment physical status; it combines a Nash equilibrium model to generate charging and discharging competition behavior characteristics, quantitatively analyzing the interaction between market behavior and physical constraints; and it locates abnormal energy storage nodes and generates monitoring reports with blockchain evidence to ensure reliable traceability of abnormal behavior.

[0057] Furthermore, this application also extracts electricity price incentive parameters and energy storage response parameters from time series databases, transforms stress electromagnetic coupling fingerprints into frequency stability indicators and mechanical strength attenuation indicators, and inputs them into the supply-side constraint sub-model and demand-side game sub-model of the Nash equilibrium model, respectively, to generate a set of feasible behavioral strategies and a set of expected behavioral strategies, and finally outputs charging and discharging competition behavior characteristics that reflect the actual competitive situation.

[0058] Furthermore, through the collaborative analysis of the two models, the dynamic game relationship between the economic behavior of energy storage resources and physical state constraints under the incentive of electricity prices can be accurately captured, providing a quantitative basis for anomaly detection that combines market adaptability and equipment safety.

[0059] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 A flowchart illustrating an anomaly monitoring method for a distributed base station energy storage resource virtual power plant provided in this application embodiment;

[0062] Figure 2 This application provides a schematic diagram of the structure of a distributed base station energy storage resource virtual power plant anomaly monitoring system provided in an embodiment of the present application;

[0063] Figure 3 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0064] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0065] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0066] Current virtual power plants primarily rely on analyzing the matching degree between historical charging and discharging data and grid commands, combined with fixed thresholds, to monitor anomalies in energy storage devices. While this method can detect obvious anomalies, it has significant shortcomings: the behavior of energy storage devices in response to electricity price fluctuations is not only influenced by market factors but also closely related to their physical conditions, such as mechanical vibration and electromagnetic environment. Existing solutions focus only on market data while neglecting physical condition monitoring, resulting in the system's inability to accurately identify abnormal behaviors caused by changes in the device's physical state. Furthermore, the fixed judgment criteria are difficult to adapt to dynamic changes in market prices.

[0067] To address this issue, this application proposes a method for anomaly monitoring of distributed base station energy storage virtual power plants. This method simultaneously collects grid frequency regulation demand, market price trends, energy storage device status data, and physical signals from device operation to construct a dual-dimensional monitoring system reflecting both market and device status. First, the method processes market data and device status data separately to form feature databases. Then, an intelligent analysis model evaluates the actual operating status of energy storage devices under the dual effects of market incentives and physical constraints. Specifically, by introducing feature analysis of device vibration and electromagnetic signals, potential anomalies that are difficult to detect using traditional methods can be captured. Finally, blockchain technology is used to ensure the authenticity and reliability of all anomaly records. This solution overcomes the limitations of traditional single-dimensional monitoring, considering both the impact of market price fluctuations and changes in device physical status, achieving comprehensive and accurate identification of abnormal behavior in energy storage devices, and effectively solving the problems of single monitoring dimensions and insufficient adaptability in existing technologies.

[0068] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0069] Figure 1 A flowchart illustrating an anomaly monitoring method for a distributed base station energy storage resource virtual power plant provided in this application embodiment is shown below. Figure 1 As shown, the method includes:

[0070] Step 101: Obtain grid frequency regulation demand information, electricity price trend data, historical data of energy storage status, mechanical stress signals, and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant.

[0071] In step 101, the grid frequency regulation demand information refers to the power regulation command data issued by the grid dispatch system to the virtual power plant, including parameters such as power magnitude and duration. Electricity price trend data reflects real-time price fluctuations in the electricity market and is used to analyze the impact of electricity prices on energy storage behavior. Historical energy storage status data records the changes in the state of charge of the energy storage device over a past period. Mechanical stress signals are collected via sensors to capture vibration waveform data during the operation of the energy storage device. Radio frequency disturbance signals monitor interference waveform data generated by changes in the electromagnetic environment surrounding the energy storage device.

[0072] In this embodiment, real-time frequency regulation command data is first obtained through the power grid dispatch system interface, while electricity price data is collected by connecting to the power market trading platform. Historical operation data is exported from the energy storage device management system. Vibration sensors are installed on the energy storage device to collect mechanical stress signals, and radio frequency monitoring devices are deployed to collect electromagnetic interference signals. Finally, the five types of data are aligned and stored according to a unified time standard.

[0073] For example, during a frequency regulation period, a virtual power plant obtains frequency regulation instructions updated every minute from the dispatch system. For instance, at 10:00, the instruction requires an increase of 0.5 MW of output for 15 minutes. Simultaneously, it collects real-time electricity price data updated every minute from the electricity market. For example, at 10:00, the electricity price is 0.68 yuan per kilowatt-hour. It reads the state of charge value recorded every minute from the energy storage device. For example, at 10:00, it is 65%. It collects mechanical vibration waveform data at 1000 points per second through vibration sensors and electromagnetic interference waveform data at the same sampling rate through radio frequency monitoring devices. All data is stored after being marked with a unified timestamp.

[0074] Step 102: Aggregate the power grid frequency regulation demand information, the electricity price trend data, and the historical data of energy storage status to form an electricity price load response time series database.

[0075] In step 102, the electricity price load response time-series database is a structured data storage format that records the correlation data of frequency regulation commands, electricity prices and energy storage status in chronological order, and is used to analyze the dynamic relationship between market incentives and energy storage response.

[0076] In this embodiment, the acquired frequency regulation command value, real-time electricity price and energy storage status value are bound at the same time point to form a basic data unit. The time window is divided according to the minimum frequency regulation cycle stipulated by the electricity market. Within each window, the data units are arranged in time sequence to form a window-level sequence. Finally, all window sequences are integrated into a three-dimensional structure database according to the time axis.

[0077] For example, frequency regulation commands, electricity prices, and energy storage status data for each minute during the 10:00-10:15 period can be bound together. For instance, the 10:00 data unit contains 0.5 MW, 0.68 yuan, and 65%. The data is divided into 3 windows with 5-minute intervals. Within each window, 5 data units are arranged in chronological order, ultimately forming a time-series database containing 15 data units.

[0078] Step 103: Fuse the mechanical stress signal and the radio frequency disturbance signal to generate a stress electromagnetic coupling fingerprint.

[0079] In step 103, the stress electromagnetic coupling fingerprint is a unique feature identifier generated by fusing device vibration and electromagnetic signals, reflecting the physical state characteristics of the energy storage device.

[0080] In this embodiment, the vibration and electromagnetic signals collected at the same time are processed by waveform alignment and superposition. The zero-crossing time interval sequence is extracted from the mixed waveform to reflect the signal frequency characteristics, and the amplitude envelope is extracted to reflect the signal intensity change. The two types of features are combined and encoded according to preset weights to generate a fingerprint.

[0081] For example, the vibration and electromagnetic signal waveforms at 10:00:00 are superimposed, the zero-crossing interval sequence is extracted, the standard deviation is calculated to be 0.001 seconds, the average value of the amplitude envelope is extracted to be 12.5 millivolts, and the fingerprint code "0.001-12.5" is generated by combining them with a weight of 3 to 7.

[0082] Step 104: Based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, and combined with the Nash equilibrium model, generate charging and discharging competition behavior characteristics.

[0083] In step 104, the charging and discharging competition behavior characteristics are quantitative indicators output by analyzing the interaction between market incentives and physical constraints through the Nash equilibrium model, reflecting the actual operating status of energy storage under the influence of multiple factors.

[0084] In this embodiment, electricity price and energy storage change rate are extracted from time series database as market parameters, and frequency stability and mechanical strength index are extracted from fingerprint as physical parameters. These are respectively input into two sub-models of the Nash equilibrium model. The demand-side sub-model outputs the theoretically optimal strategy, and the supply-side sub-model outputs the practically feasible strategy. Behavioral feature vectors are generated through strategy game interaction.

[0085] For example, the electricity price parameter of 0.68 yuan and the energy storage change rate of -2% / minute in the 10:00-10:05 window are input into the demand-side model to output a 1.2 times charging rate for 30 minutes strategy. The physical parameters are input into the supply-side model to output a feasible range of 0.8-1.0 times charging rate for no more than 20 minutes. The comparison generates a behavioral feature vector [1.2,30,0.8-1.0,20].

[0086] Step 105: Based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, locate the abnormal energy storage nodes in the virtual power plant, and generate an immutable abnormal monitoring report based on the abnormal energy storage nodes through a blockchain smart contract.

[0087] In step 105, an abnormal energy storage node refers to an energy storage device whose behavior significantly deviates from the expected equilibrium state. The anomaly monitoring report is a blockchain-based evidence file that records the abnormal behavior and analysis results.

[0088] In this embodiment, the deviation of behavioral characteristics from the expected state is calculated. When the deviation value continuously exceeds the threshold, it is marked as an abnormal node. The complete operation record of the abnormal period is extracted from the database, reorganized into decision chain data, and the data and deviation value are packaged and uploaded to the chain through a smart contract to generate a monitoring report.

[0089] For example, if the deviation value of a base station's behavior characteristics is 1.85, exceeding the threshold of 1.5, it is marked as an abnormal node. Frequency regulation, electricity price, and status data from 10:05 to 10:07 are extracted to form a decision chain, which is then packaged together with the deviation value and uploaded to the chain to generate a monitoring report containing digital signatures and timestamps.

[0090] This method achieves comprehensive monitoring of abnormal behavior of energy storage resources in virtual power plants through multi-source data acquisition and fusion analysis. The method first establishes a two-dimensional data system encompassing market response and equipment status, then analyzes the interaction between market incentives and physical constraints using an intelligent model, and finally utilizes blockchain technology to ensure the credibility of the monitoring results. This method can accurately identify complex anomalies that are difficult to detect using traditional methods, providing a reliable guarantee for the safe operation of virtual power plants.

[0091] To address the accuracy issue in monitoring energy storage resource anomalies in virtual power plants, in some embodiments, step 104: generating charging and discharging competition behavior characteristics based on the electricity price load response time-series database and the stress electromagnetic coupling fingerprint, combined with a Nash equilibrium model, includes:

[0092] Step 201: Extract the price value from the window-level response sequence of the electricity price load response time series database as the electricity price incentive parameter, and extract the rate of change of state of charge as the energy storage response parameter.

[0093] In step 201, the price value refers to the real-time electricity price updated every minute in the electricity market, directly reflecting the market's incentive for energy storage behavior. The rate of change of state of charge (SOC) is calculated by measuring the rate of change of the remaining electricity of the energy storage device at adjacent time points, characterizing the actual response of the energy storage device to market frequency regulation demands.

[0094] In this embodiment, electricity price values ​​are extracted from a pre-built time-series database according to time windows, and the average value of the energy storage state change per minute within the window is calculated as the rate of change parameter. These two parameters represent the intensity of market incentives and the actual response speed of energy storage, respectively.

[0095] Step 202: Convert the zero-crossing interval sequence in the stress electromagnetic coupling fingerprint into a frequency stability index, and convert the amplitude envelope in the stress electromagnetic coupling fingerprint into a mechanical strength attenuation index.

[0096] In step 202, the frequency stability index is a quantified value obtained by calculating the dispersion of the zero-crossing interval sequence in the stress electromagnetic coupling fingerprint, reflecting the degree of interference in the electromagnetic environment in which the energy storage device is located. The mechanical strength attenuation index is an evaluation value obtained by analyzing the downward trend of the amplitude envelope in the fingerprint, characterizing the fatigue state of the device's mechanical structure.

[0097] In this embodiment, the standard deviation of the zero-crossing interval in the fingerprint data is calculated as a frequency stability index, and the slope of the amplitude envelope is calculated as a mechanical strength attenuation index. These two physical state indices provide constraints for subsequent strategy generation.

[0098] Step 203: Input the energy storage response parameters, the frequency stability index, and the mechanical strength attenuation index into the supply-side constraint sub-model of the Nash equilibrium model to generate a set of feasible behavioral strategies.

[0099] In step 203, the feasible behavioral strategy set refers to the set of charging and discharging schemes that the energy storage device can safely execute after considering the actual physical state limitations. Specifically, it includes the actual operational boundaries such as the maximum charging and discharging power range that the base station energy storage device can withstand and the maximum continuous operating time under various mechanical vibration and electromagnetic interference conditions. These boundaries are jointly determined by the current mechanical strength attenuation degree of the energy storage device and the stability of the electromagnetic environment. For example, when strong mechanical vibration is detected, the feasible strategy will limit the charging and discharging rate to prevent equipment damage. When electromagnetic interference is large, the continuous operating time will be shortened to ensure communication stability. This strategy set essentially reflects the actual operability range of the energy storage device under the constraints of the real physical environment.

[0100] In this embodiment, the energy storage response parameters, frequency stability index, and mechanical strength attenuation index are input into the supply-side constraint sub-model. Through multi-parameter weighted calculation, a feasible strategy range including the maximum allowable charge and discharge rate and the longest duration is output.

[0101] Step 204: Input the electricity price incentive parameters into the demand-side game sub-model of the Nash equilibrium model to generate the expected behavior strategy set.

[0102] In step 204, the expected behavior strategy set is the theoretically optimal set of charging and discharging schemes for energy storage devices when only considering market electricity price incentives, reflecting the economic operation strategy under ideal conditions.

[0103] In this embodiment, real-time electricity price data is input into the demand-side game sub-model, and the theoretically optimal charging and discharging rate and duration strategy is output through profit maximization calculation.

[0104] Step 205: Based on the expected behavior strategy set and the feasible behavior strategy set, output the charging and discharging competition behavior characteristics.

[0105] In this embodiment of the application, the theoretical strategy output by the demand-side model is compared with the feasible strategy output by the supply-side model item by item, and the degree of difference between the two in terms of power change rate and duration is calculated to generate a behavioral feature vector containing conflict type and intensity.

[0106] Here is a specific example:

[0107] During the frequency regulation operation of a virtual power plant, the system generates charging and discharging competition behavior characteristics based on monitoring data from 10:00 to 10:15. First, data for the 10:00-10:05 time window is extracted from the existing time-series database. The 10:00 data unit contains a 0.5 MW frequency regulation command, a price of 0.68 yuan, and a 65% state of charge. The 10:01 data unit contains 0.5 MW, 0.67 yuan, and 63%, up to 10:04, totaling five data units. The energy storage response parameter is calculated as a 2% decrease per minute by averaging the rate of change of state of charge per minute. Simultaneously, the average price of 0.68 yuan over the five minutes within this window is used as the price incentive parameter. For physical state monitoring during the same time period, the system uses the stress electromagnetic coupling fingerprint "0.001-12.5" generated at 10:00:00. The standard deviation of the zero-crossing interval (0.001 seconds) reflects the degree of electromagnetic interference and is directly used as the frequency stability index. The average amplitude envelope (12.5 mV) reflects the mechanical vibration intensity. The mechanical intensity attenuation index of 0.5 is calculated using the formula k = (Vmax - Vmin) / t, where Vmax and Vmin represent the maximum and minimum amplitudes within the window, respectively, and t represents the time window length of 5 minutes. The energy storage response parameter (2% decrease per minute), the frequency stability index (0.001 seconds), and the mechanical intensity attenuation index (0.5) are input into the supply-side constraint sub-model of the Nash equilibrium model. Through multi-parameter weighted calculation, a feasible strategy is output: controlling the charging rate between 0.8 and 1.0 times the standard rate for a duration not exceeding 20 minutes. Simultaneously, the electricity price incentive parameter of 0.68 yuan is input into the demand-side game sub-model. Based on the principle of maximizing returns, the expected strategy is to charge at 1.2 times the standard rate for 30 minutes. Finally, the system compared the two strategy sets and found that the expected strategy's rate of 1.2 times exceeded the feasible range by 0.8-1.0 times, and the 30-minute duration exceeded the 20-minute limit. Therefore, a feature vector of charging and discharging competition behavior [1.2,30,0.8-1.0,20] was generated for this time window. The first two terms represent the theoretical expected strategy parameters, and the last two terms represent the actual feasible strategy range parameters. This feature vector will be used for subsequent anomaly detection. The whole process ensured the organic combination of market incentive analysis and physical state assessment.

[0108] In this embodiment, a comprehensive assessment of the operating status of energy storage devices is achieved by establishing a two-dimensional analysis framework combining market incentives and physical conditions. The method can accurately identify complex abnormal behaviors caused by the combined effects of market and equipment factors, providing a reliable basis for the refined management and safe operation of virtual power plants.

[0109] To further improve the accuracy of charging and discharging behavior feature analysis, in some embodiments, step 205: outputting charging and discharging competition behavior features based on the expected behavior strategy set and the feasible behavior strategy set includes:

[0110] Step 301: Sort the expected behavior strategy set according to the time window of the scheduling cycle to construct a first strategy matrix, where each first row vector corresponds to a theoretical charging and discharging strategy for a time window.

[0111] In step 301, the time window of the scheduling cycle specifically refers to the fixed time unit (e.g., 5 minutes / 15 minutes) for the virtual power plant to participate in electricity market frequency regulation, which is forcibly set by the grid scheduling rules; while the "analysis window" is a manually divided time period during data processing. The two are essentially the same but emphasize different technical attributes (the former emphasizes scheduling constraints, while the latter emphasizes data processing). The first strategy matrix is ​​a data structure arranged in chronological order. Each row vector stores the theoretically optimal charging and discharging strategy determined by market factors within a time window, including key parameters such as charging rate and duration. The theoretical charging and discharging strategy refers to the optimal charging and discharging scheme calculated using the Nash equilibrium model, considering only electricity market price incentives. It includes the standard charging rate and duration that the energy storage device should adopt to obtain maximum economic benefits under ideal market conditions, such as charging continuously for 30 minutes at 1.2 times the standard rate. This strategy does not consider the actual physical limitations of the equipment.

[0112] In this embodiment, the expected strategies for each time period output by the demand-side game sub-model are arranged in chronological order, with each window corresponding to a row of data, forming a matrix structure that reflects the theoretically optimal strategy.

[0113] Step 302: Sort the set of feasible behavior strategies according to the time windows of the same scheduling period to construct a second strategy matrix, where each second row vector corresponds to an executable charging and discharging strategy in the same time window.

[0114] In step 302, the second strategy matrix is ​​a data structure time-aligned with the first strategy matrix. Each row vector stores the range of actually executable charging and discharging strategies within the corresponding time period after considering physical constraints, including parameters such as the upper and lower limits of the rate and the maximum duration. The executable charging and discharging strategy is a safe operation plan generated by a Nash equilibrium model after comprehensively evaluating physical constraints such as the current mechanical strength degradation of the energy storage device and the stability of the electromagnetic environment. It includes the maximum charging and discharging rate range that the device can withstand under actual physical conditions and the threshold for the longest sustainable operation time, such as charging at 0.8 to 1.0 times the rate for no more than 20 minutes.

[0115] In this embodiment of the application, the feasible strategies for each time period output by the supply-side constraint sub-model are arranged in the same time order to ensure that the theoretical strategy and the actual constraint for each window can be compared one by one.

[0116] Step 303: For each time window, perform strategy matching between the theoretical charge / discharge strategy and the executable charge / discharge strategy to generate a set of conflict strategy units.

[0117] In step 303, the conflict strategy unit set is the result of recording the difference between theoretical strategies and actual feasible strategies by comparing them item by item, and includes two basic types: power conflict and duration conflict.

[0118] In this embodiment of the application, the theoretical strategy parameters and feasible strategy parameters of each time window are compared one by one. When the theoretical charging rate exceeds the feasible range or the theoretical duration exceeds the maximum limit, the window is recorded as a conflict unit.

[0119] Step 304: Based on the distribution density and temporal continuity of the conflict strategy unit set, generate charging and discharging competition behavior characteristics.

[0120] In step 304, distribution density refers to the proportion of conflict time windows to the total number of windows. It is calculated by the ratio of the number of marked windows in the conflict strategy unit set to the total number of scheduling windows (e.g., 10 conflict windows / 100 total windows = density 0.1), reflecting the frequency characteristics of anomalies. Temporal continuity refers to the maximum number of consecutive occurrences of conflict windows, obtained by detecting the length of consecutively marked window sequences in the conflict strategy unit set (e.g., if windows 5-7 are consecutively conflicted, then continuity = 3), characterizing the persistence of anomalies. Both together constitute the spatiotemporal dimension basis for weight 5 anomaly determination.

[0121] In this embodiment of the application, the proportion of conflicts occurring in all time windows is used as the conflict density, and the number of windows with the longest consecutive conflicts is calculated as the continuity index. These two values ​​are combined to form the final behavioral feature vector.

[0122] Here is a specific example:

[0123] In the operation monitoring of a virtual power plant during the 10:00-10:30 time period, the system first constructs two strategy matrices, one for the expected strategy set and the other for the feasible strategy set. The first strategy matrix stores theoretical charging and discharging strategies in 5-minute windows. The row vectors in the 10:00-10:05 window record the 30-minute charging strategy at 1.2 times the demand-side model output, and the row vectors in the 10:05-10:10 window record the 25-minute charging strategy at 1.1 times the demand-side model output, for a total of 6 time windows. The second strategy matrix uses the same time division. The row vectors in the 10:00-10:05 window store feasible strategies at 0.8-1.0 times the demand-side model output, lasting no more than 20 minutes, and the row vectors in the 10:05-10:10 window store strategies at 0.9-1.1 times the demand-side model output, lasting no more than 22 minutes. The system compares strategy parameters window by window. For the 10:00-10:05 window, the theoretical rate of 1.2 times exceeds the feasible range by 0.8-1.0 times, and the theoretical duration of 30 minutes exceeds the 20-minute limit, thus marking it as a composite conflict unit. For the 10:05-10:10 window, the theoretical rate of 1.1 times is within the feasible range of 0.9-1.1 times, but the theoretical duration of 25 minutes exceeds the 22-minute limit, thus marking it as a duration conflict unit. A total of 4 conflict units were found across 6 windows, with two consecutive conflicting windows in the 10:00-10:10 window. The conflict density was calculated to be 4 / 6≈0.67, and the longest consecutive conflicting window count was 2. Finally, a feature vector of charging and discharging competition behavior for this period was generated [0.67, 2]. The conflict density was calculated by dividing the number of conflicting units by the total number of windows, and the continuity index was taken as the maximum number of consecutive conflicting windows.

[0124] In this embodiment, a systematic comparison mechanism between theoretical strategies and actual constraints is established to achieve a refined assessment of the operating status of energy storage devices. This method can accurately identify various conflict situations in charging and discharging behavior, and reflect the severity and duration of the conflict through quantitative indicators, providing a reliable basis for anomaly monitoring.

[0125] To further improve the accuracy of conflict detection, in some embodiments, step 303: for each time window, performing policy matching between the theoretical charge / discharge strategy and the executable charge / discharge strategy to generate a set of conflict policy units includes:

[0126] Step 401: Extract the power change rate and duration parameters from the theoretical charge-discharge strategy.

[0127] In step 401, the power change rate refers to the proportion of change between the planned charge / discharge rate of the energy storage device in the theoretical strategy and the standard rate, reflecting the ideal power adjustment range under market incentives. The duration parameter refers to the recommended continuous charge / discharge operation duration in the theoretical strategy.

[0128] In this embodiment of the application, two key values ​​are extracted from the theoretical strategy data corresponding to each time window: one is the planned change ratio of the charge and discharge rate, and the other is the recommended duration of continuous operation.

[0129] Step 402: Extract the allowable power variation range and duration threshold from the executable charge / discharge strategy.

[0130] In step 402, the allowable power variation range is a safe range for charge and discharge rates determined after considering the physical state of the equipment, including the minimum and maximum allowable variation ratios. The duration threshold is the longest continuous operating time that the equipment can safely operate in its current state.

[0131] In this application embodiment, two constraints are obtained from the feasible strategy data: one is the upper and lower limits of the allowable fluctuation of the charge and discharge rate, and the other is the longest continuous working time that the device can withstand.

[0132] Step 403: When the power change rate is less than the minimum value or greater than the maximum value of the allowable power change range, mark the current time window as a power conflict strategy unit.

[0133] In step 403, the power conflict strategy unit refers to the abnormal time window record where the theoretical charge and discharge rate exceeds the actual capacity of the equipment, indicating that the market incentive requirements do not match the equipment capacity.

[0134] In this embodiment of the application, the theoretical rate is compared with the feasible rate range, and when the theoretical value is lower than the minimum value or higher than the maximum value, the window is marked as having a power conflict.

[0135] Step 404: When the duration parameter is greater than the duration threshold, mark the current time window as a duration conflict strategy unit.

[0136] In step 404, the duration conflict strategy unit refers to the abnormal time window record where the theoretical running time exceeds the safe operating time of the equipment, reflecting that the planned operation time exceeds the physical limit.

[0137] In this embodiment of the application, the theoretical duration is compared with the maximum allowed duration. When the theoretical value is greater than the threshold, the window is marked as having a duration conflict.

[0138] Step 405: If both power conflict and duration conflict exist simultaneously, the current time window is remarked as a composite conflict strategy unit.

[0139] In step 405, the composite conflict strategy unit refers to a severe abnormal time window where both power and duration conflict simultaneously, which requires special attention.

[0140] In this embodiment of the application, each window marked as conflicting is reviewed, and when both power and duration conflict conditions are met, it is upgraded to a composite conflict type.

[0141] Step 406: After traversing all time windows, summarize the time windows that are used as power conflict strategy units, duration conflict strategy units, or composite conflict strategy units to form a conflict strategy unit set.

[0142] In this embodiment of the application, all marked conflict windows are arranged in chronological order and stored according to conflict type to form the final conflict record set.

[0143] Here is a specific example:

[0144] In the operation monitoring of a virtual power plant during the 10:00-10:30 time period, the system performed strategy matching analysis on each of the six 5-minute windows. For the 10:00-10:05 window, the system extracted the 1.2x power change rate and 30-minute duration parameters from the theoretical strategy, and the 0.8-1.0x power change allowable range and 20-minute duration threshold from the feasible strategy. Since 1.2x is greater than the 1.0x upper limit and 30 minutes exceeds the 20-minute threshold, the system first marked this window as a power conflict unit and a duration conflict unit, and then upgraded it to a composite conflict strategy unit. For the 10:05-10:10 window, the 1.1x power change rate of the theoretical strategy is within the feasible range of 0.9-1.1x, but the 25-minute duration exceeds the 22-minute threshold, so it is marked as a duration conflict strategy unit separately. The system continues to analyze the following four windows. In the 10:10-10:15 window, the theoretical strategy's power change rate of 0.9 times is lower than the feasible range of 1.0-1.2 times the minimum value, and it is marked as a power conflict strategy unit. In the 10:15-10:20 window, all parameters are within the feasible range and no marking is made. In the 10:20-10:25 window, the theoretical strategy's power change rate of 1.3 times exceeds the 1.0-1.2 times range, and the 35-minute duration exceeds the 30-minute threshold, and it is marked as a composite conflict strategy unit. There is no conflict in the 10:25-10:30 window. The final set of conflict strategy units includes the 10:00-10:05 composite conflict unit, the 10:05-10:10 duration conflict unit, the 10:10-10:15 power conflict unit, and the 10:20-10:25 composite conflict unit. The composite conflict unit records both power and duration conflicts, the power conflict unit records rate anomalies, and the duration conflict unit records timeouts.

[0145] In this embodiment, a multi-layered conflict detection mechanism is established to achieve refined classification of operational anomalies in energy storage devices. This method can distinguish between different risk levels of single and compound conflicts, providing differentiated decision-making basis for subsequent anomaly handling and effectively improving the safe operation level of the virtual power plant.

[0146] To further improve the systematicness and standardization of data processing, in some embodiments, step 102: aggregating the grid frequency regulation demand information, the electricity price trend data, and the historical energy storage status data to form an electricity price load response time-series database includes:

[0147] Step 501: Bind the instruction value in the power grid frequency regulation demand information, the price value in the electricity price trend data, and the state of charge value in the energy storage status history data according to the same time mark to generate a basic data unit.

[0148] In step 501, the instruction value refers to the specific power regulation requirement issued by the power grid dispatch system to the virtual power plant, including the amount of power to be increased or decreased. The price value refers to the real-time electricity price published in the electricity trading market, reflecting the current electricity supply and demand relationship. The state of charge (SOC) value refers to the percentage of the energy storage device's current remaining capacity relative to its total capacity, characterizing the real-time energy status of the energy storage device. A common timestamp is used to ensure the data synchronization of the frequency regulation instruction, electricity price, and SOC value (e.g., all using the power grid SCADA system timestamp). The basic data unit is a structured data packet formed by associating these three key parameters through the same time point.

[0149] In this application embodiment, the power grid frequency regulation command, market electricity price and energy storage status data at the same time are obtained from different data sources, and these three values ​​are bound with the corresponding timestamps to generate a basic data unit containing complete information.

[0150] Step 502: Divide the continuous time period of the virtual power plant in the power market into multiple analysis windows according to the minimum time granularity of the virtual power plant's participation in frequency regulation in the power market.

[0151] In step 502, the minimum time granularity is the smallest time unit in which the virtual power plant responds to frequency regulation commands in the electricity market, explicitly defined by market rules. For example, if the electricity market requires frequency regulation commands to have a minimum bidding cycle of 5 minutes, then this 5 minutes is the minimum time granularity, directly determining the basis for dividing the analysis window. The continuous time period refers to the continuous operating period during which the virtual power plant participates in electricity market frequency regulation; its continuity is determined by the continuous issuance characteristics of grid dispatch commands. For example, when a virtual power plant accepts a 15-minute-level frequency regulation task from the grid, the continuous time period is the complete cycle from the start to the end of the task (e.g., 0:00-24:00 on the same day), and its continuity is guaranteed by the timestamp sequence of the grid dispatch system. The analysis window is a standard time unit used to organize data, divided into equal-length time periods based on this minimum unit.

[0152] In this embodiment of the application, the entire operating period is divided into several analysis windows of equal length according to the minimum time unit of frequency regulation transactions stipulated by the electricity market, and each window contains a fixed number of basic data units.

[0153] Step 503: Within each analysis window, arrange the basic data units in chronological order to form a window-level response sequence.

[0154] In step 503, the time sequence means arranging the data by timestamp within the window (e.g., 1-second data within 5 minutes sorted by timestamp). The window-level response sequence refers to the data chain formed by arranging all basic data units in a time sequence within a single analysis window, reflecting the continuous changes in frequency regulation demand, electricity price, and energy storage status during that period.

[0155] In this embodiment of the application, the basic data units within each analysis window are arranged in chronological order to form sequence data that reflects the complete operating status of that period.

[0156] Step 504: Integrate the window-level response sequences of all analysis windows according to the time dimension to construct a time series database of electricity price load response.

[0157] In step 504, the time dimension represents the concatenation of all windows in the database along the time axis (e.g., combining 288 5-minute windows into a 24-hour time series structure).

[0158] In this embodiment of the application, the sequence data generated by each analysis window are integrated in chronological order to construct a complete database with a unified structure and time dimension.

[0159] Here is a specific example:

[0160] In the process of a virtual power plant participating in frequency regulation in the electricity market, when processing the operational data for the period from 10:00 to 10:15, the system first obtains the frequency regulation command values ​​updated every minute from the dispatch system. For example, the command at 10:00 requires an increase of 0.5 MW of output for 15 minutes. Simultaneously, it collects the real-time electricity price data released every minute by the electricity market, such as the price at 10:00 being 0.68 yuan per kWh. It also reads the state of charge percentage values ​​recorded every minute from the energy storage device, such as 65% at 10:00. These three values ​​are bound together using a unified time stamp "10:00:00" to generate the first basic data unit containing 0.5 MW, 0.68 yuan, and 65%. Following the 5-minute minimum frequency regulation cycle stipulated by the electricity market, this 15-minute period is divided into three analysis windows: 10:00-10:05, 10:05-10:10, and 10:10-10:15. Within the first analysis window (10:00-10:05), the system arranges the five basic data units generated every minute in chronological order, including 0.5 MW, 0.68 yuan, and 65% at 10:00; 0.5 MW, 0.67 yuan, and 63% at 10:01; and so on, up to 0.5 MW, 0.65 yuan, and 57% at 10:04, forming the complete response sequence for this window. After performing the same operation on the subsequent two analysis windows, the response sequences of the three windows are integrated in chronological order. The final constructed electricity price load response time-series database contains 15 data units arranged strictly in chronological order. Each data unit contains three key parameters: frequency regulation command accurate to the minute, real-time electricity price, and energy storage status.

[0161] In this embodiment, a standardized data processing workflow was used to construct an electricity price-load response database containing complete time-series characteristics and multi-dimensional correlations. This method ensures accurate matching between market data and equipment status, providing a reliable data foundation for subsequent analysis and effectively supporting the refined management and anomaly monitoring of virtual power plants.

[0162] To further improve the accuracy of equipment condition monitoring, in some embodiments, step 103: fusing the mechanical stress signal and the radio frequency disturbance signal to generate a stress electromagnetic coupling fingerprint includes:

[0163] Step 601: Perform waveform sampling on the mechanical stress signal and the radio frequency disturbance signal respectively to obtain stress time-domain waveform data and radio frequency time-domain waveform data.

[0164] In step 601, the stress time-domain waveform data is the mechanical vibration signal of the energy storage device acquired by the vibration sensor, reflecting the structural stress situation during device operation. The radio frequency time-domain waveform data is the electromagnetic interference signal around the device acquired by the electromagnetic sensor, characterizing the electromagnetic environment state of the device.

[0165] In this embodiment, a high-precision sensor is used to synchronously acquire the vibration waveform and electromagnetic interference waveform of the device at a fixed sampling frequency, ensuring the time alignment and data consistency of the two signals.

[0166] Step 602: Superimpose the stress time-domain waveform data and the radio frequency time-domain waveform data at the same time mark to generate mixed waveform data.

[0167] In step 602, the same time stamp refers to the precise time synchronization point (e.g., millisecond-level timestamp alignment) of the mechanical stress signal and the radio frequency disturbance signal during acquisition, used to ensure that the instantaneous waveforms of the two signals can be accurately superimposed; while the "continuous time period" is the market cycle division for the virtual power plant's participation in frequency regulation (e.g., a scheduling window of 5 minutes), used to organize historical data into a time-series database. Both serve two independent technical purposes: signal fusion and data aggregation, with different time scales (milliseconds vs. minutes) and application scenarios. Hybrid waveform data is a new waveform generated by numerically superimposing the mechanical vibration waveform and electromagnetic interference waveform acquired at the same time, containing composite characteristics of the two physical signals.

[0168] In this embodiment of the application, the vibration signal value and electromagnetic signal value at the same time point are added point by point to generate hybrid waveform data containing dual physical characteristics.

[0169] Step 603: Extract the zero-crossing interval sequence and amplitude envelope from the mixed waveform data.

[0170] In step 603, the zero-crossing interval sequence refers to the time difference sequence between two adjacent waveforms crossing the zero-level point in the mixed waveform data, reflecting the frequency stability of the signal. The amplitude envelope refers to the line connecting the peaks of the mixed waveform in each period, reflecting the signal intensity change trend.

[0171] In this embodiment, the time difference between adjacent zero-crossing points is calculated by detecting the moment when the mixed waveform crosses the zero level, and the amplitude envelope is formed by tracking the highest point of each fluctuation cycle.

[0172] Step 604: Combine the standard deviation of the zero-crossing interval sequence and the average value of the amplitude envelope according to a preset weight ratio to generate a stress electromagnetic coupling fingerprint.

[0173] In step 604, the preset weight ratio refers to the allocation of importance of two physical features (frequency stability and signal strength) in the fingerprint encoding based on the device's operating characteristics. For example, a weight ratio of "3:7" indicates that the signal strength feature (70%) has a greater impact on the device status than the frequency stability feature (30%). The specific calculation is as follows: fingerprint value = 0.3 × zero-crossing interval standard deviation + 0.7 × amplitude envelope average value. If the measured standard deviation is 0.001 seconds and the average value is 12.5 millivolts, then the fingerprint value is 0.3 × 0.001 + 0.7 × 12.5 = 8.7503. After standardization, the final fingerprint code "0.001-12.5" is generated.

[0174] In this embodiment, the standard deviation of the zero-crossing interval sequence is calculated to characterize frequency stability, and the average value of the amplitude envelope is calculated to characterize signal strength. The two indicators are combined according to preset weights to form the final fingerprint code.

[0175] Here is a specific example:

[0176] During the frequency regulation operation of a virtual power plant, when monitoring the equipment status at 10:00:00, the system first collects mechanical vibration waveform data at 1000 points per second using vibration sensors, and simultaneously collects electromagnetic interference waveform data at the same sampling rate using an RF monitoring device. After aligning the two waveform data according to the same time points, the corresponding values ​​are added to generate a mixed waveform data, where the value at each point equals the sum of the vibration data value and the electromagnetic interference value. The time intervals between two adjacent zero-crossing points are extracted from the mixed waveform to form a sequence [0.010 seconds, 0.009 seconds, 0.011 seconds]. The standard deviation of this sequence is calculated to be 0.001 seconds, reflecting the stability of the signal frequency. Simultaneously, the peak values ​​of the waveform in each cycle are extracted to form an amplitude envelope [12.5 mV, 13.2 mV, 11.8 mV], and the average value is calculated to be 12.5 mV, reflecting the signal strength level. According to a pre-set weighting ratio of 3 to 7, the frequency stability index of 0.001 seconds and the signal strength index of 12.5 millivolts are combined for calculation, with frequency stability accounting for 30% of the weight and signal strength accounting for 70% of the weight. Using the formula fingerprint value = 0.3 × frequency stability + 0.7 × signal strength, we get 0.3 × 0.001 + 0.7 × 12.5 = 8.7503. After standardization, the stress electromagnetic coupling fingerprint code "0.001-12.5" for that moment is generated.

[0177] In this embodiment, a feature fingerprint that comprehensively reflects the operating status of the equipment is generated by fusing both mechanical vibration and electromagnetic interference physical signals. This method overcomes the limitations of single-signal monitoring, enables multi-dimensional assessment of the physical state of energy storage equipment, and provides a more reliable physical feature basis for anomaly detection.

[0178] To further improve the accuracy and reliability of anomaly monitoring, in some embodiments, step 105: Based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, locating the abnormal energy storage node in the virtual power plant, and generating an immutable anomaly monitoring report based on the abnormal energy storage node through a blockchain smart contract, includes:

[0179] Step 701: Calculate the deviation quantization value between the charging and discharging competition behavior characteristics and the characteristics corresponding to the expected equilibrium state.

[0180] In step 701, the deviation quantification value is a numerical index calculated using mathematical methods to determine the degree of difference between the actual behavioral characteristics and the expected state of the model, reflecting the severity of the deviation of the energy storage device's operating state from the ideal situation.

[0181] In this embodiment, a weighted distance algorithm is used to compare each dimension of the behavioral feature vector with the expected state vector, and a comprehensive deviation score is generated after comprehensively considering the differences in each dimension. Specifically, the generated charging / discharging competition behavioral feature vector is compared with the expected equilibrium state vector output by the Nash equilibrium model, and the weighted sum of the squares of the differences in each dimension is taken as the deviation quantization value. The weight coefficients are preset according to the frequency regulation performance indicators of the virtual power plant. In this embodiment, when the behavioral feature vector of a base station's energy storage is [charging rate 1.2C, discharge depth 80%], and the expected equilibrium state is [1.0C, 75%], with a rate weight of 0.6 and a depth weight of 0.4, the deviation quantization value is calculated as sqrt(0.6*(1.2-1.0)). 2 +0.4*(80-75) 2 =2.02, exceeding the set threshold of 1.5, triggers an anomaly detection.

[0182] Step 702: When the deviation quantization value of any distributed base station exceeds the preset deviation threshold N times in a row, the distributed base station is marked as an abnormal energy storage node, where N is greater than or equal to 2.

[0183] In step 702, the preset deviation threshold is an acceptable deviation limit set based on historical data and device characteristics.

[0184] In this embodiment of the application, each energy storage base station is continuously monitored. When its deviation score exceeds the safety threshold multiple times in a row, the system marks it as an abnormal node and triggers subsequent processing procedures.

[0185] Step 703: Extract the key window-level response sequence corresponding to the abnormal energy storage node from the electricity price load response time series database.

[0186] In step 703, the critical window level response sequence refers to the complete operational data record of the abnormal node during the abnormal period, including the time-series changes of key parameters such as frequency regulation, electricity price, and status.

[0187] In this embodiment of the application, all window data of the abnormal node during the period of the abnormality are extracted from the constructed time series database to ensure the integrity and traceability of the abnormality analysis.

[0188] Step 704: Reassemble the state of charge values, electricity price values, and frequency regulation command values ​​in the key window level response sequence into anomaly decision chain data in chronological order.

[0189] In step 704, the abnormal decision chain data is a structured record formed by reorganizing the key parameters of the abnormal period in chronological order, which clearly shows the trajectory of decision changes during the abnormal occurrence process.

[0190] In this embodiment of the application, the frequency regulation commands, real-time electricity prices and energy storage status data for each minute during the abnormal period are arranged in chronological order to form a complete chain of evidence for abnormal behavior.

[0191] Step 705: Through a blockchain smart contract, the abnormal decision chain data and the corresponding deviation quantification value are reliably stored and an anomaly monitoring report is generated.

[0192] In this embodiment of the application, abnormal decision chain data and deviation scores are packaged by smart contracts to generate a monitoring report with timestamps and digital signatures and written to the blockchain.

[0193] Here is a specific example:

[0194] In the operation monitoring of a virtual power plant from 10:00 to 11:00, the system continuously tracks the charging and discharging behavior characteristics of each base station. For base station A, the system first calculates the deviation between its generated behavioral feature vector [1.2,30,0.8-1.0,20] from the expected equilibrium state [1.0,25,0.9-1.1,25] during the period from 10:00 to 10:15, using the weighted Euclidean distance formula d=sqrt[w1(a1-b1)]. 2 +w2(a2-b2) 2The calculated deviation value is 1.85, where w1 = 0.6 and w2 = 0.4 are preset weights, a1 = 1.2 and a2 = 30 are actual feature values, and b1 = 1.0 and b2 = 25 are expected values. When the deviation values ​​of this base station in the subsequent three consecutive periods of 10:15-10:30 and 10:30-10:45 are 1.78, 1.82, and 1.89 respectively, all exceeding the preset threshold of 1.5, the system will mark base station A as an abnormal energy storage node. Subsequently, key data for this node during the abnormal period of 10:00-10:45 was extracted from the time-series database. This included frequency regulation commands recorded every minute, such as 0.8 MW at 10:05, electricity price, such as 0.68 yuan at 10:05, and state of charge, such as 65% at 10:05. These values ​​were then reorganized into an abnormal decision chain "10:00:0.5:0.68:65 to 10:01:0.5:0.67:63 to ... to 10:44:0.9:0.72:42". Finally, through a smart contract deployed on the blockchain, the decision chain is packaged with the average deviation of 1.84 over four time periods to generate an anomaly monitoring report containing the timestamp "2023-05-01 11:00:00" and a digital signature. The average deviation of 1.84 is the arithmetic mean of the deviations of 1.85, 1.78, 1.82, and 1.89 over the four time periods. This report is permanently recorded on the blockchain for subsequent traceability and verification. The entire process ensures the consistency and reliability of data throughout the entire process from anomaly detection to evidence storage.

[0195] In this embodiment, by establishing strict anomaly detection criteria and a blockchain-based evidence storage mechanism, reliable identification and tamper-proof recording of abnormal behavior of energy storage devices are achieved. This method ensures both the accuracy of anomaly detection and the reliability of monitoring results, providing effective support for the safe operation and management of virtual power plants.

[0196] Figure 2 This application provides a schematic diagram of the structure of a distributed base station energy storage resource virtual power plant anomaly monitoring system, as shown in the embodiment of the present application. Figure 2 As shown, the system includes:

[0197] The acquisition module 21 is used to acquire grid frequency regulation demand information, electricity price trend data, energy storage status historical data, mechanical stress signals and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant.

[0198] The aggregation module 22 is used to aggregate the power grid frequency regulation demand information, the electricity price trend data, and the energy storage status historical data to form an electricity price load response time series database.

[0199] The fusion module 23 is used to fuse the mechanical stress signal and the radio frequency disturbance signal to generate a stress electromagnetic coupling fingerprint.

[0200] The generation module 24 is used to generate charging and discharging competition behavior features based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model.

[0201] The positioning module 25 is used to locate abnormal energy storage nodes in the virtual power plant based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, and to generate an immutable abnormal monitoring report based on the abnormal energy storage nodes through a blockchain smart contract.

[0202] Figure 2 The aforementioned distributed base station energy storage resource virtual power plant anomaly monitoring system can perform... Figure 1 The implementation principle and technical effects of the distributed base station energy storage resource virtual power plant anomaly monitoring method described in the illustrated embodiment will not be repeated here. The specific operation methods of each module and unit in the distributed base station energy storage resource virtual power plant anomaly monitoring system described in the above embodiments have been described in detail in the relevant method embodiments, and will not be elaborated upon here.

[0203] In one possible design, Figure 2 The distributed base station energy storage resource virtual power plant anomaly monitoring system of the embodiment shown can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0204] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0205] The processing component 32 is used to perform the above. Figure 1 The embodiment describes a method for monitoring anomalies in a distributed base station energy storage resource virtual power plant.

[0206] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0207] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0208] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0209] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0210] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0211] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0212] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a method for anomaly monitoring of a distributed base station energy storage resource virtual power plant.

[0213] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0214] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0215] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0216] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for anomaly monitoring of distributed base station energy storage resource virtual power plant, characterized in that, include: Acquire grid frequency regulation demand information, electricity price trend data, historical energy storage status data, mechanical stress signals, and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant; The power grid frequency regulation demand information, the electricity price trend data, and the historical energy storage status data are aggregated to form an electricity price load response time-series database. By fusing the mechanical stress signal and the radio frequency disturbance signal, a stress electromagnetic coupling fingerprint is generated; Based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model, charging and discharging competition behavior characteristics are generated; Based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, abnormal energy storage nodes in the virtual power plant are located. Based on the abnormal energy storage nodes, an immutable abnormal monitoring report is generated through a blockchain smart contract.

2. The method according to claim 1, characterized in that, The generation of charging and discharging competition behavior characteristics based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model, includes: The price value is parsed from the window-level response sequence of the electricity price load response time series database as the electricity price incentive parameter, and the rate of change of state of charge is parsed as the energy storage response parameter. The zero-crossing interval sequence in the stress electromagnetic coupling fingerprint is converted into a frequency stability index, and the amplitude envelope in the stress electromagnetic coupling fingerprint is converted into a mechanical strength attenuation index. The energy storage response parameters, the frequency stability index, and the mechanical strength attenuation index are input into the supply-side constraint sub-model of the Nash equilibrium model to generate a set of feasible behavioral strategies. The electricity price incentive parameters are input into the demand-side game sub-model of the Nash equilibrium model to generate a set of expected behavioral strategies. Based on the expected behavior strategy set and the feasible behavior strategy set, the charging and discharging competition behavior characteristics are output.

3. The method according to claim 2, characterized in that, The step of outputting charging and discharging competition behavior characteristics based on the expected behavior strategy set and the feasible behavior strategy set includes: The expected behavior strategy set is sorted according to the time window of the scheduling cycle to construct a first strategy matrix, where each first row vector corresponds to a theoretical charging and discharging strategy for a time window. The feasible behavior strategy set is sorted according to the time window with the same scheduling period to construct a second strategy matrix, where each second row vector corresponds to an executable charging and discharging strategy in the same time window; For each time window, the theoretical charging / discharging strategy is matched with the executable charging / discharging strategy to generate a set of conflicting strategy units; Based on the distribution density and temporal continuity of the conflict strategy unit set, charging and discharging competition behavior characteristics are generated.

4. The method according to claim 3, characterized in that, For each time window, the theoretical charge / discharge strategy is matched with the executable charge / discharge strategy to generate a set of conflicting strategy units, including: The power change rate and duration parameters are extracted from the theoretical charge-discharge strategy. Extract the allowable power variation range and duration threshold from the executable charge / discharge strategy; When the power change rate is less than the minimum value or greater than the maximum value of the allowable power change range, the current time window is marked as a power conflict strategy unit; When the duration parameter is greater than the duration threshold, the current time window is marked as a duration conflict strategy unit; If both power conflict and duration conflict exist simultaneously, the current time window will be remarked as a composite conflict strategy unit. After traversing all time windows, the time windows that are used as power conflict strategy units, duration conflict strategy units, or composite conflict strategy units are summarized to form a conflict strategy unit set.

5. The method according to claim 1, characterized in that, The aggregation of the power grid frequency regulation demand information, the electricity price trend data, and the historical data of energy storage status forms an electricity price load response time-series database, including: The command value in the power grid frequency regulation demand information, the price value in the electricity price trend data, and the state of charge value in the energy storage status history data are bound together with three parameters according to the same time mark to generate a basic data unit. Based on the minimum time granularity of virtual power plants participating in electricity market frequency regulation, the continuous time period of virtual power plants in the electricity market is divided into multiple analysis windows; Within each analysis window, the basic data units are arranged in chronological order to form a window-level response sequence; The window-level response sequences of all analysis windows are integrated according to the time dimension to construct a time-series database of electricity price load response.

6. The method according to claim 1, characterized in that, The process of fusing the mechanical stress signal and the radio frequency perturbation signal to generate a stress electromagnetic coupling fingerprint includes: The mechanical stress signal and the radio frequency disturbance signal are sampled respectively to obtain stress time-domain waveform data and radio frequency time-domain waveform data; The stress time-domain waveform data and the radio frequency time-domain waveform data marked at the same time are superimposed to generate hybrid waveform data; Extract the zero-crossing interval sequence and amplitude envelope from the mixed waveform data; The standard deviation of the zero-crossing interval sequence and the average value of the amplitude envelope are combined according to a preset weight ratio to generate a stress electromagnetic coupling fingerprint.

7. The method according to claim 1, characterized in that, The process involves locating abnormal energy storage nodes in the virtual power plant based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, and generating an immutable anomaly monitoring report based on the abnormal energy storage nodes via a blockchain smart contract, including: Calculate the deviation quantization value between the charging and discharging competition behavior characteristics and the characteristics corresponding to the expected equilibrium state; When the deviation quantization value of any distributed base station exceeds the preset deviation threshold N times in a row, the distributed base station is marked as an abnormal energy storage node, where N is greater than or equal to 2. Extract the key window-level response sequence corresponding to the abnormal energy storage node from the electricity price load response time-series database; The state of charge values, electricity price values, and frequency regulation command values ​​in the key window-level response sequence are reorganized into anomaly decision chain data in chronological order. Through blockchain smart contracts, the abnormal decision chain data and the corresponding deviation quantification value are reliably stored and an anomaly monitoring report is generated.

8. A distributed base station energy storage resource virtual power plant anomaly monitoring system, characterized in that, include: The acquisition module is used to acquire grid frequency regulation demand information, electricity price trend data, energy storage status historical data, mechanical stress signals, and radio frequency disturbance signals from multiple distributed base stations in the virtual power plant. The aggregation module is used to aggregate the power grid frequency regulation demand information, the electricity price trend data, and the historical data of energy storage status to form an electricity price load response time-series database. The fusion module is used to fuse the mechanical stress signal and the radio frequency disturbance signal to generate a stress electromagnetic coupling fingerprint; The generation module is used to generate charging and discharging competition behavior features based on the electricity price load response time series database and the stress electromagnetic coupling fingerprint, combined with the Nash equilibrium model; The positioning module is used to locate abnormal energy storage nodes in the virtual power plant based on the charging and discharging competition behavior characteristics and the expected equilibrium state defined by the Nash equilibrium model, and to generate an immutable abnormal monitoring report based on the abnormal energy storage nodes through a blockchain smart contract.

9. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the method for abnormal monitoring of distributed base station energy storage resource virtual power plant as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The system contains a computer program, which, when executed by a computer, implements a method for abnormal monitoring of distributed base station energy storage resource virtual power plants as described in any one of claims 1 to 7.

Images