Energy storage safety early warning system of smart power grid

Abstract

The invention discloses an energy storage safety early warning system of a smart power grid, and relates to the technical field of smart power grids, the system comprises a parameter acquisition module, a data preprocessing module, a fusion analysis module, a dynamic threshold adjustment module and an early warning output module; multi-mode physical parameters of vibration frequency, sound characteristics and electromagnetic field change of the energy storage equipment are introduced, traditional temperature and electric quantity parameters are combined, comprehensive and deep monitoring of the operation state of the equipment is achieved, the analysis mode of multi-parameter fusion not only enriches data dimensions, but also improves the reliability of the equipment. Meanwhile, by means of the artificial intelligence algorithm, the early warning threshold value can be dynamically adjusted according to equipment operation data and environment data which are collected in real time, so that the method can be more flexibly adapted to changes of the equipment operation state, and the comprehensive and accurate equipment state evaluation and early warning mechanism can be used for improving the accuracy and reliability of equipment state evaluation. And a powerful guarantee is provided for safe operation of an intelligent power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grids, and in particular to an energy storage safety early warning system for a smart grid. Background Art

[0002] With the continuous development of smart grid technology, the safety warning system of energy storage equipment has gradually become a key link in ensuring the stable operation of the power grid. Energy storage equipment plays an important role in energy storage and regulation in smart grids, and its operating status directly affects the overall performance and safety of the power grid. However, due to the complex and changeable working environment of energy storage equipment and the various physical and chemical processes inside it, its operating status is difficult to accurately predict and evaluate. Therefore, it is particularly important to develop a system that can monitor and warn of potential safety hazards of energy storage equipment in real time.

[0003] Traditional energy storage equipment safety warning systems mainly rely on the monitoring of small physical parameters such as temperature and power. Although these parameters can reflect the operating status of the equipment to a certain extent, due to the limited data dimension, it is often difficult to achieve a comprehensive and accurate assessment of the equipment status. In recent years, with the rapid development of sensor technology and artificial intelligence algorithms, it has become possible to introduce multimodal physical parameters such as vibration frequency, sound characteristics, and electromagnetic field changes of energy storage equipment. However, traditional technologies still have shortcomings in the deep fusion analysis of these parameters. On the one hand, the correlation between different parameters has not been fully explored and utilized, resulting in information redundancy and misjudgment. On the other hand, there is a lack of efficient and accurate algorithms to integrate multimodal data to achieve accurate assessment of equipment status.

[0004] In addition, the traditional method of setting warning thresholds also has limitations. Fixed warning thresholds are often unable to adapt to changes in the operating status of energy storage equipment in different environments and under different load conditions, resulting in limited accuracy and reliability of the warning system. Therefore, a method is needed that can dynamically adjust the warning threshold based on real-time data to improve the adaptability and accuracy of the warning system.

[0005] Therefore, developing a smart grid energy storage safety early warning system will effectively improve the safety early warning capabilities of energy storage equipment in the smart grid and provide strong guarantees for the stable operation of the power grid. Summary of the Invention

[0006] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a smart grid energy storage safety early warning system. It can collect multi-modal physical parameters such as temperature, power, vibration frequency, sound characteristics, and electromagnetic field changes through high-precision sensors to achieve comprehensive monitoring of the operating status of energy storage equipment. It also uses artificial intelligence algorithms to perform data fusion analysis, build an equipment operating status model, and dynamically adjust the early warning threshold, thereby improving the accuracy and flexibility of the early warning.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a smart grid energy storage safety early warning system, the system includes the following components:

[0008] The system includes: parameter acquisition module, data preprocessing module, fusion analysis module, dynamic threshold adjustment module and early warning output module;

[0009] The parameter acquisition module: a variety of high-precision sensors are installed on the energy storage device to collect physical parameters such as temperature, power, vibration frequency, sound characteristics, electromagnetic field changes, pressure and humidity. At the same time, a communication connection is established between the sensors and the data processing center through wireless transmission technology to transmit the data to the data processing center;

[0010] The data preprocessing module preprocesses the data from the parameter acquisition module and adopts corresponding filtering algorithms and anomaly detection methods for different types of parameters;

[0011] The fusion analysis module extracts relevant feature data from different parameters through a feature extraction algorithm, and determines the fusion weight according to the importance of different parameter features through a feature fusion weight formula based on the characteristics and importance of the feature data. The formula is: Among them, W i represents the feature fusion weight of the i-th parameter, F i represents the feature importance metric of the i-th parameter, m represents the total number of parameters, and the equipment operation status evaluation formula is selected according to the actual data characteristics and requirements to build the equipment operation status model. The formula is: Among them, S represents the equipment operation status evaluation value, W i is the feature fusion weight, E i Represents the value of the i-th element in the fused feature vector, trains and optimizes the model, and applies the optimized model in practice;

[0012] The dynamic threshold adjustment module: conducts in-depth analysis of the real-time acquisition equipment operating parameters and environmental parameters, explores potential relationships and rules, and selects a dynamic threshold adjustment formula to adjust the dynamic threshold. The formula is: T new =T old ×(1+λ×D), where T new Represents the adjusted warning threshold, T old represents the original warning threshold, λ is the adjustment coefficient, and D represents the measurement value of the change in the environment or equipment operating status. A dynamic threshold adjustment mechanism is established to automatically adjust the warning threshold based on real-time data and algorithm output. At the same time, the adjustment process is monitored and evaluated, and the results are regularly analyzed and summarized to adjust the mechanism and strategy.

[0013] The early warning output module receives the output results of the fusion analysis module and the dynamic threshold adjustment module, determines whether the equipment operating status exceeds the early warning threshold, and when the equipment operating status is abnormal or the parameters exceed the early warning threshold, issues an alarm through sound and light alarms and text messages, and at the same time, records in detail the time when the abnormality occurred and information about the parameters involved.

[0014] Furthermore, the parameter acquisition module includes a temperature sensor, a power sensor, a vibration sensor, a sound sensor, an electromagnetic field change sensor, a pressure sensor and a humidity sensor.

[0015] Furthermore, the data preprocessing module performs filtering on the source data by using a variable weight mean filtering formula, which is: where X filtered Represents the filtered data, X i is the original data point, n is the total number of data points, X is the mean of the original data, and γ is the adjustment coefficient.

[0016] Furthermore, the data preprocessing module determines the detection range of outliers by using the outlier detection threshold formula, which is: AV =μ±k×σ, where T AV It represents the outlier detection threshold range, μ represents the mean of the data, σ represents the standard deviation of the data, and k is the adjustment coefficient, which is adjusted according to the distribution of the actual data and the tolerance of outliers. When the data exceeds this threshold range, it is judged as an outlier.

[0017] Furthermore, the specific steps of training and optimizing the constructed equipment operation status model in the fusion analysis module are as follows:

[0018] (1) Collect sensor data, environment and equipment operating parameters, and perform data cleaning and preprocessing;

[0019] (2) Select a model based on the task characteristics and data type, determine the structure and initialize the parameters;

[0020] (3) Define a loss function to measure the difference between prediction and reality, select an optimization algorithm to update parameters, divide the data set for training, and use the validation set to adjust hyperparameters;

[0021] (4) Evaluate the model performance using the test set and analyze the results to determine whether they meet the requirements;

[0022] (5) Adjust hyperparameters based on the evaluation results and continuously update the model.

[0023] Furthermore, the dynamic threshold adjustment module analyzes the correlation between the equipment operating parameters and the environmental parameters through the correlation coefficient, and the formula is: Where R represents the correlation coefficient between the two parameter data, n represents the number of data points, that is, the number of samples involved in calculating the correlation coefficient, and X i and Y i Represents the i-th data value of the equipment operating parameters and environmental parameters respectively, and Represent the means of the two parameter data respectively.

[0024] Furthermore, the original warning threshold value T in the dynamic threshold adjustment formula in the dynamic threshold adjustment module old ,According to the experimental data, the adjustment coefficient λ is set,first, an initial value is set. By evaluating the system performance under different adjustment coefficients, the measurement value D of the change of environment or equipment operation status is calculated according to the specific environment and equipment operation status parameters, considering the change of ambient temperature, set Where T current is the current ambient temperature, T base As the reference ambient temperature, considering the equipment load change, set Among them L current is the current equipment load, L base is the baseline equipment load.

[0025] Furthermore, the determination of the original warning threshold in the dynamic threshold adjustment module is based on the operation data of the energy storage equipment collected over a period of time. According to the operating status of the equipment and the time period factor, the data set D is divided into several groups. Assuming there are K days in total, the data of each day is a group. For the data of each group, the mean value μ of each parameter is calculated. j and standard deviation σ j , where j represents the parameter type. According to the statistical results of each parameter in each group, the preliminary warning threshold is set. For parameter j, the lower limit of the preliminary warning threshold is T lower,j and upper limit T upper,j Set as: T lower,j =μ j -A×σ j , T upper,j =μ j +A×σ j Where A is a coefficient determined experimentally.

[0026] Furthermore, the warning output module determines the alarm level through the alarm level determination formula, which is: L = B × S + C × (TT new ) Where L represents the alarm level, S is the equipment operation status evaluation value, T represents the current parameter value, that is, the value of a parameter collected in real time, T newis the adjusted warning threshold, B and C are weight coefficients, then it is judged as a low-level alarm, indicating that the equipment operating status has a slight abnormality, but does not pose a serious threat for the time being; it is judged as a medium-level alarm, there are certain potential problems that need to be paid attention to and timely measures should be taken to investigate and deal with them; it is considered a high-level alarm, indicating that the equipment operating status is seriously abnormal.

[0027] Furthermore, the thresholds L1 and L2 in the warning output module are calculated by statistically analyzing all the calculated L values ​​and calculating the mean μ of the L values. L and standard deviation σ L , set L1 = μ L -k1σ L , L2=μ L +k2σ L , where k1 and k2 are coefficients determined according to actual needs.

[0028] Compared with the existing technology, this smart grid energy storage safety early warning system has the following beneficial effects:

[0029] 1. The present invention achieves comprehensive and in-depth monitoring of the equipment's operating status by introducing the multimodal physical parameters of energy storage devices, including vibration frequency, sound characteristics, and electromagnetic field changes, combined with traditional temperature and power parameters. This multi-parameter fusion analysis method not only enriches the data dimension but also improves the accuracy and reliability of equipment status assessment. At the same time, the use of artificial intelligence algorithms can dynamically adjust the warning threshold based on real-time collected equipment operation data and environmental data, thereby more flexibly adapting to changes in equipment operating status. This comprehensive and accurate equipment status assessment and warning mechanism provides a strong guarantee for the safe operation of smart grids and effectively reduces the risk of equipment failures and safety accidents.

[0030] 2. The present invention significantly improves the intelligence and adaptability of the early warning system by introducing multimodal physical parameters and a dynamic threshold setting mechanism. It can monitor equipment operation data and environmental data in real time, and dynamically adjust the early warning threshold based on these data, thereby more accurately reflecting the actual safety status of the equipment. This intelligent early warning system not only improves the accuracy and timeliness of the early warning, but also can make adaptive adjustments according to changes in the equipment's operating status to ensure that the early warning system always remains in the best working state. This improvement in intelligence and adaptability enables the patent of this invention to play an important role in the safe operation of smart grids.

[0031] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0033] Figure 1 An operation diagram of an energy storage safety early warning system for a smart grid;

[0034] Figure 2 This is the flow chart of the fusion analysis module;

[0035] Figure 3 Flowchart of the dynamic threshold adjustment module. DETAILED DESCRIPTION

[0036] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0037] Example 1: Security Early Warning for Energy Storage Systems in Large Data Centers

[0038] A large data center is equipped with a large-scale energy storage system to provide backup power during power outages or power fluctuations, ensuring the continuous operation of critical server equipment. The energy storage system consists of multiple large battery packs, power conversion equipment, and cooling systems, distributed across different areas of the data center.

[0039] The parameter acquisition module installs high-precision temperature sensors at key locations of each battery pack. The measurement data is recorded as T, which can accurately measure the temperature changes of battery cells and the entire battery pack within a range of -20°C to 80°C. The power sensor is installed in the charge and discharge circuit of the battery pack to monitor the battery power level, charge and discharge current, and voltage in real time. The measurement data is recorded as Q. The vibration sensor is installed on the motor moving parts of the power conversion equipment and cooling system. The detection frequency range is 0-1000Hz, and the detection data is recorded as V. The sound sensors are distributed around the energy storage system to capture the sound characteristics emitted by the equipment during operation. The frequency response range is 0-1000Hz. The frequency range is 20Hz-20kHz, and the sound data is recorded as S. The electromagnetic field change sensor is installed near the battery pack and power conversion equipment to monitor the changes in the electromagnetic field intensity. The electromagnetic field data is recorded as E. The pressure sensor is installed in the cooling system pipeline to measure the pressure change of the cooling medium. The range is 0-10MPa, and the pressure data is recorded as P. The humidity sensor is installed in the computer room where the energy storage system is located to monitor the ambient humidity. The range is 0%-100% RH, and the humidity data is recorded as H. All sensors transmit data to the data processing center in real time through wireless Wi-Fi communication technology. The transmission frequency is 1 time per second to ensure the timeliness of the data.

[0040] After receiving the collected data, the data preprocessing module uses the variable weight mean filtering formula for the temperature data T: Perform filtering to effectively remove noise interference in temperature measurement. For the power data Q, use the abnormal value detection threshold formula: T AV =μ±k×σ, detect abnormal values ​​to ensure the accuracy of power data. For vibration data V and sound data S, combine spectrum analysis and time domain feature extraction methods to identify characteristic patterns under normal operating conditions for subsequent abnormality judgment.

[0041] The fusion analysis module extracts key features of each parameter, such as the rate of change of temperature T, the charge and discharge efficiency of power Q, the spectrum peak of vibration V, the frequency distribution characteristics of sound S, the intensity fluctuation of electromagnetic field change E, the change trend of pressure P, and the influence coefficient of humidity H on the equipment. According to the feature fusion weight formula: Calculate the fusion weight of each feature and use the equipment operation status evaluation formula: Build a device operation status model and device operation status evaluation formula to build a device operation status model and initialize the parameters. Collect data from the past month as a training set, of which normal operation status data accounts for 70% and abnormal data (including battery overheating, abnormal power drop, and abnormal equipment vibration) accounts for 30%. Define a loss function to measure the difference between prediction and reality, select an optimization algorithm to update parameters, divide the data set for training and use the validation set to adjust hyperparameters, use the test set to evaluate model performance, analyze the results to determine whether they meet the requirements, adjust hyperparameters based on the evaluation results, continuously update the model, and apply the optimized model to practice.

[0042] Dynamic threshold adjustment module, through the correlation coefficient analysis formula: For the correlation between the real-time acquisition equipment operating parameters and environmental parameters, let the temperature data be X and the battery power consumption speed related data be Y. It is found that there is a certain positive correlation between the temperature T and the battery power consumption speed. When the temperature rises, the battery power consumption is slightly accelerated. Let the vibration frequency data be X and the load change data of the power conversion equipment be Y. The vibration frequency V has a strong positive correlation with the load change of the power conversion equipment. When the load increases, the vibration intensifies. According to the relationship between the equipment operating parameters and the environmental parameters, the warning threshold is adjusted through the dynamic threshold adjustment formula. For example, when the ambient temperature rises from 25℃ to 35℃, let the current ambient temperature be T current , the reference ambient temperature is T base , according to the dynamic threshold adjustment formula: T new =T old ×(1+λ×D),λ=0.1, The temperature warning threshold is reduced by 10% to provide early warning of battery overheating risks. When the device load increases from 50% to 80%, the current device load is set to L current , the benchmark equipment load is L base, The warning thresholds for current and voltage are increased by 15% accordingly, calculated based on load changes.

[0043] The early warning output module determines the alarm level according to the formula: L=B×S+C×(TT new ), judge the alarm level, and assume that the equipment operation status evaluation value S is calculated at a certain moment ′ =0.7, current temperature parameter value T = 40 °C, adjusted temperature warning threshold T new =38℃, then L=0.6×0.7+0.4×(40-38)=0.42+0.8=1.22. If L1=0.5 and L2=1.5 are set, it is judged as a medium-level alarm. The system notifies the operation and maintenance personnel through SMS and displays the alarm information in the data center monitoring system, prompting the operation and maintenance personnel to check whether the battery pack and cooling system are working properly.

[0044] In summary, in large-scale data center energy storage systems, by installing a variety of high-precision sensors at key locations, comprehensive collection of physical parameters such as temperature, power, vibration, sound, electromagnetic field changes, pressure, and humidity is achieved. The data preprocessing module uses the variable weight mean filtering formula and the outlier detection threshold formula to effectively improve data quality. The fusion analysis module uses the feature fusion weight formula and the equipment operation status evaluation formula to construct an accurate operation status model. The dynamic threshold adjustment module flexibly adjusts the warning threshold based on the dynamic threshold adjustment formula in combination with changes in ambient temperature and equipment load. The warning output module issues alarms in a timely manner through the alarm level determination formula, ensuring the safe and stable operation of the data center energy storage system, effectively preventing potential safety issues, and improving the reliability of system operation and timely maintenance.

[0045] Example 2:

[0046] Distributed photovoltaic power station energy storage safety warning

[0047] A distributed photovoltaic power station is equipped with an energy storage system to store excess electricity generated by photovoltaic power generation during the day and provide power to users at night or when there is insufficient sunlight. The energy storage system consists of multiple distributed energy storage units, which are distributed at different locations in the photovoltaic power station and work in conjunction with photovoltaic panels and inverter equipment.

[0048] Parameter acquisition module, a temperature sensor is installed on the battery module of each energy storage unit, the measurement data is recorded as T, and the measurement range is -10℃ to 50℃. The power sensor is used to accurately monitor the battery power, and the power data is recorded as Q. The vibration sensor is installed at the connection between the inverter and the energy storage unit, with a detection frequency range of 0-500Hz and vibration data recorded as V. The sound sensor is installed on the outer casing of the energy storage unit to collect sound characteristics, with a frequency response range of 50Hz-15kHz and sound data recorded as S. The electromagnetic field change sensor is close to the battery and inverter to monitor changes in electromagnetic field intensity, and the electromagnetic field data is recorded as E. The pressure sensor is installed at the sealing part of the energy storage unit to detect internal pressure changes, with a range of 0-5MPa and pressure data recorded as P. The humidity sensor is installed in the surrounding environment of the power station, with a range of 096-9596RH and humidity data recorded as H. The sensor data is transmitted to the data processing center via wireless communication with a transmission rate of 9600bps to ensure stable data transmission.

[0049] The data preprocessing module processes the collected data. For the temperature data T, the variable weight mean filter formula is used: Where γ = 0.3) is used for filtering to effectively remove noise interference in temperature measurement. For the power data Q, the abnormal value detection threshold formula is used: T AV =μ±k×σ, and perform outlier screening.

[0050] The fusion analysis module extracts the characteristics of various parameters, such as the mean temperature T, the slope of change of the power Q, the energy mean of the vibration V, the harmonic components of the sound S, the standard deviation of the electromagnetic field intensity E, the fluctuation range of the pressure P, and the influence factor of the humidity H on the battery performance. According to the feature fusion weight formula: Calculate the fusion weight of each feature and use the equipment operation status evaluation formula: Build a device operation status model and collect data from the past three months as a training set (normal operation data accounts for 80%, abnormal data accounts for 20%, abnormal conditions include battery aging and inverter failure). Define a loss function to measure the difference between prediction and actual situation, select an optimization algorithm to update parameters, divide the data set for training and use the validation set to adjust hyperparameters, use the test set to evaluate model performance, analyze the results to determine whether they meet the requirements, adjust hyperparameters based on the evaluation results, continuously update the model, and apply the optimized model in practice.

[0051] The dynamic threshold adjustment module extracts features from vibration data V and sound data S, such as the energy features of vibration and the harmonic features of sound. Let the vibration data be X and the sound data be Y, and use the correlation coefficient analysis formula: Analyze the correlation between the two to better understand the operating status of the equipment. Analyze the correlation of other parameters. Let the electromagnetic field strength data be X and the inverter working status data be Y. It is found that the change in electromagnetic field strength E is closely related to the working status of the inverter. When the inverter works abnormally, the electromagnetic field strength will fluctuate. Let the humidity data be X and the battery self-discharge rate data be Y. The humidity change H has a certain impact on the battery self-discharge rate. The battery self-discharge increases slightly in a high humidity environment. Adjust the threshold according to the ambient temperature and the equipment operation time. For example, when the ambient temperature is high in summer, let the current ambient temperature be T. current , the reference ambient temperature is T base , according to the dynamic threshold adjustment formula: T new =T old ×(1+λ×D), calculate λ is set based on actual conditions. When the temperature exceeds 30°C, the temperature warning threshold is reduced by 8%. As the device runs for longer, the power warning threshold is dynamically adjusted based on battery degradation. After one year of operation, the power warning threshold is reduced by 5%.

[0052] The early warning output module determines the alarm level. Assume that the equipment operation status evaluation value S = 0.65 at a certain moment, the current power parameter value Q = 30%, and the adjusted power warning threshold T new =28%, according to the alarm level determination formula (L=B×S+B×(QT new), B=0.6, B=0.4), L=0.6×0.65+0.4×(30-28)=0.39+0.8=1.19. If L1=0.4, L2=1.2, it is judged as a medium-level alarm. The system notifies the operation and maintenance personnel through sound and light alarms and text messages, and generates a detailed abnormality report, including abnormal parameters, possible causes and recommended treatment measures, such as checking battery health status and inverter operating parameters. The operation and maintenance personnel perform maintenance and processing in a timely manner according to the alarm information to ensure the safe and stable operation of the distributed photovoltaic power station energy storage system.

[0053] In summary, the distributed photovoltaic power station energy storage system uses various deployed sensors to accurately collect parameter data including temperature, power, vibration, sound, electromagnetic field changes, pressure, and humidity. The data preprocessing module uses corresponding formulas to process data to ensure data reliability. The fusion analysis module builds a model based on the feature fusion weight formula and the equipment operation status assessment formula, and optimizes the dynamic threshold adjustment module through the support vector machine algorithm, using the dynamic threshold adjustment formula to adjust the threshold according to the ambient temperature and equipment operation time. The early warning output module relies on the alarm level determination formula to determine the alarm level, and can quickly notify operation and maintenance personnel in the event of an anomaly, effectively ensuring the normal operation of the distributed photovoltaic power station energy storage system, reducing the risk of failure, and improving the overall performance and safety of the system.

[0054] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A smart grid energy storage safety early warning system, characterized in that: The system includes: parameter acquisition module, data preprocessing module, fusion analysis module, dynamic threshold adjustment module and early warning output module; The parameter acquisition module: a variety of high-precision sensors are installed on the energy storage device to collect physical parameters such as temperature, power, vibration frequency, sound characteristics, electromagnetic field changes, pressure and humidity. At the same time, a communication connection is established between the sensors and the data processing center through wireless transmission technology to transmit the data to the data processing center; The data preprocessing module preprocesses the data from the parameter acquisition module and adopts corresponding filtering algorithms and anomaly detection methods for different types of parameters; The fusion analysis module extracts relevant feature data from different parameters through a feature extraction algorithm, and determines the fusion weight according to the importance of different parameter features through a feature fusion weight formula based on the characteristics and importance of the feature data. The formula is: Among them, W i represents the feature fusion weight of the i-th parameter, F i represents the feature importance metric of the i-th parameter, m represents the total number of parameters, and the equipment operation status evaluation formula is selected according to the actual data characteristics and requirements to build the equipment operation status model. The formula is: Among them, S represents the equipment operation status evaluation value, W i is the feature fusion weight, E i Represents the value of the i-th element in the fused feature vector, trains and optimizes the model, and applies the optimized model in practice; The dynamic threshold adjustment module: conducts in-depth analysis of the real-time acquisition equipment operating parameters and environmental parameters, explores potential relationships and rules, and selects a dynamic threshold adjustment formula to adjust the dynamic threshold. The formula is: T new =T old ×(1+λ×D), where T new Represents the adjusted warning threshold, T old represents the original warning threshold, λ is the adjustment coefficient, and D represents the measurement value of the change in the environment or equipment operating status. A dynamic threshold adjustment mechanism is established to automatically adjust the warning threshold based on real-time data and algorithm output. At the same time, the adjustment process is monitored and evaluated, and the results are regularly analyzed and summarized to adjust the mechanism and strategy. The early warning output module receives the output results of the fusion analysis module and the dynamic threshold adjustment module, determines whether the equipment operating status exceeds the early warning threshold, and when the equipment operating status is abnormal or the parameters exceed the early warning threshold, issues an alarm through sound and light alarms and text messages, and at the same time, records in detail the time when the abnormality occurred and information about the parameters involved.

2. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The parameter acquisition module includes a temperature sensor, a power sensor, a vibration sensor, a sound sensor, an electromagnetic field change sensor, a pressure sensor and a humidity sensor.

3. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The data preprocessing module filters the source data using a variable weight mean filter formula, which is: where X filtered Represents the filtered data, X i is the original data point, n is the total number of data points, X is the mean of the original data, and γ is the adjustment coefficient.

4. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The data preprocessing module determines the detection range of outliers by using the outlier detection threshold formula, which is: AV =μ±k×σ, where T AV It represents the outlier detection threshold range, μ represents the mean of the data, σ represents the standard deviation of the data, and k is the adjustment coefficient, which is adjusted according to the distribution of the actual data and the tolerance of outliers. When the data exceeds this threshold range, it is judged as an outlier.

5. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The specific steps for training and optimizing the constructed equipment operation status model in the fusion analysis module are as follows: (1) Collect sensor data, environment and equipment operating parameters, and perform data cleaning and preprocessing; (2) Select a model based on the task characteristics and data type, determine the structure and initialize the parameters; (3) Define a loss function to measure the difference between prediction and reality, select an optimization algorithm to update parameters, divide the data set for training, and use the validation set to adjust hyperparameters; (4) Evaluate the model performance using the test set and analyze the results to determine whether they meet the requirements; (5) Adjust hyperparameters based on the evaluation results and continuously update the model.

6. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The dynamic threshold adjustment module analyzes the correlation between the equipment operating parameters and the environmental parameters through the correlation coefficient, and the formula is: Where R represents the correlation coefficient between the two parameter data, n represents the number of data points, that is, the number of samples involved in calculating the correlation coefficient, and X i and Y i Represents the i-th data value of the equipment operating parameters and environmental parameters respectively, and Represent the mean of the two parameter data respectively.

7. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The original warning threshold T in the dynamic threshold adjustment formula of the dynamic threshold adjustment module old ,According to the experimental data, the adjustment coefficient λ is set,first, an initial value is set. By evaluating the system performance under different adjustment coefficients, the measurement value D of the change of environment or equipment operation status is calculated according to the specific environment and equipment operation status parameters, considering the change of ambient temperature, set Where T current is the current ambient temperature, T base As the reference ambient temperature, considering the equipment load change, set Among them L current is the current equipment load, L base is the baseline equipment load.

8. The energy storage safety early warning system for a smart grid according to claim 7, characterized in that: The determination of the original warning threshold in the dynamic threshold adjustment module is based on the operation data of the energy storage equipment collected over a period of time. According to the operating status of the equipment and the time period factor, the data set D is divided into several groups. Assuming there are K days in total, the data of each day is a group. For the data of each group, the mean value μ of each parameter is calculated. j and standard deviation σ j , where j represents the parameter type. According to the statistical results of each parameter in each group, the preliminary warning threshold is set. For parameter j, the lower limit of the preliminary warning threshold is T lower,j and upper limit T upper,j Set as: T lower,j =μ j -A×σ j , T upper,j =μ j +A×σ j Where A is a coefficient determined experimentally.

9. The energy storage safety early warning system for a smart grid according to claim 1, characterized in that: The warning output module determines the alarm level through the alarm level determination formula, which is: L = B × S + C × (TT new ) Where L represents the alarm level, S is the equipment operation status evaluation value, T represents the current parameter value, that is, the value of a parameter collected in real time, T new is the adjusted warning threshold, B and C are weight coefficients, then it is judged as a low-level alarm, indicating that the equipment operating status has a slight abnormality, but does not pose a serious threat for the time being; it is judged as a medium-level alarm, there are certain potential problems that need to be paid attention to and timely measures should be taken to investigate and deal with them; it is considered a high-level alarm, indicating that the equipment operating status is seriously abnormal.

10. The energy storage safety early warning system for a smart grid according to claim 9, characterized in that: The thresholds L1 and L2 in the warning output module are calculated, and statistical analysis is performed on all the calculated L values ​​to calculate the mean μ of the L values. L and standard deviation σ L , set L1 = μ L -k1σ L , L2=μ L +k2σ L , where k1 and k2 are coefficients determined according to actual needs.

Images