A remote security monitoring method and system for shared power bank cabinets

Abstract

This invention relates to the field of shared equipment security monitoring technology, and discloses a remote security monitoring method and system for shared power bank cabinets. The method includes: collecting and preprocessing operational data; extracting peak temperature rise and usage duration to group equipment status; screening abnormal equipment and extracting abnormal temperature rise intervals, calculating the correlation characteristics between temperature and current, and generating a preliminary abnormal list; acquiring historical and real-time fluctuation sequences and calculating correlation strength scores; calculating a comprehensive risk index based on the scores and fluctuation amplitudes, and generating a priority monitoring list; increasing the data collection frequency and analyzing dynamic health status; when the health status remains abnormal, generating and verifying alarm signals to trigger equipment control commands; executing commands and updating the risk index to generate a high-risk equipment monitoring list. This invention achieves accurate early warning and closed-loop management of security risks in shared power bank cabinets through multi-dimensional data analysis and dynamic risk assessment.

Description

Technical Field

[0001] This invention relates to the field of shared charging equipment monitoring technology, and in particular to a remote security monitoring method and system for shared power bank cabinets. Background Technology

[0002] Currently, the remote security monitoring capabilities of shared power bank cabinets are directly related to user service experience and public safety. With the continuous expansion of equipment deployment and the increasing complexity of application scenarios, traditional monitoring methods face severe challenges in terms of response speed and judgment accuracy when dealing with security risks in industrial big data environments.

[0003] One existing technology primarily employs a monitoring scheme that combines fixed threshold alarms with periodic manual inspections. This scheme collects operational data by deploying basic sensors within the cabinet and sets static safety thresholds for key parameters; alarms are triggered when data exceeds these limits. Simultaneously, maintenance personnel conduct on-site inspections of the equipment at predetermined intervals to confirm its operational status. However, this method struggles to address the demands of fusing and analyzing multi-source heterogeneous data and lacks a deep understanding of the equipment's operational status.

[0004] This monitoring method, which relies on static rules and manual intervention, has inherent limitations. Because it lacks the ability to effectively extract and analyze the characteristics of time-series data, it cannot effectively capture early features that characterize the gradual degradation of equipment performance, nor can it perform in-depth correlation analysis of multi-dimensional operational data. This results in insufficient sensitivity and low specificity in identifying complex failure modes.

[0005] Therefore, the core engineering problem facing existing technologies is how to achieve accurate assessment of battery health status and timely early warning of safety risks in a large-scale deployment of shared power bank clusters through intelligent analysis of multi-source heterogeneous operating data, so as to overcome the shortcomings of traditional fixed threshold monitoring methods, such as slow response and high false alarm rate. Summary of the Invention

[0006] This invention provides a remote security monitoring method and system for shared power bank cabinets, which solves the technical problems of existing technologies that rely on fixed thresholds and manual inspections, resulting in delayed early warning of potential risks, high misjudgment rate of complex faults, and difficulty in supporting refined monitoring of large-scale equipment clusters.

[0007] Firstly, to address the aforementioned technical problems, a remote security monitoring method for shared power bank cabinets includes: The operation data of the shared power bank cabinet is collected and preprocessed to obtain an operation data set, which includes battery temperature data, charging current data, usage time data and ambient temperature data. Extract the battery temperature rise peak and usage time data from the operating data set, group the device into status groups based on the temperature rise peak and usage time data, and generate grouping results; wherein the temperature rise peak is calculated based on the battery temperature data; From the grouping results, devices in abnormal condition are screened, the abnormal temperature rise range of the devices in abnormal condition is extracted, and the correlation characteristics between temperature change and current change within the abnormal temperature rise range are calculated. Potential abnormal devices are marked according to the correlation characteristics to generate a preliminary abnormal list. Obtain the historical fluctuation sequence and the first real-time fluctuation sequence of the temperature rise data of the equipment in the preliminary anomaly list, and calculate the correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence; Based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, a comprehensive risk index is calculated, high-risk devices are marked and listed as priority monitoring targets, and a priority monitoring list is generated. For the devices in the priority monitoring list, increase the data collection frequency, obtain the second real-time fluctuation sequence, calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determine the dynamic health status based on the similarity. When the dynamic health status shows a continuous abnormality, environmental correction is performed on the data that triggered the continuous abnormality based on the ambient temperature data. An alarm signal is generated based on the corrected data and a reliability verification is performed. If the verification is successful, a device control command is generated based on the alarm signal. Execute the equipment control command, collect the operating data of the controlled equipment and update the comprehensive risk index to obtain an updated risk index, mark high-risk equipment according to the updated risk index, and obtain a high-risk equipment monitoring list.

[0008] In one optional implementation, the collection and preprocessing of operational data from the shared power bank cabinet yields an operational data set, which includes battery temperature data, charging current data, usage duration data, and ambient temperature data, including: The shared power bank cabinets use built-in sensors to collect data on battery temperature, charging current, usage time, and ambient temperature. The collected data is formatted and outliers are removed to form a standardized set of operational data.

[0009] In one optional implementation, the step of extracting the battery's peak temperature rise and usage duration data from the operational data set, and grouping the device into statuses based on the peak temperature rise and usage duration data to generate grouping results includes: Extract the battery's peak temperature rise and usage duration data from the operational data set; Based on the peak temperature rise and usage duration data, K-means clustering is used to perform cluster analysis on the equipment. According to the clustering results, the equipment is divided into different state groups to obtain the grouping results.

[0010] In one optional implementation, the step of filtering abnormal state devices from the grouping results, extracting the abnormal temperature rise range of the abnormal state devices, calculating the correlation characteristics between temperature changes and current changes within the abnormal temperature rise range, marking potential abnormal devices based on the correlation characteristics, and generating a preliminary abnormal list includes: Filter devices in abnormal condition from the grouping results; Acquire the temperature rise change data of the abnormal state device, and identify the interval where the temperature rise change exceeds the preset change threshold and continues to exist, as the abnormal temperature rise interval; Extract the current change data of the abnormal temperature rise range from the charging current data, calculate the Pearson correlation coefficient between the temperature change data and the current change data in the abnormal temperature rise range, and obtain the correlation characteristics between temperature change and current change. Potentially abnormal devices are marked based on the abnormal temperature rise range and the associated characteristics, and a preliminary list of abnormalities is generated.

[0011] In one optional implementation, the step of obtaining historical fluctuation sequences of devices in the preliminary anomaly list, generating a first real-time fluctuation sequence based on currently collected temperature data, and generating a correlation strength score by analyzing the differences between the first real-time fluctuation sequence and the historical fluctuation sequence through time series comparison analysis includes: The temperature rise fluctuation data of the equipment in the preliminary anomaly list within the historical normal cycle is obtained from the equipment historical database to obtain the historical fluctuation sequence. Based on the temperature rise change data, a first real-time fluctuation sequence is generated; The correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence is calculated using a dynamic time warping algorithm.

[0012] In one optional implementation, the step of calculating a comprehensive risk index based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, marking high-risk devices as priority monitoring targets, and generating a priority monitoring list includes: The correlation strength score is weighted with the fluctuation amplitude of the first real-time fluctuation sequence to generate a comprehensive risk index; Based on the comprehensive risk index, the devices in the preliminary anomaly list are sorted in descending order to obtain the risk ranking result; Based on the risk ranking results, devices whose comprehensive risk index exceeds a preset risk threshold are marked as high-risk devices and listed as priority monitoring targets, generating a priority monitoring list.

[0013] In one optional implementation, calculating the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determining the dynamic health status based on the similarity, includes: Based on the second real-time fluctuation sequence and the historical fluctuation sequence, the similarity value between the two is calculated using cosine similarity. Based on the preset range to which the similarity value belongs, the deviation level of the device's health status is determined; A corresponding dynamic health status is generated based on the deviation level.

[0014] In one optional implementation, when the dynamic health status shows a persistent abnormality, environmental correction is performed on the data triggering the persistent abnormality based on the ambient temperature data. An alarm signal is generated based on the corrected data and its reliability is verified. If the verification passes, a device control command is generated based on the alarm signal, including: Based on the ambient temperature data, the temperature rise data that triggers the continuous abnormality is compensated and corrected, and the corrected data is compared with the preset alarm threshold to generate an alarm signal. The data sequence corresponding to the alarm signal is verified for data integrity and data continuity. When the verification result meets the preset requirements, the alarm signal is determined to be reliable, and a device control command is generated based on the alarm signal.

[0015] In one optional implementation, the step of executing the equipment control command, collecting the operating data of the controlled equipment and updating the comprehensive risk index to obtain an updated risk index, and marking high-risk equipment according to the updated risk index to obtain a high-risk equipment monitoring list, includes: Execute the equipment control command, collect the operating data of the controlled equipment, and obtain updated operating data; Based on the updated operational data, the comprehensive risk index of the device is updated to obtain the updated risk index; Devices whose updated risk index exceeds a preset high-risk threshold are marked as high-risk devices, and a high-risk device monitoring list is generated.

[0016] Secondly, the present invention provides a remote security monitoring system for shared power bank cabinets, comprising: The data acquisition module is used to collect and preprocess the operating data of the shared power bank cabinets to obtain an operating data set. The status grouping module is used to extract the battery's temperature rise peak and the usage time data from the operating data set, and to group the device status based on the temperature rise peak and usage time data to generate grouping results; An abnormal feature module is used to filter devices in abnormal state from the grouping results, extract the abnormal temperature rise range of the abnormal devices, calculate the correlation features between temperature change and current change within the abnormal temperature rise range, and mark potential abnormal devices according to the correlation features to generate a preliminary abnormal list. The correlation assessment module is used to obtain the historical fluctuation sequence and the first real-time fluctuation sequence of the temperature rise data of the equipment in the preliminary anomaly list, and to calculate the correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence. The risk ranking module is used to calculate a comprehensive risk index based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, mark high-risk devices as priority monitoring targets, and generate a priority monitoring list. The health analysis module is used to increase the data collection frequency for devices in the priority monitoring list, obtain a second real-time fluctuation sequence, calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determine the dynamic health status based on the similarity. The alarm control module is used to perform environmental correction on the data that triggers the continuous abnormality based on the ambient temperature data when the dynamic health status shows a continuous abnormality, generate an alarm signal based on the corrected data and perform reliability verification, and generate equipment control instructions based on the alarm signal if the verification is successful. The closed-loop management module is used to execute the equipment control commands, collect the operating data of the controlled equipment and update the comprehensive risk index to obtain an updated risk index, mark high-risk equipment according to the updated risk index, and obtain a high-risk equipment monitoring list.

[0017] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the remote security monitoring method for shared power bank cabinets as described in any one of the above.

[0018] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the remote security monitoring method for shared power bank cabinets described in any one of the above-mentioned methods.

[0019] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention constructs an early risk identification mechanism based on multi-source data fusion by acquiring multi-dimensional operational data and grouping states, combined with the extraction of temperature rise anomaly intervals and current correlation features. The technical derivation of this method lies in the fact that existing technologies rely on fixed thresholds and cannot perceive the gradual degradation of equipment performance. This invention, through cluster analysis and correlation feature calculation, can accurately locate potential abnormal individuals from a large number of devices, thereby solving the problem of delayed early risk warning in traditional methods and significantly improving the foresight of safety monitoring.

[0020] (2) This invention establishes a closed-loop decision-making system from risk identification to accurate alarm by analyzing the correlation strength through time series comparison, combined with dynamic health status assessment and alarm signal reliability verification. The technical derivation of this method lies in the fact that a single indicator is prone to misjudgment under complex operating environments. This invention achieves accurate identification and reliable early warning of high-risk equipment through multi-dimensional risk assessment and signal cross-verification, thereby effectively reducing the false alarm rate and missed alarm rate of the system.

[0021] (3) This invention forms an adaptive and optimized long-term monitoring mechanism by dynamically adjusting monitoring strategies and risk assessment thresholds. The technical derivation of this method lies in the fact that the equipment cluster is large in scale and its operating status changes dynamically. This invention realizes intelligent allocation of monitoring resources and continuous optimization of risk models through hierarchical monitoring and risk index updates, thereby supporting the refined security management of large-scale equipment clusters and significantly improving operation and maintenance efficiency. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the remote security monitoring method for shared power bank cabinets provided in the first embodiment of the present invention; Figure 2 This is a flowchart of the core algorithm of the remote security monitoring method for shared power bank cabinets provided in the first embodiment of the present invention; Figure 3 This is a schematic diagram of the remote security monitoring method for shared power bank cabinets provided in the second embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Reference Figure 1 The first embodiment of the present invention provides a remote security monitoring method for shared power bank cabinets, comprising the following steps: S11. Collect and preprocess the operating data of the shared power bank cabinet to obtain an operating data set; wherein the operating data includes battery temperature data, charging current data, usage time data and ambient temperature data. S12, extract the battery temperature rise peak and the usage time data from the operating data set, group the device into states based on the temperature rise peak and the usage time data, and generate grouping results; wherein the temperature rise peak is calculated based on the battery temperature data; S13, filter out abnormal state devices from the grouping results, extract the abnormal temperature rise range of the abnormal state devices, calculate the correlation characteristics between temperature change and current change within the abnormal temperature rise range, mark potential abnormal devices according to the correlation characteristics, and generate a preliminary abnormal list. S14, obtain the historical fluctuation sequence and the first real-time fluctuation sequence of the temperature rise data of the equipment in the preliminary anomaly list, and calculate the correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence; S15, Calculate the comprehensive risk index based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, mark high-risk devices as priority monitoring targets, and generate a priority monitoring list; S16, for the devices in the priority monitoring list, increase the data acquisition frequency, obtain the second real-time fluctuation sequence, calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determine the dynamic health status based on the similarity; S17, when the dynamic health status shows a continuous abnormality, environmental correction is performed on the data that triggered the continuous abnormality based on the ambient temperature data, an alarm signal is generated based on the corrected data and reliability verification is performed, and if the verification is successful, an equipment control command is generated based on the alarm signal. S18, execute the equipment control command, collect the operating data of the controlled equipment and update the comprehensive risk index to obtain an updated risk index, mark high-risk equipment according to the updated risk index, and obtain a high-risk equipment monitoring list.

[0025] In step S11, the operating data of the shared power bank cabinet is collected and preprocessed to obtain an operating data set, including: The shared power bank cabinets use built-in sensors to collect data on battery temperature, charging current, usage time, and ambient temperature. The collected data is formatted and outliers are removed to form a standardized set of operational data.

[0026] First, raw operational data is collected through the sensor network built into the shared power bank cabinet. A high-precision digital temperature sensor in each power bank slot collects battery temperature data, with a measurement range of -55℃ to +125℃ and an accuracy of ±0.5℃. The sampling frequency is set to 1Hz, based on the monitoring requirements of the initial temperature rise rate in battery thermal runaway, ensuring effective capture of temperature changes of safety significance. Simultaneously, an INA219 current sensor is used to collect charging current data, with a measurement range of ±3.2A and an accuracy of ±0.5%. The sampling frequency is set to 10Hz, to accurately capture instantaneous current fluctuations caused by poor contact or changes in battery internal resistance during charging. Usage time data is recorded in real-time by a timing module on the cabinet's main control board. This module starts accumulating charging time when the power bank is inserted into the slot and pauses timing when the user removes it. The accumulated usage time data is stored in non-volatile memory to ensure data integrity in the event of a power outage. The usage time data acquisition cycle is synchronized with the temperature data acquisition, set to 1Hz, and is packaged and sent uniformly within each data transmission cycle. Ambient temperature data is collected by a DHT22 temperature and humidity sensor deployed in the cabinet's ventilation area. Its temperature measurement accuracy is ±0.5℃, and the sampling frequency is set to 0.1Hz, a frequency chosen to account for the relatively slow changes in ambient temperature. All sensor data is transmitted to the edge computing node within the cabinet via the CAN bus protocol, with a bus baud rate set to 250kbps and a data transmission cycle of 5 seconds.

[0027] Subsequently, the collected data underwent format standardization and outlier removal. Format standardization included: resampling sensor data from different sampling frequencies to a 1Hz timestamp sequence using linear interpolation; and encapsulating all data fields into JSON format. Outlier detection was performed based on the 3σ principle, and detected outliers were removed. For blank data points following removed outliers, linear interpolation was used to fill in the gaps between the two valid data points before and after the removed outlier. After the above processing, a standardized set of operational data was finally formed and temporarily stored in a local SQLite database within the server rack.

[0028] In step S12, the peak temperature rise of the battery and the usage time data are extracted from the operating data set. Based on the peak temperature rise and usage time data, the device is grouped by state to generate grouping results, including: Extract the battery's peak temperature rise and usage duration data from the operational data set; Based on the peak temperature rise and usage duration data, K-means clustering is used to perform cluster analysis on the equipment. According to the clustering results, the equipment is divided into different state groups to obtain the grouping results.

[0029] First, the battery temperature rise peak and usage duration data are extracted from the operational data set. For example, the temperature rise peak is calculated as follows: for each power bank, the battery temperature data sequence and ambient temperature data sequence within the most recent fixed time window are taken, with the time window length set to ten minutes. This ten-minute window is validated because it is sufficient to cover the typical initial rapid temperature rise phase, and its duration is designed to filter out meaningless instantaneous fluctuations, thereby ensuring that the extracted temperature rise peak is statistically significant and capable of characterizing early anomalies. The instantaneous temperature rise value at each sampling moment within this time window is calculated using the following formula:

[0030] in Indicates the first The instantaneous temperature rise at each sampling moment, Indicates the first Battery temperature measurements at each sampling time. Indicates the first The ambient temperature was measured at each sampling time. Then, the maximum value was selected from the calculated sequence of all instantaneous temperature rise values ​​as the peak temperature rise of the battery within the current time window. The usage duration data is directly taken from the cumulative usage duration recorded in the runtime dataset. The data, expressed in hours, reflects the historical wear and tear of the battery.

[0031] Subsequently, based on the peak temperature rise and usage duration data, K-means clustering algorithm is used to perform cluster analysis on the device. For example, the cluster analysis includes the following steps: first, Z-score standardization is performed on the peak temperature rise and usage duration data; then, the core parameter K value for clustering is determined using the elbow rule; after determining the K value, K cluster centers are initialized using the K-means++ algorithm; iterative clustering is performed using Euclidean distance as the distance metric; the iteration termination condition is set when the moving distance of all cluster center points is less than a preset threshold. exemplarily Set as Or, it can reach the maximum number of iterations of 100.

[0032] Finally, based on the clustering results, the devices are divided into different state groups. For example, when K=3, the clustering algorithm naturally divides the devices into three clusters. Combining the business implications of the two features, peak temperature rise and usage time, the coordinates of the center point of each cluster are analyzed to assign a state label to each cluster: clusters with low peak temperature rise and short usage time are labeled as "normal state group," characterized by peak temperature rise typically below 5℃ and usage time less than 100 hours; clusters with medium peak temperature rise and medium usage time are labeled as "observation state group," characterized by peak temperature rise between 5℃ and 10℃ and usage time between 100 and 300 hours; clusters with high peak temperature rise and long usage time are labeled as "abnormal state group," characterized by peak temperature rise exceeding 10℃ and usage time exceeding 300 hours. Finally, a grouping result containing each device ID and its corresponding state group is generated.

[0033] In step S13, abnormal state devices are screened from the grouping results, the abnormal temperature rise range of the abnormal state devices is extracted, and the correlation characteristics between temperature changes and current changes within the abnormal temperature rise range are calculated. Potential abnormal devices are marked based on the correlation characteristics, and a preliminary abnormality list is generated, including: Filter devices in abnormal condition from the grouping results; Acquire the temperature rise change data of the abnormal state device, and identify the interval where the temperature rise change exceeds the preset change threshold and continues to exist, as the abnormal temperature rise interval; Extract the current change data of the abnormal temperature rise range from the charging current data, calculate the Pearson correlation coefficient between the temperature change data and the current change data in the abnormal temperature rise range, and obtain the correlation characteristics between temperature change and current change. Potentially abnormal devices are marked based on the abnormal temperature rise range and the associated characteristics, and a preliminary list of abnormalities is generated.

[0034] First, devices in abnormal states are filtered from the grouping results. This step directly reads the grouping results generated in step S12 and filters out all devices marked as "abnormal state groups" as the candidate device set for subsequent in-depth analysis.

[0035] Subsequently, the temperature rise change data of the abnormal state device is acquired, and the interval where the temperature rise change exceeds a preset change threshold and persists is identified as the abnormal temperature rise interval. The temperature rise change data is obtained by calculating the derivative of battery temperature with respect to time, using the central difference method. The calculation formula is as follows:

[0036] in Indicates the first The temperature rise rate at each sampling time is expressed in °C / min. and They represent and The battery temperature value at any given time; Indicates a time interval, for example, =1min. The preset change threshold is set to 0.5℃ / min, which is determined based on safety specifications for the maximum permissible temperature rise rate of lithium-ion batteries during normal charging. The duration threshold is set to 3 minutes, which is designed to filter out transient, harmless temperature fluctuations while capturing continuous, potentially dangerous temperature rise trends. The identification algorithm traverses the temperature rise rate sequence of each abnormal state device within the most recent 30-minute time window, and selects all devices that meet the criteria... The range of temperature rise >0.5℃ / min and continuous duration ≥3 minutes is marked as the abnormal temperature rise range. The threshold of 0.5℃ / min is determined based on the statistical analysis of the temperature rise rate of lithium-ion batteries under normal and abnormal operating conditions, which can effectively distinguish between normal charging heat generation and abnormal heat generation caused by potential faults.

[0037] Next, the current change data of the abnormal temperature rise range is extracted from the charging current data, and the Pearson correlation coefficient between the temperature change data and the current change data in the abnormal temperature rise range is calculated to obtain the correlation characteristics between temperature change and current change.

[0038] Finally, potential abnormal devices are marked based on the abnormal temperature rise range and the associated characteristics, and a preliminary list of abnormalities is generated. A threshold for the associated characteristics is set. >0.7. For any device in an abnormal state, as long as it has at least one abnormal temperature rise interval, and the correlation coefficient calculated within that interval is [value missing]. If the value is greater than 0.7, the device is marked as a potentially malfunctioning device. This threshold... The determination of 0.7 is based on statistical analysis of historical fault data, which can effectively distinguish between abnormal temperature rises caused by internal faults and those caused by external environmental factors or normal high-current charging. Finally, information on all marked potentially abnormal devices is compiled to generate a preliminary anomaly list. This list is a structured data table containing device ID, rack number, start and end times of the identified abnormal temperature rise interval, maximum temperature rise rate, and associated characteristics. The numerical values ​​and timestamps provide accurate input for subsequent steps.

[0039] In step S14, historical fluctuation sequences of devices in the preliminary anomaly list are obtained, and a first real-time fluctuation sequence is generated based on the currently collected temperature data. The differences between the first real-time fluctuation sequence and the historical fluctuation sequence are analyzed through time series comparison to generate a correlation strength score, including: The temperature rise fluctuation data of the equipment in the preliminary anomaly list within the historical normal cycle is obtained from the equipment historical database to obtain the historical fluctuation sequence. Based on the temperature rise change data, a first real-time fluctuation sequence is generated; The correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence is calculated using a dynamic time warping algorithm.

[0040] First, the historical fluctuation sequence and the first real-time fluctuation sequence of the devices in the preliminary anomaly list are obtained. For example, the historical fluctuation sequence is constructed as follows: From the device's historical operation database, all temperature rise data marked as "normal charging cycles" for the device in the past thirty days are retrieved. A "normal charging cycle" refers to a complete charging period where the charging process is stable, there are no alarm records, and the charging efficiency is between 85% and 115% of the rated range. For each normal charging cycle, the temperature rise data sequence from the start to the end of charging is extracted. Subsequently, a dynamic time warping algorithm is used to align the temperature rise sequences of all normal charging cycles, and their average sequence is calculated. This average sequence is used as the historical fluctuation sequence of the device, denoted as sequence. The method for obtaining the first real-time fluctuation sequence is as follows: For devices in the preliminary anomaly list, extract the temperature rise data that was most recently identified as being in the "abnormal temperature rise range". This data is a sequence of the difference between battery temperature and ambient temperature, denoted as sequence. Historical fluctuation sequence This represents the typical temperature rise behavior pattern of the device under healthy conditions, while the first real-time fluctuation sequence This indicates the current abnormal temperature rise state of the equipment.

[0041] Subsequently, a dynamic time warping algorithm is used to calculate the warped path distance between the two sequences. This algorithm constructs a... Cumulative cost matrix To find the optimal regular path. Matrix elements. The calculation is performed iteratively, and the recursive formula is:

[0042] in, Represents the i-th real-time sequence point With the j-th historical sequence point The local distance between them is calculated using Euclidean distance, i.e. . This indicates taking the minimum cumulative cost among the three adjacent values ​​to the left, below, and bottom left of the current position. The matrix initialization condition is... From the top left corner of the matrix. Start calculating until the bottom right corner. ,final The value is the cumulative normalized path distance between the two sequences.

[0043] After obtaining the cumulative regularized path distance Then, it is converted into a correlation strength score S. The conversion formula is:

[0044] in, This represents the association strength score, and its value ranges from (0,1]. Represents the natural exponential function; This is a scaling factor, exemplarily set to 10. This value was determined through grid search and cross-validation on a historical dataset containing multiple known failure modes, ultimately selecting the value that maximizes the correlation strength score. The value with the highest differentiation between devices in different health states; It is the minimum cumulative path distance calculated by the dynamic time warping algorithm; and These are the lengths of the real-time sequence and the historical sequence, respectively. This rating... This reflects the degree of deviation between the current real-time fluctuation sequence and the device's own historical health pattern. The closer the value is to 1, the more similar the current state is to the historical health pattern, and the stronger the correlation. The closer the value is to 0, the greater the deviation, the weaker the correlation, and the higher the risk of equipment malfunction. This correlation strength score provides a quantitative basis for subsequent comprehensive risk assessment.

[0045] In step S15, a comprehensive risk index is calculated based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence. High-risk devices are marked and listed as priority monitoring targets, generating a priority monitoring list, including: A comprehensive risk index is calculated based on the correlation strength score and the volatility amplitude of the first real-time volatility sequence. The devices in the preliminary anomaly list are sorted according to the comprehensive risk index; High-risk devices are marked and listed as priority monitoring targets, generating a priority monitoring list.

[0046] First, a comprehensive risk index is calculated based on the correlation strength score and the volatility amplitude of the first real-time volatility sequence. For example, the volatility amplitude of the first real-time volatility sequence is obtained by subtracting the maximum and minimum values ​​in the first real-time volatility sequence.

[0047] Overall Risk Index The calculation uses a weighted geometric mean. First, the fluctuation amplitude is... The value is then dimensionless, converting it to a value relative to a reference amplitude. The ratio. The reference amplitude. Based on historical statistical data, the temperature rise fluctuation amplitude of all devices marked as "normal state group" during their normal charging cycles was used as the 95th percentile of the distribution of these amplitudes. Based on the analysis of over 100,000 historical normal charging records, Set to 10℃. The dimensionless fluctuation amplitude is denoted as... Its calculation formula is .

[0048] Overall Risk Index The calculation formula is:

[0049] in This indicates the association strength score calculated in step S14, with a value range of (0,1]. This represents the dimensionless amplitude of the fluctuation. and These are weighted indices, exemplarily set as follows: =2, =1. This weighting scheme was determined through logistic regression analysis of historical fault data. The results show that... The contribution of the association strength score to the predicted equipment risk is approximately [missing information]. This represents twice the fluctuation amplitude. (In the formula...) The correlation strength score is converted into a deviation score, with a higher value indicating a greater deviation from a healthy state. The weighted geometric mean can simultaneously consider the influence of two indicators; when either indicator is significantly abnormal, it will lead to an increase in the overall risk index.

[0050] Subsequently, the devices in the preliminary anomaly list are sorted according to the comprehensive risk index. For example, all devices in the preliminary anomaly list are sorted according to their comprehensive risk index. The values ​​are sorted in descending order. Overall Risk Index The larger the value, the more the device's current state deviates from its historical health pattern, and the more drastic the temperature fluctuations, thus indicating a higher potential risk and a higher ranking in the sorting.

[0051] Finally, high-risk devices are marked and listed as priority monitoring targets, generating a priority monitoring list. For example, a high-risk threshold is set. =0.75. This threshold was determined based on statistical analysis of a large amount of equipment operating data: equipment that ultimately failed was selected from historical data, and the comprehensive risk index of the last monitored equipment before the failure was calculated. The 10th percentile of these indices was used as the threshold, thus ensuring that 90% of potential failures can be effectively identified. The comprehensive risk index was then sorted. Devices exceeding the threshold of 0.75 are marked as high-risk devices and listed as priority monitoring targets. Information on all marked high-risk devices is compiled to generate a priority monitoring list. This list includes the device ID, its rack location, comprehensive risk index, calculation timestamp, and relevant temperature rise anomaly data, providing a clear target list for subsequent high-frequency monitoring and in-depth analysis.

[0052] In step S16, for the devices in the priority monitoring list, the data acquisition frequency is increased to obtain a second real-time fluctuation sequence, the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence is calculated, and the dynamic health status is determined based on the similarity, including: For the devices in the priority monitoring list, increase the data collection frequency; Obtain the second real-time fluctuation sequence; Calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence; The dynamic health status is determined based on the similarity.

[0053] First, for devices on the priority monitoring list, the data acquisition frequency is increased. For example, for devices included in the priority monitoring list, the system increases their data acquisition frequency from the usual 1Hz to 10Hz to capture high-frequency temperature fluctuation characteristics related to specific fault modes that might be masked by low-frequency sampling. This high-frequency acquisition mode is initiated by sending a control command to the edge computing node of the corresponding rack, containing the target device identifier and the new sampling frequency parameters. The duration of high-frequency acquisition is set to 15 minutes, sufficient to capture multiple complete charge-discharge cycle waveforms, providing a robust data foundation for subsequent detailed analysis. Simultaneously with increasing the acquisition frequency, the system synchronously adjusts the data transmission cycle to 1 second to ensure that high-frequency data is uploaded to the monitoring center in real time.

[0054] Subsequently, a second real-time fluctuation sequence is acquired. For example, during high-frequency data acquisition, the system continuously records the battery temperature and ambient temperature of the target device and calculates the instantaneous temperature rise, forming a high-frequency temperature rise data stream. The second real-time fluctuation sequence is extracted from this high-frequency temperature rise data stream, with a length of 512 data points corresponding to a 51.2-second observation window. This length is determined based on the analysis of typical fault characteristic frequencies and can cover the main transient processes. Sequence preprocessing includes smoothing using a first-order low-pass digital filter with the following transfer function:

[0055] in Indicates the first The original temperature rise value of each sampling point This represents the filtered output value. This is the smoothing coefficient, exemplarily set to 0.2. This preprocessing effectively suppresses measurement noise while preserving the true characteristics of temperature changes.

[0056] Next, the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence is calculated. For example, a cosine similarity algorithm is used to calculate the similarity between the two sequence vectors. It should be noted that if the second real-time fluctuation sequence and the historical fluctuation sequence have different dimensions, linear interpolation is used to increase the dimension of the lower-dimensional sequence to be the same as the higher-dimensional sequence before calculating the similarity.

[0057] Finally, the dynamic health status is determined based on the similarity. For example, three health status level thresholds are set, with the thresholds (0.85, 0.6) determined based on the cluster center distance analysis of the three status clusters "Healthy," "Observation," and "Abnormal" in the device's historical operating data. When the similarity... When the device is in a "healthy" state, it is determined that the device is in a "healthy" state; when When the similarity is high, the device is determined to be in "observation" mode; when the similarity is low... When the system determines that an equipment is in an "abnormal" state, it uses cluster analysis of historical operating data to accurately reflect the degree of equipment degradation. The system compares the currently calculated similarity value with these thresholds to automatically determine the equipment's dynamic health status level. This status information, along with the equipment identifier and timestamp, is then updated to the equipment status database, providing real-time data for subsequent early warning decisions.

[0058] In step S17, when the dynamic health status shows a continuous abnormality, environmental correction is performed on the data triggering the continuous abnormality based on the ambient temperature data. An alarm signal is generated based on the corrected data and its reliability is verified. If the verification passes, a device control command is generated based on the alarm signal, including: When the dynamic health status shows a continuous abnormality, environmental correction is performed on the data that triggered the continuous abnormality based on the ambient temperature data. An alarm signal is generated based on the corrected data; The reliability of the alarm signal is verified; If the verification is successful, a device control command is generated based on the alarm signal.

[0059] First, when the dynamic health status shows a persistent abnormality, environmental correction is performed on the data triggering the persistent abnormality based on the ambient temperature data. For example, the criterion for determining the persistent abnormality is: the device's dynamic health status is determined to be abnormal for five consecutive monitoring cycles, with each monitoring cycle lasting 3 minutes. The purpose of environmental correction is to eliminate the influence of ambient temperature on battery temperature rise assessment. The correction formula uses a linear compensation model:

[0060] in This indicates the battery temperature value after environmental correction. This represents the actual measured battery temperature value, and k is the temperature compensation coefficient, which is set to 0.15 for example. This coefficient is obtained by testing the temperature rise of a typical battery module under different ambient temperatures in a controlled temperature chamber and by linear regression fitting such as the least squares method. This is based on the ambient temperature and is set to 25℃. This is the actual measured ambient temperature value. This correction converts battery temperatures measured under different ambient temperatures to equivalent values ​​under a standard ambient temperature, ensuring the accuracy of subsequent analyses.

[0061] Subsequently, an alarm signal is generated based on the corrected data. For example, the environmentally corrected temperature rise data is compared with preset multi-level alarm thresholds. Three alarm thresholds are set: the first level is a temperature rise exceeding 15°C for 2 minutes; the second level is a temperature rise exceeding 25°C for 1 minute; and the third level is a temperature rise exceeding 35°C for 30 seconds. When the corrected temperature rise data meets any of the alarm threshold conditions, a preliminary alarm signal of the corresponding level is generated. The alarm signal includes key information such as the device identifier, alarm level, trigger time, actual temperature rise value, and environmental correction amount.

[0062] Next, the alarm signal undergoes reliability verification. Exemplarily, reliability verification includes data integrity verification and data continuity verification. Data integrity verification requires that the integrity rate of the original data sequence on which the alarm signal is based is not less than 90%, where integrity rate equals the number of valid data points divided by the number of data points to be collected multiplied by 100%. Data continuity verification requires that the maximum consecutive missing period in the data sequence does not exceed 10 seconds. The alarm signal is deemed to have passed reliability verification if and only if the integrity rate is ≥90% and the maximum consecutive missing period is ≤10 seconds.

[0063] Finally, if the verification passes, a device control command is generated based on the alarm signal. For example, corresponding control commands are generated according to the alarm level: for a level 1 alarm, a command to "reduce charging current to 1A" is generated, where 1A is a verified safe current that can mitigate overheating risks and maintain partial charging function under most abnormal temperature rise conditions; for a level 2 alarm, a command to "pause charging and start fan cooling" is generated; for a level 3 alarm, a command to "immediately cut off power and send an emergency maintenance notification" is generated. The control commands are sent to the target device through the rack communication module, and information such as the command issuance time and execution status is recorded to form a complete command execution log.

[0064] In step S18, the equipment control command is executed, the operating data of the controlled equipment is collected and the comprehensive risk index is updated to obtain an updated risk index. High-risk equipment is then marked based on the updated risk index to obtain a high-risk equipment monitoring list, including: Execute the equipment control command, collect the operating data of the controlled equipment, and obtain updated operating data; Based on the updated operational data, the comprehensive risk index of the device is updated to obtain the updated risk index; Devices whose updated risk index exceeds a preset high-risk threshold are marked as high-risk devices, and a high-risk device monitoring list is generated.

[0065] First, the system executes the device control commands. For example, the system sends control commands to the target device via the cabinet control bus, with the command format conforming to the Modbus RTU protocol. For a command to reduce charging current, the system adjusts the charging current setting from the default 2A to 1A, using a ramp-down method to uniformly reduce the current to the target value within 10 seconds, avoiding sudden current changes. For a command to pause charging, the system first disconnects the charging relay, delays for 0.5 seconds, and then starts the cooling fan, setting its operating power to 80% of its rated power. For an immediate power-off command, the system simultaneously disconnects the charging relay and the output relay, and triggers an audible and visual alarm. The execution status of all commands is fed back in real time through the device status register, which the system reads every 0.5 seconds to ensure correct command execution.

[0066] Subsequently, the system collects operational data from the controlled equipment and updates the comprehensive risk index to obtain an updated risk index. For example, during a 15-minute monitoring period after command execution, the system collects operational parameters of the equipment, such as battery temperature, charging current, and output voltage, at a frequency of 1Hz. Based on this data, the correlation strength score is recalculated. and dimensionless fluctuation amplitude The correlation strength score uses the same dynamic time warping algorithm as step S14, but the historical baseline sequence is updated to a sliding window sequence containing the most recent 24 hours of normal operation data; dimensionless fluctuation amplitude. The calculation method is the same as in step S15, updating the risk index. The calculation formula is:

[0067] in It is the comprehensive risk index before regulation. This is the forgetting factor, exemplarily set to 0.3, meaning that the current state accounts for 70% of the weight in the updated risk index. This ensures that the risk assessment can respond quickly to recent changes in equipment status. This value is determined through validation using historical data, balancing the impact of historical risks with the current state. and Maintaining the same weight settings as in step S15, namely 2 and 1 respectively, this formula achieves dynamic updates to the risk index, taking into account both the historical risk status of the equipment and incorporating the latest status information after adjustments.

[0068] Next, high-risk devices are marked according to the updated risk index. For example, a high-risk device marking threshold is set. The high-risk threshold of 0.85 corresponds to a verified high failure probability critical point. Equipment exceeding this threshold requires immediate intervention. This threshold is determined based on statistical analysis of equipment failure probabilities: when the update risk index exceeds 0.85, the probability of equipment failure within the next 72 hours exceeds 80%. The marking logic is: for all equipment executing control commands, if its update risk index... If the equipment is deemed high-risk, it will be marked as such. The marking information includes the equipment ID, marking time, updated risk index value, type of control instruction executed, and key operating parameters after control.

[0069] Finally, a high-risk equipment monitoring list is obtained. For example, the system aggregates information on all marked high-risk equipment to generate a monitoring list, which is stored in a structured database table and includes the following fields: unique equipment identifier, rack number, timestamp, update risk index, current health status level, implemented control measures, and most recent maintenance record. This list is sorted in descending order of update risk index and is set up with an automatic refresh mechanism, updating every 30 minutes. For the top 10% of equipment in the list, the system automatically generates maintenance work orders and assigns them to the on-site maintenance team, while simultaneously increasing their data collection frequency to 5Hz for enhanced monitoring.

[0070] Reference Figure 2 The first embodiment of the present invention provides a flowchart of the core algorithm for implementing the method of the present invention. This schematic diagram fully presents the key calculation process from time series correlation analysis to risk level determination.

[0071] The process first obtains the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence based on the devices in the preliminary anomaly list. These two types of data reflect the degree of matching between the device's current temperature rise behavior and its historical health pattern, as well as the intensity of temperature rise fluctuations. Subsequently, the process enters the comprehensive risk index calculation stage. By weighting and fusing the correlation strength score and fluctuation amplitude, a comprehensive risk index that quantifies the real-time risk status of the device is generated. This step is the key difference between this invention and traditional threshold judgment methods in terms of refined risk classification.

[0072] After obtaining the comprehensive risk index, the process moves to the risk equipment identification stage. By comparing the comprehensive risk index with a preset risk threshold, it is possible to accurately determine whether the equipment has a high level of safety hazard, and accordingly mark it as high-risk equipment and include it in the priority monitoring list. The entire process constitutes a technical chain of "time series assessment → index fusion calculation → risk level determination," providing core algorithmic support for the present invention to achieve accurate risk identification and hierarchical monitoring.

[0073] Reference Figure 3 The second embodiment of the present invention provides a remote security monitoring system for shared power bank cabinets, including: The data acquisition module is used to collect and preprocess the operating data of the shared power bank cabinets to obtain an operating data set. The status grouping module is used to extract the battery's temperature rise peak and the usage time data from the operating data set, and to group the device status based on the temperature rise peak and usage time data to generate grouping results; An abnormal feature module is used to filter devices in abnormal state from the grouping results, extract the abnormal temperature rise range of the abnormal devices, calculate the correlation features between temperature change and current change within the abnormal temperature rise range, and mark potential abnormal devices according to the correlation features to generate a preliminary abnormal list. The correlation assessment module is used to obtain the historical fluctuation sequence and the first real-time fluctuation sequence of the temperature rise data of the equipment in the preliminary anomaly list, and to calculate the correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence. The risk ranking module is used to calculate a comprehensive risk index based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, mark high-risk devices as priority monitoring targets, and generate a priority monitoring list. The health analysis module is used to increase the data collection frequency for devices in the priority monitoring list, obtain a second real-time fluctuation sequence, calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determine the dynamic health status based on the similarity. The alarm control module is used to perform environmental correction on the data that triggers the continuous abnormality based on the ambient temperature data when the dynamic health status shows a continuous abnormality, generate an alarm signal based on the corrected data and perform reliability verification, and generate equipment control instructions based on the alarm signal if the verification is successful. The closed-loop management module is used to execute the equipment control commands, collect the operating data of the controlled equipment and update the comprehensive risk index to obtain an updated risk index, mark high-risk equipment according to the updated risk index, and obtain a high-risk equipment monitoring list.

[0074] It should be noted that the remote security monitoring device for shared power bank cabinets provided in this embodiment of the invention is used to execute all the process steps of the remote security monitoring method for shared power bank cabinets in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0075] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps described in the method embodiments above.

[0076] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0077] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0078] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0079] The memory can be used to store the computer programs and modules. The processor implements various functions of the electronic device by running or executing the computer programs and modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0080] Wherein, if the modules / units integrated in the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0081] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and 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. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0082] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A remote security monitoring method for shared power bank cabinets, characterized in that, include: Collect and preprocess the operational data of the shared power bank cabinets to obtain an operational data set; The operational data mentioned above includes battery temperature data, charging current data, usage time data, and ambient temperature data. Extract the battery's peak temperature rise and usage time data from the operational data set, group the devices by status based on the peak temperature rise and usage time data, and generate grouping results; The peak temperature rise is calculated based on the battery temperature data. From the grouping results, devices in abnormal condition are screened, the abnormal temperature rise range of the devices in abnormal condition is extracted, and the correlation characteristics between temperature change and current change within the abnormal temperature rise range are calculated. Potential abnormal devices are marked according to the correlation characteristics to generate a preliminary abnormal list. Obtain the historical fluctuation sequence and the first real-time fluctuation sequence of the temperature rise data of the equipment in the preliminary anomaly list, and calculate the correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence; Based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, a comprehensive risk index is calculated, high-risk devices are marked and listed as priority monitoring targets, and a priority monitoring list is generated. For the devices in the priority monitoring list, increase the data collection frequency, obtain the second real-time fluctuation sequence, calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determine the dynamic health status based on the similarity. When the dynamic health status shows a continuous abnormality, environmental correction is performed on the data that triggered the continuous abnormality based on the ambient temperature data. An alarm signal is generated based on the corrected data and a reliability verification is performed. If the verification is successful, a device control command is generated based on the alarm signal. Execute the equipment control command, collect the operating data of the controlled equipment and update the comprehensive risk index to obtain an updated risk index, mark high-risk equipment according to the updated risk index, and obtain a high-risk equipment monitoring list.

2. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, The process of collecting and preprocessing the operational data of the shared power bank cabinets yields an operational data set, including: The shared power bank cabinets use built-in sensors to collect data on battery temperature, charging current, usage time, and ambient temperature. The collected data is formatted and outliers are removed to form a standardized set of operational data.

3. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, The step involves extracting the battery's peak temperature rise and usage duration data from the operational data set, grouping the device into status groups based on the peak temperature rise and usage duration data, and generating grouping results, including: Extract the battery's peak temperature rise and usage duration data from the operational data set; Based on the peak temperature rise and the usage time data, K-means clustering is used to perform cluster analysis on the equipment. According to the clustering results, the equipment is divided into different state groups to obtain the grouping results.

4. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, The process involves filtering out devices in abnormal condition from the grouping results, extracting the abnormal temperature rise range of the devices in abnormal condition, calculating the correlation characteristics between temperature changes and current changes within the abnormal temperature rise range, marking potential abnormal devices based on the correlation characteristics, and generating a preliminary abnormality list, including: Filter devices in abnormal condition from the grouping results; Acquire the temperature rise change data of the abnormal state device, and identify the interval where the temperature rise change exceeds the preset change threshold and continues to exist, as the abnormal temperature rise interval; Extract the current change data of the abnormal temperature rise range from the charging current data, calculate the Pearson correlation coefficient between the temperature change data and the current change data in the abnormal temperature rise range, and obtain the correlation characteristics between temperature change and current change. Potentially abnormal devices are marked based on the abnormal temperature rise range and the associated characteristics, and a preliminary list of abnormalities is generated.

5. The remote security monitoring method for shared power bank cabinets according to claim 4, characterized in that, The process involves obtaining historical fluctuation sequences of devices in the preliminary anomaly list, generating a first real-time fluctuation sequence based on currently collected temperature data, and analyzing the differences between the first real-time fluctuation sequence and the historical fluctuation sequences through time series comparison to generate a correlation strength score, including: The temperature rise fluctuation data of the equipment in the preliminary anomaly list within the historical normal cycle is obtained from the equipment historical database to obtain the historical fluctuation sequence. Based on the temperature rise change data, a first real-time fluctuation sequence is generated; The correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence is calculated using a dynamic time warping algorithm.

6. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, The comprehensive risk index is calculated based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence. High-risk devices are marked and listed as priority monitoring targets, generating a priority monitoring list, including: The correlation strength score is weighted with the fluctuation amplitude of the first real-time fluctuation sequence to generate a comprehensive risk index; Based on the comprehensive risk index, the devices in the preliminary anomaly list are sorted in descending order to obtain the risk ranking result; Based on the risk ranking results, devices whose comprehensive risk index exceeds a preset risk threshold are marked as high-risk devices and listed as priority monitoring targets, generating a priority monitoring list.

7. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, The step of calculating the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determining the dynamic health status based on the similarity, includes: Based on the second real-time fluctuation sequence and the historical fluctuation sequence, the similarity value between the two is calculated using cosine similarity. Based on the preset range to which the similarity value belongs, the deviation level of the device's health status is determined; A corresponding dynamic health status is generated based on the deviation level.

8. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, When the dynamic health status shows a continuous abnormality, environmental correction is performed on the data triggering the continuous abnormality based on the ambient temperature data. An alarm signal is generated based on the corrected data and its reliability is verified. If the verification passes, a device control command is generated based on the alarm signal, including: Based on the ambient temperature data, the temperature rise data that triggers the continuous abnormality is compensated and corrected, and the corrected data is compared with the preset alarm threshold to generate an alarm signal. The data sequence corresponding to the alarm signal is verified for data integrity and data continuity. When the verification result meets the preset requirements, the alarm signal is determined to be reliable, and a device control command is generated based on the alarm signal.

9. The remote security monitoring method for shared power bank cabinets according to claim 1, characterized in that, The process involves executing the equipment control commands, collecting operational data from the controlled equipment, updating the comprehensive risk index to obtain an updated risk index, marking high-risk equipment based on the updated risk index, and obtaining a high-risk equipment monitoring list, including: Execute the equipment control command, collect the operating data of the controlled equipment, and obtain updated operating data; Based on the updated operational data, the comprehensive risk index of the device is updated to obtain the updated risk index; Devices whose updated risk index exceeds a preset high-risk threshold are marked as high-risk devices, and a high-risk device monitoring list is generated.

10. A remote security monitoring system for shared power bank cabinets, characterized in that, include: The data acquisition module is used to collect and preprocess the operating data of the shared power bank cabinets to obtain an operating data set. The status grouping module is used to extract the battery's temperature rise peak and the usage time data from the operating data set, and to group the device status based on the temperature rise peak and usage time data to generate grouping results; An abnormal feature module is used to filter devices in abnormal state from the grouping results, extract the abnormal temperature rise range of the abnormal devices, calculate the correlation features between temperature change and current change within the abnormal temperature rise range, and mark potential abnormal devices according to the correlation features to generate a preliminary abnormal list. The correlation assessment module is used to obtain the historical fluctuation sequence and the first real-time fluctuation sequence of the temperature rise data of the equipment in the preliminary anomaly list, and to calculate the correlation strength score between the first real-time fluctuation sequence and the historical fluctuation sequence. The risk ranking module is used to calculate a comprehensive risk index based on the correlation strength score and the fluctuation amplitude of the first real-time fluctuation sequence, mark high-risk devices as priority monitoring targets, and generate a priority monitoring list. The health analysis module is used to increase the data collection frequency for devices in the priority monitoring list, obtain a second real-time fluctuation sequence, calculate the similarity between the second real-time fluctuation sequence and the historical fluctuation sequence, and determine the dynamic health status based on the similarity. The alarm control module is used to perform environmental correction on the data that triggers the continuous abnormality based on the ambient temperature data when the dynamic health status shows a continuous abnormality, generate an alarm signal based on the corrected data and perform reliability verification, and generate equipment control instructions based on the alarm signal if the verification is successful. The closed-loop management module is used to execute the equipment control commands, collect the operating data of the controlled equipment and update the comprehensive risk index to obtain an updated risk index, mark high-risk equipment according to the updated risk index, and obtain a high-risk equipment monitoring list.

Images