An automatic monitoring and data analysis system and method for a power distribution room based on an internet of things

Abstract

The application discloses a power distribution room automatic monitoring and data analysis system and method based on the Internet of Things, collects multi-source parameters of electrical operation and environmental characteristics of a power distribution room, and pushes the parameters to a digital twin virtual model of the power distribution room to realize virtual mapping of a physical state after timestamp alignment and heterogeneous data standardization processing. Based on historical health data of a time series database, an early warning threshold that is adaptive to equipment operation length and is associated with monitoring points is determined through a sliding time window and an exponential weighting algorithm, parameters are monitored in real time, a multi-parameter continuous deviation trend is identified, and a comprehensive deviation degree of a power distribution room operation state relative to a historical health benchmark is calculated. When the comprehensive deviation degree exceeds the threshold or a single parameter exceeds a limit in a continuous sampling period, a fault early warning signal containing monitoring points and an abnormal type is generated and is synchronized to the digital twin model and a specified early warning receiving end. The application realizes early fault predictive early warning of the power distribution room, improves real-time monitoring and intelligent level, and guarantees safe and stable operation of the power distribution room.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring technology for power distribution rooms, specifically to an automatic monitoring and data analysis system and method for power distribution rooms based on the Internet of Things. Background Technology

[0002] As a key power distribution node in the power system, the operation stability of the power distribution room is directly related to the safety and reliability of power supply. The power distribution room contains a large number of power distribution equipment such as transformers, switch cabinets, and busbars. At the same time, its operating environment has a significant impact on the lifespan of the equipment and safe operation. It is necessary to conduct comprehensive and real-time monitoring of electrical operating parameters and environmental characteristic parameters to prevent power supply accidents caused by equipment failure, insulation aging, environmental anomalies, etc.

[0003] Currently, power distribution room monitoring mainly employs a combination of fixed threshold alarms and regular manual inspections. On one hand, fixed threshold alarms use static thresholds, which cannot adapt to performance degradation and parameter baseline drift caused by long-term operation of power distribution equipment. They struggle to distinguish between normal aging fluctuations and early-stage faults, leading to false alarms and missed alarms. On the other hand, manual inspections suffer from long cycles and poor real-time performance, failing to capture dynamic parameter trends and lacking effective analytical methods for complex anomalies caused by multi-parameter coupling. Furthermore, existing monitoring systems mostly monitor single parameters, lacking multi-source data fusion analysis, resulting in insufficient assessment of the overall operating status of the power distribution room and difficulty in achieving accurate early fault warnings. Therefore, how to achieve intelligent acquisition, dynamic early warning, and comprehensive analysis of multi-dimensional parameters in power distribution rooms based on IoT technology, and capture early, subtle anomaly trends, has become a pressing problem for the industry. Summary of the Invention

[0004] Based on this, the present invention provides an automatic monitoring and data analysis system and method for power distribution rooms based on the Internet of Things, which realizes intelligent sensing of multi-source parameters, dynamic threshold early warning and multi-parameter fusion analysis of power distribution rooms, accurately captures early abnormal trends and improves the intelligence and real-time level of power distribution room monitoring.

[0005] In a first aspect, the present invention provides an automatic monitoring and data analysis method for power distribution rooms based on the Internet of Things, comprising the following steps:

[0006] The electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room are collected to form a multi-source monitoring parameter set;

[0007] The multi-source monitoring parameter set is timestamped and heterogeneous data is standardized. The processed parameter data is then pushed to the digital twin virtual model corresponding to the power distribution room, so as to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space.

[0008] Based on the historical health operation data of the power distribution room stored in the time series database associated with the digital twin virtual model, a sliding time window mechanism combined with an exponential weighting algorithm is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter, and then determine the upper and lower thresholds of the early warning for each monitoring parameter that are adaptive with the cumulative running time of the equipment and associated with the monitoring point identifier for traceability.

[0009] Real-time monitoring of the parameter values ​​of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, calculate the comprehensive deviation of the current power distribution room operation status from the historical health benchmark.

[0010] When the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper or lower warning threshold within a continuous sampling period, a power distribution room fault warning signal containing the monitoring point identifier and the anomaly type is generated and synchronized to the digital twin virtual model and the designated warning receiver.

[0011] As a further improvement to the technical solution of the present invention, the timestamp alignment and heterogeneous data standardization processing of the multi-source monitoring parameter set specifically includes:

[0012] Obtain structured data packets with local timestamps and monitoring point identifiers corresponding to all monitoring parameters. The structured data packets contain the original values ​​of electrical operating parameters, the original values ​​of environmental characteristic parameters, and the data acquisition device number.

[0013] Based on the IEEE 1588 precision clock synchronization protocol, the local timestamps of all structured data packets are uniformly calibrated to obtain multi-source parameter data after timestamp alignment. According to the preset parameter per-unit formula, the heterogeneous parameter data after timestamp alignment is normalized to eliminate dimensional differences and obtain a standardized parameter dataset.

[0014] As a further improvement to the technical solution of this invention, the processed parameter data is pushed to the digital twin virtual model corresponding to the power distribution room, realizing the real-time mapping and visualization of the physical operating status of the power distribution room in the virtual space, specifically including:

[0015] The standardized parameter dataset is sent to the data access layer of the digital twin virtual model via the MQTT industrial communication protocol, and the parameter data is simultaneously written to the time series database for persistent storage. Based on the monitoring point identifier and data type of the parameter data, the status attributes of the corresponding virtual monitoring nodes, power distribution equipment and power distribution room space in the digital twin virtual model are updated.

[0016] The visualization engine that drives the digital twin virtual model renders and displays a dynamic parameter dashboard, equipment operation status color mark, and parameter change trend curve of the 3D virtual scene of the power distribution room based on real-time changes in status attributes.

[0017] As a further improvement to the technical solution of this invention, based on the historical health operation data of the power distribution room stored in the time-series database associated with the digital twin virtual model, a sliding time window mechanism combined with an exponential weighting algorithm is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter, specifically including:

[0018] Extract historical healthy operation data that match the fault-free operation period of the power distribution room within a preset time range from the time series database, and remove abnormal outliers to obtain a clean historical dataset.

[0019] For each monitoring parameter, the window length and sliding step size are set according to the operating cycle of the power distribution room equipment, and the sliding time window data sequence based on the time dimension is extracted from the clean historical dataset;

[0020] The exponentially weighted moving average algorithm is used to calculate the exponentially weighted moving average of the data series for each sliding time window as the dynamic statistical benchmark value, and the exponentially weighted moving standard deviation is used to calculate the volatility coefficient.

[0021] As a further improvement to the technical solution of this invention, the upper and lower thresholds for early warning, which are adaptive to the cumulative operating time of the equipment and associated with the monitoring point identifier for traceability, specifically include:

[0022] For each monitoring parameter, based on the preset confidence coefficient, the product of the confidence coefficient and the fluctuation coefficient is added to and subtracted from the dynamic statistical benchmark value to obtain the upper and lower warning limits of the parameter.

[0023] Obtain the cumulative running time of the power distribution equipment corresponding to the current monitoring point and the unique identifier of the monitoring point. Associate and store the upper and lower warning limits with the cumulative running time of the equipment and the identifier of the monitoring point to form an adaptive warning threshold library with traceability information.

[0024] As a further improvement to the technical solution of this invention, the real-time monitoring of the parameter values ​​of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds specifically include:

[0025] The real-time parameter values ​​of each monitoring node in the power distribution room are collected and processed in real time, and the timestamp alignment and standardization are completed.

[0026] Using the monitoring point identifier and the current cumulative running time of the equipment as query conditions, the latest upper and lower thresholds for each parameter's real-time value are matched and obtained from the adaptive early warning threshold library.

[0027] As a further improvement to the technical solution of this invention, when multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, the calculation of the comprehensive deviation of the current power distribution room operating status relative to the historical health benchmark specifically includes:

[0028] Set a warning zone ratio for each monitoring parameter, calculate the warning zone boundary outside the upper and lower warning thresholds, and determine whether the real-time value of the parameter enters the warning zone;

[0029] For monitoring parameters entering the warning zone, extract the time series data within the most recent preset sampling period, and use linear regression to fit the trend slope. If the absolute value of the slope exceeds the preset trend threshold and the direction is far away from the normal range, it is determined to show a continuous deviation trend.

[0030] When two or more monitoring parameters show a continuous deviation trend, the historical health mean and covariance matrix of such parameters are extracted, the real-time values ​​of the parameters are used to form the current observation vector, and the Mahalanobis distance of this vector relative to the historical health benchmark is calculated as the comprehensive deviation.

[0031] As a further improvement to the technical solution of this invention, generating a power distribution room fault early warning signal that includes monitoring point identifiers and anomaly types, and synchronizing it to the digital twin virtual model and the designated early warning receiving end specifically includes:

[0032] When the overall deviation exceeds the preset deviation threshold or the continuous sampling period of a single parameter exceeds the warning threshold, the abnormal trigger type is determined, and information such as the trigger timestamp, monitoring point identifier, abnormal parameter value, and warning threshold is collected.

[0033] The above information is encapsulated according to the JSON data architecture to form a structured fault warning signal, which includes the anomaly type, source information and handling prompts;

[0034] The fault warning signal is pushed to the designated warning receiving end such as the operation and maintenance terminal and the power distribution room monitoring screen through the enterprise message bus. At the same time, the signal is sent through the internal interface of the digital twin virtual model to drive the alarm visualization display of the corresponding point in the virtual scene.

[0035] As a further improvement to the technical solution of the present invention, the electrical operating parameters include the effective value of three-phase voltage, the effective value of three-phase current, power factor, switch opening and closing status, and bus temperature, and the environmental characteristic parameters include the ambient temperature of the power distribution room, relative humidity, dust concentration, and SF6 gas concentration.

[0036] Secondly, the present invention provides an automatic monitoring and data analysis system for power distribution rooms based on the Internet of Things, comprising:

[0037] The Internet of Things (IoT) data acquisition module is used to collect electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room, forming a multi-source monitoring parameter set;

[0038] The data processing module is used to perform timestamp alignment and heterogeneous data standardization on the multi-source monitoring parameter set, and push the processed parameter data to the digital twin virtual model corresponding to the power distribution room to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space.

[0039] The data processing module is also used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter by using a sliding time window mechanism combined with an exponential weighting algorithm to determine the historical health operation data of the power distribution room stored in the time series database associated with the digital twin virtual model. In this way, the upper and lower thresholds of the early warning for each monitoring parameter are determined to be adaptive with the cumulative running time of the equipment and associated with the monitoring point identifier for traceability.

[0040] The data processing module is also used to monitor the real-time values ​​of parameters of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, the module calculates the comprehensive deviation of the current power distribution room operating status from the historical health benchmark.

[0041] The alarm execution module is used to generate a power distribution room fault warning signal containing the monitoring point identifier and the abnormality type when the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper and lower warning thresholds within a continuous sampling period, and synchronize it to the digital twin virtual model and the designated warning receiver.

[0042] Compared with the prior art, the present invention has the following advantages:

[0043] This invention establishes a comprehensive, unified, and intuitive data and visualization foundation for power distribution room monitoring by performing timestamp alignment and heterogeneous data standardization on multi-source electrical and environmental monitoring parameters, and combining this with a digital twin virtual model to achieve real-time virtual mapping of physical operating status. Simultaneously, based on a sliding time window and exponential weighting algorithm, and combined with the cumulative operating time of equipment, it constructs an adaptive early warning threshold for correlation and traceability with monitoring points, overcoming the technical limitations of traditional fixed thresholds that cannot adapt to equipment aging, and improving the tolerance for fluctuations in normal operation and the accuracy of early anomaly identification. Furthermore, it captures the continuous deviation trend of multi-parameter collaboration and calculates the comprehensive deviation... The deviation measurement system enables quantitative assessment of the overall operating status of the power distribution room, effectively identifying complex anomalies and early weak fault trends that are difficult to detect with single-parameter monitoring. Combined with dual early warning criteria of continuous single-parameter exceedance and comprehensive deviation exceeding the threshold, it generates fault early warning signals with monitoring point and anomaly type tracing information and realizes virtual and real linkage alarms. This not only upgrades the power distribution room from manual inspection and static alarm to automatic monitoring and predictive early warning, significantly improving the real-time performance and intelligence of monitoring, but also allows maintenance personnel to accurately locate anomaly points, quickly determine anomaly types, improve fault handling efficiency, and comprehensively ensure the safe and stable operation of the power distribution room. Attached Figure Description

[0044] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0045] Figure 1 This is an exemplary flowchart of an IoT-based automatic monitoring and data analysis method for power distribution rooms, as shown in some embodiments of the present invention.

[0046] Figure 2 This is a schematic diagram illustrating an application scenario of an automatic monitoring and data analysis system for a power distribution room, as shown in some embodiments of the present invention.

[0047] Figure 3 This is a schematic diagram of the structure of an Internet of Things-based automatic monitoring and data analysis system for power distribution rooms, as shown in some embodiments of the present invention.

[0048] Figure 4 This is a schematic diagram of the structure of a computer device for implementing an Internet of Things-based automatic monitoring and data analysis method for power distribution rooms, as shown in some embodiments of the present invention. Detailed Implementation

[0049] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0050] The present invention will be further described in detail below with reference to the accompanying drawings.

[0051] Reference Figure 1 In a first aspect, the present invention provides an automatic monitoring and data analysis method for power distribution rooms based on the Internet of Things, comprising the following steps:

[0052] The electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room are collected to form a multi-source monitoring parameter set;

[0053] The multi-source monitoring parameter set is timestamped and heterogeneous data is standardized. The processed parameter data is then pushed to the digital twin virtual model corresponding to the power distribution room, so as to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space.

[0054] Based on the historical health operation data of the power distribution room stored in the time series database associated with the digital twin virtual model, a sliding time window mechanism combined with an exponential weighting algorithm is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter, and then determine the upper and lower thresholds of the early warning for each monitoring parameter that are adaptive with the cumulative running time of the equipment and associated with the monitoring point identifier for traceability.

[0055] Real-time monitoring of the parameter values ​​of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, calculate the comprehensive deviation of the current power distribution room operation status from the historical health benchmark.

[0056] When the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper or lower warning threshold within a continuous sampling period, a power distribution room fault warning signal containing the monitoring point identifier and the anomaly type is generated and synchronized to the digital twin virtual model and the designated warning receiver.

[0057] This invention collects comprehensive electrical operation and environmental monitoring parameters of a power distribution room using IoT sensing technology. First, it performs unified processing, including timestamp alignment and standardization, on multi-source heterogeneous parameters. The processed data is then pushed to a digital twin virtual model of the power distribution room to achieve a virtual real-time mapping of the physical state. Based on historical health data from a time-series database associated with the digital twin, and combined with a sliding time window mechanism and an exponential weighting algorithm, an adaptive early warning threshold is constructed that dynamically adjusts with the cumulative operating time of the equipment and is associated with the monitoring point. Real-time parameter values ​​are matched with the corresponding early warning thresholds to identify continuous deviation trends in multi-parameter collaboration and to calculate the comprehensive deviation of the power distribution room's operating state relative to the historical health benchmark. Dual early warning criteria are set: exceeding the comprehensive deviation threshold and exceeding the continuous sampling period limit for a single parameter. When either criterion is met, a fault early warning signal containing the monitoring point and anomaly type is generated and simultaneously pushed to the digital twin virtual model and a designated early warning receiver, achieving fault early warning through virtual-real linkage.

[0058] This invention enables the comprehensive acquisition and standardized processing of multi-source monitoring parameters in power distribution rooms, establishing a unified spatiotemporal and virtual-real interconnected monitoring data foundation. It overcomes the technical limitations of traditional fixed thresholds that cannot adapt to aging equipment, achieving dynamic adaptive adjustment and precise source tracing of warning thresholds. Through multi-parameter collaborative deviation trend identification and comprehensive deviation calculation, it can capture complex anomalies and early weak fault trends that are difficult to detect with single-parameter monitoring. Combined with dual warning criteria, it achieves predictive warnings of power distribution room operating status, replacing traditional manual inspections and static alarm modes. This significantly improves the real-time performance and intelligence level of power distribution room monitoring. Simultaneously, through warning signals with source tracing information and virtual-real linked alarms, maintenance personnel can accurately locate abnormal points and quickly determine the type of abnormality, improving fault handling efficiency and comprehensively ensuring the safe and stable operation of the power distribution room.

[0059] In some embodiments, performing timestamp alignment and heterogeneous data standardization on the multi-source monitoring parameter set specifically includes:

[0060] Obtain structured data packets with local timestamps and monitoring point identifiers corresponding to all monitoring parameters. The structured data packets contain the original values ​​of electrical operating parameters, the original values ​​of environmental characteristic parameters, and the data acquisition device number.

[0061] Based on the IEEE 1588 precision clock synchronization protocol, the local timestamps of all structured data packets are uniformly calibrated to obtain multi-source parameter data after timestamp alignment. According to the preset parameter per-unit formula, the heterogeneous parameter data after timestamp alignment is normalized to eliminate dimensional differences and obtain a standardized parameter dataset.

[0062] In practice, the system first acquires a structured data packet containing a local timestamp, monitoring point identifier, original values ​​of electrical and environmental parameters, and data acquisition device number. Based on the IEEE 1588 precision clock synchronization protocol, the local timestamps of all data packets are uniformly calibrated to eliminate time differences between multi-source data. Then, according to the preset parameter normalization formula, the heterogeneous parameter data after timestamp alignment is normalized to eliminate dimensional differences between different parameters, resulting in a standardized multi-source parameter dataset.

[0063] This invention achieves high-precision time synchronization of multi-source heterogeneous monitoring parameters, eliminating time differences between data from different acquisition devices and monitoring points, and ensuring spatiotemporal consistency of data. Through per-unit normalization processing, it eliminates dimensional differences between electrical and environmental parameters, realizing the standardization and unification of heterogeneous data. This provides standardized and comparable basic data for subsequent data analysis, threshold calculation, and status assessment, avoiding analytical errors caused by differences in data format and dimensions, and improving the accuracy and reliability of subsequent data processing.

[0064] In some embodiments, the processed parameter data is pushed to the digital twin virtual model corresponding to the power distribution room to realize the real-time mapping and visualization of the physical operating status of the power distribution room in the virtual space. Specifically, this includes:

[0065] The standardized parameter dataset is sent to the data access layer of the digital twin virtual model via the MQTT industrial communication protocol, and the parameter data is simultaneously written to the time series database for persistent storage. Based on the monitoring point identifier and data type of the parameter data, the status attributes of the corresponding virtual monitoring nodes, power distribution equipment and power distribution room space in the digital twin virtual model are updated.

[0066] The visualization engine that drives the digital twin virtual model renders and displays a dynamic parameter dashboard, equipment operation status color mark, and parameter change trend curve of the 3D virtual scene of the power distribution room based on real-time changes in status attributes.

[0067] It should be noted that the standardized parameter dataset is sent to the data access layer of the digital twin virtual model via the MQTT industrial communication protocol, and the parameter data is simultaneously written to the time-series database for persistent storage. Based on the monitoring point identifiers and data types of the parameter data, the status attributes of the corresponding virtual monitoring nodes, power distribution equipment, and power distribution room space in the digital twin virtual model are accurately updated. The visualization engine of the digital twin virtual model is driven to render and display the dynamic parameter dashboard, equipment operation status color marks, and parameter change trend curves in the 3D virtual scene according to the real-time changes in status attributes, thereby realizing the visual mapping of the physical operation status.

[0068] This invention employs the lightweight MQTT protocol to achieve efficient transmission of parameter data to the digital twin model, while simultaneously achieving persistent data storage, thus balancing the real-time nature of data transmission with the integrity of storage. Through precise correlation and updates between parameter data and the state attributes of virtual entities, real-time data synchronization between the physical power distribution room and the virtual model is realized, constructing a high-fidelity digital mirror of the power distribution room. With the help of multi-dimensional visualization, maintenance personnel can intuitively and comprehensively grasp the real-time operating status of each point and device in the power distribution room, replacing traditional numerical list-based monitoring and improving the intuitiveness and convenience of monitoring.

[0069] In some embodiments, the historical health operation data of the power distribution room stored in the time-series database associated with the digital twin virtual model are used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter by employing a sliding time window mechanism combined with an exponential weighting algorithm. Specifically, this includes:

[0070] Extract historical healthy operation data that match the fault-free operation period of the power distribution room within a preset time range from the time series database, and remove abnormal outliers to obtain a clean historical dataset.

[0071] For each monitoring parameter, the window length and sliding step size are set according to the operating cycle of the power distribution room equipment, and the sliding time window data sequence based on the time dimension is extracted from the clean historical dataset;

[0072] The exponentially weighted moving average algorithm is used to calculate the exponentially weighted moving average of the data series for each sliding time window as the dynamic statistical benchmark value, and the exponentially weighted moving standard deviation is used to calculate the volatility coefficient.

[0073] It should be noted that historical healthy operation data matching the fault-free operation period of the power distribution room within a preset time range is extracted from the time-series database associated with the digital twin. Abnormal outliers in the data are removed to obtain a clean historical dataset. For each monitoring parameter, an appropriate sliding time window length and sliding step size are set according to the operating cycle of the power distribution room equipment. The sliding time window data sequence based on the time dimension is extracted from the clean historical dataset. The exponentially weighted moving average algorithm is used to calculate the exponentially weighted moving average of the data sequence of each window as the dynamic statistical benchmark value of the parameter. The exponentially weighted moving standard deviation is calculated using the exponentially weighted moving variance algorithm as the fluctuation coefficient of the parameter.

[0074] This invention ensures the purity of historical analysis data by screening data from fault-free periods and removing outliers, providing a reliable health data foundation for subsequent threshold calculations. By setting sliding time window parameters based on equipment operating cycles, the data analysis more closely aligns with the actual operating patterns of the power distribution room. The exponential weighting algorithm assigns higher weights to recent data, enabling the calculated dynamic statistical benchmark and fluctuation coefficients to sensitively reflect the recent operating status and fluctuation characteristics of the parameters. This accurately captures the slow changing trend of parameters as the equipment operates, providing a scientific and dynamic statistical basis for constructing adaptive early warning thresholds.

[0075] In some embodiments, determining the upper and lower thresholds for early warning of each monitoring parameter that adapts to the cumulative operating time of the device and is associated with the monitoring point identifier specifically includes:

[0076] For each monitoring parameter, based on the preset confidence coefficient, the product of the confidence coefficient and the fluctuation coefficient is added to and subtracted from the dynamic statistical benchmark value to obtain the upper and lower warning limits of the parameter.

[0077] Obtain the cumulative running time of the power distribution equipment corresponding to the current monitoring point and the unique identifier of the monitoring point. Associate and store the upper and lower warning limits with the cumulative running time of the equipment and the identifier of the monitoring point to form an adaptive warning threshold library with traceability information.

[0078] It should be noted that, for each monitoring parameter, based on a preset confidence coefficient, the dynamic statistical benchmark value is calculated by adding and subtracting the product of the confidence coefficient and the fluctuation coefficient, respectively, to obtain the upper and lower warning limits of the parameter; the cumulative running time of the power distribution equipment corresponding to the current monitoring point and the unique identifier of the monitoring point are obtained, and the upper and lower warning limits are associated and stored with the cumulative running time of the equipment and the identifier of the monitoring point to form an adaptive warning threshold library with traceability information, so as to achieve precise binding of thresholds with equipment operating status and monitoring points.

[0079] This invention calculates early warning thresholds based on dynamic statistical benchmarks and fluctuation coefficients, allowing the thresholds to be dynamically adjusted according to the operating status and fluctuation characteristics of parameters. Combined with the setting of confidence coefficients, it balances the security and flexibility of monitoring. By associating and storing the early warning thresholds with the cumulative running time of the equipment and the identification of the monitoring point, it achieves adaptive adjustment of the early warning thresholds as the equipment ages, overcoming the drawbacks of the traditional fixed threshold approach. At the same time, it enables precise traceability of the thresholds, allowing for quick querying of the early warning thresholds for any monitoring point and any equipment running period. This provides data support for anomaly analysis and fault tracing, improving the adaptability and traceability of the early warning thresholds.

[0080] In some embodiments, real-time monitoring of the parameter values ​​and corresponding upper and lower thresholds of each monitoring node in the power distribution room specifically includes:

[0081] The real-time parameter values ​​of each monitoring node in the power distribution room are collected and processed in real time, and the timestamp alignment and standardization are completed.

[0082] Using the monitoring point identifier and the current cumulative running time of the equipment as query conditions, the latest upper and lower thresholds for each parameter's real-time value are matched and obtained from the adaptive early warning threshold library.

[0083] It should be noted that the raw parameter values ​​of each monitoring node in the power distribution room, collected in real time by the IoT acquisition terminal, are sequentially timestamped and standardized to obtain real-time parameter values ​​with the same format as historical analysis data. Using the monitoring point identifier and the current cumulative operating time of the device as core query conditions, the latest upper and lower warning thresholds corresponding to each parameter's real-time value are accurately matched from the adaptive warning threshold library, achieving one-to-one matching monitoring between real-time parameter values ​​and dynamic warning thresholds. This ensures the format consistency of real-time parameter values ​​with historical data and threshold calculation data, enabling comparable connection between real-time monitoring data and warning thresholds. Through the dual query conditions of monitoring point identifier and cumulative operating time of the device, precise matching between real-time parameter values ​​and corresponding warning thresholds is achieved, ensuring that parameters at each monitoring point and for each operating period can be matched with suitable dynamic thresholds. This avoids false alarms and missed alarms caused by threshold mismatches, improving the accuracy and reliability of real-time monitoring.

[0084] In some embodiments, when multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, calculating the comprehensive deviation of the current power distribution room operating status relative to the historical health benchmark specifically includes:

[0085] Set a warning zone ratio for each monitoring parameter, calculate the warning zone boundary outside the upper and lower warning thresholds, and determine whether the real-time value of the parameter enters the warning zone;

[0086] For monitoring parameters entering the warning zone, extract the time series data within the most recent preset sampling period, and use linear regression to fit the trend slope. If the absolute value of the slope exceeds the preset trend threshold and the direction is far away from the normal range, it is determined to show a continuous deviation trend.

[0087] When two or more monitoring parameters show a continuous deviation trend, the historical health mean and covariance matrix of such parameters are extracted, the real-time values ​​of the parameters are used to form the current observation vector, and the Mahalanobis distance of this vector relative to the historical health benchmark is calculated as the comprehensive deviation.

[0088] It should be noted that an appropriate warning zone ratio is set for each monitoring parameter, and the warning zone boundary outside the upper and lower warning thresholds is calculated. The real-time value of the parameter is compared with the warning zone boundary to determine whether the parameter has entered the warning zone. For monitoring parameters that have entered the warning zone, the time series data within the most recent preset sampling period is extracted, and the data trend slope is obtained by linear regression fitting. If the absolute value of the slope exceeds the preset trend threshold and the direction of change is far from the normal range, the parameter is determined to show a continuous deviation trend. When two or more monitoring parameters show a continuous deviation trend, the historical health mean and covariance matrix of such parameters are extracted, the real-time value of the parameter is used to construct the current observation vector, and the Mahalanobis distance of the vector relative to the historical health benchmark is calculated. This Mahalanobis distance is used as the comprehensive deviation of the substation operating status.

[0089] This invention enables early identification of parameter anomalies by setting warning zones, allowing time for handling fault warnings. It employs linear regression fitting to identify the continuous deviation trend of parameters, accurately capturing early anomalies where parameters are still within the warning zone but continuously deteriorating, avoiding misjudgments caused by fluctuations at a single sampling point. Based on Mahalanobis distance calculation of the comprehensive deviation, it fully considers the correlation between multiple parameters, enabling quantitative assessment of the deviation degree of the overall operating status of the power distribution room. This effectively identifies multi-parameter collaborative anomalies that are difficult to detect with single-parameter monitoring, achieving a scientific quantitative assessment of the overall operating status of the power distribution room and improving the sensitivity and accuracy of early fault warnings.

[0090] In some embodiments, generating a power distribution room fault early warning signal that includes monitoring point identifiers and anomaly types, and synchronizing it to the digital twin virtual model and the designated early warning receiver specifically includes:

[0091] When the overall deviation exceeds the preset deviation threshold or the continuous sampling period of a single parameter exceeds the warning threshold, the abnormal trigger type is determined, and information such as the trigger timestamp, monitoring point identifier, abnormal parameter value, and warning threshold is collected.

[0092] The above information is encapsulated according to the JSON data architecture to form a structured fault warning signal, which includes the anomaly type, source information and handling prompts;

[0093] The fault warning signal is pushed to the designated warning receiving end such as the operation and maintenance terminal and the power distribution room monitoring screen through the enterprise message bus. At the same time, the signal is sent through the internal interface of the digital twin virtual model to drive the alarm visualization display of the corresponding point in the virtual scene.

[0094] It should be noted that continuous monitoring is performed to check whether the overall deviation exceeds the preset deviation threshold and whether any monitoring parameter exceeds the upper or lower threshold of the warning within the continuous sampling period. When any monitoring condition is met, the abnormal trigger type is first determined, and then key information such as the trigger timestamp, monitoring point identifier, parameter abnormal value, and corresponding warning threshold are collected comprehensively. The above key information is encapsulated according to the JSON data architecture to form a structured fault warning signal containing the abnormal type, source information, and handling prompts. The fault warning signal is pushed to designated warning receiving terminals such as operation and maintenance terminals and power distribution room monitoring screens through the enterprise message bus. At the same time, the warning signal is sent through the internal interface of the digital twin virtual model to drive the corresponding points in the virtual scene to realize the alarm visualization display.

[0095] This invention constructs fault warning signals through structured encapsulation, ensuring the integrity, standardization, and transmissibility of warning information. The included traceability information enables maintenance personnel to quickly locate abnormal points, and the handling prompts provide initial guidance for on-site maintenance. The use of an enterprise message bus enables efficient push of warning signals to multiple terminals, ensuring real-time delivery of warning information. Through the linkage alarm of a digital twin virtual model, the synchronous display of physical anomalies and virtual scenes is achieved, allowing maintenance personnel to intuitively view the location and details of anomalies in the virtual model. This improves the efficiency of warning information transmission and the response speed of fault handling, realizing accurate and efficient multi-terminal linkage fault warning.

[0096] In some embodiments, the electrical operating parameters include the effective values ​​of three-phase voltage, the effective values ​​of three-phase current, the power factor, the switching status, and the bus temperature; the environmental characteristic parameters include the ambient temperature of the distribution room, relative humidity, dust concentration, and SF6 gas concentration.

[0097] It should be noted that, for the core electrical status of the power distribution equipment in the power distribution room and the overall operating environment of the power distribution room, the effective value of three-phase voltage, the effective value of three-phase current, power factor, switch opening and closing status, and bus temperature are selected as core electrical operating parameters. The ambient temperature, relative humidity, dust concentration, and SF6 gas concentration of the power distribution room are selected as key environmental characteristic parameters. The above parameters are used as the core monitoring indicators for power distribution room monitoring, and the accurate collection of each parameter is achieved through corresponding Internet of Things sensors.

[0098] This invention precisely selects key parameters that directly reflect the core operating status of power distribution equipment and the safety status of the power distribution room environment, achieving comprehensive and core-dimensional monitoring of the power distribution room's operating status. It covers key performance indicators of equipment electrical operation while also taking into account the impact of environmental factors on equipment safety operation, avoiding redundancy and omissions in monitoring parameters. By collecting various parameters through targeted sensors, the accuracy and effectiveness of monitoring data are ensured, providing core and reliable monitoring basis for subsequent status analysis and fault early warning, thus improving the comprehensiveness and relevance of power distribution room monitoring.

[0099] Reference Figure 3 Secondly, the present invention provides an automatic monitoring and data analysis system for power distribution rooms based on the Internet of Things, comprising:

[0100] The Internet of Things (IoT) data acquisition module is used to collect electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room, forming a multi-source monitoring parameter set;

[0101] The data processing module is used to perform timestamp alignment and heterogeneous data standardization on the multi-source monitoring parameter set, and push the processed parameter data to the digital twin virtual model corresponding to the power distribution room to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space.

[0102] The data processing module is also used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter by using a sliding time window mechanism combined with an exponential weighting algorithm to determine the historical health operation data of the power distribution room stored in the time series database associated with the digital twin virtual model. In this way, the upper and lower thresholds of the early warning for each monitoring parameter are determined to be adaptive with the cumulative running time of the equipment and associated with the monitoring point identifier for traceability.

[0103] The data processing module is also used to monitor the real-time values ​​of parameters of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, the module calculates the comprehensive deviation of the current power distribution room operating status from the historical health benchmark.

[0104] The alarm execution module is used to generate a power distribution room fault warning signal containing the monitoring point identifier and the abnormality type when the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper and lower warning thresholds within a continuous sampling period, and synchronize it to the digital twin virtual model and the designated warning receiver.

[0105] It should be noted that an IoT acquisition module is set up to collect electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room through various IoT sensors and acquisition terminals, forming a multi-source monitoring parameter set; a data processing module is set up to complete the timestamp alignment of multi-source monitoring parameters and the standardization of heterogeneous data, and push the processed data to the digital twin virtual model to realize the virtual mapping and visualization of physical state. At the same time, relying on historical health data in the time series database, an adaptive early warning threshold is constructed through a sliding time window and exponential weighting algorithm to match the real-time value of the parameter with the early warning threshold in real time, identify the continuous deviation trend of multiple parameters and calculate the comprehensive deviation degree; an alarm execution module is set up to monitor the comprehensive deviation degree and the single parameter over-limit situation, and generate a fault early warning signal with traceability information when the early warning criteria are met, and push it synchronously to the digital twin virtual model and the designated early warning receiver to realize virtual and real linkage alarm.

[0106] This invention, through a modular system architecture design, achieves independent implementation and collaborative linkage of data acquisition, data processing, status analysis, and fault early warning functions, improving the system's scalability and maintainability, and facilitating subsequent function upgrades and module replacements. The IoT acquisition module enables full-domain, real-time acquisition of multi-source parameters in the power distribution room, the data processing module completes standardized data processing and intelligent analysis, constructing a dynamic and adaptive early warning system, and the alarm execution module achieves accurate and efficient multi-terminal linkage early warning. The collaborative work of these modules forms a fully intelligent monitoring system from data acquisition to fault early warning, replacing the traditional manual monitoring and static alarm mode, significantly improving the automation and intelligence level of power distribution room monitoring, and providing comprehensive system protection for the safe and stable operation of the power distribution room.

[0107] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific embodiments.

[0108] refer to Figure 1 The figure is an exemplary flowchart of an IoT-based automatic monitoring and data analysis method for power distribution rooms according to some embodiments of the present invention. The method mainly includes the following steps:

[0109] In step 101, electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room are collected to form a multi-source monitoring parameter set.

[0110] In practice, IoT data acquisition terminals are deployed at key locations of power distribution equipment such as transformers, switchgear, and busbars in the power distribution room. These terminals include electrical parameter acquisition devices such as voltage and current sensors, temperature sensors, power factor analyzers, and switch status sensors, as well as environmental parameter acquisition devices such as temperature and humidity sensors, dust concentration sensors, and SF6 gas sensors. All acquisition devices are equipped with standard RS485 / 5G communication interfaces and unique monitoring point identifiers. Raw data is collected according to a preset sampling frequency (e.g., 50ms / time for electrical parameters and 5s / time for environmental parameters). The collected electrical operating parameters, such as the effective values ​​of three-phase voltage, effective values ​​of three-phase current, power factor, switch opening and closing status, and busbar temperature, as well as environmental characteristic parameters such as ambient temperature, relative humidity, dust concentration, and SF6 gas concentration in the power distribution room, are encapsulated into structured data packets with local timestamps, monitoring point identifiers, and acquisition device numbers. All structured data packets are aggregated to form a multi-source monitoring parameter set, awaiting subsequent processing.

[0111] It should be noted that the monitoring point identifier in this invention is a unique code for the deployment location of each data acquisition device in the power distribution room, which includes information such as device type, area number, and point serial number, and is used to achieve accurate correlation and traceability between parameter data and physical points.

[0112] In some embodiments, reference Figure 2 As shown in the figure, this is a schematic diagram of an application scenario of an automatic monitoring and data analysis system for a power distribution room according to some embodiments of the present invention. The figure includes an IoT acquisition terminal, an edge computing gateway, a cloud server, a time-series database, a digital twin visualization platform, and an operation and maintenance early warning terminal. The IoT acquisition terminal collects multi-source parameters of the power distribution room and sends them to the edge computing gateway. After the edge computing gateway completes data preprocessing, it uploads the data to the cloud server. The cloud server realizes data storage, analysis, and early warning calculation. The time-series database is used to store historical healthy operation data and real-time monitoring data. The digital twin visualization platform realizes the virtual mapping and visualization of the operating status of the power distribution room. The operation and maintenance early warning terminal receives fault early warning signals and realizes alarm prompts.

[0113] In step 102, the multi-source monitoring parameter set is timestamped and heterogeneous data is standardized. The processed parameter data is then pushed to the digital twin virtual model corresponding to the power distribution room, so as to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space.

[0114] In some embodiments, the timestamp alignment and heterogeneous data standardization processing of the multi-source monitoring parameter set specifically includes: obtaining structured data packets with local timestamps and monitoring point identifiers corresponding to all monitoring parameters, wherein the structured data packets contain the original values ​​of electrical operating parameters, the original values ​​of environmental characteristic parameters, and the data acquisition device number; uniformly calibrating the local timestamps of all structured data packets based on the IEEE 1588 precision clock synchronization protocol to obtain timestamp-aligned multi-source parameter data; and normalizing the timestamp-aligned heterogeneous parameter data according to a preset parameter per-unit formula to eliminate dimensional differences and obtain a standardized parameter dataset.

[0115] In practice, the edge computing gateway deploys a data aggregation service, which subscribes to and receives structured data packets sent by all IoT acquisition terminals via the Modbus-RTU protocol, and performs temporary caching and queue management on the data packets. The edge computing gateway and all acquisition terminals are connected to an industrial network based on the IEEE 1588 precision clock synchronization protocol, which has a unified nanosecond-level high-precision clock source. The data aggregation service converts the local timestamp of each data packet into a unified absolute timestamp, sets a fixed time alignment period (such as 1 second), aggregates all data packets within the same period, and fills in missing data by interpolating the valid value of the previous period to complete the timestamp alignment.

[0116] For the aligned heterogeneous parameters, a per-unit formula is used. Normalization is performed, where x is the original value of the parameter. , The maximum and minimum values ​​of this parameter are defined. After processing, all parameter values ​​are mapped to the [0,1] interval to eliminate dimensional differences and obtain a standardized parameter dataset.

[0117] In some embodiments, pushing the processed parameter data to the corresponding digital twin virtual model of the power distribution room to realize the real-time mapping and visualization of the physical operating status of the power distribution room in the virtual space specifically includes: sending the standardized parameter dataset to the data access layer of the digital twin virtual model through the MQTT industrial communication protocol, and simultaneously writing the parameter data into a time-series database for persistent storage; updating the status attributes of the corresponding virtual monitoring nodes, power distribution equipment, and power distribution room space in the digital twin virtual model according to the monitoring point identifiers and data types of the parameter data; driving the visualization engine of the digital twin virtual model to render and display the parameter dynamic dashboard, equipment operating status color mark, and parameter change trend curve of the three-dimensional virtual scene of the power distribution room according to the real-time changes of the status attributes.

[0118] In practical implementation, the edge computing gateway encapsulates the standardized parameter dataset into synchronous data frames containing frame sequence numbers and aligned timestamps, and publishes them to a designated topic via the MQTT protocol. The data access layer of the digital twin virtual model, acting as a subscriber, receives the synchronous data frames in real time. Simultaneously, the synchronous data frames are written to a time-series database for persistent storage via a JDBC interface. The data access layer parses the synchronous data frames, extracting monitoring point identifiers, parameter types, and standardized values. The simulation kernel of the digital twin virtual model locates the corresponding virtual monitoring nodes and power distribution equipment based on the monitoring point identifiers and assigns the parameter values ​​to them. Predefined status attribute variables, such as assigning bus temperature to the "operating temperature" attribute of the virtual bus and SF6 concentration to the "SF6 concentration" attribute of the virtual space of the power distribution room; the visualization engine monitors the changes of status attributes in real time, drives 3D rendering according to preset rules, displays dynamic parameter value dashboards next to virtual points, and marks equipment models as green (normal), yellow (warning), and red (abnormal) according to whether the parameter values ​​are within the normal range. The virtual monitoring panel draws parameter change trend curves for the past 1 hour / 24 hours, realizing real-time virtual mapping and visualization of the physical operating status of the power distribution room.

[0119] It should be noted that the digital twin virtual model in this invention is a 1:1 three-dimensional high-fidelity model built based on the physical structure and equipment layout of the power distribution room. It includes virtual entities of the power distribution room space, all power distribution equipment, and monitoring nodes, and realizes real-time data synchronization with the physical power distribution room, serving as a digital mirror of the power distribution room's operating status.

[0120] In step 103, based on the historical health operation data of the power distribution room stored in the time-series database associated with the digital twin virtual model, a sliding time window mechanism combined with an exponential weighting algorithm is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter, and then determine the upper and lower threshold values ​​of the early warning for each monitoring parameter that are adaptive with the cumulative running time of the equipment and are associated with the monitoring point identifier for traceability.

[0121] In some embodiments, the historical healthy operation data of the power distribution room stored in the time-series database associated with the digital twin virtual model is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter by employing a sliding time window mechanism combined with an exponential weighting algorithm. Specifically, this includes: extracting historical healthy operation data within a preset time range that matches the fault-free operation period of the power distribution room from the time-series database, and removing outliers to obtain a clean historical dataset; for each monitoring parameter, setting the window length and sliding step size according to the operating cycle of the power distribution room equipment, and extracting a time-dimensional sliding time window data sequence from the clean historical dataset; using an exponential weighted moving average algorithm to calculate the exponentially weighted moving average of each sliding time window data sequence as the dynamic statistical benchmark value, and using an exponentially weighted moving variance algorithm to calculate the exponentially weighted moving standard deviation as the fluctuation coefficient.

[0122] In practice, the cloud server extracts nearly six months of power distribution room monitoring data from the time-series database. Combined with power distribution room operation and maintenance records, it filters out healthy operating periods without faults or maintenance. The 3σ criterion is used to remove outliers from the data, resulting in a clean historical dataset. For each monitoring parameter, a sliding time window of 30 days and a sliding step of 1 day are set based on the rated operating cycle of the power distribution equipment. The parameter data sequence for each time window is extracted from the clean historical dataset. A smoothing coefficient α = 0.7 (with newer data having greater weight) is set, and an exponentially weighted moving average algorithm is used to recursively calculate the exponentially weighted moving average of the data sequence. As a dynamic statistical benchmark value, among which The parameter value at the current time. This is the average value from the previous moment;

[0123] Based on the mean, the exponentially weighted moving standard deviation is calculated using the exponentially weighted moving variance algorithm.

[0124] As a volatility coefficient, it quantifies the recent volatility of the parameter.

[0125] In some embodiments, determining the upper and lower warning thresholds for each monitoring parameter that are adaptive with the cumulative operating time of the equipment and associated with the monitoring point identifier for traceability specifically includes: for each monitoring parameter, based on a preset confidence coefficient, adding and subtracting the product of the confidence coefficient and the fluctuation coefficient from the dynamic statistical benchmark value respectively to obtain the upper and lower warning thresholds for the parameter; obtaining the cumulative operating time of the power distribution equipment corresponding to the current monitoring point and the unique identifier of the monitoring point, and associating and storing the upper and lower warning thresholds with the cumulative operating time of the equipment and the monitoring point identifier to form an adaptive warning threshold library with traceability information.

[0126] In practice, based on the safety requirements of the power distribution room operation, a confidence coefficient is set for each monitoring parameter (e.g., 3.0 for voltage and current parameters, and 2.5 for environmental parameters). The upper limit of the warning is calculated as: dynamic statistical benchmark value + confidence coefficient × fluctuation coefficient, and the lower limit of the warning is calculated as: dynamic statistical benchmark value - confidence coefficient × fluctuation coefficient. The cloud server obtains the cumulative running time (accurate to the hour) of the power distribution equipment corresponding to the current monitoring point through the equipment management module, as well as the unique identifier of the monitoring point. The record "monitoring point identifier - cumulative running time of equipment - upper limit of warning - lower limit of warning - calculation timestamp" is stored in the Redis adaptive warning threshold library to realize the correlation and traceability between the warning threshold and the equipment running time and the monitoring point. When the cumulative running time of the equipment increases by 100 hours, the warning threshold is automatically recalculated and updated to achieve adaptive adjustment of the threshold.

[0127] It should be noted that the upper and lower thresholds for early warning in this invention are the dynamic normal operating boundaries of each monitoring parameter. They are dynamically adjusted with the cumulative operating time of the equipment to adapt to the parameter baseline drift caused by the degradation of equipment performance. Compared with fixed thresholds, they are more in line with the actual operating status of the power distribution room.

[0128] In step 104, the real-time values ​​of parameters of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds are monitored in real time. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, the comprehensive deviation of the current operating status of the power distribution room from the historical health benchmark is calculated.

[0129] In some embodiments, real-time monitoring of the parameter values ​​of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds specifically includes: real-time acquisition and processing of the parameter values ​​of each monitoring node in the power distribution room, and completion of timestamp alignment and standardization processing; using the monitoring point identifier and the current cumulative running time of the equipment as query conditions, matching and obtaining the latest upper and lower warning thresholds corresponding to the real-time values ​​of each parameter from the adaptive warning threshold library.

[0130] In practice, the IoT data acquisition terminal collects raw parameter values ​​in real time. After the edge computing gateway completes the timestamp alignment and standardization processing, the data is uploaded to the cloud server to obtain the real-time parameter values. The cloud server extracts the monitoring point identifier corresponding to the real-time parameter values, queries the device management module to obtain the current cumulative runtime of the device corresponding to that point, and uses the monitoring point identifier and the device cumulative runtime as joint query conditions to match and obtain the upper and lower thresholds of the warning closest to the current runtime from the adaptive warning threshold library, thereby achieving accurate matching between the real-time parameter values ​​and dynamic thresholds.

[0131] In some embodiments, when multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, calculating the comprehensive deviation of the current power distribution room operating status relative to the historical health benchmark specifically includes: in step 301, setting a warning zone ratio for each monitoring parameter, calculating the warning zone boundary outside the upper and lower warning thresholds, and determining whether the real-time value of the parameter enters the warning zone; in step 302, for the monitoring parameters that have entered the warning zone, extracting their time-series data within the most recent preset sampling period, and using linear regression fitting to obtain the trend slope, if the absolute value of the slope exceeds the preset trend threshold and the direction is far from the normal range, it is determined to show a continuous deviation trend; in step 303, when there are two or more monitoring parameters showing a continuous deviation trend, extracting the historical health mean and covariance matrix of such parameters, constructing the current observation vector from the real-time values ​​of the parameters, and calculating the Mahalanobis distance of the vector relative to the historical health benchmark as the comprehensive deviation.

[0132] In practice, a warning zone ratio of 0.9 is set for each monitoring parameter. The upper boundary of the warning zone is calculated as the upper warning limit / 0.9, and the lower boundary is calculated as the lower warning limit × 0.9. If the real-time value of a parameter is greater than or equal to the upper warning limit and less than or equal to the upper boundary of the warning zone, or less than or equal to the lower warning limit and greater than or equal to the lower boundary of the warning zone, then the parameter is determined to have entered the warning zone. For parameters that have entered the warning zone, time-series data from the most recent 50 sampling periods are extracted, and linear regression fitting is performed using the least squares method to obtain the trend slope k. If |k| > a preset trend threshold (e.g., 0.02), and k is positive (warning zone on the upper warning limit side) or negative (warning zone on the lower warning limit side), then the parameter is determined to have a continuous deviation trend. When two or more parameters are detected to have a continuous deviation trend, the historical health mean of such parameters is extracted from the clean historical dataset, a historical mean vector μ is constructed, the covariance matrix Σ between parameters is calculated, and the current real-time value of the current parameter is used to construct the current observation vector X. The Mahalanobis distance formula is then used. The overall deviation is calculated to quantify the degree to which the overall operating status of the power distribution room deviates from the historical health benchmark.

[0133] It should be noted that the comprehensive deviation in this invention is a comprehensive indicator reflecting the multi-parameter coordinated anomalies in the power distribution room. It can effectively identify composite anomalies that cannot be detected by single-parameter monitoring and is the core criterion for realizing the overall operation status assessment and early fault warning of the power distribution room.

[0134] In step 105, when the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper or lower warning threshold within a continuous sampling period, a power distribution room fault warning signal containing the monitoring point identifier and the abnormality type is generated and synchronized to the digital twin virtual model and the designated warning receiver.

[0135] In some embodiments, generating a power distribution room fault early warning signal containing monitoring point identifiers and anomaly types, and synchronizing it to the digital twin virtual model and a designated early warning receiver specifically includes: when the overall deviation exceeds a preset deviation threshold or the continuous sampling period of a single parameter exceeds an early warning threshold, determining the anomaly trigger type, and collecting information such as trigger timestamps, monitoring point identifiers, parameter anomaly values, and early warning thresholds; encapsulating the above information according to a JSON data architecture to form a structured fault early warning signal, the signal containing anomaly type, source tracing information, and handling prompts; pushing the fault early warning signal to designated early warning receivers such as operation and maintenance terminals and power distribution room monitoring screens through an enterprise message bus, and simultaneously sending signals through the internal interface of the digital twin virtual model to drive the alarm visualization display of the corresponding points in the virtual scene.

[0136] In practice, the cloud server runs two monitoring threads simultaneously. The first thread continuously compares the overall deviation with a preset deviation threshold (calibrated through historical fault cases, such as the 95th percentile of Mahalanobis distance). If the deviation exceeds the threshold for three consecutive calculation cycles, an overall anomaly warning is triggered. The second thread maintains a continuous over-limit counter for each parameter. If the real-time value of a parameter exceeds the upper or lower threshold of the warning for five consecutive sampling cycles, the counter is triggered and the monitoring point identifier is recorded, triggering a single-parameter anomaly warning. When either thread triggers an warning, the cloud server's warning generation module immediately determines the anomaly trigger type (overall anomaly / single-parameter anomaly) and collects the trigger timestamp and monitoring point identifier. Information such as abnormal parameter values, warning thresholds, and cumulative equipment runtime are encapsulated into structured fault warning signals according to a preset JSON architecture. The signals include handling prompts based on the type of anomaly, such as "Immediately turn on the ventilation system and check for leaks" when SF6 concentration exceeds the limit. The warning signals are pushed to designated warning receivers such as the mobile APP of maintenance personnel, the power distribution room monitoring screen, and the power dispatching platform through the Kafka enterprise message bus. At the same time, the warning signals are sent through the internal event interface of the digital twin virtual model, which drives the corresponding monitoring points and equipment in the virtual scene to flash red alarm lights and pop up an alarm details window on the virtual panel to achieve virtual-real linkage alarm.

[0137] It should be noted that the preset deviation threshold is calibrated through historical failure cases. In the early stage of system deployment, a comprehensive deviation sample before the failure occurs is extracted, and the 95th percentile is taken as the initial threshold. Subsequently, it is regularly optimized and updated based on new failure cases and healthy operation data to ensure the sensitivity and accuracy of the early warning.

[0138] Furthermore, in another aspect, in some embodiments, the present invention provides an Internet of Things-based automatic monitoring and data analysis system for power distribution rooms, as referenced. Figure 3 The figure is a schematic diagram of an IoT-based automatic monitoring and data analysis system for a power distribution room according to some embodiments of the present invention. The system includes: an IoT acquisition module, a data processing module, and an alarm execution module, which are described below:

[0139] The IoT acquisition module is used to collect electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room to form a multi-source monitoring parameter set. This module consists of various IoT acquisition terminals and communication modules, which realize the full-domain, real-time acquisition of multi-dimensional parameters and encapsulate them into structured data packets.

[0140] The data processing module performs timestamp alignment and heterogeneous data standardization on the multi-source monitoring parameter set, and pushes the processed parameter data to the digital twin virtual model corresponding to the power distribution room to realize real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space. The data processing module also uses a sliding time window mechanism combined with an exponential weighting algorithm to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter based on the historical health operation data of the power distribution room stored in the time-series database associated with the digital twin virtual model. This determines the upper and lower warning thresholds for each monitoring parameter, which are adaptive with the cumulative operating time of the equipment and traceable by association with the monitoring point identifier. The data processing module also monitors the real-time parameter values ​​and corresponding upper and lower warning thresholds of each monitoring node in the power distribution room. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, the module calculates the comprehensive deviation of the current power distribution room operation status from the historical health benchmark. This module is deployed on an edge computing gateway and a cloud server to realize functions such as data preprocessing, dynamic threshold calculation, and operation status analysis.

[0141] The alarm execution module is used to generate a power distribution room fault warning signal containing the monitoring point identifier and the type of abnormality when the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper or lower warning threshold within a continuous sampling period, and synchronize it to the digital twin virtual model and the designated warning receiving terminal. This module realizes the generation, release and virtual-real linkage alarm of the warning signal, and completes the multi-terminal push of the warning information.

[0142] Each module in the above system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0143] In another embodiment, the present invention provides a computer device, which may be a cloud server, and its internal structure diagram may be as follows. Figure 4As shown, the computer device includes a processor, memory, network interface, and time-series database interface connected via a system bus. The processor provides computing and control capabilities, supporting data analysis and early warning calculations for the entire system. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a device management database, while the internal memory provides an environment for the operating system and computer programs in the non-volatile storage media to run. The time-series database interface establishes a data connection with the time-series database, enabling the reading and writing of historical and real-time data. The network interface communicates with edge computing gateways, digital twin visualization platforms, and operation and maintenance early warning terminals via a network connection. When the computer program is executed by the processor, it implements the IoT-based automatic monitoring and data analysis method for power distribution rooms described in this invention.

[0144] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0145] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above embodiment of the Internet of Things-based automatic monitoring and data analysis method for power distribution rooms.

[0146] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps described in the embodiment of the Internet of Things-based automatic monitoring and data analysis method for power distribution rooms.

[0147] In one embodiment, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps described in the embodiment of the IoT-based automatic monitoring and data analysis method for power distribution rooms.

[0148] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0149] The technical solutions provided by the embodiments disclosed in this invention have the following beneficial effects:

[0150] This invention achieves full-domain acquisition of multi-dimensional parameters of electrical operation and environmental characteristics of the power distribution room through an IoT acquisition module. Combined with timestamp alignment and heterogeneous data standardization processing, it eliminates the time difference and dimensional differences of multi-source data, providing a unified and standardized data foundation for subsequent analysis. At the same time, it realizes virtual mapping and visualization of the physical power distribution room through a digital twin virtual model, allowing operation and maintenance personnel to intuitively grasp the overall operating status of the power distribution room.

[0151] Based on the sliding time window and exponential weighting algorithm, the dynamic statistical benchmark value and fluctuation coefficient of each parameter are determined. An adaptive early warning threshold library is constructed in combination with the cumulative running time of the equipment. This realizes the dynamic adjustment of the early warning threshold as the equipment ages, overcomes the drawback of a one-size-fits-all fixed threshold, improves the tolerance of normal aging fluctuations of the equipment and the accuracy of early fault and anomaly identification, and reduces false alarms and missed alarms.

[0152] By identifying the coordinated deviation trend of multiple parameters and calculating the comprehensive deviation degree, a quantitative assessment of the overall operating status of the power distribution room is achieved. It can capture composite anomalies and early weak fault trends that are difficult to detect by single parameter monitoring, and advance the early warning time from after the fault occurs to the fault initiation stage, thus achieving predictive early warning.

[0153] The generated fault warning signal contains complete monitoring point source information and anomaly type, and realizes linkage alarm between the warning signal and the digital twin virtual model, enabling operation and maintenance personnel to accurately locate abnormal points, quickly determine the anomaly type, and improve the efficiency and accuracy of fault handling.

[0154] The system integrates data acquisition, processing, analysis, early warning, and visualization, realizing the intelligent and automated monitoring of the entire power distribution room process. It replaces the traditional manual inspection mode, greatly improves the real-time performance and intelligence level of power distribution room monitoring, and ensures the safe and stable operation of the power distribution room.

[0155] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things, characterized in that, Includes the following steps: The electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room are collected to form a multi-source monitoring parameter set; The multi-source monitoring parameter set is timestamped and heterogeneous data is standardized. The processed parameter data is then pushed to the digital twin virtual model corresponding to the power distribution room, so as to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space. Based on the historical health operation data of the power distribution room stored in the time series database associated with the digital twin virtual model, a sliding time window mechanism combined with an exponential weighting algorithm is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter, and then determine the upper and lower thresholds of the early warning for each monitoring parameter that are adaptive with the cumulative running time of the equipment and associated with the monitoring point identifier for traceability. Real-time monitoring of the parameter values ​​of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, calculate the comprehensive deviation of the current power distribution room operation status from the historical health benchmark. When the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper or lower warning threshold within a continuous sampling period, a power distribution room fault warning signal containing the monitoring point identifier and the anomaly type is generated and synchronized to the digital twin virtual model and the designated warning receiver.

2. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, The specific steps for timestamp alignment and heterogeneous data standardization of the multi-source monitoring parameter set include: Obtain structured data packets with local timestamps and monitoring point identifiers corresponding to all monitoring parameters. The structured data packets contain the original values ​​of electrical operating parameters, the original values ​​of environmental characteristic parameters, and the data acquisition device number. Based on the IEEE 1588 precision clock synchronization protocol, the local timestamps of all structured data packets are uniformly calibrated to obtain multi-source parameter data after timestamp alignment. According to the preset parameter normalization formula, the heterogeneous parameter data after timestamp alignment is normalized to eliminate the difference in dimensions and obtain a standardized parameter dataset.

3. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, The processed parameter data is pushed to the corresponding digital twin virtual model of the power distribution room to realize the real-time mapping and visualization of the physical operating status of the power distribution room in the virtual space. Specifically, this includes: The standardized parameter dataset is sent to the data access layer of the digital twin virtual model via the MQTT industrial communication protocol, and the parameter data is written to the time series database in parallel for persistent storage. Based on the monitoring point identifiers and data types of the parameter data, update the status attributes of the corresponding virtual monitoring nodes, power distribution equipment, and power distribution room space in the digital twin virtual model; The visualization engine that drives the digital twin virtual model renders and displays a dynamic parameter dashboard, equipment operation status color mark, and parameter change trend curve of the 3D virtual scene of the power distribution room based on real-time changes in status attributes.

4. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, Based on the historical health operation data of the power distribution room stored in the time-series database associated with the digital twin virtual model, a sliding time window mechanism combined with an exponential weighting algorithm is used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter, specifically including: Extract historical healthy operation data that match the fault-free operation period of the power distribution room within a preset time range from the time series database, and remove abnormal outliers to obtain a clean historical dataset. For each monitoring parameter, the window length and sliding step size are set according to the operating cycle of the power distribution room equipment, and the sliding time window data sequence based on the time dimension is extracted from the clean historical dataset; The exponentially weighted moving average algorithm is used to calculate the exponentially weighted moving average of the data series for each sliding time window as the dynamic statistical benchmark value, and the exponentially weighted moving standard deviation is used to calculate the volatility coefficient.

5. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, The specific upper and lower thresholds for early warning, which are determined by adapting each monitoring parameter to the cumulative operating time of the equipment and being linked to the monitoring point identifier for traceability, include: For each monitoring parameter, based on the preset confidence coefficient, the product of the confidence coefficient and the fluctuation coefficient is added to and subtracted from the dynamic statistical benchmark value to obtain the upper and lower warning limits of the parameter. Obtain the cumulative running time of the power distribution equipment corresponding to the current monitoring point and the unique identifier of the monitoring point. Associate and store the upper and lower warning limits with the cumulative running time of the equipment and the identifier of the monitoring point to form an adaptive warning threshold library with traceability information.

6. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, The real-time monitoring of parameters and corresponding upper and lower thresholds for each monitoring node in the power distribution room includes: The real-time parameter values ​​of each monitoring node in the power distribution room are collected and processed in real time, and the timestamp alignment and standardization are completed. Using the monitoring point identifier and the current cumulative running time of the equipment as query conditions, the latest upper and lower thresholds for each parameter's real-time value are matched and obtained from the adaptive early warning threshold library.

7. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower thresholds and show a continuous deviation trend, the comprehensive deviation of the current power distribution room's operating status from the historical health benchmark is calculated, specifically including: Set a warning zone ratio for each monitoring parameter, calculate the warning zone boundary outside the upper and lower warning thresholds, and determine whether the real-time value of the parameter enters the warning zone; For monitoring parameters entering the warning zone, extract the time series data within the most recent preset sampling period, and use linear regression to fit the trend slope. If the absolute value of the slope exceeds the preset trend threshold and the direction is far away from the normal range, it is determined to show a continuous deviation trend. When two or more monitoring parameters show a continuous deviation trend, the historical health mean and covariance matrix of such parameters are extracted, the real-time values ​​of the parameters are used to form the current observation vector, and the Mahalanobis distance of this vector relative to the historical health benchmark is calculated as the comprehensive deviation.

8. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, Generating a power distribution room fault early warning signal that includes monitoring point identifiers and anomaly types, and synchronizing it to the digital twin virtual model and the designated early warning receiver specifically includes: When the overall deviation exceeds the preset deviation threshold or the continuous sampling period of a single parameter exceeds the warning threshold, the abnormal trigger type is determined, and information including the trigger timestamp, monitoring point identifier, abnormal parameter value, and warning threshold is collected. The information is encapsulated according to the JSON data architecture to form a structured fault warning signal, which includes the anomaly type, source information and handling prompts; The fault warning signal is pushed to the designated warning receiving end, including the operation and maintenance terminal or the power distribution room monitoring screen, through the enterprise message bus. At the same time, the signal is sent through the internal interface of the digital twin virtual model to drive the alarm visualization display of the corresponding point in the virtual scene.

9. The method for automatic monitoring and data analysis of power distribution rooms based on the Internet of Things as described in claim 1, characterized in that, The electrical operating parameters include the effective values ​​of three-phase voltage, the effective values ​​of three-phase current, the power factor, the switch opening and closing status, and the bus temperature. The environmental characteristic parameters include the ambient temperature of the power distribution room, relative humidity, dust concentration, and SF6 gas concentration.

10. An automatic monitoring and data analysis system for power distribution rooms based on the Internet of Things, characterized in that, include: The Internet of Things (IoT) data acquisition module is used to collect electrical operating parameters and environmental characteristic parameters of each monitoring node in the power distribution room, forming a multi-source monitoring parameter set; The data processing module is used to perform timestamp alignment and heterogeneous data standardization on the multi-source monitoring parameter set, and push the processed parameter data to the digital twin virtual model corresponding to the power distribution room to realize the real-time mapping and visualization of the physical operation status of the power distribution room in the virtual space. The data processing module is also used to determine the dynamic statistical benchmark value and fluctuation coefficient of each monitoring parameter by using a sliding time window mechanism combined with an exponential weighting algorithm to determine the historical health operation data of the power distribution room stored in the time series database associated with the digital twin virtual model. In this way, the upper and lower thresholds of the early warning for each monitoring parameter are determined to be adaptive with the cumulative running time of the equipment and associated with the monitoring point identifier for traceability. The data processing module is also used to monitor the real-time values ​​of parameters of each monitoring node in the power distribution room and the corresponding upper and lower warning thresholds. When multiple monitoring parameters enter the warning zone outside the corresponding upper and lower warning thresholds and show a continuous deviation trend, the module calculates the comprehensive deviation of the current power distribution room operating status from the historical health benchmark. The alarm execution module is used to generate a power distribution room fault warning signal containing the monitoring point identifier and the abnormality type when the overall deviation exceeds the preset deviation threshold, or when any monitoring parameter exceeds the corresponding upper and lower warning thresholds within a continuous sampling period, and synchronize it to the digital twin virtual model and the designated warning receiver.

Images