Power supply monitoring method based on Internet of Things distributed architecture

Abstract

The invention belongs to the technical field of Internet of Things and power supply monitoring, and particularly relates to a power supply monitoring method based on an Internet of Things distributed architecture, which is realized based on a distributed monitoring terminal, an edge computing node and a cloud monitoring platform, and comprises the following steps: collecting multiple types of parameters of a power supply to be monitored and packaging the parameters into initial monitoring data; after preprocessing, environment parameter fusion calibration is carried out, and terminal identification and position data are associated to construct a data set; hierarchically judging power supply working and battery states, and dynamically optimizing a threshold value; generating a monitoring report and constructing a monitoring topological graph; and updating the state and triggering early warning based on a multi-dimensional condition. According to the invention, data processing pressure is dispersed, monitoring precision is improved, real-time monitoring, precise early warning and efficient operation and maintenance of the distributed power supply are realized, the problems of high delay, threshold stiffness, inaccurate early warning and the like in the prior art are solved, and the method is adaptive to a distributed power supply dispersed deployment scene.

Description

Technical Field

[0001] This invention belongs to the field of Internet of Things (IoT) and power monitoring technology, and particularly relates to a power monitoring method based on an IoT distributed architecture. Background Technology

[0002] With the rapid development of IoT technology, the distributed deployment of various electronic and industrial equipment is becoming increasingly widespread. As a result, the number of power supplies to be monitored (such as industrial control power supplies, IoT terminal backup power supplies, distributed energy storage power supplies, etc.) has increased significantly, and the deployment scenarios are becoming more decentralized and complex. This places higher demands on the real-time monitoring, accurate early warning, and efficient operation and maintenance of power supply status.

[0003] Currently, most existing power supply monitoring technologies adopt a centralized monitoring architecture, which involves collecting operating parameters from multiple power supplies through a single monitoring terminal and then aggregating them to a monitoring platform for processing. This architecture has significant drawbacks: First, the monitoring range is limited and it is difficult to adapt to distributed power supply scenarios. When power supplies are distributed in different areas, the data transmission latency is high and the stability is poor, making real-time monitoring impossible. Second, the data processing accuracy is insufficient. Existing technologies mostly use raw monitoring data directly for status determination without considering the interference of environmental factors such as temperature, humidity, and electromagnetic interference on the accuracy of electrical variable parameters and battery parameters. This results in a large error in status determination and is prone to misjudgment and missed judgment. Third, the threshold setting is rigid, and fixed thresholds are mostly used for status determination. It is impossible to dynamically optimize the threshold based on the historical data of the power supply's long-term operation, adapt to the operating characteristics and aging status of different power supplies, and further reduce the accuracy of monitoring. Fourth, the data correlation is low. The location data of the power supply, monitoring parameters, status judgment results and threshold data are not effectively correlated, making it difficult to generate a comprehensive monitoring report. Furthermore, the distribution and operating status of distributed power supplies cannot be presented intuitively through visualization, making it difficult for maintenance personnel to quickly locate faulty power supplies and carry out maintenance work efficiently. Fifth, the early warning mechanism is imperfect, triggering warnings only based on a single abnormal parameter, without combining multiple dimensions such as parameter mutation and duration of abnormality for comprehensive early warning. This results in low accuracy of early warnings and makes it difficult for maintenance personnel to address faults in a targeted manner.

[0004] Furthermore, in existing distributed monitoring technologies, edge nodes only handle data transmission and do not participate in data preprocessing and fusion calibration. This results in excessive data processing pressure on the cloud platform, further reducing the response speed of the monitoring system. Simultaneously, existing technologies lack a robust multi-source data fusion calibration mechanism, failing to effectively eliminate the impact of environmental interference on monitoring data and thus making it difficult to meet the demands of high-precision power supply monitoring.

[0005] To address the shortcomings of the existing technologies, this invention aims to provide a power monitoring method based on an IoT distributed architecture, which solves the technical problems of high data transmission latency, low monitoring accuracy, rigid thresholds, low data correlation, inaccurate early warning, and low operation and maintenance efficiency in distributed power monitoring, thereby achieving real-time, accurate, efficient monitoring and intelligent operation and maintenance of distributed power. Summary of the Invention

[0006] The purpose of this invention is to provide a power monitoring method based on an IoT distributed architecture to solve the technical problems of high data transmission latency, low monitoring accuracy, rigid thresholds, low data correlation, inaccurate early warning, and low operation and maintenance efficiency in distributed power monitoring, thereby achieving real-time, accurate, efficient monitoring and intelligent operation and maintenance of distributed power.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A power monitoring method based on an IoT distributed architecture, the method comprising several distributed monitoring terminals, edge computing nodes, and a cloud monitoring platform, the method comprising the following steps: S1: Collect real-time location data, real-time electrical variable parameters, surrounding environmental parameters, and real-time battery parameters of the power supply to be monitored, and package all data into initial monitoring data; S2: Preprocess the initial monitoring data to obtain standardized monitoring data; fuse and calibrate other standardized monitoring data (excluding environmental parameters) based on environmental parameters to obtain calibrated standardized monitoring data; associate each calibrated standardized monitoring data with the corresponding terminal identifier and real-time location data to construct a distributed power source monitoring dataset; S3: Based on preset electrical variable thresholds and preset battery parameter thresholds, the distributed power source monitoring dataset is hierarchically judged to obtain the working status judgment results of the power source to be monitored and the built-in battery status judgment results, and the preset electrical variable thresholds and preset battery parameter thresholds are dynamically iterated and updated to achieve adaptive optimization of the thresholds; S4: Associate the real-time location data, working status determination results and built-in battery status determination results of the power supply to be monitored, and combine them with the optimized threshold output by the dynamic threshold self-learning module to generate a distributed monitoring report for each power supply to be monitored. At the same time, construct a distributed power supply monitoring topology map based on the location data of each power supply to be monitored. S5: Based on the calibrated standardized monitoring data and the dynamically updated preset thresholds, update the working status judgment results of the power supply to be monitored and the built-in battery status judgment results; when the judgment results meet the preset warning conditions, generate warning information including the real-time location data of the power supply to be monitored, the abnormal status type, the abnormal parameters and the warning information.

[0008] Preferably, the specific process of fusing and calibrating other standardized monitoring data besides environmental parameters based on environmental parameters in step S2 to obtain calibrated standardized monitoring data is as follows: S21: Record the standardized real-time electrical variable parameters as X, the real-time battery parameters as Y, as the core monitoring parameters, and the standardized environmental parameters as Z, as the interference calibration parameters, including temperature T, humidity H, and electromagnetic interference intensity E. S22: Calculate the interference weights of temperature, humidity, and electromagnetic interference intensity on electrical variable parameters and battery parameters, respectively; S23: Adaptive calibration of standardized electrical variable parameters and battery parameters based on comprehensive interference weights; S24: Associate the calibrated electrical variable parameters and battery parameters with the standardized real-time location data to obtain the calibrated standardized monitoring data.

[0009] Preferably, the specific process of step S22 is as follows: S221: Calculate the temperature disturbance weight, the specific formula is as follows: ; ; in, w T,X The single disturbance weight of temperature on the electrical variable parameter X. w T,Y The single disturbance weight of temperature on battery parameter Y; k T Temperature interference coefficient; T The standardized real-time ambient temperature parameters; T 0 represents the standard operating temperature of the power supply to be monitored; S222: Calculate the humidity interference weight, the specific formula is as follows: ; ; in, w H,X The single disturbance weight of humidity on the electrical variable parameter X. w H,Y The single disturbance weight of humidity on battery parameter Y; k H H represents the humidity interference coefficient; H is the standardized real-time environmental humidity parameter. H 0 represents the standard operating humidity of the power supply to be monitored; S223: Calculate the interference weight for electromagnetic interference intensity. The specific formula is as follows: ; ; in, w E,X The electromagnetic interference intensity is the single interference weight of the electrical variable parameter X. w E,Y The single interference weight of electromagnetic interference intensity on battery parameter Y; k E E is the electromagnetic interference intensity coefficient; E is the standardized real-time electromagnetic interference intensity parameter. E 0 represents the standard operating electromagnetic interference intensity of the power supply to be monitored; S224: Calculate the comprehensive interference weight of the core monitoring parameters. The specific formula is as follows: w X =α· w T,X + β · w H,X +γ· w E,X ; w Y =α· w T,Y + β · w H,Y +γ· w E,Y ; in, w X The combined disturbance weights for electrical variable parameters, w Y The overall interference weights for battery parameters; α, β γ and γ are the weighting coefficients for temperature, humidity, and electromagnetic interference intensity, respectively.

[0010] Preferably, the specific process of step S23 is as follows: S231: Perform electrical variable parameter calibration, the specific formula is as follows: X cal = X ·(1- w X )+ X 0· w X ; in, X cal These are the standardized electrical variable parameters after calibration. X These are the standardized real-time electrical variable parameters; w X The comprehensive disturbance weights for electrical variable parameters; X0 represents the standard reference value for electrical variable parameters; S232: Perform battery parameter calibration. The specific formula is as follows: Y cal = Y ·(1- w Y )+ Y 0· w Y ; in, Y cal Standardized battery parameters after calibration; Y These are the standardized real-time battery parameters; w Y The overall interference weight for battery parameters; Y 0 represents the standard baseline value for battery parameters.

[0011] Preferably, the specific process of constructing the distributed power source monitoring dataset by associating each calibrated standardized monitoring data with the corresponding terminal identifier and real-time location data in step S2 is as follows: S25: Extract the terminal identifier marked when each distributed monitoring terminal uploaded the initial monitoring data, and denote it as... ID i The standardized real-time location data after preprocessing is denoted as... L i Including latitude and longitude coordinates ( Lon i , Lat i ) and altitude Alt i Extract the standardized monitoring data after calibration and record it as follows: D cal,i ; S26: Employ a key-value pair mapping algorithm, using terminal identifiers. ID i Using a unique primary key, establish and calibrate standardized monitoring data. D cal,i One-to-one association; S27: Based on terminal identifier ID i The uniqueness of the terminal identifier-calibrated monitoring data pair is determined by linking it with the corresponding standardized real-time location data. L i A secondary correlation is performed to form a three-dimensional correlated data system consisting of terminal identifier, location data, and calibrated monitoring data. Assoc ( ID i ); S28: Associate the three-dimensional data of all terminals Assoc (ID i Summarize the data, sort by terminal identifier, and supplement with monitoring timestamps. t i We constructed a structured distributed power monitoring dataset DS.

[0012] Preferably, the specific process for stratifying the calibrated and standardized monitoring data in the distributed power source monitoring dataset in step S3 is as follows: S31: From the distributed power source monitoring dataset DS, sorted by terminal identifier ID i Extract the corresponding calibrated and standardized monitoring data one by one, and match them with the historical monitoring data of the corresponding terminals. DS his,i This includes parameters calibrated in the past 30 days and historical judgment results; S32: Retrieve the set of initial preset electrical variable thresholds stored in the cloud monitoring platform. Th X ={ Th X1 ,Th X2 ,..., Th Xk Initial preset battery parameter threshold set Th Y ={Th Y1 , Th Y2 ,..., Th Ym}; S33: Working Status Layer Determination: Comparison One by One X cal,i Each dimension parameter and its corresponding Th X Thresholds are used to determine the operating status of the power supply to be monitored. S34: Built-in battery status layering determination: comparison one by one Y cal,i Each dimension parameter and its corresponding Th Y Threshold to determine the status of the built-in battery; S35: Layered determination result output: Summarize the operating status of each monitored power supply. Status X,i Built-in battery status Status Y,i Associate with the corresponding terminal identifier ID i and real-time location data L i This forms the associated data of terminal identifier, location, and dual-state determination results.

[0013] Preferably, the specific process of dynamically iteratively updating the preset electrical variable threshold and the preset battery parameter threshold in step S3 to achieve adaptive optimization of the threshold is as follows: S36: Obtain the set of judgment results Statu si ={ Status X,i ,Status Y,i}, Calibrated parameters in the current distributed power source monitoring dataset DS X cal,i , Y cal,i Historical monitoring dataset DS his This includes calibration parameters and corresponding judgment results for the past 30 days. S37: Employ an outlier removal algorithm to filter calibrated parameters whose results are normal from historical monitoring data and remove parameters whose results are abnormal. S38: Employs a weighted moving average iterative algorithm, combining historical valid data with current calibration data, to dynamically update the preset electrical variable threshold. Th X Preset battery parameter thresholds Th Y ; S39: Verify the deviation between the updated threshold and the initial threshold. If the deviation is within 5%, it will take effect directly and replace the original preset threshold. If the deviation exceeds 5%, it will be corrected in conjunction with the rated parameters of the power supply to be monitored, and will take effect after correction. The optimized threshold after taking effect will be synchronized to the edge computing node.

[0014] Preferably, the specific process of step S4 is as follows: S41: From the distributed power source monitoring dataset DS, sorted by terminal identifier ID i Extract standardized real-time location data one by one L i Standardized electrical variable parameters after calibration X cal,i Standardized battery parameters after calibration Y cal,i Extract the first i Dual-state determination results of the power supply to be monitored Status i The optimized threshold output by the dynamic threshold self-learning module Th i,new ; by terminal identifier ID i Using a unique primary key, location data, status determination results, optimized thresholds, and calibrated monitoring data are bound together in four dimensions to form a complete associated dataset for a single power supply to be monitored.Data i ; S42: For those that pass the verification Data i Extract and organize basic information, including terminal identifiers. ID i Power supply model to be monitored, monitoring timestamp t i Real-time location raw data; Data i Standardized monitoring data after calibration D cal,i , and the optimized threshold Th i,new Compare the parameters and calculate the deviation rate between each parameter and the threshold. S43: Generate a distributed monitoring report for a single monitored power supply based on the deviation rate of each parameter from the threshold. Report i It includes a basic information module, a monitoring parameter module, a status determination module, a threshold module, and an analysis and suggestion module; S44: Based on location data, construct a visual topology map to intuitively display the status of multiple power sources.

[0015] Preferably, the specific process of step S44 is as follows: S441: Data on the actual location of all power sources to be monitored L real,i Convert to Cartesian coordinates required for topological graph drawing ( X map,i ,Y map,i ); S442: In plane rectangular coordinates ( X map,i ,Y map,i ) is the node location. Draw the topology nodes of each power supply to be monitored, and use different colors and shapes of nodes to distinguish the working status and the built-in battery status. S443: Label the terminal identifier next to each topology node. ID i The core monitoring parameters and status determination results are labeled using a node offset labeling algorithm to ensure that the labels do not overlap.

[0016] The beneficial effects of this invention include: 1. Adopting an IoT distributed architecture, relevant parameters of each power source to be monitored are collected through several distributed monitoring terminals. Edge computing nodes are responsible for data preprocessing, fusion calibration and dataset construction, while the cloud platform is responsible for status determination, threshold optimization, report generation and early warning push. This effectively distributes the data processing pressure, reduces data transmission latency, adapts to the distributed deployment scenario of distributed power sources, realizes real-time monitoring of the power sources to be monitored, and solves the technical problems of limited monitoring range and slow response speed of the existing centralized architecture.

[0017] 2. A multi-source data fusion calibration algorithm is introduced to calculate the interference weights on electrical variable parameters and battery parameters based on environmental parameters (temperature, humidity, electromagnetic interference intensity). Then, the standardized core monitoring parameters are calibrated through an adaptive calibration formula. This effectively eliminates the interference of environmental factors on the accuracy of monitoring data acquisition, improves the accuracy of monitoring data, and provides reliable data support for subsequent status determination and early warning. It solves the technical problems of existing technologies that do not consider environmental interference and have low monitoring accuracy.

[0018] 3. By using key-value pair mapping and secondary association, the system achieves precise binding of calibrated monitoring data, terminal identifiers, and real-time location data, constructing a structured distributed power supply monitoring dataset. This ensures the correlation and integrity of multi-dimensional data, laying the foundation for subsequent hierarchical judgment, report generation, and topology map construction, and solving the technical problem of low data correlation in existing technologies. At the same time, the regularization and sorting of the dataset facilitates the cloud platform to quickly retrieve and process monitoring data from each terminal, improving data processing efficiency.

[0019] 4. A layered judgment method is adopted to independently judge the working status of the power supply and the status of the built-in battery. Combined with preset thresholds, the state classification is achieved with precision. Compared with the existing single state judgment method, the judgment results are more comprehensive and accurate. At the same time, based on historical valid data and current judgment results, the preset threshold is dynamically updated through a weighted moving average iterative algorithm to achieve adaptive optimization of the threshold, adapting to the operating characteristics and aging status of different power supplies. This avoids the problems of misjudgment and omission caused by fixed thresholds, and further improves the accuracy of state judgment. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the power monitoring method based on an IoT distributed architecture according to the present invention.

[0021] Figure 2 This is a schematic diagram of the fusion calibration process of the present invention.

[0022] Figure 3 This is a schematic diagram of the layer determination process of the present invention. Detailed Implementation

[0023] The following is in conjunction with the appendixFigures 1 - 3 The present invention will be further described in detail below: Example 1 See appendix Figure 1 As shown, a power monitoring method based on an IoT distributed architecture is applied to a distributed power online monitoring system. The method is implemented based on an IoT distributed architecture, which includes several distributed monitoring terminals, edge computing nodes, and a cloud monitoring platform. Each distributed monitoring terminal corresponds one-to-one with the power source to be monitored and has a built-in positioning module, electrical parameter acquisition module, battery status acquisition module, and environmental sensing module. The method includes the following steps: S1: The distributed monitoring terminal initializes, completes encrypted communication connection with the edge computing node, and simultaneously collects real-time location data of the power supply to be monitored through the built-in positioning module, real-time electrical variable parameters of the power supply to be monitored through the electrical parameter acquisition module, real-time battery parameters of the built-in battery of the power supply to be monitored through the battery status acquisition module, and environmental parameters around the power supply to be monitored through the environmental perception module. The real-time location data, real-time electrical variable parameters, real-time battery parameters and environmental parameters are packaged into initial monitoring data, marked with the terminal identifier and uploaded to the edge computing node.

[0024] S2: The edge computing node receives the initial monitoring data uploaded by each distributed monitoring terminal, and preprocesses the initial monitoring data, including data denoising, outlier removal, and data standardization, to obtain standardized monitoring data. At the same time, a multi-source data fusion calibration algorithm is introduced to fuse and calibrate the standardized real-time electrical variable parameters, real-time battery parameters, and environmental parameters to eliminate the interference of environmental factors on the acquisition accuracy of electrical and battery parameters, and obtain calibrated standardized monitoring data. Each calibrated standardized monitoring data is associated with the corresponding terminal identifier and real-time location data to construct a distributed power supply monitoring dataset, and the distributed power supply monitoring dataset is synchronized to the cloud monitoring platform.

[0025] S3: The cloud monitoring platform receives the distributed power monitoring dataset synchronized by the edge computing nodes. Based on preset electrical variable thresholds and preset battery parameter thresholds, it performs hierarchical judgment on the calibrated standardized monitoring data in the distributed power monitoring dataset to obtain the working status judgment result and the built-in battery status judgment result of the power supply to be monitored. At the same time, it starts the dynamic threshold self-learning module, which dynamically iteratively updates the preset electrical variable thresholds and preset battery parameter thresholds based on historical monitoring data and judgment results to achieve adaptive optimization of the thresholds. The working status judgment result is determined based on the comparison result of the calibrated real-time electrical variable parameters and the preset electrical variable thresholds, and the built-in battery status judgment result is determined based on the comparison result of the calibrated real-time battery parameters and the preset battery parameter thresholds.

[0026] S4: The cloud monitoring platform associates the real-time location data, working status judgment results, and built-in battery status judgment results of the power supply to be monitored, and combines the optimized threshold output by the dynamic threshold self-learning module to generate a distributed monitoring report for each power supply to be monitored. At the same time, it constructs a distributed power supply monitoring topology map based on the location data of each power supply to be monitored, and pushes the distributed monitoring report and distributed power supply monitoring topology map to the operation and maintenance terminal.

[0027] S5: The cloud monitoring platform receives updated and calibrated standardized monitoring data synchronized from edge computing nodes in real time. Based on the updated and calibrated standardized monitoring data and dynamically updated preset thresholds, it updates the working status judgment results and built-in battery status judgment results of the power supply to be monitored. When the judgment results meet the preset warning conditions, it generates warning information containing the real-time location data of the power supply to be monitored, the abnormal status type, and abnormal parameters, pushes it to the operation and maintenance terminal, and triggers the edge computing node to link with the corresponding distributed monitoring terminal to start the local alarm.

[0028] The real-time electrical variable parameters include the output voltage, output current, power, and frequency of the power supply to be monitored; the real-time battery parameters include the terminal voltage, charging and discharging current, remaining power, and cycle count of the built-in battery; and the environmental parameters include the temperature, humidity, and electromagnetic interference intensity around the power supply to be monitored.

[0029] In this embodiment, see Figure 2 In step S2, a multi-source data fusion calibration algorithm is introduced to fuse and calibrate the standardized real-time electrical variable parameters, real-time battery parameters, and environmental parameters. The specific process is as follows: S21: The standardized real-time electrical variable parameter is denoted as X, the real-time battery parameter is denoted as Y, and they are used as core monitoring parameters. The standardized environmental parameter is denoted as Z, and it is used as interference calibration parameter, including temperature T, humidity H, and electromagnetic interference intensity E. Among them, X, Y, and Z are dimensionless parameters obtained after min-max standardization, and their values ​​range from [0,1].

[0030] S22: Calculate the interference weights of environmental parameters on core monitoring parameters: calculate the interference weights of temperature, humidity, and electromagnetic interference intensity on electrical variable parameters and battery parameters respectively.

[0031] S23: Fusion calibration calculation of core monitoring parameters: Based on the comprehensive interference weight, adaptive calibration is performed on the standardized electrical variable parameters and battery parameters to eliminate environmental interference.

[0032] S24: Associate the calibrated electrical variable parameters and battery parameters with the standardized real-time location data to obtain the calibrated standardized monitoring data.

[0033] The specific process of step S22 is as follows: S221: Calculate the temperature disturbance weight, the specific formula is as follows: ; ; in, w T,X The single disturbance weight of temperature on the electrical variable parameter X. w T,Y The single disturbance weight of temperature on battery parameter Y is taken as [0,1]. k T Temperature interference coefficient; T The standardized real-time ambient temperature parameters; T 0 represents the standard operating temperature of the power supply to be monitored.

[0034] S222: Calculate the humidity interference weight, the specific formula is as follows: ; ; in, w H,X The single disturbance weight of humidity on the electrical variable parameter X. w H,Y The single disturbance weight of humidity on battery parameter Y is taken as [0,1]. k H H represents the humidity interference coefficient; H is the standardized real-time environmental humidity parameter. H 0 represents the standard operating humidity of the power supply to be monitored.

[0035] S223: Calculate the interference weight for electromagnetic interference intensity. The specific formula is as follows: ; ; in, w E,X The electromagnetic interference intensity is the single interference weight of the electrical variable parameter X. w E,Y The value of each parameter Y is a single interference weight of the electromagnetic interference intensity, and all values ​​are [0,1]. k E E is the electromagnetic interference intensity coefficient; E is the standardized real-time electromagnetic interference intensity parameter. E 0 represents the standard operating electromagnetic interference intensity of the power supply to be monitored.

[0036] S224: Calculate the comprehensive interference weight of the core monitoring parameters. The specific formula is as follows: w X =α· w T,X +β · w H,X +γ· w E,X ; w Y =α· w T,Y + β · w H,Y +γ· w E,Y ; in, w X The combined disturbance weights for electrical variable parameters, w Y The comprehensive interference weights for battery parameters are all set to [0,1]; α, β α and γ are the weighting coefficients for temperature, humidity, and electromagnetic interference intensity, respectively, satisfying α+ β +γ=1.

[0037] The specific process of step S23 is as follows: S231: Perform electrical variable parameter calibration, the specific formula is as follows: X cal = X ·(1- w X )+ X 0· w X ; in, X cal These are the standardized electrical variable parameters after calibration, dimensionless, and take values ​​[0,1]. X These are the standardized real-time electrical variable parameters; w X The comprehensive disturbance weights for electrical variable parameters; X 0 represents the standard reference value for electrical variable parameters; S232: Perform battery parameter calibration. The specific formula is as follows: Y cal = Y ·(1- w Y )+ Y 0· w Y ; in, Y cal These are the standardized battery parameters after calibration, dimensionless, and take values ​​[0,1]. Y These are the standardized real-time battery parameters; wY The overall interference weight for battery parameters; Y 0 represents the standard baseline value for battery parameters.

[0038] In another embodiment of this example, the specific process of constructing the distributed power source monitoring dataset by associating each calibrated standardized monitoring data with the corresponding terminal identifier and real-time location data in step S2 is as follows: S25: Extract the terminal identifier marked when each distributed monitoring terminal uploaded the initial monitoring data, and denote it as... ID i , i =1,2,..., n , n The total number of distributed monitoring terminals is matched with the standardized real-time location data collected in the preceding step S1 and after preprocessing, denoted as . L i Including latitude and longitude coordinates ( Lon i , Lat i ) and altitude Alt i Simultaneously, the standardized monitoring data after calibration in step S2 is extracted and recorded as follows: D cal,i , including X cal,i , Y cal,i ,make sure ID i , L i , D cal,i One-to-one correspondence.

[0039] S26: Correlation mapping between terminal identifier and calibrated monitoring data: A key-value pair mapping algorithm is used, with the terminal identifier as the key. ID i Using a unique primary key, establish and calibrate standardized monitoring data. D cal,i This one-to-one association ensures that the calibrated parameters of each terminal can be quickly retrieved through its unique identifier. The association mapping formula is as follows: Map ( ID i )= D cal,i ={ X cal,i1 , X cal,i2 ,..., X cal,ik , Y cal,i1 , Ycal,i2 ,..., Y cal,im}; in, Map (·) is a key-value pair mapping function that maps terminal identifiers to calibrated monitoring data; ID i For the first i The unique identifier of each distributed monitoring terminal is generated by combining the terminal hardware address and the ID of the power supply to be monitored, and the format is a string. D cal,i For the first i The set of calibrated and standardized monitoring data corresponding to each terminal; X cal,i1 ,..., X cal,ik For the first i Standardized electrical variable parameters after terminal calibration k Several dimensions, such as output voltage and output current; Y cal,i1 ,..., Y cal,im For the first i Standardized battery parameters after terminal calibration m Several dimensions, such as terminal voltage and remaining power.

[0040] S27: Secondary association between location data and associated data: based on terminal identifier ID i To ensure uniqueness, the "terminal identifier - calibrated monitoring data" pair obtained in step S26 is associated with the corresponding standardized real-time location data. L i A secondary association is performed to form three-dimensional associated data: "terminal identifier - location data - calibrated monitoring data". The association formula is as follows: Assoc ( ID i )={ L i , Map ( ID i )}={( Lon i , Lat i , Alt i ),{ X cal,i1 ,..., X cal,ik , Y cal,i1 ,..., Y cal,im}}; in, Assoc (·) represents the three-dimensional data association function; L i For the first i Standardized real-time location data corresponding to each terminal. Lon i For the standardized longitude parameters, Lat i For standardized latitude parameters, Alt i The elevation parameters are standardized and are dimensionless, with values ​​ranging from [0,1]. They are obtained from the original location data through min-max standardization.

[0041] S28: Construction and organization of distributed power source monitoring dataset: 3D correlation data of all terminals Assoc ( ID i The dataset is aggregated, a dataset normalization algorithm is introduced, and the data is sorted by terminal identifier, with monitoring timestamps added. t i (Consistent with the initial monitoring data acquisition time), construct a structured distributed power source monitoring dataset DS, with the following regularization formula: in, DS This is the final distributed power source monitoring dataset; For set aggregation operators, n The associated data from each terminal is aggregated into a unified dataset; t i For the first i The initial monitoring data collection timestamps for each terminal are in the format YYYY-MM-DDHH:MM:SS, accurate to the second.

[0042] Example 2 Based on Example 1, see Figure 3 The specific process of performing stratified determination on the calibrated and standardized monitoring data in the distributed power source monitoring dataset in step S3 is as follows: S31: Dataset Extraction and Matching: The cloud monitoring platform extracts data from the distributed power source monitoring dataset DS, categorized by terminal identifier. ID i Extract the corresponding standardized monitoring data after calibration one by one, i.e., the first i A set of calibrated electrical variable parameters of a power supply to be monitored X cal,i ={ X cal,i1 , X cal,i2 ,..., Xcal,ik}, k For electrical variable parameters, including output voltage, output current, etc., it is a set of calibrated battery parameters. Y cal,i ={ Y cal,i1 , Y cal,i2 ,..., Y cal,im}, m This includes battery parameters such as terminal voltage and remaining capacity, and is also matched with historical monitoring data corresponding to this terminal. DS his,i It includes parameters after calibration in the past 30 days and historical judgment results.

[0043] S32: Preset Threshold Initialization: Retrieve the set of initial preset electrical variable thresholds stored in the cloud monitoring platform. Th X ={ Th X1 ,Th X2 ,...,Th Xk Initial preset battery parameter threshold set Th Y ={Th Y1 , Th Y2 ,..., Th Ym}

[0044] Each parameter dimension corresponds to a set of upper and lower thresholds, i.e. Th Xj =[ Th Xj,min , Th Xj , max ], j=1,2,..., k , Th Yl =[ Th Yl ,min}, Th Y l [max] l =1,2,..., m The initial threshold is set based on the rated parameters of the power supply to be monitored and industry standards; S33: Working Status Layer Determination: Comparison One by One X cal,i Each dimension parameter and its corresponding Th X The threshold is used to determine the operating status of the power supply to be monitored. The specific determination formula is as follows: ; in, Status X,i For the first i The working status judgment result of the power supply to be monitored is set to {1,0,-1}, where 1 represents "normal", 0 represents "slight abnormality" and -1 represents "severe abnormality". : Logical product operator, used to synthesize a single decision result from all electrical variable parameters; δ (·) is the parameter determination function, specifically defined as: ; Where, Δ X j For the first j The slight abnormality threshold deviation of each electrical variable parameter is set to 0.05×( Th Xj,max - Th Xj,min This represents the permissible deviation of a parameter from the standard range that does not affect normal operation. X cal,ij : No. i The power supply to be monitored j calibrated standardized electrical variable parameters in each dimension; Th Xj,min , Th Xj,max : No. j Preset lower and upper threshold values ​​for individual electrical variable parameters; S34: Built-in battery status layering determination: Employs a comprehensive determination logic consistent with the working status determination, comparing one by one. Y cal,i Each dimension parameter and its corresponding Th Y The threshold is used to determine the status of the built-in battery. The specific formula for this determination is as follows: ; in, Status Y,i : No. i The status judgment result of the built-in battery of the power supply to be monitored is set to {1,0,-1}, where 1 represents "normal", 0 represents "slight abnormality" and -1 represents "severe abnormality". Y cal,il : No. i The power supply to be monitored has a built-in battery. l Standardized battery parameters after calibration in each dimension; Th Yl,min , Th Yl,max : No. lPreset lower and upper threshold values ​​for individual battery parameters; δ (·) is a single-parameter decision function, which is logically consistent with the electrical variable parameter decision function, only replacing the parameter with the battery parameter, where the slight abnormal threshold deviation Δ Y l =0.05×( Th Yl,max - Th Yl,,min ).

[0045] S35: Layered determination result output: Summarize the results for each power source to be monitored. Status X,i (Work status) Status Y,i (Built-in battery status), associated with the corresponding terminal identifier. ID i and real-time location data L i This generates associated data of terminal identifier, location, and dual-state determination results, providing a foundation for generating monitoring reports in step S4 and triggering early warnings in step S5. At the same time, the determination results are synchronized to the dynamic threshold self-learning module.

[0046] In this embodiment, the specific process of dynamically iteratively updating the preset electrical variable threshold and the preset battery parameter threshold in step S3 to achieve adaptive optimization of the threshold is as follows: S36: Self-learning input data preparation: The dynamic threshold self-learning module receives three parts of data, namely: the set of judgment results output from the current S3 step. Statu si ={ Status X,i ,Status Y,i}, i=1,2,...,n, the calibrated parameters in the current distributed power source monitoring dataset DS. X cal,i , Y cal,i Historical monitoring dataset It includes calibration parameters and corresponding judgment results for the past 30 days.

[0047] S37: Historical Valid Data Screening: Employing an outlier removal algorithm, filtering historical monitoring data that has been determined to be "normal" (… Status X,i =1 and Status Y,i After calibration, parameters with a result of 0 or -1 are discarded to avoid abnormal data affecting the threshold optimization accuracy. The selection formula is as follows: DS his,valid ={(X cal,his,ij ,Y cal,his,il )| Status his,i ={1,1}}; in, DS his,valid This is the filtered set of historical valid calibration data; X cal,his,ij , Y cal,his,il : respectively the first i The first in the historical monitoring of the power supply to be monitored j The first electrical variable parameter, the first l The calibrated and standardized values ​​of each battery parameter; Status his,i : No. i The dual-state judgment result of the historical monitoring of the power supply to be monitored is given. A value of {1,1} indicates that the historical data is valid.

[0048] S38: Dynamic Threshold Iterative Update Calculation: A weighted moving average iterative algorithm is used, combining historical valid data with current calibration data, to dynamically update the preset electrical variable threshold. Th X Preset battery parameter thresholds Th Y To ensure that the threshold value closely matches the actual operating state of the power supply being monitored, the specific update formula is as follows: The formula for updating the threshold of electrical variable parameters is as follows: ; The formula for updating battery parameter thresholds is as follows: in, Th Xj,new 、Th Yl,new Each is the updated version of the first. j The first electrical variable parameter, the first l Preset thresholds for each battery parameter, including upper and lower limits, with "+" corresponding to the upper limit and "-" corresponding to the lower limit; w Threshold update weight coefficient, with a value range of 0.7~0.9 and a default of 0.8, representing the weight distribution between historical data and current data; N his : No. j The first electrical variable parameter / the first l The number of historical valid data points for each battery parameter (default) N his ≥100, to ensure statistical validity; X cal,his,jt The t-th historical valid calibration data for the j-th electrical variable parameter;Y cal,his,lt : The t-th historical valid calibration data for the l-th battery parameter; This is the threshold correction coefficient, ranging from 1.2 to 1.5, adjusted based on the 3σ principle of normal distribution to ensure that the threshold covers more than 95% of normal data; σ Xj 、σ Yl : No. j The first electrical variable parameter, the first l The standard deviation of historical valid data for each battery parameter is used to characterize the degree of data dispersion, and the formula is: ; This is the average of historical valid data.

[0049] S39: Threshold Optimization Verification and Implementation: Verify the deviation between the updated threshold and the initial threshold. If the deviation is within 5%, it will take effect directly and replace the original preset threshold. If the deviation exceeds 5%, it will be corrected in conjunction with the rated parameters of the power supply to be monitored, and will take effect after correction. The optimized threshold after taking effect will be synchronized to the edge computing node for the result update of the subsequent S5 step and the hierarchical determination of the next round of monitoring data, so as to realize the cyclic adaptive optimization of the threshold.

[0050] In the above process, the hierarchical judgment is based on the distributed power source monitoring dataset constructed in the preceding S2, and the dynamic threshold self-learning is based on the current hierarchical judgment result and historical monitoring data. The two major links are executed in conjunction to ensure the accuracy of the current state judgment and to optimize the subsequent judgment threshold through self-learning, thereby improving the adaptability and long-term monitoring accuracy of the entire monitoring system.

[0051] Example 3 Based on Example 1 or Example 2, the specific process of step S4 is as follows: S41: Perform multi-dimensional data association to achieve precise binding of location, status, and threshold: Related data extraction: The cloud monitoring platform extracts data from the distributed power source monitoring dataset DS, categorized by terminal identifier. ID i , i =1,2,..., n , n For the total number of distributed monitoring terminals, extract the corresponding data for each terminal, including: standardized real-time location data. L i =( Lon i ,Lat i ,Alt i ), calibrated standardized electrical variable parametersX cal,i Standardized battery parameters after calibration Y cal,i Simultaneously extract the first i Dual-state determination results of the power supply to be monitored Status i ={ Status X,i ,Status Y,i The optimized threshold output by the dynamic threshold self-learning module. Th i,new ={ Th X,new,i ,Th Y,new,i Match the terminal identifier to ensure that the threshold of a single power supply matches its own parameters.

[0052] Multi-dimensional data association operations: Employing structured data association algorithms, based on terminal identifiers. ID i Using a unique primary key, location data, status determination results, optimized thresholds, and calibrated monitoring data are bound together in four dimensions to form a complete associated dataset for a single power supply to be monitored. Data i The related formula is as follows: Data i ={ ID i ,L i ,Status i ,Th i,new ,D cal,i}; The expansion is as follows: Data i ={ ID i ,( Lon i , Lat i , Alt i ),( Status X,i , Status Y,i ),( Th X,new,i ,Th Y,new,i ),( X cal,i, Y cal,i )}; in,Data i : No. i A complete four-dimensional correlation dataset of a power source to be monitored; ID i : No. i The unique identifier of each distributed monitoring terminal adopts the format of terminal MAC address-power supply ID to be monitored; L i : No. i Standardized real-time location data of a power source to be monitored. Lon i Standardized longitude; [[ID= i Standardized latitude; ​ i Standardized altitude, all are dimensionless parameters in the [0,1] interval; ​ i : No. i A set of dual-state determination results for a power supply to be monitored. ​ X,i : Working status, with values ​​{1, 0, -1}; ​ Y,i : Built-in battery status, with values ​​{1, 0, -1}; ​ X,new,i : Optimized threshold values ​​for electrical variable parameters, including upper and lower limits for each dimension). ​ Y,new,i Optimized battery parameter thresholds, including upper and lower limits for each dimension; D cal,i : No. i A standardized monitoring data set of a power supply under test after calibration, including calibrated electrical variable parameters. X cal,i, and calibrated battery parameters Y cal,i .

[0053] S42: Generate distributed monitoring reports: Basic information integration for reports: For those that have passed verification ​ i Extract and organize basic information, including: terminal identifier. ​ i The model of the power supply to be monitored (associated) ​ (i retrieves from cloud database) and monitors timestamps t i (with S2 step dataset) ​ (Timestamps are consistent), real-time location raw data (standardized) L i (Inverse standardization is converted to actual latitude, longitude, and altitude).

[0054] Inverse standardization formula: L real = L std ×( L max - L min )+ L min ; in, L real These are the original values ​​for the location parameters: longitude / latitude: °, altitude: m; L std For standardized position parameters ( ​ i or ​ i or ​ i ); L max This is the maximum actual value for this location parameter, such as longitude: 180°, latitude: 90°, altitude: set according to the monitoring area, default 0~5000m; L min This represents the minimum actual value of the location parameter, such as longitude: -180°, latitude: -90°, altitude: 0m.

[0055] Comparative analysis of monitoring parameters and thresholds: ​ i Standardized monitoring data after calibration D cal,i , and the optimized threshold ​ i,new The comparison is performed, and the deviation rate of each parameter from the threshold is calculated for anomaly analysis in the report. The formula for calculating the deviation rate is as follows: Electrical variable parameter deviation rate: ; Battery parameter deviation rate: ; in, ​ X,ij Let be the deviation rate of the j-th electrical variable parameter of the i-th power supply to be monitored. ​ Y,il : No. i The power supply to be monitored has a built-in battery. l Deviation rate (%) of each parameter; X cal,ij : No. i The power supply to be monitored j One calibrated electrical variable parameter; Y cal,il : No. iThe power supply to be monitored has a built-in battery. l One calibrated battery parameter; ​ X,new,ij,avg : No. i The power supply to be monitored j The average value of the threshold after optimization of individual electrical variable parameters: ​ X,new,ij,avg =( ​ X,new,ij,min ​ X,new,ij,max ) / 2; S43: Structured Report Generation: Generate a distributed monitoring report for a single power source under monitoring, according to power industry monitoring report standards. ​ i The report contains 5 core modules, and the generation formula is as follows: ​ i ={ ​ i ​ i ​ i ​ i ​ i}; Specific content of each module: ​ i Basic information module, including ​ i Power supply model, monitoring time t i Actual location data L real,i ; ​ i Monitoring parameter module, including X cal,i 、Y cal,i The original acquired values ​​(after inverse normalization) and the calibrated values; ​ i Status determination module, including ​ X,i , ​ Y,i And the criteria for judgment (corresponding parameter deviation rate); ​ i Threshold module, including optimized thresholds ​ i,new Initial threshold, threshold update magnitude; ​i Analysis and suggestion module, based on deviation rate ​ X,ij , ​ Y,il Based on the status assessment results, maintenance suggestions are generated. For example, for minor anomalies, regular inspections are recommended; for serious anomalies, immediate shutdown and repair are recommended. S44: Distributed Power Source Monitoring Topology Map Construction: Based on location data, a visual topology map is constructed to provide an intuitive display of the status of multiple power sources.

[0056] The specific process of step S44 is as follows: S441: Location Data Coordinate Transformation: Transform the actual location data of all power sources to be monitored. L real,i =( ​ real,i , ​ real,i ​ real,i ), converted to Cartesian coordinates required for topological graph drawing ( X map,i ,Y map,i The Gauss-Kruger projection transformation algorithm is used, and the transformation formula is as follows: ; in,( X map,i ,Y map,i ): No. i The Cartesian coordinates of the power source to be monitored in the topology diagram; ​ real,i , ​ real,i : respectively the first i The actual longitude and latitude of the power source to be monitored; 6378137m: Earth's equatorial radius; S442: Topology graph node drawing: using Cartesian coordinates ( X map,i ,Y map,i The topology nodes of each power source to be monitored are drawn, with different colors and shapes of nodes used to distinguish between the working state and the built-in battery state.

[0057] Working status is normal ( ​ X,i =1) and the battery status is normal ( ​ Y,i =1): Green circular node; Any state is slightly abnormal (0): yellow square node; Any state that is severely abnormal (-1): Red triangle node; S443: Topology Graph Association Information Labeling: Label the terminal identifier next to each topology node. ​ i The core monitoring parameters (output voltage, remaining power) and status determination results are labeled using a node offset labeling algorithm to ensure that the labels do not overlap. The label offset formula is as follows: ; in,( X label,i ,Y label,i ): No. i The coordinates of the node's labeled information; : Annotate offset distance, default value is 50, unit: pixels, adaptively adjusts according to the size of the topology map; Offset angle, which is the direction of the angle bisector of the angle between the node and the line connecting the node to the adjacent node.

[0058] In another embodiment, the specific process of step S5 is as follows: The cloud-based monitoring platform establishes a real-time data receiving channel to receive updated, standardized monitoring data synchronized from edge computing nodes. This updated, standardized monitoring data is generated from initial monitoring data collected in real-time by edge computing nodes based on distributed monitoring terminals, after preprocessing including data denoising, outlier removal, and standardization, and is synchronously tagged with the corresponding terminal identifier. ​ i and data update timestamp t update,i .

[0059] Following the S3 hierarchical judgment logic, the working status and built-in battery status judgment results are updated based on updated data and optimized thresholds: Update data to match thresholds: Update data ​ update,i According to terminal identification ​ i Optimized threshold matching S3 output ​ i,new ={ ​ X,new,i ​ Y,new,i If the threshold is updated in a new iteration, the latest optimized threshold is used, and the previous state determination result of the corresponding terminal is retrieved. ​ last,i ={ ​ X,last,i ​ Y,last,i}, used for subsequent state change analysis.

[0060] Working status determination result update: Using the same determination logic as S3, the updated standardized electrical variable parameters are updated. X update,i ={ X update,i1 ,X update,i2 ,...,X update,ik}, and the corresponding optimized electrical variable threshold ​ X,new,i Compare each item individually and update the work status assessment results.

[0061] Built-in battery status determination result update: Using the same determination logic as S3, the updated standardized battery parameters will be updated. Y update,i ={ Y update,i1 ,Y update,i2 ,...,Y update,im} and the corresponding optimized battery parameter thresholds ​ Y,new,i Compare each one individually and update the built-in battery status determination results.

[0062] The updated judgment results are associated with and stored along with the terminal identifier, updated location data, and update timestamp. At the same time, the distributed monitoring report and monitoring topology map are updated to ensure that the status information of the report and the topology map are synchronized in real time.

[0063] Preset warning condition initialization: Retrieve the preset warning condition set W={W1,W2} stored in the cloud monitoring platform. 2, W3}, the specific warning conditions are as follows: W1: Single-state anomaly with continuous warning condition: Updated working status ​ X,update,i =0 (minor anomaly) or -1 (serious anomaly), or the updated battery status. ​ Y,update,i =0 or -1, and the duration of this abnormal state is... t duration,i Warning threshold for duration of abnormal state greater than or equal to t warn .

[0064] W2: Low battery warning condition: Battery status after update ​ Y,update,i =0 or -1, and the updated standardized battery remaining capacity parameter Y update,ires Less than or equal to the remaining battery power warning threshold Y warn .

[0065] W3: Parameter mutation anomaly warning condition: The mutation magnitude Δ of the updated electrical variable parameter or battery parameter compared to the previous updated parameter. P i Greater than or equal to the warning threshold Δ for mutation magnitude P warn， And the updated status determination result is abnormal (0 or -1).

[0066] Early warning condition determination calculation: for each terminal ​ i Each updated state and parameter is checked to determine whether it meets the above warning conditions. The comprehensive judgment formula is as follows: ; in, ​ i : No. i The warning judgment result of the power supply to be monitored is as follows: a value of 1 indicates that the warning condition is met (warning is triggered), and a value of 0 indicates that it is not met (warning is not triggered); ∨: logical OR operator, a warning is triggered if any warning condition is met.

[0067] Warning level classification: targeting ​ i For terminals with a value of 1 (triggered an alert), the alert level is determined based on the type of alert condition met, according to the following rules: Level 1 warning (priority handling): Meets W3 (parameter mutation anomaly), or simultaneously meets W1 and W2; Level 2 warning (routine handling): Only meets W1 (single state abnormality continues); Level 3 warning (reminder): Only meets W2 (low battery abnormality).

Claims

1. A power monitoring method based on an IoT distributed architecture, characterized in that, The method is based on an IoT distributed architecture and includes several distributed monitoring terminals, edge computing nodes, and a cloud monitoring platform. The method includes the following steps: S1: Collect real-time location data, real-time electrical variable parameters, surrounding environmental parameters, and real-time battery parameters of the power supply to be monitored, and package all data into initial monitoring data; S2: Preprocess the initial monitoring data to obtain standardized monitoring data; fuse and calibrate other standardized monitoring data (excluding environmental parameters) based on environmental parameters to obtain calibrated standardized monitoring data; associate each calibrated standardized monitoring data with the corresponding terminal identifier and real-time location data to construct a distributed power source monitoring dataset; S3: Based on preset electrical variable thresholds and preset battery parameter thresholds, the distributed power source monitoring dataset is hierarchically judged to obtain the working status judgment results of the power source to be monitored and the built-in battery status judgment results, and the preset electrical variable thresholds and preset battery parameter thresholds are dynamically iterated and updated to achieve adaptive optimization of the thresholds; S4: Associate the real-time location data, working status determination results and built-in battery status determination results of the power supply to be monitored, and combine them with the optimized threshold output by the dynamic threshold self-learning module to generate a distributed monitoring report for each power supply to be monitored. At the same time, construct a distributed power supply monitoring topology map based on the location data of each power supply to be monitored. S5: Based on the calibrated standardized monitoring data and the dynamically updated preset thresholds, update the working status judgment results of the power supply to be monitored and the built-in battery status judgment results; when the judgment results meet the preset warning conditions, generate warning information including the real-time location data of the power supply to be monitored, the abnormal status type, and the abnormal parameters.

2. The power monitoring method based on an IoT distributed architecture according to claim 1, characterized in that, The specific process of fusing and calibrating other standardized monitoring data besides environmental parameters based on environmental parameters in step S2 to obtain calibrated standardized monitoring data is as follows: S21: Record the standardized real-time electrical variable parameters as X, the real-time battery parameters as Y, as the core monitoring parameters, and the standardized environmental parameters as Z, as the interference calibration parameters, including temperature T, humidity H, and electromagnetic interference intensity E. S22: Calculate the interference weights of temperature, humidity, and electromagnetic interference intensity on electrical variable parameters and battery parameters, respectively; S23: Adaptive calibration of standardized electrical variable parameters and battery parameters based on comprehensive interference weights; S24: Associate the calibrated electrical variable parameters and battery parameters with the standardized real-time location data to obtain the calibrated standardized monitoring data.

3. The power monitoring method based on an IoT distributed architecture according to claim 2, characterized in that, The specific process of step S22 is as follows: S221: Calculate the temperature disturbance weight, the specific formula is as follows: ； ； in, w T,X The single disturbance weight of temperature on the electrical variable parameter X. w T,Y The single disturbance weight of temperature on battery parameter Y; k T This refers to the temperature interference coefficient. T The standardized real-time ambient temperature parameters; T 0 represents the standard operating temperature of the power supply to be monitored; S222: Calculate the humidity interference weight, the specific formula is as follows: ； ； in, w H,X The single disturbance weight of humidity on the electrical variable parameter X. w H,Y The single disturbance weight of humidity on battery parameter Y; k H H represents the humidity interference coefficient; H is the standardized real-time environmental humidity parameter. H 0 represents the standard operating humidity of the power supply to be monitored; S223: Calculate the interference weight for electromagnetic interference intensity. The specific formula is as follows: ； ； in, w E,X The single interference weight of electromagnetic interference intensity on the electrical variable parameter X. w E,Y The single interference weight of electromagnetic interference intensity on battery parameter Y; k E E is the electromagnetic interference intensity coefficient; E is the standardized real-time electromagnetic interference intensity parameter. E 0 represents the standard operating electromagnetic interference intensity of the power supply to be monitored; S224: Calculate the comprehensive interference weight of the core monitoring parameters. The specific formula is as follows: w X =a· w T,X + β · w H,X +g· w E,X ; w Y =a· w T,Y + β · w H,Y +g· w E,Y ; in, w X The combined disturbance weights for electrical variable parameters, w Y The overall interference weights for battery parameters; α, β γ and γ are the weighting coefficients for temperature, humidity, and electromagnetic interference intensity, respectively.

4. The power monitoring method based on an IoT distributed architecture according to claim 3, characterized in that, The specific process of step S23 is as follows: S231: Perform electrical variable parameter calibration, the specific formula is as follows: X cal = X ·(1- w X )+ X 0· w X ； in, X cal These are the standardized electrical variable parameters after calibration. X These are the standardized real-time electrical variable parameters; w X The comprehensive disturbance weights for electrical variable parameters; X 0 represents the standard reference value for electrical variable parameters; S232: Perform battery parameter calibration. The specific formula is as follows: Y cal = Y ·(1- w Y )+ Y 0· w Y ； in, Y cal Standardized battery parameters after calibration; Y These are the standardized real-time battery parameters; w Y The overall interference weight for battery parameters; Y 0 represents the standard baseline value for battery parameters.

5. The power monitoring method based on an IoT distributed architecture according to claim 3, characterized in that, The specific process of constructing the distributed power source monitoring dataset by associating each calibrated and standardized monitoring data with the corresponding terminal identifier and real-time location data in step S2 is as follows: S25: Extract the terminal identifier marked when each distributed monitoring terminal uploaded the initial monitoring data, and denote it as... ID i The standardized real-time location data after preprocessing is denoted as... L i Including latitude and longitude coordinates ( Lon i , Lat i ) and altitude Alt i Extract the standardized monitoring data after calibration and record it as follows: D cal,i ; S26: Employ a key-value pair mapping algorithm, using terminal identifiers. ID i Using a unique primary key, establish and calibrate standardized monitoring data. D cal,i A one-to-one association; S27: Based on terminal identifier ID i The uniqueness of the terminal identifier-calibrated monitoring data pair is determined by linking it with the corresponding standardized real-time location data. L i A secondary correlation is performed to form a three-dimensional correlated data system consisting of terminal identifier, location data, and calibrated monitoring data. Assoc ( ID i ); S28: Associate the three-dimensional data of all terminals Assoc ( ID i Summarize the data, sort by terminal identifier, and supplement with monitoring timestamps. t i We constructed a structured distributed power monitoring dataset DS.

6. The power monitoring method based on an IoT distributed architecture according to claim 5, characterized in that, The specific process for layer determination in step S3 is as follows: S31: From the distributed power source monitoring dataset DS, sorted by terminal identifier ID i Extract the corresponding calibrated and standardized monitoring data one by one, and match them with the historical monitoring data of the corresponding terminals. DS his,i This includes parameters calibrated in the past 30 days and historical judgment results; S32: Retrieve the set of initial preset electrical variable thresholds stored in the cloud monitoring platform. Th X ={ Th X1 ,Th X2 ,...,Th Xk Initial preset battery parameter threshold set Th Y ={ Th Y1 , Th Y2 ,..., Th Ym }; S33: Working Status Layer Determination: Comparison One by One X cal,i Each dimension parameter and its corresponding Th X Thresholds are used to determine the operating status of the power supply to be monitored. S34: Built-in battery status layering determination: comparison one by one Y cal,i Each dimension parameter and its corresponding Th Y Threshold to determine the status of the built-in battery; S35: Layered determination result output: Summarize the operating status of each monitored power supply. Status X,i Built-in battery status Status Y,i Associate with the corresponding terminal identifier ID i and real-time location data L i This forms the associated data of terminal identifier, location, and dual-state determination results.

7. The power monitoring method based on an IoT distributed architecture according to claim 6, characterized in that, The specific process of dynamically iteratively updating the preset electrical variable threshold and the preset battery parameter threshold in step S3 to achieve adaptive optimization of the threshold is as follows: S36: Obtain the set of judgment results Statu si ={ Status X,i Status Y,i }, Calibrated parameters in the current distributed power source monitoring dataset DS X cal,i , Y cal,i Historical monitoring dataset DS his This includes calibration parameters and corresponding judgment results for the past 30 days. S37: Employ an outlier removal algorithm to filter calibrated parameters whose results are normal from historical monitoring data and remove parameters whose results are abnormal. S38: Employs a weighted moving average iterative algorithm, combining historical valid data with current calibration data, to dynamically update the preset electrical variable threshold. Th X Preset battery parameter thresholds Th Y ; S39: Verify the deviation between the updated threshold and the initial threshold. If the deviation is within 5%, it will take effect directly and replace the original preset threshold. If the deviation exceeds 5%, it shall be corrected in conjunction with the rated parameters of the power supply to be monitored, and the correction shall take effect. Once the optimization threshold takes effect, it will be synchronized to the edge computing nodes.

8. The power monitoring method based on an IoT distributed architecture according to claim 1, characterized in that, The specific process of step S4 is as follows: S41: From the distributed power source monitoring dataset DS, sorted by terminal identifier ID i Extract standardized real-time location data one by one L i Standardized electrical variable parameters after calibration X cal,i Standardized battery parameters after calibration Y cal,i Extract the first i Dual-state determination results of the power supply to be monitored Status i The optimized threshold output by the dynamic threshold self-learning module Th i,new ; by terminal identifier ID i Using a unique primary key, location data, status determination results, optimized thresholds, and calibrated monitoring data are bound together in four dimensions to form a complete associated dataset for a single power supply to be monitored. Data i ; S42: For those that pass the verification Data i Extract and organize basic information, including terminal identifiers. ID i Power supply model to be monitored, monitoring timestamp t i Real-time location raw data; Data i Standardized monitoring data after calibration D cal,i , and the optimized threshold Th i,new Compare the parameters and calculate the deviation rate between each parameter and the threshold. S43: Generate a distributed monitoring report for a single monitored power supply based on the deviation rate of each parameter from the threshold. Report i It includes a basic information module, a monitoring parameter module, a status determination module, a threshold module, and an analysis and suggestion module; S44: Based on location data, construct a visual topology map to intuitively display the status of multiple power sources.

9. A power monitoring method based on an IoT distributed architecture according to claim 8, characterized in that, The specific process of step S44 is as follows: S441: Data on the actual location of all power sources to be monitored L real,i Convert to Cartesian coordinates required for topological graph drawing ( X map,i ,Y map,i ); S442: In plane rectangular coordinates ( X map,i ,Y map,i ) is the node location. Draw the topology nodes of each power supply to be monitored, and use different colors and shapes of nodes to distinguish the working status and the built-in battery status. S443: Label the terminal identifier next to each topology node. ID i The core monitoring parameters and status determination results are labeled using a node offset labeling algorithm to ensure that the labels do not overlap.

Images