Power adapter intelligent monitoring method and system based on Internet of Things

Abstract

The invention relates to the technical field of power adapter monitoring, and discloses a power adapter intelligent monitoring method and system based on the Internet of Things, and the method comprises the steps: collecting multi-dimensional operation data of a power adapter under an operation condition, and recognizing a target monitoring region of the power adapter; performing adaptive spectral decomposition on the multi-dimensional operation data to obtain target analysis data, and positioning abnormal nodes in the target monitoring area to obtain an abnormal circuit area; acquiring infrared thermal imaging data and electromagnetic radiation data of the power adapter, analyzing thermal stress distribution characteristics of the power adapter, and performing frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum; and performing three-dimensional topology analysis on the abnormal circuit region to obtain a fault form feature vector, and performing health state analysis on the power adapter to obtain an equipment health analysis report. According to the invention, the accuracy of detecting the abnormal problem of the power adapter in a complex environment can be improved.

Description

Technical Field

[0001] The present invention relates to an intelligent monitoring method and system for a power adapter based on the Internet of Things, and belongs to the technical field of power adapter monitoring. Background Art

[0002] In the context of the booming development of the modern Internet of Things, as a key component for stable power supply of various electronic devices, the effective monitoring of the operating state of the power adapter is of great significance. From ensuring the continuous normal operation of various smart home appliances in the smart home system to ensuring the stable power support of production equipment in the industrial Internet of Things environment, the reliable operation of the power adapter is the cornerstone of the stable operation of the entire Internet of Things system, directly related to the normal order and efficiency of production and life. Timely and accurately grasping the operating state of the power adapter can detect potential faults in advance, prevent equipment downtime, and ensure the uninterrupted operation of the system.

[0003] Currently, for the intelligent monitoring of power adapters, most rely on traditional basic monitoring means. Traditional power adapter monitoring systems mostly adopt a combination of threshold alarm and regular manual inspection, and determine the device state through static thresholds of basic parameters such as voltage / current. However, with the high-frequency operation of new power electronic devices and the complexity of the topology structure, equipment failures present complex characteristics such as multi-physical field coupling (electrical-thermal-mechanical stress interaction), dynamic load mutation (millisecond-level power fluctuation), and hidden degradation accumulation (gradual change of the capacitance value of electrolytic capacitors), making static threshold determination often inaccurate (that is, there are abnormal circuit points but problems cannot be found), resulting in the inability to accurately identify problems existing in the power adapter. Summary of the Invention

[0004] The present invention provides an intelligent monitoring method and system for a power adapter based on the Internet of Things, and its main purpose is to improve the accuracy of detecting abnormal problems of the power adapter in a complex environment.

[0005] To achieve the above object, the intelligent monitoring method for a power adapter based on the Internet of Things provided by the present invention includes: Deploying Internet of Things devices at key nodes of the power adapter to collect multi-dimensional operation data under the operating conditions of the power adapter by using the Internet of Things devices, calculating the parameter deviation degree of the multi-dimensional operation data, extracting key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, and using the key parameter points to identify the target monitoring area of the power adapter; Based on the target monitoring area, performing adaptive spectral decomposition on the multi-dimensional operation data to obtain target analysis data, using the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter, and using the multi-dimensional electrical characteristics to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area; Collect the infrared thermal imaging data and electromagnetic radiation data of the power adapter, perform non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, identify the temperature gradient characteristics of the calibrated thermal distribution map, analyze the thermal stress distribution characteristics of the power adapter using the temperature gradient characteristics, and perform frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum; Perform three-dimensional topology analysis on the abnormal circuit area to obtain a fault form feature vector, fuse the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a fused diagnosis data set, and use the fused diagnosis data set to analyze the health status of the power adapter to obtain an equipment health analysis report.

[0006] Optionally, calculating the parameter deviation degree of the multi-dimensional operation data includes: Identify the parameter measurement type of the multi-dimensional operation data; Classify the multi-dimensional operation data according to the parameter measurement type to obtain classified data; Query the measured value of the classified data and the observation time corresponding to the measured value; Based on the measured value, calculate the observation value deviation degree of the classified data within the observation time using the following formula: ; Wherein, represents the observation value deviation degree, t represents the observation time, represents the measured value of the i-th classified data at time t, represents the corresponding reference value, represents the deviation weight coefficient of the i-th classified data; Determine the parameter deviation degree of the multi-dimensional operation data according to the observation value deviation degree.

[0007] Optionally, using the key parameter points to identify the target monitoring area of the power adapter includes: Perform three-dimensional coordinate system mapping on the key parameter points to obtain a standardized coordinate data set; Perform time synchronization interpolation operation on the standardized coordinate data set to obtain a spatio-temporal alignment data set; Perform data clustering on the spatio-temporal alignment data set to obtain an initial monitoring cluster set; Perform spatio-temporal correlation verification on the initial monitoring cluster set to extract the effective data set in the initial monitoring cluster set using the verification result of the spatio-temporal correlation verification; Based on the effective data set, perform dynamic risk analysis on the power adapter to identify the priority monitoring area of the power adapter; Perform regional boundary processing on the priority monitoring area to obtain the target monitoring area.

[0008] Optionally, based on the target monitoring area, perform adaptive spectral decomposition on the multi-dimensional operation data to obtain target analysis data, including: Identify the regional characteristics of the target monitoring area; Based on the regional characteristics, perform data screening on the multi-dimensional operation data to obtain screened data; Calculate the signal Shannon entropy of the screened data; Based on the signal Shannon entropy, perform dynamic layer decomposition on the screened data to obtain a time-frequency matrix; Perform time-frequency sub-generation optimization on the time-frequency matrix to obtain target analysis data.

[0009] Optionally, use the multi-dimensional electrical characteristics to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area, including: Perform feature screening on the multi-dimensional electrical characteristics to obtain screened features; Use the screened features to perform anomaly detection on the target monitoring area to obtain anomaly detection data; Perform matrix transformation on the anomaly detection data to obtain an anomaly detection matrix; Use the anomaly detection matrix to identify the abnormal range of the target monitoring area to obtain a preliminary identification area; Calculate the anomaly similarity of the electrical data in the preliminary identification area; Based on the anomaly similarity, perform regional segmentation on the target monitoring area to obtain an abnormal segmentation area; Identify the regional center point of the abnormal segmentation area; Use the regional center point to perform anomaly localization on the target monitoring area to obtain an abnormal circuit area.

[0010] Optionally, perform non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, including: Perform bad pixel detection on the infrared thermal imaging data to obtain bad pixel coordinates; Perform matrix transformation on the bad pixel coordinates to obtain a bad pixel coordinate matrix; Perform neighborhood interpolation on the bad pixel coordinate matrix to obtain an interpolated image; Use the following formula to calculate the gain matrix of the interpolated image: ; Among them, represents the gain matrix, represents the pixel of the interpolated image the original gray value at the high calibration temperature, represents the pixel of the interpolated image the original gray value at the low calibration temperature, represents the high calibration temperature, represents the low calibration temperature; Based on the gain matrix, use the following formula to calculate the offset matrix of the interpolated image: ; Among them, represents the offset matrix, represents the pixel of the interpolated image the original gray value at the low calibration temperature, represents the low calibration temperature; Based on the gain matrix and the offset matrix, perform linear correction on the infrared thermal imaging data to obtain a linearly corrected image; Perform scene adaptive correction on the linearly corrected image to obtain a calibrated thermal distribution map.

[0011] Optionally, the recognition of the temperature gradient feature of the calibrated thermal distribution map includes: Perform an effective area masking operation on the thermal distribution map to obtain a masked image; Perform morphological closing operation on the masked image to obtain an effective image Calculate the gradient matrix of the effective image; Perform polar coordinate transformation on the gradient matrix to obtain a gradient amplitude matrix; Perform adaptive threshold segmentation on the gradient amplitude matrix to obtain a binary image; Use the binary image to recognize the temperature gradient feature of the calibrated thermal distribution map.

[0012] Optionally, the three-dimensional topological analysis of the abnormal circuit area to obtain a fault morphology feature vector includes: Perform three-dimensional finite element modeling on the abnormal circuit area to obtain a visualized circuit area; Identify the current density distribution data, thermal stress data, and electromagnetic field intensity distribution data in the visualized circuit area; Perform vectorization processing on the current density distribution data, the thermal stress data, and the electromagnetic field intensity distribution data to obtain a fault morphology feature vector.

[0013] Optionally, the multi-source data fusion of the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a fusion diagnosis data set includes: Perform matrix transformation on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a first matrix, a second matrix, and a third matrix; Perform tensor combination on the first matrix, the second matrix, and the third matrix to obtain a high-order tensor; Perform multilinear decomposition on the high-order tensor to obtain a low-rank matrix; Perform matrix solution on the low-rank matrix to obtain matrix elements; Perform element fusion on the matrix elements to obtain a fusion diagnosis data set.

[0014] To solve the above problems, the present invention also provides an intelligent monitoring system for a power adapter based on the Internet of Things. The system includes: A target area recognition module for deploying Internet of Things devices at key nodes of the power adapter to collect multi-dimensional operation data under the operating conditions of the power adapter by using the Internet of Things devices, calculate the parameter deviation degree of the multi-dimensional operation data, extract key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, and use the key parameter points to identify the target monitoring area of the power adapter; An abnormal area recognition module for performing adaptive spectral decomposition on the multi-dimensional operation data based on the target monitoring area to obtain target analysis data, using the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter, and using the multi-dimensional electrical characteristics to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area; A multi-dimensional data processing module for collecting infrared thermal imaging data and electromagnetic radiation data of the power adapter, performing non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, identifying the temperature gradient characteristics of the calibrated thermal distribution map, analyzing the thermal stress distribution characteristics of the power adapter by using the temperature gradient characteristics, and performing frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum; A state analysis module for performing three-dimensional topology analysis on the abnormal circuit area to obtain a fault form feature vector, performing multi-source data fusion on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a fusion diagnosis data set, and using the fusion diagnosis data set to perform a health state analysis on the power adapter to obtain a device health analysis report.

[0015] Compared with the problems described in the background art, in the embodiments of the present invention, Internet of Things devices are deployed at key nodes of the power adapter, integrating multiple sensor modules such as temperature, humidity, and vibration, to collect multi-dimensional operation data under operating conditions in real time, laying a data foundation for subsequent analysis, and calculating the parameter deviation degree of the multi-dimensional operation data to quickly identify abnormal parameters, so as to screen out the points with a deviation degree higher than the preset threshold, focus on the target monitoring area, and improve the analysis efficiency; further, the present invention realizes the positioning of the abnormal circuit area in the target monitoring area by identifying the multi-dimensional electrical characteristics of the target monitoring area, facilitating the accurate positioning of the abnormal circuit area and facilitating the key monitoring of potential hazards; further, the present invention collects the infrared thermal imaging data and electromagnetic radiation data of the power adapter to understand the surface temperature and electromagnetic energy emission situation, and deeply understand the thermo-mechanical performance and working stability of the power adapter; even further, the present invention performs three-dimensional topological analysis on the abnormal circuit area, generates a visualized circuit area by means of finite element modeling, and then obtains a fusion diagnosis data set of the visualized circuit area to evaluate the health status of the power adapter from multiple dimensions such as thermal performance, electromagnetic compatibility, and circuit faults. Therefore, the present invention can improve the accuracy of detecting abnormal problems of the power adapter in a complex environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic flow chart of a method for intelligent monitoring of a power adapter based on the Internet of Things provided by an embodiment of the present invention; Figure 2 is a schematic diagram of modules of a system for realizing the intelligent monitoring of a power adapter based on the Internet of Things provided by an embodiment of the present invention.

[0017] The implementation, functional features, and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] The embodiments of the present application provide a method for intelligent monitoring of a power adapter based on the Internet of Things. The execution subject of the method for intelligent monitoring of a power adapter based on the Internet of Things includes but is not limited to at least one of electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiments of the present application. In other words, the method for intelligent monitoring of a power adapter based on the Internet of Things can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server, or a cloud server cluster, etc.

[0020] Embodiment 1: Refer to Figure 1As shown in the figure, it is a schematic flowchart of an intelligent monitoring method for a power adapter based on the Internet of Things provided by an embodiment of the present invention. In this embodiment, the intelligent monitoring method for a power adapter based on the Internet of Things includes: S1. Deploy Internet of Things devices at key nodes of the power adapter to collect multi-dimensional operation data of the power adapter under its operating conditions by using the Internet of Things devices, calculate the parameter deviation degree of the multi-dimensional operation data, extract key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, and use the key parameter points to identify the target monitoring area of the power adapter.

[0021] In the embodiment of the present invention, by deploying Internet of Things devices at key nodes of the power adapter to collect multi-dimensional operation data of the power adapter under its operating conditions by using the Internet of Things devices, various data during the operation of the power adapter can be collected in real time and accurately, thereby providing basic data support for subsequent analysis and decision-making.

[0022] Among them, the power adapter refers to an electronic device whose main function is to convert the alternating current provided by the power grid into direct current suitable for various electronic devices and stably supply power to the devices. The Internet of Things device refers to a device integrated with multiple sensor modules for real-time sensing, collecting information related to the operation of the power adapter, and having the function of transmitting this information through the network, such as temperature sensors, humidity sensors, and vibration sensors. The multi-dimensional operation data refers to a series of data sets reflecting the operation state of the power adapter from multiple dimensions, such as temperature, humidity, and vibration data.

[0023] Optionally, the multi-dimensional operation data can be obtained by deploying Internet of Things devices at parts such as the power input port, power output port, and core power conversion module of the power adapter, and then collecting the data generated during the operation of the power adapter by using the deployed Internet of Things devices when the power adapter is operating.

[0024] Furthermore, in the embodiment of the present invention, by calculating the parameter deviation degree of the multi-dimensional operation data, it can quickly identify which parameters have abnormal changes, thereby helping users quickly screen out which parameters have abnormal changes.

[0025] Among them, the parameter deviation degree is an index used to measure the deviation degree of each parameter in the multi-dimensional operation data relative to the reference value.

[0026] As an embodiment of the present invention, calculating the parameter deviation degree of the multi-dimensional operation data includes: identifying the parameter measurement type of the multi-dimensional operation data, classifying the multi-dimensional operation data according to the parameter measurement type to obtain classified data, querying the measured value of the classified data and the observation time corresponding to the measured value, and calculating the observation value deviation degree of the classified data within the observation time based on the measured value by using the following formula: ; Wherein, represents the observation value deviation degree, t represents the observation time, represents the measured value of the i-th classified data at time t, represents the corresponding reference value, represents the deviation weight coefficient of the i-th classified data; Determine the parameter deviation degree of the multi-dimensional operation data according to the observation value deviation degree.

[0027] Wherein, the parameter measurement type refers to the measurement attribute category of different parameters in the multi-dimensional operation data under the operating conditions of the power adapter. For example, voltage, current, temperature, frequency, etc. The measured value refers to the actual measurement value of each parameter collected by various sensors deployed on the key nodes of the power adapter through Internet of Things devices at a specific observation time. For example, at a certain moment, the output voltage of the power adapter measured by the voltage sensor is 20V, and this 20V is the measured value of the voltage parameter. The observation time refers to the specific moment or time period when the measured value in the multi-dimensional operation data is collected.

[0028] Optionally, the parameter measurement type can analyze each parameter included in the multi-dimensional operation data to determine its measurement type, such as obtaining different types such as voltage, current, temperature, and humidity. The classified data can be obtained by dividing the multi-dimensional operation data according to the parameter measurement type and classifying the data with the same measurement type into one category. The measured value can be obtained by extracting the specific measured value of each classified data from the classified data. The observation time can be obtained by querying the test time corresponding to the measured value using MySQL. The parameter deviation degree can be obtained by calculating the observation value deviation degree of different parameters in the multi-dimensional operation data according to the calculation formula of the observation value deviation degree.

[0029] It should be further noted that the calculation formula of the observed value deviation degree highlights the deviation degree between the two by calculating the square of the difference between the measured value and the reference value. The larger the difference, the more obvious the deviation. Dividing by the deviation weight coefficient takes into account the different importance of different classification data on the result, and can quantify the deviation of the multi-dimensional operation data of the power adapter from the reference state, helping users quickly judge whether its operation is normal, so as to evaluate the stability and reliability of the operation state of the power adapter, and timely discover potential failure risks.

[0030] In the embodiment of the present invention, by extracting the key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, the data volume can be reduced, focusing on the most valuable information, thereby improving the analysis efficiency.

[0031] Among them, the key parameter points refer to the parameter points with a relatively high deviation degree screened out according to the parameter deviation degree index in the multi-dimensional operation data of the power adapter. These points represent the key information with a large difference from the normal reference state during the operation of the power adapter, and can reflect the possible abnormalities or the operation states that need to be focused on of the power adapter.

[0032] Optionally, the key parameter points can be obtained by extracting the parameter points corresponding to the parameter deviation degree higher than the preset parameter deviation degree threshold in the multi-dimensional operation data.

[0033] Furthermore, the preset parameter deviation degree threshold can be set according to actual application requirements. If the data volume is large, the preset parameter deviation degree threshold can be set to 0.6. If the data volume is small, the preset parameter deviation degree threshold can be set to 0.4. Specifically, it needs to be set in combination with the actual application.

[0034] In the embodiment of the present invention, by using the key parameter points to identify the target monitoring area of the power adapter, targeted monitoring and maintenance can be carried out, improving the efficiency of fault discovery and handling, and reducing the maintenance cost.

[0035] Among them, the target monitoring area refers to the area of the power adapter selected as the key observation and detection area, such as transformers, capacitors, etc.

[0036] As an embodiment of the present invention, the identification of the target monitoring area of the power adapter by using the key parameter points includes: performing a three-dimensional coordinate system mapping on the key parameter points to obtain a standardized coordinate data set, performing a time synchronization interpolation operation on the standardized coordinate data set to obtain a spatio-temporal alignment data set, performing data clustering on the spatio-temporal alignment data set to obtain an initial monitoring cluster set, performing a spatio-temporal correlation verification on the initial monitoring cluster set to extract an effective data set from the initial monitoring cluster set based on the verification result of the spatio-temporal correlation verification, performing a dynamic risk analysis on the power adapter based on the effective data set to identify a priority monitoring area of the power adapter, and performing a regional boundary processing on the priority monitoring area to obtain a target monitoring area.

[0037] Among them, the standardized coordinate data set refers to a data set formed by mapping key parameter points to a three-dimensional coordinate system and making the coordinate values in a unified scale through normalization and other processes. The spatio-temporal alignment data set refers to a data set obtained by synchronizing and interpolating the standardized coordinate data set according to time stamps so that the data corresponds neatly in both time and space dimensions. The initial monitoring cluster set refers to a set formed by dividing data points into different clusters after processing the spatio-temporal alignment data set. The effective data set refers to a data set composed of clusters that meet the rules after verifying and screening the initial monitoring cluster set according to spatio-temporal correlation rules. The priority monitoring area refers to an area that needs to be monitored preferentially in the monitoring area of the power adapter.

[0038] Optionally, the standardized coordinate data set can first determine the three-dimensional space reference coordinate system of the power adapter, convert the physical position information corresponding to the key parameter points into coordinate values in this coordinate system, and then perform normalization processing on the coordinate values to obtain. The spatio-temporal alignment data set can be obtained by using the interpolation function library in Python to sort and interpolate the standardized coordinate data set according to the time stamp. The initial monitoring cluster set can be obtained by clustering the spatio-temporal alignment data set using DBSCAN clustering. The effective data set can first define the spatio-temporal correlation judgment rules for the initial monitoring cluster set, such as continuity in time, proximity in space, etc., to verify each cluster in the initial monitoring cluster set, and extract the clusters that meet the spatio-temporal correlation requirements as the effective data set. The priority monitoring area can use a decision tree to construct a dynamic risk analysis model, combine factors such as the numerical change trend of key parameter points, the size and density of clusters, etc., evaluate the risk level of the area corresponding to each effective data set, and determine the order of the areas according to the risk level. For example, for the power adapter monitoring, three areas A, B, and C are obtained. In area A, the voltage continues to rise significantly, the cluster is large and dense, and the regression analysis shows that the voltage rise slope is 0.5, and the decision tree determines that the risk is high; in area B, the current fluctuates regularly, the cluster and density are moderate, and the risk is evaluated as medium through time series and neural network; in area C, the temperature fluctuates normally, the cluster is small and sparse, and the statistical judgment shows that the risk is low. According to the risk level, area A has the highest priority for key monitoring, area B is monitored regularly, and area C has a reduced frequency. The target monitoring area can perform boundary smoothing processing on the priority monitoring area, remove the noise points and irregular parts on the boundary, and make the area boundary clearer to obtain.

[0039] S2. Based on the target monitoring area, perform adaptive spectral decomposition on the multi-dimensional operation data to obtain target analysis data, use the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter, and use the multi-dimensional electrical characteristics to locate the abnormal nodes in the target monitoring area to obtain the abnormal circuit area.

[0040] In the embodiment of the present invention, by performing adaptive spectral decomposition on the multi-dimensional operation data based on the target monitoring area to obtain target analysis data, the information of different frequency characteristics in the data can be highlighted, making the data characteristics more obvious, and thus better capturing the subtle changes and hidden information in the operation data of the power adapter, providing a more valuable data basis for accurately identifying electrical characteristics.

[0041] As an embodiment of the present invention, the adaptive spectrum decomposition of the multi-dimensional operation data based on the target monitoring area to obtain target analysis data includes: identifying the area characteristics of the target monitoring area, based on the area characteristics, performing data screening on the multi-dimensional operation data to obtain screened data, calculating the signal Shannon entropy of the screened data, based on the signal Shannon entropy, performing dynamic layer decomposition on the screened data to obtain a time-frequency matrix, and performing time-frequency sub-generation optimization on the time-frequency matrix to obtain target analysis data.

[0042] Among them, the signal Shannon entropy refers to a measure for measuring the uncertainty or information content in a signal, and the time-frequency matrix refers to a matrix that represents a signal in two dimensions of time and frequency.

[0043] Optionally, the area characteristics are obtained by conducting a detailed survey and analysis of the target monitoring area, collecting area information such as the physical characteristics and electrical characteristics of this area, and then analyzing the area information using tools such as sensor networks and circuit design drawings. The screened data can determine the screening criteria and conditions according to the identified area characteristics. For example, if the area characteristics indicate that the area mainly focuses on signals within a certain frequency range, then this frequency range is used as the screening condition to screen out the data that meets the requirements from the multi-dimensional operation data. The calculation of the signal Shannon entropy of the screened data can be performed using the Shannon algorithm. The time-frequency matrix can be obtained by using a wavelet analysis library to automatically adjust the decomposition layer according to the signal Shannon entropy and performing wavelet decomposition on the screened data. The target analysis data can evaluate and rank each time-frequency sub-generation in the time-frequency matrix by calculating indicators such as the energy and correlation of the time-frequency sub-generation. According to the ranking results, the optimal time-frequency sub-generation is selected as the target analysis data, such as selecting the time-frequency sub-generation with the largest energy, the strongest correlation, etc.

[0044] Furthermore, by using the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter in the embodiment of the present invention, the operation state of the power adapter can be comprehensively understood, so as to facilitate the user to discover potential fault hazards of the power adapter.

[0045] Among them, the multi-dimensional electrical characteristics refer to the characteristic performances of the power adapter in multiple dimensions in terms of electrical performance, such as voltage characteristics, current characteristics, and power characteristics, etc.

[0046] Optionally, the multi-dimensional electrical characteristics can be obtained by identifying data such as the voltage value, current value, and power value of the power adapter in the target analysis data.

[0047] Furthermore, in the embodiment of the present invention, by using the multi-dimensional electrical features to locate the abnormal nodes in the target monitoring area, the abnormal circuit area can be obtained, and the area where the power adapter has an abnormality can be accurately located, so that this area can be monitored key points to timely discover potential hidden dangers or faults that may exist.

[0048] Among them, the abnormal circuit area refers to the area where abnormal operation exists in the power adapter, such as voltage abnormality at a certain interface or power output port.

[0049] As an embodiment of the present invention, using the multi-dimensional electrical features to locate the abnormal nodes in the target monitoring area to obtain the abnormal circuit area includes: performing feature screening on the multi-dimensional electrical features to obtain screening features, using the screening features to perform abnormal detection on the target monitoring area to obtain abnormal detection data, performing matrix transformation on the abnormal detection data to obtain an abnormal detection matrix, using the abnormal detection matrix to identify the abnormal range of the target monitoring area to obtain a preliminary identification area, calculating the abnormal similarity of the electrical data in the preliminary identification area, based on the abnormal similarity, performing area segmentation on the target monitoring area to obtain an abnormal segmentation area, identifying the area center point of the abnormal segmentation area, and using the area center point to perform abnormal positioning on the target monitoring area to obtain the abnormal circuit area.

[0050] Among them, the abnormal detection matrix refers to a matrix used to represent the abnormal situation of electrical data in the target monitoring area.

[0051] Optionally, the screening features can be obtained by using the correlation analysis algorithm to calculate the correlation coefficient of each electrical feature with the known abnormal situation, and according to the size of the coefficient, combined with expert experience, key features such as voltage fluctuation features and current distortion features are screened out. The abnormal detection data can be obtained by monitoring and comparing the real-time electrical data of each node in the target monitoring area with the screened electrical features as indicators, and screening out the data that deviates from the normal range. The abnormal detection matrix can be obtained by filling the abnormal detection results (such as boolean values indicating whether it is abnormal, abnormal degree values, etc.) of each node at different times into the corresponding positions of the matrix according to the node number and time sequence. The preliminary identification area can be determined by analyzing the abnormal detection matrix and finding the range of rows (corresponding to nodes) and columns (corresponding to time) where abnormal data is concentrated. The calculation of the abnormal similarity of the electrical data in the preliminary identification area can be calculated using the Euclidean distance. The abnormal segmentation area is obtained by using a clustering algorithm, such as K-means clustering, with the abnormal similarity value as the clustering basis to divide the preliminary identification area into multiple segmentation areas. The area center point can be identified by using the area center method.

[0052] S3. Collect the infrared thermal imaging data and electromagnetic radiation data of the power adapter, perform non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, identify the temperature gradient characteristics of the calibrated thermal distribution map, analyze the thermal stress distribution characteristics of the power adapter using the temperature gradient characteristics, and perform frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum.

[0053] In the embodiment of the present invention, by collecting the infrared thermal imaging data and electromagnetic radiation data of the power adapter, the temperature distribution on the surface of the power adapter and the emission characteristics of electromagnetic energy can be understood.

[0054] Among them, the infrared thermal imaging data refers to the data related to the thermal radiation information of an object obtained through infrared thermal imaging technology, and the electromagnetic radiation data refers to the data of the electromagnetic radiation characteristics generated by an electrical device during operation.

[0055] Optionally, the infrared thermal imaging data can be obtained by collecting the infrared radiation of the power adapter during operation using an infrared imager, and the electromagnetic radiation data can be obtained by collecting the electromagnetic energy of the power adapter during operation using an electromagnetic radiation tester.

[0056] Furthermore, in the embodiment of the present invention, by performing non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, the quality and reliability of the infrared thermal imaging data can be improved, ensuring that the obtained calibrated thermal distribution map can truly reflect the heat generation situation of the power adapter.

[0057] Among them, the calibrated thermal distribution map refers to an image that presents the temperature distribution on the surface of an object in an intuitive and visual way after being processed and optimized.

[0058] As an embodiment of the present invention, performing non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map includes: detecting bad points in the infrared thermal imaging data to obtain bad point coordinates, performing matrix conversion on the bad point coordinates to obtain a bad point coordinate matrix, performing neighborhood interpolation on the bad point coordinate matrix to obtain an interpolated image, and calculating the gain matrix of the interpolated image using the following formula: ; Among them, represents the gain matrix, represents the pixel point of the interpolated image The original gray value at the high calibration temperature, represents the pixel point of the interpolated image The original gray value at the low calibration temperature, represents the high calibration temperature, represents the low calibration temperature; Based on the gain matrix, calculate the offset matrix of the interpolated image using the following formula: ; where, represents the offset matrix, represents the pixel of the interpolated image at the original grayscale value under the low calibration temperature, represents the low calibration temperature; Based on the gain matrix and the offset matrix, perform linear correction on the infrared thermal imaging data to obtain a linearly corrected image, and perform scene adaptive correction on the linearly corrected image to obtain a calibrated thermal distribution map.

[0059] Among them, the bad pixel coordinates refer to those pixel points in the infrared thermal imaging data whose grayscale values deviate significantly from the normal range due to various reasons (such as sensor failure, noise interference, etc.). The interpolated image refers to the image obtained after repairing the bad pixels in the infrared thermal imaging data. The gain matrix refers to the matrix obtained after linearly correcting the infrared thermal imaging data. The offset matrix is calculated based on the interpolated image and the gain matrix and is used to further adjust the infrared thermal imaging data. The high calibration temperature refers to a relatively high temperature value set during the calibration of the infrared thermal imaging device, and the low calibration temperature refers to a relatively low temperature value set during the calibration of the infrared thermal imaging device.

[0060] Optionally, the bad pixel coordinates can be obtained by using a bad pixel detection algorithm based on statistical analysis and traversing the pixel points and performing grayscale value comparison operations using an image processing library (such as OpenCV). The bad pixel coordinate matrix can be obtained by using matrix operation functions in a programming language (such as Python) to convert the bad pixel coordinate list into a two-dimensional matrix, where the rows of the matrix can correspond to the bad pixel numbers and the columns correspond to the abscissa and ordinate. The interpolation of the bad pixel coordinate matrix to obtain the interpolated image can be realized by using the bilinear interpolation method. The linearly corrected image can be obtained by performing matrix operations and image processing techniques, through matrix multiplication and addition operations, to perform operations on the gain matrix, the offset matrix, and the original infrared thermal imaging data matrix to correct the grayscale value of each pixel point. The calibrated thermal distribution map can be obtained by using an adaptive correction algorithm based on scene analysis, such as using the histogram statistical information, edge features, etc. of the image, combined with the adaptive threshold adjustment technology, to perform pixel-by-pixel optimization processing on the linearly corrected image.

[0061] It should be further noted that in the above gain matrix calculation formula, by comparing the gray-scale changes of pixels at high and low calibration temperatures, the change rate of gray-scale when the temperature changes can be calculated. Combining the change rates of each pixel gives the gain matrix, so that the change rate of the gray-scale value of each pixel point can be adjusted, enabling the gray-scale value of the pixel point to more accurately reflect the actual temperature change at different temperatures, thereby improving the accuracy and uniformity of the infrared thermal imaging image. After knowing the gray-scale change rate reflected by the gain matrix, the offset matrix calculation formula can determine the starting offset in the linear relationship between pixel gray-scale and temperature in combination with the low calibration temperature and the corresponding gray-scale, and further adjust the gray-scale value of each pixel point so that at the low calibration temperature, the gray-scale value of the pixel point can accurately correspond to the actual low calibration temperature.

[0062] Further, by identifying the temperature gradient characteristics of the calibrated thermal distribution map in the embodiment of the present invention, the severity and direction of the surface temperature change of the power adapter can be understood.

[0063] Among them, the temperature gradient characteristics refer to a set of characteristics used to describe the temperature change situation on the surface of an object or in space, which can intuitively reflect the change trend and severity of the temperature at different positions and in different directions.

[0064] Optionally, the temperature gradient characteristics can be identified by calculating the temperature difference and direction between adjacent pixel points in the calibrated thermal distribution map.

[0065] As an embodiment of the present invention, the identification of the temperature gradient characteristics of the calibrated thermal distribution map includes: performing an effective area masking operation on the thermal distribution map to obtain a masked image, performing a morphological closing operation on the masked image to obtain an effective image, calculating the gradient matrix of the effective image, performing a polar coordinate transformation on the gradient matrix to obtain a gradient amplitude matrix, performing an adaptive threshold segmentation on the gradient amplitude matrix to obtain a binary image, and using the binary image to identify the temperature gradient characteristics of the calibrated thermal distribution map.

[0066] Among them, the masked image refers to the image obtained by operating a specific mask on the original thermal distribution map. The effective image refers to the image obtained by performing a morphological closing operation on the masked image. The gradient matrix refers to the matrix that records the gray-scale change rates of each pixel point in the horizontal and vertical directions in the effective image. The gradient amplitude matrix refers to the matrix composed of extracting the gradient amplitudes of all pixel points in the gradient matrix.

[0067] Optionally, the mask image can first determine the effective region in the heat distribution map, for example, excluding the background region that has nothing to do with the analysis target. Then, a mask with the same size as the heat distribution map is created. The corresponding part of the mask in the effective region is set to a specific value (such as 1), and the invalid region is set to other values (such as 0). The mask and the heat distribution map are multiplied pixel by pixel to obtain it. The effective image can be obtained by performing a dilation operation on the mask image and then an erosion operation. The gradient matrix of the effective image can be calculated using a gradient operator, such as the Sobel operator. The gradient magnitude matrix can be obtained by performing a polar coordinate transformation on the gradient matrix using a function in a mathematical library (such as NumPy). The binary image can be obtained by performing adaptive threshold segmentation on the gradient magnitude matrix using an adaptive threshold segmentation function in an image processing library. The temperature gradient feature can be obtained by analyzing the distribution of different regions in the binary image, such as the size, shape, and position of connected regions, etc.

[0068] In an embodiment of the present invention, by analyzing the thermal stress distribution characteristics of the power adapter using the temperature gradient feature, the thermo-mechanical performance of the power adapter can be deeply understood to understand the stability during the operation of the power adapter.

[0069] Among them, the thermal stress distribution characteristic refers to the distribution of stress in the entire adapter when the power adapter is working, which is caused by different temperature distributions due to uneven heat generation in each part inside, and then generated.

[0070] Optionally, the thermal stress distribution characteristic can be obtained by establishing a correlation model between the temperature gradient feature and thermal stress, and based on the temperature gradient magnitude, direction, and change situation, combined with the material characteristics and structural parameters of the power adapter, analyzing the magnitude, distribution region, and change trend of the internal thermal stress.

[0071] Furthermore, in an embodiment of the present invention, by performing frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum, abnormal electromagnetic radiation frequency components can be avoided from interfering, so as to more clearly understand the frequency characteristics of the electromagnetic radiation of the power adapter.

[0072] Among them, the standard frequency domain spectrum refers to the spectrum obtained by adjusting and standardizing the frequency components of the original electromagnetic radiation data and eliminating the interference of abnormal frequency components.

[0073] Optionally, the standard frequency domain spectrum can be obtained by converting the electromagnetic radiation data from the time domain to the frequency domain, and then adjusting and scaling the magnitude of the frequency domain data according to the statistical characteristics of the data or a specific normalization algorithm to make it have a unified scale and range.

[0074] S4. Perform three-dimensional topological analysis on the abnormal circuit area to obtain a fault morphology feature vector. Fuse the thermal stress distribution characteristics, standard frequency domain spectrum, and fault morphology feature vector to obtain a fused diagnostic data set. Use the fused diagnostic data set to analyze the health status of the power adapter and obtain a device health analysis report.

[0075] In the embodiment of the present invention, by performing three-dimensional topological analysis on the abnormal circuit area to obtain a fault morphology feature vector, the structure and morphology of the abnormal circuit area can be described in detail from a spatial perspective, which can provide accurate information about the physical form of the fault for the fault diagnosis of the abnormal circuit area.

[0076] The fault morphology feature vector refers to a set of numerical feature collections obtained by performing three-dimensional topological analysis on the abnormal circuit area and used to characterize the fault morphology of this area.

[0077] As an embodiment of the present invention, performing three-dimensional topological analysis on the abnormal circuit area to obtain a fault morphology feature vector includes: performing three-dimensional finite element modeling on the abnormal circuit area to obtain a visualized circuit area, identifying the current density distribution data, thermal stress data, and electromagnetic field strength distribution data in the visualized circuit area, and performing vectorization processing on the current density distribution data, the thermal stress data, and the electromagnetic field strength distribution data to obtain a fault morphology feature vector.

[0078] The visualized circuit area refers to a three-dimensional diagram that shows the structure, layout, and spatial relationship of each part of the circuit. The current density distribution data refers to a set of data that describes the current distribution in the visualized circuit area. The thermal stress data refers to the data that reflects the stress distribution caused by temperature changes in the visualized circuit area. The electromagnetic field strength distribution data refers to the data that represents the spatial distribution of the electromagnetic field strength in the visualized circuit area.

[0079] Optionally, performing three-dimensional finite element modeling on the abnormal circuit area to obtain a visualized circuit area can be achieved by using professional finite element modeling software such as ANSYS. Further, in the embodiment of the present invention, by fusing the thermal stress distribution characteristics, standard frequency domain spectrum, and fault morphology feature vector to obtain a fused diagnostic data set, the advantages of various data can be fully utilized, complementing and corroborating each other, and improving the accuracy and reliability of fault diagnosis.

[0080] As an embodiment of the present invention, the multi-source data fusion of the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vectors to obtain a fusion diagnosis data set includes: performing matrix transformation on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vectors to obtain a first matrix, a second matrix, and a third matrix; performing tensor combination on the first matrix, the second matrix, and the third matrix to obtain a high-order tensor; performing multilinear decomposition on the high-order tensor to obtain a low-rank matrix; performing matrix solution on the low-rank matrix to obtain matrix elements; and performing element fusion on the matrix elements to obtain a fusion diagnosis data set.

[0081] Among them, the first matrix refers to the matrix obtained by matrix transformation of the thermal stress distribution characteristics, the second matrix refers to the matrix obtained by matrix transformation of the standard frequency domain spectrum, the third matrix refers to the matrix transformed from the fault form feature vectors, and the low-rank matrix refers to the matrix obtained by performing multilinear decomposition on the high-order tensor.

[0082] Optionally, the matrix transformation of the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vectors can be implemented using the NumPy library in Python. The high-order tensor can convert the first matrix, the second matrix, and the third matrix into a dimensional slice representation of the tensor, and then stack and combine the dimensional slices in a specific order, such as the size of the tensor on a new dimension. The low-rank matrix can be obtained by performing Tucker decomposition on the high-order tensor. The matrix solution of the low-rank matrix to obtain matrix elements can be implemented using the NumPy library. The fusion diagnosis data set can perform weighted average processing on all elements obtained from the matrix solution of the low-rank matrix, and arrange them in order according to the element values after weighted average processing to form a data set.

[0083] In the embodiment of the present invention, by using the fusion diagnosis data set to analyze the health status of the power adapter, an equipment health analysis report can be obtained to evaluate and judge the overall health status of the power adapter, and at the same time provide a detailed report on the health status of the power adapter for users.

[0084] Optionally, the process of using the fusion diagnosis data set to analyze the health status of the power adapter to obtain an equipment health analysis report is as follows: using machine learning algorithms (such as support vector machines, random forests, etc.) to train and predict the fusion diagnosis data set, evaluating the health status of the power adapter from multiple dimensions such as thermal performance, electromagnetic compatibility, and circuit faults based on the prediction results, and then constructing a health analysis report of the power adapter using the result data of the multi-dimensional evaluation.

[0085] Embodiment 2: Such asFigure 2 As shown, it is a functional module diagram of the intelligent monitoring system for power adapters based on the Internet of Things according to the present invention.

[0086] The intelligent monitoring system 200 for power adapters based on the Internet of Things according to the present invention can be installed in an electronic device. According to the functions achieved, the intelligent monitoring system for power adapters based on the Internet of Things can include a target area recognition module 201, an abnormal area recognition module 202, a multi-dimensional data processing module 203, and a status analysis module 204. The modules in the present invention can also be referred to as units, which refer to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, and are stored in the memory of the electronic device.

[0087] In the embodiment of the present invention, the functions of each module / unit are as follows: The target area recognition module 201 is used to deploy Internet of Things devices at key nodes of the power adapter to collect multi-dimensional operation data under the operating conditions of the power adapter by using the Internet of Things devices, calculate the parameter deviation degree of the multi-dimensional operation data, extract key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, and use the key parameter points to identify the target monitoring area of the power adapter; The abnormal area recognition module 202 is used to perform adaptive spectral decomposition on the multi-dimensional operation data based on the target monitoring area to obtain target analysis data, use the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter, and use the multi-dimensional electrical characteristics to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area; The multi-dimensional data processing module 203 is used to collect infrared thermal imaging data and electromagnetic radiation data of the power adapter, perform non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, identify the temperature gradient characteristics of the calibrated thermal distribution map, analyze the thermal stress distribution characteristics of the power adapter by using the temperature gradient characteristics, and perform frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum; The status analysis module 204 is used to perform three-dimensional topology analysis on the abnormal circuit area to obtain a fault form feature vector, perform multi-source data fusion on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a fusion diagnosis data set, and use the fusion diagnosis data set to perform a health status analysis on the power adapter to obtain an equipment health analysis report.

[0088] Specifically, each module in the intelligent monitoring system 200 for power adapters based on the Internet of Things in the embodiment of the present invention adopts the same as the above-mentioned Figure 1The technical means are the same as those described in the intelligent monitoring method of the power adapter based on the Internet of Things, and the same technical effects can be achieved, which will not be elaborated here.

[0089] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. The intelligent monitoring method for a power adapter based on the Internet of Things, characterized in that, The method includes: Deploying Internet of Things devices at key nodes of the power adapter to collect multi-dimensional operation data under the operating conditions of the power adapter, calculating the parameter deviation degree of the multi-dimensional operation data, extracting key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, and using the key parameter points to identify the target monitoring area of the power adapter; Based on the target monitoring area, performing adaptive spectral decomposition on the multi-dimensional operation data to obtain target analysis data, using the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter, and using the multi-dimensional electrical characteristics to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area; Collecting infrared thermal imaging data and electromagnetic radiation data of the power adapter, performing non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, identifying the temperature gradient characteristics of the calibrated thermal distribution map, analyzing the thermal stress distribution characteristics of the power adapter using the temperature gradient characteristics, and performing frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum; Performing three-dimensional topology analysis on the abnormal circuit area to obtain a fault form feature vector, fusing multi-source data of the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a fusion diagnosis data set, and using the fusion diagnosis data set to perform a health status analysis on the power adapter to obtain an equipment health analysis report.

2. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, wherein, The calculating the parameter deviation degree of the multi-dimensional operation data includes: Identifying the parameter measurement type of the multi-dimensional operation data; Classifying the multi-dimensional operation data according to the parameter measurement type to obtain classified data; Querying the measured value of the classified data and the observation time corresponding to the measured value; Calculating the observation value deviation degree of the classified data within the observation time based on the measured value using the following formula: ； Among them, represents the deviation degree of the observed value, t represents the observation time, represents the measured value of the i-th classified data at time t, represents the corresponding reference value, represents the deviation weight coefficient of the i-th classified data; Determining the parameter deviation degree of the multi-dimensional operation data according to the observation value deviation degree.

3. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, wherein, The using the key parameter points to identify the target monitoring area of the power adapter includes: Performing three-dimensional coordinate system mapping on the key parameter points to obtain a standardized coordinate data set; Performing time synchronization interpolation operation on the standardized coordinate data set to obtain a spatio-temporal alignment data set; Performing data clustering on the spatio-temporal alignment data set to obtain an initial monitoring cluster set; Performing spatio-temporal correlation verification on the initial monitoring cluster set to extract an effective data set from the initial monitoring cluster set using the verification result of the spatio-temporal correlation verification; Performing dynamic risk analysis on the power adapter based on the effective data set to identify the priority monitoring area of the power adapter; Performing regional boundary processing on the priority monitoring area to obtain the target monitoring area.

4. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, characterized in that The performing adaptive spectral decomposition on the multi-dimensional operation data based on the target monitoring area to obtain target analysis data includes: Identifying the regional characteristics of the target monitoring area; Performing data screening on the multi-dimensional operation data based on the regional characteristics to obtain screened data; Calculating the signal Shannon entropy of the screened data; Based on the Shannon entropy of the signal, perform dynamic layer decomposition on the screened data to obtain a time-frequency matrix; Perform time-frequency sub-generation optimization on the time-frequency matrix to obtain target analysis data.

5. The intelligent monitoring method of the power adapter based on the Internet of Things according to claim 1, characterized in that The using the multi-dimensional electrical features to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area, including: Perform feature screening on the multi-dimensional electrical features to obtain screened features; Use the screened features to perform abnormal detection on the target monitoring area to obtain abnormal detection data; Perform matrix transformation on the abnormal detection data to obtain an abnormal detection matrix; Use the abnormal detection matrix to identify the abnormal range of the target monitoring area to obtain a preliminary identification area; Calculate the abnormal similarity of the electrical data in the preliminary identification area; Based on the abnormal similarity, perform area segmentation on the target monitoring area to obtain an abnormal segmentation area; Identify the area center point of the abnormal segmentation area; Use the area center point to perform abnormal positioning on the target monitoring area to obtain an abnormal circuit area.

6. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, wherein The performing non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, including: Perform bad pixel detection on the infrared thermal imaging data to obtain bad pixel coordinates; Perform matrix conversion on the bad pixel coordinates to obtain a bad pixel coordinate matrix; Perform neighborhood interpolation on the bad pixel coordinate matrix to obtain an interpolated image; Use the following formula to calculate the gain matrix of the interpolated image: ； Among them, represents the gain matrix, represents the pixel points of the interpolated image and the original gray value at the high calibration temperature, represents the pixel points of the interpolated image and the original gray value at the low calibration temperature, represents the high calibration temperature, represents the low calibration temperature; Based on the gain matrix, use the following formula to calculate the offset matrix of the interpolated image: ； Among them, represents the offset matrix, represents the pixel points of the interpolated image with the original gray value at the low calibration temperature, represents the low calibration temperature; Based on the gain matrix and the offset matrix, perform linear correction on the infrared thermal imaging data to obtain a linearly corrected image; Perform scene adaptive correction on the linearly corrected image to obtain a calibrated thermal distribution map.

7. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, characterized in that The identifying the temperature gradient feature of the calibrated thermal distribution map, including: Perform an effective area masking operation on the thermal distribution map to obtain a masked image; Perform morphological closing operation on the masked image to obtain an effective image; Calculate the gradient matrix of the effective image; Perform polar coordinate transformation on the gradient matrix to obtain a gradient amplitude matrix; Perform adaptive threshold segmentation on the gradient amplitude matrix to obtain a binary image; Use the binary image to identify the temperature gradient feature of the calibrated thermal distribution map.

8. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, characterized in that The performing three-dimensional topology analysis on the abnormal circuit area to obtain a fault morphology feature vector, including: Perform three-dimensional finite element modeling on the abnormal circuit area to obtain a visualized circuit area; Identify the current density distribution data, thermal stress data, and electromagnetic field intensity distribution data in the visualized circuit area; Perform vectorization processing on the current density distribution data, the thermal stress data, and the electromagnetic field intensity distribution data to obtain a fault morphology feature vector.

9. The intelligent monitoring method for a power adapter based on the Internet of Things according to claim 1, wherein The performing multi-source data fusion on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault morphology feature vector to obtain a fusion diagnosis data set, including: Perform matrix transformation on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault morphology feature vector to obtain a first matrix, a second matrix, and a third matrix; Perform a tensor combination on the first matrix, the second matrix, and the third matrix to obtain a high-order tensor; Perform a multilinear decomposition on the high-order tensor to obtain low-rank matrices; Solve the matrix for the low-rank matrix to obtain matrix elements; Perform element fusion on the matrix elements to obtain a fused diagnostic data set.

10. The intelligent monitoring system for power adapters based on the Internet of Things is characterized in that The system includes: A target area recognition module for deploying Internet of Things devices at key nodes of a power adapter to collect multi-dimensional operation data under the operating conditions of the power adapter using the Internet of Things devices, calculate the parameter deviation degree of the multi-dimensional operation data, extract key parameter points of the multi-dimensional operation data based on the parameter deviation degree index, and use the key parameter points to identify the target monitoring area of the power adapter; An abnormal area recognition module for performing adaptive spectral decomposition on the multi-dimensional operation data based on the target monitoring area to obtain target analysis data, using the target analysis data to identify the multi-dimensional electrical characteristics of the power adapter, and using the multi-dimensional electrical characteristics to locate abnormal nodes in the target monitoring area to obtain an abnormal circuit area; A multi-dimensional data processing module for collecting infrared thermal imaging data and electromagnetic radiation data of the power adapter, performing non-uniformity correction on the infrared thermal imaging data to obtain a calibrated thermal distribution map, identifying the temperature gradient characteristics of the calibrated thermal distribution map, analyzing the thermal stress distribution characteristics of the power adapter using the temperature gradient characteristics, and performing frequency domain normalization processing on the electromagnetic radiation data to obtain a standard frequency domain spectrum; A state analysis module for performing three-dimensional topology analysis on the abnormal circuit area to obtain a fault form feature vector, performing multi-source data fusion on the thermal stress distribution characteristics, the standard frequency domain spectrum, and the fault form feature vector to obtain a fused diagnostic data set, and using the fused diagnostic data set to perform a health state analysis on the power adapter to obtain a device health analysis report.

Images