A method and system for condition monitoring of distributed distribution transformers
By performing initial fault screening and early warning at the edge processing layer, combined with cloud-based diagnostics and historical trend analysis, and dynamically updating the fault identification threshold, the problems of real-time early warning timeliness and false positives/false negatives in distributed distribution transformers are solved, thereby improving the operational reliability of the distribution network.
Patent Information
- Application Number
- CN202611237076.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-14
- Publication Date
- 2026-11-10
AI Technical Summary
In existing technologies, the condition monitoring of distributed distribution transformers suffers from problems such as insufficient real-time early warning and misjudgment or missed judgment due to the lengthy data transmission links and fixed fault discrimination thresholds that cannot be adapted to different operating conditions.
A three-layer collaborative architecture (terminal perception layer, edge processing layer, and cloud analysis layer) is adopted. The edge processing layer performs initial screening and early warning of faults. By extracting operational characteristic parameters and comparing them with preset fault discrimination thresholds, abnormal data is uploaded in real time, and diagnosis and historical trend analysis are performed in the cloud to dynamically update the fault discrimination thresholds.
It shortens the data link, improves real-time performance and accuracy, reduces false positives and false negatives, and enhances the operational reliability of the power distribution network.
Smart Images

Figure CN123073767A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of power equipment condition monitoring technology, and more specifically, relates to a condition monitoring method and system for distributed distribution transformers. Background Technology
[0002] With the rapid advancement of the power Internet of Things (IoT) and the large-scale integration of distributed energy resources, the structure of power distribution networks is becoming increasingly complex, and the number and distribution range of distributed distribution transformers are increasing significantly. As the terminal power distribution equipment that directly faces users in the power distribution network, the operating status of distribution transformers directly affects the reliability of power supply and power quality.
[0003] Currently, the condition monitoring of distribution transformers mainly adopts a centralized cloud processing model, where operational data is collected by sensors and uploaded to a cloud platform for analysis and diagnosis. However, with the surge in the number of distributed distribution transformers, the data upload, queuing, and result feedback links are lengthy, making it difficult to meet the timeliness requirements of real-time fault early warning. Furthermore, most existing methods use fixed fault discrimination thresholds to uniformly determine the faults of distribution transformers of different models and operating conditions. Due to the individual differences, operating environments, and aging levels of each transformer, fixed thresholds are difficult to adapt to the dynamic changes in actual operating scenarios, easily leading to misjudgments or missed judgments. Summary of the Invention
[0004] This application provides a method and system for monitoring the status of distributed distribution transformers to solve the problems of long data transmission links making it difficult to meet the timeliness requirements of real-time early warning, and fixed fault discrimination thresholds being unable to adapt to different operating conditions, leading to misjudgments or missed judgments.
[0005] The technical solutions provided in this application are as follows: Firstly, a method for monitoring the state of a distributed distribution transformer is provided, applied to a controller. This controller includes a terminal sensing layer, an edge processing layer, and a cloud analysis layer. The method includes: The terminal perception layer obtains the status data of each distributed distribution transformer in the target area and sends the status data to the edge processing layer. For each distributed distribution transformer, the edge processing layer extracts the operating characteristic parameters from the status data corresponding to the distributed distribution transformer and compares the operating characteristic parameters with the preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer according to the preset period. If the comparison result is an abnormal state, a local early warning is triggered and an abnormal event data packet is generated and sent to the cloud analysis layer. The status data of each distribution transformer in the target group is retrieved by the cloud analysis layer, and the status data is compared with the abnormal event data packet to obtain the diagnostic results. Obtain historical status data of the distributed distribution transformer and predict the trend of the status data based on the historical status data. The edge processing layer updates the preset fault identification threshold based on the diagnostic results and the trend of status data, and updates the status label of the distributed distribution transformer.
[0006] Secondly, a condition monitoring system for distributed distribution transformers is provided, which is installed within a controller. This system includes: The status data acquisition module is used to acquire the status data of each distributed distribution transformer in the target area through the terminal perception layer, and send the status data to the edge processing layer. The feature parameter comparison module is used to extract operating feature parameters from the status data corresponding to each distributed distribution transformer through the edge processing layer, and compare the operating feature parameters with the preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer according to the preset period. If the comparison result is an abnormal state, a local early warning is triggered and an abnormal event data packet is generated and sent to the cloud analysis layer. The diagnostic result acquisition module is used to retrieve the status data of each distribution transformer in the target group through the cloud analysis layer, and compare the status data with the abnormal event data packet to obtain the diagnostic result; The trend acquisition module is used to acquire historical status data of the distributed distribution transformer and predict the trend of the status data based on the historical status data. The status monitoring module is used to update the preset fault identification threshold based on the diagnostic results and the trend of status data through the edge processing layer, and to update the status label of the distributed distribution transformer.
[0007] Thirdly, embodiments of this application also provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the state monitoring method for distributed distribution transformers provided in the first aspect.
[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the state monitoring method for distributed distribution transformers provided in the first aspect.
[0009] The beneficial effects of the technical solution provided in this application are as follows: Compared with related technologies, the distributed distribution transformer status monitoring method and system provided in this application embodiment adopts a three-layer collaborative architecture (e.g., terminal sensing layer, edge processing layer, and cloud analysis layer) to push fault screening and early warning down to the edge side, shortening the data link, alleviating cloud latency, and improving real-time performance. Furthermore, the edge processing layer periodically uploads normal data, and if abnormal data is detected, it immediately uploads abnormal event data packets, ensuring priority handling of abnormalities while reducing cloud load and enabling centralized comparison and diagnosis in the cloud, thus improving accuracy. In this embodiment, cloud diagnostic results are combined with historical trend predictions and fed back to the edge processing layer to dynamically update the preset fault discrimination threshold and corresponding status labels. This overcomes the shortcomings of fixed thresholds that cannot adapt to individual differences and aging changes in distribution transformers, achieving threshold self-adaptation, reducing false positives and false negatives, and improving the reliability of distribution network operation. Attached Figure Description
[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below.
[0011] Figure 1 A flowchart illustrating the state monitoring method for distributed distribution transformers provided in this application embodiment; Figure 2 This is a structural block diagram of the condition monitoring system for distributed distribution transformers provided in an embodiment of this application; Figure 3 A schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation
[0012] The embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the embodiments described below with reference to the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions of the embodiments of this application.
[0013] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the terms “comprising” and “including” as used in embodiments of this application mean that the corresponding feature can be implemented as the presented feature, information, data, step, operation, element, and / or component, but do not exclude implementation as other features, information, data, step, operation, element, component, and / or combinations thereof supported by the art. It should be understood that when we say that an element is “connected” or “coupled” to another element, the one element can be directly connected or coupled to the other element, or it can mean that the one element and the other element establish a connection relationship through an intermediate element. Furthermore, “connected” or “coupled” as used herein can include wireless connection or wireless coupling. The term “and / or” as used herein indicates at least one of the items defined by the term; for example, “A and / or B” can be implemented as “A,” or as “B,” or as “A and B.” When describing multiple (two or more) items, if the relationship between the multiple items is not explicitly defined, the multiple items can refer to one, several or all of the multiple items. For example, the description of "parameter A includes A1, A2, A3" can be implemented as parameter A includes A1 or A2 or A3, or it can be implemented as parameter A includes at least two of the three items A1, A2 and A3.
[0014] This application provides a method for monitoring the status of distributed distribution transformers. This method can be executed by a controller, which includes a terminal sensing layer, an edge processing layer, and a cloud analysis layer. The terminal sensing layer refers to the data acquisition front end, including multiple sensors deployed on each distributed distribution transformer and in its operating environment. The edge processing layer refers to edge computing gateways or smart terminal devices deployed locally or near the distribution transformer, which have data storage and computing capabilities and are used to process and analyze the raw status data uploaded by the terminal sensing layer in real time. The cloud analysis layer refers to a server cluster or cloud service platform deployed in a cloud computing center, which is used to receive and store the status data uploaded by the edge processing layer, perform cross-device group comparison diagnosis, trend prediction, and management of the operating characteristic benchmark library, and feed back the diagnostic results to the edge processing layer.
[0015] like Figure 1 As shown, the method may include: S101: Obtain the status data of each distributed distribution transformer in the target area through the terminal perception layer, and send the status data to the edge processing layer.
[0016] In this embodiment, the distribution transformer refers to a dry-type distribution transformer.
[0017] In this embodiment, the target area refers to a pre-defined geographical area or equipment group where distribution transformer status monitoring is required, such as the entire power supply area covered by an industrial park or a residential community. Status data refers to the raw dataset reflecting the real-time operating conditions of the distribution transformer. For example, this status data includes at least winding temperature data, core temperature data, vibration amplitude data, vibration frequency data, partial discharge data, three-phase voltage data, three-phase current data, and environmental status data.
[0018] In this embodiment, the terminal perception layer collects the status data corresponding to each distributed distribution transformer according to a preset sampling period (e.g., once every 10 seconds), and then sends the status data to the edge processing layer through a preset communication method (e.g., wired Ethernet or wireless communication network).
[0019] Furthermore, to ensure the quality and integrity of the data sent to the edge processing layer and to avoid invalid data caused by sensor noise, communication interference, or momentary interruptions affecting the accuracy of feature extraction, in one embodiment of this application, after obtaining the status data corresponding to each distributed distribution transformer in the target area through the terminal perception layer, a preprocessing step for the status data is further included, specifically including: The status data is cleaned by removing invalid data frames to obtain the cleaned status data. Data interpolation is performed on the cleaned state data according to the acquisition time sequence to obtain the interpolated state data.
[0020] In this embodiment, the status data is first cleaned to remove invalid data frames, resulting in cleaned status data. For example, the integrity of each parameter in each frame of status data is checked. If a key parameter is missing in a frame (e.g., the current value of any one of the three phases is missing), the frame of status data is determined to be invalid. Alternatively, the numerical range of each parameter in the status data is checked. If the value of a parameter is significantly outside the reasonable range (e.g., the three-phase voltage value exceeds ±20% of the rated voltage, or the temperature value exceeds the range of -40℃ to 200℃), the frame of status data is determined to be invalid. Or, if multiple consecutive frames of data are completely identical, it is highly likely that the data is invalid due to a sensor malfunction, and therefore it is removed.
[0021] Then, data interpolation is performed on the cleaned state data according to the acquisition time sequence to obtain the interpolated state data. Specifically, based on the preset data acquisition time interval (e.g., acquisition period of 10 seconds), missing time points in the data sequence are identified; for the data at the missing time points, linear interpolation is used to fill them, that is, the values of the corresponding parameters in the two adjacent valid data frames before and after the missing point are taken, and their arithmetic mean is calculated as the data value of the missing point; if the missing point is the start point or the end point of the data sequence, the value in the nearest adjacent valid data frame is taken as the data value of the missing point.
[0022] S102: For each distributed distribution transformer, the edge processing layer extracts the operating characteristic parameters from the status data corresponding to the distributed distribution transformer, and compares the operating characteristic parameters with the preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer according to the preset cycle. If the comparison result is an abnormal state, a local early warning is triggered and an abnormal event data packet is generated and sent to the cloud analysis layer.
[0023] In this embodiment, the status data includes winding temperature data, core temperature data, vibration amplitude data, vibration frequency data, partial discharge data, three-phase voltage data, and three-phase current data. Operating characteristic parameters refer to key physical quantities extracted from the original status data that can quantitatively characterize the operating status of the distribution transformer. In this embodiment, the characteristic operating parameters include temperature change rate, peak vibration frequency, average partial discharge value, voltage imbalance, current imbalance, and real-time load rate. The preset fault discrimination threshold refers to a set of critical values pre-stored in the local memory of the edge processing layer to determine whether each operating characteristic parameter is within its normal range. In this embodiment, during system initialization, the preset fault discrimination threshold uses a broad factory default value, used only to capture extreme outliers. In subsequent operation, the preset fault discrimination threshold is dynamically updated by the cloud analysis layer based on the normal parameter fluctuation range of the target group, so that the edge layer warning threshold is wider than the group fluctuation range boundary, avoiding premature alarm triggering by the edge layer.
[0024] In one embodiment of this application, operating characteristic parameters are extracted from the state data corresponding to the distributed distribution transformer through an edge processing layer, including: Based on the winding temperature data and core temperature data, determine the rate of temperature change; based on the vibration amplitude data and vibration frequency data, determine the peak value of the dominant vibration frequency; based on the partial discharge quantity data, determine the average partial discharge value; based on the three-phase voltage data and three-phase current data in the status data, determine the voltage imbalance and current imbalance; based on the three-phase current data and the rated current of the distributed distribution transformer, determine the real-time load rate.
[0025] In this embodiment, the rate of temperature change refers to the rate of temperature change of the windings, i.e., the amount of temperature change per unit time, used to characterize the dynamic trend of the transformer's thermal state. The peak value of the dominant vibration frequency refers to the amplitude corresponding to the frequency component with the largest amplitude in the vibration spectrum. Voltage imbalance refers to the percentage ratio of the negative-sequence component to the positive-sequence component of the three-phase voltage, used to characterize the symmetry of the three-phase voltage; current imbalance refers to the percentage ratio of the negative-sequence component to the positive-sequence component of the three-phase current, used to characterize the symmetry of the three-phase load. Real-time load factor refers to the percentage ratio of the current operating current of the distribution transformer to its rated current, used to reflect the current load level of the distribution transformer.
[0026] In this embodiment, the edge processing layer reads the winding temperature data and the core temperature data from the interpolated state data according to a preset time window (e.g., 6 temperature data points continuously collected within the most recent minute, with a collection period of 10 seconds). Then, the least squares method is used to linearly fit the temperature values within the time window to obtain a fitted straight line of temperature change over time. The slope of this fitted straight line is the rate of temperature change.
[0027] From the interpolated state data, the vibration amplitude and frequency data corresponding to the current moment are read. The original vibration time-domain signal is transformed from the time domain to the frequency domain by performing a Fast Fourier Transform on it through the edge processing layer to obtain the spectral distribution of the vibration signal; then, the frequency point with the largest amplitude is searched in the spectrum, and the frequency value corresponding to this frequency point is the vibration dominant frequency, and the amplitude corresponding to this frequency point is the peak value of the vibration dominant frequency.
[0028] From the interpolated state data, read multiple instantaneous values of partial discharge within a preset statistical period (e.g., 10 consecutive power frequency cycles), calculate the arithmetic mean of all instantaneous values of partial discharge within the preset statistical period, and obtain the mean value of partial discharge.
[0029] From the interpolated state data, the effective values of the three-phase phase voltages (Ua, Ub, Uc) and the effective values of the three-phase line currents (Ia, Ib, Ic) at the same moment are read. The three-phase voltages are calculated using the symmetrical component method, decomposing them into positive-sequence, negative-sequence, and zero-sequence components. The voltage unbalance is then calculated using the following formula: Voltage Unbalance = (Negative-sequence Voltage Amplitude / Positive-sequence Voltage Amplitude) × 100%. In this embodiment, the three-phase currents are decomposed into positive-sequence and negative-sequence components using the symmetrical component method, and the current unbalance is calculated using the following formula: Current Unbalance = (Negative-sequence Current Amplitude / Positive-sequence Current Amplitude) × 100%.
[0030] From the interpolated state data, read the effective values of the three-phase line currents (Ia, Ib, Ic) at the same moment, calculate the average effective value of the three-phase line currents, and then read the rated current value (Ie) of the distributed distribution transformer from the local storage of the edge processing layer; calculate the real-time load rate according to the following real-time load rate formula, where the real-time load rate formula can be: Real-time load rate = (average three-phase current / rated current) × 100%.
[0031] During the operation of a distribution transformer, its characteristic parameters such as winding temperature, core temperature, vibration characteristics, and partial discharge are significantly affected by external environmental factors. For example, under the same load current conditions, there is a large difference in the absolute value of winding temperature between high-temperature environments in summer and low-temperature environments in winter; when the ambient humidity is too high, the measured value of partial discharge may fluctuate due to the increased air humidity.
[0032] To further reduce interference from environmental factors, in one embodiment of this application, after extracting operating characteristic parameters from the state data corresponding to the distributed distribution transformer through the edge processing layer, the operating characteristic parameters are further corrected. For example, the operating characteristic parameters are corrected according to the environmental state data to obtain the corrected operating characteristic parameters.
[0033] Specifically, in the edge processing layer, the current ambient temperature value is obtained, and the ambient temperature value from the previous moment is retrieved to calculate the rate of change of the ambient temperature (unit: °C / min). The rate of change of the winding temperature is subtracted from the rate of change of the ambient temperature to obtain the corrected rate of change of the winding temperature; similarly, the rate of change of the core temperature is subtracted from the rate of change of the ambient temperature to obtain the corrected rate of change of the core temperature.
[0034] Based on a pre-stored correction factor table, the correction factor corresponding to the current ambient temperature is determined. For example, in the correction factor table, with 20℃ as the reference temperature, the correction factor at this reference temperature is set to 1.0; when the ambient temperature is 0℃, the correction factor is 1.15 (i.e., the measured value is about 15% lower than the reference state, requiring amplification compensation); when the ambient temperature is 40℃, the correction factor is 0.88 (i.e., the measured value is about 12% higher than the reference state, requiring reduction adjustment). The current ambient temperature value is read through the edge processing layer, and the corresponding correction factor is obtained by querying the above correction factor table; the product of the vibration dominant frequency peak value and the correction factor is taken as the corrected vibration dominant frequency peak value.
[0035] Based on the pre-stored humidity-partial discharge correction curve, the correction factor corresponding to the current ambient humidity is determined. For example, with 50%RH as the baseline humidity, the correction factor at that humidity is set to 1.0. When the ambient humidity is 80%RH, since the measured value tends to be too high in high humidity environments, the correction factor can be set to 0.85 (that is, the measured value is converted to 85% to eliminate the positive deviation introduced by humidity). When the ambient humidity is 30%RH, the correction factor can be set to 1.05 (the measured value is more stable when the environment is dry, and a small correction is sufficient). The current ambient humidity value is read through the edge processing layer, the corresponding correction factor is obtained by querying the above correction curve, and then the product of the average partial discharge value and the correction factor is used as the corrected average partial discharge value.
[0036] Since voltage and current imbalances reflect the symmetry of three-phase electrical quantities, their physical nature is weakly correlated with external environmental temperature and humidity; environmental changes will not substantially affect the symmetry of these electrical quantities. Therefore, in this embodiment, no environmental condition data correction is performed on these two characteristic parameters, voltage and current imbalances. The voltage imbalance value is used as the corrected voltage imbalance value, and the current imbalance value is used as the corrected current imbalance value.
[0037] Since the real-time load factor is calculated as the ratio of three-phase current data to rated current, it represents the electrical load level of the transformer. Changes in environmental conditions (such as increased ambient temperature) may affect the transformer's heat dissipation and permissible overload capacity, but they do not alter the proportional relationship between the current and rated current reflected in the load factor. Therefore, the real-time load factor itself does not require numerical correction based on environmental condition data; the real-time load factor value is used as the corrected real-time load factor.
[0038] In this embodiment, the corrected normal state data is sent to the cloud analysis layer at a preset period, so that the cloud analysis layer can establish a runtime feature benchmark library based on the corrected normal state data.
[0039] As can be seen from the above, this embodiment uses environmental state data to correct the operating characteristic parameters that are susceptible to environmental factors, thereby eliminating the interference of external environmental changes on the monitoring data. This allows the corrected characteristic parameters to truly reflect the operating status of the distribution transformer itself, avoiding misjudgment or missed judgment of faults caused by seasonal or sudden changes in environmental temperature and humidity, and improving the accuracy and reliability of fault discrimination threshold comparison.
[0040] In this embodiment, after obtaining the corrected operating characteristic parameters, the corrected operating characteristic parameters are compared with the preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer according to the preset period. If the comparison result is an abnormal state, a local warning is triggered and an abnormal event data packet is generated and sent to the cloud analysis layer.
[0041] Specifically, the edge processing layer sequentially judges whether the corrected winding temperature change rate, the corrected core temperature change rate, the corrected peak value of the main vibration frequency, the corrected average value of partial discharge, the voltage imbalance, the current imbalance, and the real-time load rate are within the corresponding preset fault judgment threshold range. If all parameters are within their respective preset fault judgment threshold ranges, the current state is determined to be normal. If any parameter exceeds its corresponding preset fault judgment threshold range, the current state is determined to be abnormal.
[0042] If the comparison result indicates a normal state, the edge processing layer sends the normal state data to the cloud analysis layer at a preset cycle (e.g., every 15 minutes). This normal state data includes the device identifier, timestamp information, and corrected operational characteristic parameters at the current moment. Upon receiving this data, the cloud analysis layer stores it as a normal sample in the operational characteristic benchmark library for the corresponding target group, which is used to subsequently update the statistical parameters of the benchmark library.
[0043] If the comparison result is abnormal, the edge processing layer immediately triggers a local early warning, notifying the site of the abnormal situation through audible and visual alarms or by pushing notifications to the mobile terminals of maintenance personnel; and generates an abnormal event data packet, which includes at least the following: the device ID and device model of the abnormal device, the name of the parameter that triggered the abnormality and its real-time value and corresponding threshold, the timestamp of the abnormality trigger, and the continuous status data within a preset time period (e.g., 5 minutes) before the abnormality was triggered. After the data packet is encapsulated, it is sent to the cloud analysis layer so that the cloud analysis layer can perform subsequent group comparison diagnosis in a timely manner.
[0044] S103: Retrieve the status data of each distribution transformer in the target group through the cloud analysis layer, compare the status data with the abnormal event data packet for diagnosis, and obtain the diagnosis result.
[0045] In one embodiment of this application, before retrieving the status data of each distribution transformer within the target group through the cloud analytics layer, the method further includes: The cloud-based analysis layer uses the equipment model, commissioning period, installation area, and load condition of each distributed distribution transformer as group classification tags. Based on these group classification tags, distributed transformers with the same equipment model, commissioning period, installation area, and load condition are grouped into the same target group. Based on the normal status data of each distributed distribution transformer received by the cloud-based analysis layer, a corresponding operating characteristic benchmark library is established for each target group. The status data of each distribution transformer within the target group is retrieved through the cloud analytics layer, including: Read the abnormal device model from the abnormal event data packet, determine the target group corresponding to the abnormal device based on the abnormal device model, and retrieve the status data of other distribution transformers from the operating characteristic benchmark library of the target group.
[0046] In this embodiment, "segmentation of the same commissioning age" refers to segments whose commissioning age falls within the same timeframe, such as 0-3 years, 3-8 years, and over 8 years. "Target group" refers to a set of distribution transformers with highly similar operating characteristics selected from all distributed distribution transformers within a target area, categorized by group labels. "Operating characteristic benchmark library" refers to a dataset for each target group, built upon the status data uploaded by each distribution transformer within that target group under normal operating conditions, used to characterize the typical values and fluctuation ranges of various operating characteristic parameters of that target group under normal operating conditions.
[0047] In this embodiment, the cloud analysis layer reads the equipment model of each distributed distribution transformer, retrieves the commissioning date of each transformer from local storage, calculates its service life, and categorizes it according to the aforementioned segmentation rules. It also retrieves the installation area information of each transformer from local storage, calculates the average load rate of each transformer during peak hours over the past 7 days from received historical load rate data, and determines its load condition classification according to the aforementioned classification rules. The cloud analysis layer uses the equipment model, service life segmentation, installation area category, and load condition classification information as group classification tags for each transformer.
[0048] The grouping labels of all distributed distribution transformers in the target area are compared. All transformers with the same equipment model, the same commissioning year segment, the same installation area category, and the same load condition classification are clustered into the same set, and a unique group identifier is assigned to this set to complete the division of the target group.
[0049] This embodiment extracts all normal status data for each distribution transformer within a target group over the past 30 days (i.e., the preset statistical period) from the stored normal status data for each target group. For each operational characteristic parameter, the data is grouped by hourly time periods (dividing a day into 24 time periods). Within each time period, the arithmetic mean μ and standard deviation σ of all normal status data points for all transformers in the group are calculated by the cloud analysis layer, and the fluctuation range of the normal parameter is calculated according to the setting of μ±3σ. The calculated mean μ, standard deviation σ, normal parameter fluctuation range, and update time of each parameter are stored in the operational characteristic benchmark library using the target group identifier and time period as a joint index.
[0050] Furthermore, after the cloud analysis layer completes the initial establishment of the benchmark database for each target group, it multiplies the upper limit of the normal parameter fluctuation range for each operating characteristic parameter in each group by 1.5 times, using this as the edge layer early warning threshold for the corresponding parameter in that group. This threshold is then sent to the edge processing layer of each distributed distribution transformer within that group via a preset downlink communication interface. Upon receiving this threshold, the edge processing layer updates its locally stored preset fault discrimination threshold with the value sent from the cloud, thereby ensuring that the edge layer early warning threshold is wider than the fluctuation range of the normal parameters within the group. Subsequently, every preset update cycle (e.g., 30 days), the cloud analysis layer recalculates the statistical parameters of the benchmark database and sends the latest threshold again at a multiplier of 1.5, thus achieving periodic calibration of the edge layer threshold.
[0051] In this embodiment, after receiving the abnormal event data packet reported by the edge processing layer, the cloud analysis layer reads the device model of the abnormal device from the abnormal event data packet. Based on the device model, it searches and matches it among all target groups to determine the target group corresponding to the abnormal device. After determining the target group, the cloud analysis layer retrieves the normal state data of other distribution transformers in the target group during the same time period as the abnormal trigger time, which are of the same type as the abnormal parameters, and their corresponding normal parameter fluctuation ranges, from the operational characteristic benchmark library of the target group, as a reference benchmark for subsequent comparative diagnosis.
[0052] As can be seen from the above, the group division and benchmark library establishment method provided in this embodiment uses four dimensions—equipment model, service life segmentation, installation area category, and load condition classification—as joint division labels. This ensures that transformers within the same target group have high consistency in inherent design attributes, aging stages, external environmental conditions, and load stress levels, eliminating the coupling interference of individual differences, aging degree differences, environmental differences, and operating condition differences of distribution transformers on operating characteristic parameters. On this basis, the edge processing layer is used to divide a dynamic normal fluctuation range for each target group, effectively avoiding the inherent defect that a single fixed threshold cannot adaptively adjust with changes in operating conditions.
[0053] In one embodiment of this application, comparing and diagnosing status data with abnormal event data packets includes: Read the device model, abnormal parameter type, and real-time value of the abnormal parameter from the abnormal event data packet; Based on the equipment model of the abnormal equipment, determine the target group corresponding to the abnormal equipment, and retrieve the same type of status data of all distribution transformers in the same group during the same time period from the operation characteristic benchmark library of the target group; Calculate the mean and standard deviation of each type of state data, and determine the normal parameter fluctuation range of the target group based on the mean and standard deviation; Diagnose the problem by comparing the real-time values of abnormal parameters with the fluctuation range of normal parameters; This includes comparing and diagnosing abnormal parameters with the fluctuation range of normal parameters in real time, including: If the real-time value of the abnormal parameter falls within the normal parameter fluctuation range, then the abnormal event data packet is marked as a false alarm. If the real-time value of the abnormal parameter does not fall within the normal parameter fluctuation range, it is determined that the distributed distribution transformer has a fault in the equipment itself. Based on the type of abnormal parameter and the degree of deviation of the real-time value, the fault type and cause corresponding to the abnormal parameter are determined.
[0054] In this embodiment, after receiving the abnormal event data packet reported by the edge processing layer, the cloud analysis layer reads the device model (e.g., "S11-M-800 / 10"), abnormal parameter type (e.g., "corrected winding temperature change rate"), and the real-time value of the abnormal parameter (e.g., "3.8℃ / min") from the abnormal event data packet.
[0055] Based on the read abnormal device model, a search and matching process is performed across all target groups to determine the target group corresponding to the abnormal device (e.g., "S11-M-800 / 10_Youthful_Outdoor_Medium Load"). After determining the target group, all status data with the same abnormal parameter type within the same time period as the abnormality trigger time (e.g., 10:00 AM to 11:00 AM) are retrieved from the operational characteristic benchmark library of that target group. This includes normal status data of the same type for other distribution transformers within the same group during the same historical time period.
[0056] Then, the arithmetic mean and standard deviation of the retrieved data of the same type are calculated, and the normal parameter fluctuation range of the target group during that time period is determined based on the mean and standard deviation. The real-time values of abnormal parameters are compared with the normal parameter fluctuation range for diagnosis. Specifically: If the real-time value of the abnormal parameter falls within the normal parameter fluctuation range, for example, the corrected winding temperature change rate of the abnormal transformer is 1.5℃ / min, which is within the preset fault discrimination threshold (e.g., [-0.7℃ / min, 2.3℃ / min]), then the abnormal event data packet is marked as a false alarm. This indicates that although the parameter of the distribution transformer exceeds its own preset fault discrimination threshold, it is still within the normal fluctuation range of the same type of equipment in the same group, and is not a real equipment failure.
[0057] If the real-time value of an abnormal parameter does not fall within the normal parameter fluctuation range, for example, if the corrected winding temperature change rate of the abnormal transformer is 3.8℃ / min, which is greater than the upper limit of the range of 2.3℃ / min, then the distributed distribution transformer is determined to have a fault in its equipment. In this case, based on the type of abnormal parameter and the degree of deviation of the real-time value, the fault type and cause corresponding to the abnormal parameter are determined.
[0058] As can be seen from the above, this embodiment reads abnormal device information from abnormal event data packets and determines its target group. It then retrieves similar status data from other devices within the same group during the same time period as a diagnostic reference. A normal parameter fluctuation range is constructed using the average and standard deviation obtained from a large number of normal samples. Whether the real-time abnormal value falls within this range is used as the basis for judging the true fault. This solves the problem of misjudgment caused by the traditional fixed threshold method ignoring individual device differences. Furthermore, for anomalies confirmed as true faults, the fault type and cause are further determined based on the degree of deviation. This upgrades the diagnostic result from a simple "whether it is abnormal" to a multi-level output that "determines the anomaly category, fault type, and severity," improving the accuracy and operability of cloud-based diagnosis and solving the problems of frequent false alarms and insufficient diagnostic information.
[0059] S104: Obtain the historical status data of the distributed distribution transformer and predict the trend of the status data based on the historical status data.
[0060] In this embodiment, based on the equipment model of the distributed distribution transformer, a search operation is performed in the historical database to extract all historical status data records of the distribution transformer within a preset time range. The preset time range refers to a specified number of days tracing back from the current moment.
[0061] In one embodiment of this application, predicting the trend of state data based on historical state data includes: The basic trend prediction data of the distributed distribution transformer within a preset number of days is retrieved from historical status data. The basic trend prediction data includes winding temperature data, core temperature data, vibration amplitude data, vibration frequency data, and partial discharge data. Based on the trend prediction data, the rate of change and fluctuation of each core monitoring parameter within a preset number of days are determined; the core monitoring parameters include winding temperature, core temperature, vibration amplitude, vibration frequency and partial discharge. Based on the rate of change and the amplitude of fluctuation, predict the trend of state data changes over a preset period of time in the future.
[0062] In this embodiment, the trend prediction base data refers to the basic dataset extracted from historical state data, used to analyze the changing patterns of various operating characteristic parameters. The core monitoring parameters refer to the physical quantities extracted from the trend prediction base data, used to characterize the key operating states of the transformer. The future preset duration refers to the future time period covered by the trend prediction, i.e., how long from the current moment it is necessary to predict the trend of state data changes.
[0063] In this embodiment, after obtaining the historical status data of the distributed distribution transformer through the cloud analysis layer, the basic data for trend prediction of the distributed distribution transformer in the past preset number of days (e.g., 7 days) is retrieved from the historical status data. Specifically, the data includes winding temperature data, core temperature data, vibration amplitude data, vibration frequency data, and partial discharge data. The retrieved data is arranged in chronological order to form a time series dataset of each parameter.
[0064] Based on trend prediction data, the rate of change and fluctuation amplitude of each core monitoring parameter over 7 days are determined. Taking winding temperature as an example, the cloud analysis layer compiles a sequence of winding temperature values at the same time each day (e.g., 10:00 AM) over 7 days. The least squares method is used to linearly fit this sequence, and the slope of the fitted line represents the rate of change of the winding temperature. A positive slope indicates an upward trend in winding temperature, while a negative slope indicates a downward trend. The maximum and minimum values of the winding temperature data over 7 days are also determined, and the difference between the maximum and minimum values is taken as the fluctuation amplitude.
[0065] Based on the rate of change and fluctuation amplitude of each core monitoring parameter, the trend of change of each state data within a preset time period (e.g., the next 24 hours) is predicted. Specifically, starting from the real-time value of each core monitoring parameter at the current moment, the rate of change is used as the slope for linear extrapolation to obtain the predicted value of each parameter for the next 24 hours. The fluctuation amplitude is added to the prediction result as the uncertainty range of the predicted value. For example, if the current real-time value of the winding temperature is 85.0℃, the rate of change is +0.5℃ / day, and the fluctuation amplitude is ±2.0℃, then the predicted winding temperature after 24 hours is approximately 85.5℃ (current real-time value of winding temperature + rate of change × number of days), and the predicted fluctuation range is 83.5℃ to 87.5℃.
[0066] S105: The edge processing layer updates the preset fault discrimination threshold based on the diagnostic results and the trend of status data, and updates the status label of the distributed distribution transformer.
[0067] In one embodiment of this application, updating a preset fault discrimination threshold based on diagnostic results and trends in status data includes: Acquire the trend of status data changes over a preset time period; If the diagnosis result is a fault in the equipment itself, the preset fault discrimination threshold is adjusted according to the fault type, fault cause and trend, and the preset threshold adjustment rules. If the diagnosis result is a false alarm, the preset fault discrimination threshold remains unchanged.
[0068] In this embodiment, the edge processing layer receives diagnostic results (including conclusions of equipment failure or false alarms, as well as information such as fault type and fault cause) from the cloud analysis layer, and obtains the trend data of the changes of each core monitoring parameter of the distributed distribution transformer in the next 24 hours.
[0069] If the diagnosis indicates a fault in the device itself, the fault type is further determined through the edge processing layer. Specifically: If the fault type is a thermal fault (corresponding to an abnormal rate of temperature change), and the trend shows that the parameter will continue to rise in the next 24 hours, then the local temperature change rate preset fault judgment threshold will be increased according to the preset threshold adjustment rules, and the updated threshold will be stored in the local configuration file.
[0070] If the fault type is mechanical fault (corresponding to abnormal vibration main frequency peak value), and the trend shows that the vibration amplitude will increase in the next 24 hours, then the local vibration main frequency peak value upper limit threshold will be increased, and the updated threshold will be stored in the local configuration file.
[0071] In this embodiment, if the diagnostic result is "equipment body failure" and the failure type is thermal or mechanical, then the rate of change is predicted based on the next 24 hours. v Update the preset fault detection threshold. The update method can be: Updated preset fault detection threshold = Preset fault detection threshold + v × T × γ ,in, T Indicates the forecast duration (unit: days). γ This represents the preset safety margin coefficient (1.5 for thermal failures and 1.2 for mechanical failures); and the updated threshold must not exceed +3σ of the average value of the group to which the device belongs. If it exceeds this range, it will be truncated with the group average value +3σ as the upper limit.
[0072] If the fault type is an insulation fault (corresponding to an abnormal partial discharge mean), the upper limit threshold of the partial discharge mean will remain unchanged, and the current diagnostic result will be marked as "insulation degradation alarm".
[0073] If the trend in the diagnostic results indicates that the parameter will decrease or stabilize in the future, the threshold adjustment operation will not be performed, and the original threshold will remain unchanged.
[0074] If the diagnosis result is a false alarm, the preset fault discrimination threshold remains unchanged, no threshold correction operation is performed, and the false alarm event and its corresponding diagnosis conclusion are recorded.
[0075] After completing the above threshold correction or maintenance operation, this embodiment updates the status label of the distributed distribution transformer according to the diagnosis results. For example, it can be marked as "thermal fault warning", "mechanical fault warning", "insulation deterioration alarm" or maintain the "normal" status label according to the fault type, so that maintenance personnel can quickly understand the current status of the equipment when querying locally or viewing in the cloud.
[0076] As can be seen from the above, this embodiment achieves adaptive correction of the preset fault discrimination threshold by acquiring the change trend of status data over the next 24 hours and combining it with the diagnostic conclusions of the cloud analysis layer. For thermal and mechanical faults, the updated corresponding thresholds are obtained according to the update formula, so that the new thresholds can adapt to the normal parameter drift caused by equipment aging or changes in operating conditions. For insulation faults, considering that an increase in partial discharge directly reflects a decline in insulation performance, a strategy of maintaining the threshold unchanged and marking alarms is adopted, which reduces invalid alarms while ensuring that the real risks are not masked. For false alarms, the threshold remains unchanged to avoid threshold distortion caused by a single accidental event. This embodiment enables the preset fault discrimination threshold to dynamically adapt to the actual operating status of each transformer, effectively reducing the false alarm rate and the missed alarm rate, and improving the adaptive capability and long-term operational reliability of the condition monitoring system.
[0077] Based on the same principle as the state monitoring method for distributed distribution transformers provided in the embodiments of this application, the embodiments of this application also provide a state monitoring system for distributed distribution transformers, such as... Figure 2 As shown, the state monitoring system 20 for the distributed distribution transformer may specifically include: a state data acquisition module 21, a characteristic parameter comparison module 22, a diagnostic result acquisition module 23, a trend acquisition module 24, and a state monitoring module 25.
[0078] Among them, the status data acquisition module 21 is used to acquire the status data corresponding to each distributed distribution transformer in the target area through the terminal perception layer, and send the status data to the edge processing layer. The feature parameter comparison module 22 is used to extract operating feature parameters from the status data corresponding to each distributed distribution transformer through the edge processing layer, and compare the operating feature parameters with the preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer according to the preset period. If the comparison result is an abnormal state, a local early warning is triggered and an abnormal event data packet is generated and sent to the cloud analysis layer. The diagnostic result acquisition module 23 is used to retrieve the status data of each distribution transformer in the target group through the cloud analysis layer, and compare the status data with the abnormal event data packet to obtain the diagnostic result; Trend acquisition module 24 is used to acquire historical status data of the distributed distribution transformer and predict the trend of the status data based on the historical status data. The status monitoring module 25 is used to update the preset fault discrimination threshold based on the diagnostic results and the trend of status data through the edge processing layer, and to update the status label of the distributed distribution transformer.
[0079] In one embodiment of this application, when retrieving status data of each distribution transformer within a target group through a cloud analysis layer, the diagnostic result acquisition module 23 is specifically used for: The cloud-based analytics layer categorizes each distributed distribution transformer into groups based on its equipment model, years of operation, installation area, and load condition. Based on the group classification labels, distributed transformers with the same equipment model, the same commissioning period, the same installation area, and the same load conditions are classified into the same target group. Based on the normal status data of each distributed distribution transformer received by the cloud analysis layer, establish the corresponding operation characteristic benchmark library for each target group; Read the abnormal device model from the abnormal event data packet, determine the target group corresponding to the abnormal device based on the abnormal device model, and retrieve the status data of other distribution transformers from the operating characteristic benchmark library of the target group.
[0080] In one embodiment of this application, when predicting the trend of state data based on historical state data, the trend acquisition module 24 is specifically used for: The basic trend prediction data of the distributed distribution transformer within a preset number of days is retrieved from historical status data. The basic trend prediction data includes winding temperature data, core temperature data, vibration amplitude data, vibration frequency data, and partial discharge data. Based on the trend prediction data, the rate of change and fluctuation of each core monitoring parameter within a preset number of days are determined; the core monitoring parameters include winding temperature, core temperature, vibration amplitude, vibration frequency and partial discharge. Based on the rate of change and the amplitude of fluctuation, predict the trend of state data changes over a preset period of time in the future.
[0081] In one embodiment of this application, when extracting operational characteristic parameters from the state data corresponding to the distributed distribution transformer through the edge processing layer, the characteristic parameter comparison module 22 is specifically used for: The rate of temperature change is determined based on the winding temperature data and the core temperature data. Determine the peak value of the dominant vibration frequency based on the vibration amplitude data and vibration frequency data; Determine the mean value of partial discharge based on the partial discharge quantity data; Based on the three-phase voltage and three-phase current data in the status data, determine the voltage imbalance and current imbalance. The real-time load rate is determined based on the three-phase current data and the rated current of the distributed distribution transformer.
[0082] In one embodiment of this application, when comparing and diagnosing status data with abnormal event data packets, the diagnosis result acquisition module 23 is specifically used for: Read the device model, abnormal parameter type, and real-time value of the abnormal parameter from the abnormal event data packet; Based on the equipment model of the abnormal equipment, determine the target group corresponding to the abnormal equipment, and retrieve the same type of status data of all distribution transformers in the same group during the same time period from the operation characteristic benchmark library of the target group; Calculate the mean and standard deviation of each type of state data, and determine the normal parameter fluctuation range of the target group based on the mean and standard deviation; The abnormal parameter's real-time value is compared with the normal parameter's fluctuation range for diagnosis.
[0083] In one embodiment of this application, when comparing the real-time value of the abnormal parameter with the fluctuation range of the normal parameter for diagnosis, the diagnosis result acquisition module 23 is specifically used for: If the real-time value of the abnormal parameter falls within the normal parameter fluctuation range, then the abnormal event data packet is marked as a false alarm. If the real-time value of the abnormal parameter does not fall within the normal parameter fluctuation range, it is determined that the distributed distribution transformer has a fault in the equipment itself. Based on the type of abnormal parameter and the degree of deviation of the real-time value, the fault type and cause corresponding to the abnormal parameter are determined.
[0084] In one embodiment of this application, when updating the preset fault discrimination threshold based on the diagnostic results and the trend of status data, the status monitoring module 25 is specifically used for: Acquire the trend of status data changes over a preset time period; If the diagnosis result is a fault in the equipment itself, the preset fault discrimination threshold is adjusted according to the fault type, fault cause and trend, and the preset threshold adjustment rules. If the diagnosis result is a false alarm, the preset fault discrimination threshold remains unchanged.
[0085] In one embodiment of this application, the state data further includes environmental state data, and the device further includes a correction module, which is specifically used for: The operating characteristic parameters are corrected based on environmental status data to obtain the corrected operating characteristic parameters.
[0086] In one embodiment of this application, the device further includes a data processing module, which is specifically used for: The status data is cleaned by removing invalid data frames to obtain the cleaned status data. Data interpolation is performed on the cleaned state data according to the acquisition time sequence to obtain the interpolated state data.
[0087] The system of this application embodiment can execute the method provided in this application embodiment. The implementation principle is similar. The actions performed by each module in the system of each embodiment of this application correspond to the steps in the method of each embodiment of this application. For detailed functional descriptions of each module of the system, please refer to the descriptions in the corresponding methods shown above, which will not be repeated here.
[0088] Figure 3 A schematic diagram of the structure of an electronic device to which this application embodiment applies is shown, such as... Figure 3 As shown, the electronic device can be used to implement the methods provided in any embodiment of this application.
[0089] like Figure 3 As shown, the electronic device 300 may primarily include at least one processor 301. Figure 3 The diagram shows components such as a memory 302, a communication module 303, and an input / output interface 304. Optionally, these components can be connected and communicate with each other via a bus 305. It should be noted that... Figure 3 The structure of the electronic device 300 shown is merely illustrative and does not constitute a limitation on the electronic devices to which the methods provided in the embodiments of this application are applicable.
[0090] The memory 302 can be used to store operating systems and applications, etc. The applications can include computer programs that implement the methods shown in the embodiments of this application when invoked by the processor 301, and can also include programs for implementing other functions or services. The memory 302 can be ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices that can store information and computer programs, or it can be EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.
[0091] Processor 301 is connected to memory 302 via bus 305 and implements corresponding functions by calling the application programs stored in memory 302. Processor 301 can be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 301 can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.
[0092] Electronic device 300 can connect to a network via communication module 303 (which may include, but is not limited to, components such as a network interface) to communicate with other devices (such as user terminals or servers) through the network and achieve data interaction, such as sending data to or receiving data from other devices. Communication module 303 may include wired network interfaces and / or wireless network interfaces, meaning the communication module may include at least one of wired or wireless communication modules.
[0093] The electronic device 300 can connect to necessary input / output devices, such as a keyboard and display device, via the input / output interface 304. The electronic device 300 itself may have a display device, and other external display devices can also be connected via the input / output interface 304. Optionally, a storage system, such as a hard drive, can also be connected via the input / output interface 304 to store data from the electronic device 300 into or retrieve data from the storage system, and data can also be stored in the memory 302. It is understood that the input / output interface 304 can be a wired interface or a wireless interface. Depending on the actual application scenario, the device connected to the input / output interface 304 can be a component of the electronic device 300 or an external device connected to the electronic device 300 when needed.
[0094] The bus 305 used to connect the components may include a path for transmitting information between the components. The bus 305 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Depending on its function, the bus 305 may be divided into an address bus, a data bus, a control bus, etc.
[0095] Optionally, for the solution provided in the embodiments of this application, the memory 302 can be used to store a computer program that executes the solution of this application, and the processor 301 runs the computer program. When the processor 301 runs the computer program, it implements the operation of the method or system provided in the embodiments of this application.
[0096] Based on the same principle as the method provided in the embodiments of this application, the embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the corresponding content of the aforementioned method embodiments.
[0097] It should be noted that the terms "first," "second," "third," "fourth," "1," "2," etc. (if present) in this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than that shown in the figures or text descriptions.
[0098] In the embodiments of this application, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.
[0099] It should be understood that although arrows indicate various operation steps in the flowcharts of this application's embodiments, the order in which these steps are implemented is not limited to the order indicated by the arrows. Unless explicitly stated herein, in some implementation scenarios of this application's embodiments, the implementation steps in each flowchart can be executed in other orders as required. Furthermore, some or all steps in each flowchart, based on the actual implementation scenario, may include multiple sub-steps or multiple stages. Some or all of these sub-steps or stages can be executed at the same time, and each sub-step or stage can also be executed at different times. In scenarios where execution times differ, the execution order of these sub-steps or stages can be flexibly configured according to requirements, and this application's embodiments do not limit this.
[0100] The above are only optional implementation methods for some implementation scenarios of this application. It should be noted that for those skilled in the art, other similar implementation methods based on the technical concept of this application, without departing from the technical concept of this application, also fall within the protection scope of the embodiments of this application.
Claims
1. A method for condition monitoring of distributed distribution transformers, characterized in that, Applied to a controller, which includes a terminal perception layer, an edge processing layer, and a cloud analytics layer, the method includes: The terminal perception layer acquires the status data corresponding to each distributed distribution transformer in the target area, and sends the status data to the edge processing layer. For each distributed distribution transformer, the edge processing layer extracts operating characteristic parameters from the status data corresponding to the distributed distribution transformer, and compares the operating characteristic parameters with a preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer according to a preset period. If the comparison result is an abnormal state, a local early warning is triggered and an abnormal event data packet is generated, which is then sent to the cloud analysis layer. The cloud analysis layer retrieves the status data of each distribution transformer in the target group, compares the status data with the abnormal event data packet, and obtains the diagnostic results. Obtain historical status data of the distributed distribution transformer, and predict the trend of the status data based on the historical status data; The edge processing layer updates the preset fault discrimination threshold based on the diagnostic results and the trend of the status data, and updates the status label of the distributed distribution transformer.
2. The method as described in claim 1, characterized in that, Before retrieving the status data of each distribution transformer within the target group through the cloud analysis layer, the method further includes: The cloud-based analysis layer uses the equipment model, years of operation, installation area, and load condition of each distributed distribution transformer as group classification tags. Based on the group classification labels, distributed transformers with the same equipment model, the same commissioning period, the same installation area, and the same load conditions are classified into the same target group. Based on the normal status data of each distributed distribution transformer received by the cloud analysis layer, establish a corresponding operating characteristic benchmark library for each target group; The process of retrieving status data of each distribution transformer within the target group through the cloud analysis layer includes: The abnormal device model is read from the abnormal event data packet, the target group corresponding to the abnormal device is determined according to the abnormal device model, and the status data of other distribution transformers are retrieved from the operating characteristic benchmark library of the target group.
3. The method as described in claim 1, characterized in that, The step of predicting the trend of the state data based on the historical state data includes: The trend prediction basic data of the distributed distribution transformer within a preset number of days is retrieved from the historical status data. The trend prediction basic data includes winding temperature data, core temperature data, vibration amplitude data, vibration frequency data, and partial discharge data. Based on the trend prediction data, the rate of change and fluctuation amplitude of each core monitoring parameter within the preset number of days are determined; the core monitoring parameters include winding temperature, core temperature, vibration amplitude, vibration frequency, and partial discharge quantity. Based on the rate of change and the amplitude of fluctuation, the trend of change of the state data within a preset time period in the future is predicted.
4. The method as described in claim 3, characterized in that, The extraction of operational characteristic parameters from the status data corresponding to the distributed distribution transformer through the edge processing layer includes: The rate of temperature change is determined based on the winding temperature data and the core temperature data. Based on the vibration amplitude data and the vibration frequency data, the peak value of the dominant vibration frequency is determined; Based on the partial discharge quantity data, the average partial discharge value is determined; Based on the three-phase voltage data and three-phase current data in the state data, determine the voltage imbalance and current imbalance. The real-time load rate is determined based on the three-phase current data and the rated current of the distributed distribution transformer.
5. The method as described in claim 1, characterized in that, The step of comparing and diagnosing the status data with the abnormal event data packet includes: Read the device model, abnormal parameter type, and real-time value of the abnormal parameter from the abnormal event data packet; Based on the equipment model of the abnormal equipment, determine the target group corresponding to the abnormal equipment, and retrieve the same type of status data of all distribution transformers in the same group within the same time period from the operation characteristic benchmark library of the target group; Calculate the average value and standard deviation of each type of state data, and determine the normal parameter fluctuation range of the target group based on the average value and standard deviation; The abnormal parameter's real-time value is compared with the normal parameter's fluctuation range for diagnosis.
6. The method as described in claim 5, characterized in that, The step of comparing and diagnosing the abnormal parameter's real-time value with the normal parameter's fluctuation range includes: If the real-time value of the abnormal parameter falls within the fluctuation range of the normal parameter, then the abnormal event data packet is marked as a false alarm. If the real-time value of the abnormal parameter does not fall within the fluctuation range of the normal parameter, it is determined that the distributed distribution transformer has a fault in the equipment itself, and the fault type and cause corresponding to the abnormal parameter are determined according to the type of abnormal parameter and the degree of deviation of the real-time value.
7. The method as described in claim 6, characterized in that, The step of updating the preset fault discrimination threshold based on the diagnostic results and the trend of the status data includes: Obtain the trend of the change of the state data within a preset time period in the future; If the diagnostic result is a fault in the device itself, then the preset fault discrimination threshold is corrected according to the fault type, the fault cause, and the trend of change, in accordance with the preset threshold adjustment rules. If the diagnostic result is a false alarm, the preset fault discrimination threshold remains unchanged.
8. The method as described in claim 1, characterized in that, The status data also includes environmental status data; After extracting the operating characteristic parameters from the state data corresponding to the distributed distribution transformer through the edge processing layer, the method further includes: The operating characteristic parameters are corrected based on the environmental state data to obtain the corrected operating characteristic parameters.
9. The method as described in claim 1, characterized in that, After obtaining the status data of each distributed distribution transformer within the target area through the terminal sensing layer, the method further includes: The state data is cleaned to remove invalid data frames and obtain cleaned state data. The cleaned state data is interpolated according to the acquisition time sequence to obtain interpolated state data.
10. A condition monitoring system for distributed distribution transformers, characterized in that, Located within the controller, the system includes: The status data acquisition module is used to acquire the status data corresponding to each distributed distribution transformer in the target area through the terminal perception layer, and send the status data to the edge processing layer. The feature parameter comparison module is used to extract operating feature parameters from the status data corresponding to each distributed distribution transformer through the edge processing layer, and compare the operating feature parameters with a preset fault discrimination threshold. If the comparison result is a normal state, the normal state data is sent to the cloud analysis layer at a preset period. If the comparison result is an abnormal state, a local early warning is triggered and an abnormal event data packet is generated and sent to the cloud analysis layer. The diagnostic result acquisition module is used to retrieve the status data of each distribution transformer in the target group through the cloud analysis layer, and compare the status data with the abnormal event data packet to obtain the diagnostic result; The trend acquisition module is used to acquire historical state data of the distributed distribution transformer and predict the trend of the state data based on the historical state data. The status monitoring module is used to update the preset fault discrimination threshold based on the diagnostic results and the trend of the status data through the edge processing layer, and to update the status label of the distributed distribution transformer.
