Transformer monitoring method and system based on sensor network technology

Abstract

The invention relates to a transformer monitoring method and system based on a sensor network technology. The method comprises the following steps: acquiring an initial physical parameter matrix collected by a heterogeneous sensor cluster corresponding to a target transformer; based on the distribution height of each sensor and the sensing node distance in the heterogeneous sensor cluster, performing space-time normalization processing on the initial physical parameter matrix to obtain a target physical parameter matrix of the target transformer; and determining the operation state of the target transformer according to the target physical parameter matrix of the target transformer. The heterogeneous sensor cluster includes a plurality of different types of sensors. By adopting the method, the accuracy of transformer state monitoring results can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of power systems, and in particular to a transformer monitoring method and system based on sensor network technology. Background Art

[0002] With the development of power systems, the types and number of power equipment in power grid systems have increased dramatically, and the pressure on equipment operation and maintenance has become increasingly prominent. Based on this, monitoring the status of power grid equipment is crucial to the normal operation of the power grid system.

[0003] Taking the transformer in the power grid system as an example, in related technologies, when monitoring the operating status of the transformer, the electrical parameters of the transformer, such as voltage, current or power parameters, are usually collected, and then the collected electrical parameters are statistically analyzed to obtain the operating status of the transformer.

[0004] However, the accuracy of transformer status monitoring results in related technologies is low. Summary of the Invention

[0005] Based on this, it is necessary to provide a transformer monitoring method and system based on sensor network technology that can improve the accuracy of transformer status monitoring results in response to the above technical problems.

[0006] In a first aspect, the present application provides a transformer monitoring method based on sensor network technology, comprising:

[0007] Obtaining an initial physical parameter matrix collected by a heterogeneous sensor cluster corresponding to a target transformer; the heterogeneous sensor cluster includes multiple sensors of different types;

[0008] Based on the distribution height of each sensor in the heterogeneous sensor cluster and the distance between sensor nodes, the initial physical parameter matrix is ​​normalized in time and space to obtain the target physical parameter matrix of the target transformer.

[0009] The operating state of the target transformer is determined according to the target physical parameter matrix of the target transformer.

[0010] In one embodiment, the initial physical parameter matrix includes initial mechanical parameters, initial oil chromatographic parameters, and multiple initial electrical parameters. Based on the distribution heights of the sensors in the heterogeneous sensor cluster and the distances between the sensor nodes, the initial physical parameter matrix is ​​subjected to spatiotemporal normalization to obtain a target physical parameter matrix of the target transformer, including:

[0011] Based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial mechanical parameters, initial oil chromatographic parameters, and initial electrical parameters are temporally and spatially aligned to obtain the corresponding target mechanical parameters, target oil chromatographic parameters, and target electrical parameters.

[0012] The target mechanical parameters, target oil chromatography parameters and target electrical parameters are integrated to obtain the target physical parameter matrix of the target transformer.

[0013] In one embodiment, the target mechanical parameters, target oil chromatographic parameters, and target electrical parameters are integrated to obtain a target physical parameter matrix of the target transformer, including:

[0014] Determine the winding loosening conduction path of the target transformer according to the mapping relationship between the target mechanical parameters and the target oil chromatography parameters;

[0015] According to the preset weights, the target electrical parameters are integrated to obtain the component area identification of the target transformer;

[0016] The winding loosening conduction path of the target transformer and the component area identification are associated to obtain the target physical parameter matrix of the target transformer.

[0017] In one embodiment, determining the operating state of the target transformer according to the target physical parameter matrix of the target transformer includes:

[0018] The target physical parameter matrix of the target transformer is input into a preset state prediction model. The fault state features in the target physical parameter matrix are extracted through the state prediction model, and the operating state of the target transformer is determined based on the associated quantitative values ​​between the fault state features.

[0019] In one embodiment, the method further comprises:

[0020] When the target transformer is in a fault state, the target transformer's alarm information is displayed, including the target transformer's fault location, transformer position, and fault cause.

[0021] In the second aspect, the present application also provides a condition monitoring system, which includes: a processor, a target transformer, and a heterogeneous sensor cluster corresponding to the target transformer; the heterogeneous sensor cluster includes multiple sensors of different types; the processor is used to execute the steps of the method in any one of the embodiments of the first aspect above.

[0022] In one embodiment, a surface of each sensor in a heterogeneous sensor cluster includes a shielding composite layer.

[0023] In a third aspect, the present application further provides a condition monitoring device, comprising:

[0024] A data acquisition module is used to obtain the initial physical parameter matrix collected by the heterogeneous sensor cluster corresponding to the target transformer; the heterogeneous sensor cluster includes multiple sensors of different types;

[0025] The data processing module is used to perform spatiotemporal normalization on the initial physical parameter matrix based on the distribution height of each sensor in the heterogeneous sensor cluster and the distance between sensor nodes to obtain the target physical parameter matrix of the target transformer;

[0026] The state determination module is used to determine the operating state of the target transformer according to the target physical parameter matrix of the target transformer.

[0027] In a fourth aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method in any one of the embodiments of the first aspect are implemented.

[0028] In a fifth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in any one of the embodiments of the first aspect above.

[0029] In a sixth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the method in any one of the embodiments of the first aspect above.

[0030] The above-mentioned transformer monitoring method and system based on sensor network technology first obtains an initial physical parameter matrix collected by a heterogeneous sensor cluster corresponding to the target transformer. Then, based on the distribution heights and sensor node distances of each sensor in the heterogeneous sensor cluster, the initial physical parameter matrix is ​​subjected to spatiotemporal normalization to obtain a target physical parameter matrix for the target transformer. Finally, the operating status of the target transformer is determined based on the target physical parameter matrix of the target transformer. The heterogeneous sensor cluster includes multiple different types of sensors. In this method, the physical parameters of the target transformer are collected using multiple different types of sensors to obtain the initial physical parameter matrix, providing a diversified reference for the condition monitoring of the target transformer. Furthermore, the initial physical parameter matrix of the target transformer is subjected to spatiotemporal normalization based on the distribution heights and sensor node distances of the sensors. This achieves the coordinated capture of multiple physical quantities, such as electromagnetic, mechanical, and thermodynamic quantities, avoiding the existence of isolated data in the transformer monitoring data. This provides a comprehensive and reliable data base for transformer condition monitoring, improves the accuracy of the target physical parameter matrix, and thus improves the accuracy of the results of determining the operating status of the target transformer based on the target physical parameter matrix. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0032] Figure 1 2. A diagram showing an application environment of a transformer monitoring method based on sensor network technology in one embodiment;

[0033] Figure 2 1 is a flow chart of a transformer monitoring method based on sensor network technology in one embodiment;

[0034] Figure 3 Schematic diagram of a process for obtaining a parameter matrix in one embodiment;

[0035] Figure 4 A schematic flow chart of a parameter matrix acquisition step in another embodiment;

[0036] Figure 5 Schematic diagram of a transformer monitoring method based on sensor network technology in one embodiment;

[0037] Figure 6 A schematic diagram of the architecture of a condition monitoring system in one embodiment;

[0038] Figure 7 is a structural block diagram of a condition monitoring device in one embodiment;

[0039] Figure 8 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0041] With the transformation of power systems, transformers, as key nodes in power network topologies, face multiple coupled risks in their safe operation. The high volatility of renewable energy sources triggers atypical load shocks, causing winding stress distribution to deviate from design baseline values. High-frequency harmonic interference exacerbates core vibration, increasing the rate of electrothermal aging of insulation materials by 2-3 times. Bidirectional interaction between source and load causes abnormal oil temperature gradients, increasing the false negative rate of traditional fault diagnosis models. These characteristics, combined with the transformer's inherent multi-physics coupling (electromagnetic, thermal, and mechanical synergy), can induce hidden defects such as winding deformation and insulation breakdown.

[0042] In practical applications, transformers, as the most core equipment in power grid systems, have a direct impact on power supply reliability and power quality. Transformer operation, especially in the UHV sector, often faces high AC / DC and composite electric field harmonics and heavy workloads, placing stringent demands on equipment performance and safe operation. Furthermore, with the completion and commissioning of numerous UHV AC / DC projects, the number of transformers in operation has increased dramatically, placing increasing pressure on equipment operation and maintenance. Transformer operational failures can cause grid paralysis, equipment burnout, and other accidents.

[0043] Taking into account the problem of low accuracy of transformer status monitoring results in the existing technology, the present application provides a transformer monitoring method and system based on sensor network technology. By establishing a transformer monitoring system based on sensor network technology, the perception accuracy of the transformer operating status is enhanced, multi-dimensional integrated perception of the transformer is realized, and a technological leap from discrete parameter monitoring to system health assessment is achieved, thereby improving the accuracy of transformer status monitoring results.

[0044] The transformer monitoring method based on sensor network technology provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown in FIG. , the processor 102 interacts with the oil gas concentration sensor probe 104, the infrared temperature sensor probe 106, the partial discharge monitoring sensor probe 108, the vibration sensor probe 110, and the noise sensor probe 112 through the network to receive the transformer physical parameters sent by each sensor node. Figure 1 It can be seen that due to factors such as terrain, the distribution heights of the sensor nodes may be inconsistent, and the distances between the sensor nodes may be inconsistent. The processor 102 may process the received information to determine the operating status of the target transformer.

[0045] For example, processor 102 transmits information to the data analysis and intelligent diagnosis platform, which then evaluates the transformer's operating status and generates data reports such as the transformer's vibration spectrum, thereby digitizing the transformer. The data analysis and intelligent diagnosis platform can also issue instructions to processor 102 through the data analysis and intelligent diagnosis platform to control each sensor node. The data analysis and intelligent diagnosis platform can also store information such as the operating status evaluation report or transmit it to a terminal, where it can display the operating status evaluation information. Terminals include, but are not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projectors. Portable wearable devices can include smart watches, smart bracelets, head-mounted devices, and the like. Head-mounted devices can include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, and the like.

[0046] In an exemplary embodiment, Figure 2 As shown, a state monitoring method is provided, which is applied to Figure 1 The following steps are taken as an example to illustrate the processor in the example:

[0047] S201 , obtaining an initial physical parameter matrix collected by a heterogeneous sensor cluster corresponding to a target transformer; the heterogeneous sensor cluster includes multiple sensors of different types.

[0048] A transformer is a device that transforms voltage, current, and impedance through electromagnetic mutual induction. A transformer consists of a primary coil, a secondary coil, and an iron core (or magnetic core) and is used for voltage conversion, current conversion, impedance conversion, electrical isolation, and voltage stabilization. Transformers in power grid systems can be categorized by their intended use as distribution transformers, rectifier transformers, oil-immersed transformers, and so on. The target transformer in the embodiments of this application refers to any transformer in a power grid system.

[0049] It should be noted that transformers are multi-physics field coupling devices, and their fault characteristics exhibit significant systemic characteristics. For example, winding deformation can change the local electric field distribution and induce polarization loss in the insulating medium; mechanical vibration generated by loose core clamps can accelerate the degradation of the oil-paper composite insulation interface, etc. Based on this, in order to collect more comprehensive operating data, the embodiments of this application use multiple different types of sensors to collect physical parameters of the same target transformer, such as partial discharge data collected by electrical parameter sensors, winding vibration data collected by mechanical sensors, and oil chromatography data collected by chemical sensors.

[0050] After obtaining the physical parameters of the target transformer collected by each sensor, multiple groups of physical parameters of the target transformer are summarized according to the type of physical parameters to obtain an initial physical parameter matrix corresponding to the target transformer.

[0051] Optionally, for each sensor, the initial physical parameters collected by the sensor are sorted according to the reception time to obtain a physical parameter sequence corresponding to the sensor. Then, multiple groups of physical parameter sequences corresponding to each sensor are aggregated to obtain an initial physical parameter matrix.

[0052] Optionally, multiple different types of initial physical parameters of the target transformer at the same time are taken as a set of sequences to obtain the physical parameter sequences of the target transformer at different times, and then the physical parameter sequences at different times are summarized according to the receiving time to obtain the initial physical parameter matrix.

[0053] S202 , based on the distribution height of each sensor in the heterogeneous sensor cluster and the sensor node distance, perform spatiotemporal normalization processing on the initial physical parameter matrix to obtain a target physical parameter matrix of the target transformer.

[0054] It should be noted that due to factors such as terrain, the distribution height of sensor nodes may be inconsistent, which means that the physical parameters collected by different sensors may deviate due to differences in altitude. Furthermore, the distance between sensor nodes and the processor may be inconsistent, which means that the processor may experience a delay in the acquisition time of the physical parameters collected by each sensor. Spatiotemporal normalization refers to the correction of various physical parameters of the target transformer based on both time and space to ensure temporal and spatial consistency of different physical parameters.

[0055] When performing spatiotemporal normalization on the initial physical parameter matrix, the initial physical parameter matrix can be first spatially normalized based on the distribution height of each sensor in the heterogeneous sensor cluster, and then the spatially normalized matrix can be temporally normalized based on the sensor node distance of each sensor in the heterogeneous sensor cluster to obtain the target physical parameter matrix of the target transformer; or the initial physical parameter matrix can be first temporally normalized based on the sensor node distance of each sensor in the heterogeneous sensor cluster, and then the spatially and temporally normalized matrix can be spatially normalized based on the distribution height of each sensor in the heterogeneous sensor cluster to obtain the target physical parameter matrix of the target transformer.

[0056] When spatially correcting the initial physical parameter matrix, the distributed heights of each sensor can be compared with the actual elevation of the target transformer to determine the target sensors that are aligned with the target transformer and the sensors to be adjusted that are not aligned with the target transformer. The physical parameters collected by the target sensors remain unchanged in the initial physical parameter matrix, and the physical parameters of the sensors to be adjusted are proportionally processed based on the ratio of their distributed heights to the actual elevation of the target transformer, resulting in a spatially corrected physical parameter matrix.

[0057] When performing time correction on the initial physical parameter matrix, the actual acquisition time of the physical parameters collected by each sensor can be determined based on the physical distance between each sensor and the processor in the heterogeneous sensor cluster. Then, the physical parameters collected by different sensors are time-aligned to obtain the time-corrected physical parameter matrix.

[0058] S203: Determine the operating state of the target transformer according to the target physical parameter matrix of the target transformer.

[0059] The operating status of the target transformer may include the current status of the target transformer and / or the predicted status of the target transformer at a certain moment in the future.

[0060] Optionally, the operating state of the target transformer refers to a current state. Different types of parameters in the target physical parameter matrix are compared with preset parameter threshold intervals. If the parameters in the target physical parameter matrix are not within the corresponding parameter threshold intervals, it is determined that the target transformer is currently in a fault state; otherwise, it is determined that the target transformer is currently in a normal operating state.

[0061] Optionally, the operating state of the target transformer refers to a predicted state. In one embodiment, determining the operating state of the target transformer based on a target physical parameter matrix of the target transformer includes: inputting the target physical parameter matrix of the target transformer into a preset state prediction model, extracting fault state features from the target physical parameter matrix using the state prediction model, and determining the operating state of the target transformer based on correlation quantization values ​​between the fault state features.

[0062] The state prediction model is trained based on a historical dataset and is used to determine the predicted state of the target transformer at a future time. The historical dataset includes the transformer's physical parameters during a first historical time period and its operating status during a second historical time period, where the first historical time period is earlier than the second historical time period.

[0063] The determination logic of the state prediction model can be: screen out abnormal parameters in the target physical parameter matrix through the state prediction model to obtain at least one group of fault state characteristics, then calculate the correlation quantitative values ​​between the fault state characteristics of different groups, such as the Pearson correlation coefficient, and then determine the operating state corresponding to the correlation quantitative value based on the mapping relationship between the preset correlation quantitative value and the operating state as the operating state of the target transformer.

[0064] In an embodiment of the present application, a preset state prediction model is used to extract features from the target physical parameter matrix of the target transformer to identify the operating conditions of the internal main components, such as discharge and overheating, during the operation of the transformer, and then determine the predicted state of the target transformer to achieve early warning of potential equipment failures.

[0065] In an exemplary embodiment, the method further includes:

[0066] When the target transformer is in a fault state, the target transformer's alarm information is displayed, including the target transformer's fault location, transformer position, and fault cause.

[0067] When the operating status represents the current status, if the target transformer is currently in a fault state, the first alarm information is displayed. The first alarm information includes the fault location of the target transformer to facilitate subsequent maintenance; it also includes the transformer location to facilitate locating the faulty transformer; it also includes the cause of the fault to facilitate investigation of related fault influencing factors and prevent the fault from further expanding.

[0068] In the case where the operating status represents the future predicted status, if the target transformer is currently in a fault state, the second alarm information is displayed. The second alarm information includes the fault location of the target transformer to facilitate subsequent maintenance; the transformer location to facilitate locating the faulty transformer; and the cause of the fault to facilitate investigation of related fault influencing factors to prevent further expansion of the fault.

[0069] In an embodiment of the present application, when the operating state of the target transformer is in a fault state, the fault location, transformer position and fault cause of the target transformer are displayed, and an alarm is issued in time, thereby quickly restoring the normal operation of the transformer and avoiding equipment loss.

[0070] In an embodiment of the present application, the initial physical parameter matrix collected by the heterogeneous sensor cluster corresponding to the target transformer is first obtained, and then the initial physical parameter matrix is ​​subjected to spatiotemporal normalization based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster to obtain the target physical parameter matrix of the target transformer. Finally, the operating state of the target transformer is determined based on the target physical parameter matrix of the target transformer, wherein the heterogeneous sensor cluster includes multiple sensors of different types. In this method, multiple sensors of different types are used to collect the physical parameters of the target transformer to obtain the initial physical parameter matrix, which provides a diversified reference basis for the state monitoring of the target transformer. On this basis, based on the distribution height and sensor node distance of the sensors, the initial physical parameter matrix of the target transformer is subjected to spatiotemporal normalization to achieve the coordinated capture of multiple physical quantities such as electromagnetic, mechanical, and thermodynamics, avoid the existence of isolated data in the transformer monitoring data, provide a comprehensive and reliable data base for transformer state monitoring, improve the accuracy of the target physical parameter matrix, and thus improve the accuracy of the operating state result determined according to the target physical parameter matrix.

[0071] In an exemplary embodiment, the initial physical parameter matrix includes initial mechanical parameters, initial oil chromatography parameters and a plurality of initial electrical parameters; Figure 3 As shown, in the aforementioned embodiment, S202 “performing spatiotemporal normalization processing on the initial physical parameter matrix based on the distribution height of each sensor in the heterogeneous sensor cluster and the sensor node distance to obtain the target physical parameter matrix of the target transformer” includes the following steps:

[0072] S301 , based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, perform spatiotemporal alignment processing on the initial mechanical parameters, initial oil chromatographic parameters, and initial electrical parameters to obtain corresponding target mechanical parameters, target oil chromatographic parameters, and target electrical parameters.

[0073] Among them, the initial mechanical parameters include winding temperature, vibration spectrum, bushing dielectric loss parameters, etc.; the initial oil chromatogram parameters include: the concentration of characteristic gases such as H2 and C2H2; the electrical parameters include partial discharge parameters, core grounding current and other parameters.

[0074] The spatiotemporal alignment processing includes spatiotemporal normalization processing and time alignment processing, and the spatiotemporal normalization processing includes time normalization processing and space normalization processing.

[0075] Based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial mechanical parameters, initial oil chromatographic parameters, and initial electrical parameters are spatiotemporally normalized. Then, the physical parameters after spatiotemporal normalization are aligned in the time dimension to obtain the target mechanical parameters, target oil chromatographic parameters, and target electrical parameters.

[0076] In actual application scenarios, the spatiotemporal alignment processing includes filtering processing. In this case, the initial mechanical parameters, initial oil chromatogram parameters and initial electrical parameters can be filtered first, and then the filtered initial mechanical parameters, initial oil chromatogram parameters and initial electrical parameters can be spatiotemporally aligned.

[0077] S302 , fusing target mechanical parameters, target oil chromatographic parameters, and target electrical parameters to obtain a target physical parameter matrix of the target transformer.

[0078] The target mechanical parameters, target oil chromatographic parameters and target electrical parameters are standardized respectively, and the standardized parameters are summarized to obtain the target physical parameter matrix of the target transformer.

[0079] In an embodiment of the present application, based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial mechanical parameters, initial oil chromatogram parameters and initial electrical parameters are sequentially subjected to spatiotemporal alignment and fusion processing to obtain the target physical parameter matrix of the target transformer.

[0080] In an exemplary embodiment, Figure 4 As shown in the figure, the target mechanical parameters, target oil chromatography parameters and target electrical parameters are integrated to obtain the target physical parameter matrix of the target transformer, including:

[0081] S401 : Determine a winding loosening conduction path of a target transformer according to a mapping relationship between target mechanical parameters and target oil chromatography parameters.

[0082] The target mechanical parameters include a mechanical vibration curve, and the target oil chromatogram parameters include an oil chromatogram data curve. The nonlinear harmonic components in the mechanical vibration curve are cross-domain mapped with the oil chromatogram data curve to determine the transmission path from winding looseness to insulation degradation in the target transformer.

[0083] S402: fusing target electrical parameters according to preset weights to obtain component region identifiers of the target transformer.

[0084] The target electrical parameters include the partial discharge phase distribution and the temperature gradient field. By assigning weights to the partial discharge phase distribution and the temperature gradient field and establishing a coupling relationship between them, the system can locate potential faulty components in the target transformer and identify the faulty component's location relative to the target transformer.

[0085] S403 : Associating the winding loosening conduction path of the target transformer with the component area identifier to obtain a target physical parameter matrix of the target transformer.

[0086] For each winding looseness conduction path, the components in the path are replaced with the corresponding component area identifiers according to the conduction order to obtain a set of target physical parameter vectors. Similarly, the target physical parameter vectors corresponding to multiple winding looseness conduction paths are obtained and summarized to obtain the target physical parameter matrix of the target transformer.

[0087] Taking the case where the winding looseness conduction paths include two, namely component A, component B, component C and component D; component E, component F, component G and component H; the component area identification corresponding to component A is 1, the component area identification corresponding to component B is 2, the component area identification corresponding to component C is 3, and the component area identification corresponding to component D is 4; the component area identification corresponding to component E is 5, the component area identification corresponding to component F is 6, and the component area identification corresponding to component H is 7 as an example, the target physical parameter matrix includes two groups of row vectors, the elements of the first row vector are: 1, 2, 3, 4; the elements of the second row vector are: 5, 6, 7, 8.

[0088] In an embodiment of the present application, according to the mapping relationship between the target mechanical parameters and the target oil chromatography parameters, the loose winding conduction path of the target transformer is determined, and according to the preset weights, the target electrical parameters are fused to obtain the component area identification of the target transformer. Then, the loose winding conduction path of the target transformer and the component area identification are associated to obtain the target physical parameter matrix of the target transformer, so as to intuitively represent the actual operating status of each component in the target transformer.

[0089] In one embodiment, a transformer monitoring method based on sensor network technology is provided, comprising:

[0090] (1) Obtain the initial physical parameter matrix collected by the heterogeneous sensor cluster corresponding to the target transformer; the heterogeneous sensor cluster includes multiple sensors of different types; the initial physical parameter matrix includes initial mechanical parameters, initial oil chromatography parameters and multiple initial electrical parameters.

[0091] (2) Based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial mechanical parameters, initial oil chromatographic parameters and initial electrical parameters are spatiotemporally aligned to obtain the corresponding target mechanical parameters, target oil chromatographic parameters and target electrical parameters.

[0092] (3) According to the mapping relationship between the target mechanical parameters and the target oil chromatography parameters, the winding loosening conduction path of the target transformer is determined.

[0093] (4) According to the preset weights, the target electrical parameters are integrated to obtain the component area identification of the target transformer.

[0094] (5) The winding loosening conduction path of the target transformer and the component area identification are associated to obtain the target physical parameter matrix of the target transformer.

[0095] (6) The target physical parameter matrix of the target transformer is input into the preset state prediction model, the fault state characteristics in the target physical parameter matrix are extracted through the state prediction model, and the operating state of the target transformer is determined based on the correlation quantitative values ​​between the fault state characteristics.

[0096] (7) When the target transformer is in a fault state, the alarm information of the target transformer is displayed, including the fault location of the target transformer, the transformer location and the cause of the fault.

[0097] In the embodiment of the present application, sensor network technology is used to complete the integrated acquisition and data fusion of multiple parameters of the transformer, so as to realize comprehensive, real-time and accurate monitoring of the transformer operating status and ultra-early accurate warning and diagnosis of faults. Solve the island problem of transformer fault monitoring data under the transformation of the new power system, perceive the status of the transformer in real time, enhance the perception accuracy of the transformer operating status, and build a digital and transparent transformer. Through the digital monitoring system with deep integration of distributed sensor networks and intelligent communication protocols, based on multi-source heterogeneous sensor clusters, real-time acquisition of multiple types of heterogeneous parameters such as winding vibration, oil chromatography, partial discharge, etc., relying on adaptive networking algorithms to build low-latency data transmission channels, and using deep confidence networks to realize feature decoupling of electromagnetic-mechanical-thermal multi-field data. The transformer monitoring method based on sensor network technology provided in the embodiment of the present application breaks through the limitations of traditional single-point monitoring, and by constructing a three-dimensional spatiotemporal evolution model of the equipment status, the early recognition rate of faults such as insulation aging and mechanical looseness is improved, providing a dynamic quantitative basis for the health management of the transformer throughout its life cycle.

[0098] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0099] Based on the same inventive concept, an embodiment of the present application also provides a condition monitoring system for implementing the above-mentioned transformer monitoring method based on sensor network technology, the system comprising: a processor, a target transformer, and a heterogeneous sensor cluster corresponding to the target transformer; the heterogeneous sensor cluster comprises multiple sensors of different types; the processor is used to execute the steps in the above-mentioned method embodiments.

[0100] See Figure 5 , Figure 5 This is a schematic diagram of the framework of the condition monitoring system, which includes the perception layer, network layer, and application layer. The perception layer is used for data acquisition, the network layer is used for data transmission, and the application layer is used for data analysis and intelligent diagnosis. At the perception layer, multi-source heterogeneous sensor nodes achieve the coordinated capture of multiple physical quantities such as electromagnetic, mechanical, and thermodynamic through a high-precision synchronous acquisition mechanism, solving the problem of isolated transformer monitoring data and breaking through the perception blind spots of traditional single-dimensional monitoring. At the network layer, deterministic network protocols and adaptive anti-interference technology are used to ensure the integrity and real-time performance of high-frequency signals under complex working conditions, providing a reliable data base for subsequent analysis. At the application layer, the intelligent diagnosis core relies on a digital twin model coupled with multiple physical fields. By integrating deep learning algorithms with dynamic probabilistic reasoning, it accurately analyzes implicit state parameters such as winding stress distribution and insulating medium polarization characteristics, and constructs a spatiotemporal evolution model of equipment health.

[0101] from Figure 5 It can be seen that the perception layer in the embodiment of the present application includes a variety of different types of sensor probes, such as partial discharge (PD) sensor probes, current sensor probes, oil level sensor probes, etc., to comprehensively monitor the multi-physical quantities of the transformer. Among them, multi-physical quantities refer to the integrated collection and fusion analysis of multiple physical quantities. The condition monitoring system builds a multi-dimensional perception system through a distributed sensor network, synchronously collecting 12 key parameters such as oil chromatography (characteristic gases such as H2, C2H2), winding temperature, partial discharge, vibration spectrum, bushing dielectric loss, core grounding current, etc., to achieve holographic data capture of multiple physical quantities such as electrical, mechanical, and chemical. At the fusion analysis level, a spatiotemporal registration algorithm is used to eliminate the time delay and spatial deviation of multi-source heterogeneous data. Through feature-level fusion, a multi-state quantity correlation model such as insulation degradation, mechanical deformation, and discharge intensity is established. The digital twin technology is integrated with the long short-term memory network (LSTM) to realize the dynamic deduction of the transformer aging process. Experimental data show that compared with the traditional single-dimensional monitoring method, the accuracy of the transformer health assessment monitored by the condition monitoring system provided in the embodiment of the present application is improved to more than 93%, providing panoramic data support for equipment condition maintenance.

[0102] The framework of the condition monitoring system includes the following technical features: (1) Multi-dimensionality: integrated acquisition and fusion analysis of multiple physical quantities; High integration: perception-computing-transmission-storage multi-function integration; (2) Digitalization: equipped with NPU. Possessing edge intelligent decision-making capabilities; (3) Multimodality: Coordinated processing capabilities of multi-source heterogeneous data; (4) MEMS: Using micro-nano processing technology to prepare high-precision, wide-range, wide-frequency domain, and high-consistency sensor perception probes; (5) Collaboration: Multi-dimensional data fusion analysis to jointly describe the equipment status; (6) High reliability: Using advanced process material technology to improve the reliable service time of sensors in complex environments with high voltage and strong magnetic fields. Below, the implementation principles of the above features are explained separately:

[0103] (1) Multi-dimensional: integrated acquisition and fusion analysis of multiple physical quantities. The integrated acquisition and fusion analysis technology of multiple physical quantities of transformers based on sensor network technology realizes panoramic perception of equipment status by constructing a "space-time-parameter" three-dimensional monitoring system. In the spatial dimension, distributed optical fiber temperature measurement, ultrasonic array sensing and multi-probe networking technology of dissolved gas in oil are used to cover key parts such as the transformer body, bushing, cooling system, etc., and synchronously collect 12 types of physical parameters such as winding hot spot temperature, partial discharge ultrasonic signal, oil chromatogram (8 characteristic gases such as H2, C2H2, CO), forming a three-dimensional monitoring matrix of electrical, mechanical and chemical parameters. At the analytical level, the improved Kalman filter algorithm is used to eliminate the temporal and spatial benchmark differences of multiple sensors, establish a multi-scale correlation model between characteristic parameters such as insulation aging index, mechanical stress distribution, and thermal balance coefficient, and combine with the transfer learning algorithm to realize cross-dimensional mapping of equipment degradation modes under different operating conditions. Engineering measurements show that this technology upgrades the traditional single-parameter threshold alarm to multi-parameter coupling diagnosis, increases the defect identification dimension, and provides a multi-dimensional data fusion analysis paradigm for the full life cycle management of transformers.

[0104] (2) High integration: perception-computation-transmission-storage multifunctional integration. The transformer monitoring system based on sensor network technology has achieved a major breakthrough in functional integration. Its core feature is the deep integration of the four major functional modules of perception, calculation, transmission and storage into the system. At the hardware level, heterogeneous integrated packaging technology is used to embed MEMS manufacturing process vibration sensors, broadband pulse current detection units and LoRa / 5G dual-mode communication modules in the miniaturized structure to achieve synchronous acquisition and processing of multiple signals such as vibration signals, leakage currents, and partial discharge pulses. The segmented compression storage of the original data stream is achieved through adaptive dynamic framing technology. Combined with the ring buffer design, the storage efficiency is improved by 60% compared with the traditional architecture. It can still operate continuously for 72 hours without external power supply, providing a highly reliable and low-power integrated solution for real-time status perception of transformers in complex electromagnetic environments.

[0105] (3) Digitalization: Equipped with a neural network processor (NPU). Possessing edge intelligent decision-making capabilities. The transformer monitoring system based on sensor network technology achieves edge intelligent leap in digital upgrade. Its core feature is the integration of NPU to build a closed-loop system of "end-side perception-real-time decision-making". The hardware architecture adopts a heterogeneous computing design, embedding an 8-core NPU chip (computing power of 4TOPS) in the edge node, equipped with a lightweight inference engine, which can process multi-dimensional heterogeneous data streams such as vibration spectrum, partial discharge pulse, and dissolved gas concentration in oil in parallel. By pre-setting the depthwise separable convolutional network (DCS) and the gated recurrent unit (GRU) fusion model, millisecond-level online diagnosis of multiple typical working conditions such as partial discharge pattern recognition and winding hot spot temperature prediction can be achieved. The system adopts dynamic quantization coding technology to increase the original data compression rate to more than 90%. Combined with the adaptive incremental learning algorithm, it can complete the iterative update of the diagnostic model without interrupting operation. This technology can significantly improve the intelligent operation and maintenance level of transformers.

[0106] (4) Multimodality: Ability to coordinate and process multi-source heterogeneous data. The transformer multimodal monitoring system based on sensor network technology breaks through the bottleneck of traditional monitoring through a heterogeneous data fusion architecture. Its core capability is reflected in the coordinated processing of multi-source heterogeneous data such as electrical, mechanical, and chemical data. The system integrates multiple sensor probes such as high-frequency current sensor probes, partial discharge monitoring probes, vibration acceleration probes, and gas sensitive sensor probes (with ppb-level detection sensitivity), and synchronously captures multiple types of heterogeneous data streams such as partial discharge pulse waveforms, mechanical vibration spectra, and dissolved gas concentrations in oil, covering microsecond-level transient signals and hour-level slow-moving data. At the data processing level, a tensor spatiotemporal alignment algorithm is used to eliminate the time base deviation of multi-sensor sampling. The time-frequency characteristics of discharge pulses are extracted through an improved wavelet packet transform. A cross-domain correlation model between vibration modes and insulation degradation is constructed in combination with a graph convolutional neural network (GCN). The system integrates the transformer monitoring digital annualization technology to achieve coordinated diagnosis of multiple types of complex faults such as excessive moisture content and winding looseness, providing a multi-dimensional data fusion paradigm for complex transformer defect diagnosis.

[0107] (5) MEMS (Microelectromechanical systems, MEMS) manufacturing process: microstructure processing technology from nanometer scale to millimeter scale. Micro-nano processing technology is used to prepare high-precision, wide-range, wide-frequency domain, and high-consistency sensor probes. The transformer MEMS monitoring system based on sensor network technology innovates the performance of sensor units through micro-nano processing technology. Its core advantages are reflected in the miniaturization, multi-parameter integration and dynamic characteristic optimization of sensor devices. Deep reactive ion etching (DRIE) and wafer-level packaging technology are used to construct three-dimensional microstructure sensors on single-crystal silicon substrates. Micron-level feature size control is achieved through wafer-level batch manufacturing, which ensures the performance of the probe while reducing the effective area of ​​the probe. Through the optimization design of the resonant cavity, high-precision signal capture is achieved over a wide frequency band. The dynamic baseline correction algorithm is used to eliminate multi-physical field coupling interference, and the adaptive variable sampling rate technology (1kHz-10MHz intelligent switching) is combined to ensure the complete capture of wide-frequency domain signals and improve the signal-to-noise ratio. It provides a highly reliable micro-nano sensing solution for densely distributed transformer monitoring.

[0108] (6) Collaboration: Multi-dimensional data fusion analysis to jointly describe the equipment status. The transformer monitoring system based on sensor network technology builds a three-dimensional evaluation model for equipment status through multi-dimensional data collaborative analysis. The system aligns heterogeneous data such as vibration spectrum, dissolved gas species concentration in oil, and partial discharge pulses in time and space, and extracts high-order correlation characteristics of each physical quantity: the nonlinear harmonic components in mechanical vibration form a cross-domain mapping with oil chromatography data, revealing the conduction path of winding looseness to insulation degradation; the partial discharge phase distribution and the temperature gradient field establish a coupling relationship through dynamic weight distribution to locate the core area of ​​latent defect evolution. The data fusion layer introduces a multi-scale feature pyramid structure to construct a state representation vector in the time domain, frequency domain, and spatial domain, and identifies the correlation strength of abnormal modes through an adaptive clustering algorithm. This collaborative analysis mechanism breaks through the one-sidedness of traditional single-parameter diagnosis, improves the causal chain tracing accuracy of complex faults, and provides multi-dimensional evidence chain support for the dynamic deduction of equipment health status.

[0109] (7) High reliability: Using advanced process material technology to improve the reliable service time of sensors in complex environments with high voltage and strong magnetic fields. The transformer high reliability monitoring system based on sensor network technology achieves stable sensing under complex working conditions through material and process innovation. Its core breakthrough lies in the use of sensor packaging technology with synergistic strengthening of electromagnetic, thermal and chemical resistance. The surface of each sensor in the heterogeneous sensor cluster includes a shielding composite layer. For example, the sensor body adopts a composite structure of aluminum nitride ceramic substrate and metal magnetic shielding layer, and integrates the sensor array, piezoelectric vibration sensor unit and anti-saturation electric field sensor through magnetron sputtering process. A high-temperature bonding process is used to achieve a fully sealed package of the sensor unit and signal conditioning circuit, and the adaptive LMS filtering algorithm is used to effectively suppress strong electromagnetic interference. This technology enables the sensor to effectively reduce the baseline drift rate, extend the corrosion resistance life in the oil-immersed environment, reduce the fault false alarm rate, and provide high robustness hardware guarantee for transformer full life cycle monitoring.

[0110] In summary, the state monitoring system provided by the embodiment of the present application reconstructs the equipment state perception paradigm through the fusion of multi-dimensional technologies, and its core features are reflected in the deep combination of multi-parameter perception, edge intelligence and collaborative analysis. The system integrates miniaturized sensing units, covers the synchronous acquisition of multi-dimensional physical quantities such as electromagnetic field distortion, mechanical deformation, and thermodynamic parameters, and realizes real-time feature extraction and cross-domain correlation analysis with the help of embedded intelligent nodes. Multimodal data is mapped to a unified feature space through an adaptive fusion model, and the implicit correlation between winding stress concentration and insulation degradation is revealed in combination with a dynamically corrected digital twin system. The sensing unit adopts a special packaging process to maintain stable measurement performance in a strong electromagnetic environment, significantly extending the service life of the equipment. This technology upgrades traditional discrete monitoring to a holographic perception system, and realizes a decision-making closed loop from abnormal warning to life prediction through fault evolution modeling, providing reliable support for intelligent operation and maintenance of power equipment. In addition, upgrading the traditional discrete threshold criterion to a multi-dimensional state space assessment not only enables early identification of complex defects such as mechanical looseness and partial discharge, but also forms a decision-making closed loop from abnormal warning to life prediction through dynamic deduction of the fault evolution path, promoting the upgrade of the power equipment operation and maintenance model from periodic maintenance to status prediction.

[0111] See Figure 6 , Figure 6This is a schematic diagram of the transformer monitoring system architecture. The entire architecture is built according to the "sensing-computing-transmission-storage" principle. Various sensor probes transmit transformer operating status information to the data acquisition unit via the communication bus. The data acquisition unit transmits processed data to the central processing unit (CPU). The CPU, in conjunction with the NPU and graphics processing unit (GPU), performs data cleaning, feature extraction, and fusion analysis on the transformer operating status information to complete transformer status assessment and generate assessment reports. The CPU and memory interact with each other. On the one hand, the memory stores transformer operating status information, transformer status assessment reports, and sensor network node (sensor probe) information. On the other hand, the CPU can call the information in the memory. The memory interacts with the cloud server through the network card to upload and download information. The cloud server can access the monitoring system through the network card. The client accesses the monitoring system using the display to complete control of the entire system.

[0112] In summary, the sensor network-based transformer monitoring method, system, and framework provided in the embodiments of this application provide an excellent solution for realizing digital and transparent transformers. By utilizing sensor network technology and deploying comprehensive sensor nodes for transformers, operations and maintenance personnel can use clients and cloud platforms to observe various status parameter data for each transformer in a region, including oil temperature, winding temperature, dissolved gas in oil, partial discharge, vibration, core grounding current, position, and dimensions. They can also view transformer operating status assessment reports and implement early maintenance of transformer faults based on system alarm information, thereby improving transformer safety.

[0113] Based on the same inventive concept, embodiments of the present application also provide a condition monitoring device for implementing the aforementioned sensor network-based transformer monitoring method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations in one or more of the following condition monitoring device embodiments can be found in the aforementioned limitations of the sensor network-based transformer monitoring method and will not be further elaborated here.

[0114] In an exemplary embodiment, Figure 7 As shown, a state monitoring device is provided, including: a data acquisition module 701, a data processing module 702 and a state determination module 703, wherein:

[0115] The data acquisition module 701 is used to obtain the initial physical parameter matrix collected by the heterogeneous sensor cluster corresponding to the target transformer; the heterogeneous sensor cluster includes multiple sensors of different types;

[0116] The data processing module 702 is used to perform spatiotemporal normalization processing on the initial physical parameter matrix based on the distribution height of each sensor and the sensor node distance in the heterogeneous sensor cluster to obtain the target physical parameter matrix of the target transformer;

[0117] The state determination module 703 is used to determine the operating state of the target transformer according to the target physical parameter matrix of the target transformer.

[0118] In an exemplary embodiment, the initial physical parameter matrix includes initial mechanical parameters, initial oil chromatography parameters, and multiple initial electrical parameters; the data processing module 702 includes: an alignment processing unit and a data fusion unit, wherein:

[0119] an alignment processing unit, configured to perform spatiotemporal alignment processing on the initial mechanical parameters, the initial oil chromatogram parameters, and the initial electrical parameters based on the distribution heights and sensor node distances of the sensors in the heterogeneous sensor cluster, to obtain corresponding target mechanical parameters, target oil chromatogram parameters, and target electrical parameters;

[0120] The data fusion unit is used to fuse the target mechanical parameters, target oil chromatographic parameters and target electrical parameters to obtain the target physical parameter matrix of the target transformer.

[0121] In an exemplary embodiment, the data fusion unit includes: a path determination subunit, a component determination subunit, and an information association subunit, wherein:

[0122] a path determination subunit, configured to determine a winding loosening conduction path of a target transformer according to a mapping relationship between target mechanical parameters and target oil chromatography parameters;

[0123] The component determination subunit is used to fuse the target electrical parameters according to the preset weights to obtain the component area identification of the target transformer;

[0124] The information association subunit is used to associate the winding loosening conduction path of the target transformer with the component area identification to obtain the target physical parameter matrix of the target transformer.

[0125] In an exemplary embodiment, the state determination module 703 is also used to input the target physical parameter matrix of the target transformer into a preset state prediction model, extract the fault state characteristics in the target physical parameter matrix through the state prediction model, and determine the operating state of the target transformer based on the associated quantitative values ​​between the fault state characteristics.

[0126] In an exemplary embodiment, the condition monitoring device further includes: an information display module, which is used to display the alarm information of the target transformer when the operating status of the target transformer is in a fault state, including the fault location of the target transformer, the transformer location and the cause of the fault.

[0127] Each module in the aforementioned condition monitoring device may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0128] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 8 As shown. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, memory and I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store condition monitoring data. The I / O interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a transformer monitoring method based on sensor network technology is implemented.

[0129] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0130] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0131] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0132] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0133] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0134] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0135] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A transformer monitoring method based on sensor network technology, characterized in that: The method comprises: Obtaining an initial physical parameter matrix collected by a heterogeneous sensor cluster corresponding to a target transformer; the heterogeneous sensor cluster includes multiple sensors of different types; Based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial physical parameter matrix is ​​subjected to spatiotemporal normalization processing to obtain a target physical parameter matrix of the target transformer; An operating state of the target transformer is determined according to a target physical parameter matrix of the target transformer.

2. The method according to claim 1, characterized in that The initial physical parameter matrix includes initial mechanical parameters, initial oil chromatogram parameters, and multiple initial electrical parameters. Based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial physical parameter matrix is ​​subjected to spatiotemporal normalization processing to obtain the target physical parameter matrix of the target transformer, including: Based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster, the initial mechanical parameters, the initial oil chromatogram parameters, and each initial electrical parameter are subjected to spatiotemporal alignment processing to obtain corresponding target mechanical parameters, target oil chromatogram parameters, and target electrical parameters; The target mechanical parameters, the target oil chromatographic parameters and the target electrical parameters are fused to obtain a target physical parameter matrix of the target transformer.

3. The method according to claim 2, characterized in that The step of fusing the target mechanical parameters, the target oil chromatographic parameters, and the target electrical parameters to obtain a target physical parameter matrix of the target transformer includes: determining a winding loosening conduction path of the target transformer according to a mapping relationship between the target mechanical parameter and the target oil chromatogram parameter; According to preset weights, the target electrical parameters are integrated to obtain a component area identification of the target transformer; The winding loosening conduction path of the target transformer is associated with the component area identifier to obtain a target physical parameter matrix of the target transformer.

4. The method according to any one of claims 1 to 3, characterized in that The determining the operating state of the target transformer according to the target physical parameter matrix of the target transformer includes: The target physical parameter matrix of the target transformer is input into a preset state prediction model, the fault state characteristics in the target physical parameter matrix are extracted through the state prediction model, and the operating state of the target transformer is determined based on the associated quantitative values ​​between the fault state characteristics.

5. The method according to any one of claims 1 to 3, characterized in that The method further comprises: When the operating state of the target transformer is in a fault state, alarm information of the target transformer is displayed, including the fault location of the target transformer, the transformer position and the fault cause.

6. A condition monitoring system, characterized in that: The system includes: a processor, a target transformer, and a heterogeneous sensor cluster corresponding to the target transformer; the heterogeneous sensor cluster includes multiple sensors of different types; and the processor is configured to execute the method according to any one of claims 1 to 5.

7. The system according to claim 6, characterized in that A surface of each sensor in the heterogeneous sensor cluster includes a shielding composite layer.

8. A condition monitoring device, characterized in that: The device comprises: A data acquisition module is used to obtain an initial physical parameter matrix collected by a heterogeneous sensor cluster corresponding to a target transformer; the heterogeneous sensor cluster includes multiple sensors of different types; a data processing module, configured to perform spatiotemporal normalization processing on the initial physical parameter matrix based on the distribution height and sensor node distance of each sensor in the heterogeneous sensor cluster to obtain a target physical parameter matrix of the target transformer; The state determination module is used to determine the operating state of the target transformer according to the target physical parameter matrix of the target transformer.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images