Power equipment health state diagnosis method and system based on multi-modal data fusion

Abstract

The application discloses a kind of multi-modal data fusion's electric power equipment health state diagnosis method and system, belong to electric power equipment state monitoring field, wherein method includes: S100, the infrared thermal imaging data of electric power equipment, vibration signal, current harmonic data and partial discharge signal are collected, and time synchronization and space registration are carried out.S200, obtain infrared thermal imaging data, vibration signal, current harmonic data and partial discharge signal, input dynamic weight fusion model, obtain equipment health state score and fault type.S300, according to equipment health state score and fault type, trigger hierarchical alarm.The application can carry out timely early warning to potential fault hidden danger of electric power equipment.

Description

Technical Field

[0001] This invention relates to the field of power equipment condition monitoring, and in particular to a method and system for diagnosing the health status of power equipment through multimodal data fusion. Background Technology

[0002] In recent years, power equipment has played a key role in various industries because it is an important part of maintaining production and life. In power distribution systems, switchgear is a key component for protecting and controlling equipment, ensuring the reliable operation of the power system. However, the heat generated by power equipment during operation and potential faults can lead to a decline in equipment performance, an increase in downtime, and even fire and safety risks.

[0003] In existing technologies, single-sensor diagnosis is susceptible to interference (such as infrared imaging being affected by environmental radiation), and fixed-weight fusion models are difficult to adapt to the operating characteristics of different devices such as transformers and circuit breakers. Summary of the Invention

[0004] This invention aims to address at least one of the technical problems existing in the prior art. To this end, this invention proposes a method and system for diagnosing the health status of power equipment based on multimodal data fusion, which can provide timely early warning of potential faults in power equipment and quickly locate the causes of faults.

[0005] This invention provides a method for diagnosing the health status of power equipment using multimodal data fusion, comprising the following steps: S100, acquiring infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals of the power equipment, and performing time synchronization and spatial registration; S200, acquiring infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals, inputting them into a dynamic weighted fusion model, and obtaining a health status score and fault type for the equipment; the dynamic weighted fusion model includes: a feature extraction layer, used to extract features from the infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals to obtain temperature gradient features, high-frequency component features of the vibration signal, and current... The system includes: odd-order harmonic characteristics and pulse characteristics of partial discharge; a condition perception layer for real-time monitoring of equipment condition parameters and prediction of current condition type using an LSTM network; the condition parameters include real-time load rate, ambient temperature and humidity, and cooling fan speed; the condition type includes at least one of steady state, overload, high temperature, and mechanical wear; an attention weight allocation layer for calculating dynamic weights of each mode based on condition type and feature correlation using a multi-head attention mechanism; a fusion decision layer for generating equipment health status scores and fault probability distributions through weighted summation, ultimately outputting equipment status level and locating abnormal parts; and S300 triggering graded alarms based on equipment health status scores and fault types.

[0006] According to some embodiments of the present invention, step S200 includes: S210, extracting the temperature gradient matrix, high-frequency energy ratio, THD distortion rate, and odd harmonic amplitude ratio based on infrared thermal imaging data, vibration signal, current harmonic data, and partial discharge signal; S220, acquiring operating condition parameters; predicting the current operating condition type based on the long-term dependency relationship of the operating condition parameters captured by bidirectional LSTM; S230, calculating the dynamic weights of each mode based on an improved multi-head attention structure, wherein the multi-head attention structure includes an operating condition attention head and a feature attention head; the operating condition attention head outputs an operating condition weight vector based on the operating condition parameters, and the feature attention head outputs feature correlation weights based on the feature vectors of each mode, the calculation formula being: Where W is the feature relevance weight; The operating condition adaptive factor is generated by passing the hidden state of BiLSTM through a fully connected layer and sigmoid activation; Q is the query matrix, representing the current window operating condition parameters; K is the key matrix, representing the historical window operating condition sequence. The intermodal energy function; The dimension of the key vector; ;in, The hidden state of the BiLSTM at the current time step; These are trainable parameters that control the intensity of the influence of operating conditions on attention. For the Sigmoid function; ;in, For the feature vectors of different modalities, the trainable matrix We is a diagonal matrix whose diagonal elements are optimized by gradient descent and initialized to follow a normal distribution with a mean of 0 and a standard deviation of 0.02; i,j are modality indices; S240, dynamically fuse the working condition weight vector and feature correlation weights to obtain the final weights; S250, generate a weight allocation heatmap to show the contribution ratio of each modality in the decision-making; output a SHAP value interpretation report.

[0007] According to some embodiments of the present invention, step S220 includes: S221, inputting multi-source operating condition parameters within a sliding window into a bidirectional LSTM network, extracting historical and future potential trend features through the forward and backward layers of the bidirectional LSTM network respectively; and dynamically weighting the parameters at key time points within the window based on an attention mechanism; S222, calculating the probability distribution of each operating condition type through a Softmax function, and determining a continuous abnormal operating condition if three consecutive sliding windows predict the same non-steady-state operating condition; S223, calculating the KL divergence between the current operating condition data distribution and the historical benchmark in real time, and freezing the underlying parameters of the BiLSTM and fine-tuning the attention layer weights if the KL divergence exceeds a threshold of 0.1.

[0008] According to some embodiments of the present invention, the time synchronization and spatial registration include: timestamping multi-sensor data using GPS or PTP protocol; and aligning the infrared thermal image with the physical location of the device through feature point matching.

[0009] According to some embodiments of the present invention, the method for extracting the temperature gradient features includes: performing super-resolution reconstruction on the infrared thermal image to generate a high-precision temperature distribution map; and using the Sobel operator to calculate the temperature gradient and identify micron-level crack hot spots.

[0010] According to some embodiments of the present invention, the dynamic weight fusion model is optimized by reinforcement learning, including: dynamically adjusting the decision weight of vibration data in mechanical wear faults based on historical diagnostic accuracy feedback.

[0011] According to some embodiments of the present invention, the method further includes early detection of weak faults: constructing a generative adversarial network to synthesize abnormal samples using normal data; and conducting joint frequency domain analysis of the implicit correlation between infrared temperature fluctuations and vibration spectra.

[0012] According to some embodiments of the present invention, the method further includes: automatically triggering a robot to perform bolt tightening operation when the diagnostic result indicates that the connector is loose; verifying the temperature drop rate with a second infrared scan after maintenance, and escalating the alarm if the expected result is not achieved.

[0013] According to some embodiments of the present invention, the method further includes: constructing a three-dimensional thermal field simulation model of the device, inputting real-time data to simulate the heat diffusion path; and locating the abnormal coordinates of the heat source through a backpropagation algorithm.

[0014] Another embodiment of the present invention provides a power equipment health status diagnosis system based on multimodal data fusion, comprising: a data acquisition module for integrating an infrared thermal imager, a vibration sensor, a current transformer, and a partial discharge detector; and a cloud analysis platform for executing the method described above and outputting diagnostic results.

[0015] The embodiments of this invention achieve at least the following beneficial effects: Through multimodal data fusion and dynamic weight allocation, the embodiments of this invention can more accurately reflect the health status of equipment, improving the accuracy of diagnosis; they can adapt to the operating characteristics and working conditions of different equipment, improving the breadth and practicality of diagnosis. Automated and intelligent diagnosis and alarm processes reduce the burden of manual monitoring and improve management efficiency. By early detection of minor faults and the construction of three-dimensional thermal field simulation models of equipment, potential faults can be detected and prevented in a timely manner, ensuring the safe and reliable operation of equipment.

[0016] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0017] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0018] Figure 1 This is a flowchart illustrating the method according to an embodiment of the present invention;

[0019] Figure 2 This is a schematic block diagram of the system modules according to an embodiment of the present invention.

[0020] Figure label:

[0021] Data acquisition module 100, cloud analysis platform 200. Detailed Implementation

[0022] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0023] In the description of this invention, "several" means one or more, "multiple" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0024] This invention utilizes infrared thermal imaging technology to monitor the temperature of critical components of power equipment in real time, such as: temperature monitoring of busbars and copper busbar connections in distribution cabinets (switch cabinets); temperature monitoring of cable joints in distribution cabinets and distribution boxes; temperature monitoring of power equipment such as transformers, surge arresters, and voltage transformers; temperature monitoring of the main body and mechanical connection parts of large motors and operating equipment; temperature monitoring of connection and plug parts of enclosed busbars; and so on.

[0025] This invention addresses the limitations of single sensors, poor adaptability to operating conditions, and difficulty in detecting early faults in power equipment health status monitoring through dynamic fusion and intelligent diagnosis of multimodal data. Traditional methods rely on single sensor data such as infrared or vibration, which are susceptible to environmental noise interference leading to false alarms. This solution integrates four types of heterogeneous data: infrared thermal imaging, vibration signals, current harmonics, and partial discharge. A dynamic weight allocation model quantifies the contribution of different modes, significantly improving diagnostic reliability. To address operating condition interference such as equipment load fluctuations and environmental temperature changes, an LSTM operating condition perception layer is introduced to identify the operating status in real time. A multi-head attention mechanism is used to dynamically adjust the weights of each mode, ensuring high accuracy in different scenarios such as substations and wind power. Through high-resolution temperature gradient feature extraction and time-frequency analysis of partial discharge signals, combined with a threshold alarm strategy optimized by reinforcement learning, potential defects such as insulation aging and bearing microcracks, which are difficult to detect using traditional methods, can be identified, providing early warning of faults.

[0026] Reference Figure 1 This invention provides a method for diagnosing the health status of power equipment based on multimodal data fusion, comprising the following steps:

[0027] S100 collects infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals of power equipment, and performs time synchronization and spatial registration.

[0028] S200 acquires infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals, inputs them into a dynamic weighted fusion model, and obtains equipment health status scores and fault types.

[0029] S300: Triage alarms are triggered based on the equipment health status score and fault type.

[0030] The dynamic weight fusion model in this embodiment of the invention includes:

[0031] The feature extraction layer is used to extract features from infrared thermal imaging data, vibration signals, and partial discharge signals from current harmonic datasets, obtaining temperature gradient features, high-frequency component features of vibration signals, odd harmonic features of current harmonics, and pulse features of partial discharge.

[0032] The operating condition perception layer is used to monitor the operating condition parameters of the equipment in real time and predict the current operating condition type through an LSTM network. The operating condition parameters include real-time load rate, ambient temperature and humidity, and cooling fan speed. The operating condition type includes at least one of steady state, overload, high temperature, and mechanical wear.

[0033] The attention weight allocation layer is used to calculate the dynamic weights of each modality based on the working condition type and feature correlation using a multi-head attention mechanism.

[0034] The fusion decision layer is used to generate equipment health status scores and fault probability distributions through weighted summation, and finally output the equipment status level and locate the abnormal parts.

[0035] In some embodiments, step S200 includes:

[0036] S210. Based on infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals, extract the temperature gradient matrix, high-frequency energy ratio, THD distortion rate, and odd harmonic amplitude ratio.

[0037] S220. Obtain operating condition parameters; based on bidirectional LSTM, capture the long-term dependencies of operating condition parameters and predict the current operating condition type.

[0038] S230. Calculate the dynamic weights of each modality based on the improved multi-head attention structure. The multi-head attention structure includes a condition attention head and a feature attention head. The condition attention head outputs a condition weight vector based on the condition parameters, and the feature attention head outputs feature correlation weights based on the feature vectors of each modality. The calculation formula is as follows:

[0039] ;

[0040] Where W is the feature relevance weight; The operating condition adaptive factor is generated by passing the hidden state of BiLSTM through a fully connected layer and sigmoid activation; Q is the query matrix, representing the current window operating condition parameters; K is the key matrix, representing the historical window operating condition sequence. The intermodal energy function; The dimension of the key vector;

[0041] ;

[0042] in, The hidden state of the BiLSTM at the current time step; These are trainable parameters that control the intensity of the influence of operating conditions on attention. For the Sigmoid function;

[0043] ;

[0044] in, For the feature vectors of different modalities, the trainable matrix We is a diagonal matrix whose diagonal elements are optimized by gradient descent and initialized to follow a normal distribution with a mean of 0 and a standard deviation of 0.02; i,j are modality indices.

[0045] S240. Dynamically fuse the working condition weight vector and the feature correlation weight to obtain the final weight.

[0046] S250: Generate a weighted heatmap to show the contribution ratio of each modality in the decision-making process; output a SHAP value interpretation report.

[0047] In some embodiments of the present invention, step S220 includes:

[0048] S221. Input the multi-source operating condition parameters within the sliding window into a bidirectional LSTM network, and extract historical and future potential trend features through the forward and reverse layers of the bidirectional LSTM network respectively; and dynamically weight the parameters at key time points within the window based on the attention mechanism.

[0049] S222. Calculate the probability distribution of each working condition type using the Softmax function. If three consecutive sliding windows predict the same unsteady working condition, it is determined to be a continuous abnormal working condition.

[0050] S223. Calculate the KL divergence between the current operating condition data distribution and the historical benchmark in real time. If the KL divergence exceeds the threshold of 0.1, freeze the underlying parameters of BiLSTM and fine-tune the attention layer weights.

[0051] In some embodiments, time synchronization and spatial registration of the present invention include: timestamping multi-sensor data using GPS or PTP protocol; and aligning infrared thermal images with the physical location of the device through feature point matching.

[0052] In some embodiments, the method for extracting temperature gradient features according to the present invention includes: performing super-resolution reconstruction on an infrared thermal image to generate a high-precision temperature distribution map; and using the Sobel operator to calculate the temperature gradient and identify micron-level crack hot spots.

[0053] In some embodiments, the method of the present invention includes: optimizing a dynamic weight fusion model through reinforcement learning, including: dynamically adjusting the decision weight of vibration data in mechanical wear faults based on historical diagnostic accuracy feedback.

[0054] In some embodiments, the method of the present invention further includes early detection of weak faults: constructing a generative adversarial network to synthesize abnormal samples using normal data; and conducting joint frequency domain analysis of the implicit correlation between infrared temperature fluctuations and vibration spectra.

[0055] In some embodiments, the method of the present invention further includes: automatically triggering a robot to perform bolt tightening operation when the diagnostic result indicates that the connector is loose; verifying the temperature drop rate with a second infrared scan after maintenance, and escalating the alarm if the temperature drop does not meet expectations.

[0056] In some embodiments, the method of the present invention further includes: constructing a three-dimensional thermal field simulation model of the device, inputting real-time data to simulate the thermal diffusion path; and locating the abnormal coordinates of the heat source through a backpropagation algorithm.

[0057] Reference Figure 2 Another embodiment of the present invention provides a power equipment health status diagnosis system based on multimodal data fusion, comprising:

[0058] The data acquisition module 100 is used to integrate an infrared thermal imager, a vibration sensor, a current transformer, and a partial discharge detector.

[0059] The cloud-based analytics platform 200 is used to execute the power equipment health status diagnosis method described above using multimodal data fusion, and output the diagnosis results.

[0060] Although specific embodiments are described herein, those skilled in the art will recognize that many other modifications or alternative embodiments are also within the scope of this disclosure. For example, any of the functions and / or processing capabilities described in connection with a particular device or component can be performed by any other device or component. Furthermore, while various exemplary embodiments and architectures have been described according to embodiments of this disclosure, those skilled in the art will recognize that many other modifications to the exemplary embodiments and architectures described herein are also within the scope of this disclosure.

[0061] The foregoing description, with reference to block diagrams and flowcharts of systems, methods, systems, and / or computer program products according to exemplary embodiments, has described certain aspects of this disclosure. It should be understood that one or more blocks in the block diagrams and flowcharts, as well as combinations of blocks in the block diagrams and flowcharts, can be implemented by executing computer-executable program instructions, respectively. Similarly, according to some embodiments, some blocks in the block diagrams and flowcharts may not need to be executed in the order shown, or may not all need to be executed. Furthermore, additional components and / or operations beyond those shown in the blocks in the block diagrams and flowcharts may exist in some embodiments.

[0062] Therefore, blocks in block diagrams and flowcharts support combinations of means for performing a specified function, combinations of elements or steps for performing a specified function, and program instruction means for performing a specified function. It should also be understood that each block in a block diagram and flowchart, and combinations of blocks in block diagrams and flowcharts, can be implemented by a dedicated hardware computer system or a combination of dedicated hardware and computer instructions that performs a specific function, element, or step.

[0063] The program modules, applications, etc., described herein may include one or more software components, including, for example, software objects, methods, data structures, etc. Each such software component may include computer-executable instructions that, in response to execution, cause at least a portion of the functionality described herein (e.g., one or more operations of the exemplary methods described herein) to be performed.

[0064] Software components can be coded using any of a variety of programming languages. An exemplary programming language could be a low-level programming language, such as assembly language associated with a specific hardware architecture and / or operating system platform. Software components including assembly language instructions may need to be converted into executable machine code by an assembler before being executed by the hardware architecture and / or platform. Another exemplary programming language could be a higher-level programming language that is portable across multiple architectures. Software components including higher-level programming languages ​​may need to be converted into an intermediate representation by an interpreter or compiler before execution. Other examples of programming languages ​​include, but are not limited to, macro languages, shell or command languages, job control languages, scripting languages, database query or search languages, or report writing languages. In one or more exemplary embodiments, a software component containing instructions from one of the above-described programming language examples can be executed directly by the operating system or other software components without first being converted into another form.

[0065] Software components can be stored as files or other data storage structures. Software components of similar type or related function can be stored together in a specific directory, folder, or library. Software components can be static (e.g., pre-defined or fixed) or dynamic (e.g., created or modified at runtime).

[0066] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for power equipment health state diagnosis based on multi-modal data fusion, characterized in that, Includes the following steps: S100: Collects infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals of power equipment, and performs time synchronization and spatial registration. S200: Acquire infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals, input them into a dynamic weighted fusion model, and obtain equipment health status scores and fault types. The dynamic weight fusion model includes: The feature extraction layer is used to extract features from infrared thermal imaging data, vibration signals, current harmonic datasets, and partial discharge signals to obtain temperature gradient features, high-frequency component features of vibration signals, odd harmonic features of current harmonics, and pulse features of partial discharge. The operating condition perception layer is used to monitor the operating condition parameters of the equipment in real time and predict the current operating condition type through an LSTM network. The operating condition parameters include real-time load rate, ambient temperature and humidity, and cooling fan speed. The operating condition type includes at least one of steady state, overload, high temperature, and mechanical wear. The attention weight allocation layer is used to calculate the dynamic weights of each modality based on the working condition type and feature correlation using a multi-head attention mechanism. The fusion decision layer is used to generate equipment health status scores and fault probability distributions through weighted summation, and finally output the equipment status level and locate the abnormal parts; S300: Trigger graded alarms based on equipment health status score and fault type; Step S200 includes: S210. Based on infrared thermal imaging data, vibration signals, current harmonic data, and partial discharge signals, extract the temperature gradient matrix, high-frequency energy ratio, THD distortion rate, and odd harmonic amplitude ratio. S220. Obtain operating condition parameters; based on bidirectional LSTM, capture the long-term dependencies of operating condition parameters and predict the current operating condition type; S230. Calculate the dynamic weights of each modality based on the improved multi-head attention structure, wherein the multi-head attention structure includes a condition attention head and a feature attention head; the condition attention head outputs a condition weight vector based on the condition parameters and the condition type predicted in S220, and the feature attention head outputs feature correlation weights based on the feature vectors of each modality, calculated using the following formula: ； wherein W is a feature correlation weight; is a working condition adaptive factor, generated by BiLSTM hidden state through a full connection layer and a sigmoid activation; Q is a query matrix, representing the current window working condition parameters; K is a key matrix, representing the historical window working condition sequence; is an inter-modal energy function; is the dimension of the key vector; ； wherein, is the current time step BiLSTM hidden state; is a trainable parameter that controls the strength of the influence of the working condition on the attention; is a sigmoid function; ； wherein, are feature vectors of different modalities, the trainable matrix Weis a diagonal matrix whose diagonal elements are optimized by gradient descent method, initialized to follow a normal distribution with mean 0 and standard deviation 0.02; i, j are modal indices; S240. Dynamically fuse the working condition weight vector and the feature correlation weight to obtain the final weight.

2. The method of claim 1, wherein, Step S200 further includes: S250: Generate a weighted heatmap to show the contribution ratio of each modality in the decision-making process; output a SHAP value interpretation report.

3. The method of claim 1, wherein, Step S220 includes: S221. Input the multi-source operating condition parameters within the sliding window into a bidirectional LSTM network, and extract historical and future potential trend features through the forward and reverse layers of the bidirectional LSTM network respectively; and dynamically weight the parameters at key time points within the window based on the attention mechanism. S222. Calculate the probability distribution of each working condition type using the Softmax function. If three consecutive sliding windows predict the same unsteady working condition, then it is determined to be a continuous abnormal working condition. S223. Calculate the KL divergence between the current operating condition data distribution and the historical benchmark in real time. If the KL divergence exceeds the threshold of 0.1, freeze the underlying parameters of BiLSTM and adjust the attention layer weights.

4. The method of power equipment health condition diagnosis using multi-modal data fusion as claimed in claim 1, wherein, The time synchronization and spatial registration include: Timestamping multi-sensor data with GPS or PTP protocol; aligning infrared thermal image with device physical location through feature point matching.

5. The method of power equipment health condition diagnosis using multi-modal data fusion as claimed in claim 1, wherein, The temperature gradient feature extraction method comprises: Super-resolution reconstruction of infrared thermal image to generate high-precision temperature distribution map; Sobel operator is used to calculate temperature gradient to identify micron-level crack hot spots.

6. The method of power equipment health condition diagnosis using multi-modal data fusion as claimed in claim 1, wherein, The method comprises: Optimizing the dynamic weight fusion model through reinforcement learning, comprising:

7. The method of power equipment health condition diagnosis using multi-modal data fusion as claimed in claim 1, wherein, According to the historical diagnosis accuracy feedback, dynamically adjusting the decision weight of vibration data in mechanical wear and tear fault. The method also includes early perception of weak faults: Constructing a generative adversarial network to synthesize abnormal samples using normal data; 8. The method of power equipment health condition diagnosis using multi-modal data fusion as claimed in claim 1, wherein, Through frequency domain joint analysis, the implicit correlation between infrared temperature fluctuation and vibration spectrum is analyzed. The method also comprises: When the diagnosis result is loose connection, automatically trigger the robot to perform bolt tightening operation; 9. The method of power equipment health condition diagnosis using multi-modal data fusion as claimed in claim 1, wherein, After maintenance, the secondary infrared scanning verifies the temperature drop amplitude, and if it does not reach the expected value, the alarm is upgraded. The method also comprises: Constructing a three-dimensional thermal field simulation model of the device, inputting real-time data to simulate the heat diffusion path; 10. A multi-modal data fusion based power equipment health condition diagnosis system characterized in that, Locating the abnormal coordinates of the heat source through the back propagation algorithm. Comprise: A data acquisition module for integrating an infrared thermal imager, a vibration sensor, a current transformer, and a partial discharge detector; A cloud analysis platform for executing the method as claimed in any one of claims 1-9, outputting a diagnosis result.

Images