Transformer state intelligent monitoring and cloud management method, system and device and storage medium

Abstract

The invention discloses a transformer state intelligent monitoring and cloud management method, system and device and a storage medium, and relates to the field of power system automation and intelligent monitoring, and the method comprises the steps: employing a multi-sensor heterogeneous data fusion algorithm and a self-adaptive sampling mechanism, and collecting transformer operation data in real time; establishing a deep learning model, and performing state evaluation and fault prediction on the acquired transformer operation data; a cloud-side collaborative architecture and a rule engine are constructed based on results of state evaluation and fault prediction, and dynamic task allocation and real-time early warning are realized; after the rule engine triggers early warning, a fault diagnosis and maintenance scheme is provided in combination with the expert knowledge base and association rule mining; according to the invention, the acquired data is accurate and diverse, the deep learning model is combined, the prediction is accurate, the cloud-side collaborative architecture is efficient, a diagnosis and maintenance scheme is provided by means of an expert knowledge base and rule mining, and the operation reliability and the operation and maintenance efficiency of the transformer are improved.

Description

Technical Field

[0001] The present invention relates to the field of power system automation and intelligent monitoring, and in particular to a method, system, device and storage medium for intelligent monitoring and cloud management of transformer status. Background Art

[0002] In the power system, transformers are key equipment, and their stable operation is crucial to ensuring the reliability and safety of power supply.

[0003] However, in the existing technology, transformer operation monitoring mainly relies on traditional offline detection and manual inspection methods. Traditional offline detection methods are usually carried out after a transformer fault occurs, and are unable to capture changes in the operating status of the equipment in real time, resulting in delayed fault detection. Manual inspections are limited by the experience and subjective judgment of inspectors, and data collection may not be accurate and comprehensive enough to fully reflect the actual operating status of the transformer. At the same time, the lack of advanced data processing and analysis technology makes it impossible to deeply explore and utilize the collected data, making it difficult to provide accurate fault warnings and maintenance recommendations. In addition, frequent on-site inspections and maintenance require a lot of manpower, material resources and time, resulting in high maintenance costs. In addition, due to the lack of real-time data analysis methods, it is difficult to optimize the configuration and scheduling of power resources according to the operating status of the transformer. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: how to develop a system that can monitor the operating status of the transformer in real time, accurately analyze data, and provide accurate fault warnings and maintenance suggestions.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a method for intelligent monitoring and cloud management of transformer status, including:

[0008] Adopt multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism to collect transformer operation data in real time;

[0009] Establish a deep learning model to perform condition assessment and fault prediction on the collected transformer operation data;

[0010] Based on the results of status assessment and fault prediction, a cloud-edge collaborative architecture and rule engine are built to achieve dynamic task allocation and real-time warning.

[0011] When the rule engine triggers an early warning, it combines the expert knowledge base with association rule mining to provide fault diagnosis and maintenance solutions.

[0012] As a preferred solution for intelligent transformer status monitoring and cloud management, it:

[0013] The multi-sensor heterogeneous data fusion algorithm includes:

[0014] Confidence fusion of sensor data: define the reliability factor of each sensor and dynamically adjust it through the entropy weight method.

[0015] The beneficial effects of this preferred technical solution are: by performing confidence fusion on sensor data and dynamically adjusting the sensor reliability factor using the entropy weight method, the characteristics and reliability of different sensors can be comprehensively considered, so that the fused data can more accurately reflect the actual operating status of the transformer, thereby improving the quality and credibility of data acquisition.

[0016] As a preferred solution for intelligent transformer status monitoring and cloud management, it:

[0017] The multi-sensor heterogeneous data fusion algorithm also includes:

[0018] If the fusion confidence is lower than the preset threshold, the data is judged to be abnormal and the following operations are initiated: activating the adaptive sampling mechanism and pushing a preliminary warning to the edge.

[0019] As a preferred solution for intelligent transformer status monitoring and cloud management, it:

[0020] The adaptive sampling mechanism includes:

[0021] Parameters are selected to dynamically calculate the health index, a dynamic optimization method with regular updates is used for the health index, and adaptive sampling is performed based on the health index.

[0022] The beneficial effects of this preferred technical solution are: Dynamically calculating the health index by selecting parameters and employing a regularly updated dynamic optimization method can more comprehensively and accurately reflect the health status of the transformer, and the health index calculation method can be promptly adjusted according to changes in the transformer's actual operating conditions. Adaptive sampling based on the health index allows for reasonable adjustment of the sampling interval based on the varying health states of the transformer, reducing data acquisition costs and resource consumption while ensuring data validity.

[0023] As a preferred solution for intelligent transformer status monitoring and cloud management, it:

[0024] The adaptive sampling based on the health index includes:

[0025] If the health index is greater than the first threshold, it is determined to be in normal state, and the sampling interval is set to 10 minutes;

[0026] If the health index is greater than the second threshold and less than or equal to the first threshold, it is determined to be in a warning state, and the sampling interval is the larger value between 2 minutes and the original interval × the third threshold;

[0027] If the health index is less than or equal to the second threshold, it is determined to be a fault state and 1-second continuous sampling of all sensors is started.

[0028] As a preferred solution for intelligent transformer status monitoring and cloud management, it:

[0029] The establishment of a deep learning model to perform status assessment and fault prediction on the collected transformer operation data includes:

[0030] Preprocess and extract features from the collected transformer operation data to generate feature vectors;

[0031] A dynamic model of the transformer's normal operating status is established through a long short-term memory network, and the deviation between the current status and the baseline is compared in real time to evaluate the operating status.

[0032] The feature vector is input into the deep learning model for condition assessment and fault prediction.

[0033] As a preferred solution for intelligent transformer status monitoring and cloud management, it:

[0034] The cloud-edge collaborative architecture and rule engine based on the results of status assessment and fault prediction are constructed to achieve dynamic task allocation and real-time warning, including:

[0035] Set specific alarm logic and response measures according to different operating scenarios and environmental conditions;

[0036] A decision value is generated based on network conditions and data volume. If the decision value is greater than the set threshold, edge processing is initiated and a fast Fourier transform is performed to analyze the vibration spectrum.

[0037] If the decision value is less than or equal to the set threshold, it is uploaded to the cloud for further fault prediction.

[0038] In a second aspect, an embodiment of the present invention provides a transformer status intelligent monitoring and cloud management system, including:

[0039] Real-time acquisition module, which uses multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism to collect transformer operation data in real time;

[0040] The evaluation and prediction module is used to establish a deep learning model to perform status evaluation and fault prediction on the collected transformer operation data;

[0041] The early warning module is used to build a cloud-edge collaborative architecture and rule engine based on the results of status assessment and fault prediction, realizing dynamic task allocation and real-time early warning;

[0042] The solution generation module is used to provide fault diagnosis and maintenance solutions by combining the expert knowledge base and association rule mining when the rule engine triggers an early warning.

[0043] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0044] memory and processor;

[0045] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the one or more programs are executed by the one or more processors, the one or more processors implement the transformer status intelligent monitoring and cloud management method as described in any embodiment of the present invention.

[0046] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the intelligent transformer status monitoring and cloud management method.

[0047] Beneficial effects of the present invention: The present invention combines a multi-sensor heterogeneous data fusion algorithm with an adaptive sampling mechanism to fuse data from different sensors and reduce errors. Adaptive sampling can adjust the frequency as needed, reducing the frequency to reduce redundancy during stability and increasing the frequency to capture key data during fluctuations, thereby improving data quality and efficiency and enhancing real-time performance. Deep learning models can mine patterns from large amounts of data to accurately assess transformer status and predict faults. Dynamic task allocation is achieved through a cloud-edge collaborative architecture and a rule engine, and local processing on edge devices reduces pressure on the cloud and rationally allocates resources. The rule engine monitors in real time, quickly responds to anomalies, and ensures the real-time performance of the system. Combined with the expert knowledge base and association rule mining, it accurately locates the cause of the fault and improves diagnostic efficiency. It can provide targeted and actionable maintenance solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0049] Figure 1 This is the overall flow chart of the transformer status intelligent monitoring and cloud management method described in the present invention. DETAILED DESCRIPTION

[0050] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0051] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a method for intelligent monitoring and cloud management of transformer status, including:

[0052] S1: Uses multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism to collect transformer operation data in real time;

[0053] S2: Build a deep learning model to perform status assessment and fault prediction on the collected transformer operation data;

[0054] S3: Based on the results of status assessment and fault prediction, a cloud-edge collaborative architecture and rule engine are built to achieve dynamic task allocation and real-time warning.

[0055] S4: When the rule engine triggers an early warning, it combines the expert knowledge base with association rule mining to provide fault diagnosis and maintenance solutions.

[0056] It should be noted that, through steps S1-S4, this embodiment achieves efficient and accurate collection of transformer operating data through a multi-sensor heterogeneous data fusion algorithm and an adaptive sampling mechanism. The use of deep learning models to perform status assessment and fault prediction on the collected data can detect possible problems with the transformer in advance. The construction of a cloud-edge collaborative architecture and a rule engine not only realizes dynamic task allocation and improves the system's response speed and processing efficiency, but also can issue real-time warnings in a timely manner. After the rule engine triggers the warning, the fault diagnosis and maintenance solutions provided by the expert knowledge base and association rule mining provide strong support for operation and maintenance personnel to solve transformer faults.

[0057] Example 2, reference Figure 1 , which is an embodiment of the present invention, provides a transformer status intelligent monitoring and cloud management method based on the previous embodiment, including:

[0058] In this embodiment, the multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism are used in the above step S1 to collect transformer operation data in real time, including:

[0059] High-precision fiber optic sensing technology and wireless sensor networks are used to monitor key parameters such as transformer voltage, current, oil temperature, and internal conditions (such as temperature, humidity, insulator status, etc.) in real time.

[0060] It should be noted that these sensors offer high sensitivity, high precision, long-term stability, immunity to electromagnetic interference, and suitability for high-voltage environments. They can monitor temperature, strain, and partial discharge (e.g., fiber grating (FBG) technology). Furthermore, they can utilize low-power wide-area networks (LoRa / NB-IoT) or mesh networks (Zigbee) to achieve simultaneous multi-node data transmission, accurately reflecting the transformer's operating status.

[0061] Multi-sensor heterogeneous data fusion algorithms include:

[0062] The confidence fusion of the data of fiber Bragg grating (FBG), wireless vibration sensor, and gas chromatography sensor is performed: the reliability factor α of each sensor is defined (temperature factor α1 = 0.9, vibration α2 = 0.85, gas α3 = 0.8), and it is dynamically adjusted through the entropy weight method. The fusion formula is:

[0063]

[0064] Where S is the fusion confidence, μ i is the mean of historical data of each sensor, α i is the standard deviation of historical data of each sensor.

[0065] If the fusion confidence is lower than the preset threshold (such as 0.7), the data is judged to be abnormal and the following operations are initiated: the adaptive sampling mechanism is activated and a preliminary warning is pushed to the edge (such as "the temperature data confidence is insufficient and needs to be reviewed").

[0066] The adaptive sampling mechanism includes:

[0067] Dynamically calculate the health index (HI), the parameters selected are: oil temperature (T), H2 concentration (C), vibration spectrum energy (E), humidity (H), and the calculation formula is expressed as:

[0068] HI=0.3*Sigmoid(T norm )+0.25*tanh(C / 50ppm)+0.2*log(E / 1*10 -5 g) + 0.25*(1-H / 100%)

[0069] Among them, T norm is the normalized oil temperature.

[0070] The initial weight values ​​are oil temperature (30%), H2 concentration (25%), vibration (20%), and humidity (25%);

[0071] The dynamic optimization method (updated every 24 hours) is:

[0072] Set the objective function: minimize the HI misjudgment rate of historical fault data;

[0073] Adopting the gradient descent method: adjusting the weight coefficient so that HI decreases significantly before a fault occurs;

[0074] For example, if the recent H 2 concentration contributes more to the fault warning, its weight is increased to 28%, and the weight of the oil temperature is reduced to 27%.

[0075] Adaptive sampling based on health index:

[0076] If the health index is greater than 0.8, it is considered normal and the sampling interval is set to 10 minutes;

[0077] If the health index is greater than 0.6 and less than or equal to 0.8, it is judged to be in a warning state, and the sampling interval is the larger value between 2 minutes and the original interval × 0.7;

[0078] If the health index is less than or equal to 0.6, it is determined to be a fault state and 1-second continuous sampling of all sensors is started.

[0079] In this embodiment, the deep learning model is established in the above step S2 to perform status assessment and fault prediction on the collected transformer operation data, including:

[0080] Preprocess and extract features from the collected transformer operation data to generate feature vectors;

[0081] Specifically, an adaptive filtering algorithm is used for denoising to ensure data accuracy and reliability. Wavelet transform is used to perform multi-scale analysis on vibration signals to extract fault characteristic frequencies. Principal component analysis is used to highlight key features of dissolved gas analysis data.

[0082] A dynamic model of the transformer's normal operating status is established through a long short-term memory network, and the deviation between the current status and the baseline is compared in real time to evaluate the operating status.

[0083] For example, the model structure used is graph TD;

[0084] A[input layer: 12-dimensional sensor data]-->B{LSTM layer: 64 units};

[0085] B-->C [Attention layer: calculate time step weight];

[0086] C-->D [fully connected layer: Softmax output];

[0087] And introduce the working condition adaptive gating mechanism.

[0088] Feature vectors are fed into a deep learning model for condition assessment and fault prediction. When a potential fault is predicted or an abnormal pattern is identified, the system automatically updates the dynamic baseline model and rule engine to ensure the accuracy of the assessment and prediction. All processing and analysis results are uploaded to the cloud, where big data analytics are used to continuously optimize algorithms and model parameters, enabling the system to evolve.

[0089] Based on the collected transformer operation data, a joint criterion of local density ρ and distance δ is defined. When C2H2 in oil is greater than 5ppm and δ is greater than 3σ, it is judged as an arc discharge anomaly, where σ is the bandwidth parameter of the Gaussian kernel function.

[0090] In another possible implementation, in addition to the joint criterion of local density ρ and distance δ, seasonal decomposition methods from time series analysis can also be introduced. Transformer operating data may exhibit seasonal variations, and seasonal decomposition can be used to decompose the data into trend components, seasonal components, and residual components. Abnormal fluctuations in the residual component can also serve as an indicator for fault prediction. For example, in a specific season of each year, the oil temperature of the transformer will experience certain seasonal variations. If the residual component suddenly increases during that season, it may indicate a potential fault in the transformer.

[0091] In this embodiment, in step S3 above, based on the results of status assessment and fault prediction, a cloud-edge collaborative architecture and rule engine are constructed to implement dynamic task allocation and real-time warning, including:

[0092] It should be noted that traditional cloud-based monitoring systems may face data transmission delays and bandwidth limitations. To address this, a cloud-edge collaborative architecture is being adopted, offloading some data processing and analysis tasks to edge devices. Edge devices can process critical data in real time, perform preliminary analysis, and issue fault warnings, reducing reliance on the cloud and improving response speed.

[0093] The cloud-edge collaborative architecture and rule engine allow operators to set monitoring and alarm rules based on specific needs. For example, specific alarm logic and response measures can be set according to different operating scenarios and environmental conditions, improving the flexibility and adaptability of the system.

[0094] Specifically, define the network delay weight ω = 0.6 and the data volume weight θ = 0.4; divide the current network delay (unit: milliseconds) by 100ms, and then multiply it by the weight ω (0.6); divide the data volume (unit: MB) by 1MB, and then multiply it by the weight θ (0.4); add the two together to get the decision value.

[0095] If the decision value is greater than 1.2, it indicates that the network latency is high or the data volume is large, making it unsuitable for real-time uploading to the cloud. Edge processing is initiated, and a fast Fourier transform (FFT) is performed to analyze the vibration spectrum.

[0096] If the decision value is less than or equal to 1.2, it means that the network condition is good, the data volume is moderate, and it is suitable for uploading to the cloud for in-depth analysis. In this case, the data is uploaded to the cloud and the LSTM-ATT model is used for more accurate fault prediction.

[0097] The system uses the Rete algorithm to achieve efficient multi-condition rule matching, and automatically triggers corresponding alarms and plans when specific fault characteristics are met.

[0098] It should be noted that this design enables the transformer monitoring system to adapt to different network environments and accurately identify and quickly respond to various fault risks.

[0099] Another possible implementation involves introducing a fog computing layer, building upon the existing cloud-edge collaborative architecture. Fog computing, situated between edge devices and the cloud, provides computing and storage capabilities closer to the data source. When edge devices have limited processing capabilities, some data can be sent to the fog computing layer for processing. For example, for complex data analysis tasks requiring significant computing resources, edge devices can send data to the fog computing layer, which then processes the data and feeds the results back to the edge device or uploads them to the cloud. This further reduces the burden on the cloud and improves overall system performance.

[0100] In this embodiment, when the rule engine triggers an early warning in step S4, it combines the expert knowledge base with association rule mining to provide fault diagnosis and maintenance solutions, including:

[0101] When equipment parameters exceed normal ranges or show abnormal trends, the system automatically triggers a fault warning and provides detailed repair recommendations to operators. These recommendations include possible causes, repair steps, and required materials, helping operators quickly locate and resolve the problem.

[0102] The system uses an interactive dashboard to display key operating parameters of the transformer (such as temperature, humidity, vibration, gas composition, etc.) and health index (HI) in real time. Through a combination of different colors and graphics, it intuitively reflects the operating status of the equipment.

[0103] Use time series charts to display the historical change trends of key parameters. When parameters fluctuate abnormally or exceed the normal range, the chart will highlight them, making it easier for operators to quickly identify potential problems.

[0104] Based on the results of condition assessment and fault prediction, a failure probability heat map is generated to show the possibility of failure of different components, helping operators focus on high-risk areas.

[0105] In another possible implementation, combining the expert knowledge base with association rule mining to provide fault diagnosis and maintenance solutions may also include:

[0106] The system dynamically adjusts the normal range of various parameters based on the transformer's historical operating data and environmental conditions. When real-time data exceeds the adaptive threshold, an early warning mechanism is triggered.

[0107] By mining association rules, we analyze the relationships between multiple parameters. For example, a temperature increase accompanied by an increase in the concentration of a specific gas may indicate a specific type of fault. The system makes comprehensive judgments based on these association rules, improving the accuracy of early warnings.

[0108] The system has a built-in fault pattern library containing characteristic patterns of different fault types. When the degree of match between real-time data and a fault pattern exceeds a set threshold, the system automatically identifies the fault type and provides appropriate repair recommendations.

[0109] Combining the knowledge and experience of power experts, a series of rules are developed. For example, "If the acetylene content in the dissolved gas in the oil increases significantly and is accompanied by an abnormal increase in temperature, an arc discharge fault may exist." The system makes judgments and issues warnings based on these rules.

[0110] Example 3. The above is a schematic scheme of the intelligent transformer status monitoring and cloud management method of this embodiment. It should be noted that the technical scheme of the intelligent transformer status monitoring and cloud management system and the technical scheme of the intelligent transformer status monitoring and cloud management method described above are based on the same concept. For details not described in detail in the technical scheme of the intelligent transformer status monitoring and cloud management system in this embodiment, please refer to the description of the technical scheme of the intelligent transformer status monitoring and cloud management method described above.

[0111] This embodiment also provides a method for intelligent monitoring and cloud management of transformer status, including:

[0112] Real-time acquisition module, which uses multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism to collect transformer operation data in real time;

[0113] The evaluation and prediction module is used to establish a deep learning model to perform status evaluation and fault prediction on the collected transformer operation data;

[0114] The early warning module is used to build a cloud-edge collaborative architecture and rule engine based on the results of status assessment and fault prediction, realizing dynamic task allocation and real-time early warning;

[0115] The solution generation module is used to provide fault diagnosis and maintenance solutions by combining the expert knowledge base and association rule mining when the rule engine triggers an early warning.

[0116] This embodiment further provides an electronic device applicable to the transformer status intelligent monitoring and cloud management method, including:

[0117] Memory and processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent monitoring and cloud management method of transformer status proposed in the above embodiment.

[0118] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for intelligent monitoring and cloud management of transformer status proposed in the above embodiment is implemented.

[0119] The storage medium proposed in this embodiment and the transformer status intelligent monitoring and cloud management method proposed in the above embodiment belong to the same inventive concept. For technical details not fully described in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for intelligent monitoring and cloud management of transformer status, characterized in that: include: Adopt multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism to collect transformer operation data in real time; Establish a deep learning model to perform condition assessment and fault prediction on the collected transformer operation data; Based on the results of status assessment and fault prediction, a cloud-edge collaborative architecture and rule engine are built to achieve dynamic task allocation and real-time warning. When the rule engine triggers an early warning, it combines the expert knowledge base with association rule mining to provide fault diagnosis and maintenance solutions.

2. The transformer status intelligent monitoring and cloud management method according to claim 1, characterized in that: The multi-sensor heterogeneous data fusion algorithm includes: Confidence fusion of sensor data: define the reliability factor of each sensor and dynamically adjust it through the entropy weight method.

3. The transformer status intelligent monitoring and cloud management method according to claim 2, characterized in that: The multi-sensor heterogeneous data fusion algorithm also includes: If the fusion confidence is lower than the preset threshold, the data is judged to be abnormal and the following operations are initiated: activating the adaptive sampling mechanism and pushing a preliminary warning to the edge.

4. The transformer status intelligent monitoring and cloud management method according to claim 3, characterized in that: The adaptive sampling mechanism includes: Parameters are selected to dynamically calculate the health index, a dynamic optimization method with regular updates is used for the health index, and adaptive sampling is performed based on the health index.

5. The transformer status intelligent monitoring and cloud management method according to claim 4, characterized in that: The adaptive sampling based on the health index includes: If the health index is greater than the first threshold, it is determined to be in normal state, and the sampling interval is set to 10 minutes; If the health index is greater than the second threshold and less than or equal to the first threshold, it is determined to be in a warning state, and the sampling interval is the larger value between 2 minutes and the original interval × the third threshold; If the health index is less than or equal to the second threshold, it is determined to be a fault state and 1-second continuous sampling of all sensors is started.

6. The transformer status intelligent monitoring and cloud management method according to claim 5, characterized in that: The establishment of a deep learning model to perform status assessment and fault prediction on the collected transformer operation data includes: Preprocess and extract features from the collected transformer operation data to generate feature vectors; A dynamic model of the transformer's normal operating status is established through a long short-term memory network, and the deviation between the current status and the baseline is compared in real time to evaluate the operating status. The feature vector is input into the deep learning model for condition assessment and fault prediction.

7. The transformer status intelligent monitoring and cloud management method according to claim 6, characterized in that: The cloud-edge collaborative architecture and rule engine based on the results of status assessment and fault prediction are constructed to achieve dynamic task allocation and real-time warning, including: Set specific alarm logic and response measures according to different operating scenarios and environmental conditions; A decision value is generated based on network conditions and data volume. If the decision value is greater than the set threshold, edge processing is initiated and a fast Fourier transform is performed to analyze the vibration spectrum. If the decision value is less than or equal to the set threshold, it is uploaded to the cloud for further fault prediction.

8. A transformer status intelligent monitoring and cloud management system, applying the method according to any one of claims 1 to 7, characterized in that: include: Real-time acquisition module, which uses multi-sensor heterogeneous data fusion algorithm and adaptive sampling mechanism to collect transformer operation data in real time; The evaluation and prediction module is used to establish a deep learning model to perform status evaluation and fault prediction on the collected transformer operation data; The early warning module is used to build a cloud-edge collaborative architecture and rule engine based on the results of status assessment and fault prediction, realizing dynamic task allocation and real-time early warning; The solution generation module is used to provide fault diagnosis and maintenance solutions by combining the expert knowledge base and association rule mining when the rule engine triggers an early warning.

9. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images