Electric transducer fault detection system and method based on big data

Abstract

The invention discloses a big data-based transformer fault detection system and method, and relates to the technical field of fault detection, and the method comprises the following steps: carrying out the feature extraction of multi-dimensional data of a transformer, and obtaining feature data; performing abnormal mode recognition according to the feature data to obtain an abnormal recognition result; time sequence trend analysis is carried out according to the abnormity identification result to obtain time sequence trend data; performing abnormal reason tracing according to the time sequence trend data to obtain abnormal reason data; performing fault factor comprehensive analysis according to the abnormal reason data to obtain fault factor associated information; performing fault mode modeling according to the fault factor association information to obtain a fault mode; performing health degree evaluation according to the fault mode to obtain an evaluation result; and carrying out fault classification according to an evaluation result. According to the method, the evolution trend of the fault can be effectively captured, the comprehensiveness and accuracy of fault prediction are improved, the health condition of the transformer can be reflected more accurately, and the accuracy and adaptability of fault diagnosis are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to a transformer fault detection system and method based on big data. Background Art

[0002] In modern power systems, transformers are key equipment whose operating status directly affects the security and stability of the power grid. Therefore, fault detection and prediction for transformers has become an important research direction.

[0003] In existing technologies, there are deficiencies in predicting the development of faults: existing fault detection technologies are mainly based on static data analysis, that is, by real-time monitoring of the physical and electrical parameters of the transformer, setting thresholds or using machine learning algorithms to determine whether there are any abnormalities at the moment. They can only analyze abnormalities in the current state, but cannot predict the development trend of abnormalities, and lack early warning capabilities.

[0004] Existing fault analysis techniques have shortcomings: Existing transformer fault analysis methods often rely on single-factor analysis, analyzing parameters such as temperature, current, voltage, and oil and gas composition separately and using feature extraction techniques to determine if these parameters are abnormal. However, in actual operation, transformer failures are often caused by the combined effects of multiple factors. For example, overload can cause temperature increases, which in turn affect oil and gas composition, reducing the accuracy of single-factor analysis methods. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a transformer fault detection system and method based on big data to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a transformer fault detection system and method based on big data, comprising the following steps:

[0008] S1, extracting the multi-dimensional data features of the transformer to obtain feature data;

[0009] S2. Perform abnormal pattern recognition based on the feature data to obtain abnormal recognition results;

[0010] S3. Perform time series trend analysis based on the anomaly recognition results to obtain time series trend data;

[0011] S4. Trace the cause of the abnormality based on the time series trend data to obtain abnormality cause data;

[0012] S5. Perform a comprehensive analysis of fault factors based on the abnormal cause data to obtain fault factor correlation information;

[0013] S6. Perform fault mode modeling based on fault factor correlation information to obtain a fault mode;

[0014] S7. Perform health assessment based on the failure mode and obtain assessment results;

[0015] S8. Classify the faults according to the evaluation results.

[0016] To further optimize this technical solution, the abnormal pattern recognition in S2 includes:

[0017] Based on the characteristic data, the support vector machine (SVM) technology is used to identify and classify the normal operating state and abnormal state of the transformer. When the data is classified as an abnormal mode, the specific category of the abnormality is further analyzed.

[0018] To further optimize this technical solution, the time series trend analysis in S3 includes:

[0019] According to the anomaly identification results, the fault prediction model is used to analyze the development trend of transformer faults, analyze the time evolution law of transformer abnormal data, calculate the change rate of abnormal parameters, and infer the future degree of abnormality.

[0020] To further optimize this technical solution, the fault prediction model includes:

[0021] ;

[0022] in:

[0023] : predicted value of the fault factor at time t;

[0024] : sensor data at time ti;

[0025] : External environmental factors that affect the health of transformers;

[0026] : Current abnormal data monitoring value;

[0027] , , : weight coefficient;

[0028] , : Attenuation coefficient of historical data and external environmental factors.

[0029] To further optimize this technical solution, the multi-fault correlation analysis in S5 includes:

[0030] According to the abnormal cause data, a comprehensive fault factor analysis model is used to analyze the correlation between multiple different factors that cause the fault and obtain the correlation information of the fault factors.

[0031] To further optimize this technical solution, the comprehensive analysis model of fault factors includes:

[0032] ;

[0033] in:

[0034] : comprehensive fault index at time t;

[0035] : The measured value of the i-th fault factor at time t;

[0036] : The measured value of the jth related factor at time t;

[0037] : The interaction weight coefficient between the i-th fault factor and the j-th related factor;

[0038] , : are the weight of the i-th fault factor and the influence coefficient of the external environmental factors respectively;

[0039] : External environmental factors at time t.

[0040] To further optimize this technical solution, the failure mode modeling in S6 includes:

[0041] According to the correlation information of fault factors, the hidden Markov model (HMM) technology is used to model the fault mode. By defining different operating states of the transformer, the transition probability between different states is calculated to describe the development process of the fault.

[0042] To further optimize this technical solution, the health evaluation in S7 includes:

[0043] Based on the obtained failure modes, the health assessment model is used to evaluate the health of the transformer. Combined with the transformer's real-time status, failure history, and external environmental factors, the health of the transformer is comprehensively calculated to quantify its health status.

[0044] To further optimize this technical solution, the health assessment model includes:

[0045] ;

[0046] in:

[0047] , : Health index at time t and t-1;

[0048] : Time decay factor of health status;

[0049] : represents the risk value of the i-th fault factor;

[0050] : the influence attenuation coefficient of the i-th fault factor;

[0051] : The dynamic weight of the i-th fault factor;

[0052] : the positive contribution value of the jth health impact factor;

[0053] : represents the impact attenuation coefficient of the jth health impact factor;

[0054] : The dynamic weight of the jth health impact factor.

[0055] A transformer fault detection system based on big data is constructed based on the above-mentioned transformer fault detection method based on big data, and includes the following functional modules:

[0056] Feature extraction module, abnormal pattern recognition module, trend analysis module, comprehensive analysis module of fault factors, and health assessment module.

[0057] In a second aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a transformer fault detection system and method based on big data as described in the first aspect of the present invention are implemented.

[0058] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a transformer fault detection system and method based on big data as described in the first aspect of the present invention are implemented.

[0059] Compared with the existing technology, the present invention provides a transformer fault detection system and method based on big data, which has the following beneficial effects:

[0060] This big data-based transformer fault detection system and method uses a fault prediction model to combine multi-dimensional data and the dynamic changes in the transformer health status to perform time series trend analysis, which can effectively capture the evolution trend of faults and improve the comprehensiveness and accuracy of fault prediction.

[0061] Through the comprehensive analysis model of fault factors, the correlation between multiple different factors that cause transformer failure can be considered, thereby more accurately reflecting the health status of the transformer and improving the accuracy and adaptability of fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] 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 work.

[0063] Figure 1 This is a flow chart of a transformer fault detection method based on big data proposed by the present invention;

[0064] Figure 2 This is a flow chart of a fault prediction model for a transformer fault detection method based on big data proposed by the present invention;

[0065] Figure 3 This is a flow chart of a comprehensive analysis model of fault factors in a transformer fault detection method based on big data proposed by the present invention;

[0066] Figure 4 This is a flow chart of a health assessment model for a transformer fault detection method based on big data proposed in the present invention;

[0067] Figure 5 This is a module schematic diagram of a transformer fault detection system based on big data proposed in the present invention. DETAILED DESCRIPTION

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0069] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0070] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0071] Example 1:

[0072] Reference Figures 1 to 4 , which is the first embodiment of the present invention, provides a transformer fault detection method based on big data, comprising the following steps:

[0073] S1. Extract multi-dimensional data features of transformer to obtain feature data.

[0074] In this embodiment, the multidimensional data feature extraction includes:

[0075] Principal component analysis (PCA) is used to extract key features from various monitoring data of transformers, including current, voltage, temperature, vibration, partial discharge signals, and gas content in oil. This removes redundant information, makes key features easier to identify, and screens out the characteristic components that have the greatest impact on the health status of the transformer. This reduces data dimensions, improves computational efficiency, and forms a low-dimensional feature data set.

[0076] S2. Perform abnormal pattern recognition based on the feature data to obtain abnormal recognition results.

[0077] In this embodiment, the abnormal pattern recognition includes:

[0078] Based on the characteristic data obtained in step S1, support vector machine (SVM) technology (existing technology) is used to identify and classify the normal operating state and abnormal state of the transformer. When the data is classified as an abnormal mode, the specific category of the abnormality is further analyzed, including partial discharge, winding deformation, and abnormal oil temperature.

[0079] S3. Perform time series trend analysis based on the anomaly recognition results to obtain time series trend data.

[0080] In this embodiment, the time series trend analysis includes:

[0081] Based on the anomaly identification results, a fault prediction model is used to analyze the development trend of transformer faults. The model analyzes the temporal evolution of transformer anomaly data, calculates the rate of change of anomaly parameters, and infers the severity of future anomalies. For example, for an oil temperature anomaly, the model predicts the likely temperature change over the next few hours or days, thereby inferring whether the transformer will enter an overheating failure state. This method overcomes the limitations of traditional static anomaly detection, enabling dynamic prediction of fault development and improving transformer fault early warning capabilities.

[0082] Furthermore, the fault prediction model includes:

[0083] ;

[0084] in:

[0085] : The predicted value of the fault factor at time t, including the predicted values ​​of winding temperature, partial discharge intensity, and oil and gas concentration;

[0086] : Sensor data at time ti, including winding temperature, partial discharge data, and oil and gas concentration;

[0087] : External environmental factors that affect the health of the transformer, including ambient temperature and humidity;

[0088] : Current abnormal data monitoring values, used to supplement real-time data to help adjust trend forecasts;

[0089] , , : weight coefficient;

[0090] , : The attenuation coefficient of historical data and external environmental factors is used to control the degree of influence of historical data and external environment.

[0091] The model describes how to predict failures by combining historical data, real-time data and external factors.

[0092] Traditional methods mainly rely on static anomaly detection and cannot fully consider the evolution process of faults. This model combines multi-dimensional data and the dynamic changes in the health status of transformers to predict faults, making fault prediction more comprehensive and accurate.

[0093] The steps for using this model include:

[0094] Data input: According to the characteristic data obtained in step S1, historical sensor data is obtained after processing and external environmental data , input the data into the model;

[0095] Real-time data correction: monitoring values ​​through current abnormal data The system uses real-time data as input to flexibly adjust predictions of future faults. When a fault state suddenly changes (such as sudden overheating or a sudden increase in partial discharge intensity), real-time data will be used as feedback to correct the model's prediction, ensuring that the fault prediction is always consistent with the current state.

[0096] Trend prediction: perform trend prediction based on input data and obtain prediction results , giving a fault prediction value at a certain moment in the future, helping maintenance personnel take measures in advance. For example, if the intensity of partial discharge has shown an upward trend in the past 10 minutes, and the predicted value calculated by the model indicates that this trend will continue to intensify in the next 30 minutes, then this information will help operation and maintenance personnel take measures in advance.

[0097] S4. Trace the cause of the abnormality based on the time series trend data to obtain abnormality cause data.

[0098] In this embodiment, the abnormality cause tracing includes:

[0099] Based on time-series trend data, a decision tree approach (existing technology) is used to construct a tree-like model of key factors affecting the transformer's operating status. For example, oil temperature changes may be caused by load fluctuations, cooling system failures, or ambient temperature changes. Combined with historical data, this model matches current abnormal trends, identifies the primary factors contributing to the abnormality, and analyzes the root causes of the abnormality. The advantage of the decision tree approach lies in its high interpretability. It can clearly demonstrate the path of abnormal conditions and help understand the logical relationships that lead to faults. For example, if the root cause of an abnormal pattern is determined to be winding overload, further load analysis can be conducted to confirm whether there is a problem of long-term overload operation.

[0100] S5. Perform a comprehensive analysis of the fault factors based on the abnormal cause data to obtain fault factor correlation information.

[0101] In this embodiment, the multi-fault correlation analysis includes:

[0102] Based on the anomaly cause data, a comprehensive fault factor analysis model is used to analyze the relationships between multiple factors that lead to the failure, thereby obtaining correlation information about the fault factors. This step reveals the impact of potential interactions between different factors on the failure risk, improving the accuracy of fault detection and diagnosis.

[0103] Furthermore, the comprehensive analysis model of fault factors includes:

[0104] ;

[0105] in:

[0106] : The comprehensive fault index at time t, reflecting the degree of fault risk after the interaction of multiple factors;

[0107] : The measured value of the i-th fault factor at time t, including partial discharge, winding temperature, and gas concentration;

[0108] : The measured value of the jth related factor at time t, including load, grid voltage, etc., indicating the factors related to the fault factor;

[0109] : The interaction weight coefficient between the i-th fault factor and the j-th related factor, reflecting the intensity of the interaction between them;

[0110] , : are the weight of the i-th fault factor and the influence coefficient of the external environmental factors respectively;

[0111] : External environmental factors at time t, including ambient temperature and humidity.

[0112] This model describes the impact of the correlation between different fault factors on the fault by modeling the mutual influence between various factors.

[0113] Traditional fault analysis methods often focus on the impact of a single fault, ignoring the potential correlations between multiple factors, resulting in reduced accuracy in fault prediction and analysis. This model comprehensively considers multiple factors and their interactions, avoiding the limitation of traditional methods that treat each factor as an independent event, thereby improving the accuracy of fault diagnosis. The weight coefficients in the model can be dynamically adjusted based on real-time data, allowing fault analysis to more accurately reflect current operating conditions and demonstrate strong adaptability.

[0114] Uses of the above models include:

[0115] Data acquisition: From step S4, multiple factors causing the failure are obtained, including the failure factor , related factors , external environmental factors , as the input of the model;

[0116] Interaction relationship calculation: Fault factors and related factors , through the interaction weight coefficient Calculate their interactions, learned from historical data and fault records;

[0117] Calculation of comprehensive fault index: Calculate the comprehensive fault index based on the acquired data and the interaction weight coefficient. , reflecting the failure risk under the interaction of multiple factors. Higher values ​​indicate a greater risk of failure, for example if partial discharge With high temperature The interaction between the two is strong, and external factors such as temperature changes also affect the failure, so It will rise significantly, indicating that the transformer may fail.

[0118] S6. Fault mode modeling is performed based on the fault factor correlation information to obtain the fault mode.

[0119] In this embodiment, the failure mode modeling includes:

[0120] Based on the correlation information of fault factors, a Hidden Markov Model (HMM) (existing technology) is used to model fault modes. By defining different operating states of the transformer—normal, mildly abnormal, moderately abnormal, and severely faulted—the transition probabilities between these states are calculated to describe the fault's progression. For example, a winding overheating condition may develop into a winding damage condition with a certain probability, while a partial discharge anomaly may evolve into an insulation breakdown. This method quantifies the evolution of fault modes through probabilistic modeling, enabling more accurate fault prediction.

[0121] S7. Perform health assessment based on the failure mode and obtain an assessment result.

[0122] In this embodiment, the health assessment includes:

[0123] Based on the obtained failure modes, the health assessment model is used to evaluate the health of the transformer. Combined with the transformer's real-time status, failure history, and external environmental factors, the health of the transformer is comprehensively calculated to quantify its health status.

[0124] Furthermore, the health assessment model includes:

[0125] ;

[0126] in:

[0127] , : Health index at time t and t-1, ranging from [0,1]. The larger the value, the better the health status.

[0128] : Time decay factor of health status, which controls the influence of past health status on current health, ranging from (0,1);

[0129] : represents the risk value of the i-th fault factor, including the abnormal value of winding temperature, the value of partial discharge intensity, and the change value of oil and gas concentration;

[0130] : The influence attenuation coefficient of the i-th fault factor, used to adjust its influence on the health level;

[0131] : The dynamic weight of the i-th fault factor, which is adjusted according to the real-time monitoring data to reflect its contribution to the health;

[0132] : The positive contribution value of the jth health influencing factor, including the cooling system operating status, maintenance quality, and grid load stability;

[0133] : represents the influence attenuation coefficient of the jth health influencing factor, which is used to control its influence range on health;

[0134] : The dynamic weight of the jth health influencing factor, reflecting its positive impact on health status.

[0135] The model comprehensively calculates the health of the transformer by combining the real-time status, fault history and external environmental factors of the transformer.

[0136] Traditional health assessment methods usually use simple weighting or scoring methods based on expert experience, but these methods have problems such as strong subjectivity and difficulty in dynamically adapting to changes in operating status. Compared with traditional methods, this model integrates multiple influencing factors and can dynamically adjust weights. It can more accurately describe the dynamic changes in health, avoid the limitations of linear models, maintain the continuity of health assessment, and make the health assessment results more comprehensive and objective.

[0137] The steps for using the model include:

[0138] Data input: The risk value of the failure factor obtained in step S6 and health impact factors The positive contribution value of is used as the input of the model;

[0139] Health calculation: combined with historical health index and dynamically adjusted weights and Calculate and obtain the current health index ;

[0140] Health assessment: Based on the calculated health index The system can analyze the decline rate of the transformer's health status and its changing trends, and predict the possible health level in the future to make maintenance decisions in advance.

[0141] S8. Classify the faults according to the evaluation results.

[0142] In this embodiment, the fault classification includes:

[0143] Based on the health assessment results, the K-means clustering algorithm (existing technology) is used to classify the data into different fault level categories, including normal, slightly abnormal, moderately abnormal, and severely abnormal. The fault severity of the transformer is classified and transformers with similar health status are grouped together to facilitate maintenance and management.

[0144] Example 2:

[0145] Reference Figure 5 , which is the second embodiment of the present invention, provides a transformer fault detection system based on big data, including the following functional modules:

[0146] Feature extraction module: extracts features from the collected multi-dimensional data of the transformer, extracts key features, and removes redundant information;

[0147] Abnormal pattern recognition module: classifies the normal operating state and abnormal state of the transformer and analyzes the specific category of the abnormality;

[0148] Trend analysis module: Analyzes the development trend of transformer faults, analyzes the time evolution of transformer abnormal data, calculates the rate of change of abnormal parameters, and estimates the degree of future abnormalities;

[0149] Comprehensive failure factor analysis module: Analyzes the correlation between different factors that lead to failures, and obtains the impact of potential interactions between different factors on failure risks;

[0150] Health Assessment Module: Evaluates the health of the transformer, comprehensively calculates the health of the transformer and quantifies its health status based on its real-time status, fault history, and external environmental factors.

[0151] Example 3:

[0152] This embodiment also provides a computer device, which is suitable for a transformer fault detection system and method based on big data, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement a transformer fault detection system and method based on big data as proposed in the above embodiment.

[0153] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, a transformer fault detection system and method based on big data as proposed in the above embodiment is implemented.

[0154] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0155] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0156] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0157] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0158] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0159] 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 transformer fault detection method based on big data, characterized in that: The following steps are involved: S1, extracting the multi-dimensional data features of the transformer to obtain feature data; S2. Perform abnormal pattern recognition based on the feature data to obtain abnormal recognition results; S3. Perform time series trend analysis based on the anomaly recognition results to obtain time series trend data; S4. Trace the cause of the abnormality based on the time series trend data to obtain abnormality cause data; S5. Perform a comprehensive analysis of fault factors based on the abnormal cause data to obtain fault factor correlation information; S6. Perform fault mode modeling based on fault factor correlation information to obtain a fault mode; S7. Perform health assessment based on the failure mode and obtain assessment results; S8. Classify the faults according to the evaluation results.

2. The transformer fault detection method based on big data according to claim 1, characterized in that: The abnormal pattern recognition in S2 includes: Based on the characteristic data, the support vector machine (SVM) technology is used to identify and classify the normal operating state and abnormal state of the transformer. When the data is classified as an abnormal mode, the specific category of the abnormality is further analyzed.

3. The transformer fault detection method based on big data according to claim 1, characterized in that: The time series trend analysis in S3 includes: According to the anomaly identification results, the fault prediction model is used to analyze the development trend of transformer faults, analyze the time evolution law of transformer abnormal data, calculate the change rate of abnormal parameters, and infer the future degree of abnormality.

4. The transformer fault detection method based on big data according to claim 3 is characterized in that: The fault prediction model includes: ； in: : predicted value of the fault factor at time t; : sensor data at time ti; : External environmental factors that affect the health of transformers; : Current abnormal data monitoring value; , , : weight coefficient; , : Attenuation coefficient of historical data and external environmental factors.

5. The transformer fault detection method based on big data according to claim 1, characterized in that: The multi-fault correlation analysis in S5 includes: According to the abnormal cause data, a comprehensive fault factor analysis model is used to analyze the correlation between multiple different factors that cause the fault and obtain the correlation information of the fault factors.

6. The transformer fault detection method based on big data according to claim 5, characterized in that: The comprehensive analysis model of fault factors includes: ； in: : comprehensive fault index at time t; : The measured value of the i-th fault factor at time t; : The measured value of the jth related factor at time t; : The interaction weight coefficient between the i-th fault factor and the j-th related factor; , : are the weight of the i-th fault factor and the influence coefficient of the external environmental factors respectively; : External environmental factors at time t.

7. The transformer fault detection method based on big data according to claim 1, characterized in that: The failure mode modeling in S6 includes: According to the correlation information of fault factors, the hidden Markov model (HMM) technology is used to model the fault mode. By defining different operating states of the transformer, the transition probability between different states is calculated to describe the development process of the fault.

8. The transformer fault detection method based on big data according to claim 1, characterized in that: The health assessment in S7 includes: Based on the obtained failure modes, the health assessment model is used to evaluate the health of the transformer. Combined with the transformer's real-time status, failure history, and external environmental factors, the health of the transformer is comprehensively calculated to quantify its health status.

9. The transformer fault detection method based on big data according to claim 8, characterized in that: The health assessment model includes: ； in: , : Health index at time t and t-1; : Time decay factor of health status; : represents the risk value of the i-th fault factor; : the influence attenuation coefficient of the i-th fault factor; : dynamic weight of the i-th fault factor; : the positive contribution value of the jth health impact factor; : represents the impact attenuation coefficient of the jth health impact factor; : The dynamic weight of the jth health impact factor.

10. A transformer fault detection system based on big data, constructed based on the transformer fault detection method based on big data according to any one of claims 1 to 9, characterized in that: Includes the following functional modules: Feature extraction module, abnormal pattern recognition module, trend analysis module, comprehensive analysis module of fault factors, and health assessment module.

Images