Fault identification diagnosis method and device based on energy storage system transformer, equipment and medium thereof
Patent Information
- Application Number
- CN202610428388.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-02
- Publication Date
- 2026-08-18
AI Technical Summary
[0008]本发明要解决的技术问题是：为了解决现有储能系统变压器难以构造输入量与输出量之间映射模型而影响故障识别诊断的精确度以及效率的技术问题，本发明提供一种基于储能系统变压器的故障识别诊断方法、装置、设备及其介质，通过对基于储能系统变压器的故障识别诊断方式的改进，以构建输入量与输出量之间的映射模型，从而提高储能系统变压器故障识别诊断的精确度以及效率
[0034]1、本发明通过构建输入量与输出量之间的映射模型，以实现储能系统变压器的故障识别诊断模型，相比于现有未构建的方式，该方式通过分析变压器健康状态与特征之间的映射关系，能够提高储能系统变压器故障识别诊断的精确度以及效率；此外，在进行识别诊断时，能够先将输入数据进行分类，再进行诊断识别，如此，能够进一步提高储能系统变压器故障识别诊断的精确度。
Smart Images
Figure CN122594792A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of transformer fault identification and diagnosis technology, and in particular to a fault identification and diagnosis method, device, equipment and medium based on transformers in energy storage systems. Background Technology
[0002] With the grid connection of various new energy power generation methods, the power grid signal exhibits strong transient characteristics, placing higher demands on the capacity, heat dissipation performance, and insulation performance of energy storage system transformers. Specifically, differences in insulation performance can lead to voltage breakdown and short circuits in the windings, resulting in major safety accidents. Furthermore, heat dissipation performance is crucial for the safe operation of the transformer. Excessive transformer temperature accelerates the insulation aging process, directly damaging the windings and ultimately exacerbating transformer failures. Because transformers operate under the influence of multiple factors, including electrical, thermal, mechanical forces, and environmental factors, their basic insulation material composition and heat dissipation performance are affected and damaged by internal and external factors, leading to a decline in transformer reliability. Therefore, we urgently need a fault identification and diagnosis method based on energy storage system transformers.
[0003] Currently, fault diagnosis methods for transformers in energy storage systems include:
[0004] 1. A power transformer fault diagnosis method based on classification decision tree: support vector machine is used to classify sample points close to the decision boundary, and decision tree is used to classify the remaining samples;
[0005] 2. The original signal is denoised using singular entropy, and a fault diagnosis method based on correlation vector machine optimized by quantum particle swarm optimization algorithm is proposed.
[0006] 3. Obtain transformer vibration and noise signals through a vibration and noise detection system, calculate feature values through FFT transformation, and predict transformer fault types by combining them with a BP neural network;
[0007] However, the above method makes it difficult to construct a mapping model between input and output quantities, which will affect the accuracy and efficiency of transformer fault identification and diagnosis in energy storage systems. Summary of the Invention
[0008] The technical problem to be solved by this invention is: to address the difficulty in constructing a mapping model between input and output quantities in existing energy storage system transformers, which affects the accuracy and efficiency of fault identification and diagnosis. This invention provides a fault identification and diagnosis method, device, equipment, and medium based on energy storage system transformers. By improving the fault identification and diagnosis method based on energy storage system transformers, a mapping model between input and output quantities is constructed, thereby improving the accuracy and efficiency of fault identification and diagnosis of energy storage system transformers.
[0009] The technical solution adopted by this invention to solve its technical problem is: a fault identification and diagnosis method based on an energy storage system transformer, comprising the following steps:
[0010] S1. Data Acquisition: Acquire the input and output data of the energy storage system transformer;
[0011] S2. Model Construction: Construct a fault identification and diagnosis model based on the transformer of the energy storage system;
[0012] S3, Training Model: Obtain the input data and output data in S1 and the fault identification and diagnosis model in S2, and use the input data and output data to train the fault identification and diagnosis model to obtain the optimal fault identification and diagnosis model.
[0013] S4. Real-time fault identification and diagnosis of energy storage system transformer: The input data of the energy storage system transformer is acquired in real time, and the optimal fault identification and diagnosis model in S3 is obtained. The real-time data is then input into the optimal model. The type of the input data is first determined, and then the fault identification and diagnosis result corresponding to the input data is output according to the optimal model.
[0014] Therefore, by constructing a mapping model between input and output quantities, a fault identification and diagnosis model for energy storage system transformers can be realized. Compared with existing methods that do not construct such models, this method can improve the accuracy and efficiency of fault identification and diagnosis of energy storage system transformers by analyzing the mapping relationship between the transformer's health status and characteristics. In addition, during the identification and diagnosis process, the input data can be classified first before the diagnosis and identification are performed, which can further improve the accuracy of fault identification and diagnosis of energy storage system transformers.
[0015] Furthermore, in S1, the input data includes: a large amount of high-dimensional monitoring data, a small amount of fault mode monitoring data, and unknown fault monitoring data; the output data is: the fault identification and diagnosis type of the energy storage system transformer.
[0016] Furthermore, in S2, the fault identification and diagnosis model based on the energy storage system transformer includes: a fault identification and diagnosis model based on single-class learning of comparative representation, a fault identification and diagnosis model based on few samples of comparative representation, and a fault identification and diagnosis model based on unknown faults.
[0017] Further, S2 includes the following steps:
[0018] S2-1. Obtaining the Sample Set: Obtaining the Positive Sample Set negative sample set ;
[0019] S2-2, Sample Set Feature Extraction: Construct the first feature extractor based on a deep neural network to obtain the positive sample set in S2-1. negative sample set And utilize this positive sample set negative sample set The first feature extractor is trained to obtain the trained first feature extractor;
[0020] Wherein: In S2-2, the training objective function of the first feature extractor is... The expression is:
[0021] ;
[0022] in: Indicates a sample, This represents numbers greater than or equal to. This represents the mapping function in the first feature extractor. Represents the negative sample set The number of negative samples. This enables early fault identification and diagnosis of transformers in energy storage systems.
[0023] Furthermore, in S2-1, for each positive sample data, negative samples are constructed using signal enhancement methods, and / or, time-domain and frequency-domain features are extracted using a squeezing transform as negative samples. Thus, in S2-1, the data constructed through data enhancement is used as negative samples. Because negative samples and positive samples are similar, they are closer in the feature space. This distinguishes positive and negative samples, making the boundaries of positive sample features in the feature space more compact, which is beneficial for early fault condition identification and diagnosis of transformers in energy storage systems. In S2-2, the training objective is to minimize the similarity between positive and negative sample features.
[0024] Further, S2 includes the following steps:
[0025] S2-A, Obtain the sample set: Treat all samples belonging to this fault mode as positive samples and all samples not belonging to this fault type as negative samples;
[0026] S2-B, Model Construction: A fault identification and diagnosis model based on comparative representation of a small number of samples is constructed using the distance between features. A rotating identification and diagnosis model is built for each fault mode. This enables fault identification and diagnosis of a small number of early-stage samples in energy storage system transformers.
[0027] Further, S2 includes the following steps:
[0028] S2-a, Feature Extraction: Obtain input data of unknown faults, mark them as new fault modes, and then use contrastive coding to extract features;
[0029] S2-b, Update Model: Obtain the new fault model from S2-b, construct an anomaly identification model, train it, and add the trained anomaly identification model to the anomaly identification model set to achieve incremental updates of the anomaly identification model set. This enables fault identification and diagnosis of unknown faults in energy storage system transformers.
[0030] A fault identification and diagnosis device based on an energy storage system transformer includes: a data acquisition module, a first fault identification and diagnosis module, and a second fault identification and diagnosis module. The data acquisition module is used to acquire input data and output data of the energy storage system transformer. The first fault identification and diagnosis module is used to construct a fault identification and diagnosis model based on the energy storage system transformer. The second fault identification and diagnosis module is used to train the first fault identification and diagnosis module, determine the type of the input data, and output the fault identification and diagnosis result corresponding to the input data.
[0031] An apparatus includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the fault identification and diagnosis method based on an energy storage system transformer.
[0032] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the aforementioned fault identification and diagnosis method based on an energy storage system transformer.
[0033] Compared with the prior art, the beneficial effects of the present invention are:
[0034] 1. This invention constructs a mapping model between input and output quantities to realize a fault identification and diagnosis model for transformers in energy storage systems. Compared with existing methods that do not construct such a model, this method can improve the accuracy and efficiency of fault identification and diagnosis of transformers in energy storage systems by analyzing the mapping relationship between the transformer's health status and characteristics. In addition, during the identification and diagnosis process, the input data can be classified first before the diagnosis and identification are performed, which can further improve the accuracy of fault identification and diagnosis of transformers in energy storage systems.
[0035] 2. In S2-1 of this invention, the data constructed through data augmentation is used as a negative sample. Since the negative sample is similar to the positive sample, they are close to each other in the feature space. Distinguishing between positive and negative samples makes the boundary of the positive sample features in the feature space more compact, which is beneficial for the early fault state identification and diagnosis of the transformer in the energy storage system.
[0036] 3. In S2-2 of this invention, the purpose of training is to minimize the similarity between the features of positive and negative samples. Attached Figure Description
[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0038] Figure 1 This is a flowchart of the fault identification and diagnosis method based on the transformer of an energy storage system according to the present invention;
[0039] Figure 2 This is a flowchart of the fault identification and diagnosis model based on contrastive representation single-class learning of the present invention;
[0040] Figure 3 The image shows the effect of the fault identification and diagnosis model based on contrastive representation single-class learning of the present invention.
[0041] Figure 4 This is a flowchart of the fault identification and diagnosis model based on comparative characterization of a small number of samples according to the present invention;
[0042] Figure 5 The image shows the effect of the fault identification and diagnosis model based on comparative representation of a small number of samples in this invention.
[0043] Figure 6 This is a flowchart of the fault identification and diagnosis model based on unknown faults of the present invention;
[0044] Figure 7 The image shows the effect of the fault identification and diagnosis model based on unknown faults of the present invention.
[0045] Figure 8 This is a schematic diagram of the fault identification and diagnosis device based on an energy storage system transformer according to the present invention;
[0046] Figure 9 This is a schematic diagram of the device of the present invention.
[0047] In the diagram: 1. Fault identification and diagnosis device;
[0048] 101. Data acquisition module; 102. First fault identification and diagnosis module; 103. Second fault identification and diagnosis module;
[0049] 2. Equipment;
[0050] 201. Memory; 202. Processor. Detailed Implementation
[0051] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.
[0052] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, features defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.
[0053] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0054] like Figures 1 to 7 As shown, a fault identification and diagnosis method based on an energy storage system transformer includes the following steps:
[0055] S1. Data Acquisition: Acquire the input and output data of the energy storage system transformer;
[0056] S2. Model Construction: Construct a fault identification and diagnosis model based on the transformer of the energy storage system;
[0057] S3, Training Model: Obtain the input data and output data in S1 and the fault identification and diagnosis model in S2, and use the input data and output data to train the fault identification and diagnosis model to obtain the optimal fault identification and diagnosis model.
[0058] S4. Real-time Fault Identification and Diagnosis of Energy Storage System Transformer: Input data of the energy storage system transformer is acquired in real time, and the optimal fault identification and diagnosis model from S3 is obtained. The real-time data is then input into the optimal model. The type of the input data is first determined, and then the fault identification and diagnosis result corresponding to the input data is output based on the optimal model. Thus, by constructing a mapping model between input and output quantities, a fault identification and diagnosis model for the energy storage system transformer is realized. Compared to existing methods without such a model, this method improves the accuracy and efficiency of fault identification and diagnosis of the energy storage system transformer by analyzing the mapping relationship between the transformer's health status and characteristics. Furthermore, during identification and diagnosis, the input data can be classified first before diagnostic identification, further improving the accuracy of fault identification and diagnosis of the energy storage system transformer.
[0059] In this embodiment, in S1, the input data includes: a large amount of high-dimensional monitoring data, a small amount of fault mode monitoring data, and unknown fault monitoring data; the output data is: the fault identification and diagnosis type of the energy storage system transformer.
[0060] In this embodiment, in S2, the fault identification and diagnosis model based on the energy storage system transformer includes: a fault identification and diagnosis model based on single-class learning of comparative representation, a fault identification and diagnosis model based on few samples of comparative representation, and a fault identification and diagnosis model based on unknown faults.
[0061] In this embodiment, step S2 includes the following steps:
[0062] S2-1. Obtaining the Sample Set: Obtaining the Positive Sample Set negative sample set ;
[0063] S2-2, Sample Set Feature Extraction: Construct the first feature extractor based on a deep neural network to obtain the positive sample set in S2-1. negative sample set And utilize this positive sample set negative sample set The first feature extractor is trained to obtain the trained first feature extractor;
[0064] Wherein: In S2-2, the training objective function of the first feature extractor is... The expression is:
[0065] ;
[0066] in: Indicates a sample, This represents numbers greater than or equal to. This represents the mapping function in the first feature extractor. Represents the negative sample set The number of negative samples;
[0067] In S2-1, for each positive sample data, negative samples are constructed using signal enhancement methods, and / or, time-domain and frequency-domain features are extracted using squeeze transform as negative samples. This enables early fault identification and diagnosis of transformers in energy storage systems. In S2-1, the data constructed through data augmentation is used as negative samples. Because negative samples and positive samples are similar, they are close in the feature space. Distinguishing between positive and negative samples makes the boundaries of positive sample features in the feature space more compact, which is beneficial for early fault state identification and diagnosis of transformers in energy storage systems. In S2-2, the training objective is to minimize the similarity between positive and negative sample features.
[0068] Specifically, if the monitoring data is large and high-dimensional, then sensitive features of health status are extracted from the monitoring data to construct a fault identification and diagnosis model based on contrastive representation single-class learning, and S2-1 and S2-2 are executed for early fault identification and diagnosis of transformers in energy storage systems.
[0069] Specifically, in actual use, energy storage system transformers only fail after long-term operation, so early fault identification and diagnosis of energy storage system transformers is essentially a single-class learning problem.
[0070] Specifically, contrastive learning is a type of deep learning self-supervised feature extraction method. Its core idea is to start from the data itself, and extract the features that make a certain data different from other data by comparing a certain data with other data in the dataset, without relying on supervised learning tasks such as classification.
[0071] Specifically, in S2-1, one-dimensional signal enhancement methods include, but are not limited to: adding noise, flipping, and random cropping; the negative samples constructed by the enhancement methods are similar to but different from the positive samples; the extracted time-domain features and frequency-domain features include, but are not limited to, voltage signals, current signals, and vibration signals; furthermore, the constructed negative samples and the negative samples formed by extracting features together constitute the final negative sample set. .
[0072] Specifically, in S2-1, the positive sample set All data in this document are health status monitoring data.
[0073] Specifically, in S2-2, for Train a first feature extractor as the mapping function. (soon) Mapping to its features ), and make:
[0074] ;
[0075] in: As a preset value, Indicates a sample, This represents a function that measures the similarity (i.e., distance) between sample features (the larger the value, the higher the similarity).
[0076] Specifically, in S2-2, the first feature extractor trains the objective function. middle, and The similarity of each negative sample is very small, approaching 0, while... and When the sum of the distances to the negative samples is large, it is greater than 0. Therefore, the training objective of the first feature extractor is to optimize its parameters so that the objective function... Obtain the minimum value.
[0077] In this embodiment, step S2 includes the following steps:
[0078] S2-A, Obtain the sample set: Treat all samples belonging to this fault mode as positive samples and all samples not belonging to this fault type as negative samples;
[0079] S2-B, Model Construction: A fault identification and diagnosis model based on comparative representation of a small number of samples is constructed using the distance between features. A rotating identification and diagnosis model is built for each fault mode. This enables fault identification and diagnosis of a small number of early-stage samples in energy storage system transformers.
[0080] Specifically, if there is limited monitoring data, a fault identification and diagnosis model based on comparative representation of a small number of samples is constructed, and S2-A and S2-B are executed for fault identification and diagnosis of transformers in energy storage systems (i.e., since there are few data samples for each fault mode, it is difficult to train a classifier and extract fault mode features using supervised training).
[0081] Specifically, in S2-A, all samples that do not belong to this fault type include: health status data samples.
[0082] Specifically, this method extracts the features that distinguish each fault mode from other fault modes and health states, which are used to identify whether a sample is in that fault mode. By integrating all fault mode recognition models, new data samples can be classified, and a fault identification and diagnosis model based on comparative representation of a small number of samples can be used for fault identification and diagnosis.
[0083] Specifically, a dedicated fault identification and diagnosis model is constructed for each fault mode. Since the commonly used single-classification model is not suitable for a small number of samples, the distance between features is directly used for fault mode identification and diagnosis. For all known fault modes, this method is used to construct the identification and diagnosis model, and the results are combined to perform fault identification and diagnosis.
[0084] In this embodiment, step S2 includes the following steps:
[0085] S2-a, Feature Extraction: Obtain input data of unknown faults, mark them as new fault modes, and then use contrastive coding to extract features;
[0086] S2-b, Update Model: Obtain the new fault model from S2-b, construct an anomaly identification model, train it, and add the trained anomaly identification model to the anomaly identification model set to achieve incremental updates of the anomaly identification model set. This enables fault identification and diagnosis of unknown faults in energy storage system transformers.
[0087] Specifically, in S2-a, the extracted features are: effective features characterizing the degradation of transformers in energy storage systems.
[0088] Specifically, after the energy storage system transformer starts operating, early fault identification and diagnosis detection is implemented until an abnormality is detected in the transformer. Fault diagnosis and identification detection are performed on the abnormality of the transformer based on known fault modes. If the fault is found to be not belonging to any known fault mode, then S2-a and S2-b are executed.
[0089] like Figure 8 As shown, a fault identification and diagnosis device based on an energy storage system transformer includes: a data acquisition module, a first fault identification and diagnosis module, and a second fault identification and diagnosis module. The data acquisition module is used to acquire input data and output data of the energy storage system transformer. The first fault identification and diagnosis module is used to construct a fault identification and diagnosis model based on the energy storage system transformer. The second fault identification and diagnosis module is used to train the first fault identification and diagnosis module, determine the type of the input data, and output the fault identification and diagnosis result corresponding to the input data.
[0090] like Figure 9 As shown, an apparatus includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a fault identification and diagnosis method based on an energy storage system transformer.
[0091] Specifically, device 2 can be a desktop computer, laptop, handheld computer, or cloud server, etc.; device 2 includes, but is not limited to, a memory 201 and a processor 202; those skilled in the art will understand that... Figure 5This is merely an example of device 2 and does not constitute a limitation on device 2. It may include more or fewer components than shown in the figure, or combine certain components, or different components (e.g., it may also include: input device 2, output device 2, network access device 2).
[0092] Specifically, the processor 202 can be a central processing unit (CPU), or it can be other general-purpose processors 202, digital signal processors 202 (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.; the general-purpose processor 202 can be a microprocessor 202 or any conventional processor 202, etc.
[0093] Specifically, in some embodiments, memory 201 may be an internal storage unit of device 2 (e.g., a hard disk or memory of device 2); in other embodiments, memory 201 may be an external storage device 2 of device 2 (e.g., a plug-in hard disk, smart media card (SMC), secure digital card (SD), or flash card equipped on device 2); further, memory 201 may include both internal storage units and external storage devices 2 of device 2; memory 201 is used to store operating systems, applications, bootloaders, data, and other programs (e.g., program code of computer programs); memory 201 may also be used to temporarily store data that has been output or will be output.
[0094] Specifically, those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or software. In addition, the specific names of each functional unit and module are only for easy differentiation and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.
[0095] Specifically, in the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0096] Specifically, those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware; whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution; those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0097] Specifically, in the embodiments provided in this application, it should be understood that the disclosed methods, apparatus, and device 2 can be implemented in other ways (for example, the apparatus and device 2 embodiments described above are merely illustrative; for example, the division of modules or units is merely a logical functional division, and there may be other division methods in actual implementation; for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed); another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, devices, or units, and can be electrical, mechanical, or other forms.
[0098] Specifically, the units described as discrete components may or may not be physically separate, and the parts shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0099] Specifically, in addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of software functional units.
[0100] Specifically, if an integrated module or unit 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, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable medium. When the computer program is executed by the processor 202, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory 201, a read-only memory (ROM), or a random access memory (RAM). Computer-readable media include memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in computer-readable media may be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction (for example, in some jurisdictions, according to legislation and patent practice, computer-readable media may not include electrical carrier signals and electrical signals).
[0101] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements a fault identification and diagnosis method based on an energy storage system transformer.
[0102] In summary, this invention constructs a mapping model between input and output quantities to realize a fault identification and diagnosis model for transformers in energy storage systems. Compared with existing methods that do not construct such a model, this method improves the accuracy and efficiency of fault identification and diagnosis of transformers in energy storage systems by analyzing the mapping relationship between the transformer's health status and characteristics. Furthermore, during the identification and diagnosis process, the input data can be classified first before the diagnosis and identification are performed, which further improves the accuracy of fault identification and diagnosis of transformers in energy storage systems.
[0103] The above description is based on the preferred embodiments of the present invention. Through the above description, those skilled in the art can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the contents of the specification, but must be determined by the scope of the claims.
Claims
1. A method for diagnosing a fault of a transformer based on an energy storage system, characterized by, Includes the following steps: S1. Data Acquisition: Acquire the input and output data of the energy storage system transformer; S2. Model Construction: Construct a fault identification and diagnosis model based on the transformer of the energy storage system; S3, Training Model: Obtain the input data and output data in S1 and the fault identification and diagnosis model in S2, and use the input data and output data to train the fault identification and diagnosis model to obtain the optimal fault identification and diagnosis model. S4. Real-time fault identification and diagnosis of energy storage system transformer: The input data of the energy storage system transformer is acquired in real time, and the optimal fault identification and diagnosis model in S3 is obtained. The real-time data is then input into the optimal model. The type of the input data is first determined, and then the fault identification and diagnosis result corresponding to the input data is output according to the optimal model.
2. The fault identification diagnostic method for an energy storage system based transformer according to claim 1, characterized by, In S1, the input data includes: A large amount of high-dimensional monitoring data, a small amount of fault mode monitoring data, and unknown fault monitoring data; The output data is: Types of transformer fault identification and diagnosis in energy storage systems.
3. The fault identification and diagnosis method based on an energy storage system transformer according to claim 1, characterized in that, In S2, the fault identification and diagnosis model based on the energy storage system transformer includes: Fault identification and diagnosis models based on contrastive representation single-class learning, fault identification and diagnosis models based on contrastive representation with few samples, and fault identification and diagnosis models based on unknown faults.
4. The fault identification and diagnosis method based on an energy storage system transformer according to claim 1, characterized in that, S2 includes the following steps: S2-1. Obtaining the Sample Set: Obtaining the Positive Sample Set negative sample set ; S2-2, Sample Set Feature Extraction: Construct the first feature extractor based on a deep neural network to obtain the positive sample set in S2-1. negative sample set And utilize this positive sample set negative sample set The first feature extractor is trained to obtain the trained first feature extractor; Wherein: In S2-2, the training objective function of the first feature extractor is... The expression is: ； in: Indicates a sample, This represents numbers greater than or equal to. This represents the mapping function in the first feature extractor. Represents the negative sample set The number of negative samples.
5. The fault identification and diagnosis method based on an energy storage system transformer according to claim 1, characterized in that, In S2-1, for each positive sample data, a negative sample is constructed using a signal enhancement method; And / or, The squeeze transform extracts time-domain and frequency-domain features as negative samples.
6. The fault identification and diagnosis method based on an energy storage system transformer according to claim 1, characterized in that, S2 includes the following steps: S2-A, Obtain the sample set: Treat all samples belonging to this fault mode as positive samples and all samples not belonging to this fault type as negative samples; S2-B, Model Construction: A fault identification and diagnosis model based on comparative representation of a small number of samples is constructed using the distance between features. A rotational identification and diagnosis model is constructed for each fault mode.
7. The fault identification and diagnosis method based on an energy storage system transformer according to claim 1, characterized in that, S2 includes the following steps: S2-a, Feature Extraction: Obtain input data of unknown faults, mark them as new fault modes, and then use contrastive coding to extract features; S2-b, Update Model: Obtain the new fault model in S2-b, construct the anomaly recognition model, train it, and add the trained anomaly recognition model to the anomaly recognition model set to achieve incremental update of the anomaly recognition model set.
8. A fault identification and diagnosis device based on an energy storage system transformer, characterized in that, include: The data acquisition module is used to acquire the input and output data of the transformer in the energy storage system. The first fault identification and diagnosis module is used to construct a fault identification and diagnosis model based on the transformer of the energy storage system. The second fault identification and diagnosis module is used to train the first fault identification and diagnosis module, determine the type of the input data, and output the fault identification and diagnosis result corresponding to the input data.
9. A device, characterized in that, include: The memory is used to store computer programs; A processor, when executing the computer program, implements the fault identification and diagnosis method based on an energy storage system transformer as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the fault identification and diagnosis method based on an energy storage system transformer as described in any one of claims 1-7.