Fault diagnosis method and system for operating mechanism of high-voltage circuit breaker

Abstract

The invention discloses a fault diagnosis method and system for a high-voltage circuit breaker operating mechanism. The method comprises the following steps: S1, collecting current-time data on an opening and closing coil in the opening and closing process of a high-voltage circuit breaker; s2, carrying out Grubm angle field conversion on the current-time data to obtain an image data set; and S3, constructing and training a fault classification model, and inputting the image data set into the trained model to obtain a fault diagnosis result. The system comprises a data acquisition module, a data conversion module and a fault diagnosis module. According to the method, the problem that the fault identification accuracy is relatively low due to the fact that an existing diagnosis method lacks information identification of the whole process of the current-time curve in the switching-on and switching-off process and only selects curve inflection point data is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis of a high-voltage circuit breaker operating mechanism, and in particular to a method and system for fault diagnosis of a high-voltage circuit breaker operating mechanism. Background Art

[0002] High-voltage circuit breakers are extremely important equipment in power systems, playing a role in controlling and protecting power systems. The normal operation of high-voltage circuit breakers is of great significance to improving the stability of power systems. As an important part of high-voltage circuit breakers, the health status of the operating mechanism directly affects the success of the circuit breaker opening and closing. With the increase of the operating time of high-voltage circuit breakers and the increase of the number of opening and closing, the operating mechanism will gradually deteriorate, wear, and become stuck, resulting in the failure of the high-voltage circuit breaker to open and close normally, affecting the normal operation of the power system.

[0003] Existing research on the health status of high-voltage circuit breaker operating mechanisms mainly focuses on the research parameters of the circuit breaker opening and closing coil current signal, the circuit breaker mechanical vibration signal and the circuit breaker travel-time characteristic curve.

[0004] Based on the above-mentioned health status diagnosis method of the high-voltage circuit breaker operating mechanism, there are the following defects and shortcomings: 1. Mechanical vibration signals are easily disturbed by environmental noise.

[0005] 2. It is difficult to collect the travel-time characteristic curve of the circuit breaker.

[0006] 3. Existing research on current signals based on opening and closing coils mainly focuses on the inflection point data of the current-time curve, lacks information extraction of the entire opening and closing process, and the accuracy of the diagnostic results is low. Summary of the invention

[0007] To solve the above problems, the present invention provides a fault diagnosis method and system for a high-voltage circuit breaker operating mechanism, which solves the problem that the existing diagnostic methods lack the identification of the entire process information of the current-time curve of the opening and closing process, and only select the curve inflection point data, resulting in low fault identification accuracy due to inaccurate point selection.

[0008] The present invention provides a fault diagnosis method for a high-voltage circuit breaker operating mechanism, and the specific technical solution is as follows: S1: Collect the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker; S2: Perform Gram angle field conversion on the current-time data to obtain an image data set; S3: construct and train a fault classification model, and input the image data set into the trained model to obtain a fault diagnosis result.

[0009] Furthermore, the construction and training of the fault classification model is as follows: M1: Obtain the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker, and perform data preprocessing; M2: Divide the image dataset obtained after data preprocessing into training set, validation set and test set according to the set ratio; M3: Build a fault classification model and input the training set and validation set into the model for training; M4: During the iterative training of the model, a data enhancement layer is added to the model for optimization according to the characteristics of the Gram matrix; M5: After iterative optimization, the obtained model is tested for accuracy on the test set to obtain the model with the highest fault identification accuracy, and integrated with the data preprocessing module. The sampled data is input into the integrated model to obtain the output high-voltage circuit breaker operating mechanism fault diagnosis results.

[0010] Furthermore, the data preprocessing is specifically performed as follows: M101: The current-time data on the opening and closing coils during the opening and closing process of the high-voltage circuit breaker are labeled according to the state of the circuit breaker operating mechanism represented by them to obtain a sequence data set; M102: Convert sequence datasets into image datasets.

[0011] Further, the state of the circuit breaker operating mechanism includes fault and normal; Fault data representing the fault status of the circuit breaker operating mechanism. Depending on the fault, the corresponding labels include abnormal voltage, abnormal circuit resistance, coil turn-to-turn short circuit, deformation of the moving iron core, and jamming of the moving iron core.

[0012] Furthermore, in step M102, the marked current-time data set is converted into a Gram matrix using the Gram angle field to obtain a converted image data set.

[0013] Furthermore, the specific process of converting the sequence data set into the image data set is as follows: Scale the time series values ​​and scale the data values ​​of the closing and opening coil current waveform to [-1,1]; For each data point after scaling, the polar coordinates corresponding to each data point are generated using time as the radius and the arc cosine of the data value as the angle; According to the vector inner product relationship of the opening and closing coil current waveform in polar coordinates, the polar coordinate expression of the opening and closing coil current waveform is converted into a Gram matrix, and an image is generated based on the Gram matrix.

[0014] Furthermore, in step M3, the fault classification model is constructed using ResNet as the model framework, a model with a batch size of 32 is used as the final classification model, and the optimizer is Adam.

[0015] The present invention further provides a fault diagnosis system for a high-voltage circuit breaker operating mechanism based on the above method, the system comprising a data acquisition module, a data conversion module and a fault diagnosis module; The data acquisition module is connected to the data conversion module, and the data acquisition module collects the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker, and transmits it to the data conversion module; The data conversion module is used to perform Gram angle field conversion on the current-time data to obtain an image data set; The fault diagnosis module is a trained fault classification model obtained based on the above method, receives image data, and performs fault diagnosis output.

[0016] The beneficial effects of the present invention are as follows: The present invention converts a two-dimensional current-time signal into an image signal, uses image data as a model input to implement fault diagnosis, retains the full process information of the opening and closing coil current signal, avoids the problem of only using inflection point current data when using the opening and closing coil current signal for fault diagnosis, which is susceptible to noise interference, easy to misjudge and has low accuracy, and improves the accuracy of fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the fault diagnosis process of the present invention.

[0018] Figure 2 It is a schematic diagram of the construction process of the fault classification model of the present invention.

[0019] Figure 3 This is a schematic diagram of the closing and opening coil current waveform under normal circumstances before scaling.

[0020] Figure 4 This is a schematic diagram of the scaled current waveform of the opening and closing coil under normal conditions.

[0021] Figure 5 It is a schematic diagram of the expression of the current waveform of the opening and closing coil in polar coordinates; Figure 6 This is a schematic diagram of the convolutional neural network structure after the image enhancement layer is introduced.

[0022] Figure 7 This is the confusion matrix diagram of the ResNet-Adam-32 model. DETAILED DESCRIPTION

[0023] The following description clearly and completely describes the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] In the description of the embodiments of the present invention, it should be noted that the indicated orientation or positional relationship is based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the invention product is usually placed when in use, or the orientation or positional relationship commonly understood by those skilled in the art, or the orientation or positional relationship in which the invention product is usually placed when in use, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first" and "second" are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0025] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0026] Example 1 Embodiment 1 of the present invention discloses a method for diagnosing a fault of a high voltage circuit breaker operating mechanism, such as Figure 1 As shown, the specific steps are as follows: S1: Collect the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker; S2: Perform Gram angle field conversion on the current-time data to obtain an image data set; Specifically, the current-time data set is converted into a Gram matrix using the Gram angle field, that is, the sequence data set is converted into an image data set; the specific conversion steps are described in step M102 below.

[0027] S3: construct and train a fault classification model, and input the image data set into the trained model to obtain a fault diagnosis result.

[0028] As a preferred embodiment, Figure 2 As shown in the figure, the construction and training process of the fault classification model is as follows: M1: Obtain the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker, and perform data preprocessing; As a preferred embodiment, the data preprocessing is specifically performed as follows: M101: The current-time data on the opening and closing coils during the opening and closing process of the high-voltage circuit breaker are labeled according to the state of the circuit breaker operating mechanism represented by them to obtain a sequence data set; As a preferred embodiment, the state of the circuit breaker operating mechanism includes fault and normal; Fault data representing the fault status of the circuit breaker operating mechanism. Depending on the fault, the corresponding labels include abnormal voltage, abnormal circuit resistance, coil turn-to-turn short circuit, deformation of the moving iron core, and jamming of the moving iron core.

[0029] M102: Use the Gram angular field to convert the labeled current-time dataset into a Gram matrix, that is, convert the sequence dataset into an image dataset.

[0030] As a preferred embodiment, the specific steps of the conversion include scaling, polar coordinate conversion and generating a Gram matrix, as follows: First, in order to make the inner product between the generated vectors not biased towards the maximum observed value, the time series value needs to be scaled. In this embodiment, the Min-Max Scaler is used to scale the time series value, and the data value of the closing coil current waveform is scaled to [-1,1]. The closing coil current signal before and after scaling is as follows: Figure 3 and Figure 4 shown.

[0031] For each data point after scaling, the polar coordinates corresponding to each data point are generated with time as the radius and the arc cosine of the data value as the angle, such as Figure 5 As shown; Suppose there is a set of vectors , take the matrix composed of the inner product of each pair of vectors, which is the Gram matrix, expressed as follows:

[0032] Each polar coordinate point is regarded as a vector in the vector group, and the Gram matrix is ​​calculated using the above formula. However, since the polar coordinate radius of each vector is determined by the time at which the data point is located, that is, the norm of each vector depends on time, the polar coordinate radius of the vector will increase with time, so that the inner product between vectors at two different times will be biased towards the vector generated by the data point at the closer time. In order to avoid such bias, the following inner product relationship is defined:

[0033]

[0034] in, and for and Angle expressed in polar coordinates.

[0035] It is the inner product formula of Gramian Angular Summation Fields, which means the inner product formula of summation, and focuses on the dynamic characteristics of time series; It is the Gramian Angular Difference Fields inner product formula, which means the difference inner product formula, and focuses on the changing trend of the time series.

[0036] According to the sum inner product formula and the difference inner product formula, the polar coordinate expression of the closing and opening coil current waveform is converted into a Gram matrix, and an image is generated based on the Gram matrix.

[0037] M2: Divide the image dataset obtained after data preprocessing into training set, validation set and test set according to 6:2:2.

[0038] M3: Build a fault classification model and input the training set and validation set into the model for training; M4: During the iterative training of the model, according to the characteristics of the Gram matrix, add appropriate data enhancement layers to the model for optimization, such as Figure 6 as stated; Specifically, the convolutional neural network model, optimizer and training batch size were screened, and a grid search comparison was used to comprehensively consider the accuracy and training time, and the ResNet network model, Adam optimizer and batch size of 32 were selected as the final classification model.

[0039] M5: After iterative optimization, the obtained model is tested for accuracy on the test set to obtain the model with the highest fault identification accuracy; Specifically, the test set data is input into the fault classification model ResNet-Adam-32 based on Gram angle field and convolutional neural network respectively, and the confusion matrix is ​​generated according to the classification results of the model, as shown in Figure 7 As shown in the figure, F1~F7 represent 7 types of faults, namely low voltage fault, high voltage fault, low resistance fault, high resistance fault, coil turn short circuit fault, iron core stuck fault, and iron core movement too long fault. The ordinate represents the real type of the predicted sample, and the abscissa represents the type predicted by the model for the data in the test set.

[0040] The trained convolutional neural network model is integrated with the data preprocessing part, and the sampled data is input into the integrated model to obtain the output fault diagnosis result of the high-voltage circuit breaker operating mechanism.

[0041] Example 2 Embodiment 2 of the present invention discloses a fault diagnosis system for a high voltage circuit breaker operating mechanism, such as Figure 1 As shown, the system includes a data acquisition module, a data conversion module and a fault diagnosis module; The data acquisition module is connected to the data conversion module, and the data acquisition module collects the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker, and transmits it to the data conversion module; The data conversion module is used to perform Gram angle field conversion on the current-time data to obtain an image data set; The fault diagnosis module is a trained fault classification model obtained based on the method described in the above embodiment 1, receives image data, and performs fault diagnosis output.

[0042] The present invention is not limited to the above-mentioned specific embodiments, but extends to any new features or any new combination disclosed in this specification, as well as any new method or process steps or any new combination disclosed.

Claims

1. A method for diagnosing a fault of a high voltage circuit breaker operating mechanism, characterized in that: include: S1: Collect the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker; S2: Perform Gram angle field conversion on the current-time data to obtain an image data set; S3: construct and train a fault classification model, and input the image data set into the trained model to obtain a fault diagnosis result.

2. The fault diagnosis method for a high-voltage circuit breaker operating mechanism according to claim 1, characterized in that: The construction and training of the fault classification model is as follows: M1: Obtain the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker, and perform data preprocessing; M2: Divide the image dataset obtained after data preprocessing into training set, validation set and test set according to the set ratio; M3: Build a fault classification model and input the training set and validation set into the model for training; M4: During the iterative training of the model, a data enhancement layer is added to the model for optimization according to the characteristics of the Gram matrix; M5: After iterative optimization, the obtained model is tested for accuracy on the test set to obtain the model with the highest fault identification accuracy, and integrated with the data preprocessing module. The sampled data is input into the integrated model to obtain the output high-voltage circuit breaker operating mechanism fault diagnosis results.

3. The fault diagnosis method for a high-voltage circuit breaker operating mechanism according to claim 2, characterized in that: The data preprocessing process is as follows: M101: The current-time data on the opening and closing coils during the opening and closing process of the high-voltage circuit breaker are labeled according to the state of the circuit breaker operating mechanism represented by them to obtain a sequence data set; M102: Convert sequence datasets into image datasets.

4. The fault diagnosis method for a high-voltage circuit breaker operating mechanism according to claim 3, characterized in that: The status of the circuit breaker operating mechanism includes fault and normal; Fault data representing the fault status of the circuit breaker operating mechanism. Depending on the fault, the corresponding labels include abnormal voltage, abnormal circuit resistance, coil turn-to-turn short circuit, deformation of the moving iron core, and jamming of the moving iron core.

5. The fault diagnosis method for a high-voltage circuit breaker operating mechanism according to claim 3, characterized in that: In step M102, the marked current-time data set is converted into a Gram matrix using the Gram angle field to obtain a converted image data set.

6. The fault diagnosis method for a high-voltage circuit breaker operating mechanism according to claim 5, characterized in that: The specific process of converting the sequence data set into an image data set is as follows: Scale the time series values ​​and scale the data values ​​of the closing and opening coil current waveform to [-1,1]; For each data point after scaling, the polar coordinates corresponding to each data point are generated using time as the radius and the arc cosine of the data value as the angle; According to the vector inner product relationship of the opening and closing coil current waveform in polar coordinates, the polar coordinate expression of the opening and closing coil current waveform is converted into a Gram matrix, and an image is generated based on the Gram matrix.

7. The fault diagnosis method for a high-voltage circuit breaker operating mechanism according to claim 2, characterized in that: In step M3, the fault classification model is constructed using ResNet as the model framework, a model with a batch size of 32 is used as the final classification model, and the optimizer is Adam.

8. A fault diagnosis system for a high voltage circuit breaker operating mechanism, characterized in that: The system includes a data acquisition module, a data conversion module and a fault diagnosis module; The data acquisition module is connected to the data conversion module, and the data acquisition module collects the current-time data on the opening and closing coil during the opening and closing process of the high-voltage circuit breaker, and transmits it to the data conversion module; The data conversion module is used to perform Gram angle field conversion on the current-time data to obtain an image data set; The fault diagnosis module is a trained fault classification model obtained based on any one of the methods described in claims 2-7, receives image data, and performs fault diagnosis output.

Images