CN115439667A - Method and system for transformer mechanical fault diagnosis based on sound field distribution map - Google Patents

Abstract

The invention discloses a transformer mechanical fault diagnosis method and system based on a sound field distribution diagram, wherein the transformer mechanical fault diagnosis method comprises the following steps: acquiring a sound field distribution diagram of a transformer to be diagnosed and a reference sound field distribution diagram of a transformer of the same type as the transformer to be diagnosed; acquiring the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram; acquiring a Pearson correlation coefficient between the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram; and judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient. The transformer mechanical fault diagnosis method based on the sound field distribution diagram can realize diagnosis of the transformer mechanical fault in a non-contact mode, and the transformer does not need to be stopped in the diagnosis process.

Description

Method and system for transformer mechanical fault diagnosis based on sound field distribution map

technical field

The invention belongs to the technical field of electric equipment fault diagnosis, relates to the field of transformer fault diagnosis, in particular to a transformer mechanical fault diagnosis method and system based on a sound field distribution map.

Background technique

In power systems, the vibration and noise generated by transformer faults are mainly measured and analyzed based on vibration signals, and a complete theoretical system and methods have been formed, such as: spectrum analysis technology, amplitude parameter analysis, shock pulse technology and resonance demodulation technology etc.

In the above-mentioned fault diagnosis method based on vibration signals in the prior art, the sensor must be arranged on the surface of the vibrating equipment; Analysis of the vibration signals of several isolated measuring points on the vibrating surface can only reflect the local vibration information of the equipment. In the case of installing a vibration sensor, shutting down the installation will bring greater economic losses; moreover, due to the diversity of equipment failures, the failure characteristics are also different, and the vibration characteristics are not obvious under some failures, while other characteristics (such as Acoustic features) are more obvious. To sum up, it is urgent to find an effective non-contact monitoring and analysis method.

Contents of the invention

The object of the present invention is to provide a transformer mechanical fault diagnosis method and system based on the sound field distribution diagram, so as to solve one or more technical problems mentioned above. The transformer mechanical fault diagnosis method based on the sound field distribution diagram provided by the present invention can realize the diagnosis of the transformer mechanical fault in a non-contact manner, and the transformer does not need to be shut down during the diagnosis process.

To achieve the above object, the present invention adopts the following technical solutions:

A method for diagnosing mechanical faults of transformers based on sound field distribution diagrams provided by the present invention comprises the following steps:

Obtaining a sound field distribution map of the transformer to be diagnosed, and a reference sound field distribution map of a transformer of the same type as the transformer to be diagnosed;

Obtaining the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map;

Obtaining the Pearson correlation coefficient between the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map;

The mechanical state of the transformer is judged based on the obtained Pearson correlation coefficient.

The further improvement of the present invention lies in that in the process of acquiring the sound field distribution map of the transformer to be diagnosed and the reference sound field distribution map of a transformer of the same type as the transformer to be diagnosed,

The steps of obtaining the sound field distribution map of the transformer to be diagnosed include: during the inspection process of the transformer, detecting the noise generated by the front, rear, left and right sides of the transformer through a microphone array, and obtaining multi-channel noise signals; The obtained multi-channel noise signal is drawn to obtain the sound field distribution map of the transformer to be diagnosed;

The step of obtaining the reference sound field distribution map of the transformer of the same type as the transformer to be diagnosed includes: during the normal operation of the transformer, the microphone array is used to detect the front, rear, left, and right sides of the transformer of the same type as the transformer to be diagnosed. The noise signal of the multi-channel noise signal is obtained; according to the multi-channel noise signal obtained in normal operation, the sound field distribution diagram of the different surfaces of the transformer is drawn, and the reference sound field distribution diagram of the transformer of the same type as the transformer to be diagnosed is obtained.

A further improvement of the present invention is that, during the detection process by the microphone array, the distance between the microphone array and the outline of the transformer box is 5 times the length of the outer facade of the corresponding transformer outline.

A further improvement of the present invention lies in that, in the process of acquiring the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map,

The HOG feature vector extraction method includes the following steps:

(1) Perform grayscale processing on the image of the input sound field distribution map to obtain a grayscale image; wherein, when performing grayscale processing, the gray value calculation expression is Gray=0.3×R+0.59×G+ 0.11×B (1);

In the formula, Gray represents the gray value of a pixel in the grayscale sound field distribution map; R, G, and B represent the red, green, and blue intensity values of the corresponding pixel in the source image, respectively;

(2) Calculate the horizontal and vertical gradients and directions of each pixel in the grayscaled image to obtain contour information;

Among them, the calculation expression of the gradient at the pixel point (x, y) is formula (2), (3),

G _x (x,y)=H(x+1,y)-H(x-1,y) (2);

G _y (x,y)=H(x,y+1)-H(x,y-1) (3);

In the formula, G _x (x, y), G _y (x, y), and H (x, y) are the horizontal gradient, vertical gradient, and gray value of the image after gray scale, respectively;

Among them, the calculation expressions of the gradient magnitude and direction at the pixel point (x, y) are formulas (4), (5),

(3) based on the contour information, divide the image into a plurality of cells, and obtain the feature vector of each cell;

(4) A preset number of cells is formed into a block; wherein, the eigenvectors of all cells in each block are concatenated to obtain the HOG eigenvector of the block;

(5) The block traverses the entire image with a step size of one cell, and finally the HOG feature vectors of all blocks in the image are concatenated to obtain the HOG feature vector of the image.

A further improvement of the present invention is that in the process of obtaining the Pearson correlation coefficient between the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map,

The calculation expression of the Pearson correlation coefficient between HOG eigenvectors is formula (6):

依次分别为待诊断变压器的声场分布图的HOG特征向量和基准声场分布图的HOG特征向量的样本平均值，σ_X,σ_Y依次分别为待诊断变压器的声场分布图的HOG特征向量和基准声场分布图的HOG特征向量的样本标准差，PCC值域为[-1,1]。In the formula, n is the number of HOG eigenvector data, X _i and Y _i are respectively the i-th sample of the HOG eigenvector of the sound field distribution diagram of the transformer to be diagnosed and the HOG eigenvector of the reference sound field distribution diagram,

are respectively the sample average values of the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map, σ _X , σ _Y are respectively the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the reference sound field The sample standard deviation of the HOG eigenvector of the distribution map, and the PCC value range is [-1,1].

A further improvement of the present invention is that in the process of judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient,

(1) When 0.95≤PCC, it means that the current sound field distribution diagram of the transformer to be diagnosed is strongly correlated with the reference sound field distribution diagram, and the mechanical condition is good;

(2) When 0.85≤PCC<0.95, it means that the correlation between the current sound field distribution map of the transformer to be diagnosed and the reference sound field distribution map is average, and the mechanical state has changed;

(3) When PCC<0.85, it means that the current sound field distribution diagram of the transformer to be diagnosed is weakly correlated with the reference sound field distribution diagram, and the mechanical state has changed seriously.

A transformer mechanical fault diagnosis system based on the sound field distribution diagram provided by the present invention includes:

A distribution map acquisition module, configured to obtain a sound field distribution map of a transformer to be diagnosed, and a reference sound field distribution map of a transformer of the same type as the transformer to be diagnosed;

An eigenvector acquisition module, configured to acquire the HOG eigenvector of the sound field distribution diagram of the transformer to be diagnosed and the HOG eigenvector of the reference sound field distribution diagram;

Pearson correlation coefficient acquisition module, used to obtain the Pearson correlation coefficient between the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map;

The mechanical state acquisition module is used for judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient.

A further improvement of the present invention is that, in the process of obtaining the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map by the feature vector acquisition module,

The HOG feature vector extraction method includes the following steps:

(1) Perform grayscale processing on the image of the input sound field distribution map to obtain a grayscale image; wherein, when performing grayscale processing, the gray value calculation expression is Gray=0.3×R+0.59×G+ 0.11×B (1);

In the formula, Gray represents the gray value of a pixel in the grayscale sound field distribution map; R, G, and B represent the red, green, and blue intensity values of the corresponding pixel in the source image, respectively;

(2) Calculate the horizontal and vertical gradients and directions of each pixel in the grayscaled image to obtain contour information;

Among them, the calculation expression of the gradient at the pixel point (x, y) is formula (2), (3),

G _x (x,y)=H(x+1,y)-H(x-1,y) (2);

G _y (x,y)=H(x,y+1)-H(x,y-1) (3);

In the formula, G _x (x, y), G _y (x, y), and H (x, y) are the horizontal gradient, vertical gradient, and gray value of the image after gray scale, respectively;

Among them, the calculation expressions of the gradient magnitude and direction at the pixel point (x, y) are formulas (4), (5),

(3) based on the contour information, divide the image into a plurality of cells, and obtain the feature vector of each cell;

(4) A preset number of cells is formed into a block; wherein, the eigenvectors of all cells in each block are concatenated to obtain the HOG eigenvector of the block;

(5) The block traverses the entire image with a step size of one cell, and finally the HOG feature vectors of all blocks in the image are concatenated to obtain the HOG feature vector of the image.

A further improvement of the present invention is that the Pearson correlation coefficient acquisition module implements the process of obtaining the Pearson correlation coefficient between the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map,

The calculation expression of the Pearson correlation coefficient between HOG eigenvectors is formula (6):

依次分别为待诊断变压器的声场分布图的HOG特征向量和基准声场分布图的HOG特征向量的样本平均值，σ_X,σ_Y依次分别为待诊断变压器的声场分布图的HOG特征向量和基准声场分布图的HOG特征向量的样本标准差，PCC值域为[-1,1]。In the formula, n is the number of HOG eigenvector data, X _i and Y _i are respectively the i-th sample of the HOG eigenvector of the sound field distribution diagram of the transformer to be diagnosed and the HOG eigenvector of the reference sound field distribution diagram,

are respectively the sample average values of the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map, σ _X , σ _Y are respectively the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the reference sound field The sample standard deviation of the HOG feature vector of the distribution map, and the PCC value range is [-1,1].

A further improvement of the present invention lies in that the mechanical state acquisition module implements the process of judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient,

(1) When 0.95≤PCC, it means that the current sound field distribution diagram of the transformer to be diagnosed is strongly correlated with the reference sound field distribution diagram, and the mechanical condition is good;

(2) When 0.85≤PCC<0.95, it means that the correlation between the current sound field distribution map of the transformer to be diagnosed and the reference sound field distribution map is average, and the mechanical state has changed;

(3) When PCC<0.85, it means that the current sound field distribution diagram of the transformer to be diagnosed is weakly correlated with the reference sound field distribution diagram, and the mechanical state has changed seriously.

Compared with the prior art, the present invention has the following beneficial effects:

In the transformer mechanical fault diagnosis method based on the sound field distribution diagram specifically provided by the present invention, the transformer is used as the sound source, and the noise signals generated under various mechanical states of the transformer are detected through the microphone array in a non-contact manner, and the noise signals are drawn according to the detected noise signals. The sound field distribution under the transformer, so as to obtain the HOG characteristics of the sound field distribution map under the transformer; in the process of judging the mechanical state of the transformer to be tested, it is only necessary to collect the noise signal generated during the working process of the transformer to be tested through the microphone array, and use the comparison method , the mechanical state of the transformer to be tested can be known, so as to determine whether the transformer to be tested has a mechanical failure, and there is no need to shut down the transformer to be tested, and the detection cost is low, and the sensitivity and accuracy are high.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art; obviously, the accompanying drawings in the following description are For some embodiments of the present invention, those skilled in the art can also obtain other drawings based on these drawings without creative effort.

Fig. 1 is a schematic flow chart of a method for diagnosing a mechanical fault of a transformer based on a sound field distribution diagram provided by an embodiment of the present invention;

Fig. 2 is a schematic diagram of the HOG feature extraction of the transformer sound field distribution diagram in the embodiment of the present invention;

Fig. 3 is a schematic diagram of generation and propagation of transformer vibration noise in an embodiment of the present invention.

detailed description

In order to enable those skilled in the art to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only It is an embodiment of a part of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

It should be noted that the terms "first" and "second" in the description and claims of the present invention and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the invention described herein can be practiced in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having", as well as any variations thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or device comprising a sequence of steps or elements is not necessarily limited to the expressly listed instead, may include other steps or elements not explicitly listed or inherent to the process, method, product or apparatus.

The present invention is described in further detail below in conjunction with accompanying drawing:

Referring to Fig. 1, a method for diagnosing a mechanical fault of a transformer based on a sound field distribution diagram provided by an embodiment of the present invention includes the following steps:

1) During the normal working process of the transformer, the noise signals generated on the front, rear, left, and right sides of a certain type of transformer are detected through the microphone array; the sound field distribution diagrams of different sides of the transformer are drawn according to the multi-channel noise signals, and the transformer under normal conditions is established. The sample library of sound field distribution map, which is used as the reference sound field distribution map of this type of transformer;

2) During the transformer inspection process, obtain the sound field distribution diagram of the transformer to be tested in the current state according to the method in step 1), and compare it with the reference sound field distribution diagram to judge the mechanical state of the transformer;

3) Obtain its directional gradient histogram feature vector (Histogram of Oriented Gradient) for the transformer sound field distribution map obtained in step 1) and step 2) respectively, referred to as HOG feature vector;

4) Step 3) obtains the Pearson correlation coefficient between the HOG eigenvectors under the different states of the transformer obtained;

5) Judging the current mechanical state of the transformer according to the Pearson correlation coefficient.

In step 1) and step 2) of the embodiment of the present invention, the sound field distribution diagram of the four faces of the transformer is obtained through the acoustic microphone array, wherein the distance between the microphone array and the outline of the transformer box is related to the size of the transformer, and the distance is selected as the corresponding transformer 5 times the length of the outer facade of the contour line to ensure that one surface of the entire transformer can be covered by the collection range of the microphone array, and the height of the sensor is half of the height of the transformer box.

In step 3) of the embodiment of the present invention, the HOG feature vector extraction method of the transformer sound field distribution map is as follows:

(1) Grayscale the image of the input sound field distribution map according to formula (1). R, G, and B respectively represent the red, green, and blue intensity values in the source image. The result after grayscale is shown in Figure 2.

Gray＝0.3×R+0.59×G+0.11×B (1)

In addition, in order to reduce the influence caused by local shadows and illumination changes of the image, and at the same time suppress the interference of noise, the Gamma correction method is generally used to normalize the color space of the input image. However, the sound field distribution diagram in the present invention is not a photographed image, and there is no such influence, so Gamma correction may not be performed.

(2) Calculate the horizontal and vertical gradients and directions of each pixel in the grayscale image, so that the contour information can be captured. The calculation method of the gradient at the pixel point (x, y) in the figure is:

G _x (x,y)=H(x+1,y)-H(x-1,y) (2)

G _y (x,y)=H(x,y+1)-H(x,y-1) (3)

In the formula, G _x (x, y), G _y (x, y), H (x, y) are the horizontal gradient, vertical gradient, and pixel value respectively in turn;

Then the gradient magnitude and direction at the pixel point are:

(3) Divide the image into many small cells (cells), as shown in Figure 2, the present invention specifically exemplifies the 8*8 pixels of the sound field distribution map into a cell, and then constructs a weight for each cell Gradient direction histogram; the direction of the gradient can be divided into several blocks, usually choose to divide 360 degrees into 9 direction block ranges. As a specific example of the embodiment of the present invention, if the gradient direction at a certain pixel point is 36 degrees, and the gradient value is 5, then the value of the direction block of 0-40 degrees in the histogram is increased by 5. Traverse all the pixels in the cell in turn, and finally get the 9-dimensional feature vector of the cell, as shown in Figure 2, the Orientationhistogram bin is the polar coordinate expression of the histogram.

(4) Every few cells form a block. In the present invention, 2*2 cells form a block. The eigenvectors of all the cells in a block are concatenated to obtain the HOG eigenvector of the block.

(5) Traverse the block from left to right and from top to bottom in the whole image with a step of one cell, and finally concatenate the HOG feature vectors of all blocks in the image to obtain the HOG feature vector of the image. After obtaining the image After the HOG eigenvector of , it can be used as a mathematical description of the sound field distribution map.

In step 4) of the embodiment of the present invention, the calculation method of the Pearson correlation coefficient of the HOG eigenvector of the transformer sound field distribution diagram in different states is as follows:

Find the Pearson correlation coefficient between the HOG eigenvectors of the sound field distribution diagrams of different working conditions:

为样本平均值，σ_X为样本标准差，PCC值域为[-1,1]。具体示例性的，一般认为相关系数大于0.9则具有强相关性。In the formula,

is the sample mean, σ _X is the sample standard deviation, and the PCC value range is [-1,1]. As a specific example, it is generally believed that the correlation coefficient is greater than 0.9, which means there is a strong correlation.

In step 5) of the embodiment of the present invention, the method for judging the current mechanical state of the transformer according to the Pearson correlation coefficient is as follows:

According to the Pearson correlation coefficient of the sound field distribution diagram of different surfaces, the mechanical status of the four surfaces of the same transformer is judged respectively, so as to roughly judge the location of the mechanical failure.

(1) When 0.95≤PCC, it means that the current sound field distribution map of the transformer to be tested is strongly correlated with the reference sound field distribution map, and the current mechanical condition is good;

(2) When 0.85≤PCC<0.95, it means that the current sound field distribution map of the transformer to be tested has a general correlation with the reference sound field distribution map, and the current mechanical state has changed, which needs to be monitored by combining various means;

(3) When PCC<0.85, it means that the current sound field distribution diagram of the transformer to be tested is weakly correlated with the reference sound field distribution diagram, and the current mechanical state has undergone serious changes. A certain side of the transformer whose sound field distribution diagram changes beyond the standard, combined with multiple methods for fault diagnosis. At the same time, it accepts detailed procedure tests to avoid large-scale power outages caused by sudden transformer failures.

In the specific operation of the transformer mechanical fault diagnosis method based on acoustic imaging technology described in the present invention, since the noise signals generated by the transformer are different, the present invention uses the transformer as the sound source, and detects various mechanical states of the transformer in a non-contact manner through the microphone array The noise signal generated under the transformer, and the sound field distribution under the transformer is drawn according to the detected noise signal, so as to obtain the HOG characteristics of the sound field distribution map under the transformer. The noise signal generated during the working process can be compared with the mechanical state of the transformer to be tested, so as to determine whether there is a mechanical failure in the transformer to be tested, and there is no need to stop the transformer to be tested, and the cost of detection at the same time lower, and higher sensitivity and accuracy.

The feasibility analysis of the diagnostic method provided by the foregoing embodiments of the present invention is as follows:

Transformer vibration is caused by the vibration of the transformer body and the vibration of the cooling device. The transformer body mainly includes the iron core and windings. The vibration generation and propagation process is shown in Figure 3. Among them, the vibration of the winding is mainly caused by the dynamic electromagnetic force of the coil with the alternating current in the leakage magnetic field, and the vibration of the iron core is mainly caused by the magnetostriction of the silicon steel sheet and the electromagnetic force caused by the eddy current between the silicon steel sheets. It can be seen from Figure 3 that the vibration and noise signals of the transformer can most directly reflect the change of its internal mechanical structure, so it is self-evident to use it as the basis for the diagnosis of the transformer state in terms of scientificity and reliability.

Fault diagnosis of transformers is to identify whether the operating state of the transformer is normal, then determine the nature, location and cause of the fault, and finally propose measures to solve the fault. The purpose of fault diagnosis is to judge the mechanical state of the system through mapping according to the measurable eigenvectors. The conventional acoustic diagnosis technology mainly measures the acoustic signal through a limited number of measuring points, and then uses the signal processing technology to extract the characteristics of the acoustic signal for fault diagnosis. Therefore, the conventional acoustic diagnosis technology can be regarded as pattern recognition based on the acoustic signal, and the noise signal is a vector signal. Therefore, the recognition feature is extracted from the vector signal, and the sound field volume distribution of the equipment sound field is obtained by using array measurement and acoustic imaging technology. When performing feature extraction, it is necessary to use matrix or image feature extraction technology. Therefore, using acoustic The essence of image fault diagnosis is fault diagnosis based on audio-visual pattern recognition. In the method provided by the embodiment of the present invention, when performing noise signal detection, the microphone array is first fixed on the array support, and then the array support is fixed around the transformer, and the noise signal generated by the transformer is detected by the microphone array, and then the data acquisition system is used for each The noise signals collected by the microphones are collected synchronously, and then input into the computer. The computer uses acoustic imaging technology to draw the sound field distribution of the transformer to be tested according to the noise signals collected by each microphone and the position of each microphone. The data acquisition system collects the electrical signals of multi-channel microphones synchronously. In addition to the requirements of dynamic range, number of digits, anti-aliasing filter and built-in IEPE excitation that need to be paid attention to in individual sound collection, the application of microphone arrays also requires the inter-channel The phase consistency is good, and the acquisition device with independent ADC for each channel is generally selected.

The following are device embodiments of the present invention, which can be used to implement the method embodiments of the present invention. For details not omitted in the device embodiment, please refer to the method embodiment of the present invention.

An embodiment of the present invention provides a transformer mechanical fault diagnosis system based on a sound field distribution diagram, including:

A distribution map acquisition module, configured to obtain a sound field distribution map of a transformer to be diagnosed, and a reference sound field distribution map of a transformer of the same type as the transformer to be diagnosed;

An eigenvector acquisition module, configured to acquire the HOG eigenvector of the sound field distribution diagram of the transformer to be diagnosed and the HOG eigenvector of the reference sound field distribution diagram;

Pearson correlation coefficient acquisition module, used to obtain the Pearson correlation coefficient between the HOG feature vector of the sound field distribution map of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution map;

The mechanical state acquisition module is used for judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient.

Those skilled in the art should understand that the embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowcharts and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the present invention can still be Any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention shall fall within the protection scope of the claims of the present invention.

Claims (10)

1. A transformer mechanical fault diagnosis method based on a sound field distribution diagram is characterized by comprising the following steps:

acquiring a sound field distribution diagram of a transformer to be diagnosed and a reference sound field distribution diagram of a transformer of the same type as the transformer to be diagnosed;

acquiring the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram;

acquiring a Pearson correlation coefficient between the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram;

and judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient.

2. The method for diagnosing the mechanical fault of the transformer based on the sound field distribution diagram according to claim 1, wherein in the process of acquiring the sound field distribution diagram of the transformer to be diagnosed and the reference sound field distribution diagram of the transformer with the same model as the transformer to be diagnosed,

the step of acquiring the sound field distribution map of the transformer to be diagnosed comprises the following steps: in the transformer inspection process, detecting the noise generated by the front, the back, the left and the right surfaces of the transformer by a microphone array to obtain a multi-channel noise signal; drawing and obtaining a sound field distribution diagram of the transformer to be diagnosed based on the multi-channel noise signals obtained in the routing inspection process;

the step of acquiring the reference sound field distribution map of the transformer with the same model of the transformer to be diagnosed comprises the following steps: in the normal working process of the transformer, detecting noise signals generated on the front, rear, left and right surfaces of the transformer with the same type as the transformer to be diagnosed by a microphone array to obtain a multi-channel noise signal; and drawing sound field distribution diagrams of different surfaces of the transformer according to the multi-channel noise signals obtained in normal work, and obtaining the reference sound field distribution diagram of the transformer with the same type of the transformer to be diagnosed.

3. The method for diagnosing the mechanical fault of the transformer based on the sound field distribution diagram according to claim 2, wherein in the process of detection through the microphone array, the distance between the microphone array and the contour line of the transformer tank is 5 times of the length of the facade of the contour line of the corresponding transformer.

4. The method for diagnosing the mechanical fault of the transformer based on the sound field distribution diagram according to claim 1, wherein in the process of obtaining the HOG feature vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution diagram,

the HOG feature vector extraction method comprises the following steps:

(1) Carrying out graying processing on the input image of the sound field distribution diagram to obtain a grayed image; wherein, when the graying processing is performed, the Gray value calculation expression is Gray =0.3 × R +0.59 × G +0.11 × B (1);

in the formula, gray represents the Gray value of a certain pixel point of the grayed sound field distribution diagram; r, G and B respectively represent red, green and blue intensity values of corresponding pixel points in the source image;

(2) Calculating the transverse and longitudinal gradients and directions of each pixel point in the grayed image to obtain contour information;

wherein, the calculation expressions of the gradient at the pixel point (x, y) are formulas (2) and (3),

G _x (x,y)＝H(x+1,y)-H(x-1,y) (2)；

G _y (x,y)＝H(x,y+1)-H(x,y-1) (3)；

in the formula, G _x (x,y)、G _y (x, y) and H (x, y) are respectively the gray scale of the image after horizontal gradient, longitudinal gradient and gray scale in sequenceA value;

wherein, the calculation expressions of the gradient amplitude and the direction at the pixel point (x, y) are formulas (4) and (5),

(3) Dividing the image into a plurality of cells based on the contour information, and acquiring a feature vector of each cell;

(4) Forming a block by a preset number of unit cells; the method comprises the following steps that feature vectors of all cells in each block are connected in series to obtain HOG feature vectors of the block;

(5) Traversing the block through the whole image by the step length of one cell, and finally connecting HOG characteristic vectors of all blocks in the image in series to obtain the HOG characteristic vector of the image.

5. The method for diagnosing the mechanical fault of the transformer based on the sound field distribution diagram according to claim 4, wherein in the process of obtaining the Pearson correlation coefficient between the HOG eigenvector of the sound field distribution diagram of the transformer to be diagnosed and the HOG eigenvector of the reference sound field distribution diagram,

the computational expression of Pearson correlation coefficients between HOG feature vectors is formula (6):

wherein n is the number of HOG feature vector data, X _i ,Y _i Sequentially and respectively obtaining the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the ith sample of the HOG characteristic vector of the reference sound field distribution diagram,

sequentially and respectively obtaining the average value, sigma, of the samples of the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram _X ,σ _Y Sequentially and respectively obtaining sample standard deviations of the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram, wherein the PCC value range is [ -1,1]。

6. The method for diagnosing the mechanical fault of the transformer based on the sound field distribution diagram according to claim 5, wherein in the process of judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient,

(1) When the PCC is not more than 0.95, the current sound field distribution diagram of the transformer to be diagnosed is strongly related to the reference sound field distribution diagram, and the mechanical state is good;

(2) When the PCC is more than or equal to 0.85 and less than 0.95, the correlation between the current sound field distribution diagram and the reference sound field distribution diagram of the transformer to be diagnosed is general, and the mechanical state changes;

(3) When PCC is less than 0.85, the current sound field distribution diagram of the transformer to be diagnosed is weakly correlated with the reference sound field distribution diagram, and the mechanical state is seriously changed.

7. A transformer mechanical fault diagnosis system based on sound field distribution diagram is characterized by comprising:

the distribution diagram acquisition module is used for acquiring a sound field distribution diagram of the transformer to be diagnosed and a reference sound field distribution diagram of the transformer with the same model as the transformer to be diagnosed;

the characteristic vector acquisition module is used for acquiring the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram;

a Pearson correlation coefficient obtaining module, configured to obtain a Pearson correlation coefficient between the HOG feature vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution diagram;

and the mechanical state acquisition module is used for judging the mechanical state of the transformer based on the acquired Pearson correlation coefficient.

8. The system of claim 7, wherein the feature vector obtaining module implements obtaining the HOG feature vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG feature vector of the reference sound field distribution diagram,

the HOG feature vector extraction method comprises the following steps:

(1) Performing graying processing on the image of the input sound field distribution diagram to obtain a grayed image; wherein, when the graying processing is performed, the grayscale value calculation expression is Gray =0.3 × R +0.59 × G +0.11 × B (1);

in the formula, gray represents the Gray value of a certain pixel point of the grayed sound field distribution diagram; r, G and B respectively represent red, green and blue intensity values of corresponding pixel points in the source image;

(2) Calculating the transverse and longitudinal gradients and directions of each pixel point in the grayed image to obtain contour information;

wherein, the calculation expressions of the gradient at the pixel point (x, y) are formulas (2) and (3),

G _x (x,y)＝H(x+1,y)-H(x-1,y) (2)；

G _y (x,y)＝H(x,y+1)-H(x,y-1) (3)；

in the formula, G _x (x,y)、G _y (x, y) and H (x, y) are respectively the gray values of the image after horizontal gradient, longitudinal gradient and gray level in sequence;

wherein, the calculation expressions of the gradient amplitude and the direction at the pixel point (x, y) are formulas (4) and (5),

(3) Dividing the image into a plurality of cells based on the contour information, and acquiring a feature vector of each cell;

(4) Forming a block by a preset number of unit cells; the method comprises the following steps that feature vectors of all cells in each block are connected in series to obtain HOG feature vectors of the block;

(5) Traversing the block through the whole image by the step length of one cell, and finally connecting HOG characteristic vectors of all blocks in the image in series to obtain the HOG characteristic vector of the image.

9. The system of claim 8, wherein the Pearson correlation coefficient obtaining module implements the process of obtaining the Pearson correlation coefficient between the HOG eigenvector of the sound field distribution diagram of the transformer to be diagnosed and the HOG eigenvector of the reference sound field distribution diagram,

the computational expression of Pearson correlation coefficient between HOG feature vectors is formula (6):

wherein n is the number of HOG feature vector data, X _i ,Y _i Sequentially and respectively obtaining the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the ith sample of the HOG characteristic vector of the reference sound field distribution diagram,

sequentially and respectively obtaining the average value, sigma, of samples of the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram _X ,σ _Y Sequentially and respectively obtaining sample standard deviations of the HOG characteristic vector of the sound field distribution diagram of the transformer to be diagnosed and the HOG characteristic vector of the reference sound field distribution diagram, wherein the PCC value range is [ -1,1]。

10. The system for diagnosing the mechanical fault of the transformer based on the sound field distribution diagram according to claim 9, wherein the mechanical state obtaining module is implemented in a process of judging the mechanical state of the transformer based on the obtained Pearson correlation coefficient,

(1) When the PCC is not more than 0.95, the current sound field distribution diagram of the transformer to be diagnosed is strongly related to the reference sound field distribution diagram, and the mechanical state is good;

(2) When PCC is more than or equal to 0.85 and less than 0.95, the correlation between the current sound field distribution diagram and the reference sound field distribution diagram of the transformer to be diagnosed is general, and the mechanical state changes;

(3) When PCC is less than 0.85, the current sound field distribution diagram of the transformer to be diagnosed is weakly correlated with the reference sound field distribution diagram, and the mechanical state is seriously changed.

Images