Fault diagnosis method and system based on pin pressure signal feature extraction

Abstract

The invention discloses a fault diagnosis method and system based on pin pressure signal feature extraction, and relates to the field of power transmission line fault diagnosis, and the method comprises the steps: arranging a pressure sensor at a power transmission line pin connection part, and collecting a pressure signal of a pin; smooth denoising processing is carried out on the pressure signals, feature extraction is carried out on the processed pressure signals, the extracted features are analyzed, and pressure distribution of the pin is obtained; and constructing a fault diagnosis model based on historical data, and loading the pressure distribution of the pins into the model to perform fault diagnosis on the power transmission line. The signal is decomposed by using the ensemble empirical mode decomposition (EEMD) method, the mode aliasing problem is reduced, more accurate signal features are obtained, the fault diagnosis model is constructed by using the machine learning algorithm, the normal state and the fault state are identified, the fault diagnosis precision and reliability are remarkably improved, and the fault diagnosis accuracy and reliability are improved. And the method has relatively high application value and popularization prospect.

Description

Technical Field

[0001] The present invention relates to the field of power transmission line fault diagnosis, and in particular to a fault diagnosis method and system based on pin pressure signal feature extraction. Background Art

[0002] During the operation of the transmission line, the connection components such as pins are subjected to complex changes in mechanical loads and environmental stresses. These changes will be reflected in the pressure signal of the pins. Effective monitoring and analysis of the pressure signal can help to detect potential equipment failures in advance and prevent accidents. Especially for high-voltage transmission lines, the reliability of its connection components is particularly important, and any minor failure may lead to serious consequences.

[0003] At present, the methods for fault diagnosis of the connection components of transmission lines mainly rely on regular manual inspections and simple electrical parameter monitoring. However, these methods have problems such as long detection cycles, low accuracy, and poor real-time performance, and cannot meet the needs of modern smart grids for real-time monitoring of equipment health status. Therefore, there is an urgent need for an efficient fault diagnosis method that can detect faults in a timely manner and perform maintenance by extracting and analyzing the pressure signal characteristics of the connection components.

[0004] Traditional signal processing methods such as Fourier transform and short-time Fourier transform can analyze the frequency characteristics of pressure signals to a certain extent, but the processing effect of non-stationary signals and mutation signals is poor. In recent years, methods based on wavelet transform and empirical mode decomposition (EMD) have been gradually applied to the field of signal processing, which have good time-frequency localization characteristics and adaptability. However, the EMD method has the problem of modal aliasing, which affects the accuracy of feature extraction.

[0005] In view of the above problems, the present invention proposes a fault diagnosis method based on pin pressure signal feature extraction. Through smoothing, wavelet threshold denoising and ensemble empirical mode decomposition (EEMD), the intrinsic mode function (IMF) reflecting the change of pin pressure is extracted, and these features are combined to establish a fault diagnosis model to perform fault diagnosis, thereby improving the accuracy and reliability of diagnosis. Summary of the invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] Therefore, the problem to be solved by the present invention is that traditional manual fault diagnosis has the problems of long detection cycle, low accuracy and poor real-time performance, and in signal processing, the existing EMD method has the problem of modal aliasing, which affects the accuracy of feature extraction.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: a fault diagnosis method based on pin pressure signal feature extraction, which includes setting up a pressure sensor at the pin connection of the transmission line to collect the pressure signal of the pin; smoothing and denoising the pressure signal, extracting features from the processed pressure signal, analyzing the extracted features to obtain the pressure distribution of the pin; constructing a fault diagnosis model based on historical data, loading the pressure distribution of the pin into the model to perform fault diagnosis on the transmission line.

[0009] As a preferred solution of the fault diagnosis method based on pin pressure signal feature extraction described in the present invention, the smoothing and denoising processing includes reducing the high-frequency noise in the pressure signal based on the Savitzky-Golay denoising algorithm, decomposing the pressure signal into approximate coefficients and detail coefficients through wavelet transform, suppressing or eliminating the noise part in the detail coefficient through threshold processing, and then obtaining the denoised pressure signal by inverse wavelet transform; the approximate coefficient reflects the large-scale characteristics of the pressure signal, and the detail coefficient contains the local detail information and noise of the pressure signal.

[0010] As a preferred solution of the fault diagnosis method based on pin pressure signal feature extraction described in the present invention, the feature extraction includes adding white noise of different amplitudes to the denoised pressure signal to generate multiple groups of noisy signals, performing EMD decomposition on each noisy signal to obtain a group of IMFs and a residual; the IMF is an intrinsic mode function that contains the frequency information, time-varying characteristics and amplitude information of the noise signal; the residual is the remaining part of the noisy signal after removing all IMFs; the corresponding IMFs obtained by decomposing all noisy signals are averaged point by point to obtain the final IMF.

[0011] As a preferred solution of the fault diagnosis method based on pin pressure signal feature extraction of the present invention, wherein: the pressure distribution of the pin is obtained, based on IMF, the tension curve, shear force curve and bending moment curve of the pin are obtained, and the force state of the pin is inferred through the tension formula, shear force formula and bending moment formula; the tension F of the pin is t The tensile force formula is obtained by fitting the deadweight of the transmission line and the electromagnetic force on the pins:

[0012]

[0013] Where L is the length of the transmission line, ρ w(x) represents the mass density per unit length of the transmission line, g represents the acceleration of gravity, μ0 represents the magnetic permeability of vacuum, I(x,t) represents the current at time t at position x, and r(x) represents the distance between the pin and the transmission line; the shear force on the pin is obtained by fitting the wind force on the pin and the electromagnetic force on the pin. The shear force formula is expressed as follows:

[0014]

[0015] Among them, ρ a It is represented as air density, V(x) is represented as wind speed at position x, A(x) is represented as wind-exposed area of ​​the pin at position x, and B(x,t) is represented as magnetic induction intensity at time t at position x. The bending moment on the pin is obtained by fitting the wind force on the pin and the deadweight of the transmission line. The bending moment formula is expressed as,

[0016]

[0017] As a preferred solution of the fault diagnosis method based on pin pressure signal feature extraction described in the present invention, the reverse deduction of the stress state of the pin includes calculating the first-order difference of the tension curve, the shear force curve and the bending moment curve to obtain the force change rate of the pin over time, setting a minimum change rate threshold to eliminate the influence of errors, and when the tension change and shear force change of the pin show an increasing trend, and the bending moment change rate is less than the minimum change rate threshold, it is determined that the electromagnetic force on the pin increases; when the tension change and bending moment change of the pin show an increasing trend, and the shear force change rate is less than the minimum change rate threshold, it is determined that the deadweight of the transmission line increases. ; When the changes in tension and shear force on the pin show a decreasing trend, and the bending moment is less than the minimum change rate threshold, it is determined that the electromagnetic force on the pin is reduced; when the changes in tension and bending moment on the pin show a decreasing trend, and the shear force change rate is less than the minimum change rate threshold, it is determined that the deadweight of the transmission line is reduced; when the changes in bending moment and shear force on the pin show an increasing trend, and the bending tension change rate is less than the minimum change rate threshold, it is determined that the wind force on the pin is increased; when the changes in bending moment and shear force on the pin show a decreasing trend, and the bending moment is less than the minimum change rate threshold, it is determined that the wind force on the pin is reduced.

[0018] As a preferred solution of the fault diagnosis method based on pin pressure signal feature extraction described in the present invention, wherein: the fault diagnosis of the transmission line includes initial diagnosis of transmission line faults and pin fault diagnosis; the initial diagnosis of the transmission line fault includes setting a maximum change rate threshold to reflect the abnormal state, when the force state is judged as an increase in the electromagnetic force on the pin, and the rate of change of the increase in the electromagnetic force is greater than the maximum change rate threshold, it indicates that the current and voltage on the transmission line are increased, and it is diagnosed as a short circuit fault in the transmission line; when the force state is judged as a decrease in the electromagnetic force on the pin, and the rate of change of the decrease in the electromagnetic force is greater than the maximum change rate threshold, it indicates that the current and voltage on the transmission line are reduced, and it is diagnosed as a disconnection fault or a grounding fault in the transmission line; when the force state is judged as an increase in the self-weight of the transmission line, and the rate of change of the increase in the self-weight is greater than the maximum change rate threshold, it indicates that there is a foreign body load on the transmission line, and additional personnel need to be dispatched to clean it up.

[0019] As a preferred solution of the fault diagnosis method based on pin pressure signal feature extraction of the present invention, the pin fault diagnosis includes dividing the pin historical fault data into a training set and a test set and assigning labels, wherein the pin historical fault data includes the force magnitude and change rate when the pin fails; standardizing the features of the training set and the test set to eliminate the influence between different dimensions; selecting the SVM model of the linear kernel function for training, and finding the best classification hyperplane that separates the data, which is expressed as follows:

[0020]

[0021] Among them, w represents the weight vector, b represents the bias, and C represents the penalty parameter. N Represented as the total number of training sets, y i Represents the label of the training sample, x i It is represented as the standardized training feature vector, and the superscript T represents the matrix transpose; the trained model is used to predict the test set to obtain the predicted label, and the confusion matrix between the predicted label and the true label is calculated. The accuracy, precision, recall and F1 score are calculated based on the confusion matrix to evaluate the model; the performance of the model under different parameter combinations is evaluated through cross-validation to avoid overfitting. The formula is expressed as,

[0022]

[0023] in, represents the penalty parameter optimized by cross-validation, K represents the number of folds in cross-validation, and D k Denotes the K-th fold validation set, Expressed as a loss function; grid search is used to optimize the hyperparameters of SVM to find the best parameter combination. The formula is expressed as,

[0024]

[0025] in, Represented as the kernel function parameter obtained by grid search optimization, α i is represented by the Lagrange multiplier of the support vector machine, γ is represented by the kernel function parameter, ‖x i -x‖ represents the Euclidean distance between samples; the tensile force curve, shear force curve and bending moment curve data are loaded into the trained model to obtain the prediction results when When the value is 1, it means that the pin is not damaged. A value of -1 indicates that the pin is broken.

[0026] Another object of the present invention is to provide a fault diagnosis system based on pin pressure signal feature extraction, which can extract the pressure signal features of the pin in real time and perform fault diagnosis on the transmission line and the pin.

[0027] In order to solve the above technical problems, the present invention provides the following technical solutions: a system of a fault diagnosis method based on pin pressure signal feature extraction, comprising: a data acquisition module, a pressure distribution analysis module and a fault diagnosis module; the data acquisition module acquires the pressure signal of the pin; the pressure distribution analysis module performs smoothing and denoising on the pressure signal, performs feature extraction on the processed pressure signal, analyzes the extracted features, and obtains the pressure distribution of the pin; the fault diagnosis module constructs a fault diagnosis model based on historical data, loads the pressure distribution of the pin into the model to perform fault diagnosis on the transmission line.

[0028] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the fault diagnosis method based on pin pressure signal feature extraction when executing the computer program.

[0029] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the fault diagnosis method based on pin pressure signal feature extraction as described above.

[0030] The beneficial effects of the present invention are as follows: the present invention performs smoothing processing on the collected pin pressure signal, reduces the high-frequency noise in the signal, and improves the stability of the signal. Secondly, the wavelet threshold denoising technology is applied to perform noise reduction processing on the signal through wavelet decomposition and reconstruction, suppressing the noise while retaining the important features of the signal.

[0031] The first-order difference of the processed pressure signal is calculated to capture the dynamic change characteristics of the signal. The signal is decomposed into several intrinsic mode functions (IMFs) using the ensemble empirical mode decomposition (EEMD) method. Through multiple decomposition and averaging processing, the modal aliasing problem is reduced and more accurate signal characteristics are obtained.

[0032] In the feature extraction stage, for each IMF component, the statistical features such as the mean, standard deviation, and peak-to-valley difference of the first-order difference are calculated. These features can effectively reflect the changing rules and patterns of the pin pressure signal. Finally, fault diagnosis is performed by analyzing the extracted features, and a fault diagnosis model is constructed using a machine learning algorithm to identify normal and fault states. The method of the present invention can accurately extract features from the pin pressure signal under complex working conditions, significantly improving the accuracy and reliability of fault diagnosis, and has high application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0034] Figure 1 This is a flow chart of the fault diagnosis method based on pin pressure signal feature extraction in Example 1.

[0035] Figure 2 Schematic diagram of pressure signals before and after smoothing processing of the fault diagnosis method based on pin pressure signal feature extraction in Example 1.

[0036] Figure 3 This is a schematic diagram of the characteristic curve after EEMD extraction of the fault diagnosis method based on pin pressure signal feature extraction in Example 1.

[0037] Figure 4 This is a module structure diagram of the fault diagnosis system based on pin pressure signal feature extraction in Example 3. DETAILED DESCRIPTION

[0038] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0039] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0040] Example 1, reference Figure 1-Figure 3 , which is the first embodiment of the present invention, and which provides a fault diagnosis method based on pin pressure signal feature extraction, including: Figure 1 As shown:

[0041] Step 1: Set up a pressure sensor at the pin connection of the transmission line to collect the pressure signal of the pin.

[0042] Step 2: Smooth and denoise the pressure signal, extract features from the processed pressure signal, analyze the extracted features, and obtain the pressure distribution of the pin.

[0043] Smoothing and denoising include reducing high-frequency noise in the pressure signal based on the Savitzky-Golay denoising algorithm, decomposing the pressure signal into approximate coefficients and detail coefficients through wavelet transform, suppressing or eliminating the noise part in the detail coefficients through threshold processing, and then using inverse wavelet transform to obtain the denoised pressure signal. The signal before and after smoothing is as follows Figure 2 As shown in the figure, Original signal is the original signal, and smoothed pressure signal is the signal after smoothing.

[0044] The approximate coefficient reflects the large-scale characteristics of the pressure signal, while the detail coefficient contains the local detail information and noise of the pressure signal.

[0045] The feature extraction using the EEMD method includes adding white noise of different amplitudes to the denoised pressure signal to generate multiple groups of noisy signals, performing EMD decomposition on each noisy signal to obtain a group of IMFs and a residual.

[0046] The process of EMD decomposition includes the following steps: Local extreme point detection: Identify all local maxima and minima in the signal. Envelope construction: Through cubic spline interpolation, fit the local maxima and minima into upper and lower envelopes respectively. Calculate the average value: Calculate the average value of the upper and lower envelopes to obtain the local mean curve of the signal. Obtain the intrinsic mode function: Subtract the local mean curve from the signal to obtain a preliminary IMF. If the IMF condition is met (that is, the difference between the number of zero crossing points and the number of extreme points in the entire data interval does not exceed 1), an IMF is obtained. Otherwise, repeat the above steps until the IMF condition is met. The characteristic curve extracted by EEMD is as follows Figure 3 shown.

[0047] IMF is the intrinsic mode function that contains the frequency information, time-varying characteristics and amplitude information of the noise signal. The residual is the remaining part of the noisy signal after removing all IMFs.

[0048] The corresponding IMFs obtained by decomposing all noisy signals are averaged point by point to obtain the final IMF.

[0049] Based on IMF, the tension curve, shear force curve and bending moment curve of the pin are obtained, and the stress state of the pin is deduced through the tension formula, shear force formula and bending moment formula.

[0050] The tensile force F on the pin t The tensile force formula is obtained by fitting the deadweight of the transmission line and the electromagnetic force on the pins:

[0051]

[0052] Where L is the length of the transmission line, ρ w (x) represents the mass density per unit length of the transmission line, g represents the acceleration due to gravity, μ0 represents the magnetic permeability of vacuum, I(x,t) represents the current at position x at time t, and r(x) represents the distance between the pin and the transmission line.

[0053] The shear force on the pin is obtained by fitting the wind force on the pin and the electromagnetic force on the pin. The shear force formula is expressed as:

[0054]

[0055] Among them, ρ a It is represented as air density, V(x) is represented as wind speed at position x, A(x) is represented as wind-exposed area of ​​the pin at position x, and B(x,t) is represented as magnetic induction intensity at time t at position x.

[0056] The bending moment on the pin is obtained by fitting the wind force on the pin and the deadweight of the transmission line. The bending moment formula is expressed as:

[0057]

[0058] The first-order differences of the tension curve, shear force curve and bending moment curve are calculated to obtain the force change rate of the pin over time. The minimum change rate threshold is set to eliminate the influence of errors. When the tension change and shear force change on the pin show an increasing trend and the bending moment change rate is less than the minimum change rate threshold, it is determined that the electromagnetic force on the pin increases.

[0059] When the change in tension and bending moment on the pin shows an increasing trend, and the shear force change rate is less than the minimum change rate threshold, it is determined that the deadweight of the transmission line has increased.

[0060] When the change in the tensile force and the shear force on the pin shows a decreasing trend, and the bending moment is less than the minimum change rate threshold, it is determined that the electromagnetic force on the pin is reduced.

[0061] When the change in the tension and bending moment on the pin shows a decreasing trend, and the shear force change rate is less than the minimum change rate threshold, it is determined that the deadweight of the transmission line has decreased.

[0062] When the change in bending moment and shear force on the pin shows an increasing trend, and the bending tension change rate is less than the minimum change rate threshold, it is determined that the wind force on the pin is increasing.

[0063] When the changes in the bending moment and the shear force on the pin show a decreasing trend, and the bending moment is less than the minimum change rate threshold, it is determined that the wind force on the pin is reduced.

[0064] Step 3: Build a fault diagnosis model based on historical data and load the pressure distribution of the pins into the model to perform fault diagnosis on the transmission line.

[0065] Before building the fault diagnosis model, the transmission line can be initially diagnosed with faults, including setting a maximum change rate threshold to reflect abnormal conditions. When the stress state is judged as an increase in the electromagnetic force on the pin, and the rate of change of the increase in the electromagnetic force is greater than the maximum change rate threshold, it indicates that the current and voltage on the transmission line have increased, and the transmission line is diagnosed as a short circuit fault.

[0066] When the stress state is judged as a decrease in the electromagnetic force on the pin, and the rate of change of the electromagnetic force decrease is greater than the maximum rate of change threshold, it indicates that the current and voltage on the transmission line are reduced, and it is diagnosed as a line break or ground fault in the transmission line.

[0067] When the stress state is judged as an increase in the deadweight of the transmission line, and the rate of change of the deadweight increase is greater than the maximum rate of change threshold, it indicates that there is a foreign load on the transmission line, and additional personnel need to be dispatched to clean it up.

[0068] Building a fault diagnosis model based on historical data to diagnose pin faults includes dividing the pin historical fault data into a training set and a test set and assigning labels. The pin historical fault data includes the force magnitude and change rate when the pin fails; standardizing the features of the training set and the test set to eliminate the influence between different dimensions.

[0069] Select the SVM model with linear kernel function for training and find the best classification hyperplane that separates the data. The formula is:

[0070]

[0071] Among them, w represents the weight vector, b represents the bias, and C represents the penalty parameter. N Represented as the total number of training sets, y i Represents the label of the training sample, x i It is represented as the standardized training feature vector, and the superscript T represents the matrix transpose.

[0072] Use the trained model to predict the test set, get the predicted labels, calculate the confusion matrix between the predicted labels and the true labels, and evaluate the model by calculating the accuracy, precision, recall and F1 score based on the confusion matrix.

[0073] The performance of the model under different parameter combinations is evaluated through cross-validation to avoid overfitting. The formula is expressed as:

[0074]

[0075] in, represents the penalty parameter optimized by cross-validation, K represents the number of folds in cross-validation, and D k Denotes the K-th fold validation set, It is expressed as a loss function.

[0076] Use grid search to optimize the hyperparameters of SVM to find the best parameter combination. The formula is expressed as:

[0077]

[0078] in, Represented as the kernel function parameter obtained by grid search optimization, α i is represented by the Lagrange multiplier of the support vector machine, γ is represented by the kernel function parameter, ‖x i -x‖ is represented as the Euclidean distance between samples.

[0079] Load the tensile force curve, shear force curve and bending moment curve data into the trained model to obtain the prediction results when When the value is 1, it means that the pin is not damaged. A value of -1 indicates that the pin is broken.

[0080] Embodiment 2 is the second embodiment of the present invention, which is different from the first embodiment in that: the fault diagnosis method and system based on pin pressure signal feature extraction also includes: in order to verify and illustrate the technical effect adopted in this method, this embodiment adopts the traditional technical scheme and the method of the present invention for comparative testing, and compares the test results by means of scientific demonstration to verify the real effect of this method.

[0081] The experimental design of this embodiment includes two parts:

[0082] Comparison of fault diagnosis accuracy: Compare the accuracy, precision and reliability of our invented method with traditional manual methods in fault diagnosis.

[0083] Comparison of signal processing effects: Compare the effects of our invented method with the traditional empirical mode decomposition (EMD) method in signal processing, including noise suppression capability and feature extraction accuracy.

[0084] The experimental data obtained are shown in Table 1 and Table 2:

[0085] Table 1: Comparison of fault diagnosis accuracy

[0086]

[0087]

[0088] Table 2: Comparison of signal processing effects

[0089]

[0090] Comparison of fault diagnosis accuracy: Improved accuracy: Our invented method has greatly improved the accuracy of fault diagnosis through advanced signal processing and feature extraction technology. In all tested fault types, the accuracy of our invented method is higher than that of traditional manual methods, especially in the diagnosis of comprehensive faults, the accuracy has increased from 76.4% to 92.5%.

[0091] Shortened diagnostic time: Due to the application of automated signal processing and diagnostic models, our invented method significantly reduces the diagnostic time, with an average diagnostic time of only 24 seconds.

[0092] Comparison of signal processing effects: signal stability and noise suppression: Through smoothing and wavelet threshold denoising technology, our invented method is superior to the traditional EMD method in terms of signal stability and noise suppression. The signal stability index has increased by nearly 17 percentage points, and the noise suppression effect has increased by 8.9 percentage points.

[0093] IMF component decomposition accuracy and modal aliasing: Using the EEMD method combined with multiple decompositions and averaging, our invented method effectively reduced the modal aliasing problem, increasing the IMF component decomposition accuracy to 92.8% and reducing the modal aliasing rate to 7.3%.

[0094] Feature extraction accuracy: Our invented method also showed significant advantages in feature extraction accuracy, which increased by about 12 percentage points to 91.5%.

[0095] Example 3, reference Figure 4 , which is the third embodiment of the present invention, and which is different from the first two embodiments in that: a system for fault diagnosis method based on pin pressure signal feature extraction comprises a data acquisition module, a pressure distribution analysis module and a fault diagnosis module; the data acquisition module acquires the pressure signal of the pin;

[0096] The pressure distribution analysis module smoothes and denoises the pressure signal, extracts features from the processed pressure signal, and analyzes the extracted features to obtain the pressure distribution of the pin; the fault diagnosis module builds a fault diagnosis model based on historical data, loads the pressure distribution of the pin into the model, and performs fault diagnosis on the transmission line.

[0097] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0098] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0099] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0100] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A fault diagnosis method based on pin pressure signal feature extraction is characterized by: include, A pressure sensor is set up at the pin connection of the transmission line to collect the pressure signal of the pin; The pressure signal is smoothed and denoised, and features are extracted from the processed pressure signal. The extracted features are analyzed to obtain the pressure distribution of the pin; A fault diagnosis model is constructed based on historical data, and the pressure distribution of the pins is loaded into the model to perform fault diagnosis on the transmission line.

2. The fault diagnosis method based on pin pressure signal feature extraction according to claim 1 is characterized in that: The smoothing and denoising process includes reducing high-frequency noise in the pressure signal based on the Savitzky-Golay denoising algorithm, decomposing the pressure signal into approximate coefficients and detail coefficients by wavelet transform, suppressing or eliminating the noise part in the detail coefficients by threshold processing, and obtaining the denoised pressure signal by inverse wavelet transform; The approximate coefficient reflects the large-scale characteristics of the pressure signal, and the detail coefficient contains the local detail information and noise of the pressure signal.

3. The fault diagnosis method based on pin pressure signal feature extraction according to claim 2 is characterized in that: The feature extraction includes adding white noise of different amplitudes to the denoised pressure signal to generate multiple groups of noisy signals, and performing EMD decomposition on each noisy signal to obtain a group of IMFs and a residual; The IMF is an intrinsic mode function that contains the frequency information, time-varying characteristics and amplitude information of the noise signal; The residual is the remaining part of the noisy signal after removing all IMFs; The corresponding IMFs obtained by decomposing all noisy signals are averaged point by point to obtain the final IMF.

4. The fault diagnosis method based on pin pressure signal feature extraction according to claim 3 is characterized in that: Obtaining the pressure distribution of the pin includes obtaining the tension curve, shear force curve and bending moment curve of the pin based on the IMF, and inferring the stress state of the pin through the tension formula, shear force formula and bending moment formula; The tensile force F on the pin t The tensile force formula is obtained by fitting the deadweight of the transmission line and the electromagnetic force on the pins: Where L is the length of the transmission line, ρ w (x) represents the mass density per unit length of the transmission line, g represents the acceleration due to gravity, μ0 represents the magnetic permeability of vacuum, I(x,t) represents the current at time t at position x, and r(x) represents the distance between the pin and the transmission line; The shear force on the pin is obtained by fitting the wind force on the pin and the electromagnetic force on the pin. The shear force formula is expressed as: Among them, ρ a It is represented as air density, V(x) is represented as wind speed at position x, A(x) is represented as windward area of ​​pin at position x, and B(x,t) is represented as magnetic induction intensity at time t at position x; The bending moment on the pin is obtained by fitting the wind force on the pin and the deadweight of the transmission line. The bending moment formula is expressed as:

5. The fault diagnosis method based on pin pressure signal feature extraction according to claim 4 is characterized in that: The inverse deduction of the stress state of the pin includes calculating the first-order difference of the tension curve, the shear force curve and the bending moment curve to obtain the force change rate of the pin over time, setting a minimum change rate threshold to eliminate the influence of errors, and when the tension change and the shear force change of the pin show an increasing trend, and the bending moment change rate is less than the minimum change rate threshold, it is determined that the electromagnetic force on the pin increases; When the change in the tension and bending moment of the pins shows an increasing trend, and the shear force change rate is less than the minimum change rate threshold, it is determined that the deadweight of the transmission line has increased; When the change in the tensile force and the shear force on the pin shows a decreasing trend, and the bending moment is less than the minimum change rate threshold, it is determined that the electromagnetic force on the pin is reduced; When the change in the tension and bending moment of the pin shows a decreasing trend, and the shear force change rate is less than the minimum change rate threshold, it is determined that the deadweight of the transmission line is reduced; When the bending moment change and shear force change on the pin show an increasing trend, and the bending tension change rate is less than the minimum change rate threshold, it is determined that the wind force on the pin increases; When the changes in the bending moment and the shear force on the pin show a decreasing trend, and the bending moment is less than the minimum change rate threshold, it is determined that the wind force on the pin is reduced.

6. The fault diagnosis method based on pin pressure signal feature extraction according to claim 5, characterized in that: The fault diagnosis of the power transmission line includes the initial diagnosis of the power transmission line fault and the pin fault diagnosis; The initial diagnosis of the transmission line fault includes setting a maximum change rate threshold to reflect an abnormal state. When the force state is judged as an increase in the electromagnetic force on the pin, and the rate of increase in the electromagnetic force is greater than the maximum change rate threshold, it indicates that the current and voltage on the transmission line increase, and a short circuit fault is diagnosed as occurring in the transmission line. When the force state is judged as the electromagnetic force on the pin is reduced, and the rate of change of the electromagnetic force reduction is greater than the maximum rate of change threshold, it indicates that the current and voltage on the transmission line are reduced, and it is diagnosed as a line break fault or a ground fault in the transmission line; When the stress state is judged as an increase in the deadweight of the transmission line, and the rate of change of the deadweight increase is greater than the maximum rate of change threshold, it indicates that there is a foreign load on the transmission line, and additional personnel need to be dispatched to clean it up.

7. The fault diagnosis method based on pin pressure signal feature extraction according to claim 6 is characterized in that: The pin fault diagnosis includes dividing the pin historical fault data into a training set and a test set and assigning labels, wherein the pin historical fault data includes the force magnitude and change rate when the pin fails; standardizing the features of the training set and the test set to eliminate the influence between different dimensions; Select the SVM model with linear kernel function for training and find the best classification hyperplane that separates the data. The formula is: Among them, w represents the weight vector, b represents the bias, and C represents the penalty parameter. N Represented as the total number of training sets, y i Represents the label of the training sample, x i It is represented as the normalized training feature vector, and the superscript T represents the matrix transpose; Use the trained model to predict the test set, get the predicted labels, calculate the confusion matrix between the predicted labels and the true labels, and evaluate the model by calculating the accuracy, precision, recall and F1 score based on the confusion matrix; The performance of the model under different parameter combinations is evaluated through cross-validation to avoid overfitting. The formula is expressed as: in, represents the penalty parameter optimized by cross-validation, K represents the number of folds in cross-validation, and D k Denotes the K-th fold validation set, Expressed as a loss function; Use grid search to optimize the hyperparameters of SVM to find the best parameter combination. The formula is expressed as: in, Represented as the kernel function parameter obtained by grid search optimization, α i is represented by the Lagrange multiplier of the support vector machine, γ is represented by the kernel function parameter, ‖x i -x‖ represents the Euclidean distance between samples; Load the tensile force curve, shear force curve and bending moment curve data into the trained model to obtain the prediction results when When the value is 1, it means that the pin is not damaged. A value of -1 indicates that the pin is broken.

8. A system using the fault diagnosis method based on pin pressure signal feature extraction as claimed in any one of claims 1 to 7, characterized in that: It includes data acquisition module, pressure distribution analysis module and fault diagnosis module; The data acquisition module acquires the pressure signal of the pin; The pressure distribution analysis module performs smoothing and denoising processing on the pressure signal, extracts features from the processed pressure signal, and analyzes the extracted features to obtain the pressure distribution of the pin; The fault diagnosis module constructs a fault diagnosis model based on historical data, and loads the pressure distribution of the pins into the model to perform fault diagnosis on the power transmission line.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the fault diagnosis method based on pin pressure signal feature extraction according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fault diagnosis method based on pin pressure signal feature extraction according to any one of claims 1 to 7 are implemented.

Images