Method and system for identifying internal defects of composite insulator

Abstract

The invention provides a composite insulator internal defect identification method and system, and the method comprises the steps: collecting an ultrasonic detection signal in a detection region selected on the outer surface of a composite insulator through an ultrasonic probe, and constructing a detection signal sample set; preprocessing the signals in the detection signal sample set by adopting a wavelet transform denoising algorithm to obtain preprocessed signals; carrying out empirical mode decomposition on the preprocessed signal, extracting characteristic parameters in the signal, and constructing a characteristic data set after carrying out dimension reduction processing on the characteristic parameters; and analyzing the feature data set by using a random forest algorithm to obtain an identification result of the internal defect of the composite insulator. Therefore, the problems that an existing manual defect identification method is high in subjectivity, more in misjudgment and the like are solved, a guiding function can be provided for factory detection, defect prevention and batch replacement of the composite insulator by means of a defect identification result, the defect identification precision is high, the anti-interference capacity is high, and automatic nondestructive detection on the interior of the composite insulator is achieved.

Description

Technical Field

[0001] The present invention relates to the field of power electronics technology, and in particular to a method and system for identifying internal defects of a composite insulator. Background Art

[0002] Composite insulators are increasingly widely used in power systems due to their small size, light weight, excellent hydrophobicity and hydrophobic migration, good internal insulation performance, high mechanical strength, small corona, and low electromagnetic interference. However, there are still a small number of composite insulators with defects that lead to accidents. In order to ensure the safe and stable operation of composite insulators, they need to be tested to verify the quality. The tests are divided into non-destructive testing and destructive testing. Destructive testing will damage composite insulators and can only be used as a means of detection and cannot be used on a large scale. Non-destructive testing refers to the location and judgment of defects in composite insulators through methods such as sound, light, electricity, and magnetism, without causing changes in the physical and chemical properties of the composite insulators being tested. There are many methods for non-destructive testing of composite insulators, such as infrared detection, ultraviolet detection, microwave reflection, ultrasonic detection, etc. Based on these defect detection methods, defect identification methods such as expert experience subjective identification, mathematical statistics methods, and artificial intelligence methods have been derived.

[0003] Ultrasonic testing technology is a convenient and efficient non-destructive testing technology. It can not only detect the internal structural information of the workpiece to be tested, but also has fast response time, safe use, long service life, good electrical insulation and anti-electromagnetic interference capabilities, and is particularly suitable for fault detection of power system equipment. Among them, the defect recognition method based on ultrasonic testing technology has made further progress in the intelligent recognition of internal defects of composite insulators by virtue of the easy availability of ultrasonic testing data and the richness of structural information.

[0004] However, although there have been studies on ultrasonic detection technology in the field of internal defect detection of composite insulators, they have only remained at the stage of successfully verifying the feasibility of ultrasonic detection of internal defects of composite insulators. Most studies only conduct subjective judgment on the ultrasonic detection signals of defective insulators, and there are cases of misjudgment and difficulty in distinguishing, while there is a lack of intelligent identification of defect signals and processing methods for ultrasonic detection signals of composite insulators. Therefore, ultrasonic detection of internal defects of composite insulators and intelligent identification of defect signals are one of the topics that need to be strengthened at present. Summary of the invention

[0005] In view of the defects in the prior art, the object of the present invention is to provide a method and system for identifying internal defects of a composite insulator.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying internal defects of a composite insulator, comprising:

[0007] Step 1: Collect ultrasonic detection signals in a detection area selected on the outer surface of the composite insulator through an ultrasonic probe, and construct a detection signal sample set;

[0008] Step 2: preprocessing the signal in the detection signal sample set by using a wavelet transform denoising algorithm to obtain a preprocessed signal;

[0009] Step 3: Performing empirical mode decomposition on the preprocessed signal to extract characteristic parameters in the signal, and constructing a characteristic data set after dimensionality reduction processing on the characteristic parameters;

[0010] Step 4: Analyze the characteristic data set using a random forest algorithm to obtain identification results of internal defects of the composite insulator.

[0011] Optionally, the ultrasonic probe used in step 1 is a small-diameter single-crystal straight probe capable of detecting the composite insulator material and independently transmitting and receiving ultrasonic waves;

[0012] The detection area is selected according to the specification parameters of the composite insulator and the location where internal defects are likely to occur.

[0013] Optionally, the step 2 includes:

[0014] Decomposing the signal in the detection signal sample set by wavelet decomposition to obtain a decomposed signal;

[0015] The decomposed signal is subjected to denoising, and the denoised signal is reconstructed and superimposed to obtain a preprocessed signal.

[0016] Optionally, the step 3 includes:

[0017] Step 3.1: Decompose the preprocessed signal into several intrinsic mode functions, perform envelope processing on each intrinsic mode function, and extract characteristic parameters in the signal;

[0018] Step 3.2: Use principal component analysis technology to reduce the dimension of feature parameters and construct a feature data set.

[0019] Optionally, the step 3.1 includes:

[0020] Step 3.1.1: The preprocessed signal is recorded as x(t), and all the maximum and minimum values ​​of x(t) are obtained;

[0021] Step 3.1.2: Use the cubic spline interpolation method to fit the maximum and minimum points of x(t) into smooth curves to form the upper envelope spectrum h of the decomposed signal. max (t) and the lower envelope spectrum h min (t), and calculate the average value of the envelope spectrum m(t) = (hmax (t)+h min (t)) / 2;

[0022] Step 3.1.3: Solve for the difference d between x(t) and the average of the upper and lower envelope spectra i (t) = x(t) - m(t), and determine d i (t) Whether it satisfies the criterion of the intrinsic mode function. If it does, extract the component corresponding to the intrinsic mode function and let I i (t) = d i (t), and let x(t) = x(t) - I i (t); if not, let x(t) = d i (t);

[0023] Step 3.1.4: Determine whether i is equal to n. If i is less than n, increment i by 1 and return to step 3.1.1. If i = n, execute step 3.1.5. Where n is the number of decomposition components of the eigenmode function and the initial value of i is 1.

[0024] Step 3.1.5: Characterize the preprocessed signal as Among them, I i (t) represents the component of the i-th intrinsic mode function, r(t) represents the remaining residual, and r(t) eventually tends to 0;

[0025] Step 3.1.6: Extract characteristic parameters from the signal.

[0026] Optionally, step 4 includes:

[0027] Dividing the feature data set into a training group data and a test group data;

[0028] Use the Bagging method to extract the training group data with replacement to obtain multiple training subsets;

[0029] Obtain a random subspace of the feature parameter group as a parameter set, and generate a corresponding feature subset based on each training subset;

[0030] A plurality of decision trees are constructed according to the feature subsets, and identification results of internal defects of the composite insulator are obtained through construction, identification and voting of the decision trees.

[0031] Optionally, the method further includes step 5: visually displaying the identification results of internal defects of the composite insulator.

[0032] In a second aspect, an embodiment of the present application provides a composite insulator internal defect identification system, which applies the composite insulator internal defect identification method as described in any one of the first aspects, and the system includes: a signal receiving module, a signal preprocessing module, a feature extraction module, a defect identification module and a comprehensive visualization module, wherein:

[0033] The signal receiving module is used to receive the ultrasonic detection signal collected by the ultrasonic detector and transmit the ultrasonic detection signal to the signal preprocessing module;

[0034] The signal preprocessing module uses a wavelet transform denoising algorithm to preprocess the signal in the detection signal sample set to obtain a preprocessed signal;

[0035] The feature extraction module is used to perform empirical mode decomposition on the preprocessed signal, extract characteristic parameters in the signal, and construct a feature data set after dimensionality reduction processing on the characteristic parameters;

[0036] The defect recognition module is used to analyze the characteristic data set to obtain the recognition result of the internal defects of the composite insulator;

[0037] The comprehensive visualization module is used to display the identification results of internal defects of the composite insulator.

[0038] In a third aspect, an embodiment of the present application provides a composite insulator internal defect identification device, including: a processor and a memory, wherein the memory stores executable program instructions, and when the processor calls the program instructions in the memory, the processor is used to:

[0039] Execute the steps of the composite insulator internal defect identification method as described in any one of the first aspects.

[0040] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium for storing a program, which, when executed, implements the steps of the composite insulator internal defect identification method as described in any one of the first aspects.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] In the present application, the nonlinear and non-stationary signals of the ultrasonic detection signals are decomposed, and the empirical mode decomposition method and the random forest algorithm are combined to identify the effectiveness of the ultrasonic detection signals of the internal defects of the composite insulator (pores, cracks, debonding, and no defects). The performance of the random forest algorithm is very good and it has a high anti-interference ability and strong anti-overfitting ability. It solves the problems of the existing manual defect identification methods such as strong subjectivity and many misjudgments. The defect identification results can provide guidance for the factory inspection, defect prevention and control, and batch replacement of composite insulators, and the identification accuracy is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings in the following descriptions are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without creative work. By reading the detailed description of the non-limiting embodiments with reference to the following drawings, other features, purposes and advantages of the present invention will become more obvious:

[0044] Figure 1 A schematic diagram of a flow chart of a composite insulator internal defect identification method provided in an embodiment of the present application;

[0045] Figure 2 A schematic diagram of the principle of the wavelet transform denoising algorithm provided in the embodiment of the present application;

[0046] Figure 3 It is a schematic diagram of a simulated waveform of an ultrasonic signal containing noise;

[0047] Figure 4 It is a schematic diagram of the simulated waveform of the ultrasonic signal after wavelet denoising;

[0048] Figure 5 A schematic diagram of the process flow of the empirical mode decomposition method provided in this embodiment;

[0049] Figure 6 A schematic diagram of the flow of the random forest algorithm provided in the embodiment of the present application;

[0050] Figure 7 A schematic diagram of the jth decision tree in the random forest provided in an embodiment of the present application;

[0051] FIG8( a ) is a schematic diagram of the original signal before filtering;

[0052] FIG8( b ) is a schematic diagram of a signal after filtering by a denoising algorithm based on wavelet transformation;

[0053] Figure 9(a) is a scatter plot of the identification of the empirical mode decomposition-random forest method;

[0054] Figure 9(b) is a schematic diagram of the confusion matrix of the empirical mode decomposition-random forest method;

[0055] Fig.10 This is a schematic diagram comparing the defect recognition effects of different methods;

[0056] Fig.11 Block diagram of the intelligent diagnosis and identification system for internal defects of composite insulators. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0058] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there can be a central component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there can be a central component at the same time.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0060] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein, for example, can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0061] The following specific embodiments are used to describe in detail the technical solution of the present invention and how the technical solution of the present application solves the above technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0062] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0063] Figure 1A schematic diagram of a flow chart of a composite insulator internal defect identification method provided in an embodiment of the present application, such as Figure 1 As shown, the method in this embodiment may include:

[0064] Step S101: collecting ultrasonic detection signals in a detection area selected on the outer surface of a composite insulator through an ultrasonic probe, and constructing a detection signal sample set.

[0065] In this embodiment, the ultrasonic probe used is a small-diameter single-crystal straight probe capable of detecting the composite insulator material and capable of transmitting and receiving ultrasonic waves independently, and the detection area is selected according to the specifications of the composite insulator and the location where internal defects are likely to occur. Specifically, a coupling agent is applied to the surface of the ultrasonic probe to closely contact the outer surface of the composite insulator, so that the ultrasonic wave can be transmitted to the inside of the insulator without loss.

[0066] Step S102: preprocessing the signals in the detection signal sample set using a wavelet transform denoising algorithm to obtain preprocessed signals.

[0067] In this embodiment, the signal in the detection signal sample set is decomposed by wavelet decomposition to obtain a decomposed signal; the decomposed signal is denoised, and the denoised signal is reconstructed and superimposed to obtain a preprocessed signal.

[0068] Figure 2 The schematic diagram of the principle of the wavelet transform denoising algorithm provided in the embodiment of the present application is as follows: Figure 2 As shown, the ultrasonic detection signal will be affected by the background noise generated by the center of the material and the electrical noise in the detection instrument, which will affect the recognition result. In order to eliminate these effects. In this embodiment, the wavelet change denoising algorithm is used to filter the ultrasonic detection signal and test the denoising effect. Exemplarily, the signal and noise are first decomposed by wavelet decomposition, and then denoising is performed according to the different characteristics of the decomposed signal, and finally the signal is reconstructed and superimposed.

[0069] Exemplarily, the ultrasonic detection signal is decomposed into a superposition of a series of basic wavelet functions after wavelet transformation. The calculation formula of wavelet transformation is shown in the following formula.

[0070]

[0071] In the formula, ψ(t) is the basic wavelet function, τ is the translation of the wavelet function, a is the scale expansion of the wavelet function, and the basic wavelet function satisfies the following formula.

[0072]

[0073] Figure 3 is a schematic diagram of the simulated waveform of an ultrasonic signal containing noise. Figure 4 Schematic diagram of the simulated waveform of the ultrasonic signal after wavelet denoising.

[0074] Optionally, wavelet packet decomposition and other methods can be used to extract features from ultrasonic detection signal data to obtain richer time-frequency features of ultrasonic detection signals. A convolutional neural network is designed using deep learning technology for feature extraction of ultrasonic signals. The convolution moving average method is used to filter the ultrasonic detection signal.

[0075] Step S103: performing empirical mode decomposition on the preprocessed signal, extracting characteristic parameters from the signal, and constructing a characteristic data set after dimensionality reduction processing on the characteristic parameters.

[0076] In this embodiment, the preprocessed signal is first decomposed into several intrinsic mode functions, and each intrinsic mode function is enveloped to extract characteristic parameters in the signal; then, the characteristic parameters are reduced in dimension using principal component analysis technology to construct a characteristic data set.

[0077] In this embodiment, since the ultrasonic detection signal is a non-stationary, nonlinear signal, it can be considered to be composed of a finite number of intrinsic mode functions (IMFs). The idea of ​​the empirical mode decomposition (EMD) method is to decompose any signal into several IMFs, and then perform envelope processing, i.e., Hilbert transform, on each IMF component. Each IMF component after decomposition satisfies two criteria:

[0078] 1) At any time, the mean of the upper and lower envelope spectra composed of the maximum and minimum values ​​of the IMF is 0;

[0079] 2) In the entire time series, the difference between the number of maximum and minimum points does not exceed 1.

[0080] Figure 5 A flow chart of the empirical mode decomposition method provided in this embodiment is as follows: Figure 5 As shown, the method in this embodiment may include the following steps:

[0081] Step S1031: record the preprocessed signal as x(t), and find all the maximum and minimum values ​​of x(t).

[0082] Step S1032: Use the cubic spline interpolation method to fit the maximum and minimum points of x(t) into smooth curves to form the upper envelope spectrum h of the decomposed signal. max (t) and the lower envelope spectrum h min (t), and calculate the average value of the envelope spectrum m(t) = (h max (t)+h min (t)) / 2.

[0083] Step S1033: Calculate the difference d between x(t) and the average value of the upper and lower envelope spectra i (t) = x(t) - m(t), and determine d i (t) Whether it satisfies the criterion of the intrinsic mode function. If it does, extract the component corresponding to the intrinsic mode function and let I i (t) = d i (t), and let x(t) = x(t) - I i (t); if not, let x(t) = d i (t).

[0084] Step S1034: determine whether i is equal to n. If i is less than n, increment i by 1 and return to step S1031. If i=n, ​​execute step S1035. Here, n is the number of decomposition components of the eigenmode function, and the initial value of i is 1.

[0085] Step S1035: Characterize the preprocessed signal as Among them, I i (t) represents the component of the i-th eigenmode function.

[0086] Step S1036: Extract characteristic parameters from the signal.

[0087] Step S104: Analyze the feature data set using the random forest algorithm to obtain the recognition result of the internal defects of the composite insulator.

[0088] Figure 6 A flow chart of the random forest algorithm provided in the embodiment of the present application is as follows: Figure 6 As shown in the figure, the feature data set is first divided into training group data and test group data; the training group data is extracted with replacement using the Bagging method to obtain multiple training subsets; the random subspace of the feature parameter group is obtained as the parameter set, and the corresponding feature subsets are generated based on each training subset; multiple decision trees are constructed according to the feature subsets, and the recognition results of internal defects of composite insulators are obtained through the construction, recognition and voting of the decision trees.

[0089] Specifically, the Bagging algorithm refers to: for a training set L{(Lbl k ,Dt k ),k=1,…,N train}——where Dt k and Lbl k are the kth training set data and its corresponding known category respectively. For example, we can give a training set sequence consisting of samples randomly selected with replacement from L {L j}, where each training set L j The number of samples is N train Due to different Lj are independent of each other, L j Each sample in is also independent of each other, so a certain sample (Lbl k ,Dt k ) may be in the training set L j For a given classification algorithm ξ(L), using any training set L j A classifier ξ(L j ), so we can get the combined classifier {ξ(L j )}.

[0090] For random forest, although the bagging combination classifier has defects, random forest can still use the bagging algorithm to resample the training data set without replacement to obtain multiple training subsets {L j}.

[0091] Furthermore, the Bagging algorithm is used to obtain the decision tree sample training subset {L j}, obtain the corresponding parameter subset {Z j}, and use this to build several decision trees.

[0092] In this embodiment, the CART binary regression classification tree algorithm is used, that is, each node can only fork to generate two new nodes at most. For ultrasonic detection signals, most feature vectors are not discrete, so when processing such data, they need to be discretized first. For the discretized feature variables, the Gini impurity is introduced to characterize the index of each feature vector being used as a node. Among them, the calculation method of the Gini index is as follows:

[0093]

[0094] In the formula, Gini(L j ) is the sample set L j The Gini index, P i Indicates category Y i In the sample set L j The probability of appearing in. If L j Split into subsets L j,1 and L j,2 , then the Gini index obtained by this division is:

[0095]

[0096] In the formula, and Respectively represent the subset L j,1 and L j,2 In L jWhen a node is split, the above formula is used to calculate the Gini index of each data set, and the splitting rule is selected according to the minimum Gini index principle to improve the calculation efficiency and speed.

[0097] Figure 7 A schematic diagram of the jth decision tree in the random forest provided in the embodiment of the present application, such as Figure 7 As shown in the figure, unlike traditional decision trees, random forests do not require pruning and depth thresholds when constructing several decision trees in the forest. Each decision tree is a complete binary decision tree.

[0098] Therefore, the leaf node generation conditions of the decision tree in the random forest are:

[0099] 1) Current signal data sample L j,k The data types are all of the same type Y i ;

[0100] 2) The remaining samples Z of the current parameter subset j is empty;

[0101] 3) Random Forest Generalization Error.

[0102] As for the performance of the classifier, the smaller the generalization error is, the better the classification performance is. The law of large numbers shows that the generalization error of random forest has a finite supremum.

[0103] For random forest {h(U X ,θ k ),k=1,2,3...}, its margin function is defined as follows:

[0104]

[0105] Where U x represents the input vector of the random forest, Y i Indicates U x The correct category, Y k,k≠i Indicates that Y is not included i Other classifications, i.e., misclassifications; avg j (g) represents the mean function, and I(g) is the indicative function.

[0106] It can be seen that the meaning of the marginal function is the difference between the probability of the correct classification in the decision tree that casts the correct vote and the maximum probability of the correct classification in the decision tree that casts the wrong vote. Obviously, the larger the value of the marginal function, the higher the classification confidence of the random forest and the better the classification effect.

[0107] The generalization error of random forest is defined as follows:

[0108]

[0109] According to the law of large numbers, as the number of decision trees in the random forest increases, for the sequence Θ = {θ k}, the generalization error converges almost everywhere, as shown below:

[0110]

[0111] Where n t is the number of decision trees. Therefore, we can know that the generalization error PE * It will tend to an upper bound, which means that random forest has a good ability to prevent overfitting.

[0112] On the other hand, the generalization error is related to the classification performance of each tree and its expression is as follows:

[0113]

[0114] In the formula, is the average value of the correlation between decision trees, is the average classification performance of each decision tree. From this, we can see that the better the classification performance of each tree in the random forest, the lower the correlation between trees, and the smaller the generalization error of the random forest.

[0115] Exemplarily, the following steps may also be included:

[0116] The identification results of internal defects of composite insulators are visualized.

[0117] Furthermore, in order to verify whether the present invention can effectively identify the internal defects of actual composite insulators, field tests were conducted on artificially defective composite insulators and retired composite insulators of a 66kV line. Before testing, the ultrasonic testing instrument was calibrated. A 2.5MHz-5MHz ultrasonic independent transceiver and a small-diameter single crystal straight probe were selected, and the gel-type coupling agent was fully applied to eliminate air interference. The ultrasonic detection signal of the internal defects of the composite insulator was identified using this system.

[0118] Figure 8(a) is a schematic diagram of the original signal before filtering, Figure 8(b) is a schematic diagram of the signal after filtering by the denoising algorithm based on wavelet transformation, Figure 9(a) is a recognition scatter plot of the empirical mode decomposition-random forest method, and Figure 9(b) is a schematic diagram of the confusion matrix of the empirical mode decomposition-random forest method. Fig.10 Schematic diagram for comparing the defect recognition effects of different methods.

[0119] like Fig.10As shown, SVM and BP neural network are also used to identify defects on the same feature data set, where the total recognition accuracy of SVM is 86.9% and the total recognition accuracy of BP neural network is 83.1%. The recognition method proposed in this embodiment has the highest accuracy, indicating that the embodiment of this application has certain practicality and application reference value, and has achieved good results in the actual detection and identification of internal defects of composite insulators.

[0120] A schematic diagram of the effect of identifying internal defects of composite insulators based on the empirical mode decomposition method and the random forest algorithm shows that the embodiment of the present application can effectively identify ultrasonic detection signals of types such as umbrella sleeve pores, cracks, interface debonding and no defects in the composite insulator, with a total recognition rate of 95.6%. Classes 1 to 4 in the figure are no defects, umbrella sleeve pores, cracks, and interface debonding, respectively.

[0121] In this embodiment, an ultrasonic probe suitable for detecting composite insulator materials is first selected, and a detection area is selected according to the specifications of the composite insulator and the location where internal defects are likely to occur. An ultrasonic detector is used to obtain a sample set of ultrasonic detection signals. Then, a wavelet transform denoising algorithm is used to filter the ultrasonic detection signals collected by the hardware system. After that, the pre-processed ultrasonic detection signal is subjected to EMD decomposition to extract characteristic parameters from the signal. The extracted characteristic data set is analyzed using a random forest algorithm. Finally, the model training and defect recognition results are visualized through the MATLAB GUI module, thereby realizing intelligent recognition of internal defects of composite insulators.

[0122] The present embodiment also provides a composite insulator internal defect recognition system, which applies the above-mentioned composite insulator internal defect recognition method. The system includes: a signal receiving module, a signal preprocessing module, a feature extraction module, a defect recognition module and a comprehensive visualization module, wherein: the signal receiving module is used to receive the ultrasonic detection signal collected by the ultrasonic detector, and transmit the ultrasonic detection signal to the signal preprocessing module; the signal preprocessing module uses a wavelet transform denoising algorithm to preprocess the signal in the detection signal sample set to obtain a preprocessed signal; the feature extraction module is used to perform empirical mode decomposition on the preprocessed signal, extract characteristic parameters in the signal, and construct a feature data set after dimensionality reduction processing of the characteristic parameters; the defect recognition module is used to analyze the characteristic data set to obtain the recognition result of the internal defect of the composite insulator; the comprehensive visualization module is used to display the recognition result of the internal defect of the composite insulator.

[0123] Fig.11 The block diagram of the intelligent diagnosis and identification system for internal defects of composite insulators is shown in Fig.11As shown in the figure, the ultrasonic probe is closely attached to the surface of the composite insulator to detect and generate ultrasonic signals. The ultrasonic signals are received by the data receiving module of the PC after passing through the ultrasonic detector and the digital oscilloscope in turn. The ultrasonic detection signal is subjected to wavelet denoising, EMD feature extraction, and random forest recognition in turn, and finally the identification results of the internal defects of the composite insulator are displayed through the comprehensive visualization module. This overcomes the subjective problems and low recognition efficiency in the manual identification of defect ultrasonic detection signals, improves the defect recognition performance for ultrasonic detection signals, and truly realizes intelligent defect diagnosis.

[0124] It should be noted that those skilled in the art will appreciate that various aspects of the present invention may be implemented as systems, methods or program products. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to as "circuit", "module" or "platform" herein.

[0125] In addition, an embodiment of the present application further provides a computer-readable storage medium, in which computer-executable instructions are stored. When at least one processor of a user device executes the computer-executable instructions, the user device executes the above-mentioned various possible methods. Among them, the computer-readable medium includes a computer storage medium and a communication medium, wherein the communication medium includes any medium that facilitates the transmission of a computer program from one place to another. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an ASIC. In addition, the ASIC can be located in a user device. Of course, the processor and the storage medium can also exist in a communication device as discrete components.

[0126] The present application also provides a program product, which includes a computer program. The computer program is stored in a readable storage medium. At least one processor of the server can read the computer program from the readable storage medium. At least one processor executes the computer program so that the server implements any method of the above-mentioned embodiments of the present invention.

[0127] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0128] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art may make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. A method for identifying internal defects of a composite insulator, characterized in that: include: Step 1: Collect ultrasonic detection signals in a detection area selected on the outer surface of the composite insulator through an ultrasonic probe, and construct a detection signal sample set; Step 2: preprocessing the signal in the detection signal sample set by using a wavelet transform denoising algorithm to obtain a preprocessed signal; Step 3: Performing empirical mode decomposition on the preprocessed signal to extract characteristic parameters in the signal, and constructing a characteristic data set after dimensionality reduction processing on the characteristic parameters; Step 4: Analyze the characteristic data set using a random forest algorithm to obtain identification results of internal defects of the composite insulator.

2. The method for identifying internal defects of composite insulators according to claim 1, characterized in that: The ultrasonic probe used in step 1 is a small-diameter single-crystal straight probe capable of detecting the ultrasonic wave transmission and reception of the composite insulator material; The detection area is selected according to the specification parameters of the composite insulator and the location where internal defects are likely to occur.

3. The method for identifying internal defects of composite insulators according to claim 1, characterized in that: The step 2 comprises: Decomposing the signal in the detection signal sample set by wavelet decomposition to obtain a decomposed signal; The decomposed signal is subjected to denoising, and the denoised signal is reconstructed and superimposed to obtain a preprocessed signal.

4. The method for identifying internal defects of composite insulators according to claim 1, characterized in that: The step 3 comprises: Step 3.1: Decompose the preprocessed signal into several intrinsic mode functions, perform envelope processing on each intrinsic mode function, and extract characteristic parameters in the signal; Step 3.2: Use principal component analysis technology to reduce the dimension of feature parameters and construct a feature data set.

5. The method for identifying internal defects of composite insulators according to claim 4, characterized in that: The step 3.1 comprises: Step 3.1.1: The preprocessed signal is recorded as x(t), and all the maximum and minimum values ​​of x(t) are obtained; Step 3.1.2: Use the cubic spline interpolation method to fit the maximum and minimum points of x(t) into smooth curves to form the upper envelope spectrum h of the decomposed signal. max (t) and the lower envelope spectrum h min (t), and calculate the average value of the envelope spectrum m(t) = (h max (t)+h min (t)) / 2; Step 3.1.3: Solve for the difference d between x(t) and the average of the upper and lower envelope spectra i (t) = x(t) - m(t), and determine d i (t) Whether it satisfies the criterion of the intrinsic mode function. If it does, extract the component corresponding to the intrinsic mode function and let I i (t) = d i (t), and let x(t) = x(t) - I i (t); if not, let x(t) = d i (t); Step 3.1.4: Determine whether i is equal to n. If i is less than n, increment i by 1 and return to step 3.1.

1. If i = n, execute step 3.1.

5. Where n is the number of decomposition components of the eigenmode function and the initial value of i is 1. Step 3.1.5: Characterize the preprocessed signal as Among them, I i (t) represents the component of the i-th intrinsic mode function, r(t) represents the remaining residual, and r(t) eventually tends to 0; Step 3.1.6: Extract characteristic parameters from the signal.

6. The method for identifying internal defects of composite insulators according to claim 1, characterized in that: The step 4 comprises: Dividing the feature data set into a training group data and a test group data; Use the Bagging method to extract the training group data with replacement to obtain multiple training subsets; Obtain a random subspace of the feature parameter group as a parameter set, and generate a corresponding feature subset based on each training subset; A plurality of decision trees are constructed according to the feature subsets, and identification results of internal defects of the composite insulator are obtained through construction, identification and voting of the decision trees.

7. The method for identifying internal defects of a composite insulator according to any one of claims 1 to 6, characterized in that: The method also includes step 5: visually displaying the identification results of the internal defects of the composite insulator.

8. A composite insulator internal defect identification system, characterized in that: The method for identifying internal defects of composite insulators according to any one of claims 1 to 7 is applied, wherein the system comprises: a signal receiving module, a signal preprocessing module, a feature extraction module, a defect identification module and a comprehensive visualization module, wherein: The signal receiving module is used to receive the ultrasonic detection signal collected by the ultrasonic detector and transmit the ultrasonic detection signal to the signal preprocessing module; The signal preprocessing module uses a wavelet transform denoising algorithm to preprocess the signal in the detection signal sample set to obtain a preprocessed signal; The feature extraction module is used to perform empirical mode decomposition on the preprocessed signal, extract characteristic parameters in the signal, and construct a feature data set after dimensionality reduction processing on the characteristic parameters; The defect recognition module is used to analyze the characteristic data set to obtain the recognition result of the internal defects of the composite insulator; The comprehensive visualization module is used to display the identification results of internal defects of the composite insulator.

9. A composite insulator internal defect identification device, characterized in that: include: A processor and a memory, wherein the memory stores executable program instructions, and when the processor calls the program instructions in the memory, the processor is used to: Execute the steps of the method for identifying internal defects of a composite insulator as described in any one of claims 1 to 7.

10. A computer-readable storage medium for storing a program, characterized in that: When the program is executed, the steps of the method for identifying internal defects of a composite insulator according to any one of claims 1 to 7 are implemented.

Images