CN110618196A - LMD sample entropy and SVM pillar insulator fault identification method - Google Patents

Abstract

The application discloses the application provides a fault identification method based on LMD sample entropy and SVM post insulator, which comprises the following steps: collecting post insulator sound vibration signals, and collecting the post insulator sound vibration signals as samples by using post insulator vibration acoustic detection equipment; and decomposing the post insulator sound vibration signal through LMD to obtain a plurality of PF components, and then solving the sample entropy of the PF energy components. Carrying out statistical classification on the sample entropies to serve as a sample library; creating and training an SVM using the sample library data; and carrying out fault recognition on the post insulator signal to be detected by using the SVM. The LMD sample entropy and SVM pillar insulator fault identification method is suitable for analyzing and processing pillar insulators and complex non-stationary time-varying fault signals on the basis of acoustic vibration signals.

Description

LMD sample entropy and SVM pillar insulator fault identification method

Technical Field

The application relates to the technical field of fault signal analysis and processing, in particular to a fault identification method based on LMD sample entropy and SVM post insulator.

Background

The post insulator has more accidents of breakage caused by the problems of severe working environment, quality degradation, natural aging and the like, and the safe operation of the electric power system is seriously influenced, so that the research on the in-service detection of the insulator has important application value. The vibration signal of the post insulator belongs to pulse excitation vibration, and the frequency spectrum of the vibration signal is complex and has more interference.

Because the structural rigidity of the component is changed due to damage, when a pulse signal containing abundant frequency components is used as an input to excite the knob insulator, certain vibration response is generated on the structure, damage occurs, and the different change amounts of the damage degrees are different.

The existing insulator longitudinal vibration mode finite element analysis considers that the surface defect of the insulator can cause the longitudinal resonance natural frequency to change, but the natural resonance frequency of each insulator is different due to the production condition of the insulator, the frequency reduction amplitude caused by the defect is not enough to be used as the judgment basis for the defect, and the analysis of the insulator fracture fault is not accurate.

If the defects generated by the insulator can cause the strength or rigidity of the whole vibration structure (comprising the insulator string, the base and the related parts) to change, the fault recognition based on the acoustic vibration signal is more accurate and effective, and the fault recognition method based on the LMD sample entropy and the SVM post insulator is suitable for the analysis and processing of the post insulator and the complex non-stationary time-varying fault signal based on the acoustic vibration signal.

Disclosure of Invention

The application provides a fault identification method based on LMD sample entropy and SVM pillar insulators, and aims to solve the technical problem that insulator fracture fault analysis is inaccurate in the prior art.

In order to solve the technical problem, the embodiment of the application discloses the following technical scheme:

the embodiment of the application discloses a fault identification method based on LMD sample entropy and SVM post insulator, which comprises the following steps: collecting post insulator sound vibration signals, and collecting the post insulator sound vibration signals as samples by using post insulator vibration acoustic detection equipment;

decomposing the post insulator sound vibration signal through LMD to obtain a plurality of PF components, and then solving sample entropy of the PF energy components;

carrying out statistical classification on the sample entropies to construct a sample library;

creating and training an SVM using the sample library data;

and (4) carrying out the signal of the post insulator to be detected by utilizing the SVM.

Optionally, the creating and training the SVM using the sample database data includes:

carrying out normalization processing on the sample database data;

finding the optimal data property parameter and penalty factor parameter,

with the cross-validation method, the SVM is tuned and created using the radial kernel function RBF.

Optionally, the post insulator acoustic vibration signal includes a normal post insulator acoustic vibration signal and a fault post insulator acoustic vibration signal.

Optionally, the PF components are 7 PF components.

Compared with the prior art, the beneficial effect of this application is:

the application provides a fault identification method based on LMD sample entropy and SVM pillar insulator, which comprises the following steps: collecting post insulator sound vibration signals, and collecting the post insulator sound vibration signals as samples by using post insulator vibration acoustic detection equipment; and decomposing the post insulator sound vibration signal through LMD to obtain a plurality of PF components, and then solving the sample entropy of the PF energy components. The LMD method adopts a smoothing method to form a local mean function and a local envelope function, so that the phenomena of over-envelope and under-envelope generated when an upper envelope and a lower envelope are formed by adopting a cubic spline function in the EMD method can be avoided. When special signals exist, the LMD result is not affected by the end point effect, for example, the amplitude modulation and frequency modulation signals of which the end points are extreme values, so that the judgment result is more accurate. Carrying out statistical classification on the sample entropies to serve as a sample library; creating and training an SVM using the sample library data; and carrying out fault recognition on the post insulator signal to be detected by using the SVM. The SVM is to establish a hyperplane as a decision surface, so that the isolation margin between the positive and negative examples is maximized. The LMD sample entropy and SVM pillar insulator fault identification method is suitable for analyzing and processing pillar insulators and complex non-stationary time-varying fault signals on the basis of acoustic vibration signals.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.

Drawings

In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

FIG. 1 is a flow chart of a fault identification method based on LMD sample entropy and SVM strut insulators provided by the present application;

fig. 2 shows a 7-layer signal decomposed by a certain post insulator LMD according to an embodiment of the present invention.

Detailed Description

In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

As shown in fig. 1, the application discloses a fault identification method based on LMD sample entropy and SVM post insulators, comprising:

utilize post insulator vibration acoustics check out test set to gather post insulator sound vibration signal as the sample, wherein: the post insulator sound vibration signal comprises a normal post insulator sound vibration signal and a fault post insulator sound vibration signal.

In this embodiment, 30 post insulator acoustic vibration signals are collected as samples, and the first 23 definition tags 1 are used as normal samples. The last 7 definition tags 2 serve as insulator fault samples.

And decomposing the post insulator sound vibration signal through LMD to obtain a plurality of PF components, and then solving the sample entropy of the PF energy components. The LMD method separates pure envelope signals and frequency modulation signals under different scales from original signals, multiplies the two signals to obtain a PF component with an instantaneous frequency having a physical meaning, and circularly processes the PF component until all PF components are separated, so that the time-frequency distribution of the original signals can be obtained.

In the embodiment of the application, 30 collected signals are decomposed into 7 PF components through LMD. Fig. 2 shows 7 layers of signals decomposed by an LMD of a certain post insulator, 7 PF components after the LMD decomposition are obtained, the PF components actually represent frequency modulated and amplitude modulated signals, and then 30 × 7 PF components are used to find the sample entropy thereof.

The sample entropy is similar to the approximate entropy in physical meaning, the time sequence complexity is measured by measuring the probability of generating a new mode in a signal, and the larger the probability of generating the new mode is, the larger the sequence complexity is.

And carrying out statistical classification on the sample entropies to construct a sample library. In the embodiment of the application, statistical classification is used as a sample library for insulator fault feature recognition, 7 × 30 samples are obtained through classification and sorting, and the samples are stored in a mat file to be used as training set data of the SVM. As shown in table 1, 30 × 7 PF1 to PF7 sample library data tables are provided for the examples of the present application.

PF1	PF2	PF3	PF4	PF5	PF6	PF7
							0.3417	0.2058	0.2683	0.3613	0.2846	0.083	0.1933
0.3671	0.2706	0.2178	0.2432	0.2115	0.2001	0.2378
							0.2396	0.0665	0.2049	0.3016	0.3232	0.3405	0.3133
0.345	0.0561	0.2521	0.3624	0.2087	0.198	0.0997
							0.176	0.1811	0.1995	0.2638	0.3095	0.3421	0.3432
0.0555	0.2166	0.3571	0.3676	0.2761	0.1955	0.141
							0.3363	0.3473	0.2134	0.1117	0.2725	0.2611	0.2735
0.3209	0.2914	0.2681	0.267	0.2624	0.2618	0.2602
							0.317	0.2212	0.0407	0.1783	0.2813	0.3594	0.3234
0.0578	0.1659	0.1698	0.366	0.3324	0.1694	0.3492
							0.3678	0.2032	0.0877	0.0887	0.1569	0.3056	0.357
0.3375	0.3156	0.3354	0.1221	0.1772	0.2459	0.2773
							0.3607	0.3015	0.1799	0.115	0.0732	0.2513	0.2855
0.3039	0.1984	0.1495	0.1437	0.1864	0.3568	0.3613
							0.3499	0.323	0.2654	0.2258	0.2324	0.2386	0.2356
0.2252	0.1196	0.1787	0.254	0.3024	0.3399	0.3556
							0.2875	0.2389	0.2858	0.2263	0.2655	0.3126	0.3054
0.252	0.2308	0.2732	0.3032	0.3056	0.2891	0.2757
							0.2017	0.2008	0.2367	0.2653	0.2983	0.3276	0.338
0.3013	0.2659	0.2491	0.2947	0.2959	0.2952	0.2254
							0.28	0.3131	0.2666	0.3127	0.2354	0.2332	0.2802
0.2567	0.2823	0.247	0.3154	0.2849	0.262	0.2844
							0.1854	0.2789	0.3472	0.3426	0.3064	0.1824	0.1501
0.2796	0.2897	0.2819	0.2839	0.2938	0.2674	0.2434
							0.2726	0.2363	0.2714	0.2984	0.2853	0.2879	0.2854
0.2527	0.3463	0.228	0.2774	0.2382	0.3005	0.2504
							0.2171	0.3053	0.2495	0.3275	0.3265	0.2465	0.2142
0.2285	0.2564	0.3025	0.2461	0.2964	0.3025	0.293
							0.3559	0.1939	0.1123	0.1822	0.1889	0.3252	0.3533
0.196	0.204	0.2361	0.2644	0.3035	0.3348	0.3292

Creating and training an SVM using the sample library data. And taking the insulator normal and fault data of the mat file as training set data, and carrying out normalization processing on the sample database data. Finding the optimal c (penalty factor)/g ((RBF kernel parameter). adopting a cross validation method to obtain the optimal c (penalty factor) and g, and using a radial kernel function RBF to adjust and create the SVM.

And diagnosing the post insulator signal to be detected by using the trained SVM, and identifying the fault of the post insulator sound vibration signal to be detected.

Since the above embodiments are all described by referring to and combining with other embodiments, the same portions are provided between different embodiments, and the same and similar portions between the various embodiments in this specification may be referred to each other. And will not be described in detail herein.

It is noted that, in this specification, relational terms such as "first" and "second," and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such circuit structure, article, or apparatus. Without further limitation, the presence of an element identified by the phrase "comprising an … …" does not exclude the presence of other like elements in a circuit structure, article or device comprising the element.

Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.

The above-described embodiments of the present application do not limit the scope of the present application.

Claims (4)

1. A fault identification method based on LMD sample entropy and SVM pillar insulators is characterized by comprising the following steps:

collecting post insulator sound vibration signals, and collecting the post insulator sound vibration signals as samples by using post insulator vibration acoustic detection equipment;

decomposing the post insulator sound vibration signal through LMD to obtain a plurality of PF components, and then solving sample entropy of the PF energy components;

carrying out statistical classification on the sample entropies to construct a sample library;

creating and training an SVM using the sample library data;

and (4) carrying out the signal of the post insulator to be detected by utilizing the SVM.

2. The method of identifying a post insulator fault of claim 1, wherein the creating and training an SVM using the sample library data comprises:

carrying out normalization processing on the sample database data;

finding the optimal data property parameter and penalty factor parameter,

with the cross-validation method, the SVM is tuned and created using the radial kernel function RBF.

3. The post insulator fault identification method of claim 1, wherein the post insulator vibro-acoustic signals include normal post insulator vibro-acoustic signals and faulty post insulator vibro-acoustic signals.

4. The post insulator fault identification method according to claim 1, wherein the plurality of PF components is 7 PF components.