Noise-based method for evaluating pollution level of operating insulators

Abstract

The application provides a noise-based operation insulator contamination grade evaluation method, comprising the following steps: S1. obtaining noise signals of sample insulators under different environmental temperatures, different humidities and different contamination grades; S2. performing filtering processing on the noise signals; S3. extracting time domain characteristic parameters from the filtered noise signals; S4. performing wavelet transform on the filtered noise signals, and extracting wavelet characteristic parameters in a set frequency band; S5. inputting the time domain characteristic parameters, the wavelet characteristic parameters, the environmental temperature and the humidity into a kernel extreme learning machine to perform training; and S6. obtaining insulator operation information in actual working conditions in real time, and inputting the operation information into the trained kernel extreme learning machine to obtain the contamination grade of the insulator, wherein the operation information comprises the environmental temperature, the humidity and the noise signal of the insulator.

Description

TECHNICAL FIELD

[0001] The present application relates to a power equipment evaluation algorithm, in particular to a noise-based operating insulator contamination level evaluation method. BACKGROUND

[0002] Insulators are widely used in power transmission lines, power plants and substations, and are an important part of the power system. Due to the increasing environmental and atmospheric pollution, insulators are directly exposed to the air during operation, and dust particles continuously deposit on the surface of the insulator, eventually forming a layer of contamination. When the contaminated insulator operates in weather conditions with high air humidity such as fog, condensation and drizzle, its insulation performance will be greatly reduced, and under the action of normal operating voltage, a pollution flashover accident is likely to occur.

[0003] In order to prevent the occurrence of pollution flashover accidents, it is necessary to decontaminate the insulator, and before decontamination, the contamination level of the insulator needs to be accurately evaluated to provide a reference for the development of decontamination measures. In the prior art, the evaluation of the contamination level of the insulator includes manual estimation and online monitoring. Manual estimation is achieved by personnel inspection and the like, which is affected by human subjective factors and has low accuracy, high cost and low efficiency. Online methods, such as image monitoring, are easily disturbed by the environment and have low accuracy. Other methods, such as leakage current detection and pulse current detection, are easily disturbed by electromagnetic interference and have low accuracy.

[0004] Therefore, in order to solve the above technical problems, a new technical means is needed. SUMMARY

[0005] Therefore, the present application provides a noise-based operating insulator contamination level evaluation method, which uses the noise signal generated by the discharge of the insulator during operation as the judgment basis of the insulator, and combines the environmental temperature and humidity factors to judge the contamination level of the insulator by kernel extreme learning machine, thereby effectively improving the recognition accuracy and avoiding the influence of electromagnetic interference and other factors of traditional methods, and having high efficiency and simple process.

[0006] The noise-based operating insulator contamination level evaluation method provided by the present application comprises the following steps:

[0007] S1. Obtain the noise signal of a sample insulator operating under different environmental temperatures, different humidities and different contamination levels;

[0008] S2. Filter the noise signal;

[0009] S3. Extract the time domain characteristic parameters from the filtered noise signal;

[0010] S4. performing wavelet transform on the filtered noise signal, and extracting wavelet characteristic parameters in the set frequency band;

[0011] S5. inputting the time domain characteristic parameters, the wavelet characteristic parameters, the ambient temperature and the humidity into the kernel extreme learning machine to perform training;

[0012] S6. obtaining running information of the insulator in actual working conditions in real time, and inputting the running information into the trained kernel extreme learning machine to obtain the contamination level of the insulator, wherein the running information comprises the ambient temperature, the humidity and the noise signal of the insulator.

[0013] Further, the time domain characteristic parameters comprise a peak value, an average value, a standard deviation, a root mean square, a kurtosis, a peak factor, a pulse factor and a margin factor of the noise;

[0014] The wavelet characteristic parameters comprise a frequency band energy ratio and an information entropy.

[0015] Further, the pre-processing of the noise signal in step S2 comprises:

[0016] performing Fourier transform on the noise signal, and then filtering the noise signal by using a band-stop filter;

[0017] performing global filtering on the filtered noise signal by using a Gaussian weighted moving average algorithm.

[0018] Further, step S4 specifically comprises:

[0019] when performing wavelet transform on the filtered noise signal, a wavelet base function is set as db6, and a decomposition layer number is set as 3;

[0020] the wavelet-transformed noise signal is divided into 2 J energy values of each sub-frequency band are determined respectively

[0021] a ratio of the energy of the set frequency band to the total energy is calculated

[0022] wherein: Em represents the energy value of the mth set frequency band;

[0023] a waveform sequence X of the noise signal is determined, and an information entropy H(X) is calculated:

[0024] H(X) = -∑q i log a q i ; q i is a probability of occurrence of the ith sub-frequency band in the waveform sequence, and a is a set calculation coefficient.

[0025] Further, in step S5, when training the kernel extreme learning machine, a particle swarm optimization algorithm is used to determine the kernel coefficient and the penalty factor of the kernel extreme learning machine.

[0026] The present application has the following advantages: by using the noise signal generated by the discharge of the insulator during operation as the judgment basis of the insulator, and combining the environmental temperature and humidity factors to judge the contamination level of the insulator by the kernel extreme learning machine, the recognition accuracy can be effectively improved, the influence of electromagnetic interference and other factors in the traditional method can be avoided, and the efficiency is high and the process is simple. BRIEF DESCRIPTION OF DRAWINGS

[0027] The present application will be further described below in conjunction with the drawings and examples:

[0028] Figure 1 The flowchart of the present application. DETAILED DESCRIPTION

[0029] The present application will be further described below in conjunction with the drawings and examples:

[0030] The present application provides a noise-based running insulator contamination level evaluation method, comprising the following steps:

[0031] S1. obtaining the noise signal of the sample insulator running under different environmental temperatures, different humidities and different contamination levels;

[0032] S2. filtering the noise signal;

[0033] S3. extracting the time domain characteristic parameters from the filtered noise signal;

[0034] S4. performing wavelet transform on the filtered noise signal, and extracting the wavelet characteristic parameters in the set frequency band;

[0035] S5. inputting the time domain characteristic parameters, the wavelet characteristic parameters, the environmental temperature and the humidity into the kernel extreme learning machine to perform training;

[0036] S6. obtaining the running information of the insulator in the actual working condition in real time, and inputting the running information into the trained kernel extreme learning machine to obtain the contamination level of the insulator, wherein the running information includes the environmental temperature, the humidity and the noise signal of the insulator; by the above method, the noise signal generated by the discharge of the insulator during operation is used as the judgment basis of the insulator, and the kernel extreme learning machine is used to judge the contamination level of the insulator in combination with the environmental temperature and humidity factors, so as to effectively improve the recognition accuracy, avoid the influence of electromagnetic interference and other factors in the traditional method, and the efficiency is high and the process is simple.

[0037] In the above, the sample insulator is set in the laboratory, and a type of insulator is selected as the sample insulator, such as a double-umbrella type porcelain insulator, and then different environmental temperatures, environmental humidities, and different pollution levels are simulated in the laboratory, wherein:

[0038] The environmental temperature is generally set to 5-25°C, and the temperature range can be divided into multiple temperature intervals, and each temperature interval is subjected to corresponding experiments. Generally, two intervals of 5-15°C and 15-25°C can meet the sample data requirements;

[0039] The humidity refers to the relative humidity, which is generally 60%-100%. Generally, three intervals can meet the sample data requirements, i.e., 65%-75%, 75%-85%, and 85%-95%;

[0040] The pollution level is divided into five levels, i.e., 0, I, II, III, and IV. 0.03 mg / cm2, 0.05 mg / cm2, 0.08 mg / cm2, 0.15 mg / cm2, and 0.25 mg / cm2 represent the equivalent salt density of 0, I, II, III, and IV pollution levels, respectively, and the ratio of ash density to salt density is 6:1. Sodium chloride, diatomite, and silicon dioxide are used to simulate the pollutants.

[0041] Under the above conditions, a direct current voltage is applied to the insulator to obtain corresponding noise under different conditions. The specific model and process of the kernel extreme learning machine are known in the art and will not be described here. The kernel coefficient and penalty factor of the kernel extreme learning machine are solved by using the existing particle swarm optimization algorithm.

[0042] In this embodiment, the time domain characteristic parameters include the peak value, average value, standard deviation, root mean square, kurtosis, peak factor, pulse factor, and margin factor of the noise. The calculation of the time domain characteristic parameters can be performed by using the existing algorithm, which will not be described here.

[0043] The wavelet characteristic parameters include the frequency band energy ratio and information entropy.

[0044] In this embodiment, the pre-processing of the noise signal in step S2 includes:

[0045] The noise signal is subjected to Fourier transform processing, and then a band-stop filter is used for filtering. The process of Fourier transform is known in the art.

[0046] The filtered noise signal is subjected to global filtering processing by using a Gaussian weighted moving average algorithm, wherein the Gaussian weighted moving average algorithm is known in the art, and its calculation formula is:

[0047]

[0048] wherein L is the number of sliding points, h is the height of the sliding point, and y is the weight average factor. n k is the filtering result.

[0049] In this embodiment, step S4 specifically includes:

[0050] When performing wavelet transform on the filtered noise signal, the wavelet base function is set as db6, and the decomposition layer number is set as 3.

[0051] The wavelet-transformed noise signal is divided into 2 J sub-frequency bands, and the energy value of each sub-frequency band is determined.

[0052] The ratio of the energy of the set frequency band to the total energy is calculated.

[0053] wherein: E m represents the energy value of the mth set frequency band.

[0054] The waveform sequence X of the noise signal is determined, the waveform sequence X is a set composed of each sub-frequency band, and the information entropy H(X) is calculated.

[0055] H(X) = -∑q i log a q i ; q i is the probability of the ith sub-frequency band appearing in the waveform sequence, and a is a set calculation coefficient.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application 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 application can be modified or replaced equivalently without departing from the purpose and scope of the technical solutions of the present application, and they should be covered in the scope of the claims of the present application.​

Claims

1. A method for assessing the pollution level of operating insulators based on noise, characterized in that: Includes the following steps: S1. Obtain noise signals of the sample insulators under different ambient temperatures, humidity levels, and pollution levels; S2. Filter the noise signal; S3. Extract time-domain feature parameters from the filtered noise signal; S4. Perform wavelet transform on the filtered noise signal and extract the wavelet feature parameters within the set frequency band; S5. Input the feature parameter matrix composed of time-domain feature parameters, wavelet feature parameters, ambient temperature, and humidity into the kernel extreme learning machine for training; S6. Real-time acquisition of insulator operation information under actual working conditions, and input of the operation information into the nuclear extreme learning machine after training to obtain the insulator pollution level. The operation information includes ambient temperature, humidity and insulator noise signal. The time-domain characteristic parameters include the peak value, average value, standard deviation, root mean square, kurtosis, peak factor, impulse factor, and margin factor of the noise. Wavelet feature parameters include band energy ratio and information entropy; Step S4 specifically includes: When performing wavelet transform on the filtered noise signal, the wavelet basis function is set to db6 and the number of decomposition levels is set to 3. The noise signal after wavelet transform is divided into Each sub-band is used to determine the energy value of each sub-band. ; Calculate the ratio of energy in the specified frequency band to the total energy. : ;in: This represents the energy value of the m-th set frequency band; Determine the waveform sequence X of the noise signal and calculate the information entropy H(X): ; denoted as , where is the probability of the i-th sub-band appearing in the waveform sequence, and 'a' is a set calculation coefficient.

2. The method for assessing the pollution level of operating insulators based on noise according to claim 1, characterized in that: Step S2, the preprocessing of the noise signal includes: The noise signal is processed by Fourier transform and then filtered using a band-stop filter; The filtered noise signal is then subjected to global filtering using a Gaussian weighted moving average algorithm.

3. The method for assessing the pollution level of operating insulators based on noise according to claim 1, characterized in that: In step S5, when training the kernel extreme learning machine, the particle swarm optimization algorithm is used to determine the kernel coefficients and penalty factor of the kernel extreme learning machine.

Images