CN108120907A - The partial discharge diagnostic method of feature extraction under a kind of tremendously low frequency voltage based on power frequency - Google Patents

Abstract

The invention proposes a partial discharge diagnosis method based on feature extraction under power frequency to low frequency voltage, which belongs to the partial discharge detection and pattern recognition technology in the field of high voltage and insulation technology. In this method, a partial discharge test platform is first constructed and an insulation defect model is made; each insulation defect model is subjected to a partial discharge test under power frequency to low frequency voltage on the partial discharge test platform, and each insulation defect model is extracted under each frequency voltage. Then conduct a partial discharge test on the power equipment to be tested and record the discharge characteristic parameters; compare the change trend of the discharge characteristic parameters of the power equipment to be tested with the change trend of the discharge characteristic parameters of each insulation defect model, and use the evaluation function to identify the discharge characteristic parameters. Detect insulation defects of electrical equipment. In this method, the same voltage amplitude is adopted at each frequency to avoid damage to insulation caused by higher voltage amplitudes; the variation trend of characteristic parameters is used for identification, and the influence of signal attenuation is weakened.

Description

A Partial Discharge Diagnosis Method Based on Feature Extraction under Power Frequency to Low Frequency Voltage

technical field

The invention belongs to the partial discharge detection and pattern recognition technology in the technical field of high voltage and insulation, in particular to a partial discharge diagnosis method based on feature extraction under power frequency to low frequency voltage.

Background technique

Partial discharge, as a partial area discharge that occurs at the insulation defect and does not form an overall penetrating damage, is accompanied by detectable phenomena such as electricity, sound and light. means. Since the characteristics of the partial discharge signal, such as discharge volume, discharge repetition rate, discharge phase distribution and n-q curve, etc., contain information on the type of insulation defect, the partial discharge test has become an important means of fault type diagnosis. In recent years, the research on partial discharge has mainly focused on different detection methods (such as electric pulse method, ultrasonic method, ultra-high frequency method, optical signal method, etc.), discharge characteristics of different insulation defects at different development stages, and fault diagnosis algorithms. .

However, the above-mentioned fault diagnosis techniques are all based on the discharge information under the power frequency voltage (50Hz), which makes the source of information relatively single. At the same time, due to the interference of on-site measurement, the propagation and attenuation of the discharge signal, it is more difficult to identify the discharge mode under power frequency voltage. For partial discharge pattern recognition, the existing method based on voltage amplitude is to apply voltages of different amplitudes, measure the partial discharge signal of the equipment under test, and analyze the characteristics of the discharge signal under different amplitude voltages for pattern recognition. The method may cause accidental damage to electrical equipment during pressurization.

Contents of the invention

The purpose of the present invention is to solve the problem that partial discharge acquisition information is relatively single under traditional power frequency voltage, and different voltage amplitudes may cause unexpected insulation damage, and proposes a partial discharge diagnosis method based on feature extraction under power frequency to low frequency voltage. This method can obtain more discharge information, adopt consistent voltage amplitudes at each frequency, and avoid damage to insulation caused by higher voltage amplitudes. The variation trend of characteristic parameters is used for identification, which weakens the influence of signal attenuation.

A method for diagnosing partial discharge based on feature extraction under power frequency to low frequency voltage proposed by the present invention, specifically includes the following steps:

1) Build a partial discharge test platform under power frequency to low frequency voltage; the partial discharge test platform includes: high voltage power amplifier, arbitrary waveform generator, protection resistor, insulation defect model, first measurement impedance, second measurement impedance, coupling capacitance , partial discharge measurement and analyzer, high voltage probe and oscilloscope;

The arbitrary waveform generator is connected to the high-voltage power amplifier through a coaxial cable connection, the high-voltage power amplifier is connected to the protection resistor through a high-voltage cable, the protection resistor is connected to the insulation defect model, and the coupling capacitor is respectively connected through a high-voltage cable, and the high-voltage probe is connected to the coupling The capacitor is connected through a high-voltage cable, the insulation defect model is connected to the first measurement impedance through a metal wire, the coupling capacitor is connected to the second measurement impedance through a metal wire, and the first measurement impedance and the second measurement impedance are connected to the partial discharge measurement and analyzer respectively. Coaxial cable connection, the high voltage probe is connected to the oscilloscope through the coaxial cable connection;

2) Make an insulation defect model;

The insulation defect model includes: internal oil gap and surface discharge model, insulating oil uniform field discharge model, corona discharge model in oil and surface discharge model, and defect models that may exist according to the specific conditions of the equipment;

3) For each insulation defect model produced in step 2), carry out partial discharge tests at 50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz frequency voltages on the partial discharge test platform constructed in step 1); In each test, adjust the frequency and amplitude of the output waveform of the arbitrary waveform generator, record and store the data of each partial discharge test at 1.05 times the discharge initiation voltage; each insulation defect model is tested at each frequency voltage The partial discharge test is repeated 10 to 15 times, and the average value of each test data at the frequency and voltage is taken as the partial discharge test data corresponding to the insulation defect model at the selected frequency voltage, and each insulation defect model is obtained at each Partial discharge test data under frequency voltage;

4) Extract discharge characteristic parameters from the partial discharge test data of each kind of insulation defect model obtained in step 3) to form a reference characteristic library; the discharge characteristic parameters include: initial voltage U _PDIV ; positive half cycle maximum Discharge capacity q ⁺ _max and negative half-cycle maximum discharge capacity q ^- _max , positive half-cycle average discharge capacity q ⁺ _mean and negative half-cycle average discharge capacity q ^- _mean , positive half-cycle pulse repetition rate rate ⁺ and negative half-cycle pulse repetition rate rate ^- ; max Discharge phase distribution map Average discharge phase distribution map Phase distribution spectrum of discharge times The characteristic parameters of , including: positive semicircular skewness S _k and negative semicircular skewness S _k of each spectrum, positive semicircular kurtosis K _u and negative semicircular kurtosis K _u of each spectrum, asymmetry of each spectrum degree A _sy and correlation degree C _c ; the characteristic parameters of the frequency distribution spectrum nQ of discharge capacity, including: skewness S _k , kurtosis K _u , asymmetry degree _Asy , correlation degree C _c ;

5) Place the power equipment to be tested at the location of the insulation defect model in the partial discharge test platform constructed in step 1), conduct a partial discharge test on the power equipment to be tested, and record the corresponding partial discharge tests of the power equipment to be tested under each frequency and voltage Data, the voltage frequency includes 50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz, and the voltage amplitude is 1.05 times the discharge initiation voltage;

6) Extract the discharge characteristic parameters in the partial discharge test data of the power equipment to be tested under each frequency and voltage;

7) Compare the change trend of the discharge characteristic parameters of the power equipment to be tested with the change trend of the corresponding discharge characteristic parameters of each insulation defect model, and identify the insulation defects of the power equipment to be tested through the evaluation function; the specific steps are as follows:

7-1) Select any insulation defect model and any discharge characteristic parameter, take the voltage frequency of the partial discharge test as the abscissa, and the discharge characteristic parameter value as the ordinate, use the insulation defect model at each frequency voltage Based on the discharge characteristic parameter value recorded in the partial discharge test data, the change trend diagram of the selected discharge characteristic parameter of the insulation defect model from power frequency to low frequency voltage is obtained;

7-2) Repeat step 7-1) to obtain the change trend diagram of each discharge characteristic parameter of each insulation defect model from power frequency to low frequency voltage;

7-3) Repeat step 7-1) for the power equipment to be tested to obtain a trend diagram of each discharge characteristic parameter corresponding to the power equipment to be tested from power frequency to low frequency voltage;

7-4) Identify the insulation defect of the power equipment to be tested through the evaluation function;

The evaluation function is as follows:

Compare the change trend of each discharge characteristic parameter of the power equipment under test with the change trend of the corresponding discharge characteristic parameter of each insulation defect model: if the initial voltage U _PDIV of the power equipment under test changes from power frequency to low frequency voltage If the trend is consistent with the change trend of the insulation defect model, then let k=1, otherwise let k=0; compare the q ⁺ _max , q ^- _max , q ⁺ _mean , q ^- _mean , rate ⁺ and rate ^- the variation trend of the 6 discharge characteristic parameters, and the value of _mi is obtained: if the variation trend is consistent, set _mi = 1, otherwise set _mi = 0; compare the power to be measured The change trend of the characteristic parameters of each discharge spectrum corresponding to the equipment and each defect model from power frequency to low frequency voltage: if the change trend is consistent, set n _i =1, otherwise set n _i =0; thus the power equipment under test is calculated Evaluation function value compared with the variation trend of discharge characteristic parameters of each insulation defect model;

Select the insulation defect model corresponding to the maximum value of each evaluation function value, the discharge mode of the insulation defect model is the most likely insulation defect of the equipment under test, and the partial discharge diagnosis ends.

Features and beneficial effects of the present invention are:

Compared with the prior art, the present invention is based on the partial discharge test under power frequency to low frequency voltage, can obtain more information about insulation defects, and avoid accidental insulation damage caused by applying different voltage amplitudes. The variation trend of discharge characteristic parameters is used for identification, which weakens the influence of signal attenuation. In addition, for some devices under test with large capacitance, the use of partial discharge detection under low frequency voltage can effectively reduce the capacity of the test power supply.

Description of drawings

Fig. 1 is the overall flow chart of the method of the present invention.

Fig. 2 is a structural schematic diagram of the partial discharge test platform under power frequency to low frequency voltage of the present invention.

Fig. 3 is a schematic diagram of an insulation defect model of an embodiment of the present invention; wherein: (a) is a composite discharge model of an internal oil gap and along a surface, (b) is a uniform field discharge model in insulating oil, and (c) is a corona discharge in insulating oil Model, (d) is the surface discharge model of oil-paper insulation.

Detailed ways

The present invention proposes a method for diagnosing partial discharge based on feature extraction under power frequency to low frequency voltage. The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments. The described embodiments are only part of the embodiments of the present invention. , but not all examples.

A partial discharge diagnosis method based on feature extraction under power frequency to low frequency voltage proposed by the present invention involves power frequency 50Hz, and low frequency 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz. The overall process is shown in Figure 1 , including the following steps:

1) Construct a partial discharge test platform under power frequency to low frequency voltage. The composition of the partial discharge test platform of the present invention is shown in Figure 2, including: high voltage power amplifier, arbitrary waveform generator, protective resistor, insulation defect model, first measurement impedance, second measurement impedance, coupling capacitance, partial discharge measurement and analysis instrument, high voltage probe and oscilloscope. The selected components must meet the requirements of "GBT 7354-2003 Partial Discharge Measurement". The high-voltage power amplifier model selected in this embodiment is Trek model 50/12, the protection resistor is 1MΩ, the measurement impedance model is CPL 542, and the coupling capacitor is 400pF, the model of partial discharge measurement and analyzer is MPD 600.

The arbitrary waveform generator and the high-voltage power amplifier are connected by a BNC line (coaxial cable connection), the high-voltage power amplifier and the protection resistor are connected by a high-voltage cable, and the protection resistor is connected with the insulation defect model and the coupling capacitor by a high-voltage cable respectively, The high-voltage probe is connected to the coupling capacitor through a high-voltage cable, the insulation defect model is connected to the first measurement impedance through a metal wire, the coupling capacitor is connected to the second measurement impedance through a metal wire, and the first measurement impedance and the second measurement impedance are respectively connected to the partial discharge measurement It is connected with the analyzer through the BNC line, and the high-voltage probe is connected with the oscilloscope through the BNC line.

When working, the sine wave signals of different frequencies (50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz) output by the arbitrary waveform generator are boosted by the high-voltage power amplifier to obtain sinusoidal voltages of different frequencies. The sinusoidal voltage output by the high-voltage power amplifier is applied to the insulation defect model after passing through the protection resistor, and the coupling capacitor is connected in parallel with the insulation defect model. The first measurement impedance connected in series with the defect model is used to collect the discharge pulse signal, the second measurement impedance connected in series with the coupling capacitor is used to collect the voltage synchronization signal, and the two signals are input to the partial discharge storage and analysis system. Use a high voltage probe and an oscilloscope to measure the applied voltage to monitor the real-time voltage and record the discharge initiation voltage.

2) Make an insulation defect model; the insulation defect model made in this example is shown in Figure 3, including the internal oil gap and the compound discharge model along the surface, as shown in Figure 3(a), the sample consists of three layers of 1mm thick insulating paper The middle layer is dug to form an oil gap, and the high-voltage pointed electrode is against the sample; the uniform field discharge model of insulating oil, as shown in Figure 3(b), the high-voltage electrode and the ground electrode are both circular plates, and the oil The gap distance is 2.5mm; the corona discharge model in oil, as shown in Figure 3(c), the distance between the high-voltage tip electrode and the ground electrode is 5mm; the surface discharge model, as shown in Figure 3(d), consists of two hemispherical electrodes , the radius of the ball is 2.5mm, the electrode spacing is 2mm, and the high-voltage electrode and the ground electrode are close to the 1mm thick insulating cardboard.

The above four insulation defect models are the most common insulation defect models, and operators can also design and manufacture insulation defect models according to the conditions of the equipment to be tested and experimental needs.

3) For each insulation defect model produced in step 2), carry out partial discharge tests at 50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz frequency voltages on the partial discharge test platform built in step 1), The partial discharge test of each insulation defect model at each frequency and voltage is repeated 10 to 15 times, and the average value of each test data at the frequency and voltage is taken as the corresponding partial discharge test of the insulation defect model at the selected frequency and voltage. Discharge test data. In each test, adjust the frequency and amplitude of the output waveform of the arbitrary waveform generator, record and store the partial discharge signal data of each test at 1.05 times the discharge initiation voltage, and obtain each insulation defect model at each frequency and voltage Partial discharge test data;

4) Extract discharge characteristic parameters from the partial discharge test data of each insulation defect model obtained in step 3) at each frequency and voltage to form a reference characteristic library. The discharge characteristic parameters include: initial voltage U _PDIV ; maximum discharge capacity q ⁺ _max in the positive half cycle and q ^- _max in the negative half cycle, average discharge capacity in the positive half cycle q ⁺ _mean and average discharge capacity in the negative half cycle q ^- _mean , positive Half-cycle pulse repetition rate rate ⁺ and negative half-cycle pulse repetition rate rate ^- ; phase distribution map of maximum discharge capacity Average discharge phase distribution map Phase distribution spectrum of discharge times The characteristic parameters of each spectrum, including the positive semicircular skewness S _k and negative semicircular skewness S _k of each spectrum, the positive semicircular kurtosis K _u and negative semicircular kurtosis K _u of each spectrum, the asymmetry of each spectrum A _sy and correlation degree C _c ; the characteristic parameters of the frequency distribution spectrum nQ of the discharge quantity, including: skewness S _k , kurtosis _Ku , asymmetry degree _Asy , correlation degree C _c ; among them, the characteristic parameters of each discharge spectrum There are 22 in total.

S _k is the skewness parameter of the spectrum, which represents the deviation of the spectrum from the normal distribution. When S _k >0, the spectrum is left-skewed relative to the normal distribution, and when S _k <0, the spectrum is right-skewed relative to the normal distribution. Its expression is:

In the formula, x _i is the discrete value of the abscissa of the map; n is the number of x _i ; f( _xi ) is the probability of x _i ; u is the mean; σ is the standard deviation.

In the formula, y _i is the value of the ordinate corresponding to x _i .

K _u is the kurtosis parameter of the spectrum, which represents the sharpness of the spectrum relative to the normal distribution. When K _u >0, the spectrum is sharper than the normal distribution, and when K _u <0, the spectrum is flatter than the normal distribution. Its expression is:

A _sy is the degree of asymmetry. When A _sy > 1, the average value in the negative half cycle range of the spectrum is greater than the average value in the positive half cycle range. When A _sy <1, the average value in the positive half cycle range of the spectrum is greater than the average value in the negative half cycle range of the spectrum. The formula is as follows:

In the formula, y _i ⁺ and y _i ^- are the ordinate values of the spectrum in the positive and negative half cycles respectively; N ₁ and N ₂ are the phase window numbers of the positive and negative half cycles respectively, which can generally be 180, that is, the positive and negative half cycles are divided into 180 a phase window.

C _c is the correlation degree, which represents the similarity of the positive and negative half-circle contour distribution of the spectrum. If C _c tends to 1, the distribution is more similar; if C _c tends to 0, the distribution difference is greater. Its expression is:

In the formula, N is the number of phase windows of a half cycle, generally 180, that is, a half cycle is divided into 180 phase windows.

5) Place the power equipment to be tested at the position of the insulation defect model in the partial discharge test platform constructed in step 1), conduct a partial discharge test under power frequency to low frequency voltage for the power equipment to be tested, and record the voltage of the power equipment to be tested at each frequency The corresponding partial discharge test data below; the voltage frequency includes 50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz, and the voltage amplitude is 1.05 times the discharge initiation voltage.

6) Extracting the discharge characteristic parameters in the partial discharge test data of the power equipment under test at each frequency and voltage, the characteristic parameters include but not limited to all the parameters designed in step 4).

7) Compare the change trend of the discharge characteristic parameters of the power equipment to be tested with the change trend of the corresponding discharge characteristic parameters of each insulation defect model, and identify the insulation defects of the power equipment to be tested through the evaluation function; the specific steps are as follows:

7-1) Select any insulation defect model and any discharge characteristic parameter, take the voltage frequency of the partial discharge test as the abscissa, and the discharge characteristic parameter value as the ordinate, use the insulation defect model at each frequency voltage Based on the discharge characteristic parameter value recorded in the partial discharge test data, the change trend diagram of the selected discharge characteristic parameter of the insulation defect model from power frequency to low frequency voltage is obtained;

7-2) Repeat step 7-1) to obtain the change trend diagram of each discharge characteristic parameter of each insulation defect model from power frequency to low frequency voltage;

7-3) Repeat step 7-1) for the power equipment to be tested to obtain a trend diagram of each discharge characteristic parameter corresponding to the power equipment to be tested from power frequency to low frequency voltage;

7-4) Identify the insulation defect of the power equipment to be tested through the evaluation function;

The evaluation function is as follows:

Compare the change trend of each discharge characteristic parameter of the power equipment under test with the change trend of the corresponding discharge characteristic parameter of each insulation defect model: if the initial voltage U _PDIV of the power equipment under test changes from power frequency to low frequency voltage The trend is consistent with the change trend of the insulation defect model, and the change trend is consistent with the selected discharge characteristic parameter (for example: the initial voltage) with the frequency change The overall trend of increasing or decreasing is consistent, then let k=1, On the contrary, set k=0; compare the 6 discharge characteristics of q ⁺ _max , q ^- _max , q ⁺ _mean , q ^- _mean , rate ⁺ and rate ^- of the power equipment under test and each insulation defect model from power frequency to low frequency voltage The change trend of the parameter, get the value of m _i : if the change trend is consistent, set m _i = 1, otherwise set m _i = 0, where i represents the number of the discharge characteristic parameter, and the value of i is 1 to 6; compare the measured The change trend of the characteristic parameters of each discharge spectrum corresponding to the power equipment and each defect model from power frequency to low frequency voltage: if the change trend is consistent, set n _i =1, otherwise set n _i =0, i represents the discharge characteristic parameter Numbering, the value of i is 1 to 22; thereby calculating the evaluation function value of the electrical equipment to be tested and the discharge characteristic parameter variation trend comparison of each insulation defect model, the present invention obtains 4 evaluation function values altogether;

Select the insulation defect model corresponding to the maximum value of each evaluation function value, the discharge mode of the insulation defect model is the most likely insulation defect of the equipment under test, and the partial discharge diagnosis ends.

The invention has the beneficial effects of obtaining more information about insulation defects based on the partial discharge test under power frequency to low frequency voltage, and avoiding accidental insulation damage caused by applying different voltage amplitudes. Using the change trend of feature quantity to identify, weaken the influence of signal attenuation. In addition, for some devices under test with large capacitance, the use of partial discharge detection under low frequency voltage can effectively reduce the capacity of the test power supply.

The above is only a preferred embodiment of the present invention, it should be pointed out that, for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications can also be made. It should be regarded as the protection scope of the present invention.

Claims (1)

1. A partial discharge diagnosis method based on feature extraction under power frequency to low frequency voltage, it is characterized in that, the method comprises the following steps: 1) Build a partial discharge test platform under power frequency to low frequency voltage; the partial discharge test platform includes: high voltage power amplifier, arbitrary waveform generator, protection resistor, insulation defect model, first measurement impedance, second measurement impedance, coupling capacitance , partial discharge measurement and analyzer, high voltage probe and oscilloscope; The arbitrary waveform generator is connected to the high-voltage power amplifier through a coaxial cable connection, the high-voltage power amplifier is connected to the protection resistor through a high-voltage cable, the protection resistor is connected to the insulation defect model, and the coupling capacitor is respectively connected through a high-voltage cable, and the high-voltage probe is connected to the coupling The capacitor is connected through a high-voltage cable, the insulation defect model is connected to the first measurement impedance through a metal wire, the coupling capacitor is connected to the second measurement impedance through a metal wire, and the first measurement impedance and the second measurement impedance are connected to the partial discharge measurement and analyzer respectively. Coaxial cable connection, the high voltage probe is connected to the oscilloscope through the coaxial cable connection; 2) Make an insulation defect model; The insulation defect model includes: internal oil gap and surface discharge model, insulating oil uniform field discharge model, corona discharge model in oil and surface discharge model, and defect models that may exist according to the specific conditions of the equipment; 3) For each insulation defect model produced in step 2), carry out partial discharge tests at 50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz frequency voltages on the partial discharge test platform constructed in step 1); In each test, adjust the frequency and amplitude of the output waveform of the arbitrary waveform generator, record and store the data of each partial discharge test at 1.05 times the discharge initiation voltage; each insulation defect model is tested at each frequency voltage The partial discharge test is repeated 10 to 15 times, and the average value of each test data at the frequency and voltage is taken as the partial discharge test data corresponding to the insulation defect model at the selected frequency voltage, and each insulation defect model is obtained at each Partial discharge test data under frequency voltage; 4) Extract discharge characteristic parameters from the partial discharge test data of each kind of insulation defect model obtained in step 3) to form a reference characteristic library; the discharge characteristic parameters include: initial voltage U _PDIV ; positive half cycle maximum Discharge capacity q ⁺ _max and negative half-cycle maximum discharge capacity q ^- _max , positive half-cycle average discharge capacity q ⁺ _mean and negative half-cycle average discharge capacity q ^- _mean , positive half-cycle pulse repetition rate rate ⁺ and negative half-cycle pulse repetition rate rate ^- ; max Discharge phase distribution map Average discharge phase distribution map Phase distribution spectrum of discharge times The characteristic parameters of , including: positive semicircular skewness S _k and negative semicircular skewness S _k of each spectrum, positive semicircular kurtosis K _u and negative semicircular kurtosis K _u of each spectrum, asymmetry of each spectrum degree A _sy and correlation degree C _c ; the characteristic parameters of the frequency distribution spectrum nQ of discharge capacity, including: skewness S _k , kurtosis K _u , asymmetry degree _Asy , correlation degree C _c ; 5) Place the power equipment to be tested at the location of the insulation defect model in the partial discharge test platform constructed in step 1), conduct a partial discharge test on the power equipment to be tested, and record the corresponding partial discharge tests of the power equipment to be tested under each frequency and voltage Data, the voltage frequency includes 50Hz, 40Hz, 30Hz, 20Hz, 10Hz, 5Hz, 1Hz, 0.1Hz, and the voltage amplitude is 1.05 times the discharge initiation voltage; 6) Extract the discharge characteristic parameters in the partial discharge test data of the power equipment to be tested under each frequency and voltage; 7) Compare the change trend of the discharge characteristic parameters of the power equipment to be tested with the change trend of the corresponding discharge characteristic parameters of each insulation defect model, and identify the insulation defects of the power equipment to be tested through the evaluation function; the specific steps are as follows: 7-1) Select any insulation defect model and any discharge characteristic parameter, take the voltage frequency of the partial discharge test as the abscissa, and the discharge characteristic parameter value as the ordinate, use the insulation defect model at each frequency voltage Based on the discharge characteristic parameter value recorded in the partial discharge test data, the change trend diagram of the selected discharge characteristic parameter of the insulation defect model from power frequency to low frequency voltage is obtained; 7-2) Repeat step 7-1) to obtain the change trend diagram of each discharge characteristic parameter of each insulation defect model from power frequency to low frequency voltage; 7-3) Repeat step 7-1) for the power equipment to be tested to obtain a trend diagram of each discharge characteristic parameter corresponding to the power equipment to be tested from power frequency to low frequency voltage; 7-4) Identify the insulation defect of the power equipment to be tested through the evaluation function; The evaluation function is as follows: <mrow><mi>y</mi><mo>=</mo><mn>10</mn><mi>%</mi><mo>&amp;times;</mo><mi>k</mi><mo>+</mo><mfrac><mrow><mn>40</mn><mi>%</mi></mrow><mn>6</mn></mfrac><mo>&amp;times;</mo><munderover><mo>&amp;Sigma;</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mo>mn></mrow><mn>6</mn></munderover><msub><mi>m</mi><mi>i</mi></msub><mo>+</mo><mfrac><mrow><mn>50</mn><mi>%</mi></mrow><mn>22</mn></mfrac><mo>&amp;times;</mo><munderover><mo>&amp;Sigma;</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mn>22</mn></munderover><msub><mi>n</mi><mi>i</mi></msub><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>6</mn><mo>)</mo></mrow></mrow> Compare the change trend of each discharge characteristic parameter of the power equipment under test with the change trend of the corresponding discharge characteristic parameter of each insulation defect model: if the initial voltage U _PDIV of the power equipment under test changes from power frequency to low frequency voltage If the trend is consistent with the change trend of the insulation defect model, then let k=1, otherwise let k=0; compare the q ⁺ _max , q ^- _max , q ⁺ _mean , q ^- _mean , rate ⁺ and rate ^- the variation trend of the 6 discharge characteristic parameters, and the value of _mi is obtained: if the variation trend is consistent, set _mi = 1, otherwise set _mi = 0; compare the power to be measured The change trend of the characteristic parameters of each discharge spectrum corresponding to the equipment and each defect model from power frequency to low frequency voltage: if the change trend is consistent, set n _i =1, otherwise set n _i =0; thus the power equipment under test is calculated Evaluation function value compared with the variation trend of discharge characteristic parameters of each insulation defect model; Select the insulation defect model corresponding to the maximum value of each evaluation function value, the discharge mode of the insulation defect model is the most likely insulation defect of the equipment under test, and the partial discharge diagnosis ends.