Method for rapidly evaluating transformer oil paper insulation moisture degree based on dielectric loss integral spectrum

Abstract

The invention belongs to the field of transformer oil paper insulation state evaluation, and particularly provides a method for quickly evaluating transformer oil paper insulation damp degree based on dielectric loss integral spectrum.The method comprises the steps that FDS testing is conducted on oil paper insulation samples with different damp degrees, and dielectric loss values and complex dielectric constants corresponding to high-frequency sections are obtained; establishing a relational expression between a score Poynting-Thomson model parameter and a complex dielectric constant, obtaining a score Poynting-Thomson model according to FDS measured data of a high-frequency section, and further calculating a dielectric loss value of a test site corresponding to a low-frequency section; performing integral processing on the dielectric loss value, constructing a dielectric loss integral spectrum curve, selecting the dielectric loss integral value as a characteristic parameter, and establishing a function relation between the selected characteristic parameter and the moisture content of the insulation paper for evaluating the moisture state of the oil paper insulation. According to the method, the transformer oil paper insulation moisture degree can be evaluated only by measuring the frequency domain dielectric spectrum under the high frequency band, and the insulation diagnosis time is greatly shortened.

Description

Technical Field

[0001] The present invention belongs to the field of evaluating the state of oil-paper insulation of transformers. Specifically, it relates to a method for quickly evaluating the moisture content of oil-paper insulation of transformers based on dielectric loss integral spectra. Background Art

[0002] Oil-immersed transformers are key equipment in the power system, and the performance of their oil-paper insulation system directly affects the operation life of the transformers. During the long-term operation of transformers, the intrusion of external moisture and the moisture generated by the aging of the oil-paper itself will cause the water content in the insulating paper to gradually increase. The increase in water content will lead to an increase in the dielectric loss value and a decrease in the breakdown voltage, thus damaging the internal insulation structure of the transformer and affecting the normal operation of the equipment. Therefore, it is very important to propose a method for quickly and accurately evaluating the moisture content of oil-paper insulation of transformers for extending the service life of transformers and improving the stability of the power system.

[0003] Currently, the methods for measuring the moisture content of oil-paper insulation are mainly physical and chemical parameter methods and dielectric response methods. Physical and chemical parameters include extraction methods, Karl Fischer titration methods, and moisture balance curves, etc. However, these methods have the disadvantages of cumbersome operation, easy introduction of impurities into the oil-paper insulation system, and reduction of the insulation strength of the equipment, and are difficult to implement on transformers under actual working conditions. The frequency-domain dielectric spectroscopy (FDS) using dielectric response theory is more suitable for on-site diagnosis and evaluation of the insulation state of oil-immersed transformers because of its strong anti-interference ability and rich information carried by the test curve. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method for quickly evaluating the moisture content of oil-paper insulation of transformers based on dielectric loss integral spectra. Only by measuring the frequency-domain dielectric spectra in the high-frequency band can the moisture content of the oil-paper insulation of transformers be evaluated, and the insulation diagnosis time can be greatly shortened.

[0005] To solve the above technical problem, the technical solution adopted by the present invention is: A method for quickly evaluating the moisture content of oil-paper insulation of transformers based on dielectric loss integral spectra, comprising the following steps: Step 1: Prepare oil-paper insulation samples with different moisture contents, and perform FDS tests using a frequency-domain dielectric spectroscopy detection system to obtain the dielectric loss values and complex dielectric constants corresponding to the high-frequency section; Step 2: Establish the relationship between the parameters of the fractional Poynting-Thomson model and the complex dielectric constant of the oil-paper insulation system. According to the FDS measured data of the high-frequency section, the model parameters of the fractional Poynting-Thomson model are solved in combination with the PSO particle swarm algorithm to obtain the fractional Poynting-Thomson model. Substitute the frequency values ​​of the characteristic frequency points of the low-frequency section into the fractional Poynting-Thomson model, deduce and calculate the dielectric loss value of the corresponding test point in the low-frequency section, and obtain the FDS curve of the full-frequency section. Step 3: By integrating the FDS curve in step 2, a dielectric loss integral spectrum curve is constructed, and the dielectric loss integral value at a suitable frequency point is selected as a characteristic parameter for evaluating the moisture state of oil-paper insulation, and the selected characteristic parameter and the moisture content M of the insulation paper are established. c The functional relationship between Step 4: Based on the characteristic parameters established in step 3 and the moisture content M of the insulation paper c The functional relationship between them is used to verify the accuracy of using dielectric loss integral spectrum to evaluate transformer oil-paper insulation.

[0006] In a preferred embodiment, in step 1, the frequency of the high frequency section is 10 0 -10 3 Hz; in the step 2, the frequency of the low frequency segment is 10 -3 -10 0 Hz.

[0007] In a preferred solution, in step 1, the steps for obtaining the dielectric loss value and the complex dielectric constant corresponding to the high-frequency section are: S1.1. Pretreatment of insulation paper: Cut the insulation paper and place it in a drying oven for drying; S1.2. Stack multiple pre-treated insulating papers into insulating paperboards, place them in the air to allow them to absorb moisture naturally, use an electronic balance to weigh the paperboards after moisture absorption in real time, and immediately immerse them in insulating oil after obtaining insulating paperboards with different moisture contents, so that the oil and paper are fully mixed and impregnated; S1.3, place the prepared oil-paper insulation samples with different moisture levels in a three-electrode FDS test, and discharge for 1.5-2.5 hours before the test to eliminate the influence of residual charge on the test results; S1.4. Select the frequency domain dielectric spectrum detection system, set the frequency domain test range, and conduct scanning tests through the frequency domain dielectric spectrum detection system. When the test points in the high-frequency section are scanned, pause the equipment and record the test time. Calculate the complex dielectric constant values ​​at different frequency points according to the parallel plate capacitance formula, and then calculate the corresponding dielectric loss value according to the complex dielectric constant value. The parallel plate capacitance formula is: ; In the formula, is the complex capacitance of the dielectric, and ε 0 is the permittivity of free space, is the complex relative permittivity of the dielectric, S is the cross-sectional area of the parallel-plate electrodes, d is the distance between the parallel-plate electrodes, is the angular frequency.

[0008] In a preferred embodiment, the dielectric loss value is calculated by the following formula: ; In the formula, ε' is the real part of the complex relative permittivity; ε'' is the imaginary part of the complex relative permittivity, and tanδ represents the dielectric loss value.

[0009] In a preferred embodiment, in the second step, the relationship between the fractional Poynting-Thomson model parameters and the complex relative permittivity of the oil-paper insulation system is: ; ; In the formula: α, β, γ, ε a , ε b and τ are six independent fractional Poynting–Thomson model parameters, where α is the shape parameter describing the symmetric distribution of the relaxation process, β and γ are the shape parameters describing the asymmetric distribution of the relaxation process, and let 0 < β < γ < α < 1; τ represents the relaxation time; ε a is the dielectric parameter reflecting the low-frequency relaxation process; ε b is the dielectric parameter reflecting the high-frequency relaxation process; M'(ω) and M''(ω) are the real and imaginary parts of the dielectric modulus; The relationship between the complex relative permittivity and the dielectric modulus is: ; In the formula: ε' is the real part of the complex relative permittivity; ε'' is the imaginary part of the complex relative permittivity.

[0010] In a preferred embodiment, in the second step, the method for obtaining the fractional Poynting-Thomson model is as follows: Use the PSO algorithm and the least squares method for parameter identification, determine the constraint conditions of the fractional Poynting-Thomson model parameters, substitute the measured FDS data in the high-frequency section, and solve for the fractional Poynting-Thomson model parameters α, β, γ, ε a , ε b and τ the specific values of, to obtain the fractional Poynting-Thomson model, where the constraint conditions of the Poynting-Thomson model parameters are: 0 < α, β, γ < 1, 0 < ε a , ε b < 500, 0 < τ < 1000.

[0011] In a preferred embodiment, in the third step, the following steps are included: S3.1. Integrate the dielectric loss values at n different points in the full frequency band of the FDS curve and fit them into a dielectric loss integral spectrum curve that varies with frequency, where the definition formula of the dielectric loss integral S δ is: ; In the formula, f represents the frequency, n represents selecting n sample points with different values in the interval [10 3 , 10 -3 ; S3.2. Frequency point selection, and the dielectric loss integral value S δ corresponding to the selected frequency point is used as the characteristic parameter; S3.3. Calculate the characteristic parameter values of samples with different moisture contents, perform function fitting, and obtain the functional relationship between the characteristic parameter and the water content M c .

[0012] In a preferred embodiment, in the step S3.2, the moment of 10 -3 Hz is used as the characteristic frequency point, and the corresponding dielectric loss integral value S δ (10 -3 ) is used as the characteristic parameter.

[0013] In a preferred embodiment, in the fourth step, the following steps are included: S4.1. Randomly prepare two samples with different moisture contents, perform FDS tests on the two groups of samples to obtain relevant data in their high-frequency sections; calculate the dielectric loss integral spectra of the two groups of samples according to the second and third steps; S4.2. Extract the characteristic parameters in the dielectric loss integral spectrum and substitute them into the functional relationship between the characteristic parameters and the water content M c to obtain the calculated value of M c Compare the error between the calculated value of the water content and the true value to verify the accuracy of the dielectric loss integral spectrum in evaluating the oil-paper insulation of the transformer.

[0014] In the preferred solution, in step four, it further includes step S4.3, comparing the time taken to construct the dielectric loss integral spectrum and the traditional FDS full-frequency band test method to verify the efficiency of the dielectric loss integral spectrum in diagnosing oil-paper insulation.

[0015] A method for quickly evaluating the moisture content of the oil-paper insulation of a transformer based on the dielectric loss integral spectrum provided by the present invention has the following beneficial effects: 1. By constructing the relationship between the fractional Poynting–Thomson model and the complex dielectric constant of the oil-paper insulation, and combining the parameter identification of the PSO particle swarm algorithm, the present invention can calculate the dielectric loss value in the low-frequency band using the data in the high-frequency band, and construct a dielectric loss integral spectrum curve by integrating the dielectric loss values in the FDS curve.

[0016] 2. The present invention extracts the characteristic parameters of the dielectric loss integral spectrum, establishes the functional relationship between the characteristic parameters and the water content M c and verifies the accuracy of the prediction of this relationship.

[0017] 3. The present invention proves that the moisture condition analysis of the oil-paper insulation system can be completed by combining the high-frequency band data measured by FDS with the calculation of the dielectric loss integral spectrum model. It takes about 3 minutes to obtain the test data, and the calculation time of the model can be ignored. Compared with the full-frequency band sweep test that takes 38 minutes, the test time is significantly shortened, making this method suitable for on-site diagnosis of the main insulation of transformers with a short window time. Description of the Drawings

[0018] The present invention will be further described below with reference to the drawings and embodiments: Figure 1 is the fractional Poynting-Thomson model diagram; Figure 2 is the schematic diagram of the frequency-domain dielectric loss integral spectrum; Figure 3 is the frequency-domain dielectric loss spectrum diagram of samples with different moisture levels; Figure 4 is the comparison curve diagram of the calculated dielectric loss value and the FDS measured dielectric loss value; Figure 5 is the dielectric loss integral spectrum S δ - f curve diagram of samples with different water contents; Figure 6 is the characteristic parameter Sδ (10 -3 ) The fitting curve graph with M c ; Figure 7 For the dielectric loss integral spectra S of the samples with moisture contents of 3.46% and 4.65 δ - f curve graph. Specific implementation manners

[0019] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] A method for quickly evaluating the moisture content of transformer oil-paper insulation based on dielectric loss integral spectra, comprising the following steps: Step 1: Prepare oil-paper insulation samples with different moisture contents, and perform FDS tests using a frequency-domain dielectric spectroscopy detection system to obtain the dielectric loss values and complex dielectric constants corresponding to the high-frequency section in about 3 minutes. The frequency of the high-frequency section is 10 0 -10 3 Hz, including 12 characteristic frequency points, which are 0.46 Hz, 1 Hz, 2 Hz, 5 Hz, 10 Hz, 20 Hz, 40 Hz, 70 Hz, 110 Hz, 220 Hz, 470 Hz, and 1000 Hz respectively.

[0021] The steps for obtaining the dielectric loss values and complex dielectric constants corresponding to the high-frequency section are as follows: S1.1: Pretreat the insulating paper: Cut the insulating paper and place it in a drying oven to dry for 48 h for standby.

[0022] S1.2: Stack 8 pieces of pretreated insulating paper into an insulating cardboard with a thickness of about 1 mm, place it in the air to absorb moisture naturally, use an electronic balance to weigh the cardboard in real time after moisture absorption, and immediately put the obtained insulating cardboard with different moisture contents into insulating oil for oil immersion for 48 h to fully mix and impregnate the oil and paper.

[0023] In this embodiment, the moisture contents of the insulating cardboard are 1.0%, 2.0%, 3.0%, 4.0%, and 5.0% respectively.

[0024] S1.3: Place the prepared oil-paper insulation samples with different moisture contents in a three-electrode for FDS tests, and discharge for 2 h before the test to eliminate the influence of residual charges on the test results.

[0025] S1.4: Select a frequency-domain dielectric spectroscopy detection system, specifically select the FDS insulation diagnostic instrument IDAX300, and set the frequency-domain test range to 10 -3 -10 3Hz, a scanning test is carried out through a frequency-domain dielectric spectroscopy detection system. If the full frequency band is scanned, the working time is about 38 minutes. In this embodiment, when the scanning of the test points in the high-frequency section is completed, the device is paused and the test time is recorded, which is about 3 minutes. The complex dielectric constant values at different frequency points are calculated according to the parallel capacitor plate formula, and then the corresponding dielectric loss values are calculated based on the complex dielectric constant values.

[0026] The complex capacitance and the complex dielectric constant are directly related through the parallel plate capacitor formula, and this relationship can be used to analyze the frequency response and loss mechanism of dielectrics.

[0027] The parallel plate capacitor formula is: ; In the formula, is the complex capacitance of the dielectric, ε 0 is the vacuum permittivity, is the complex dielectric constant of the dielectric, S is the cross-sectional area of the parallel plate electrodes, d is the distance between the parallel plate electrodes, is the angular frequency.

[0028] The dielectric loss value is calculated by the following formula: ; In the formula, ε' is the real part of the complex dielectric constant; ε'' is the imaginary part of the complex dielectric constant, and tanδ represents the dielectric loss value.

[0029] Step 2: Establish a relationship between the fractional Poynting-Thomson model parameters and the complex dielectric constant of the oil-paper insulation system. According to the measured FDS data in the high-frequency section, combined with the PSO particle swarm optimization algorithm, solve the model parameters of the fractional Poynting-Thomson model to obtain the fractional Poynting-Thomson model. Substitute the frequency values of the characteristic frequency points in the low-frequency section into the fractional Poynting-Thomson model, and deduce and calculate the dielectric loss values at the corresponding test sites in the low-frequency section to obtain the FDS curve of the full frequency section. The frequency in the low-frequency section is 10 -3 -10 0 Hz. In this way, the tester can obtain the FDS curve of the full frequency section (10 -3 -10 3 Hz) without performing a time-consuming 35-minute scanning test in the low-frequency band.

[0030] Specifically, it includes the following steps: S2.1: Establish a relationship between the fractional Poynting-Thomson model parameters and the complex dielectric constant of the oil-paper insulation system: ; ; In the formula: α, β, γ, ε a , ε b and τ are six independent fractional Poynting–Thomson model parameters, where α is the shape parameter describing the symmetric distribution of the relaxation process, β and γ are the shape parameters describing the asymmetric distribution of the relaxation process, let 0 < β < γ < α <1; τ represents the relaxation time, which is closely related to the polarization type and the position where the relaxation peak appears; ε a is the dielectric parameter reflecting the low-frequency relaxation process, which is closely related to the peak value (maximum value) of the relaxation peak; ε b is the dielectric parameter reflecting the high-frequency relaxation process, which is closely related to the peak value (minimum value) of the relaxation peak; M'(ω) and M''(ω) are the real and imaginary parts of the dielectric modulus.

[0031] The relationship between the complex dielectric constant and the dielectric modulus is: ; In the formula: ε' is the real part of the complex dielectric constant; ε'' is the imaginary part of the complex dielectric constant.

[0032] S2.2. Use the PSO algorithm and the least squares method for parameter identification, determine the constraint conditions of the P-T model parameters, substitute the measured data in the high-frequency band of FDS, and solve for the specific values of the P-T model parameters α, β, γ, ε a , ε b and τ to obtain the fractional Poynting-Thomson model.

[0033] The constraint conditions of the P-T model parameters are respectively: 0 < α, β, γ < 1, 0 < ε a , ε b < 500, 0 < τ < 1000.

[0034] In the process of identifying the parameters of the Poynting-Thomson model using the particle swarm optimization (PSO) algorithm and the least squares method, it is first necessary to complete the setting of the initial state of the particle swarm, including the spatial position and movement speed of the particles. Subsequently, the least squares method is used to improve the accuracy of parameter identification, and the sum of the squares of the residuals between the model-fitted data and the measured data is used as the fitness evaluation index to calculate the fitness of each particle. In this process, the least squares method performs local linearization on nonlinear problems, effectively reducing the fitting error of the model parameters.

[0035] The objective function Q of the least squares method is: ; In the formula: ε' is the actual value of the real part of the complex dielectric constant; ε'' is the actual value of the imaginary part of the complex dielectric constant; ε' f is the fitted value of the real part of the complex dielectric constant, ε'' f is the fitted value of the imaginary part of the complex dielectric constant, Q represents the sum of the squares of the errors between the actual value and the fitted data.

[0036] Finally, compare the current state with the historical state to determine the best position of the individual and the global optimal position. On this basis, adjust the velocity vector and spatial coordinates of the particles through the iterative update strategy until the preset constraint conditions are met; finally, output the optimal solution of the Poynting-Thomson model parameters.

[0037] S2.3: According to the functional relationship between the P-T model parameters and the dielectric constant in step S2.1, substitute the frequency values of 8 test sites with frequencies of 0.22 Hz, 0.1 Hz, 0.046 Hz, 0.022 Hz, 0.01 Hz, 0.0046 Hz, 0.0022 Hz, and 0.001 Hz in the low-frequency band into the relationship between the fractional Poynting-Thomson model and the complex dielectric constant in step S2.1 to obtain the complex dielectric constant in the low-frequency band, and then calculate its dielectric loss value according to the relationship between the complex dielectric constant and the dielectric loss value.

[0038] Step 3: By integrating the FDS curve in step 2, construct a dielectric loss integral spectrum, select the dielectric loss integral value at an appropriate frequency point as the characteristic parameter for evaluating the moisture content state of the oil-paper insulation, and establish a functional relationship between the selected characteristic parameter and the moisture content M of the insulating paper c between.

[0039] Including the following steps: S3.1. Integrate the dielectric loss values at n different points in the full frequency band of the FDS curve and fit them into a dielectric loss integral spectrum curve that changes with frequency. Among them, the dielectric loss integral Sδ The definition formula is: ; In the formula, f represents the frequency, and n represents selecting n sample points with different numerical values in the interval [10 3 , 10 -3 .

[0040] S3.2. Selection of frequency points: Select the moment of 10 -3 Hz as the characteristic frequency point, and the dielectric loss integral value S δ (10 -3 ) corresponding to the selected frequency point is used as the characteristic parameter; S3.3. Calculate the characteristic parameter values S δ (10 -3 ) of samples with different moisture levels, and use the ExpGrol module in origin for function fitting to obtain the functional relationship between S δ (10 -3 ) and the moisture content M c .

[0041] Step Four: Based on the functional relationship between the characteristic parameters established in Step Three and the moisture content M c of the insulating paper, verify the accuracy of using the dielectric loss integral spectrum to evaluate the oil-paper insulation of the transformer.

[0042] S4.1. Randomly prepare samples with two moisture contents, conduct FDS tests on the two groups of samples to obtain relevant data in the high-frequency band; calculate the dielectric loss integral spectra of the two groups of samples according to Steps Two and Three.

[0043] In this embodiment, samples with moisture contents of 3.46% and 4.65% are randomly prepared.

[0044] S4.2. Extract the characteristic parameter S δ (10 -3 ) in the dielectric loss integral spectrum, substitute it into the functional relationship between S δ (10 -3 ) and the moisture content M c to obtain the calculated value of M c , and compare the error between the calculated value of the moisture content and the true value to verify the accuracy of using the dielectric loss integral spectrum to evaluate the oil-paper insulation of the transformer.

[0045] S4.3. Compare the time taken to construct the dielectric loss integral spectrum and the traditional FDS full-frequency band test method to verify the efficiency of diagnosing oil-paper insulation using the dielectric loss integral spectrum.

[0046] Specific implementation cases: To simulate the moisture ingress situation of the main insulation inside the transformer, oil-paper insulation samples with different moisture levels were prepared in this embodiment. First, the insulating paper was cut, stacked, and pre-dried. Then it was placed indoors to be naturally moistened, and a high-precision electronic balance was used continuously to measure the moisture-absorbed insulating cardboard until it reached the target water content. Five groups of samples with different water contents were set, namely 1.0%, 2.0%, 3.0%, 4.0%, and 5.0%. The FDS test was carried out at 25 °C using the insulation diagnostic instrument IDAX-300, the test voltage was 140 V, and the test frequency was 10 -3 -10 3 Hz, and the full-frequency band test time was 38 min.

[0047] The test was carried out from the high-frequency band to the low-frequency band. After the interval test in the high-frequency band was completed, the instrument was paused and the time was recorded. At this time, the test time was about 3 min; after the recording was completed, the instrument was turned on to complete the remaining low-frequency band test. After all the tests were completed, the frequency-domain dielectric loss spectrograms of samples with different moisture levels as shown Figure 3 were obtained.

[0048] A relationship was established between the fractional Poynting-Thomson model and the complex permittivity of the oil-paper insulation system. The dielectric loss value of the system could be further calculated through the complex permittivity. The fractional Poynting-Thomson model was as shown Figure 1 . The data of the first 12 groups of test points obtained from the FDS test were substituted into the relationship, and the PSO particle swarm algorithm was introduced to solve the multi-parameter nonlinear problem. Combining the least squares method, the Poynting-Thomson model α, β, γ, ε a , ε b and τ a total of 6 model parameters were identified. The specific values are shown in Table 1.

[0049]

[0050] Eight test frequency points within the low-frequency band (10 0 -10 -3 Hz), namely 0.22 Hz, 0.1 Hz, 0.046 Hz, 0.022 Hz, 0.01 Hz, 0.0046 Hz, 0.0022 Hz, and 0.001 Hz, were substituted into the relationship between the fractional Poynting-Thomson model and the complex permittivity of the oil-paper insulation system to obtain the real part value and the imaginary part value of the complex permittivity, and further calculate the dielectric loss values at each point in the low-frequency band. The dielectric loss values in the FDS measured curve and the calculated curve were compared, as shown specifically Figure 4 . Figure 4The measured curve in [it] is obtained from the full-band FDS scanning test, which takes 38 minutes, while the calculated curve is obtained from the high-frequency FDS scanning test combined with the P-T model calculation, taking about 3 minutes.

[0051] It can be seen that although obtaining the FDS calculated curve significantly shortens the test time, the error between the two is too large in the low-frequency band, and the calculated dielectric loss value cannot reflect the true dielectric phenomenon of oil-paper insulation. Therefore, considering fitting the dielectric loss values in the full frequency band into the dielectric loss integral spectrum curve, comparing the dielectric loss integral spectrum curve fitted from the calculated value and the FDS measured value, as Figure 5 shown. It is found that the dielectric loss integral spectrum curves constructed by the measured value and the calculated value have a high degree of similarity and are more suitable for the rapid diagnosis of the insulation performance of transformer oil-paper. This is because the definition of the dielectric loss integral value is the successive accumulation from high frequency to low frequency. When the frequency decreases, the increase in the dielectric loss integral value also decreases. When the frequency decreases to a certain extent, the dielectric loss integral value will no longer show an obvious increasing trend with the increase of frequency points and begins to tend to saturation.

[0052] In addition, the dielectric loss integral value changes significantly in the 10 0 -10 3 Hz frequency band, while in the 10 -3 -10 0 Hz frequency band, it hardly changes. It is easier to obtain the test frequency points that can characterize the degree of moisture, and establish the functional relationship between the degree of moisture and the dielectric loss integral value. Therefore, the low-frequency 10 -3 Hz moment is selected as the characteristic frequency point, and its dielectric loss integral value S δ (10 -3 )is used as the characteristic parameter to characterize the moisture state of oil-paper insulation. The measured and calculated values of S δ (10 -3 )for samples with different water contents are shown in Table 2.

[0053]

[0054] It can be seen that the measured value and the calculated value of S δ (10 -3 )are extremely close. Therefore, in origin, the calculated values of S δ (10 -3 )under each moisture content are input, and the ExpGrol module is selected for function fitting to obtain the fitting relationship and fitting curve between the water content (M c )and the calculated value of S δ (10 -3 ). The fitting curve is as Figure 6 shown, and the fitting relationship is shown in Table 3.

[0055]

[0056] As can be seen from Table 3, the fitting degree of the relational expression reaches 0.99, indicating that S δ (10 -3 ) can be used as a characteristic parameter to evaluate the moisture content of transformer oil-paper insulation. When the moisture content of the oil-paper insulation is more serious, the value of S δ (10 -3 ) is larger. To verify the accuracy of the relational expression proposed in Table 2, samples with water contents of 3.46% and 4.65% were randomly prepared in a laboratory environment, and the dielectric loss integral spectrum curve was calculated according to the above steps. The results are as Figure 7 shown.

[0057] The corresponding characteristic parameters were extracted and substituted into the fitting relational expression in Table 2, and the predicted moisture contents were calculated to be 3.55% and 4.75% respectively. The relative errors of both are within 3%. Therefore, it can be preliminarily determined that the dielectric loss integral spectrum proposed in this paper can effectively evaluate the moisture state of oil-paper insulation. The prediction results are shown in Table 4.

[0058]

[0059] By comparing the dielectric loss integral spectrum constructed by parameter identification of the P-T model through an algorithm with the frequency-domain dielectric spectrum obtained by traditional FDS sweep testing, the conclusion is drawn that the dielectric loss integral spectrum described in this paper quickly establishes an integral model by performing a small number of actual measurements on key high-frequency points, achieving the result of accurately predicting the moisture content of oil-paper insulation. The comparison of the two evaluation methods is shown in Table 5.

[0060]

[0061] It is easy for those skilled in the art to understand that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for quickly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum, characterized in that: The following steps are involved: Step 1: Prepare oil-paper insulation samples with different moisture levels, and use a frequency domain dielectric spectrum detection system to perform FDS testing to obtain the dielectric loss value and complex dielectric constant corresponding to the high-frequency section; Step 2: Establish the relationship between the parameters of the fractional Poynting-Thomson model and the complex dielectric constant of the oil-paper insulation system. According to the FDS measured data of the high-frequency section, the model parameters of the fractional Poynting-Thomson model are solved in combination with the PSO particle swarm algorithm to obtain the fractional Poynting-Thomson model. Substitute the frequency values ​​of the characteristic frequency points of the low-frequency section into the fractional Poynting-Thomson model, deduce and calculate the dielectric loss value of the corresponding test point in the low-frequency section, and obtain the FDS curve of the full-frequency section. Step 3: By integrating the FDS curve in step 2, a dielectric loss integral spectrum curve is constructed, and the dielectric loss integral value at a suitable frequency point is selected as a characteristic parameter for evaluating the moisture state of oil-paper insulation, and the selected characteristic parameter and the moisture content M of the insulation paper are established. c The functional relationship between Step 4: Based on the characteristic parameters established in step 3 and the moisture content M of the insulation paper c The functional relationship between them is used to verify the accuracy of using dielectric loss integral spectrum to evaluate transformer oil-paper insulation.

2. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 1, characterized in that: In the step 1, the frequency of the high frequency section is 10 0 -10 3 Hz; in the step 2, the frequency of the low frequency segment is 10 -3 -10 0 Hz.

3. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 2 is characterized in that: In step 1, the steps for obtaining the dielectric loss value and complex dielectric constant corresponding to the high-frequency section are: S1.

1. Pretreatment of insulation paper: Cut the insulation paper and place it in a drying oven for drying; S1.

2. Stack multiple pre-treated insulating papers into insulating paperboards, place them in the air to allow them to absorb moisture naturally, use an electronic balance to weigh the paperboards after moisture absorption in real time, and immediately immerse them in insulating oil after obtaining insulating paperboards with different moisture contents, so that the oil and paper are fully mixed and impregnated; S1.3, place the prepared oil-paper insulation samples with different moisture levels in a three-electrode FDS test, and discharge for 1.5-2.5 hours before the test to eliminate the influence of residual charge on the test results; S1.

4. Select the frequency domain dielectric spectrum detection system, set the frequency domain test range, and conduct scanning tests through the frequency domain dielectric spectrum detection system. When the test points in the high-frequency section are scanned, pause the equipment and record the test time. Calculate the complex dielectric constant values ​​at different frequency points according to the parallel plate capacitance formula, and then calculate the corresponding dielectric loss value according to the complex dielectric constant value. The parallel plate capacitance formula is: ； In the formula, is the complex capacitance of the dielectric, ε0 is the vacuum dielectric constant, is the complex dielectric constant of the dielectric, S is the cross-sectional area of ​​the parallel plate electrodes, d is the distance between the parallel plate electrodes, is the angular frequency.

4. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 3 is characterized in that: The dielectric loss value is calculated by the following formula: ； In the formula, ε' is the real part of the complex dielectric constant; ε'' is the imaginary part of the complex dielectric constant, and tanδ represents the dielectric loss value.

5. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 1, characterized in that: In the step 2, the relationship between the fractional Poynting-Thomson model parameters and the complex dielectric constant of the oil-paper insulation system is: ； ； Where: α, β, γ, ε a , ε b and τ are six independent fractional Poynting–Thomson model parameters, where α is the shape parameter describing the symmetric distribution of the relaxation process, β and γ is the shape parameter describing the asymmetric distribution of the relaxation process, let 0< β < γ < α <1; τ represents relaxation time; ε a It is the dielectric parameter reflecting the low-frequency relaxation process; ε b It is the dielectric parameter reflecting the high-frequency relaxation process; M'(ω) and M''(ω) are the real and imaginary parts of the dielectric modulus; The relationship between the complex permittivity and dielectric modulus is: ； Where: ε' is the real part of the complex dielectric constant; ε'' is the imaginary part of the complex dielectric constant.

6. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 5 is characterized in that: In step 2, the method for obtaining the fractional Poynting-Thomson model is: using the PSO algorithm and the least squares method to perform parameter identification, determine the constraints of the fractional Poynting-Thomson model parameters, substitute the FDS measured data of the high-frequency section, and solve the fractional Poynting-Thomson model parameters. α, β, γ, ε a , ε b and τ The specific value of is used to obtain the fractional Poynting-Thomson model, where the constraints of the parameters of the Poynting-Thomson model are: 0< α, β, γ <1, 0< ε a , ε b <500, 0<τ<1000.

7. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 1, characterized in that: The step three includes the following steps: S3.

1. Integrate the dielectric loss values ​​of n different points in the full frequency band of the FDS curve and fit them into a dielectric loss integral spectrum curve that varies with frequency, where the dielectric loss integral S δ The definition of is: ； In the formula, f represents the frequency, n represents the interval [10 3 , 10 -3 ] select n sample points with different values; S3.2, frequency point selection, the dielectric loss integral value S corresponding to the selected frequency point δ is the characteristic parameter; S3.

3. Calculate the characteristic parameter values ​​of samples with different moisture levels, perform function fitting, and obtain the characteristic parameters and water content M. c The functional relationship between .

8. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 7 is characterized in that: In step S3.2, 10 -3 Hz is taken as the characteristic frequency point, and the corresponding dielectric loss integral value S δ (10 -3 ) is the characteristic parameter.

9. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 1, characterized in that: The step 4 includes the following steps: S4.

1. Randomly prepare samples with two different moisture contents, and perform FDS tests on the two groups of samples to obtain relevant data of their high frequency bands; calculate the dielectric loss integral spectra of the two groups of samples according to steps 2 and 3; S4.

2. Extract characteristic parameters from dielectric loss integral spectrum and substitute the characteristic parameters and water content M c The functional relationship of M c The calculated value of moisture content is compared with the true value to verify the accuracy of dielectric loss integral spectrum in evaluating transformer oil-paper insulation.

10. The method for rapidly evaluating the degree of moisture in transformer oil-paper insulation based on dielectric loss integral spectrum according to claim 9, characterized in that: The step 4 also includes step S4.3, comparing the time of constructing the dielectric loss integral spectrum with that of the traditional FDS full-band test method, and verifying the high efficiency of the dielectric loss integral spectrum in diagnosing oil-paper insulation.

Images