CN112485609B - Raman spectrum diagnosis method for insulation aging of transformer oil paper - Google Patents
Raman spectrum diagnosis method for insulation aging of transformer oil paper Download PDFInfo
- Publication number
- CN112485609B CN112485609B CN202011117423.3A CN202011117423A CN112485609B CN 112485609 B CN112485609 B CN 112485609B CN 202011117423 A CN202011117423 A CN 202011117423A CN 112485609 B CN112485609 B CN 112485609B
- Authority
- CN
- China
- Prior art keywords
- aging
- raman spectrum
- oil paper
- fuzzy
- oil
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000032683 aging Effects 0.000 title claims abstract description 168
- 238000009413 insulation Methods 0.000 title claims abstract description 104
- 238000003745 diagnosis Methods 0.000 title claims abstract description 81
- 238000001237 Raman spectrum Methods 0.000 title claims abstract description 75
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000001069 Raman spectroscopy Methods 0.000 claims abstract description 32
- 230000002431 foraging effect Effects 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 41
- 238000012360 testing method Methods 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 12
- 230000003595 spectral effect Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000006116 polymerization reaction Methods 0.000 abstract description 34
- 238000001514 detection method Methods 0.000 abstract description 15
- 238000012544 monitoring process Methods 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 abstract description 6
- 230000008569 process Effects 0.000 abstract description 4
- 238000001228 spectrum Methods 0.000 abstract description 4
- 238000013528 artificial neural network Methods 0.000 description 14
- 239000000126 substance Substances 0.000 description 12
- 238000011160 research Methods 0.000 description 7
- CSCPPACGZOOCGX-UHFFFAOYSA-N Acetone Chemical compound CC(C)=O CSCPPACGZOOCGX-UHFFFAOYSA-N 0.000 description 6
- OKKJLVBELUTLKV-UHFFFAOYSA-N Methanol Chemical compound OC OKKJLVBELUTLKV-UHFFFAOYSA-N 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- HYBBIBNJHNGZAN-UHFFFAOYSA-N furfural Chemical compound O=CC1=CC=CO1 HYBBIBNJHNGZAN-UHFFFAOYSA-N 0.000 description 6
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 4
- 230000015556 catabolic process Effects 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 238000006731 degradation reaction Methods 0.000 description 4
- 239000000203 mixture Substances 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 229910002092 carbon dioxide Inorganic materials 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000003878 thermal aging Methods 0.000 description 2
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000001569 carbon dioxide Substances 0.000 description 1
- 229910002091 carbon monoxide Inorganic materials 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 239000011810 insulating material Substances 0.000 description 1
- 239000012774 insulation material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 230000035882 stress Effects 0.000 description 1
- 230000008646 thermal stress Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/65—Raman scattering
Landscapes
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Biochemistry (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
- Housings And Mounting Of Transformers (AREA)
Abstract
A Raman spectrum diagnosis method for aging of oil paper insulation of a transformer realizes aging diagnosis of the oil paper insulation based on a Raman spectrum technology and mainly solves the technical problems that the aging state of the oil paper insulation of the transformer lacks a field electrification monitoring means and the detection process is complicated. The invention simplifies the oil paper insulation Raman spectrogram into a characteristic vector, and defines the oil paper insulation aging state by taking the degree of polymerization as an indirect basis; and the Raman spectrum diagnosis of the oil paper insulation aging is realized by establishing a transformer oil paper insulation aging Raman spectrum diagnosis model by taking the spectrum vector as input and the oil paper insulation aging state as output. Therefore, the performance state of the transformer oil paper insulation is effectively monitored, and the operation safety of the power system is guaranteed.
Description
Technical Field
The invention belongs to the field of insulation online monitoring and fault diagnosis of electrical equipment, and particularly relates to a Raman spectrum diagnosis method for insulation aging of transformer oil paper.
Background
The transformer occupies a high proportion in the power equipment, and the aging problem is a key factor related to the safe operation of a power grid. The rapid development of the power grid puts high requirements on the evaluation of the aging state of the transformer. The service life of power transformers is generally associated with the deterioration of the insulation material. After the transformer oil paper insulation system is used for years, the transformer oil paper insulation system is aged under the action of thermal stress and electric stress, and the insulation performance of the transformer is influenced. The insulating oil and the insulating paper are decomposed to generate substances such as carbon monoxide, carbon dioxide, furfural, methanol, acetone and the like which reflect the fault property and the aging degree, and the substances are dissolved in the oil. The insulating oil contains abundant aging information, so that the method has important significance for detecting the insulating oil. In order to evaluate the aging state of oil-immersed power equipment, test results such as the furfural content in oil, dissolved gas in oil, the degree of polymerization of insulating paper, and the like are often used. However, these methods are difficult to sample due to the complex steps, and are often difficult to use for in-situ rapid aging status assessment.
Raman spectroscopy is a scattering spectrum based on the raman scattering effect found in c.v. raman. The scattering spectra of different frequencies of incident light are analyzed to obtain information of molecular vibration and rotation, and the method is applied to the research of molecular structures. Due to the fact that different substances are different in structure, property and content, different Raman signals can be generated under the irradiation of laser, and therefore the analysis of the material characteristics is achieved. The Raman spectrum technology is widely applied to the fields of petrifaction, biological environmental protection, medicine, food safety and the like, and also draws wide attention in the field of oil paper insulation aging diagnosis. Different substances can generate different Raman signals when irradiated by laser due to different structures, properties and contents, so that the analysis of the characteristics of the substances is realized. In the aging process of the oil paper insulation, the substances in the oil are various and continuously generate a series of complex chemical changes, and the Raman spectrum technology can just reflect the complex process.
Disclosure of Invention
In order to solve the problems in the prior art, the invention discloses a Raman spectrum diagnosis method for aging of oil paper insulation of a transformer, which is characterized in that the aging diagnosis of the oil paper insulation is realized based on a Raman spectrum technology, the accuracy is high, and the method is a novel method at present.
In order to achieve the above object, the present invention specifically adopts the following technical solutions.
A Raman spectrum diagnosis method for insulation aging of transformer oil paper is characterized by comprising the following steps: and predicting the aging degree of the oil paper insulation sample by utilizing the improved T-S fuzzy neural network according to the Raman spectrum data of the insulation oil.
The Raman spectrum vector of the insulating oil is connected with the input layer of the neural network, the aging state of the oil paper insulation is used as the output of the neural network, the experimental data is used for training a diagnosis model, membership functions are continuously modified, the internal mathematical relationship between the oil paper insulation Raman spectrum characteristic and the oil paper insulation aging state is mined, and the purpose of predicting the oil paper insulation aging degree by using the insulating oil Raman spectrum data is achieved.
A Raman spectrum diagnosis method for insulation aging of transformer oil paper is characterized by comprising the following steps:
step 1: simulating the aging state of a real transformer to obtain an aging insulating oil sample;
and (3) placing the real transformer in a sealed system, and simulating the aging state of the real transformer at 120 ℃ to obtain an aged insulating oil sample.
Accelerated aging oil samples with aging time of 0, 1,5,10,20,35,50 and 80 days are respectively obtained, and 40 training samples and 15 testing samples are obtained in total. Monitoring the insulation aging of the oil paper, and in order to guarantee the safe and stable operation of equipment, research focuses on more concerned aging degree. The distribution of a total of 55 samples at 0, 1,5,10,20,35,50 and 80 days is therefore: 5, 6, 11, 6. The distribution of training samples at 0, 1,5,10,20,35,50 and 80 days is: 5, 5. Test samples 15, aged 10 days 1, aged 20 days 1, aged 35 days 6, aged 50 days 6, aged 80 days 1.
And (3) for the division of the aging stage of the training sample, taking the average polymerization degree of the insulating paper in each aging time stage as a judgment basis, and calculating the aging degree of the oil paper insulating sample according to the definition so as to initially establish a training sample library.
The aging state of the oiled paper insulation sample is defined as follows:
the polymerization Degree (DP) of the insulating paper is a recognized index of the insulation aging of the oil paper, but the polymerization degree cannot be detected in an actually-operated transformer. Therefore, a plurality of oil paper insulation aging diagnosis researches are focused on finding out a substitute detection method of polymerization degree, and the invention also aims to establish indirect correlation between the polymerization degree and the oil paper insulation aging Raman spectrum data so as to achieve the diagnosis effect. And (3) detecting the polymerization degree of the insulating paper in the oil paper insulating and aging sample according to national standard detection standards by dividing the oil paper insulating sample in the aging stage, and defining the aging state of the oil paper insulating sample by taking the polymerization degree of the insulating paper in each aging time stage as a judgment basis. Typically, the degree of polymerization of the new sample is between 1200 and 1600. Thus, under this definition, the samples in good insulation state have a degradation value between 0 and 1. It should be noted that when the DP value is less than 400, the sample has been severely aged and should be noted. At this time, the aging degree value is between 3 and 4. In summary, the aging degree is defined as a range between 0 and 4, and the larger the value, the deeper the aging degree is.
Step 2: dividing the aged insulating oil sample obtained in the step 1 into a training sample and a test sample;
training a sample: test sample 8: 3, in general case 9: 1,8: 2,7: 3 are useful.
And step 3: constructing an oil paper insulation aging Raman spectrum diagnosis model;
the oil paper insulation aging Raman spectrum diagnosis model is defined by adopting an if-then rule form and is divided into 4 layers, namely an input layer, a fuzzy rule calculation layer and an output layer;
wherein the fuzzy layer adopts a membership function
Fuzzifying the input value to obtain a fuzzy membership value mu; mu is a fuzzy membership value; x is the number ofjIs the jth input variable;
a fuzzy set of a fuzzy system corresponding to the ith input variable;
the center and the width of the membership function of the jth input variable of the ith sample are respectively;
the fuzzy rule calculation layer calculates the fuzzy operator omega by adopting the following fuzzy multiplication formula:
ωiis a fuzzy operator, mu is a fuzzy membership value;
for the ith input variable pairA fuzzy set of a corresponding fuzzy system, wherein k is the number of input parameters; n is the number of fuzzy subsets.
The output layer is calculated using the following formula:
wherein,
to blur the system parameters, yiFor the output obtained according to fuzzy rules, ωiFor the fuzzy operator, n is the fuzzy subset number k is the maximum number of subscript j in fuzzy system parameters, and k is the input parameter number.
And 4, step 4: the training sample obtained in the step 2 is used for testing and training the oil paper insulation aging Raman spectrum diagnosis model constructed in the step 3, and the oil paper insulation aging Raman spectrum diagnosis model is obtained after the test sample is tested and trained;
taking the Raman spectrogram vector of a training sample as input, taking the aging state of the training sample as output, and training the oil paper insulation aging Raman spectrum diagnosis model, wherein the training method specifically comprises the following steps:
the error between the desired output and the actual output is calculated as follows:
in the formula, ydExpected output of the oil paper insulation aging Raman spectrum diagnosis model; y iscThe actual output of the oil paper insulation aging Raman spectrum diagnosis model; e is the error of the desired output and the actual output;
correcting the fuzzy system parameters according to the calculated errors:
in the formula,
is a fuzzy system parameter; alpha is the network learning rate; x is the number ofjIs the jth input variable; omegaiIs the product of the membership degree of the input parameters;
modifying the center and width of the membership function:
in the formula,
respectively the center and width of the membership function.
And 5: collecting transformer insulating oil to be diagnosed in operation, acquiring a Raman spectrogram spectral vector of the insulating oil, inputting the Raman spectrogram spectral vector as input sample data into the oil paper insulation aging Raman spectral diagnosis model obtained in the step 4 after training and testing;
step 6: and 5, judging the aging degree of the transformer insulating oil according to the output result of the oil paper insulating aging Raman spectrum diagnosis model in the step 5.
The aging state of the oiled paper insulation sample is defined as follows:
the polymerization Degree (DP) of the insulating paper is a recognized index of the insulation aging of the oil paper, but the polymerization degree cannot be detected in an actually-operated transformer. Therefore, a plurality of oil paper insulation aging diagnosis researches are focused on finding out a substitute detection method of polymerization degree, and the patent also aims to establish indirect correlation between the polymerization degree and the oil paper insulation aging Raman spectrum data so as to achieve the diagnosis effect. And (3) detecting the polymerization degree of the insulating paper in the oil paper insulating and aging sample according to national standard detection standards by dividing the oil paper insulating sample in the aging stage, and defining the aging state of the oil paper insulating sample by taking the polymerization degree of the insulating paper in each aging time stage as a judgment basis. Typically, the degree of polymerization of the new sample is between 1200 and 1600. Thus, under this definition, the samples in good insulation state have a degradation value between 0 and 1. It should be noted that when the DP value is less than 400, the sample has been severely aged and should be noted. At this time, the aging degree value is between 3 and 4. In summary, the aging degree is defined as a range between 0 and 4, and the larger the value, the deeper the aging degree is.
Compared with the prior art, the invention has the following beneficial technical effects:
the oiled paper insulating material can generate various characteristic quantities reflecting the aging state, such as furfural, methanol, acetone, CO2 and the like, in the electric or thermal aging process, and is dissolved in oil. At present, the aging diagnosis of the running transformer is mainly based on detection and analysis of the aging characteristic quantities, laboratory judgment is carried out on the corresponding guide rule threshold values, and the problems that a single characteristic quantity needs different equipment for analysis, cannot be used for effective field diagnosis and the like exist. The laser raman technology has the advantage of non-contact nondestructive analysis in the fields of material composition analysis and state diagnosis, however, researchers have few researches on raman data analysis of transformer oil at present. The invention relates to a Raman spectrum diagnosis method for aging of oil paper insulation of a transformer based on a T-S fuzzy neural network, which can realize non-contact lossless aging diagnosis of the oil paper insulation based on a Raman spectrum technology and mainly solves the technical problems that the aging state of the oil paper insulation of the transformer lacks a field charged monitoring means and the detection process is complicated. A new idea is opened up for diagnosing the insulation aging state of the transformer oil paper.
Drawings
FIG. 1 is a Raman spectrum of insulating oil;
FIG. 2 is a schematic flow chart of a Raman spectrum diagnosis method for the aging of the oil paper insulation of the transformer according to the present invention;
FIG. 3 is a graph showing the results of aging diagnosis of 10 samples with deeper aging according to the embodiment of the present invention;
FIG. 4 is a graph showing the results of aging diagnosis of 5 samples with different aging degrees according to the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. The embodiments described herein are only some embodiments of the invention, and not all embodiments. All other embodiments obtained by a person skilled in the art without any inventive step based on the spirit of the present invention are within the scope of the present invention.
The dimension of the training sample for explaining the construction of the fuzzy neural network in fig. 2 determines the number of input/output nodes of the fuzzy neural network, and in the patent, 1023-dimensional spectral data is input and 1-dimensional aging state is output. During the training of the fuzzy neural network, the center and the width of the fuzzy membership function are initialized randomly, and a diagnosis model is trained by iterating for 100 times by using error calculation and a coefficient/parameter correction method along with the training. And finally, inputting the test data into a trained diagnostic model to verify the accuracy of the prediction result.
The invention discloses a Raman spectrum diagnosis method for the insulation aging of transformer oil paper, which comprises the following steps as shown in figure 2:
firstly, simulating the aging state of a real transformer to obtain an aging insulating oil sample;
and (3) placing the real transformer in a sealed system, and simulating the aging state of the real transformer at 120 ℃ to obtain an aged insulating oil sample.
The invention defines the insulation aging state of the oiled paper by taking the polymerization degree as an indirect basis.
The aging state of the oiled paper insulation sample is defined as follows:
the polymerization Degree (DP) of the insulating paper is a recognized index of the insulation aging of the oil paper, but the polymerization degree cannot be detected in an actually-operated transformer. Therefore, a plurality of oil paper insulation aging diagnosis researches are focused on finding out a substitute detection method of polymerization degree, and the patent also aims to establish indirect correlation between the polymerization degree and the oil paper insulation aging Raman spectrum data so as to achieve the diagnosis effect. And (3) detecting the polymerization degree of the insulating paper in the oil paper insulating and aging sample according to national standard detection standards by dividing the oil paper insulating sample in the aging stage, and defining the aging state of the oil paper insulating sample by taking the polymerization degree of the insulating paper in each aging time stage as a judgment basis. Typically, the degree of polymerization of the new sample is between 1200 and 1600. Thus, under this definition, the samples in good insulation state have a degradation value between 0 and 1. It should be noted that when the DP value is less than 400, the sample has been severely aged and should be noted. At this time, the aging degree value is between 3 and 4. In summary, the aging degree is defined as a range between 0 and 4, and the larger the value, the deeper the aging degree is.
And carrying out Raman spectrum detection on the aged insulating oil to obtain a Raman spectrogram. Insulating oil is a very complex mixture containing many substances, which also results in a complex raman spectrum. Each substance theoretically has a corresponding Raman signal, and as the oil-paper insulation system ages, the substance composition and content of the insulation oil change and the Raman spectrum changes. Therefore, the aging degree of the oiled paper insulation can be judged through the difference of the Raman spectrums of the insulating oil. The raman spectrum of the insulating oil consists of a series of spectral points, as shown in fig. 1. The Raman spectrum of the insulating oil can be expressed as a two-dimensional vector { (U)1,v1),(U2,v2),…,(Un,vn) In the abscissa of the unit Raman frequency shift (cm)-1) Ordinate is Raman intensity (a.u.) when Raman detection instrument parameters are kept unchanged, the sampling interval of the spectrum points in the Raman spectrum is unchanged, namely U of each samplingiKeeping the same; therefore, the Raman spectrum of the insulating oil can be simplified into a one-dimensional vector v1,v2,…,vnAs input variables for the neural network.
Step 2: dividing the aged insulating oil sample obtained in the step 1 into a training sample and a test sample;
training a sample: test sample 8: 3, in general case 9: 1,8: 2,7: 3 are useful.
And step 3: constructing an oil paper insulation aging Raman spectrum diagnosis model based on a fuzzy neural network;
the oil paper insulation aging Raman spectrum diagnosis model takes the Raman spectrogram vector of the insulation oil as input, and takes the aging stage of the oil paper insulation sample as output to construct the diagnosis model.
The model can not only be automatically updated, but also continuously modify membership functions of fuzzy subsets, which is defined by the following form of 'if-then' rule, where the rule is RiIn the case of (2), fuzzy inference is as follows:
wherein,
a fuzzy set which is a fuzzy system;
is a fuzzy system parameter; y isiFor an output derived from a fuzzy rule, the input part (i.e., if part) is fuzzy and the output part (i.e., then part) is deterministic, the fuzzy inference representing the output as a linear combination of the inputs.
Let x be [ x ] for the input quantity x1,x2,…,xk]First, each input variable x is calculated according to a fuzzy rulejDegree of membership of:
where μ is fuzzy slaveryA genus value; x is the number ofjIs an input variable;
respectively the center and the width of the membership function; k is the number of input parameters; n is the number of fuzzy subsets.
And carrying out fuzzy calculation on each membership degree, and adopting a fuzzy operator as a continuous multiplication operator.
Mu is a fuzzy membership value;
a fuzzy set of a fuzzy system is obtained, and k is the number of input parameters; n is the number of fuzzy subsets.
Calculating the output value y of the fuzzy model according to the fuzzy calculation resulti。
The oil paper insulation aging Raman spectrum diagnosis model is divided into 4 layers of an input layer, a fuzzy layer and a fuzzy rule calculation and output layer. Input layer and input vector xiAnd connecting, wherein the number of nodes is the same as the dimension of the input vector. The fuzzy layer uses membership function
And fuzzifying the input value to obtain a fuzzy membership value mu. The fuzzy rule calculation layer calculates by adopting a fuzzy continuous multiplication formula to obtain omega:
the output layer adopts the following formula to calculate the output of the oil paper insulation aging Raman spectrum diagnosis model:
and 4, step 4: the training sample obtained in the step 2 is used for testing and training the oil paper insulation aging Raman spectrum diagnosis model constructed in the step 3, and the oil paper insulation aging Raman spectrum diagnosis model is obtained after the test sample is tested and trained;
taking the Raman spectrogram vector of a training sample as input, taking the aging state of the training sample as output, and training the oil paper insulation aging Raman spectrum diagnosis model, wherein the training method specifically comprises the following steps:
and (3) error calculation:
in the formula, ydExpected output of the oil paper insulation aging Raman spectrum diagnosis model; y iscThe actual output of the oil paper insulation aging Raman spectrum diagnosis model; e is the error between the desired output and the actual output.
And (3) coefficient correction:
in the formula,
is a fuzzy system parameter; alpha is the network learning rate; x is the number ofjInputting parameters for the network; omegatTo be transportedAnd the membership degree of the input parameter is a product of the continuous degree of membership.
Parameter correction:
in the formula,
respectively the center and width of the membership function.
And 5: collecting transformer insulating oil to be diagnosed in operation, acquiring a Raman spectrogram spectral vector of the insulating oil, inputting the Raman spectrogram spectral vector as input sample data into the oil paper insulation aging Raman spectral diagnosis model obtained in the step 4 after training and testing;
step 6: and 5, judging the aging degree of the transformer insulating oil according to the output result of the oil paper insulating aging Raman spectrum diagnosis model in the step 5.
The aging state of the oiled paper insulation sample is defined as follows:
the polymerization Degree (DP) of the insulating paper is a recognized index of the insulation aging of the oil paper, but the polymerization degree cannot be detected in an actually-operated transformer. Therefore, a plurality of oil paper insulation aging diagnosis researches are focused on finding out a substitute detection method of polymerization degree, and the patent also aims to establish indirect correlation between the polymerization degree and the oil paper insulation aging Raman spectrum data so as to achieve the diagnosis effect. And (3) detecting the polymerization degree of the insulating paper in the oil paper insulating and aging sample according to national standard detection standards by dividing the oil paper insulating sample in the aging stage, and defining the aging state of the oil paper insulating sample by taking the polymerization degree of the insulating paper in each aging time stage as a judgment basis. Typically, the degree of polymerization of the new sample is between 1200 and 1600. Thus, under this definition, the samples in good insulation state have a degradation value between 0 and 1. It should be noted that when the DP value is less than 400, the sample has been severely aged and should be noted. At this time, the aging degree value is between 3 and 4. In summary, the aging degree is defined as a range between 0 and 4, and the larger the value, the deeper the aging degree is.
In conclusion, the process of constructing the oil paper insulation aging Raman spectrum diagnosis model by using the T-S fuzzy neural network is shown in FIG. 2. The number of input/output nodes of the fuzzy neural network is determined by the dimensionality of a constructed training sample of the fuzzy neural network, 1023-dimensional spectral data is input in the patent, and 1-dimensional aging state is output. During the training of the fuzzy neural network, the center and the width of the fuzzy membership function are initialized randomly, and a diagnosis model is trained by iterating for 100 times by using error calculation and a coefficient/parameter correction method along with the training. And finally, inputting the test data into a trained diagnostic model to verify the accuracy of the prediction result.
The technical scheme of the present invention is further explained by the following examples
The invention provides a method for monitoring the aging state of the oil paper insulation by detecting the Raman signal of the transformer oil. The mechanism is as follows: when substance molecules are irradiated by laser with a certain frequency, scattering occurs, only most of light is scattered by changing the direction, and the frequency of the light is the same as that of exciting light, and the light is elastic scattering, namely Rayleigh scattering; meanwhile, a small part of light not only changes the propagation direction of the light, but also changes the frequency of scattered light, and belongs to inelastic scattering, namely Raman scattering. The frequency difference between the scattered light and the incident light becomes a raman shift, which depends on the change of the vibrational energy level of the molecule, and the molecular vibration characterized by different chemical bonds or genes can be used as the basis for judging the molecular structure. With the aging of the oil paper insulation, the aging characteristics of the oil and the paper are dissolved into the oil, and the aging state monitoring of the oil paper insulation of the transformer is realized through the Raman detection of the oil.
The technical scheme for realizing the invention is as follows: the invention simulates the aging state of a real transformer under the condition of 120 ℃ according to the IEEE guide rule in a sealing system by an accelerated thermal aging method to obtain an aged insulating oil sample. Accelerated aging oil samples with aging time of 0, 1,5,10,20,35,50 and 80 days are respectively obtained, and 40 training samples and 15 testing samples are obtained in total. And (3) dividing the aging stage of the training sample, detecting the polymerization Degree (DP) of the insulating paper in the oil paper insulating aging sample according to the national standard detection standard, and calculating the aging degree of the oil paper insulating sample according to the definition by taking the average polymerization degree of the insulating paper in each aging time stage as a judgment basis so as to initially establish a training sample library.
The oil paper insulation aging Raman spectrum diagnosis model based on the T-S fuzzy neural network is trained by using 40 training samples, the Raman spectrogram vector of the sample is used as input, the aging stage of the sample is used as output, and the construction is started by using a structure with 12 hidden nodes and 100 iterations. (sample Nos. 1 to 5 for 35 days of aging and 6 to 10 for 50 days of aging)
Monitoring the insulation aging of the oil paper, and researching the diagnosis accuracy rate when the aging degree is deep in order to ensure the safe and stable operation of equipment. Therefore, 10 samples with aging state values of about 3 were selected for the intensive monitoring, and the diagnosis results are shown in fig. 3. (sample number 1 for 10 days, sample number 2 for 20 days, sample number 3 for 35 days, sample number 4 for 50 days, and sample number 5 for 80 days)
The 5 test samples with different aging states are selected for the overall test, and the diagnosis result is shown in fig. 4.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (8)
1. A Raman spectrum diagnosis method for insulation aging of transformer oil paper is characterized by comprising the following steps:
step 1: simulating the aging state of a real transformer to obtain an aging insulating oil sample;
step 2: dividing the aged insulating oil sample obtained in the step 1 into a training sample and a test sample;
and step 3: constructing an oil paper insulation aging Raman spectrum diagnosis model;
and 4, step 4: training the oil paper insulation aging Raman spectrum diagnosis model constructed in the step 3 by using the training sample obtained in the step 2, and testing and training the test sample to obtain the oil paper insulation aging Raman spectrum diagnosis model;
and 5: collecting transformer insulating oil to be diagnosed in operation, acquiring a Raman spectrogram spectral vector of the insulating oil, inputting the Raman spectrogram spectral vector as input sample data into the oil paper insulation aging Raman spectral diagnosis model obtained in the step 4 after training and testing;
step 6: and 5, judging the aging degree of the transformer insulating oil according to the output result of the oil paper insulating aging Raman spectrum diagnosis model in the step 5.
2. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper according to claim 1, wherein the Raman spectrum diagnosis method comprises the following steps:
in step 1, a real transformer is placed in a sealed system, and the aging state of the real transformer is simulated at 120 ℃ to obtain an aged insulating oil sample.
3. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper according to claim 1 or 2, wherein the Raman spectrum diagnosis method comprises the following steps:
in step 1, accelerated aging oil samples were taken for aging times of 0, 1,5,10,20,35,50 and 80 days, respectively.
4. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper as claimed in claim 3, wherein:
in step 3, the oil paper insulation aging Raman spectrum diagnosis model is defined by adopting an if-then rule form and comprises an input layer, a fuzzy rule calculation layer and an output layer;
the input layer is the Raman spectrogram vector of the oiled paper insulation, and the output layer is the oiled paper insulation aging degree.
5. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper as claimed in claim 4, wherein the Raman spectrum diagnosis method comprises the following steps:
in the oil paper insulation aging Raman spectrum diagnosis model, the fuzzy layer adopts a membership function
Fuzzifying the input value to obtain a fuzzy membership value mu;
wherein mu is a fuzzy membership value; x is the number ofjIs the jth input variable;
a fuzzy set of a fuzzy system corresponding to the ith input variable;
the center and width of the membership function of the jth input variable of the ith sample, respectively.
6. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper as claimed in claim 5, wherein the Raman spectrum diagnosis method comprises the following steps:
the fuzzy rule calculation layer adopts the following fuzzy multiplication formula to calculate and obtain a fuzzy operator omegai：
ωiIs a fuzzy operator, mu is a fuzzy membership value;
for the ith input variable correspondingK is the number of input parameters; n is the number of fuzzy subsets.
7. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper as claimed in claim 6, wherein:
the output layer is calculated using the following formula:
wherein,
to blur the system parameters, yiFor the output obtained according to fuzzy rules, ωiFor the fuzzy operator, n is the fuzzy subset number, and k is the maximum number of subscript j in the fuzzy system parameter, i.e. the input parameter number.
8. The Raman spectrum diagnosis method for the insulation aging of the transformer oil paper as claimed in claim 7, wherein:
in step 4, training the oil paper insulation aging raman spectrum diagnosis model by taking the raman spectrogram vector of the training sample as input and the aging state of the training sample as output, and specifically comprises the following steps:
the error between the desired output and the actual output is calculated as follows:
in the formula, ydExpected output of the oil paper insulation aging Raman spectrum diagnosis model; y iscThe actual output of the oil paper insulation aging Raman spectrum diagnosis model; e is the error of the desired output and the actual output;
correcting the fuzzy system parameters, the centers and the widths of the membership functions according to the calculated errors, and then iterating further until a set iteration number is reached or the error e is smaller than a set threshold:
in the formula,
is a fuzzy system parameter; alpha is the network learning rate; x is the number ofjIs the jth input variable;
in the formula,
respectively the center and width of the membership function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011117423.3A CN112485609B (en) | 2020-10-19 | 2020-10-19 | Raman spectrum diagnosis method for insulation aging of transformer oil paper |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011117423.3A CN112485609B (en) | 2020-10-19 | 2020-10-19 | Raman spectrum diagnosis method for insulation aging of transformer oil paper |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112485609A CN112485609A (en) | 2021-03-12 |
| CN112485609B true CN112485609B (en) | 2021-11-23 |
Family
ID=74926604
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011117423.3A Active CN112485609B (en) | 2020-10-19 | 2020-10-19 | Raman spectrum diagnosis method for insulation aging of transformer oil paper |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112485609B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113987909B (en) * | 2021-09-18 | 2023-05-02 | 广东电网有限责任公司广州供电局 | Oilpaper insulation aging prediction method, device, computer equipment and storage medium |
| CN113899997A (en) * | 2021-11-10 | 2022-01-07 | 哈尔滨理工大学 | Transformer insulation state diagnosis method based on improved support vector machine |
| CN114924169A (en) * | 2022-05-09 | 2022-08-19 | 重庆大学 | Oil paper insulation aging diagnosis method based on random forest and thermalization tank |
| CN115267461A (en) * | 2022-09-13 | 2022-11-01 | 重庆邮电大学 | Oil paper insulation aging diagnosis method based on dimension expansion Raman spectrum and machine vision |
| CN117250456B (en) * | 2023-11-20 | 2024-01-30 | 山东海鲲数控设备有限公司 | Transformer insulation state monitoring system |
| CN118209537B (en) * | 2024-03-25 | 2024-09-13 | 哈尔滨理工大学 | Transformer oil paper insulation aging diagnosis method based on fusion spectrum and neural network |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408580A (en) * | 2008-11-21 | 2009-04-15 | 重庆大学 | Method for evaluating oil paper insulation ageing state based on local discharge characteristic parameter |
| CN103106544A (en) * | 2013-02-01 | 2013-05-15 | 东南大学 | Photovoltaic power generation prediction system based on T-S-type fuzzy neural network |
| CN106885978A (en) * | 2017-04-20 | 2017-06-23 | 重庆大学 | A kind of paper oil insulation Diagnosis of Aging based on insulating oil Raman spectrum wavelet-packet energy entropy |
| CN109406898A (en) * | 2018-11-07 | 2019-03-01 | 福州大学 | A method of fusion multi-characteristicquantity quantity comprehensive assessment paper oil insulation degree of aging |
| CN109842210A (en) * | 2019-02-22 | 2019-06-04 | 广东科源电气有限公司 | A kind of monitoring system and method for for transformer |
| CN110031443A (en) * | 2019-04-15 | 2019-07-19 | 重庆大学 | A kind of portable oil paper insulation ageing state Raman spectrum diagnostic device and method |
| CN110889228A (en) * | 2019-11-28 | 2020-03-17 | 国网吉林省电力有限公司电力科学研究院 | Transformer oil paper insulation aging prediction method based on chicken swarm optimization BP neural network |
| CN111537845A (en) * | 2020-04-26 | 2020-08-14 | 云南电网有限责任公司电力科学研究院 | Method for identifying aging state of oil paper insulation equipment based on Raman spectrum cluster analysis |
-
2020
- 2020-10-19 CN CN202011117423.3A patent/CN112485609B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408580A (en) * | 2008-11-21 | 2009-04-15 | 重庆大学 | Method for evaluating oil paper insulation ageing state based on local discharge characteristic parameter |
| CN103106544A (en) * | 2013-02-01 | 2013-05-15 | 东南大学 | Photovoltaic power generation prediction system based on T-S-type fuzzy neural network |
| CN106885978A (en) * | 2017-04-20 | 2017-06-23 | 重庆大学 | A kind of paper oil insulation Diagnosis of Aging based on insulating oil Raman spectrum wavelet-packet energy entropy |
| CN109406898A (en) * | 2018-11-07 | 2019-03-01 | 福州大学 | A method of fusion multi-characteristicquantity quantity comprehensive assessment paper oil insulation degree of aging |
| CN109842210A (en) * | 2019-02-22 | 2019-06-04 | 广东科源电气有限公司 | A kind of monitoring system and method for for transformer |
| CN110031443A (en) * | 2019-04-15 | 2019-07-19 | 重庆大学 | A kind of portable oil paper insulation ageing state Raman spectrum diagnostic device and method |
| CN110889228A (en) * | 2019-11-28 | 2020-03-17 | 国网吉林省电力有限公司电力科学研究院 | Transformer oil paper insulation aging prediction method based on chicken swarm optimization BP neural network |
| CN111537845A (en) * | 2020-04-26 | 2020-08-14 | 云南电网有限责任公司电力科学研究院 | Method for identifying aging state of oil paper insulation equipment based on Raman spectrum cluster analysis |
Non-Patent Citations (4)
| Title |
|---|
| "Analysis of Furfural Dissolved in Transformer Oil Based on Confocal Laser Raman Spectroscopy";Weigen Chen等;《IEEE Transactions on Dielectrics and Electrical Insulation》;20160430;第23卷(第2期);第915-921页 * |
| "Detection of Methanol Dissolved in Transformer Oil By Laser Raman Spectroscopy";Zhaoliang Gu等;《IEEE Xplore》;20161231;第1-4页 * |
| "变压器油中气体拉曼光谱检测及信号处理方法";万福 等;《仪器仪表学报》;20161130;第37卷(第11期);第2482-2488页 * |
| "变压器油中水分在线监测的神经网络计算模型";陈伟根 等;《高电压技术》;20070531;第33卷(第5期);第73-78页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112485609A (en) | 2021-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112485609B (en) | Raman spectrum diagnosis method for insulation aging of transformer oil paper | |
| CN111638028B (en) | High-voltage parallel reactor mechanical state evaluation method based on vibration characteristics | |
| Liu et al. | Moisture diagnosis of transformer oil-immersed insulation with intelligent technique and frequency-domain spectroscopy | |
| Bakar et al. | A new method to detect dissolved gases in transformer oil using NIR-IR spectroscopy | |
| CN108051364A (en) | A kind of EPR nuclear energy cable residue lifetime estimation method and prediction EPR nuclear energy cable remaining life methods | |
| CN104020401A (en) | Cloud-model-theory-based method for evaluating insulation thermal ageing states of transformer | |
| CN108959769A (en) | A kind of state evaluating method and device of insulating oil | |
| Li et al. | Optimal symbolic entropy: An adaptive feature extraction algorithm for condition monitoring of bearings | |
| Mahrukh et al. | Prediction of power transformer oil chromatography based on LSTM and RF model | |
| CN113553756A (en) | Method and system for evaluating and simulating insulation state of oil paper containing air bubbles | |
| Sezavar et al. | Risk assessment of contaminated composite insulators in pre-flashover conditions | |
| Sutikno et al. | Integration of duval pentagon to the multi-method interpretation to improve the accuracy of dissolved gas analysis technique | |
| Hussein et al. | Faults diagnosis and assessment of transformer insulation oil quality: intelligent methods based on dissolved gas analysis a-review | |
| Chen et al. | Improved Interpretation of Impulse Frequency Response Analysis for Synchronous Machine Using Life long Learning Based on iCaRL | |
| CN115468946A (en) | Transformer oil aging diagnosis method and device and storage medium | |
| Sarma et al. | A long short-term memory based prediction model for transformer fault diagnosis using dissolved gas analysis with digital twin technology | |
| Jasim et al. | Dissolved gas analysis of power transformers | |
| Yang et al. | Prediction of aging degree of oil-paper insulation based on Raman spectroscopy and fuzzy neural network | |
| Duy | Apply the Artificial Neural Network to Diagnose Potential Fault of Power Transformer Based on Dissolved Gas-in-oil Analysis Data | |
| Modi et al. | Intelligent approach to interpret incipient faults of power transformer from DGA database | |
| Rhamadhan et al. | Estimating the DP Value of the Paper Insulation of Oil-Filled Power Transformers Using an ANFIS Algorithm | |
| Wei et al. | Evaluation method of spindle performance degradation based on VMD and random forests | |
| Yang et al. | Multi frequency ultrasonic detection of water content in transformer oil with GA-BPNN | |
| CN118378037B (en) | Bridge acceleration data denoising method based on long-short-term memory depth neural network | |
| CN111948282B (en) | Method for detecting physical and chemical properties of transformer oil |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |