CN115468946A - Transformer oil aging diagnosis method and device and storage medium - Google Patents
Transformer oil aging diagnosis method and device and storage medium Download PDFInfo
- Publication number
- CN115468946A CN115468946A CN202211247588.1A CN202211247588A CN115468946A CN 115468946 A CN115468946 A CN 115468946A CN 202211247588 A CN202211247588 A CN 202211247588A CN 115468946 A CN115468946 A CN 115468946A
- Authority
- CN
- China
- Prior art keywords
- transformer oil
- aging
- raman spectrum
- sample
- detected
- 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.)
- Pending
Links
- 230000032683 aging Effects 0.000 title claims abstract description 154
- 238000003745 diagnosis Methods 0.000 title claims abstract description 78
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000001237 Raman spectrum Methods 0.000 claims abstract description 87
- 238000006116 polymerization reaction Methods 0.000 claims abstract description 48
- 230000015654 memory Effects 0.000 claims abstract description 42
- 238000003062 neural network model Methods 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000012360 testing method Methods 0.000 claims description 28
- 238000005070 sampling Methods 0.000 claims description 14
- 230000000737 periodic effect Effects 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 5
- 238000003878 thermal aging Methods 0.000 claims description 5
- 230000003203 everyday effect Effects 0.000 claims description 4
- 230000005284 excitation Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000001228 spectrum Methods 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 4
- 238000002405 diagnostic procedure Methods 0.000 claims 4
- 238000009413 insulation Methods 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 6
- 230000007787 long-term memory Effects 0.000 description 5
- 230000006403 short-term memory Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 230000003647 oxidation Effects 0.000 description 4
- 238000007254 oxidation reaction Methods 0.000 description 4
- 239000000047 product Substances 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- WRIDQFICGBMAFQ-UHFFFAOYSA-N (E)-8-Octadecenoic acid Natural products CCCCCCCCCC=CCCCCCCC(O)=O WRIDQFICGBMAFQ-UHFFFAOYSA-N 0.000 description 1
- LQJBNNIYVWPHFW-UHFFFAOYSA-N 20:1omega9c fatty acid Natural products CCCCCCCCCCC=CCCCCCCCC(O)=O LQJBNNIYVWPHFW-UHFFFAOYSA-N 0.000 description 1
- QSBYPNXLFMSGKH-UHFFFAOYSA-N 9-Heptadecensaeure Natural products CCCCCCCC=CCCCCCCCC(O)=O QSBYPNXLFMSGKH-UHFFFAOYSA-N 0.000 description 1
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 1
- ZQPPMHVWECSIRJ-UHFFFAOYSA-N Oleic acid Natural products CCCCCCCCC=CCCCCCCCC(O)=O ZQPPMHVWECSIRJ-UHFFFAOYSA-N 0.000 description 1
- 239000005642 Oleic acid Substances 0.000 description 1
- 238000001069 Raman spectroscopy Methods 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 150000001299 aldehydes Chemical class 0.000 description 1
- 150000001335 aliphatic alkanes Chemical class 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 229920002678 cellulose Polymers 0.000 description 1
- 239000001913 cellulose Substances 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 239000007857 degradation product Substances 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 229930195733 hydrocarbon Natural products 0.000 description 1
- 150000002430 hydrocarbons Chemical class 0.000 description 1
- 229910052500 inorganic mineral Inorganic materials 0.000 description 1
- 239000012774 insulation material Substances 0.000 description 1
- QXJSBBXBKPUZAA-UHFFFAOYSA-N isooleic acid Natural products CCCCCCCC=CCCCCCCCCC(O)=O QXJSBBXBKPUZAA-UHFFFAOYSA-N 0.000 description 1
- 150000002576 ketones Chemical class 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 239000011707 mineral Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- ZQPPMHVWECSIRJ-KTKRTIGZSA-N oleic acid Chemical compound CCCCCCCC\C=C/CCCCCCCC(O)=O ZQPPMHVWECSIRJ-KTKRTIGZSA-N 0.000 description 1
- 239000003209 petroleum derivative Substances 0.000 description 1
- 150000003254 radicals Chemical class 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/26—Oils; Viscous liquids; Paints; Inks
- G01N33/28—Oils, i.e. hydrocarbon liquids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Chemical Kinetics & Catalysis (AREA)
- General Chemical & Material Sciences (AREA)
- Oil, Petroleum & Natural Gas (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
- Housings And Mounting Of Transformers (AREA)
Abstract
The invention discloses a transformer oil aging diagnosis method, a device and a storage medium, wherein the method comprises the following steps: obtaining an aging sample of transformer oil; detecting the initial Raman spectrum of the aged sample and the polymerization degree of the aged sample, and denoising the initial Raman spectrum to obtain a preprocessed Raman spectrum; inputting the preprocessed Raman spectrum of the aging sample into a long-time and short-time memory neural network model as input quantity, and training the neural network model by taking the polymerization degree of the aging sample as the output quantity of the long-time and short-time memory neural network model to obtain a transformer oil aging diagnosis model; the method comprises the steps of obtaining transformer oil to be detected, detecting a Raman spectrum of the transformer oil to be detected, inputting the Raman spectrum of the transformer oil to be detected into a transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree. The method can effectively improve the accuracy of diagnosis of the aging state of the transformer oil.
Description
Technical Field
The invention relates to the technical field of power equipment monitoring, in particular to a method and a device for diagnosing oil insulation aging of a transformer and a storage medium.
Background
The power transformer is the most critical and expensive equipment in power transmission and transformation equipment, and the operation reliability of the power transformer is directly related to the economic operation and the safety and stability of a power grid. Transformer accidents are mostly caused by the deterioration of the insulation properties thereof, and the transformer oil, which is a major component of the transformer insulation material, has a crucial influence on the life of the transformer. The transformer oil is mineral insulating oil refined from petroleum products, and mainly comprises alkanes, cyclanes and aromatics, mainly hydrocarbons. Transformer oil is a main liquid insulating medium used for insulation, cooling, and arc extinguishing in electric power equipment, and is required to have excellent insulating properties because it is used as a liquid insulating medium. However, with the increase of the operation time, the transformer oil is oxidized and aged to generate degradation products such as free radicals, alcohol, aldehyde, ketone and the like, and the insulation performance of the transformer oil is seriously affected. For a transformer with an average service life of about three or forty years, the oxidation and aging phenomena of transformer oil can bring serious potential safety hazards to the safe and stable operation of the transformer.
The existing transformer oil aging diagnosis method is generally an insulating oleic acid value measurement method and an oil surface tension test method, wherein the insulating oil acid value measurement and the oil surface tension test require that a sample of insulating oil taken from a transformer is sent to a laboratory for physical and chemical property analysis during the test, and because an oil sample is further oxidized and aged when exposed to the air and is easily polluted in the transportation process, the accurate diagnosis of the polymerization degree of the transformer is difficult.
Disclosure of Invention
The invention provides a transformer oil insulation aging diagnosis method, a transformer oil insulation aging diagnosis device and a storage medium, and aims to solve the technical problems that in the existing transformer oil aging diagnosis method, insulation oil is required to be taken from a transformer and a sample is sent to a laboratory for physical and chemical property analysis during testing, and the polymerization degree of the transformer is difficult to accurately diagnose due to the fact that an oil sample is further oxidized and aged when exposed to air and is easily polluted in the transportation process.
One embodiment of the present invention provides a transformer oil aging diagnosis method, including:
obtaining an aging sample of the transformer oil;
detecting an initial Raman spectrum of the aged sample and the polymerization degree of the aged sample, and denoising the initial Raman spectrum to obtain a preprocessed Raman spectrum;
inputting the preprocessed Raman spectrum of the aged sample into a long-time and short-time memory neural network model as an input quantity, and training the neural network model by taking the polymerization degree of the aged sample as an output quantity of the long-time and short-time memory neural network model to obtain a transformer oil aging diagnosis model;
acquiring transformer oil to be detected, detecting a Raman spectrum of the transformer oil to be detected, inputting the Raman spectrum of the transformer oil to be detected into the transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree.
Further, the obtaining an aged sample of the transformer oil comprises:
the transformer oil is subjected to accelerated thermal aging tests under the conditions of different oil-paper ratios, and periodic sampling is performed in the aging tests to obtain aging samples of the transformer oil at different sampling times.
Further, the denoising processing on the initial raman spectrum to obtain a preprocessed raman spectrum includes:
identifying peaks of the initial Raman spectrum based on a derivative spectrum, and removing peaks of the initial Raman spectrum based on a cubic curve peak removal method;
and denoising the initial Raman spectrum after the peak is removed by adopting a smooth denoising method of three-point circular fast Fourier transform median to obtain a preprocessed Raman spectrum.
Further, the long-time and short-time memory neural network model comprises an input gate, a forgetting gate, an output gate and a memory unit;
an input gate: i all right angle t =σ(W t ·[h t-1 ,x t ]+b i );
Forget the door: f. of t =σ(W f ·[h t-1 ,x t ]+b f );
wherein f is t ,i t ,O t Forgetting information, input information after memory, output information at time t, W f ,W i ,W C ,W o Are weights respectively; b is a mixture of f ,b i ,b C ,b o Are all offset vectors, C t is Memory state at time t, h t For the output at time t, σ is the excitation function,is a Hadamard product.
Further, the different ratios of the oilpaper include 10.
Further, the periodic sampling in the aging test comprises:
the aging test duration was set to 30 days, and the transformer oil was sampled once every day.
An embodiment of the present invention provides a transformer oil aging diagnosis apparatus including:
the aging sample acquisition module is used for acquiring an aging sample of the transformer oil;
the aging sample detection module is used for detecting the initial Raman spectrum of the aging sample and the polymerization degree of the aging sample, and carrying out denoising treatment on the initial Raman spectrum to obtain a preprocessed Raman spectrum;
the diagnosis model training module is used for inputting the preprocessed Raman spectrum of the aged sample into the long-time and short-time memory neural network model as an input quantity, and training the neural network model by taking the polymerization degree of the aged sample as an output quantity of the long-time and short-time memory neural network model to obtain a transformer oil aging diagnosis model;
the aging diagnosis module is used for acquiring transformer oil to be detected, detecting the Raman spectrum of the transformer oil to be detected, inputting the Raman spectrum of the transformer oil to be detected into the transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree.
The embodiment of the invention provides a computer-readable storage medium, which comprises a stored computer program, wherein when the computer program runs, a device where the computer-readable storage medium is located is controlled to execute the transformer oil aging diagnosis method.
According to the embodiment of the invention, the transformer oil aging diagnosis model is constructed based on the long-time and short-time memory neural network, and the polymerization degree of the transformer oil can be diagnosed by establishing the mapping relation between the polymerization degree and the characteristic quantity of the Raman spectrum, so that the accuracy of diagnosis of the transformer oil aging state can be effectively improved; according to the embodiment of the invention, the aging state diagnosis of the transformer oil can be realized without contacting an aging sample, the problem that the transformer oil is not accurately diagnosed due to oxidation of the transformer oil when the transformer oil is exposed in the air is avoided, and the diagnosis efficiency can be effectively improved; according to the embodiment of the invention, the transformer oil aging diagnosis model is constructed based on the long-time memory neural network, the input parameters of modeling and the training time of the model can be greatly reduced, the efficiency of transformer oil diagnosis can be effectively improved, the diagnosis capability of aging samples with different oil-paper ratios can be effectively improved, and thus the aging state of the transformer oil in the aging stage with different oil-paper ratios can be comprehensively diagnosed.
Furthermore, the peak removing processing and the smoothing denoising processing are carried out on the initial Raman spectrum, so that the accuracy and the reliability of the Raman spectrum can be improved, and the accuracy of transformer oil aging diagnosis can be further improved.
Drawings
FIG. 1 is a schematic flow chart of a transformer oil aging diagnosis method provided by an embodiment of the invention;
FIG. 2 is a schematic diagram of the degree of polymerization of transformer oil at different oil-to-paper ratios according to an embodiment of the present invention;
FIG. 3 is another schematic flow chart of a transformer oil aging diagnosis method provided by the embodiment of the invention;
FIG. 4 is a schematic diagram of a long-term and short-term memory neural network model according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of an aging stage of transformer oil according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a transformer oil aging diagnosis apparatus provided in an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.
In the description of the present application, it is to be understood that the terms "first", "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present application, the meaning of "a plurality" is two or more unless otherwise specified.
In the description of the present application, it should be noted that, unless otherwise explicitly stated or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be, for example, a fixed connection, a detachable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meaning of the above terms in the present application can be understood in a specific case by those of ordinary skill in the art.
Referring to fig. 1, an embodiment of the present invention provides a transformer oil aging diagnosis method, including:
s1, obtaining an aging sample of transformer oil;
in the embodiment of the invention, the aging sample of the transformer oil can be obtained through an aging test, and the embodiment of the invention performs a 130-degree accelerated thermal aging test under the conditions of different oil-paper ratios and obtains the aging sample of the transformer oil by sampling periodically.
The oil paper material in the embodiment of the invention is 25# transformer oil-cellulose insulating paper, and different oil paper proportions can be set according to requirements, for example, the oil paper proportions are 10.
S2, detecting the initial Raman spectrum of the aged sample and the polymerization degree of the aged sample, and denoising the initial Raman spectrum to obtain a preprocessed Raman spectrum;
in the embodiment of the invention, 16 samples are sampled at each oil paper ratio to measure the initial Raman spectrum of the oil paper, the oil paper is preprocessed, and the detected Raman spectrum and the polymerization degree of an aging sample construct a Raman spectrum database for diagnosing the aging state of the transformer oil according to a transformer oil aging diagnosis model in the follow-up process. One embodiment of constructing the raman spectrum database is as follows: and arranging all the Raman spectrum data by adopting Matlab software, wherein each Raman spectrum data corresponds to a polymerization degree, and organizing all the Raman spectrum data and the corresponding polymerization degrees to construct a Raman spectrum database.
Referring to fig. 2, in the embodiment of the present invention, the aging state of the Degree of Polymerization (DP) includes good edge, early stage of aging, middle stage of aging, and final stage of aging, the degree of polymerization of the oiled paper insulation aging sample is detected according to the national standard detection standard, and the oiled paper insulation aging sample is divided into four aging states, including good insulation I: DP is more than or equal to 500; and (2) early stage of aging II: (DP more than or equal to 250 is less than 500); and (3) aging middle stage III: (DP is more than or equal to 150 and less than 250); end of aging IV (DP < 150).
According to the embodiment of the invention, the initial Raman spectrum is subjected to pretreatment such as denoising, so that the accurate pretreated Raman spectrum can be obtained, and the accuracy and reliability of subsequent transformer oil aging diagnosis can be effectively improved.
S3, inputting the preprocessed Raman spectrum of the aged sample into a long-time memory neural network model as an input quantity, and training the neural network model by taking the polymerization degree of the aged sample as an output quantity of the long-time memory (LSTM) neural network model to obtain a transformer oil aging diagnosis model;
in the embodiment of the invention, the mapping relation between the polymerization degree and the characteristic quantity of the Raman spectrum can be established based on the long-term and short-term memory neural network, so that the diagnosis result of the transformer oil aging degree can be accurately improved, and the accuracy can reach more than 90%.
S4, obtaining the transformer oil to be detected, detecting the Raman spectrum of the transformer oil to be detected, inputting the Raman spectrum of the transformer oil to be detected into the transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree.
According to the embodiment of the invention, the transformer oil aging diagnosis model is constructed based on the long-time and short-time memory neural network, and the polymerization degree of the transformer oil can be diagnosed by establishing the mapping relation between the polymerization degree and the characteristic quantity of the Raman spectrum, so that the accuracy of diagnosis of the transformer oil aging state can be effectively improved; according to the embodiment of the invention, the aging state diagnosis of the transformer oil can be realized without contacting an aging sample, the problem that the transformer oil is not accurately diagnosed due to oxidation of the transformer oil when the transformer oil is exposed in the air is avoided, and the diagnosis efficiency can be effectively improved; according to the embodiment of the invention, the transformer oil aging diagnosis model is constructed based on the long-time memory neural network, the input parameters of modeling and the training time of the model can be greatly reduced, the efficiency of transformer oil diagnosis can be effectively improved, the diagnosis capability of aging samples with different oil-paper ratios can be effectively improved, and thus the aging state of the transformer oil in the aging stage with different oil-paper ratios can be comprehensively diagnosed.
The embodiment of the invention is suitable for the aging state evaluation and fault diagnosis of the oil-immersed transformer, can effectively guide the operation and maintenance work of power equipment, is beneficial to ensuring the stable operation of a power grid, and can be widely applied to the fields of power energy, power systems and the like.
Referring to fig. 2, another flow chart of a transformer oil aging diagnosis method according to an embodiment of the present invention is shown.
In one embodiment, obtaining an aged sample of transformer oil comprises:
the transformer oil is subjected to accelerated thermal aging tests under the conditions of different oil-paper ratios, and periodic sampling is performed in the aging tests to obtain aging samples of the transformer oil at different sampling times.
In the present example, the different oilpaper ratios include 10. Periodic sampling was performed in the aging test, including: the aging test duration is set to be 30 days, the transformer oil is sampled once every day, and all 720 prepared transformer oil samples (the number of samples in each oil paper ratio is 240) in a laboratory are subjected to Raman data acquisition. It should be noted that the aging duration and the sampling interval can be adjusted according to actual needs.
In one embodiment, denoising the initial raman spectrum to obtain a preprocessed raman spectrum includes:
identifying a peak of the initial Raman spectrum based on the derivative spectrum, and removing the peak of the initial Raman spectrum based on a peak removal method of the cubic curve;
and denoising the initial Raman spectrum after the peak is removed by adopting a smooth denoising method of a three-point circular fast Fourier transform median to obtain a preprocessed Raman spectrum.
In the embodiment of the invention, the peak of the initial Raman spectrum is identified and removed, so that the influence of the peak on the aging state diagnosis can be effectively reduced, the accuracy of the transformer oil aging state diagnosis can be effectively improved, and the noise removal processing is carried out on the initial Raman spectrum by a smooth noise removal method, so that the accuracy of the transformer oil aging state diagnosis can be further improved.
Referring to fig. 4, the long-term and short-term memory neural network model includes an input gate, a forgetting gate, an output gate, and a memory unit;
an input gate: i.e. i t =σ(W t ·[h t-1 ,x t ]+b i );
Forget the door: f. of t =σ(W f ·[h t-1 ,x t ]+b f );
wherein f is t ,i t ,O t Respectively, forgetting information, input information after memory, output information at time t, W f ,W i ,W C ,W o Are weights respectively; b f ,b i ,b C ,b o Are all offset vectors, C t is Memory state at time t, h t For the output at time t, σ is the excitation function,is a Hadamard product.
In the embodiment of the invention, 240 samples of each oil-paper ratio type in the Raman spectrum database are randomly divided into 165 training samples in a training set and 75 testing samples in a testing set, wherein the training set is used for establishing a relationship model of characteristic quantity and polymerization degree of a Raman spectrum, and the testing set is used for verifying the relationship model. The framework of the long-time memory neural network adopts an input-2 LSTM layers-1 full-connection layer-1 Softmax layer-output structure, the preprocessed Raman spectrum of the aging sample is input into the long-time memory neural network model as input quantity, the degree of polymerization of the aging sample is used as output quantity of the long-time memory neural network model to train the neural network model, and the transformer oil aging diagnosis model is obtained.
Referring to tables 1-3, the degree of polymerization is obtained by detecting the transformer aging diagnosis model, and the aging state of the transformer oil can be accurately judged.
Table 1 test sample aging state diagnosis result with oilpaper ratio of 10
Table 2 results of diagnosing the aging state of test samples with an oil-to-paper ratio of 15
Table 3 results of diagnosing the aging state of test samples with a 20 oilpaper ratio
Referring to fig. 5, taking a transformer oil sample with a paper-oil ratio of 10. The aging degree of the sample is obtained through the aging diagnosis output of the transformer oil, the polymerization degree of the transformer oil is 798, the transformer oil is in a good insulation state, and the transformer oil and the polymerization degree 835 obtained through actual testing are in the same aging state.
The transformer oil aging diagnosis model is obtained based on the long-time memory neural network model training, long-term rules can be learned based on the long-time memory neural network model, and the time sequence is classified, processed and predicted through experience learning, so that an accurate transformer oil aging diagnosis result is obtained.
The embodiment of the invention has the following beneficial effects:
according to the embodiment of the invention, the transformer oil aging diagnosis model is constructed based on the long-time memory neural network, and the polymerization degree of the transformer oil can be diagnosed by establishing the mapping relation between the polymerization degree and the characteristic quantity of the Raman spectrum, so that the accuracy of diagnosis of the aging state of the transformer oil can be effectively improved; according to the embodiment of the invention, the aging state diagnosis of the transformer oil can be realized without contacting an aging sample, the problem that the transformer oil is not accurately diagnosed due to oxidation of the transformer oil when the transformer oil is exposed in the air is avoided, and the diagnosis efficiency can be effectively improved; according to the embodiment of the invention, the transformer oil aging diagnosis model is constructed based on the long-time memory neural network, the input parameters of modeling and the training time of the model can be greatly reduced, the efficiency of transformer oil diagnosis can be effectively improved, the diagnosis capability of aging samples with different oil-paper ratios can be effectively improved, and thus the aging state of the transformer oil in the aging stage with different oil-paper ratios can be comprehensively diagnosed.
Furthermore, the embodiment of the invention can improve the accuracy and reliability of the Raman spectrum by carrying out peak removing processing and smooth denoising processing on the initial Raman spectrum, thereby further improving the accuracy of transformer oil aging diagnosis.
An embodiment of the present invention provides a transformer oil aging diagnosis apparatus including:
the aging sample obtaining module 10 is used for obtaining an aging sample of the transformer oil;
the aging sample detection module 20 is used for detecting the initial raman spectrum of the aging sample and the polymerization degree of the aging sample, and performing denoising treatment on the initial raman spectrum to obtain a preprocessed raman spectrum;
the diagnosis model training module 30 is used for inputting the preprocessed raman spectrum of the aging sample into the long-term and short-term memory neural network model as input quantity, and training the neural network model by taking the polymerization degree of the aging sample as the output quantity of the long-term and short-term memory neural network model to obtain a transformer oil aging diagnosis model;
and the aging diagnosis module 40 is used for acquiring the transformer oil to be detected, detecting the raman spectrum of the transformer oil to be detected, inputting the raman spectrum of the transformer oil to be detected into the transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree.
In one embodiment, the aged sample acquisition module 10 is further configured to:
the transformer oil is subjected to accelerated thermal aging tests under the conditions of different oil-paper ratios, and periodic sampling is performed in the aging tests to obtain aging samples of the transformer oil at different sampling times.
In one embodiment, the aged sample detection module 20 is further configured to:
identifying a peak of the initial Raman spectrum based on the derivative spectrum, and removing the peak of the initial Raman spectrum based on a peak removal method of a cubic curve;
and denoising the initial Raman spectrum after the peak is removed by adopting a smooth denoising method of a three-point circular fast Fourier transform median to obtain a preprocessed Raman spectrum.
In one embodiment, the long-time memory neural network model comprises an input gate, a forgetting gate, an output gate and a memory unit;
an input gate: i all right angle t =σ(W t ·[h t- 1,x t ]+b i );
Forgetting the door: f. of t =σ(W f ·[h t- 1,x t ]+b f );
wherein f is t ,i t ,O t Forgetting information, input information after memory, output information at time t, W f ,W i ,W C ,W o Are weights respectively; b f ,b i ,b C ,b o Are all offset vectors, C t is Memory state at time t, h t For the output at time t, σ is the excitation function,is a Hadamard product.
In one embodiment, the different oilpaper ratios include 10.
In one embodiment, the periodic sampling is performed in a burn-in test comprising:
the aging test duration was set to 30 days, and the transformer oil was sampled once every day.
The embodiment of the invention provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program runs, an apparatus where the computer-readable storage medium is located is controlled to execute the transformer oil aging diagnosis method.
The foregoing is a preferred embodiment of the present invention, and it should be noted that it would be apparent to those skilled in the art that various modifications and enhancements can be made without departing from the principles of the invention, and such modifications and enhancements are also considered to be within the scope of the invention.
Claims (8)
1. A transformer oil aging diagnosis method is characterized by comprising the following steps:
obtaining an aging sample of transformer oil;
detecting the initial Raman spectrum of the aged sample and the polymerization degree of the aged sample, and carrying out denoising treatment on the initial Raman spectrum to obtain a preprocessed Raman spectrum;
inputting the preprocessed Raman spectrum of the aged sample into a long-time and short-time memory neural network model as an input quantity, and training the neural network model by taking the polymerization degree of the aged sample as an output quantity of the long-time and short-time memory neural network model to obtain a transformer oil aging diagnosis model;
the method comprises the steps of obtaining transformer oil to be detected, detecting a Raman spectrum of the transformer oil to be detected, inputting the Raman spectrum of the transformer oil to be detected into a transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree.
2. The transformer oil aging diagnostic method of claim 1, wherein the obtaining an aging sample of transformer oil comprises:
the transformer oil is subjected to accelerated thermal aging tests under the conditions of different oil-paper ratios, and periodic sampling is performed in the aging tests to obtain aging samples of the transformer oil at different sampling times.
3. The transformer oil aging diagnostic method of claim 1, wherein the denoising the initial raman spectrum to obtain a preprocessed raman spectrum comprises:
identifying peaks of the initial raman spectrum based on a derivative spectrum, and removing peaks of the initial raman spectrum based on a peak removal method of a cubic curve;
and denoising the initial Raman spectrum after the peak is removed by adopting a smooth denoising method of a three-point circular fast Fourier transform median to obtain a preprocessed Raman spectrum.
4. The transformer oil aging diagnosis method according to claim 1, wherein the long-time and short-time memory neural network model comprises an input gate, a forgetting gate, an output gate and a memory unit;
an input gate: i.e. i t =σ(W t ·[h t-1 ,x t ]+b i );
Forget the door: f. of t =σ(W f ·[h t-1 ,x t ]+b f );
wherein, f t ,i t ,O t Respectively, forgetting information, input information after memory, output information at time t, W f ,W i ,W C ,W o Are respectively weight; b f ,b i ,b C ,b o Are all offset vectors, C t is Memory state at time t, h t For the output at time t, σ is the excitation function,is a Hadamard product.
5. The transformer oil aging diagnostic method according to claim 2, wherein the different paper ratios include 10.
6. The transformer oil aging diagnostic method of claim 2, wherein the periodically sampling in the aging test comprises:
the aging test duration is set to be 30 days, and the transformer oil is sampled once every day.
7. A transformer oil aging diagnostic apparatus, characterized by comprising:
the aging sample acquisition module is used for acquiring an aging sample of the transformer oil;
the aging sample detection module is used for detecting the initial Raman spectrum of the aging sample and the polymerization degree of the aging sample, and denoising the initial Raman spectrum to obtain a preprocessed Raman spectrum;
the diagnosis model training module is used for inputting the preprocessed Raman spectrum of the aged sample into the long-time and short-time memory neural network model as an input quantity, and training the neural network model by taking the polymerization degree of the aged sample as an output quantity of the long-time and short-time memory neural network model to obtain a transformer oil aging diagnosis model;
the aging diagnosis module is used for acquiring the transformer oil to be detected, detecting the Raman spectrum of the transformer oil to be detected, inputting the Raman spectrum of the transformer oil to be detected into the transformer oil aging diagnosis model to obtain the polymerization degree of the transformer oil to be detected, and judging the aging state of the transformer oil to be detected according to the polymerization degree.
8. A computer-readable storage medium, comprising a stored computer program, wherein when the computer program runs, the computer-readable storage medium controls a device to execute the transformer oil aging diagnosis method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211247588.1A CN115468946A (en) | 2022-10-12 | 2022-10-12 | Transformer oil aging diagnosis method and device and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211247588.1A CN115468946A (en) | 2022-10-12 | 2022-10-12 | Transformer oil aging diagnosis method and device and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115468946A true CN115468946A (en) | 2022-12-13 |
Family
ID=84337053
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211247588.1A Pending CN115468946A (en) | 2022-10-12 | 2022-10-12 | Transformer oil aging diagnosis method and device and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115468946A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117909652A (en) * | 2023-12-29 | 2024-04-19 | 广东电网有限责任公司江门供电局 | High-voltage circuit breaker fault diagnosis data processing method |
-
2022
- 2022-10-12 CN CN202211247588.1A patent/CN115468946A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117909652A (en) * | 2023-12-29 | 2024-04-19 | 广东电网有限责任公司江门供电局 | High-voltage circuit breaker fault diagnosis data processing method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109116193B (en) | Electrical equipment dangerous discharge distinguishing method based on partial discharge signal comprehensive entropy | |
EP3513155B1 (en) | Method and device for determining and/or monitoring the state of a transformer oil | |
US20190041450A1 (en) | An intelligent assessment method of main insulation condition of transformer oil paper insulation | |
CN106841949B (en) | Method and device for monitoring stator insulation of three-phase asynchronous alternating current motor on line | |
CN103389430A (en) | Oil-immersed type transformer fault detection method based on Bayesian discrimination theory | |
CN104914364A (en) | Capacitance oilpaper transformer sleeve insulation state assessment method | |
Azis et al. | Operational condition assessment of in-service distribution transformers | |
CN112557834B (en) | Aging diagnosis method for oiled paper insulation equipment based on Raman spectrum | |
CN115468946A (en) | Transformer oil aging diagnosis method and device and storage medium | |
CN112179852A (en) | Method and device for predicting residual life of composite insulator | |
Bjelić et al. | Fault detection in a power transformer based on reverberation time | |
CN112485609B (en) | Raman spectrum diagnosis method for insulation aging of transformer oil paper | |
Khalilisenobari et al. | A novel numerical index for assessing results of frequency response analysis (FRA): An experimental study on electrical machines | |
Schwarz et al. | Diagnostic methods for transformers | |
CN113670987B (en) | Method, device, equipment and storage medium for identifying oil paper insulation aging state | |
CN111308286B (en) | Method for estimating service life of traction transformer under multi-parameter empowerment combination | |
JPH03179245A (en) | Diagnosis of aging of electric wire and cable | |
Yuan et al. | Lab investigation of PD development in transformer winding | |
CN112881839A (en) | Transformer diagnosis method based on mutual information of frequency concentration and vibration stability | |
CN116297300B (en) | Composite insulator core rod aging evaluation method, device and storage medium | |
Sekatane et al. | Residual life estimation of power transformer based on Karl Fischer and Adaptive neuro-fuzzy interference system | |
CN114113850B (en) | Crosslinked polyethylene cable insulation life prediction method and system based on consumption dynamics model | |
JP2667327B2 (en) | OF cable insulation deterioration diagnosis method and apparatus | |
Cui et al. | Multi-variable approach for evaluating transformer paper insulation | |
Li et al. | [Retracted] Aging Law and Life Evaluation Model of Cable Insulation Based on WOA‐SVR |
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 |