JP4888402B2 - Remaining life diagnosis method for oil-filled transformers - Google Patents

Description

  The present invention relates to a method for diagnosing the remaining life of an oil-filled transformer that estimates the average degree of polymerization of insulating paper and the remaining amount of metal container in an oil-filled transformer to diagnose the remaining life of the oil-filled transformer. .

In general, the following materials are used for oil-filled transformers.
(1) Conductive materials such as copper and aluminum (2) Insulating materials such as insulating oil, insulating paper and press board (3) Iron core materials such as silicon steel strips (4) Metal container materials such as iron and stainless steel Among the materials, it is considered that insulating materials such as insulating oil and insulating paper are subject to deterioration over time in the oil-filled transformer. Insulating oil has a function of oil deterioration prevention device (open type, air-sealed type, nitrogen-sealed type, etc.), so the deterioration is very slow, and the breakdown voltage, which is an important characteristic, is low. .

On the other hand, for insulating paper, the degree of decrease in dielectric breakdown voltage due to deterioration over time is small, but the degree of decrease in mechanical strength is large (that is, the paper becomes shabby). As the deterioration of the insulating paper proceeds, the insulating paper is cracked or broken by mechanical stress due to inrush current or electromagnetic force generated at the time of external short circuit, and the risk of dielectric breakdown increases.
Therefore, the life of the oil-filled transformer is strongly influenced by the mechanical strength of the insulating paper, particularly the deterioration state of the winding conductor insulating paper. In other words, the remaining life of the oil-filled transformer is conventionally considered to be determined by the dielectric breakdown of the winding conductor insulating paper, that is, the state of deterioration of the insulating paper (decrease in average polymerization degree).

Below, the average degree of polymerization of the insulating paper, the remaining life of the oil-filled transformer, and the deterioration diagnosis method will be discussed.
(1) Average degree of polymerization of insulating paper Insulating paper is a polymer obtained by polymerizing a large number of cellulose molecules, and the number of basic molecules constituting this cellulose is called the degree of polymerization. The average degree of polymerization in the case of new kraft paper as insulating paper is about 1000. As for this average polymerization degree, when the insulating paper is oxidized and deteriorated, the chain of cellulose molecules is cut and the molecular weight of the cellulose molecules is lowered, that is, the average polymerization degree is lowered. For example, in a transformer used for 30 years, the average polymerization degree of insulating paper is said to decrease to about 40 to 60% of the initial value (polymerization degree 400 to 600).
(2) Life of oil-filled transformer The Japan Electrical Manufacturers' Association standard JEM1463-1993 defines an evaluation standard for oil-filled transformers exceeding 1000 [kVA]. Generally, according to this standard, the average degree of polymerization is The point at which it is supposed to be 450 is defined as the life of the oil-filled transformer.

(3) Current remaining life diagnosis method for oil-filled transformers In the remaining life diagnosis of transformers, it is necessary to measure or estimate the average degree of polymerization of coil insulating paper if the standard of (2) is followed. Become. However, it is difficult to measure the coil insulation paper of the oil-filled transformer in operation because it cannot be easily collected. Therefore, the average polymerization degree of the insulating material (press board, lead insulating paper) that can be collected inside the transformer, and the amount of furfural and CO ₂ + CO, which are products of the decomposition process of the insulating paper, were measured, and the results were used. Remaining life diagnosis is performed. These remaining life diagnosis methods are described in Non-Patent Document 1 and Non-Patent Document 2, for example.

Hereinafter, various remaining life diagnosis methods will be outlined.
(A) Degree of polymerization method During inspections such as when the operation is stopped, a method of diagnosing the degree of insulation paper deterioration by collecting the press board or lead insulation paper from the inside of the transformer that does not affect insulation is used. The law. This degree of polymerization method is a method for estimating the remaining life by estimating the degree of deterioration of the coil insulating paper at the hottest portion (hot spot portion) of the winding coil from the average degree of polymerization of the collected insulating paper.
(B) CO ₂ + CO method Insulating paper generates various organic components such as water, CO ₂ , and CO due to deterioration. As effective deterioration index components, there are CO ₂ + CO having a correlation with the average degree of polymerization, and further furfural. Among these, in the CO ₂ + CO method, an in-oil gas analysis is performed, and an average polymerization degree is estimated from the amount of CO ₂ + CO, which is a final degradation product of insulating paper, and a remaining life diagnosis method is performed.

(C) Furfural method The furfural aldehyde component is produced during the decomposition process of cellulose. Even when the insulating oil is deaerated, 85% of the furfural remains in the oil and does not diffuse into the gas. For this reason, if the history of the deaeration process is known, this method can be used even if the deaeration process is performed.
In this furfural method, the average degree of polymerization is obtained from the measured amount of furfural according to the correlation obtained in advance. However, the degree of deterioration (remaining life diagnosis) of the average degree of polymerization is considerably wide with respect to the amount of furfural. The diagnosis results also have a wide range, and it is very difficult to estimate the remaining life with high accuracy.

Also, by measuring the temperature of the transformer oil predicts CO 2 ₊ CO concentration, the oil-filled electrical apparatus to detect the insulation deterioration when the difference between the actual value and the predicted value is a certain value or more An insulation diagnostic apparatus is described in Patent Document 1.
In addition, in the static induction machine abnormality diagnosis method that analyzes the insulation medium of static induction equipment and detects abnormalities such as local overheating and arc discharge from changes in the type, generation amount, and generation ratio of decomposition products, the decomposition products Patent Document 2 describes a method of diagnosing abnormalities in a static induction device using a neural network in which the amount of acetylene produced is input data and an abnormal phenomenon such as arc discharge is learned as teacher data.
“Current Status and Trends of Aging Transformer Reliability Maintenance Technology”, Aging Transformer Reliability Maintenance Technology Investigation Technical Committee, Institute of Electrical Engineers of Japan, March 10, 2003, No. 922, p. 22-27 "Part IV Oil-immersed Transformer Degradation Diagnosis", Electric Cooperative Research, Electric Cooperative Research Society, February 25, 1999, Vol. 54, No. 5 (Part 1), p. 158-168 JP-A-63-52071 (page 2, lower left column, line 14 to page 3, upper right column, line 13, line 5, etc.) JP-A-6-82405 (paragraphs [0025] to [0036], FIG. 1, FIG. 2, etc.)

  However, when an oil-filled transformer is installed in a steel factory or the like, the temperature environment such as the atmospheric temperature and the chemical environment such as ozone and sulfide contained in the atmosphere vary greatly depending on the installation location. Become a thing. These differences in temperature environment and chemical environment have a greater influence on the progress of metal container thinning than on the deterioration of insulating paper. That is, in an oil-filled transformer, the wall thickness of the partition wall of the metal container decreases with time (thinning) due to corrosion or the like, but the metal container's thickness reduction rate depends on the temperature and chemical environment. Will change drastically.

  For this reason, if the progress of thinning in the metal container is faster than the progress of deterioration of the insulating paper, the life of the oil-filled transformer is not limited by the deterioration of the insulating paper, and in many cases, the amount of thinning of the metal container It will be rate-controlled by (remaining thickness). On the contrary, if the deterioration of the insulation paper is quicker than the progress of thinning in the metal container, the life of the oil-filled transformer is not limited by the thinning amount of the metal container (remaining thickness). In many cases, the rate is limited by the deterioration of the insulating paper.

Therefore, when an oil-filled transformer is installed in a facility where the temperature and chemical environment vary greatly depending on the installation location of a steel factory, etc., the polymerization degree method, the CO ₂ + CO method or the furfural method described above. Even if the average degree of polymerization of the insulating paper is estimated by this and the remaining life of the oil-filled transformer is diagnosed based only on this estimated value, the estimated remaining life is significantly different from the actual life of the oil-filled transformer. There is a risk of it.
An object of the present invention is to provide an oil that can accurately diagnose the remaining life of an oil-filled transformer even when the oil-filled transformer is installed in a facility whose temperature and chemical environment vary greatly depending on the installation location. It is to provide a method for diagnosing the remaining life of an input transformer.

  In order to solve the above-described problem, a method for diagnosing the remaining life of an oil-filled transformer according to claim 1 of the present invention includes a metal container storing insulating oil and a coil immersed in the insulating oil in the metal container. A method for diagnosing the remaining life of an oil-filled transformer, wherein the measured values of two or more deterioration index components in the insulating oil and the measured value of the thinning amount in the metal container are input factor groups, respectively, By identifying or learning the model using the average degree of polymerization and the remaining life of the metal container estimated from the remaining amount of the metal container as an output factor, the average degree of polymerization estimation model and the remaining life estimation model of the metal container are obtained. The measured value of the deterioration index component of the oil-filled transformer to be diagnosed is input to the average polymerization degree estimation model, and the measured value of the thinning amount is input to the remaining life estimation model. Process the combination of the average polymerization degree estimated value obtained by the average polymerization degree estimation model and the remaining life estimation model and the remaining life estimation value of the metal container, and estimate the final remaining life of the oil-filled transformer It is characterized by that.

An oil-filled transformer remaining life diagnosis method according to claim 2 of the present invention is the oil-filled transformer remaining life diagnosis method according to claim 1, wherein the oil-immersed transformer remaining life diagnosis method uses an insulating oil as an input factor for the average polymerization degree estimation model. The measured value of furfural and the measured value of CO ₂ + CO amount in are used.
According to a third aspect of the present invention, there is provided a method for diagnosing the remaining life of an oil-filled transformer according to claim 1 or 2, wherein an insulation factor is used as an input factor for the average polymerization degree estimation model. It is characterized by additionally using a measured value of the amount of acetylene in oil.

An oil-filled transformer remaining life diagnosis method according to claim 4 of the present invention is the oil-filled transformer remaining life diagnosis method according to any one of claims 1 to 3, wherein the average polymerization degree estimation model and As the remaining amount estimation model, a neural network is used.
An oil-filled transformer remaining life diagnosis method according to claim 5 of the present invention is the oil-filled transformer remaining life diagnosis method according to any one of claims 1 to 3, wherein the average polymerization degree estimation model and A multiple regression equation is used as the remaining amount estimation model.

  According to the method for diagnosing the remaining life of an oil-filled transformer of the present invention configured as described above, the oil-filled transformer is installed in a facility whose temperature and chemical environment greatly change depending on the installation location. Even in this case, the remaining life of the oil-filled transformer can be diagnosed with high accuracy.

Hereinafter, a remaining life diagnosis method for an oil-filled transformer according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 shows an example of an oil-filled transformer to which the remaining life diagnosis method according to the embodiment of the present invention can be applied, and FIG. 2 shows a remaining life diagnosis of the oil-filled transformer according to the embodiment of the present invention. The method is shown as a flowchart.
As shown in FIG. 1, the oil-filled transformer 20 is arranged in a substantially cylindrical metal container 22 in which the insulating oil OL is stored, and is immersed in the insulating oil OL. A coil 24 and a tap changer 26 are provided. A circular inspection window 30 is opened in the top plate portion 23 of the metal container 22, and the inspection window 30 is closed by a lid 28. Here, the metal container 22 and the lid 28 are formed of a metal material such as stainless steel or carbon steel, and surface treatment such as rust prevention treatment is performed on the surface portion as necessary. Further, the lid 28 is formed with a bottomed cylindrical immersion portion 29, which is inserted into the metal container 22 through the inspection window 30, and at least a part on the lower end side is always in the insulating oil OL. Soaked.

A primary side bushing 32 and a secondary side bushing 34 are fixed to the outer peripheral side of the inspection window 30 on the top plate portion 23 of the metal container 22. The bushings 32 and 34 are each formed in a cylindrical shape by an insulating material such as ceramic, and are fixed in a state where the lower end side penetrates the top plate portion 23. Terminal members 36 and 38 made of a conductive metal penetrate the inner peripheral portions of the bushings 32 and 34. The primary side terminal member 36 is connected to an input terminal (not shown) of the coil 24 through the connection cable 40 and the tap changer 26 in the metal container 22. Further, the secondary side terminal member 38 is connected to an output terminal (not shown) of the coil 24 via a connection cable 42 in the middle of the metal container 22.
Here, the coil 24 includes an iron core made of a silicon steel strip or the like and insulating paper for insulating between layers of the iron core (not shown). A tap board (not shown) made of insulating paper is disposed in the tap changer 26, and this tap board insulates a plurality of taps.

  In the metal container 22, a buffer tank 44 storing unused insulating oil OL is disposed on the upper side of the top plate portion 23, and this buffer tank 44 communicates with the inside of the metal container 22 through an oil supply pipe 46. In the oil-filled transformer 20, when the insulating oil OL in the metal container 22 decreases for some reason, the insulating oil OL equal to the decrease amount is replenished from the buffer tank 44 into the metal container 22. Further, a sampling valve 50 is connected to the peripheral wall portion of the metal container 22 through an oil collection pipe 48 on the lower end side thereof. In the oil-filled transformer 20, the insulating oil OL in the metal container 22 is removed by opening the sampling valve 50. Can be collected.

Next, the remaining life diagnosis method according to this embodiment for the oil-filled transformer 20 configured as described above will be described.
The remaining life diagnosis method according to the present embodiment includes a past actual measurement data input step (S1), a degree of polymerization (average degree of polymerization) estimation model construction step (S2) regarding the amount of furfural, CO ₂ + CO, and acetylene in the insulating oil OL. ), Diagnosis target input data input step (S 3), average polymerization degree estimation step (S 4), past actual measurement data input step (S 11) relating to the amount of thinning of the metal container 22, and remaining life estimation model construction step of the metal container 22 (S12), a diagnosis target input data input step (S3), a metal container remaining life estimation step (S14), and an ensemble processing step (S15).

Hereinafter, the contents of the above steps will be described sequentially.
(1) Past actual measurement data input step regarding the amount of furfural, the amount of CO ₂ + CO, and the amount of acetylene in the insulating oil OL (S1)
In order to construct an average polymerization degree estimation model of insulating paper of the oil-filled transformer 20 by a neural network, actual measurement data of the oil-filled transformer 20 measured in the past is input as learning data of the estimation model. These measured data are the amounts of various deterioration index components (in this embodiment, the amount of furfural, the amount of CO ₂ + CO, and the amount of acetylene) obtained by analyzing the insulating oil OL, and the average degree of polymerization of the insulating paper. As various deterioration index components obtained by analyzing the insulating oil OL, one or more known index components such as a water content, an oxygen content, a hydrogen content, etc. may be appropriately added in addition to the above components. it can.

(2) Average polymerization degree estimation model construction step (S2)
Using the actual measurement data input in the actual measurement data input step (S1), an average polymerization degree estimation model is constructed by a neural network. Of the measured data described above, the average degree of polymerization is used as an output factor, and various deterioration index components are used as input factors.
(3) Input data input step for diagnosis (S3)
Next, for the oil-filled transformer 20 to be diagnosed, the corresponding input factor is input to the average polymerization degree estimation model (learned neural network). Specifically, the amount of furfural, the amount of CO ₂ + CO, and the amount of acetylene, which are deterioration index components obtained by analyzing the insulating oil OL, are respectively input to the average polymerization degree estimation model.

(4) Average polymerization degree estimation step (S4)
An average degree of polymerization corresponding to the data input in the input step (S3) is calculated by an average degree of polymerization estimation model.
(5) Past actual measurement data input step (S11) regarding the amount of metal thinning
In order to construct a remaining life estimation model of the metal container 22 of the oil-filled transformer 20 by a neural network, actual measurement data of the metal container thinning amount measured in the past is input as learning data of the estimation model.

(6) Thinning amount estimation model construction step (S12)
A remaining life estimation model is constructed by a neural network using the actual measurement data input in the actual measurement data input step (S11). Among the measured data described above, the estimated remaining life value is used as an output factor, and the measured data of the thinning amount is used as an input factor. At this time, in the remaining life estimation model, the initial thickness value and the limit remaining thickness value of the metal container 22 are preset as constant data.
(7) Input data input step for diagnosis (S13)
Next, for the oil-filled transformer 20 to be diagnosed, the corresponding input factor is input to the remaining life estimation model (learned neural network). Specifically, the thinning amount obtained by actually measuring the immersion part 29 of the lid 28 is input to the remaining life estimation model.

(8) Thinning amount estimation step (S14)
The estimated remaining life value of the metal container 22 corresponding to the data input in the input step (S13) is calculated by the remaining lifetime estimation model. At this time, the time until the thickness of the metal container 22 reaches the limit residual pressure value is calculated as the estimated remaining life value. In the present embodiment, the immersion part 29 is used as a representative point of the metal container 22, and the average value when the immersion part 29 is measured at multiple points is defined as a thinning amount, and this is used as an input factor for the remaining life estimation model. However, a portion where the thinning of the hot spot or the like in the metal container 22 is estimated to be the largest may be used as a representative point, and the thinning amount of this portion may be adopted as an input factor.

(9) Ensemble processing step (S15)
The combination of the average degree of polymerization obtained in the average degree of polymerization estimation step (S4) and the estimated remaining life value of the metal container 22 obtained in the thickness reduction amount estimation step (S14) is processed (ensemble process), and finally The remaining life of a typical oil-filled transformer 20 is estimated.
As a method of ensemble processing, for example, the remaining life L1 that can be obtained by the average degree of polymerization, the estimated amount of thinning, and the remaining life L2 that is obtained from the estimated remaining life calculated from the thickness of the metal container 22 are compared. The shorter one is adopted as the final estimated remaining life L.

Next, the remaining life diagnosis method using the neural network according to the present embodiment will be specifically described with reference to FIG.
As an average polymerization degree estimation model, a three-layer neural network was used, and an average polymerization degree estimation model was constructed (first to third layers) by appropriately setting initial values of weight coupling. As the remaining life estimation model, a two-layer neural network is used, and an initial wall thickness estimation model is constructed by presetting constant data for the initial thickness value and the limit remaining thickness value of the metal container 22 (first layer). To the second layer). In this embodiment, the average degree of polymerization estimated value obtained by the average degree of polymerization estimation model and the remaining life estimation value obtained by the remaining life estimation model are subjected to ensemble processing in the fourth layer, so that the final remaining life Get an estimate.

In the remaining life diagnosis method according to the present embodiment, measured values of two or more deterioration index components (furfural amount, CO ₂ + CO amount and acetylene amount) in the insulating oil OL and measured values of the thickness reduction in the metal container 22 are respectively obtained. The average polymerization degree is estimated by learning the model using the input factor group and the average polymerization degree of the insulating paper in the coil 24 and the remaining life of the metal container 22 estimated from the remaining amount of the metal container 22 as output factors. The model and the remaining life estimation model of the metal container 22 are respectively constructed, and the measured values of the furfural amount, the CO ₂ + CO amount, and the acetylene amount of the oil-filled transformer 20 to be diagnosed are input to the average polymerization degree estimation model and reduced. The combination of the average polymerization degree estimated value obtained by inputting the measured value of the amount of meat into the remaining life estimation model and the estimated remaining life value of the metal container 22 The ensemble process is performed, and the final remaining life of the oil-filled transformer 20 is estimated.

  Thus, for example, the oil-filled transformer 20 is installed in a facility where the temperature and chemical environment vary greatly depending on the installation location, and the reduction in the metal container 22 against the progress of deterioration of the insulating paper in the coil 24. When the progress of the meat is fast and the life of the oil-filled transformer 20 is limited by the thinning amount (remaining thickness) of the metal container 22, the remaining life of the oil-filled transformer 20 is reduced to that of the metal container 22. It can be accurately estimated based on the estimated remaining life, and the deterioration of the insulation paper of the coil 24 is accelerated with the progress of the thinning in the metal container 22, and the life of the oil-filled transformer 20 is limited by the deterioration of the insulation paper. In this case, the remaining life of the oil-filled transformer 20 can be accurately estimated based on the estimated value of the average degree of polymerization of the insulating paper in the coil 24.

As a result, according to the remaining life diagnosis method according to the present embodiment, even when the oil-filled transformer 20 is installed in a facility whose temperature and chemical environment vary greatly depending on the installation location, this oil The remaining life of the input transformer 20 can be diagnosed with high accuracy.
In the remaining life diagnosis method of the present embodiment, the measurement values of the furfural amount, the CO ₂ + CO amount, and the acetylene amount are used as the deterioration index component of the insulating paper, but the furfural amount and the CO ₂ + CO amount are used as the deterioration index component. Even if only the estimated value is used, the remaining life can be estimated with sufficiently high accuracy.

In the remaining life diagnosis method of the present embodiment, an average polymerization degree estimation model is constructed using a neural network technique, and a remaining life estimation model is constructed. However, two or more deterioration index components (furfural amount, CO ₂ + CO The amount of acetylene and the amount of acetylene) are processed by multivariable multiple regression analysis to identify the average polymerization degree estimation model, and the measurement of the thinning amount is processed by multiple regression analysis to identify the remaining life estimation model It goes without saying that a result approximate to that obtained when a neural network is used can be obtained.

It is side surface sectional drawing which shows an example of the oil-filled transformer which can apply the remaining life diagnostic method which concerns on embodiment of this invention. It is a flowchart which shows the remaining life diagnosis method which concerns on embodiment of this invention. It is explanatory drawing of each data and input factor in the case of constructing | assembling an average polymerization degree estimation model and a thinning amount estimation model by a neural network.

Explanation of symbols

20 Oil-filled transformer 22 Metal container 23 Top plate part 24 Coil 26 Tap changer 28 Lid 29 Immersion part 30 Inspection window 32, 34 Bushing 36, 38 Terminal member 40, 42 Connection cable 44 Buffer tank 46 Oil supply pipe 48 Oil collection pipe 50 Sampling valve OL insulating oil

Claims (5)

A method for diagnosing the remaining life of an oil-filled transformer having a metal container storing insulating oil and a coil immersed in the insulating oil in the metal container,
Estimated from the average polymerization degree of insulating paper in the coil and the amount of remaining metal in the metal container, with the measured values of two or more deterioration index components in the insulating oil and the measured value of the thinning amount in the metal container as input factor groups, respectively. By identifying or learning the model using the remaining life of the metal container as an output factor, the average polymerization degree estimation model and the remaining life estimation model of the metal container are respectively constructed, and the oil-filled transformer to be diagnosed is constructed. The measurement value of the deterioration index component is input to the average polymerization degree estimation model, and the measurement value of the thinning amount is input to the remaining life estimation model, and the average polymerization degree estimation model and the remaining life estimation model An oil-filled transformer characterized by processing a combination of the estimated average degree of polymerization and the estimated remaining life of a metal container to estimate the final remaining life of the oil-filled transformer Remaining life assessment methods. The method for diagnosing the remaining life of an oil-filled transformer according to claim 1, wherein the measured value of furfural in the insulating oil and the measured value of CO ₂ + CO amount are used as input factors for the average polymerization degree estimation model.   The method for diagnosing the remaining life of an oil-filled transformer according to claim 2, wherein a measured value of the amount of acetylene in the insulating oil is additionally used as an input factor for the average polymerization degree estimation model.   The residual life diagnosis method for an oil-filled transformer according to any one of claims 1 to 3, wherein a neural network is used as the average polymerization degree estimation model and the remaining amount estimation model.   The method for diagnosing the remaining life of an oil-filled transformer according to any one of claims 1 to 3, wherein multiple regression equations are used as the average polymerization degree estimation model and the remaining amount estimation model.

Images