JPH0673336B2 - Abnormality prediction system for oil-filled transformer - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an oil-filled transformer abnormality predicting system, and more particularly to an oil-filled transformer based on the amount of gas components contained in insulating oil of the oil-filled transformer. The present invention relates to a system for predicting an abnormal occurrence time of a vessel.

[0002]

2. Description of the Related Art An oil-filled transformer is an outer box filled with insulating oil (mineral oil, synthetic oil, etc.) containing a main body (structures such as an iron core and windings). Since it has a high cooling effect, it is used in most large capacity transformers. Such oil-filled transformers have arc discharge, local heating,
Deterioration or failure due to various causes such as dehydrogenation reaction, aging, and overload operation. When the internal structure of the oil-filled transformer deteriorates or fails, the insulating oil is thermally decomposed and a specific gas is generated. Since this generated gas is easily dissolved in insulating oil, it is necessary to estimate the deterioration state and the cause of failure of the oil-filled transformer by sampling the insulating oil and analyzing the gas components contained in it. You can

Conventionally, by analyzing a gas component contained in insulating oil with a gas chromatograph or the like, and determining whether or not the amount of each gas exceeds a predetermined threshold value, a person The quality of the input transformer was determined.

[0004]

As described above, conventionally, it is possible to determine the current quality of an oil-filled transformer, but to know when the oil-filled transformer will become abnormal or malfunction. I couldn't. Therefore, the oil-filled transformer must be frequently inspected, which requires a great deal of cost and time.

Therefore, an object of the present invention is to provide a system capable of predicting a future abnormal occurrence time of an oil-filled transformer and thereby reducing the cost required for the inspection work of the oil-filled transformer.

[0006]

The invention according to claim 1 is
A system that predicts the future occurrence of abnormalities in oil-filled transformers, and analyzes the amount of a given gas component contained in insulating oil sampled from oil-filled transformers on a regular or irregular basis. Gas amount analysis means, storage means for storing the time-dependent change of the gas component amount obtained by the gas amount analysis means, and the time-dependent change of the predetermined gas component amount stored in the storage means is converted into an approximate function. And a calculating means for calculating the time for the amount of the predetermined gas component to reach a predetermined threshold value based on the approximation function.

The invention according to claim 2 is characterized in that, in the invention according to claim 1, the following features are further provided. That is, in the invention according to claim 2, the calculation means converts the change with time of the amount of the predetermined gas component stored in the storage means into a plurality of different approximation functions, and based on each of these plurality of approximation functions. Then, the time required for the amount of the predetermined gas component to reach a predetermined threshold value is calculated.

The invention according to claim 3 is the same as the invention according to claim 2, further characterized by the following. That is, in the invention according to claim 3, the arithmetic means converts the change with time of the amount of the predetermined gas component stored in the storage means into a first approximation function which is assumed to change in an average state. First approximation function conversion means, and second approximation function conversion means for converting a change over time in the amount of the predetermined gas component stored in the storage means into a second approximation function that is assumed to change in the worst state. Including and

The invention according to claim 4 is characterized in that, in the invention according to claim 3, the following is further provided. That is, in the invention according to claim 4, the calculating means determines the time-dependent change pattern from the plurality of kinds of second approximation functions according to the time-dependent change pattern of the amount of the predetermined gas component stored in the storage means. The second approximation function conversion means further includes a selection means for selecting an optimum second approximation function, and the second approximation function conversion means selects the second temporal function of the amount of the predetermined gas component stored in the storage means by the selection means. Convert to an approximate function of.

The invention according to claim 5 is characterized by the following in the invention according to claim 4. That is, in the invention according to claim 5, the first approximation function conversion means converts the change with time of the amount of the predetermined gas component stored in the storage means into a linear function obtained by the least square method,
The second approximation function conversion means converts the change over time in the amount of the predetermined gas component stored in the storage means into a quadratic function obtained by the difference method or a linear function obtained by the maximum gradient.

[0011]

In the invention according to claim 1, the change with time of the amount of the predetermined gas component contained in the insulating oil sampled from the oil-filled transformer is converted into an approximate function, and based on this approximate function By calculating the time required for the amount of the predetermined gas component to reach a predetermined threshold value, the future occurrence time of an abnormality of the oil-filled transformer is predicted. According to the system of the present invention as described above, it is possible to significantly reduce the number of insulating oil analysis operations, and thus it is possible to significantly reduce the cost required for the analysis operations. Further, since the abnormality prediction of the oil-filled transformer can be performed automatically and at high speed, it is possible to shorten the working time and the working labor. Furthermore, the abnormality prediction of the oil-filled transformer can be performed without requiring special knowledge and abundant experience.

According to the second aspect of the present invention, the change with time of the amount of the predetermined gas component is converted into a plurality of different approximation functions, and the predetermined gas component of the predetermined gas component is calculated based on each of these plurality of approximation functions. The time required for the quantity to reach a predetermined threshold value is calculated. by this,
Since prediction results are obtained from different viewpoints, it is possible to make more multifaceted prediction decisions.

In the invention according to claim 3, the time-dependent change of the amount of the predetermined gas component is converted into the first approximate function which is assumed to change in an average state, and the change is made in the worst state. Is converted into a second approximation function that assumes that the prediction result is obtained from two contradictory viewpoints. In this case, the difference between the two prediction results can be used as one index indicating the deterioration state of the oil-filled transformer. For example, if the difference between the two prediction results is large, it can be estimated that some failure has occurred in the oil-filled transformer.

According to a fourth aspect of the present invention, a second approximation function that is optimal for the time-dependent change pattern is selected from a plurality of types of second approximation functions according to the time-dependent change pattern of the amount of a predetermined gas component. By selecting, the prediction accuracy is further improved.

In the invention according to claim 5, a linear function obtained by the least-squares method is adopted as the first approximation function, and a second-order function obtained by the difference method is adopted as the second approximation function.
A quadratic function or a linear function obtained by the maximum gradient method is adopted.

[0016]

1 is a block diagram showing the configuration of an oil-filled transformer abnormality prediction system according to an embodiment of the present invention.
In FIG. 1, the oil-filled transformers 11 to 1n to be diagnosed
From this, a predetermined amount of insulating oil is sampled regularly or irregularly. The sampled insulating oil is supplied to the gas amount analyzer 2 and the amount of a predetermined gas component contained in the insulating oil is analyzed. As the gas amount analyzer 2,
A so-called gas chromatography device, an air bubbling type simple analyzer, or the like is used.

The analysis result of the gas amount analyzer 2 is given to the data processing device 3. The data processing device 3 is configured by a personal computer, a microcomputer, or the like, and calculates a future abnormality occurrence time of the oil-filled transformer based on the analysis result of the gas amount analyzer 2. The calculation result of the data processing device 3 is output to the display device 4 and the printer 5, where it is displayed and printed.

Next, with reference to the flow chart shown in FIG. 2, a method of creating the gas amount data necessary for the prediction calculation of the abnormality occurrence time of the oil-filled transformer will be described.

First, in step S1, the oil-filled transformer 11
Insulation oil is sampled from any one of (1 to 1n) (first, the oil-filled transformer 11) by a predetermined amount (about 50 cc) necessary for the analysis work in the gas analyzer 2. At this time, it is necessary to take care not to collect the oil remaining in the lower part of the oil-filled transformer where heat convection does not occur. This is because such residual oil does not move convectively in the transformer, so that the gas generated due to deterioration is less likely to be melted.

Next, in step S2, step S1
The gas amount analyzer 2 analyzes the amount of a desired predetermined gas component among the various gas components contained in the insulating oil sampled in (1). At this time, gas amount analyzer 2
The gas components contained in the sampling oil are expelled and separated by vacuum deaeration or air bubbling.
Furthermore, the gas amount analyzer 2 detects the amount, that is, the volume of a predetermined gas component contained in the insulating oil per unit volume by analyzing the composition and the amount of the separated gas component. The amount of the predetermined gas component detected by the gas amount analyzer 2 is indicated in the unit of ppm, for example.

There are various causes of deterioration of the oil-filled transformer, and most of the gas components generated for each cause of deterioration are combustible gas (TCG), such as carbon monoxide ( CO), hydrogen (H ₂ ), methane (C
H ₄ ), acetylene (C ₂ H ₂ ), ethylene (C
₂ H ₄ ), ethane (C ₂ H ₆ ), propylene (C
₃ H ₆ ) and propane (C ₃ H ₈ ). Therefore, the gas components to be detected in step S2 preferably include carbon monoxide, hydrogen, methane, acetylene, ethylene, ethane, propylene and propane. In addition, one or more kinds of gas components may be selected from among these carbon monoxide, hydrogen, methane, acetylene, ethylene, ethane, propylene, and propane and may be detected.

Next, in step S3, the data processing device 3 creates a data file. In this data file, in addition to the analyzed data of each gas component amount, various attributes (capacity,
(Rated voltage, year of manufacture, serial number, manufacturer, etc.) and the date of sampling the insulating oil are included.

The above steps S1 to S3 are executed for all the oil-filled transformers 11 to 1n to be diagnosed.
Further, the work of steps S1 to S3 is performed regularly or irregularly. Therefore, in the above-mentioned data file, the change with time of the amount of the predetermined gas component contained in the insulating oil is recorded for each of the oil-filled transformers 11 to 1n.

4 to 6 are graphs showing an example of changes with time of the amount of a predetermined gas component contained in the insulating oil. 4 to 6, the horizontal axis represents the passage of time, and the vertical axis represents the amount (unit: ppm) of the predetermined gas component contained in the insulating oil. 4 to 6
As shown in (1), generally, as the oil-filled transformer deteriorates over time, the amount of gas contained in the insulating oil tends to increase. Because, as mentioned above, various gases (for example, carbon monoxide, hydrogen, methane, acetylene, ethylene, ethane, propylene, propane) are generated due to various deteriorations occurring in the oil-filled transformer, This is because it melts in insulating oil. Therefore, the amount of gas contained in the insulating oil is closely correlated with the state of deterioration of the oil-filled transformer. Therefore, the future transition of the contained gas amount in the oil-filled transformer is estimated from the historical change data of the contained gas amount, and the contained gas amount reaches a predetermined threshold value based on the estimation result. By calculating the time, it is possible to predict the future occurrence time of the oil-filled transformer.

Next, referring to the flow chart shown in FIG. 3, the operation of predicting the future occurrence time of an abnormality in the oil-filled transformer will be described.

The above prediction operation is executed by the data processing device 3. First, in step S101, one oil-filled transformer is selected from the plurality of oil-filled transformers 11 to 1n. This selection operation is performed by the operator operating the keyboard attached to the data processing device 3 while looking at the menu displayed on the display 4, for example. When any one of the oil-filled transformers is selected, the process proceeds to step S102, and the data processing device 3 reads the data regarding the corresponding transformer from the data file created in step S3.

Next, in step S103, the data processing device 3 selects one of the gas components from the time-dependent change data of a plurality of gas component amounts contained in the data read from the data file. Select the time course of quantity data. Next, the process proceeds to step S104, and the data processing device 3 determines whether or not all the change gradients of the time-dependent change data of the gas component amount selected in step S103 are 0 or negative. If all of the above-mentioned change gradients are 0 or negative, that is, if the change in the amount of the selected gas component has never shown an increasing tendency in the past, the time-dependent change data of the gas component indicates Since the abnormal occurrence time of the transformer cannot be predicted, the data processing device 3 determines the abnormal occurrence time of the oil-filled transformer as 3 for convenience.
It is determined that the time is 0 years or more, and the determination result is output to the display device 4 and the printer 5 for display and printing (step S1).
05).

On the other hand, if it is determined in step S104 that the change over time in the selected gas component amount has shown a tendency to increase even once in the past, step S106.
Then, the data processing device 3 calculates the time for the selected gas component amount to reach the predetermined threshold value by using the so-called least squares method. That is, in step S106,
Assuming that the change in the gas amount changes on average, the linear function y = ax + b as an approximate expression of the change with time of the gas component amount is obtained from the average value of the past variations in the gas amount data, and the obtained The time required for the amount of the selected gas component to reach a predetermined threshold value is calculated based on the obtained linear function. Below, referring to the graph of FIG.
The prediction process executed in step S106 will be described in more detail.

Now, in the time-dependent change data of the gas amount as shown in FIG. 4, the (i + 1) th measurement date is x _i ,
Let (i + 1) th measurement gas amount be y _i, and let the number of measurements be N
Then, the time S from the first measurement date to the last measurement date, the sum t of the squares of each measurement day, the sum u of the gas amount from the first measurement date to the last measurement date, and the sum v of the product of the measurement date and the gas amount v. Are respectively expressed by the following equations.

[0030]

[Equation 1]

Here, the data processing device 3 solves the following simultaneous equations (1) to obtain the coefficients a and b of the linear function y = ax + b which approximates the temporal change of the gas amount.

[0032]

[Equation 2]

[0033] Solving the simultaneous equations (1), a = (u -Nb) / S b = (tu-Sv) / (tN-S 2) , and the linear function y = ax + b, the following equation (2 ). y = {(u-Nb) / S} x + {(tu-Sv) / (tN-S 2)} ... (2)

Next, the data processor 3 substitutes a predetermined threshold value g (threshold value specific to the selected gas) into y in the above equation (2), and the gas amount The time x to reach the threshold value g is calculated. The calculation result x at this time is expressed by the following equation (3). x = [g - {(tu -Sv) / (tN-S 2)}] / {(u-Nb) / S} ... (3)

Next, in step S107, the data processing device 3 outputs the time calculated by the above equation (3), that is, the time x when the gas amount reaches the threshold value g, to the display 4 and the printer 5. To display and print. The prediction of the abnormality occurrence time using the method of least squares as in step S106 is particularly effective when the deterioration of the oil-filled transformer changes on average.

Next, in step S108, the data processing device 3 determines that the change with time of the selected gas is a predetermined condition,
That is, it is determined whether or not the condition of (Δy ₁ ≧ 0 and Δy ₀ > Δy ₁ ) is satisfied. Here, Δy ₀ is the change gradient between the latest gas amount measurement result and the gas amount measurement result one time before, and Δy ₁ is the gas amount measurement result one time before and the gas amount measurement two times before. It is the slope of change between the results and.

If the above condition is satisfied, that is, if the recent change in the gas amount is increasing and the rate of increase is increasing with time, step S1
In step 09, the data processing device 3 calculates the time for the selected gas component amount to reach the predetermined threshold value by using the so-called difference method. That is, in step S109,
Assuming that the gas amount is increasing at an accelerating rate due to some kind of failure, the quadratic function as an approximate expression of the change with time of the gas component amount is obtained from the latest to two previous measurement results, and the obtained quadratic function is obtained. Based on the function, the time required for the amount of the selected gas component to reach the predetermined threshold value is calculated. The prediction process executed in step S109 will be described in more detail below with reference to the graph of FIG.

Now, as shown in FIG. 5, the point P ₀ indicates the measurement data (gas amount y ₀ ) at the latest measurement date x ₀ , and the point P _0
-1 is the measurement data on the previous measurement day x _-1 (gas amount y
₋₁ ) and the point P ₋₂ indicates the measurement data (gas amount y ₋₂ ) on the measurement date x ₋₂ two times before. First, the data processing device 3 uses the following equations (4) to (6) to obtain the gradient between the data of interest. Δy ₀ = (y ₀ −y ₋₁ ) / (x ₀ −x ₋₁ ) ... (4) Δy ₋₁ = (y ₋₁ −y ₋₂ ) / (x ₋₁ −x ₋₂ ) ... (5) ) Δy = (y ₀ −y ₋₂ ) / (x ₀ −x ₋₂ ) ... (6) However, in the above formulas (4) to (6), Δy ₀ is the point P _0.
And the point P _−1, and Δy ₋₁ is the point P ₋₁ and the point P _−2.
And Δy is the slope between points P ₀ and P _-2 .

Here, the rate of increase of the gradient is expressed as Δ ² y = (Δy ₀ −Δy ₋₁ ) / {(x ₀ −x ₋₂ ) / 2}. Therefore, the change with time of the gas amount can be approximated to a quadratic function like the following expression (7). _{y = y m + Δy (x} -x m) + (Δ 2 y / 2) (x-x m) 2 ... (7) Note that in the above equation _{(7), y m = y} 0 -Δy (x 0 - x _m ) + (Δ ² y / 2) (x
_{^{0 -x m) 2 x m =}} (x 0 + x -2) is / 2.

Next, the data processing device 3 calculates the time for the gas amount y to reach the threshold value g using the above equation (7). Substituting Δy for a and Δ ² y / 2 for b, and substituting the threshold value g in the above equation (7), the following equation (8) is obtained. _{g = y m + a (x} -x m) + b (x-x m) 2 ... (8)

When the above equation (8) is transposed, expanded and rearranged,
The following equation (9) is obtained. ^{0 = x 2 -Bx + C ...} (9) However, in the above equation (9), B = (a -2b · x m) / b C = (b · x m 2 -a · x m + y m -g) / b.

Next, the data processing device 3 uses the root formula to calculate the solution of the above equation (9), that is, the time x at which the amount of gas reaches the threshold value g. The calculation result is the following expression (1
0). x = {- B + (B 2 -4C) 1/2} / 2 ... (10)

Next, in step S110, the data processing device 3 displays the time x obtained by the above equation (10), that is, the time x at which the gas amount reaches the threshold value g, on the display unit 4.
And output to the printer 5 for display and printing. The prediction of the abnormality occurrence time using the difference method as in step S109 is particularly effective when some failure occurs in the oil-filled transformer and then the gas amount increases at an accelerated rate.

On the other hand, in step S108 described above,
The change of the gas amount with time is (Δy ₁ ≧ 0 and Δy ₀ > Δ
When it is determined that the condition of y ₁ ) is not satisfied, the process proceeds to step S111, and the data processing device 3 uses the so-called maximum gradient method, and the selected gas component amount reaches a predetermined threshold value. Calculate time. This maximum gradient method
Time to find the worst case (when the increase rate is the maximum) from the past gas quantity data and assume that the worst case will recur, and the time for the quantity of the selected gas component to reach the specified threshold value. Is calculated. The prediction process executed in step S111 will be described in more detail below with reference to the graph of FIG.

As shown in FIG. 6, it is assumed that six measurements have been performed so far. In FIG. 6, point P
₆ indicates the latest measurement result, and point P ₁ indicates the first measurement result. First, the data processing device 3 obtains a gradient between points. Δy ₆ = (y ₆ −y ₅ ) / (x ₆ −x ₅ ) Δy ₅ = (y ₅ −y ₄ ) / (x ₅ −x ₄ ) Δy ₄ = (y ₄ −y ₃ ) / (x ₄ −x ₃ ) Δy ₃ = (y ₃ −y ₂ ) / (x ₃ −x ₂ ) Δy ₂ = (y ₂ −y ₁ ) / (x ₂ −x ₁ ) where Δy ₆ is a point P ₆ is the slope between point P ₅ , Δy ₅ is the slope between points P ₅ and P ₄ , Δy ₄ is the slope between points P ₄ and P ₃ , Δy
₃ is the slope between points P ₃ and P ₂ , and Δy ₂ is the point P
It is the slope between ₂ and the point P ₁ .

Next, the data processing device 3 uses the gradient Δy ₆ ~.
Find the maximum gradient Δy _max from Δy ₂ . Figure 6
In the case of, Δy ₄ is the maximum gradient. Next, in the data processing device 3, the future gas amount is the maximum gradient Δ from the point P _6.
Assuming that y _max is maintained, the time for the selected gas component amount to reach a predetermined threshold value is calculated. When the threshold value is g, g = Δy _max · x + x ₆ and the time x for reaching the threshold value is x = (g−x ₆ ) / Δy _max .

Next, in step S112, the data processing device 3 outputs the time obtained in step S112, that is, the time x at which the amount of gas reaches the threshold value g, to the display 4 and the printer 5, and displays it. And print. In the prediction of the abnormality occurrence time using the maximum gradient method as in step S112, all the past tendency of changes in the gas amount of the oil-filled transformer are investigated to express the characteristic tendency of the transformer. It can also handle sudden failures.

After step S110 or S112, the process proceeds to step S113, in which the data processor 3 determines whether or not the prediction of the threshold arrival time for all gas components is completed. Step S102
The operations of to S112 are repeated. Therefore, the prediction result of the threshold arrival time is output for each gas component. On the other hand, when the prediction of the threshold arrival time for all the gas components is completed, the operation of the data processing device 3 ends.

Next, the results of experiments when the prediction of the time of occurrence of abnormality by the system of the above-mentioned embodiment is carried out for an oil-filled transformer that is actually used are shown below. FIG. 7 shows a time-dependent change (solid curve) of the amount of carbon monoxide (CO) contained in the insulating oil of the oil-filled transformer that has been used for about 20 years. Based on the data of this oil-filled transformer from August 1975 to December 1988, the future abnormal occurrence time is predicted in both the standard case and the worst case, and each prediction result is the data of February 1990. Verified in. As a standard case, when the abnormal occurrence time of the oil-filled transformer is predicted by the least squares method, the caution arrival time and the abnormality arrival time are estimated to be 30 years or more.
In addition, in the worst case, when the abnormality occurrence time of the oil-filled transformer is predicted by the maximum gradient method, it is estimated that the caution arrival time is 7.3 years later and the abnormality arrival time is 28.1 years later. The threshold value of the amount of carbon monoxide determined to require caution was set to 300 ppm, and the threshold value of the amount of carbon monoxide determined to be abnormal was set to 800 ppm.

In FIG. 7, the dotted line shows the transition curve of the amount of carbon monoxide obtained by the least squares method, and the change curve of the alternate long and short dash line shows the transition curve of the amount of carbon monoxide obtained by the maximum gradient method. ing. Referring to the data of February 1990, it can be seen that the change in the amount of carbon monoxide in this oil-filled transformer changes in substantially the same manner as in the standard case. This is because no serious failure has occurred in the oil-filled transformer and the deterioration is gradually progressing. If a serious failure occurs in the oil-filled transformer, the change in carbon monoxide content is expected to approach the worst case change. Further, in this case, the difference between the threshold arrival time in the standard case and the threshold arrival time in the worst case will be widened. Therefore, the difference between the prediction results of both the standard case and the worst case can be one index for determining the state of the oil-filled transformer.

Generally, in a power transmission facility or a power receiving facility in which a transformer is used, it is often not permissible to stop the power transmission or the power reception due to the occurrence of an abnormality. Therefore, it is required to make a management plan in consideration of a certain safety factor. To be done. As described above, in the embodiment of the present invention, the time when the gas amount reaches the threshold value can be predicted in both the standard case and the worst case. A management plan for the oil-filled transformer can be established by setting the threshold arrival time as the longest life of the oil-filled transformer and the threshold arrival time predicted by the worst case as the shortest life of the oil-filled transformer. As a result, it becomes easier for the manager to make a management plan that allows for a safety factor.

In the present invention, the future occurrence time of the oil-filled transformer may be predicted in either the standard case or the worst case.

In the above embodiment, the threshold arrival time is calculated for each of a plurality of combustible gases (TCG) generated due to deterioration or failure.
It is also possible to predict the time when the abnormality of the oil-filled transformer occurs by obtaining the time when the total amount of TCG reaches the threshold value. FIG. 8 shows the time-dependent change (solid curve) of the total amount of TCG contained in the insulating oil of the same oil-filled transformer as in FIG. 7. As a standard case, when the abnormal occurrence time of the oil-filled transformer is predicted by the least squares method, the caution arrival time and the abnormality arrival time are estimated to be 30 years or more. In addition, when the abnormality occurrence time of the oil-filled transformer is predicted by the difference method as the worst case, it was estimated that the caution arrival time was 4.5 years later and the abnormality arrival time was 9.4 years later. In FIG. 8, the dotted line shows the transition curve of the TCG amount obtained by the least square method, and the change curve of the two-dot chain line shows the transition curve of the TCG amount obtained by the difference method.

By the way, it is known that the main internal structure that determines the life of the oil-filled transformer is insulating paper (kraft paper) wound around the windings and insulating between the windings. .
That is, the life of the oil-filled transformer is often exhausted when the mechanical strength of the insulating paper drops below a predetermined value. Generally, the mechanical strength of insulating paper can be expressed by tensile strength and average polymerization degree residual rate, but conventionally it has been necessary to sample the insulating paper from the oil-filled transformer to measure the tensile strength and average polymerization degree residual rate. It was Therefore, there is a problem in that the operation of the oil-filled transformer has to be stopped, which causes a power failure. In addition, the internal structure of the oil-filled transformer needs to be lifted by a crane or the like, which causes a problem that the sampling work is troublesome and costly. Further, there is a problem in that the portion where the insulating paper is taken must be repaired later. Furthermore, the tensile strength measurement test and the average degree of polymerization residual rate measurement test must be performed on the collected insulating paper, which causes a problem that the test work is troublesome and costly.

From the above-mentioned various problems, it is practically difficult to measure the mechanical strength of the insulating paper.
The reality is that the mechanical strength of the insulating paper was rarely examined to measure the deterioration state of the oil-filled transformer.

The inventor of the present application investigated past measurement data disclosed in various documents, and further measured 100 oil-immersed transformers. As a result, the average residual polymerization degree of insulating paper and insulating oil were measured. It has been found that there is a correlation between carbon monoxide (CO) and carbon dioxide (CO ₂ ) contained in the material.

FIG. 9 shows the correlation between the average residual polymerization degree of the insulating paper and the carbon monoxide and carbon dioxide contained in the insulating oil, which was obtained by examining the past measurement data. It is a graph shown. In FIG. 9, the horizontal axis represents carbon monoxide and carbon dioxide (CO + C) per 1 g of insulating paper.
The volume (unit: ml / g) of O ₂ ) is represented on a logarithmic scale, and the vertical axis represents the average polymerization degree residual rate (%). It is generally said that when the average residual degree of polymerization is 20% to 30%, the mechanical strength of the insulating paper is lost, and the insulating paper is practically 40%.
It can be said that ~ 60% is the life of the oil-filled transformer. From the graph of FIG. 9, the volume amount of CO + CO ₂ is 1 / g of insulating paper.
When the amount is ml (ml), the average degree of polymerization residual rate is 60.
It turns out to be%.

The inventor of the present application converts the horizontal axis of the graph of FIG. 9 from a logarithmic scale to a uniform scale and further considers the measurement data obtained for 100 oil-immersed transformers to show the correlation shown in FIG. Created a relationship graph. In FIG. 10, the change curve A is the average polymerization degree residual ratio of the insulating paper and C for 1 g of the insulating paper.
3 shows a cubic function curve that approximates the correlation with the volume of O + CO ₂ . Further, in FIG. 10, the change curve B
Shows a quadratic function curve that approximates the correlation shown in FIG. As is clear from FIG. 10, in the vicinity of the average residual polymerization degree of 60% to 40%, the cubic function curve A is closer to the above correlation than the quadratic function curve B is. .

Now, let x be CO + CO ₂ for 1 g of insulating paper.
Volume ratio (ml / g), and y is the average degree of polymerization residual rate (%)
Then, the cubic function curve A in FIG. 10 is represented by y = a · x ² + b · x ² + c · x + d (11). Then, when each coefficient of the above equation (11) is obtained based on the past measurement data and the newly obtained actual measurement data, a = −2.06 b = 17.99 c = −56.52 d = 98.00 Became. Therefore, the above equation (11) becomes y = −2.06x ² + 17.99x ² −56.52x + 98.00 (12).

Therefore, in the embodiments described below, the current average polymerization degree residual ratio of the insulating paper is indirectly calculated using the above formula (12), and CO + C for 1 g of the insulating paper is calculated.
It is characterized by predicting the time when the volume of O ₂ reaches 1 ml, that is, the time when the average polymerization degree residual ratio reaches 60% and the life of the oil-filled transformer is exhausted.

Next, an abnormality prediction system for an oil-filled transformer according to another embodiment of the present invention will be described. The system configuration of the other embodiments is similar to the system configuration of the above-described embodiment shown in FIG. 1, so only the operation will be described below.

First, insulating oil is sampled from each of the oil-filled transformers 11 to 1n, and the gas amount analyzer 2 analyzes the amount of gas components in the insulating oil. This work is substantially the same as the work shown in FIG. However, CO ₂ is added to the gas component to be analyzed.

Next, the data processing device 3 executes a predetermined process. Details of the processing executed by the data processing device 3 are shown in FIG. 11, the processing of steps S101 to S113 is exactly the same as the corresponding processing shown in FIG. Therefore, similarly to the above-described embodiment, the time when the volume of each gas component reaches a predetermined threshold value, that is, the time when an abnormality occurs in the oil-filled transformer in the future is calculated.

Step S105 or step S113
After the operation of, the process proceeds to step S201, and the data processing device 3 determines whether the total amount of combustible gas (TCG) is equal to or more than a specified value. If the total amount of TCG is more than the specified value,
Since there is a high possibility that some kind of failure has occurred inside the oil-filled transformer, the data processing device 3 does not calculate the current average polymerization degree residual ratio of the insulating paper and the time until the threshold value is reached, and The operation ends. That is, when the total amount of TCG is greater than or equal to the specified value, the life of the oil-filled transformer depends on the failure factor rather than the secular change of the insulating paper.

On the other hand, if the total amount of TCG is less than the specified value, the process proceeds to step S202, and the data processing device 3 selects the oil selected in step S101 from the data file (see FIG. 2) created in step S3. Read the data corresponding to the input transformer. The data read at this time is the CO and CO contained in the sampling oil.
_{2 is} the volume data. Data for other gas components are not read.

Next, in step S203, the data processing device 3 reads CO + C read in step S202.
Temperature correction is performed on the volume data of O ₂ . This is because it is known that CO and CO ₂ are easily absorbed by the insulating paper at low temperatures, but are not adsorbed at 80 ° C. or higher. Here, the following equations (13) and (1
4) is used to correct the volume of CO and CO _{2 to} the volume at 80 ° C. _{M (CO) = M 1 ×} exp [560 {(1 / T) -0.0028}] ... (13) M (CO 2) = M 2 × exp [2260 {(1 / T) -0.0028} ] (14) In the above formulas (13) and (14), M ₁ is the temperature t ° C.
Is the volume of CO ₂ at the time, and M ₂ is the volume of CO ₂ at the temperature of t ° C. Further, T is an absolute temperature (273 + t ° C.). From the above formulas (13) and (14), the volume amount of CO + CO ₂ at 80 ° C. is M (CO) + M (CO ₂ ).

Next, in step S204, the data processing device 3 calculates the volume amount (ml / g) of CO + CO ₂ with respect to the insulating paper 1g. That is, the volume amount M (CO) + M (CO ₂ ) of CO + CO ₂ detected from the sampled oil and temperature-corrected is converted into a volume amount with respect to the total oil amount, and the converted volume amount is the total weight of the insulating paper. By dividing by, the volume of CO + CO ₂ with respect to 1 g of insulating paper can be obtained.

Next, proceeding to step S205, the data processor 3 substitutes the volume amount of CO + CO ₂ with respect to 1 g of the insulating paper obtained in step S203 into the above equation (12) to calculate the current average degree of polymerization of the insulating paper. Calculate the residual rate. The calculation result of step S204 is displayed on the display unit 4 in step S206.
And output to the printer 5, where it is displayed and printed.

Next, in step S207, the data processor 3 reduces the time when the volume amount of CO + CO ₂ per 1 g of insulating paper reaches 1 ml / g, that is, the average residual polymerization degree of the insulating paper decreases to 60%. Calculate time. In step S207, the data processing device 3 first sets CO + C.
Calculate the generation rate of O ₂ . That is, the production rate of CO + CO ₂ can be obtained by calculating (the volume of CO + CO ₂ at the time of the first measurement / the volume of CO + CO ₂ at the latest measurement). Here, when past measurement data and actual measurement data were analyzed, it was found that the amount of CO + CO ₂ generated increased substantially in proportion to time. Therefore, based on the production rate of the CO + CO _2, it can be volume of CO + CO ₂ to the insulating sheet 1g calculates the time to reach 1 ml / g. That is, the data processing device 3
Is CO based on 1 g of insulating paper based on the following equation (15).
The time x at which the volume of + CO ₂ reaches 1 ml / g is calculated. x = (1−z) / v (15) In the above formula (15), z is the current value of the volume amount of CO + CO ₂ with respect to the insulating paper 1 g obtained in step S203, and v is the CO + CO _{2 value} . It is the generation rate. The calculation result of step S207 is output to the display 4 and the printer 5 in step S208, and is displayed and printed there.

In the above examples, the current average degree of polymerization residual ratio of the insulating paper and the average degree of polymerization residual ratio of the current insulating paper are reduced to a predetermined value or less based on the volume of CO + CO ₂ contained in the insulating oil. However, the current average polymerization residual rate and average polymerization residual rate of the insulating paper are calculated based on other gas or liquid impurity components (for example, furfural) contained in the insulating oil. You may make it calculate the time when it falls below a predetermined value.

Further, in the above-mentioned embodiment, the CO + CO ₂ volume amount is calculated from the first measurement data and the latest measurement data.
+ Calculated production rate of the CO _2, the average the polymerization degree retention are such that operation time falls below a predetermined value, CO + CO ₂ volume amounts of time course data CO + CO ₂ volume on the basis of the production rate A function that approximates the change curve of the amount may be obtained, and the time for which the average residual polymerization degree falls below a predetermined value may be calculated based on this approximate function.

[0072]

According to the first aspect of the invention, the change over time in the amount of the predetermined gas component contained in the insulating oil sampled from the oil-filled transformer is converted into an approximate function, and this approximate function is converted. Since the time for the amount of the predetermined gas component to reach the predetermined threshold value is calculated based on the above, it is possible to predict the future occurrence time of abnormality of the oil-filled transformer. According to the system of the present invention as described above, it is possible to significantly reduce the number of insulating oil analysis operations, and thus it is possible to significantly reduce the cost required for the analysis operations. Further, since the abnormality prediction of the oil-filled transformer can be performed automatically and at high speed, it is possible to shorten the working time and the working labor. Furthermore, the abnormality prediction of the oil-filled transformer can be performed without requiring special knowledge and abundant experience.

According to the second aspect of the present invention, the change over time in the amount of the predetermined gas component is converted into a plurality of different approximation functions, and the predetermined gas component is calculated based on each of these plurality of approximation functions. Since the time required for the quantity to reach a predetermined threshold value is calculated, prediction results from different viewpoints can be obtained. Therefore, it is possible to make a more multifaceted prediction judgment.

According to the third aspect of the present invention, the change over time in the amount of the predetermined gas component is converted into the first approximation function which is assumed to change in an average state, and the change in the worst state is made. Since it is converted into the second approximation function assuming that, it is possible to obtain the prediction result from two contradictory viewpoints. As a result, the difference between the two prediction results can be used as one index indicating the deterioration state of the oil-filled transformer. For example, if the difference between the two prediction results is large,
It can be presumed that some sort of failure occurred in the oil-filled transformer.

According to the fourth aspect of the invention, a plurality of types of the second type are provided according to the temporal change pattern of the amount of the predetermined gas component.
Which is the most suitable for the aging pattern from the approximate function of
Since the approximation function of is selected, the prediction accuracy can be further improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing an overall configuration of an oil-filled transformer abnormality prediction system according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure for creating gas amount data in the embodiment shown in FIG.

FIG. 3 is a flowchart showing an abnormality prediction operation executed by the data processing device in the embodiment shown in FIG.

FIG. 4 is a graph showing an example of changes over time in the volume of a predetermined gas component contained in insulating oil sampled from an oil-filled transformer.

FIG. 5 is a graph showing another example of the change over time in the volume of a predetermined gas component contained in insulating oil sampled from an oil-filled transformer.

FIG. 6 is a graph showing still another example of changes over time in the volume of a predetermined gas component contained in insulating oil sampled from an oil-filled transformer.

FIG. 7: CO in an oil-filled transformer actually used
3 is a graph for comparing the change over time in the gas generation amount with the transition curve of the CO gas generation amount predicted by the embodiment shown in FIG. 1.

FIG. 8 shows a change with time in the amount of combustible gas generated in an oil-filled transformer that is actually used, and a transition curve of the amount of combustible gas predicted in the same manner as in the example shown in FIG. It is a graph for comparing.

FIG. 9 is a graph showing the correlation between the amount of CO + CO ₂ gas generated and the average residual polymerization degree of insulating paper.

FIG. 10 is a graph showing the correlation between the amount of CO + CO ₂ gas generated and the average polymerization degree residual ratio of the insulating paper. The graph shown in FIG. 9 is converted into a uniform scale, and the oil content actually used is changed. It was obtained by measuring a transformer.

FIG. 11 is a flow chart showing an oil-filled transformer abnormality prediction operation and an insulation paper deterioration prediction operation which are executed by the data processing device in another embodiment of the present invention.

[Explanation of symbols]

 11-1n: Oil-filled transformer 2: Gas amount analyzer 3: Data processing device 4: Indicator 5: Printer

Claims (5)

[Claims] 1. A system for predicting a future abnormality occurrence time of an oil-filled transformer, wherein a predetermined amount contained in insulating oil sampled from the oil-filled transformer regularly or irregularly. A gas amount analyzing means for analyzing the amount of the gas component, a storage means for storing the change over time of the amount of the gas component obtained by the gas amount analyzing means, and a quantity of the predetermined gas component stored in the storing means. Abnormality of the oil-filled transformer, which is provided with a calculating means for converting a change with time into an approximate function and calculating a time for the amount of the predetermined gas component to reach a predetermined threshold value based on the approximate function. Prediction system. 2. The calculation means converts a change over time in the amount of a predetermined gas component stored in the storage means into a plurality of different approximation functions, and based on each of the plurality of approximation functions, The abnormality prediction system for an oil-filled transformer according to claim 1, which calculates a time required for the amount of the predetermined gas component to reach a predetermined threshold value. 3. The first calculation means converts the change over time in the amount of the predetermined gas component stored in the storage means into a first approximation function that is assumed to change in an average state. An approximate function converting means and a second approximate function converting means for converting a temporal change in the amount of the predetermined gas component stored in the storage means into a second approximate function which is assumed to change in the worst state. The abnormality prediction system for an oil-filled transformer according to claim 2, including the abnormality prediction system. 4. The calculating means is adapted to correspond to a temporal change pattern of the amount of a predetermined gas component stored in the storage means.
The second approximation function converting means further includes a selecting means for selecting a second approximation function most suitable for the temporal change pattern from a plurality of types of second approximation functions, and the second approximation function converting means stores the predetermined approximation function stored in the storage means. The time-dependent change in the amount of the gas component of is converted into a second approximation function selected by the selection means.
Abnormality prediction system for oil-filled transformer described in. 5. The first approximation function conversion means calculates the change over time in the amount of a predetermined gas component stored in the storage means,
The second approximate function conversion means converts the linear function obtained by the least-squares method into the temporal function of the amount of the predetermined gas component stored in the storage means. The abnormality prediction system for an oil-filled transformer according to claim 4, which is converted into a linear function or a linear function obtained by a maximum gradient.