JPH05264645A - Method for estimating deterioration of insulating material - Google Patents

Abstract

PURPOSE:To easily estimate the deterioration of an insulating material even when the insulating material which is an estimating target is changed. CONSTITUTION:This device has a standard value converting part 2 for converting a plurality of measurement data related to an insulating material as an estimating target into a plurality of standard data related to the estimation standard target, respectively, according a predetermined converting graph, and a fuzzy inference part 3 for determining a composed deterioration from the standard data according to a predetermined membership function. Further, it has a neural network 4 for estimating the dielectric breakdown voltage survival rate from the determined composed deterioration and learning the actual dielectric breakdown voltage survival rate obtained by a breakdown test to correct the dielectric breakdown voltage survival rate to be estimated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for estimating insulation deterioration of insulators, and more particularly to a method and apparatus for estimating deterioration of insulation suitable for estimating deterioration of different insulation materials.

[0002]

2. Description of the Related Art A coil insulator of a rotating machine has a dense structure with few voids (voids) at the time of manufacture. This is subject to various deteriorations such as electrical deterioration, heat cycle deterioration, thermal deterioration, mechanical deterioration, and environmental deterioration depending on the operation history, and many small voids and large local voids are generated in the entire insulating layer. Leading to. Furthermore, as the operating time becomes longer and the deterioration progresses, the void amount further increases and the voids become connected. That is, the insulation deterioration is the generation and expansion of voids in the insulating layer. When a high voltage is applied, the void portion is discharged and becomes a conductor, so that the insulation path is shortened and the breakdown voltage (hereinafter, BDV) is reduced. .. Therefore, it is known that the BDV decreases as the number of voids increases on average and the larger voids locally exist. Nondestructive tests such as a dielectric loss tangent test (tan δ) of an insulator, an alternating current test, and a partial discharge test are generally carried out in order to know such a state of insulation deterioration.

The value of tan δ in the dielectric loss tangent test is peculiar to the insulator and is used to represent the property of the insulator itself, but this value does not represent the local property of the specimen, but is averaged. It shows the properties that have been obtained. FIG. 17
Shows the tan δ-voltage characteristic. As the deterioration progresses and the number of voids in the insulating layer increases, partial discharge occurs, and
δ ₂ rises. This is an index of deterioration.

In the AC current test, when a voltage is applied to a good insulator, the voltage-current curve is linear, but when the insulation absorbs moisture to increase the dielectric constant or when partial discharge occurs, the current increases. The rate soars. Utilizing this characteristic,
Determine the properties of the insulator. FIG. 18 shows characteristics of alternating current-voltage. Primary and secondary current surge points Pi ₁ , P
i ₂ and the rate of increase in AC current ΔI serve as a criterion for determining the insulation deterioration status.

The partial discharge test focuses on the discharge phenomenon that partially occurs in the insulating medium when a voltage is applied to the insulator,
The maximum discharge charge amount or the number of generated predetermined charge amounts per cycle is measured as a voltage characteristic. In partial discharge, an internal corona is divided into an external corona and an external corona depending on the location where the partial discharge occurs, but the internal corona is used to know the deterioration state inside the insulating layer, that is, the generation state of voids.

FIG. 19 shows the relationship between the partial discharge Qmax and the voltage. Qmax is generally a standard voltage (E / √3
E: The maximum discharge charge amount at the rated voltage) is used.
Contrary to δ, it represents the local deterioration situation.

A method of estimating the BDV residual rate of insulation by using Δtan δ ₂ , ΔI, Qmax, which is indispensable for grasping the insulation deterioration situation described above, is disclosed in Japanese Patent Laid-Open No. 56-7466.
Japanese Patent Publication No. 4 discloses the following empirical formula when the insulator is polyester. _{V R = 100-2 (△ -0.8)} -67log (Qmax / 1.5 × 10 ~ 9) where, V _R: breakdown voltage remaining ratio △ = △ tanδ ₂ + △ in I (discharge parameter) is there.

[0008]

As the insulating material, at first, a natural resin (such as shellac or asphalt compound) was generally used, but due to the development of polymer chemistry, an excellent synthetic resin was developed. Therefore, various insulators such as polyester insulators and epoxy insulators have been used.

Therefore, in such a conventional technique, each time the insulator to be estimated changes, the constants of the empirical formula are recalculated by a complicated statistical process in consideration of the mutual relation of a plurality of factors. However, there is a problem that it is difficult to deal with the case where the insulator is changed, and the versatility is low. The present invention has been made by paying attention to such a conventional problem, and even if the insulator changes, it can be easily dealt with, and a versatile insulator deterioration estimating method, deterioration estimating device, It is also an object of the present invention to provide a remaining life diagnosis device that obtains a remaining life from a value estimated by these.

[0010]

In order to achieve the above-mentioned object, an insulation deterioration estimating apparatus converts measured data relating to an insulation subject to estimation into reference data relating to an estimation reference subject according to a predetermined conversion rule. And a estimating means for estimating the residual rate of the dielectric breakdown voltage from the reference data according to a predetermined estimation rule.

Further, another insulation deterioration estimating device for achieving the above object is as follows:
In an insulator deterioration estimating device for estimating a dielectric breakdown voltage residual rate of the insulator based on measurement data including at least a plurality of non-destructive test data, according to a conversion rule set in advance, the insulator to be estimated Reference value conversion means for converting each of the plurality of measurement data into a plurality of reference data related to an estimated reference object, and estimation means for estimating a dielectric breakdown voltage residual rate from the plurality of reference data according to a predetermined estimation rule. And a rule correction unit that corrects the estimation rule based on the actual dielectric breakdown voltage residual rate by learning at least the actual dielectric breakdown voltage residual rate obtained by the breakdown test. To do.

Here, the estimating means obtains the respective fitness degrees of deterioration for the plurality of reference data based on a plurality of predetermined membership functions, and at the same time, fuzzy reasoning for synthesizing the plurality of fitness degrees of deterioration. Means, and a neutral network for obtaining the dielectric breakdown voltage residual rate from the combined deterioration suitability, the rule correction means, the at least the actual dielectric breakdown voltage residual rate obtained by a breakdown test as at least a teacher signal. The above-mentioned neutral network may be used. In addition to the nondestructive test data, it is desirable to use operation history data as the measurement data to be input to the above-described insulation deterioration estimating device.

[0013]

The measured data on the insulator to be estimated is converted into the reference data on the estimated reference object by the reference value conversion means. The estimation means estimates the residual dielectric breakdown voltage from the reference data obtained by the conversion. In this way, once the measured data is converted to the reference data, the dielectric breakdown voltage residual rate of the estimation target is calculated, so even if the insulation of the estimation target changes, the estimation rule used for estimation It is not necessary to change according to the type of insulator, and even if the insulator changes, it can be easily dealt with. When the insulator changes, it is necessary to change the conversion rule for each measurement data. To change this change rule, the constants of the empirical formula are determined by taking into consideration the mutual relation of various measurement data. It's much easier than the task. This tendency is remarkable especially when the number of types of measurement data increases.

Further, as long as the fuzzy inference means and the neutral network are provided, even if it is desired to increase the types of measurement data, it is possible to cope with the increase by simply increasing the membership function of the fuzzy inference means. In addition, the learning function of the neutral network can improve the accuracy of the estimated dielectric breakdown voltage residual rate.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. Here, a polyester-based insulator is used as an estimation target, and Δtan δ ₂ , ΔI, and Qmax are taken in as input values from the nondestructive test value of this insulator, and the start / stop count N and operating years Y are taken in from the operation history. I shall. In addition, it is assumed that an epoxy-based insulator is selected as a target for the estimation standard.

As shown in FIG. 1, the residual life diagnosing device of the present embodiment is a reference conversion unit for equivalently converting various data relating to a polyester-based insulator which is an estimation target into data of an epoxy-based insulation which is an estimation reference target. 2, a fuzzy-inference unit 3 for synthesizing the equivalently transformed data to obtain synthetic deterioration degrees 3k, 3l, a neural network 4 for obtaining a dielectric breakdown voltage residual rate from the synthetic deterioration degrees 3k, 3l, and the obtained insulation A remaining life calculation unit 5 for obtaining a remaining life of an estimation target from the breakdown voltage remaining rate. By the way, each of these units functionally represents the configuration of the apparatus, and in the present apparatus, as shown in FIG. 2, the ROM 12 and the RAM in which data and programs for various calculations are stored are actually stored.
13, a CPU 11 that executes various calculations based on programs and the like stored therein, a floppy disk device 14, a printer 15, and a display device (CRT) 1.
6 and a keyboard 17. Therefore, the actual operation in each of the above parts 2, 3, ...
A program registered in the floppy disk device 14 or the like is called and stored in the RAM 13, and C
This is achieved by the execution of PU11. Note that the processing in the neutral network 4 is also realized by software using a so-called normal computer in this embodiment, but an actual neutral network may be used.

First, in the reference value converter 2, as input data, Δtan δ ₂ 1a, ΔI 1b, Qmax 1c, N 1
d, Y 1e is input. As shown in FIG. 3, the reference value conversion unit 2 makes the non-destructive test data and the operation history data measured on the estimation target polyester insulation equivalent to the deterioration level of the estimation reference epoxy insulation. It is to convert. That is, the membership function of the fuzzy inference unit 3 and the learning function of the neural network 4 necessary for estimating the dielectric breakdown voltage residual rate described below are
Since it is created for the epoxy-based insulator which is a basic insulator, the reference value conversion unit 2 has a function of equivalently converting all input data into the level of the epoxy-based insulator. As a conversion method, it is necessary to convert each non-destructive test value and each operation history in relation to a common factor, and in this example, the dielectric breakdown voltage (BDV) residual rate was used as this common factor.

As an example of the equivalent conversion method using FIG. 4, Q
The conversion of max will be described. In the figure, the horizontal axis represents Qmax, the vertical axis represents the dielectric breakdown voltage residual rate, the solid line represents the polyester-based insulator, and the dashed line represents the epoxy-based insulator. In the estimated polyester insulation,
If the measured value now shows QmaxP, the BDV residual rate V _R at that time can be obtained. Next, when the Qmax value of the epoxy-based insulation subject to the estimation standard corresponding to the calculated BDV residual rate V _R is calculated, it becomes QmaxE, which is the equivalent conversion value from the polyester-based insulation to the epoxy-based insulation. Treated

The relationship between the individual nondestructive test values and the operation history and the dielectric breakdown voltage residual rate used here is merely a means for equivalently converting the input data into the estimation reference object, and in the present embodiment. There is a large error in the final insulation breakdown voltage residual rate 5 which is the final conclusion to be obtained, and it cannot be used. Specifically, in the case of polyester-based insulators, generally, when Qmax reaches about 10000 pc, the insulation is at the end of its life, and deterioration is accelerated in the future, and the dielectric strength is in a region in which it rapidly decreases. On the other hand, in the epoxy-based insulator, about 30,000 pc is Qmax corresponding to the above-mentioned life period, and there is a difference in deterioration level of about three times.

Also for Δtan δ ₂ , ΔI, N, Y,
Equivalent conversion is similarly performed using a conversion graph as shown in FIG. Note that this conversion graph is a conversion rule, and it is necessary to change a part of this conversion graph in order to estimate deterioration for different insulators. Each input Δtan δ ₂ '+ ΔI' and Qm equivalently converted in this way
Ax 'is input as a non-destructive test value, and N'and Y'are input as operation histories to the fuzzy inference unit 3. Here, the respective synthetic deterioration degrees divided into the non-destructive test value and the operation history are input. Desired. Here, Δtan δ ₂ '+ ΔI' and Qmax '
And are separated and input to fuzzy inference to obtain the synthetic deterioration degree. The reason is that Δtan δ ₂ ′ + ΔI ′ is an index indicating average deterioration, whereas Qmax ′ is an index indicating local deterioration and has different physical meanings. _{Also, △ tanδ 2 ', △ I} ' than correlation to each individual breakdown voltages, △ tanδ ₂ '+
This is also because the combination with ΔI 'improves the correlation with the dielectric breakdown voltage.

A general fuzzy inference method for estimating the degree of deterioration in the fuzzy inference unit 3 will be described with reference to FIGS. 5 and 6. An example of using four types of membership functions in the condition part where data is input will be given. Figure 5
As shown in FIG. 3, the case of no deterioration 3a is expressed by efb, the deterioration small 3b is expressed by agc, the deterioration 3c is expressed by bhd, and the deterioration large 3d is expressed by cij. For example, the membership function has a goodness of fit of 1 in the sense that the input data is completely in the set of no deterioration from e to f in the range of 0 to b when the input data is 0 to b. Has become. In addition, from f to b, there is a region of a value that does not gradually belong to the set of no deterioration, and in b,
The degree of conformance is 0 because it does not belong at all. The membership functions 3b, 3c and 3d are set in the same way. On the other hand, in the conclusion part which is the output, the membership function is shown with the deterioration degree on the horizontal axis. In the case of no deterioration 3e, qrl, small deterioration 3f is ksn, deterioration 3g is mtp, and deterioration 3h is ouv. Has been done.
The membership function of the conclusion section is set substantially the same as that of the condition section.

Now, when w is entered as input data,
Membership function x during deterioration of 3c and small deterioration 3b
The y-point of will represent the corresponding goodness of fit. In the conclusion, the goodnesses of fit x and y are made to correspond to the deterioration 3g and the deterioration small 3f of the corresponding membership function, and the vertex of the deterioration 3g is x ′ and the vertex of the deterioration small 3f is y ′. , A triangle mx'p and a triangle ky'n are created. The conclusion is that a combination of both triangles with a shape of ky'zx'p is obtained, but this itself is "ambiguous" because it is represented by a membership function. The conclusion that makes this clear is called a representative value, and there are various calculation methods, but here, it is performed by the center of gravity method that takes the center of gravity of the shape obtained as the conclusion.

The center of gravity method will be described with reference to FIG. First, divide the deterioration level from k to p into minute sections,
The adaptability corresponding to the minute section is considered to be the weight applied to the point, and the point at which ky'zx'p is balanced in the horizontal axis direction is set to the representative value, that is, the deterioration degree A.

Next, Δt which is a nondestructive test value of insulation deterioration
A method for obtaining the combined deterioration degree of an δ ₂ ′ + ΔI ′ and Qmax ′ and the combined deterioration degree of the start / stop count N ′ and the operation years Y ′, which is the operation history, will be specifically described. In Figure 7,
The various membership functions when Δtan δ ₂ '+ ΔI' is used as the input data of the conditional part are shown. The membership function of the condition part is composed of no deterioration 3a, small deterioration 3b, deterioration 3c, and large deterioration 3d. Further, in the conclusion part, the deterioration degree is obtained as an output, and the membership function has the same structure as the condition part, that is, no deterioration 3e, deterioration small 3f, deterioration 3g, and deterioration 3g.
It consists of h. In this embodiment, a certain value of (Δtan δ ₂ '+
When ΔI ′) is input, the goodness of fit for the non-deteriorated 3a is 0.2 and the goodness of fit for the small degraded 3b is 0.5. Using these goodnesses of fit, in the conclusion section, the area X indicated by the diagonal lines is obtained as described above.

FIG. 8 shows various membership functions when Qmax 'is used as the input data of the conditional part. Again, in the same way as the processing of (Δtan δ ₂ '+ ΔI'),
A shaded area Y is obtained by the membership function of the condition part and the conclusion part. Area X determined in this way,
Y is treated as a fuzzy rule conformity, and as shown in FIG. 9, a region X for (Δtan δ ₂ '+ ΔI').
And the area Y for Qmax ′ are combined to obtain the combined deterioration degree 3k by the above-described centroid method.

With respect to the N and Y based on the driving history, the fuzzy rule conformity regions Z and W are obtained as shown in FIGS. 10 and 11 according to the same idea, and as shown in FIG. , The synthetic deterioration degree 3 l is obtained.

In the above description, (Δtanδ
₂ '+ △ I'), Qmax ', N', Y'membership functions of each conditional part have common classifications of no deterioration, small deterioration, medium deterioration, and large deterioration, but the function itself is It uses something completely different. Further, in the conclusion section, each obtains the deterioration degree as an output, and uses the same membership function.

The respective synthetic deterioration degrees 3k and 3l are input values of the neural network 4. What is used here is a multi-layered neural network, which operates according to the principle described below. FIG. 13 shows a neuron model forming a neural network. actually,
A plurality of the neurons are combined and output. The connection between neurons is called synaptic connection,
The signal amount between neurons is determined by the strength of the synaptic connection. A neuron can be regarded as a kind of threshold element that activates input values from many synapses above a certain value. A neuron, which is a unit forming a neural network, is modeled as a multi-input, one-output threshold element. The net input signal strength I that determines the behavior to the neuron is given by Ii = ΣWijXj as the weighted sum of the input signal X to this neuron. Here, W is a weight to the signal and corresponds to synaptic connection. When this net input exceeds a certain threshold, it outputs a signal through the output line, which output is further coupled to another neuron. In this embodiment, the sigmoid function shown in FIG. 14 is used as the neuro input / output relational expression, but a binary function of 1,0 may be used. The backpropagation method is effective for learning the neutral network in the present embodiment. In this method, a pattern of an input signal and a teacher signal is given to a neural network, and iterative learning is performed in the opposite direction from the output layer to the input layer so as to reduce the squared error between the output pattern and the actual output pattern. As a result, for the learned pattern,
If the correct output method is obtained and the pattern that is not learned yet is close to the learned pattern,
A value close to the correct output can be obtained due to the generalization ability. In the present embodiment, as the input data, a combined deterioration degree 3k of Δtan δ ₂ '+ ΔI' and Qmax ', a combined deterioration degree 3l of the number of start-stop times N'and the number of years of operation Y'are given, and as an output,
The dielectric breakdown voltage residual rate is obtained. Further, as the teacher signal value, for example, when the insulator is updated, the actual breakdown test is actually performed, and the actual dielectric breakdown voltage residual rate obtained from this result is used.

In this embodiment, the estimation rule of the estimation means is the above-mentioned membership function and the neurons and synapses forming the neutral network 4. Further, the rule correction means is configured by the neutral network 4 having a learning function.

The remaining dielectric breakdown voltage rate thus estimated is used in the remaining life calculation section 5 for remaining life calculation.
The calculation of the remaining life will be described with reference to FIGS. 15 and 16. In FIG. 15, the horizontal axis represents the years of operation, the vertical axis represents the dielectric breakdown voltage residual rate, and in FIG. 16, the horizontal axis represents the dielectric breakdown voltage residual rate and the vertical axis represents the standard deviation. If the estimated breakdown voltage residual rate is estimated to be V _R1 at present,
Assuming that the number of years of operation to date is Y ₁ , the point O ₁ determined by this is plotted in the graph of FIG.
This point O ₁ and the initial value point O ₀ are linearly connected to obtain a curve A. The estimated dielectric breakdown voltage residual rate V _R1 is obtained as an average value, and therefore the curve A means an average transition curve of the dielectric breakdown voltage residual rate. By the way, since it is preferable to manage the equipment at the lowest value in consideration of safety, the 3σ management method is adopted here. As shown in FIG. 16, the standard deviation σ is obtained by preparing a standard deviation characteristic curve from past performance data and referring to it.

The obtained standard deviation value σ is multiplied by 3 and subtracted from V _R1 . Point O corresponding to this value (V _R1 −3σ)
₂ is plotted in the graph of FIG. 15 to obtain a transition curve (A-3σ) similar to the average transition curve A. The transition curve (A-3 [sigma]) obtained in this way becomes a transition curve of the dielectric breakdown voltage residual rate as the minimum value. Here, when the remaining dielectric breakdown voltage is 40%, the operating limit, that is, the life is calculated. The number of years of operation indicated by the intersection O ₃ of the line of the remaining 40% dielectric breakdown voltage and the transition curve (A-3σ) is the life Y. It becomes ₂ . Therefore, by pulling the operating life Y ₁ to the current from this lifetime Y _2, remaining life (Y ₂ -Y ₁₎ it is obtained. Here, in order to simplify the explanation, when drawing the transition curve (A-3σ), only the dielectric breakdown voltage residual rate for the number of years of operation Y ₁ up to the present time is plotted, and this point O ₂ and the initial value are plotted. simply linearly connecting at transition curve and point O ₀ value (A-
3σ) was obtained, but before the number of years of operation Y ₁ up to the present, the estimated residual dielectric breakdown voltage residual rate was also plotted to grasp the transition of the residual dielectric breakdown voltage residual from the past to the present,
The future transition of the dielectric breakdown voltage residual rate may be determined based on this.

As described above, according to the present embodiment, since the data regarding the estimation target is equivalently converted into the estimation reference target, even if the estimation target is changed, the mutual relation of a plurality of factors is taken into consideration as in the prior art. Therefore, it is not necessary to recalculate each constant of the empirical formula by complicated statistical processing, and the dielectric breakdown voltage residual rate can be easily obtained only by the data for equivalent conversion.

Further, in the present embodiment, even if the data of the deterioration factor increases, this can be easily dealt with by increasing the membership function in the fuzzy inference unit 3. Here, the reason why the membership function can be increased relatively easily is that the final conclusion is made in the stage, and the learning function of the neutral network 4 performs the correction work in the process of outputting the value input here. Because it will be seen. Therefore, it is not preferable to combine all of the equivalently converted data by fuzzy inference in order to accurately calculate the residual voltage of the breakdown voltage. At least when the final combination is performed, the neutral network is used. Need to be used.

By the way, even if it is necessary to use the neutral network at least when performing the final synthesis, it is not preferable to carry out all the synthesis processing by the neutral network. This is because the number of synthesizing processes becomes very large and the processing time becomes very long. Therefore, in this embodiment, as the input value to the neutral network 4, only the two values of the deterioration degree 3k and the deterioration degree 3l which have already been subjected to the first-stage combination are used, and the processing speed of the entire apparatus is increased. ..

[0035]

According to the present invention, since the measured data is once converted into the reference data and the deterioration is estimated, even if the insulator to be estimated changes, the rule for estimation is the reference. Since only the data is handled, it is not necessary to change the rule (for example, empirical formula), and it is possible to easily deal with the case where the insulator changes, and it is possible to enhance versatility.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a functional configuration of a remaining life diagnosis apparatus according to an embodiment of the present invention.

FIG. 2 is a circuit block diagram of a remaining life diagnosis apparatus according to an embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a function of a reference value conversion unit according to an embodiment of the present invention.

FIG. 4 is a graph for converting measurement data Qmax into reference data according to an embodiment of the present invention.

FIG. 5 is an explanatory diagram showing a membership function used for general fuzzy inference according to an embodiment of the present invention.

FIG. 6 is an explanatory diagram illustrating a centroid method in a conclusion section of a fuzzy reasoning section according to an embodiment of the present invention.

FIG. 7 shows (Δtan δ ₂ + Δ) according to an embodiment of the present invention.
It is explanatory drawing which shows the membership function used for I).

FIG. 8 is an explanatory diagram showing a membership function used for Qmax according to an embodiment of the present invention.

FIG. 9 shows (Δtan δ ₂ + Δ) according to an embodiment of the present invention.
It is explanatory drawing for demonstrating calculation of the synthetic | combination deterioration degree of I) and Qmax.

FIG. 10 is an explanatory diagram showing a membership function used with respect to the number of times N of activation and suspension according to an embodiment of the present invention.

FIG. 11 is an explanatory diagram showing a membership function used for the driving years Y according to an embodiment of the present invention.

FIG. 12 is an explanatory diagram for explaining calculation of a combined deterioration degree of the number N of start-stops and the number of years of operation Y according to the embodiment of the present invention.

FIG. 13 is an explanatory diagram showing a neuron model according to an embodiment of the present invention.

FIG. 14 is a graph showing a sigmoid function according to an embodiment of the present invention.

FIG. 15 is a graph showing a transition curve of the dielectric breakdown voltage residual ratio of one example according to the present invention.

FIG. 16 is a graph used to obtain a standard deviation σ of the dielectric breakdown voltage residual rate of one example according to the present invention.

FIG. 17 is a graph showing tan δ-voltage characteristics.

FIG. 18 is a graph showing alternating current-voltage characteristics.

FIG. 19 is a graph showing Qmax-voltage characteristics.

[Explanation of symbols]

2 ... Standard value conversion unit, 3 ... Fuzzy inference unit, 4 ... Neutral network, 5 ... Remaining life calculation unit, 11 ... CPU,
12 ... ROM, 13 ... RAM.

Claims (8)

[Claims] 1. A method for estimating deterioration of an insulator, comprising: measuring data as a deterioration index for an insulator to be estimated; and estimating a residual rate of dielectric breakdown voltage of the insulator based on the measured data. A conversion rule for converting the measurement data regarding the target insulator to the reference data regarding the estimation reference target that is a reference when estimating the deterioration, and an estimation rule for estimating the dielectric breakdown voltage residual rate from the reference data, Prepared in advance, obtain the measurement data of the insulator to be estimated, convert the measurement data to the reference data according to the conversion rule, use the reference data obtained by the conversion, according to the estimation rule , A method for estimating deterioration of an insulating material, which comprises estimating a remaining insulation voltage rate. 2. An insulator to be estimated is obtained by measuring at least data including non-destructive test data,
A method for estimating deterioration of insulation breakdown voltage of an insulator based on this measurement data, in a method of estimating deterioration of an insulator, the conversion of converting the measurement data into reference data related to an estimated reference object that serves as a reference when estimating deterioration. Rules and
An estimation rule for estimating the dielectric breakdown voltage residual rate from the reference data is prepared in advance, and the estimation rule is corrected at any time based on at least the actual dielectric breakdown voltage residual rate obtained by the breakdown test. Every other time, the measurement data regarding the insulator to be estimated is acquired, the measurement data is converted into the reference data according to the conversion rule, and the insulation voltage remains according to the estimation rule using the reference data obtained by the conversion. A method for estimating deterioration of an insulator, which comprises estimating a rate. 3. The conversion rule is also provided with a rule for converting the driving history data of the estimation target into the reference data of the estimation reference target, and the estimation rule also includes the reference data of the driving history data. Including, the rule for obtaining the dielectric breakdown voltage residual rate is prepared, as the measurement data, the non-destructive test data and the operation history data is obtained, according to the conversion rule and the estimation rule, the The method for estimating deterioration of an insulator according to claim 2, wherein the insulation voltage residual rate is estimated from the measurement data. 4. An insulation deterioration estimating apparatus for estimating the insulation breakdown voltage residual rate of an insulation based on measurement data as a deterioration index of the insulation to be estimated, according to a conversion rule set in advance. A reference value converting means for converting the measurement data of the insulator to be estimated into reference data of an estimation reference object; and an estimating means for estimating a dielectric breakdown voltage residual rate from the reference data according to a predetermined estimation rule. And a deterioration estimating device for an insulator, comprising: 5. An insulator to be estimated based on measurement data including at least a plurality of non-destructive test data,
In an insulator deterioration estimating device for estimating the insulation breakdown voltage residual ratio of the insulator, a plurality of the measurement data items of the insulator of the estimation target are respectively converted into a plurality of measurement data items of the estimation reference object according to a predetermined conversion rule. Reference value conversion means for converting to reference data, estimation means for estimating the dielectric breakdown voltage residual rate from the plurality of reference data according to a predetermined estimation rule, and actual dielectric breakdown voltage residual rate obtained by the breakdown test And a rule correction unit that corrects the estimation rule based on the actual residual rate of the insulation breakdown voltage. 6. The fuzzy inference means for deriving a deterioration suitability for each of the plurality of reference data based on a plurality of predetermined membership functions and synthesizing the plurality of deterioration suitability values. And a neutral network for obtaining the insulation breakdown voltage residual rate from the combined deterioration suitability, wherein the rule correction means uses at least the actual breakdown voltage residual rate obtained by a breakdown test as a teacher signal. The deterioration estimation device for an insulator according to claim 5, wherein the device is the neutral network. 7. The reference value converting means has conversion rules for converting non-destructive test data and operation history data into reference data related to the estimated reference object, and the estimating means converts the reference data into reference data. 7. The insulation deterioration estimating device according to claim 4, further comprising an estimation rule for obtaining the dielectric breakdown voltage residual rate from the converted non-destructive test data and operation history data. .. 8. A dielectric breakdown according to a transition of operating years, using at least the insulation deterioration estimating device according to claim 4, 5, 6 or 7, and at least the dielectric breakdown voltage residual rate obtained by the deterioration estimating device. A transition curve obtaining unit that obtains a voltage residual rate transition curve, a lifetime calculation unit that obtains a life of an estimation target from the dielectric breakdown voltage residual rate transition curve and a dielectric breakdown voltage residual rate of a predetermined operation limit, and A remaining life diagnosing device for an insulator, comprising: a remaining life calculation means for obtaining a remaining life of the estimation target from the life and the current number of operating years of the estimation target.