Estimation method and estimation program for degradation index of insulating oil

Abstract

To provide a technology that contributes to improving the estimation accuracy of the degradation degree of insulating oil.SOLUTION: Each of correlation coefficient (correlation spectrum) between the absorbance spectrum (second derivative spectrum) and furfural content for each insulating oil sample (sampled oil, thermally accelerated degraded oil, additive oil) is calculated. A PLS regression is performed by using the absorbance at the wavenumber common to each insulating oil sample (sampled oil, thermally accelerated degraded oil, additive oil) of the absorbance for which the correlation coefficient exceeds a threshold value as an explanatory variable to generate a formula for estimating the furfural content. The generated formula for estimating the furfural content is then used to estimate the furfural content of insulating oil samples for which the furfural content is unknown. This allows us to exclude information on components that are noise (various components other than furfural) and to estimate the furfural content accurately without it being dependent on the type of oil.SELECTED DRAWING: Figure 12

Description

[Technical field]

[0001] The present invention relates to a method and a program for estimating a deterioration index of insulating oil, which estimates a deterioration index that is an index showing the degree of deterioration of insulating oil. [Background technology]

[0002] Japanese Patent Application Laid-Open Publication No. 2000-105231 (Patent Document 1) describes a method for managing unknown thermal transfer oils, in which the near-infrared spectrum of a thermal transfer oil with multiple known properties and performance is measured, multivariate analysis is performed to obtain a correlation between the properties and performance and the near-infrared spectrum, and then the near-infrared spectrum of an unknown thermal transfer oil is measured, and the performance, life, etc. of the unknown thermal transfer oil are estimated from the measurement results.

[0003] This method of managing unknown thermal transfer oils makes it possible to easily estimate the physical properties and performance of the thermal transfer oil to be managed by measuring the near-infrared spectrum, and therefore makes it possible to estimate the physical properties and performance of the thermal transfer oil without having to perform complex measurements of the physical properties and performance, and to evaluate the degree of deterioration based on this. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2000-105231 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the management method of unknown thermal transfer oil described in the above publication does not mention which wavelength of absorbance in the near-infrared spectrum of the measured thermal transfer oil should be used in the multivariate analysis. If the absorbance of a wavelength unrelated to the physical properties or performance indicating the degree of deterioration of the thermal transfer oil is included in the multivariate analysis, the absorbance of the wavelength unrelated to the physical properties or performance indicating the degree of deterioration may become noise when estimating the degree of deterioration, which may reduce the accuracy of estimating the degree of deterioration of the thermal transfer oil.

[0006] The present invention has been made in view of the above, and has as its main object to provide a technique that contributes to improving the accuracy of estimating the deterioration degree of insulating oil. [Means for solving the problem]

[0007] According to a preferred embodiment of the method for estimating a deterioration index of insulating oil according to the present invention, there is provided a method for estimating a deterioration index, which is an index showing the degree of deterioration of insulating oil, comprising the steps of: (a) acquiring, for a plurality of insulating oil samples, the deterioration index and an absorbance spectrum showing a change in absorbance for each wave number or wavelength; (b) calculating a correlation coefficient between the acquired deterioration index and the acquired absorbance spectrum for each of the plurality of insulating oil samples; (c) selecting, as explanatory variables, a plurality of absorbances at the wave number or the wavelength where the correlation coefficient exceeds a threshold value in the acquired absorbance spectrum; (d) using the deterioration index as a target variable, performing PLS regression using the plurality of explanatory variables to generate an estimation equation capable of estimating the deterioration index of the insulating oil; and (e) using the generated estimation equation to estimate the deterioration index of the insulating oil, the deterioration index of which is unknown.

[0008] According to the present invention, the correlation coefficient between the degradation index and the absorbance spectrum obtained for a plurality of insulating oil samples is calculated, and the absorbance at the wave number or wavelength where the correlation coefficient exceeds the threshold is used as an explanatory variable, that is, the absorbance at the wave number or wavelength closely related to the physical properties or performance indicating the degradation degree is used as an explanatory variable, so that the absorbance at the wave number or wavelength band that becomes noise when estimating the degradation degree is not included in the explanatory variable. This makes it possible to improve the accuracy of estimating the degradation degree of the insulating oil. Here, in the present invention, "the correlation coefficient exceeds the threshold" is defined as a mode in which the correlation coefficient is larger than the threshold when the correlation coefficient is a positive value, and is defined as a mode in which the correlation coefficient is smaller than the threshold when the correlation coefficient is a negative value.

[0009] According to a further aspect of the method for estimating a deterioration degree of insulating oil according to the present invention, step (a) is a step of calculating a second derivative spectrum by subjecting the absorbance spectrum to second derivative processing, and acquiring the second derivative spectrum as the absorbance spectrum.

[0010] According to this embodiment, it is possible to easily distinguish the magnitude of the absorbance peak by separating the absorbance peaks in the acquired absorbance spectrum. In addition, since it is possible to correct the baseline of the absorbance in the acquired absorbance spectrum, it is possible to reduce the number of principal components (factors) for constructing an estimation equation by PLS regression.

[0011] According to a further embodiment of the method for estimating the degree of deterioration of insulating oil according to the present invention, step (c) includes a step of determining a change in a regression coefficient of determination, which is a correlation coefficient between actual deterioration index values, which are deterioration indexes obtained for a plurality of insulating oil samples, and estimated deterioration index values, which are deterioration indexes for a plurality of insulating oil samples estimated using the estimation equation, when the threshold value is changed, and determining a change in the regression coefficient of determination when the number of explanatory variables is changed, selecting a maximum regression coefficient of determination, which is a regression coefficient of determination that shows a maximum value when the number of explanatory variables is equal to or less than a predetermined number, and using a threshold value corresponding to the maximum regression coefficient of determination as a threshold value for selecting explanatory variables.

[0012] As a result of intensive research, the inventors have found that the accuracy of estimating the deterioration level of insulating oil can be improved by appropriately selecting the threshold value and the number of explanatory variables. Based on the results of this research, in this embodiment, the change in the regression determination coefficient when the threshold value is changed and the change in the regression determination coefficient when the number of explanatory variables is changed are calculated, and the regression determination coefficient that shows the maximum value when the number of explanatory variables is equal to or less than a predetermined number is selected, and the threshold value corresponding to the regression determination coefficient is adopted as the threshold value used to select the absorbance at a wave number or wavelength that is closely related to the physical properties or performance that indicate the deterioration level. This makes it possible to further improve the accuracy of estimating the deterioration level of insulating oil.

[0013] According to a further embodiment of the method for estimating the degree of deterioration of insulating oil according to the present invention, step (a) is a step of acquiring a deterioration index and an absorbance spectrum of each of the collected oil and the simulated deteriorated oil, using, as insulating oil samples, a collected oil collected from an insulating oil actually used in an equipment for the purpose of cooling and / or insulating the equipment, and a simulated deteriorated oil artificially generated to simulate deterioration due to use in the equipment. Step (b) is a step of calculating a first correlation coefficient, which is a correlation coefficient between the deterioration index of the collected oil and the absorbance spectrum of the collected oil, and calculating a second correlation coefficient, which is a correlation coefficient between the deterioration index of the simulated deteriorated oil and the absorbance spectrum of the simulated deteriorated oil. Step (c) is a step of selecting, as a first explanatory variable, a plurality of absorbances at wavenumbers or wavelengths at which the first correlation coefficient exceeds a first threshold value, selecting, as a second explanatory variable, a plurality of absorbances at wavenumbers or wavelengths at which the second correlation coefficient exceeds a second threshold value, and selecting, as an explanatory variable, an absorbance at a wavenumber or wavelength common between the first explanatory variable and the second explanatory variable. Here, in the present invention, "the first correlation coefficient exceeded the first threshold" and "the second correlation coefficient exceeded the second threshold" are defined as a case in which the first correlation coefficient is greater than the first threshold and the second correlation coefficient is greater than the second threshold when the first and second correlation coefficients are positive values, and a case in which the first correlation coefficient is smaller than the first threshold and the second correlation coefficient is smaller than the second threshold when the first and second correlation coefficients are negative values.

[0014] In actual equipment, various insulating materials (insulating oil, insulating paper, insulating varnish, coverings for electric wires and cables, windings, etc.) are used, and these multiple insulating materials are involved in the deterioration of insulating oil in a complex manner, and environmental noise caused by various usage environments also has an effect, so there is a possibility that various components other than the components that indicate the deterioration index are generated in a complex manner. In this case, the explanatory variable (first explanatory variable) set using the collected oil may contain a wavenumber band or wavelength band that is not related to the components that indicate the deterioration index, and if PLS ​​regression is performed using only this explanatory variable (first explanatory variable), it may act as noise when generating an estimation equation that can estimate the deterioration index, reducing the estimation accuracy of the deterioration index. According to this embodiment, the absorbance at a wave number or wavelength common between the first explanatory variable set using the collected oil, which is an insulating oil containing components other than the components showing the deterioration index in addition to the components showing the deterioration index, and the second explanatory variable set using the simulated deteriorated oil in which the components showing the deterioration index are dominant is selected as an explanatory variable, in other words, the absorbance at a wave number or wavelength having a high correlation coefficient (correlation coefficient between the deterioration index and the absorbance spectrum) common to the collected oil and the simulated deteriorated oil is selected as an explanatory variable, so that it is possible to select an explanatory variable effective for estimating the deterioration index of insulating oil used in an actual facility (absorbance at a wave number or wavelength effective for estimating the deterioration index) while effectively avoiding the above-mentioned inconvenience, that is, the inclusion of absorbance at a wave number or wavelength band unrelated to the components showing the deterioration index in the explanatory variable. In other words, it is possible to select absorbance at a wave number or wavelength correlated with the deterioration index as an explanatory variable while excluding information on noise components (various components other than the components showing the deterioration index).

[0015] According to a further embodiment of the method for estimating the degree of deterioration of insulating oil according to the present invention, step (c) includes the steps of: determining a change in the collected oil regression coefficient of determination when the first threshold value is changed, and determining a change in the collected oil regression coefficient of determination when the number of the first explanatory variables is changed, selecting a collected oil maximum regression coefficient of determination, which is the collected oil regression coefficient of determination that shows the maximum value when the number of the first explanatory variables is equal to or less than a first predetermined number, and using the first threshold value corresponding to the collected oil maximum regression coefficient of determination as a first threshold value for selecting the first explanatory variable; and determining a change in the simulated degraded oil regression coefficient of determination when the second threshold value is changed, and determining a change in the simulated degraded oil regression coefficient of determination when the number of the second explanatory variables is changed, selecting a simulated degraded oil maximum regression coefficient of determination, which is the simulated degraded oil regression coefficient of determination that shows the maximum value when the number of the second explanatory variables is equal to or less than a second predetermined number, and using the second threshold value corresponding to the simulated degraded oil maximum regression coefficient of determination as a second threshold value for selecting the second explanatory variable. The regression coefficient of determination for sampled oil is the correlation coefficient between the actual value of the sampled oil degradation index, which is a degradation index obtained for sampled oil, and the estimated value of the sampled oil degradation index, which is a degradation index for sampled oil estimated using the estimation formula. The actual value of the simulated degraded oil degradation index is the correlation coefficient between the actual value of the simulated degraded oil degradation index, which is a degradation index obtained for simulated degraded oil, and the estimated value of the simulated degraded oil degradation index, which is a degradation index for simulated degraded oil estimated using the estimation formula.

[0016] As a result of intensive research, the inventors have found that the accuracy of estimating the deterioration level of insulating oil is improved by appropriately selecting the threshold value and the number of explanatory variables. Based on the results of this research, in this embodiment, the change in the collected oil regression determination coefficient when the first threshold value is changed, the change in the collected oil regression determination coefficient when the number of the first explanatory variables is changed, and are calculated, and the collected oil regression determination coefficient that shows the maximum value when the number of the first explanatory variables is equal to or less than a first predetermined number is selected, and the first threshold value corresponding to the collected oil regression determination coefficient is adopted as the first threshold value used for selecting the absorbance at a wave number or wavelength closely related to the physical properties and performance that indicate the deterioration level, and the change in the simulated deteriorated oil regression determination coefficient when the second threshold value is changed, and the change in the simulated deteriorated oil regression determination coefficient when the number of the second explanatory variables is changed, and are calculated, and the simulated deteriorated oil regression determination coefficient that shows the maximum value when the number of the second explanatory variables is equal to or less than a second predetermined number is selected, and the second threshold value corresponding to the simulated deteriorated oil regression determination coefficient is adopted as the second threshold value used for selecting the absorbance at a wave number or wavelength closely related to the physical properties and performance that indicate the deterioration level. This makes it possible to further improve the accuracy of estimating the deterioration level of the insulating oil.

[0017] According to a further embodiment of the method for estimating the degree of deterioration of insulating oil according to the present invention, step (a) is a step of acquiring the deterioration index and the absorbance spectrum of each of the thermally accelerated deteriorated oil and the deterioration indicator component-added oil, which are obtained by adding a component showing a deterioration index to a new insulating oil, as simulated deteriorated oils by heating the new insulating oil under an environment simulating a state of use in a facility, and the deterioration indicator component-added oil, which are obtained by adding a component showing a deterioration index to the new insulating oil, respectively, instead of the second correlation coefficient. Also, step (b) is a step of calculating a third correlation coefficient, which is a correlation coefficient between the deterioration index of the thermally accelerated deteriorated oil and the absorbance spectrum of the thermally accelerated deteriorated oil, and a fourth correlation coefficient, which is a correlation coefficient between the deterioration index of the deterioration indicator component-added oil and the absorbance spectrum of the deterioration indicator component-added oil, respectively, instead of the second correlation coefficient. Then, step (c) is a step of selecting, instead of the second explanatory variable, a plurality of absorbances at wavenumbers or wavelengths at which the third correlation coefficient exceeds a third threshold as third explanatory variables, selecting a plurality of absorbances at wavenumbers or wavelengths at which the fourth correlation coefficient exceeds a fourth threshold as fourth explanatory variables, and selecting, as explanatory variables, absorbances at wavenumbers or wavelengths common to the first explanatory variable, the third explanatory variable, and the fourth explanatory variable.

[0018] According to this embodiment, two types of simulated deteriorated oil are used: a thermally accelerated deteriorated oil artificially produced so that components indicating a deterioration index are dominant, and an additive oil artificially produced so that it contains only components indicating a deterioration index. Therefore, it is possible to more effectively select the absorbance (explanatory variable) at a wavenumber or wavelength that is correlated with the deterioration index while excluding information on noise components (various components other than the components indicating the deterioration index).

[0019] According to a further embodiment of the method for estimating the degree of deterioration of insulating oil according to the present invention, step (c) includes the steps of: determining a change in the heat-accelerated deteriorated oil regression coefficient of determination when the third threshold value is changed; determining a change in the heat-accelerated deteriorated oil regression coefficient of determination when the number of the third explanatory variables is changed; selecting a heat-accelerated deteriorated oil maximum regression coefficient of determination, which is the heat-accelerated deteriorated oil regression coefficient of determination that shows the maximum value when the number of the third explanatory variables is equal to or less than a third predetermined number; and using the third threshold value corresponding to the heat-accelerated deteriorated oil maximum regression coefficient of determination as the third threshold value for selecting the third explanatory variable; and determining a change in the deterioration indicator component added oil regression coefficient of determination when the fourth threshold value is changed; determining a change in the deterioration indicator component added oil regression coefficient of determination when the number of the fourth explanatory variables is changed; selecting a deterioration indicator component added oil maximum regression coefficient of determination, which is the deterioration indicator component added oil regression coefficient of determination that shows the maximum value when the number of the fourth explanatory variables is equal to or less than a fourth predetermined number; and using the fourth threshold value corresponding to the deterioration indicator component added oil maximum regression coefficient of determination as the fourth threshold value for selecting the fourth explanatory variable. The coefficient of determination of the thermally accelerated degraded oil regression is the correlation coefficient between the actual value of the thermally accelerated degraded oil degradation index, which is a degradation index obtained for the thermally accelerated degraded oil, and the estimated value of the thermally accelerated degraded oil degradation index, which is a degradation index of the thermally accelerated degraded oil estimated using the estimation formula. The coefficient of determination of the degradation indicator component-added oil regression is the correlation coefficient between the actual value of the degradation indicator component-added oil degradation index, which is a degradation index obtained for the degradation indicator component-added oil, and the estimated value of the degradation indicator component-added oil degradation index, which is a degradation index of the degradation indicator component-added oil estimated using the estimation formula.

[0020] As a result of extensive research, the present inventors have found that the accuracy of estimating the deterioration level of insulating oil can be improved by appropriately selecting the threshold value and the number of explanatory variables. Based on the results of this research, in this embodiment, the change in the coefficient of determination of the thermally accelerated aging oil regression when the third threshold is changed, and the change in the coefficient of determination of the thermally accelerated aging oil regression when the number of the third explanatory variables is changed are obtained, and the coefficient of determination of the thermally accelerated aging oil regression that shows the maximum value when the number of the third explanatory variables is equal to or less than a third predetermined number is selected, and the third threshold corresponding to the coefficient of determination of the thermally accelerated aging oil regression is adopted as the third threshold used for selecting the absorbance at a wave number or wavelength closely related to the physical properties and performance that indicate the degree of deterioration, and the change in the coefficient of determination of the regression of the deterioration indicator component-added oil when the fourth threshold is changed, and the change in the coefficient of determination of the regression of the deterioration indicator component-added oil when the number of the fourth explanatory variables is changed are obtained, and the coefficient of determination of the regression of the deterioration indicator component-added oil that shows the maximum value when the number of the fourth explanatory variables is equal to or less than a fourth predetermined number is selected, and the fourth threshold corresponding to the coefficient of determination of the regression of the deterioration indicator component-added oil is adopted as the fourth threshold used for selecting the absorbance at a wave number or wavelength closely related to the physical properties and performance that indicate the degree of deterioration. This makes it possible to further improve the accuracy of estimating the deterioration degree of insulating oil.

[0021] According to a further embodiment of the method for estimating the deterioration degree of insulating oil of the present invention, the deterioration index is the furfural content. Step (a) is a step of acquiring an absorbance spectrum by using infrared spectroscopy. Step (c) is a step of acquiring an absorbance spectrum by using infrared spectroscopy at a wave number of 2800 cm. -1 or 2900cm -1 , and / or 1650cm -1 or 1750cm -1 This is the step of selecting multiple absorbances at 2800 cm as explanatory variables. -1 or 2900cm -1 , and / or 1650cm -1 or 1750cm -1 is preferable, and more preferably, the wave number is 2810 cm -1 Or 2830cm -1 , and / or 1705cm -1 Or 1725cm -1Here, the term "infrared spectroscopy" in the present invention is a concept including not only near-infrared spectroscopy but also mid-infrared spectroscopy.

[0022] According to the present embodiment, the absorbance spectrum is obtained using infrared spectroscopy, which has higher sensitivity than other spectroscopy methods and is less affected by the measurement environment, so that the furfural content, which is one of the deterioration indicators of insulating oil, can be estimated more effectively.

[0023] According to a preferred embodiment of the insulating oil degradation index estimation program of the present invention, the program is for estimating a degradation index, which is an index showing the degree of degradation of insulating oil, and causes a computer to execute each step of the degradation index estimation method of the present invention according to any of the above-mentioned aspects. The program may be recorded on a computer-readable recording medium, such as a hard disk, a ROM, an SSD, a flash memory (USB memory, SD card, etc.), a floppy disk, a CD, a DVD, etc., or may be distributed from one computer to another computer via a transmission medium, such as a communication network such as the Internet or a LAN, or may be transmitted and received in any other manner.

[0024] According to the present invention, it is possible to achieve the same effects as those achieved by the deterioration index estimation method of the present invention in any of the above-mentioned aspects, such as the effect of being able to improve the accuracy of estimating the deterioration degree of insulating oil. Effect of the Invention

[0025] According to the present invention, it is possible to improve the accuracy of estimating the deterioration degree of insulating oil. [Brief description of the drawings]

[0026] [Figure 1] FIG. 1 is an explanatory diagram showing an example of the configuration of a degradation index estimation device 1 according to an embodiment of the present invention. [Diagram 2] FIG. 2 is an explanatory diagram showing a correlation coefficient spectrum between a second derivative spectrum and a furfural content measured in an insulating sample oil. [Diagram 3] FIG. 13 is an explanatory diagram showing the relationship between a threshold value and a regression determination coefficient r2. [Figure 4] FIG. 13 is an explanatory diagram showing the relationship between the number of explanatory variables and the regression determination coefficient r2. [Diagram 5] FIG. 2 is an explanatory diagram showing the correlation between the actual measured furfural content in collected oil and the estimated furfural content. [Figure 6] FIG. 2 is an explanatory diagram showing the correlation between the actual measured furfural content and the estimated furfural content in thermally accelerated deteriorated oil. [Figure 7] FIG. 2 is an explanatory diagram showing the correlation between the actual measured furfural content in added oil and the estimated furfural content. [Figure 8] FIG. 1 is an explanatory diagram showing standard partial regression coefficients of explanatory variables in collected oil. [Figure 9] FIG. 1 is an explanatory diagram showing standard partial regression coefficients of explanatory variables used in PLS regression for thermally accelerated aged oil. [Figure 10] FIG. 1 is an explanatory diagram showing standard partial regression coefficients of explanatory variables used in PLS regression in added oil. [Figure 11] 1 is an explanatory diagram showing the correlation between the actual measured value of furfural content and the estimated value of furfural content calculated using a furfural content estimation formula. FIG. [Figure 12] FIG. 13 is an explanatory diagram showing standard partial regression coefficients of explanatory variables used. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] Next, the best mode for carrying out the present invention will be described with reference to examples. EXAMPLES

[0028] As shown in FIG. 1, the deterioration index estimation device 1 for estimating the deterioration degree of insulating oil according to the present embodiment includes a holder 4 capable of holding a container 2, a light source 6 connected to the holder 4, a spectroscopic sensor 8 connected to the holder 4, and a personal computer (hereinafter referred to as "PC") 10 connected to the spectroscopic sensor 8. The container 2 can contain insulating oil or insulating sample oil. The light source 6 can project light through the holder 4 onto the container 2 containing the insulating oil or insulating sample oil. The spectroscopic sensor 8 detects the light projected onto and transmitted through the container 2 to measure the absorbance spectrum of the insulating oil or insulating sample oil. For example, an infrared spectroscopic sensor or a near-infrared spectroscopic sensor can be used as the spectroscopic sensor 8. If a near-infrared spectroscopic sensor is used as the spectroscopic sensor 8, the deterioration index estimation device 1 can be realized at a lower cost. The PC 10 calculates a second derivative spectrum by subjecting the absorbance spectrum of the insulating oil or insulating sample oil measured by the spectroscopic sensor 8 to second derivative processing, and estimates the furfural content, which is an index of deterioration of the insulating oil or insulating sample oil, using the calculated second derivative spectrum. In this embodiment, the furfural content, which is an important index of deterioration of the insulating oil or insulating sample oil, is used to estimate the degree of deterioration of the insulating oil or insulating sample oil, but this is not limiting, and the moisture content or acid value may also be used. The furfural content is estimated by performing partial least squares regression (hereinafter referred to as "PLS regression") using the second derivative spectrum actually measured in the insulating oil sample, more specifically, the second derivative value of absorbance at each wave number as the explanatory variable, and the furfural content actually measured in the insulating oil sample as the objective variable, to generate a furfural content estimation equation (a regression equation for the objective variable (furfural content)), and using the furfural content estimation equation and the second derivative spectrum of the insulating oil whose furfural content is unknown, the furfural content of the insulating oil whose furfural content is unknown is estimated. Details of the generation of the furfural content estimation equation will be described later.

[0029] [Actual measurement of furfural content] When generating the furfural content estimation formula, it is necessary to measure the furfural content in the insulating sample oil. The furfural content was measured by using high-performance liquid chromatography (HPLC), which is a conventional analytical method. In this embodiment, the insulating sample oil was configured to use collected oil, heat-accelerated aged oil, and added oil. The collected oil was actually used insulating oil collected from an oil-filled transformer (voltage 6.6 kV, capacity 30 to 1500 kVA) installed in a high-voltage power receiving equipment of a commercial facility, and 76 samples were prepared. The heat-accelerated aged oil was artificially aged insulating oil by placing a material simulating the contents of an oil-filled transformer in a headspace bottle, placing it in a constant temperature dryer (ONW-300S manufactured by AS ONE Corporation), and heating it at 90°C, and 6 samples were prepared. Furthermore, the additive oil is an insulating oil obtained by mixing new oil (high-voltage insulating oil A manufactured by JXTG Energy Corporation) with furfural (manufactured by Nacalai Tesque, Inc., purity 98%, CAS (registered trademark) registration number: 98-01-1) at concentrations of 0.4%, 0.8%, 1.2%, 1.6%, and 2.0%, and five samples were prepared. The thermally accelerated aging oil and the additive oil are examples of an embodiment corresponding to the "simulated aging oil" in the present invention.

[0030] [Absorbance spectrum measurement] The absorbance spectrum of the insulating sample oil (collected oil, heat accelerated aging oil, added oil) was measured at a wave number of 550 cm for each insulating sample oil (collected oil, heat accelerated aging oil, added oil). -1 ~4000cm -1 Within 2cm -1This was done by measuring every 100 minutes. In this embodiment, the measured absorbance spectrum was subjected to a second derivative process to calculate a second derivative spectrum, and the calculated second derivative spectrum was used as an explanatory variable to perform PLS regression. By using the second derivative spectrum as an explanatory variable for PLS regression, it is possible to separate the absorbance peaks in the acquired absorbance spectrum, making it easier to distinguish the magnitude of the absorbance peaks, and it is also possible to correct the baseline of the absorbance in the acquired absorbance spectrum, so that the number of principal components (factors) for generating the furfural content estimation formula by PLS regression can be reduced. Hereinafter, the second derivative spectrum may be simply referred to as the "absorbance spectrum", but the second derivative spectrum is still used for PLS regression and furfural content estimation.

[0031] [Selection of explanatory variables] In actual equipment, various insulating materials (insulating oil, insulating paper, insulating varnish, coverings of electric wires and cables, windings, etc.) are used, and these multiple insulating materials are involved in the deterioration of insulating oil in a complex manner. Therefore, it is presumed that in the deterioration process of insulating oil, various components are generated in a complex manner in addition to furfural, which is an important deterioration indicator of insulating oil. In addition, when generating a furfural content estimation formula, if the absorbance in a wave number band that is not related to the generation of furfural is selected as an explanatory variable, the explanatory variable may become noise in the estimation of the furfural content, leading to a decrease in the estimation accuracy of the furfural content. Therefore, in this embodiment, the correlation coefficients between the absorbance spectrum (second derivative spectrum) of the insulating sample oil and the furfural content are calculated, and the absorbance at which the correlation coefficient exceeds a threshold value is selected as an explanatory variable. In other words, the absorbance in a wave number band that is thought to be related to the generation of furfural is selected as an explanatory variable. In this embodiment, when generating the furfural content estimation equation, a so-called hold-out method was adopted, in which 62 of the 76 collected oil samples were used to select explanatory variables and the remaining 14 samples were used to verify the furfural content estimation equation.

[0032] Here, the threshold value of the correlation coefficient must be determined to an optimal value while checking the estimation accuracy of the furfural content. This is because, generally, if the threshold value is set too small (i.e., close to a value of -1), the number of explanatory variables used in the PLS regression decreases, and the estimation accuracy of the furfural content decreases, whereas if the threshold value is set too large (i.e., close to a value of 0), the number of explanatory variables used in the PLS regression increases, and information unrelated to furfural (absorbance) is used as an explanatory variable, resulting in a decrease in the estimation accuracy of the furfural content.

[0033] In this embodiment, the threshold value was determined and the explanatory variables were selected in the following procedure. That is, as shown in FIG. 3, the threshold value and the regression coefficient of determination r, which represents the estimation accuracy of the furfural content, were determined for each oil (collected oil, heat-accelerated deteriorated oil, and added oil). 2 Figure 3(a) shows the relationship between the threshold and the regression coefficient of determination r 2 FIG. 3(b) is an explanatory diagram showing the relationship between the threshold value and the regression coefficient of determination r 2 Fig. 3(c) shows the relationship between the threshold value and the regression coefficient of determination r 2 Here, the regression coefficient of determination r 2 can be calculated by performing PLS regression using the absorbance at wavenumbers exceeding a threshold as explanatory variables in the correlation coefficient spectrum (see Figure 2) between the second derivative spectrum and the furfural content measured for each oil (collected oil, heat-accelerated deteriorated oil, and added oil) to generate a furfural content estimation equation, and then taking the correlation between the furfural content measured by HPLC (actual measured value of furfural content) and the estimated value of furfural content calculated using the furfural content estimation equation. In other words, the regression determination coefficient r 2can be said to be the correlation coefficient between the measured value of furfural content and the estimated value of furfural content. Note that the amount of change in the second derivative spectrum (second derivative value) of absorbance due to an increase in the amount of substances contained in each oil (collected oil, heat-accelerated aged oil, added oil) generally appears as a peak value in the negative direction, so in this embodiment, the absorbance at the wavenumber where the correlation coefficient spectrum between the second derivative spectrum and the furfural content in each oil (collected oil, heat-accelerated aged oil, added oil) is below the threshold value was selected as the explanatory variable. In this embodiment, the threshold value was changed in the range of 0 to 1 to obtain the regression determination coefficient r 2 By calculating the threshold and the regression coefficient of determination r 2(See FIG. 3.) The correlation coefficient spectrum between the second derivative spectrum and the furfural content measured with the collected oil corresponds to the "correlation coefficient" and "first correlation coefficient" in the present invention, the correlation coefficient spectrum between the second derivative spectrum and the furfural content measured with the thermally accelerated deteriorated oil corresponds to the "correlation coefficient," "second correlation coefficient," and "third correlation coefficient" in the present invention, and the correlation coefficient spectrum between the second derivative spectrum and the furfural content measured with the added oil corresponds to the "correlation coefficient," "second correlation coefficient," and "fourth correlation coefficient" in the present invention. In addition, the actual measured value of the furfural content in the harvested oil corresponds to the ``actual measured value of the deterioration index'' and ``actual measured value of the harvested oil deterioration index'' in the present invention, the actual measured value of the furfural content in the thermally accelerated degraded oil corresponds to the ``actual measured value of the deterioration index,'' ``actual measured value of the simulated degraded oil deterioration index,'' and ``actual measured value of the thermally accelerated degraded oil deterioration index'' in the present invention, and the actual measured value of the furfural content in the added oil is an example of an embodiment configuration corresponding to the ``actual measured value of the deterioration index,'' ``actual measured value of the simulated degraded oil deterioration index,'' and ``actual measured value of the oil deterioration index added with deterioration index component'' in the present invention. Furthermore, the estimated furfural content in the harvested oil corresponds to the ``deterioration index estimated value'' and ``harvested oil deterioration index estimated value'' in the present invention, the estimated furfural content in the thermally accelerated degraded oil corresponds to the ``degradation index estimated value,'' ``simulated degraded oil deterioration index estimated value,'' and ``thermally accelerated degraded oil deterioration index estimated value'' in the present invention, and the estimated furfural content in the added oil is an example of an implementation configuration corresponding to the ``degradation index estimated value,'' ``simulated degraded oil deterioration index estimated value,'' and ``degradation index component added oil deterioration index estimated value'' in the present invention.

[0034] Next, we calculate the threshold and the regression coefficient of determination r 2 Using the relationship between the above, the threshold value used to select the absorbance at the wavenumber related to the production of furfural is determined. Specifically, in FIG. 3, the largest regression coefficient of determination r 2The threshold value corresponding to was used to select the absorbance at the wavenumber related to the production of furfural. As described above, if the number of explanatory variables used in the PLS regression is too small or too large, the estimation accuracy of the furfural content decreases, so attention should be paid to the number of explanatory variables. In this example, as shown in FIG. 4, the relationship between the number of explanatory variables and the regression coefficient of determination r 2 The relationship between the number of explanatory variables and the largest regression coefficient of determination (r) is found when the number of explanatory variables is 100 or less. 2 Select the regression coefficient of determination r 2 The threshold value corresponding to the above was adopted as the threshold value used to select the absorbance at the wavenumber related to the generation of furfural. Specifically, for the collected oil, the threshold value is -0.5 as shown in Figures 3(a) and 4(a), for the heat-accelerated deteriorated oil, the threshold value is -0.7 as shown in Figures 3(b) and 4(b), and for the spiked oil, the threshold value is -0.998 as shown in Figures 3(c) and 4(c). Figure 4(a) shows the relationship between the number of explanatory variables in the collected oil and the regression coefficient of determination r 2 FIG. 4(b) shows the relationship between the number of explanatory variables and the regression coefficient of determination r 2 Fig. 4(c) shows the relationship between the number of explanatory variables in the additive oil and the regression coefficient of determination r 2 and the threshold value -0.5 corresponds to the "first threshold" in the present invention, the threshold value -0.7 corresponds to the "second threshold" and "third threshold" in the present invention, and the threshold value -0.998 is an example of an embodiment corresponding to the "second threshold" and "fourth threshold" in the present invention. Furthermore, a value of 100 or less, which is the number of explanatory variables, is an example of an embodiment corresponding to the "predetermined number", the "first predetermined number", the "second predetermined number", the "third predetermined number", and the "fourth predetermined number" in the present invention. Furthermore, the regression coefficient of determination r 2 corresponds to the "regression coefficient of determination of collected oil" in the present invention, and the regression coefficient of determination r 2 corresponds to the "simulated deteriorated oil regression coefficient of determination" and the "thermally accelerated deteriorated oil regression coefficient of determination" in the present invention, and the regression coefficient of determination r 2is an example of an embodiment corresponding to the "simulated deteriorated oil regression coefficient of determination" and the "deterioration index component added oil regression coefficient of determination" in the present invention. In addition, the regression determination coefficient r 2 corresponds to the "maximum regression coefficient of determination" and the "maximum regression coefficient of determination for collected oil" in the present invention, and is the regression coefficient of determination r that shows the maximum value when the number of explanatory variables is 100 or less in the thermally accelerated deteriorated oil. 2 corresponds to the "maximum regression coefficient of determination", "maximum regression coefficient of determination for simulated deteriorated oil" and "maximum regression coefficient of determination for thermally accelerated deteriorated oil" in the present invention, and is the regression coefficient of determination r that shows the maximum value when the number of explanatory variables is 100 or less in the additive oil. 2 is an example of an embodiment corresponding to the "maximum regression coefficient of determination", "maximum regression coefficient of determination for simulated deteriorated oil" and "maximum regression coefficient of determination for oil containing deterioration indicator components" in the present invention.

[0035] Then, for each oil (collected oil, heat-accelerated aged oil, and added oil), the absorbance was selected using the thresholds determined above (threshold value -0.5 for collected oil, threshold value -0.7 for heat-accelerated aged oil, and threshold value -0.998 for added oil) as explanatory variables to perform PLS regression to generate an estimation equation for the furfural content. The regression coefficient of determination r 2 And obtain the standard partial regression coefficient of the explanatory variable used in the PLS regression. Figure 5 is an explanatory diagram showing the correlation between the actual value of the furfural content in the collected oil and the estimated value of the furfural content, Figure 6 is an explanatory diagram showing the correlation between the actual value of the furfural content in the heat-accelerated deterioration oil and the estimated value of the furfural content, and Figure 7 is an explanatory diagram showing the correlation between the actual value of the furfural content in the added oil and the estimated value of the furfural content. Also, Figure 8 is an explanatory diagram showing the standard partial regression coefficient of the explanatory variable in the collected oil, Figure 9 is an explanatory diagram showing the standard partial regression coefficient of the explanatory variable used in the PLS regression in the heat-accelerated deterioration oil, and Figure 10 is an explanatory diagram showing the standard partial regression coefficient of the explanatory variable used in the PLS regression in the added oil.

[0036] In Figures 5 to 7, the more points are plotted on the diagonal line, the smaller the error between the measured value of furfural content and the estimated value of furfural content. As shown in Figures 5 to 7, the highest regression coefficient of determination r 2 (value 0.7804). In actual equipment (e.g., oil-immersed transformers installed in high-voltage power receiving equipment at commercial facilities), furfural is produced from cellulose, a constituent of insulating paper, but there are multiple reaction pathways, which involve complex structural changes of organic components (functional groups). Therefore, a furfural content estimation equation generated by performing PLS regression using an additive oil in which only a furfural reagent is added to new oil may not be a general estimation equation that can be applied to various actual equipment. From this perspective, the furfural content estimation equation generated by performing PLS regression using oil collected from actual equipment should be considered as the standard, but on the other hand, if a furfural content estimation equation is generated by performing PLS regression using only information on collected oil collected from equipment in various environments, it will include explanatory variables specific to the equipment that do not contribute to furfural, and as a result, it may not be possible to generate a furfural content estimation equation with high accuracy. Therefore, the inventors of the present invention considered that in order to generate a furfural content estimation formula that is versatile and capable of estimating the furfural content with high accuracy, it would be effective to combine information obtained from oil sampled from actual equipment with information obtained from artificially created heat-accelerated aged oil and added oil that are less susceptible to disturbances and are specialized for increasing furfural, and thus adopted a configuration in which the absorbance at a wavenumber common to each oil (sampled oil, heat-accelerated aged oil, added oil) is selected as an explanatory variable. Specifically, as shown in Figures 8 to 10, the absorbance at a wavenumber common to the sampled oil, heat-accelerated aged oil, and added oil, i.e., 2800 cm -1 ~2900cm -1 and 1650~1750cm -1The absorbance at wave numbers below the threshold value of -0.5 in the collected oil corresponds to the "first explanatory variable" in the present invention, the absorbance at wave numbers below the threshold value of -0.7 in the heat accelerated deteriorated oil corresponds to the "second explanatory variable" and "third explanatory variable" in the present invention, and the absorbance at wave numbers below the threshold value of -0.998 in the additive oil corresponds to the "second explanatory variable" and "fourth explanatory variable" in the present invention.

[0037] It is generally known that during the production process of furfural, components containing aldehyde groups or ketone groups are produced. The wavenumber band related to the aldehyde group is 1720 cm -1 ~1740cm -1 Or 2820cm -1 The waveband related to the ketone group is 1705 cm -1 ~1725cm -1 That is, the wavenumber (2800 cm) common to each of the above-mentioned insulating sample oils (sample oil, heat-accelerated aged oil, and added oil) -1 ~2900cm -1 and 1650~1750cm -1 ) includes the wavenumber bands related to aldehyde and ketone groups. For this reason, the wavenumber (2800 cm) common to each of the above-mentioned insulating sample oils (sample oil, heat-accelerated aged oil, and added oil) is -1 ~2900cm -1 and 1650~1750cm -1 ) as an explanatory variable, the accuracy of estimating the furfural content can be improved. Based on the above, when generating the furfural content estimation equation, the wavenumber band of absorbance used as the explanatory variable is 2800 cm -1 ~2840cm -1 and 1700 cm -1 ~1740cm -1 It is preferable to set the -1 ~2830cm -1 and 1705 cm -1 ~1725cm -1 It is more preferable that:

[0038] [Determination of furfural content estimation formula] Using the explanatory variables selected in this way, PLS regression was performed with the furfural content measured by HPLC as the objective variable to generate a furfural content estimation equation. The furfural content estimation equation is stored in a memory unit (not shown) of the PC 10. In the deterioration index estimation device 1, the spectroscopic sensor 8 measures the absorbance spectrum of the insulating oil or insulating sample oil, and when the absorbance spectrum is input to the PC 10, the PC 10 calculates a second derivative spectrum using the absorbance spectrum, and estimates the furfural content using the second derivative spectrum and the furfural content estimation equation.

[0039] [Furfural content estimation results] The furfural content estimation equation thus generated was verified using 14 collected oil samples. Figure 11 is an explanatory diagram showing the correlation between the actual measured values ​​of furfural content and the estimated values ​​of furfural content calculated using the furfural content estimation equation, and Figure 12 is an explanatory diagram showing the standard partial regression coefficients of the explanatory variables used. In Figure 11, the more points are plotted on the diagonal line, the smaller the error between the actual measured values ​​and the estimated values. In this embodiment, the regression determination coefficient r 2 showed a relatively high value of 0.7255. In other words, it can be said that the selected explanatory variables explain the furfural content, which is the objective variable, well. The regression coefficient of determination r 2 The value of 0.7255 is the regression coefficient of determination r 2 However, in the range where the measured furfural content is greater than 5 mg / kg, i.e., in the area where deterioration has progressed further, the degree of agreement is higher than with additive oil, and it can be said to be useful for determining when to replace the oil. Here, the furfural content of insulating oil is set at 1.5 mg / kg as a warning value and 15 mg / kg as a recommended replacement value.

[0040] According to the embodiment described above, the correlation coefficient (correlation spectrum) between the absorbance spectrum (second derivative spectrum) and the furfural content of each insulating sample oil (collected oil, heat accelerated aging oil, added oil) is calculated, and the absorbance at a wave number common to each insulating sample oil (collected oil, heat accelerated aging oil, added oil) among the absorbances at which the correlation coefficient exceeds a threshold value is used as an explanatory variable to perform PLS regression to generate a furfural content estimation formula, and the furfural content of an insulating oil with an unknown furfural content is estimated using the generated furfural content estimation formula, so that the information of noise components (various components other than furfural) can be excluded and the furfural content can be accurately estimated without depending on the type of oil. Of course, in order to know the deterioration degree of an unknown insulating oil, it is not necessary to measure the furfural content using high efficiency liquid chromatography (HPLC) as in the past, so that the deterioration degree of an unknown insulating oil can be known easily in a short time and at low cost.

[0041] According to this embodiment, the wave number band (1720 cm) related to the aldehyde group and ketone group, which are components related to the production of furfural, -1 ~1740cm -1 Or 2820cm -1 , 1705cm -1 ~1725cm -1 ) (wavenumber is 2800 cm -1 or 2900cm -1 , and 1650cm -1 or 1750cm -1 Absorbance at a wave number of 2800 cm -1 or 2900cm -1 , and 1650cm -1 or 1750cm -1 More preferably, the absorbance at a wavenumber of 2810 cm -1 Or 2830cm -1 , and 1705 to 1725 cm -1 The absorbance at 100 nm is selected as an explanatory variable, which improves the accuracy of estimating the furfural content.

[0042] In addition, according to this embodiment, the regression coefficient of determination r 2 The regression coefficient of determination r when changing the number of explanatory variables and 2 The change in the regression coefficient of determination r, which is the maximum value when the number of explanatory variables is 100 or less, is calculated. 2 Select the regression coefficient of determination r 2 As a result, the number of absorbances (explanatory variables) at wavenumbers or wavelengths closely related to the generation of furfural can be appropriately selected. This makes it possible to more accurately estimate the furfural content of insulating oils whose furfural content is unknown.

[0043] In this embodiment, the sample oil, the heat accelerated aging oil, and the additive oil are used as the insulating sample oil, but the present invention is not limited to this. For example, only the sample oil and the heat accelerated aging oil, or only the sample oil and the additive oil may be used.

[0044] In this embodiment, the wave number is 2800 cm -1 or 2900cm -1 , and 1650cm -1 or 1750cm -1 Preferably, the absorbance at a wave number of 2800 cm -1 or 2900cm -1 , and 1650cm -1 or 1750cm -1 More preferably, the absorbance at a wavenumber of 2810 cm -1 Or 2830cm -1 , and 1705 to 1725 cm -1 The absorbance at the wavenumber of 1650 cm was selected as an explanatory variable, but this is not limited to this. -1 or 1750cm -1 Absorbance at a wave number of 1650 cm -1 or 1750cm -1 More preferably, the absorbance at a wave number of 1705 to 1725 cm -1 A configuration in which the absorbance at the wavenumber of 2800 cm is selected as an explanatory variable.-1 or 2900cm -1 Absorbance at a wave number of 2800 cm -1 or 2900cm -1 More preferably, the absorbance at a wavenumber of 2810 cm -1 Or 2830cm -1 The absorbance at may be selected as the explanatory variable.

[0045] In the present embodiment and the above-mentioned modified example, the present invention has been described as a method for estimating a deterioration index of insulating oil, which estimates a deterioration index that is an index showing the degree of deterioration of insulating oil. However, for example, the present invention may be applied to a program for executing the deterioration index estimation method. This program may be recorded on a computer-readable recording medium (for example, a hard disk, a ROM, an SSD, a flash memory (USB memory, SD card, etc.), a floppy disk, a CD, a DVD, etc.), or may be distributed from one computer to another computer via a transmission medium (a communication network such as the Internet or a LAN), or may be transmitted and received in any other form. By having one computer execute this program or having multiple computers share and execute each step, each step of the above-mentioned deterioration index estimation method is executed, so that the same effect as the deterioration index estimation method of the present embodiment can be obtained.

[0046] In this embodiment and the above-mentioned modified example, the regression coefficient of determination r 2 In order to select the above, the number of explanatory variables is set to 100 or less, but the number of explanatory variables may be any number up to 300.

[0047] In this embodiment and the above-mentioned modified example, the regression coefficient of determination r 2The number of explanatory variables for selecting is set to 100 or less, but is not limited to this. For example, the number of explanatory variables for collected oil is set to 100 or less, the number of explanatory variables for heat-accelerated aged oil is set to 300 or less, and the number of explanatory variables for added oil is set to 200 or less. For each oil (collected oil, heat-accelerated aged oil, added oil), the regression coefficient of determination r 2 The number of explanatory variables for selecting may be different.

[0048] The present embodiment shows an example of a mode for carrying out the present invention, and therefore the present invention is not limited to the configuration of the present embodiment.

[0049] <Additional Notes> In consideration of the above-mentioned gist of the invention, the degradation index estimation method and the degradation index estimation program according to the present invention can be configured in the following aspects. (Aspect 1) "A method for estimating an insulating oil deterioration index, which is an index showing the degree of deterioration of insulating oil, comprising: (a) obtaining the deterioration index and an absorbance spectrum showing a change in absorbance for each wave number or wavelength for a plurality of insulating oil samples; (b) calculating a correlation coefficient between the acquired deterioration index and the acquired absorbance spectrum for each of the insulating oil samples; (c) selecting, as explanatory variables, a plurality of absorbances at the wavenumbers or wavelengths at which the correlation coefficient exceeds a threshold value in the acquired absorbance spectrum; (d) performing PLS regression using the deterioration index as a target variable and a plurality of the explanatory variables to generate an estimation equation capable of estimating the deterioration index of the insulating oil; (e) estimating the deterioration index of the insulating oil, the deterioration index of which is unknown, using the generated estimation equation. A method for estimating the deterioration index of insulating oil. (Aspect 2) "The step (a) is a step of calculating a second derivative spectrum by subjecting the absorbance spectrum to second derivative processing, and acquiring the second derivative spectrum as the absorbance spectrum." The method for estimating a deterioration index of insulating oil according to the first aspect. (Aspect 3) "The step (c) includes a step of obtaining a change in a regression coefficient of determination, which is a correlation coefficient between the deterioration index actual values, which are the deterioration indexes obtained for the plurality of insulating oil samples, and the deterioration index estimated values, which are the deterioration indexes of the plurality of insulating oil samples, estimated using the estimation formula, when the threshold value is changed, obtaining a change in the regression coefficient of determination when the number of the explanatory variables is changed, selecting a maximum regression coefficient of determination, which is the regression coefficient of determination that shows a maximum value when the number of the explanatory variables is equal to or less than a predetermined number, and using the threshold value corresponding to the maximum regression coefficient of determination as the threshold value for selecting the explanatory variables." The method for estimating a deterioration index of insulating oil according to the first or second aspect. (Aspect 4) "The step (a) is a step of obtaining the deterioration index and the absorbance spectrum of each of the collected oil and the simulated deteriorated oil, using, as insulating oil samples, a sampled oil collected from insulating oil actually used in an equipment for the purpose of cooling and / or insulating the equipment, and a simulated deteriorated oil artificially generated to simulate deterioration due to use in the equipment, The step (b) is a step of calculating a first correlation coefficient which is the correlation coefficient between the deterioration index of the collected oil and the absorbance spectrum of the collected oil, and calculating a second correlation coefficient which is the correlation coefficient between the deterioration index of the simulated degraded oil and the absorbance spectrum of the simulated degraded oil, The step (c) is a step of selecting, as a first explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the first correlation coefficient exceeds a first threshold value, selecting, as a second explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the second correlation coefficient exceeds a second threshold value, and selecting, as the explanatory variable, the absorbances at the wavenumbers or the wavelengths that are common between the first explanatory variable and the second explanatory variable. The method for estimating a deterioration index of insulating oil according to any one of the first to third aspects. (Aspect 5) "The step (c) includes a step of obtaining a change in a collected oil regression determination coefficient, which is a correlation coefficient between a collected oil degradation index actual value, which is the degradation index obtained for the collected oil, and a collected oil degradation index estimated value, which is the degradation index of the collected oil estimated using the estimation formula, when the first threshold value is changed, and obtaining a change in the collected oil regression determination coefficient when the number of the first explanatory variables is changed, selecting a collected oil maximum regression determination coefficient, which is the collected oil regression determination coefficient that shows a maximum value when the number of the first explanatory variables is equal to or less than a first predetermined number, and using the first threshold value corresponding to the collected oil maximum regression determination coefficient as the first threshold value for selecting the first explanatory variable; and a step of obtaining a change in the collected oil regression determination coefficient when the number of the first explanatory variables is changed, which is a correlation coefficient between a collected oil degradation index actual value, which is the degradation index obtained for the collected oil, and an estimated value of the collected oil degradation index estimated using the estimation formula, when the first threshold value is changed, and selecting a collected oil maximum regression determination coefficient when the number of the first explanatory variables is equal to or less than a first predetermined number, and using the first threshold value corresponding to the collected oil maximum regression determination coefficient as the first threshold value for selecting the first explanatory variable; and a step of obtaining a change in the collected oil regression determination coefficient when the number of the first explanatory variables is changed, which is a correlation coefficient between a collected oil degradation index actual value, which is the degradation index obtained for the collected oil, and an estimated value of the collected oil degradation index estimated using the estimation formula. a step of determining a change in a simulated degraded oil regression coefficient of determination, which is a correlation coefficient between a simulated degraded oil degradation index actual value, which is the degradation index obtained for the simulated degraded oil, and a simulated degraded oil degradation index estimated value, which is the degradation index of the simulated degraded oil estimated using the estimation formula, when the number of the second explanatory variables is changed, determining a change in the simulated degraded oil regression coefficient of determination when the number of the second explanatory variables is changed, selecting a simulated degraded oil maximum regression coefficient of determination, which is the simulated degraded oil regression coefficient of determination that shows a maximum value when the number of the second explanatory variables is equal to or less than a second predetermined number, and using the second threshold value corresponding to the simulated degraded oil maximum regression coefficient of determination as the second threshold value for selecting the second explanatory variables. The method for estimating a deterioration index of insulating oil according to the fourth aspect. (Aspect 6) "The step (a) is a step of obtaining the degradation index and the absorbance spectrum of each of the thermally accelerated deteriorated oil and the degradation indicator component-added oil, which are obtained by artificially deteriorating the insulating oil by heating a new insulating oil under an environment simulating the usage state in the facility, as the simulated deteriorated oil, and the degradation indicator component-added oil, which are obtained by adding a component showing the degradation indicator to the new insulating oil; The step (b) is a step of calculating, instead of the second correlation coefficient, a third correlation coefficient which is a correlation coefficient between the deterioration index of the heat-accelerated deteriorated oil and the absorbance spectrum of the heat-accelerated deteriorated oil, and a fourth correlation coefficient which is a correlation coefficient between the deterioration index of the deterioration indicator component-added oil and the absorbance spectrum of the deterioration indicator component-added oil, The step (c) is a step of selecting, instead of the second explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the third correlation coefficient exceeds the third threshold as third explanatory variables, selecting a plurality of the absorbances at the wavenumbers or the wavelengths at which the fourth correlation coefficient exceeds a fourth threshold as fourth explanatory variables, and selecting, as the explanatory variable, the absorbances at the wavenumbers or the wavelengths that are common between the first explanatory variable, the third explanatory variable, and the fourth explanatory variable. The method for estimating a deterioration index of insulating oil according to the fourth or fifth aspect. (Aspect 7) "The step (c) includes a step of obtaining a change in a thermally accelerated degraded oil regression coefficient of determination, which is a correlation coefficient between a thermally accelerated degraded oil degradation index actual value, which is the degradation index obtained for the thermally accelerated degraded oil, and a thermally accelerated degraded oil degradation index estimated value, which is the degradation index of the thermally accelerated degraded oil estimated using the estimation formula, when the third threshold value is changed, and obtaining a change in the thermally accelerated degraded oil regression coefficient of determination when the number of the third explanatory variables is changed, selecting a thermally accelerated degraded oil maximum regression coefficient of determination, which is the thermally accelerated degraded oil regression coefficient of determination that shows a maximum value when the number of the third explanatory variables is equal to or less than a third predetermined number, and using the third threshold value corresponding to the thermally accelerated degraded oil maximum regression coefficient of determination as the third threshold value for selecting the third explanatory variable; and a step of obtaining a change in the thermally accelerated degraded oil regression coefficient of determination when the number of the third explanatory variables is changed, which is a correlation coefficient between a thermally accelerated degraded oil actual value, which is the degradation index obtained for the thermally accelerated degraded oil, and an estimated value of the thermally accelerated degraded oil degradation index estimated using the estimation formula, when the third threshold value is changed. a step of determining a change in a degradation indicator component added oil regression coefficient of determination, which is a correlation coefficient between a degradation indicator component added oil degradation index actual measurement value, which is the degradation index obtained for the degradation indicator component added oil, and a degradation indicator component added oil degradation index estimated value, which is the degradation index of the degradation indicator component added oil estimated using the estimation formula, when the deterioration indicator component added oil is subjected to a change in the number of the fourth explanatory variables, determining a change in the degradation indicator component added oil regression coefficient of determination when the number of the fourth explanatory variables is changed, selecting a degradation indicator component added oil maximum regression coefficient of determination, which is the degradation indicator component added oil regression coefficient of determination that shows a maximum value when the number of the fourth explanatory variables is equal to or less than a fourth predetermined number, and using the fourth threshold value corresponding to the degradation indicator component added oil maximum regression coefficient of determination as the fourth threshold value for selecting the fourth explanatory variables. The method for estimating a deterioration index of insulating oil according to the sixth aspect, which is dependent on the fifth aspect. (Aspect 8) "The deterioration indicator is the furfural content, In the step (c), the wave number is 2800 cm -1 or 2900cm -1 and / or 1650cm -1 or 1750cm -1 A step of selecting a plurality of the absorbances in the The method for estimating a deterioration index of insulating oil according to any one of the first to seventh aspects. (Aspect 9) "The step (c) is performed by determining whether the wave number is 2800 cm -1 Or 2840cm -1 and / or 1700cm -1 Or 1740cm -1 A step of selecting a plurality of the absorbances in the The method for estimating a deterioration index of insulating oil according to the eighth aspect. (Aspect 10) "The step (c) is performed by determining whether the wave number is 2810 cm -1 Or 2830cm -1 and / or 1705cm -1 Or 1725cm -1 A step of selecting a plurality of the absorbances in the The method for estimating a deterioration index of insulating oil according to the ninth aspect. (Aspect 11) "An insulating oil deterioration index estimation program for estimating a deterioration index, which is an index showing the degree of deterioration of insulating oil, A degradation index estimation program for causing a computer to execute each step of the degradation index estimation method according to any one of the first to tenth aspects. [Explanation of symbols]

[0050] 1 Deterioration index estimation device 2 containers 4 Holder 6 light source 8. Spectroscopic Sensor 10. Personal Computers

Claims

1. A method for estimating a deterioration index of insulating oil, which estimates a deterioration index that is an index indicating a degree of deterioration of insulating oil, comprising: (a) obtaining the deterioration index and an absorbance spectrum showing a change in absorbance for each wave number or wavelength for a plurality of insulating oil samples; (b) calculating a correlation coefficient between the acquired deterioration index and the acquired absorbance spectrum for each of the plurality of insulating oil samples; (c) selecting, as explanatory variables, a plurality of absorbances at the wavenumbers or wavelengths at which the correlation coefficient exceeds a threshold value in the acquired absorbance spectrum; (d) performing PLS regression using a plurality of the explanatory variables with the deterioration index as a target variable to generate an estimation equation capable of estimating the deterioration index of the insulating oil; (e) estimating a deterioration index of the insulating oil, the deterioration index of which is unknown, using the generated estimation equation. A method for estimating the deterioration index of insulating oil.

2. The step (a) is a step of calculating a second derivative spectrum by subjecting the absorbance spectrum to second derivative processing, and acquiring the second derivative spectrum as the absorbance spectrum. The method for estimating a deterioration index of insulating oil according to claim 1.

3. The step (c) includes a step of obtaining a change in a regression coefficient of determination, which is a correlation coefficient between actual degradation index values, which are the degradation indexes obtained for the plurality of insulating oil samples, and degradation index estimated values, which are the degradation indexes of the plurality of insulating oil samples, estimated using the estimation formula, when the threshold value is changed, and obtaining a change in the regression coefficient of determination when the number of the explanatory variables is changed, selecting a maximum regression coefficient of determination, which is the regression coefficient of determination that shows a maximum value when the number of the explanatory variables is equal to or less than a predetermined number, and using the threshold value corresponding to the maximum regression coefficient of determination as the threshold value for selecting the explanatory variables. The method for estimating a deterioration index of insulating oil according to claim 1 or 2.

4. The step (a) is a step of obtaining the deterioration index and the absorbance spectrum of each of the collected oil and the simulated deteriorated oil, using, as insulating oil samples, a collected oil collected from insulating oil actually used in an equipment for the purpose of cooling and / or insulating the equipment, and a simulated deteriorated oil artificially generated to simulate deterioration due to use in the equipment, The step (b) is a step of calculating a first correlation coefficient which is the correlation coefficient between the deterioration index of the collected oil and the absorbance spectrum of the collected oil, and calculating a second correlation coefficient which is the correlation coefficient between the deterioration index of the simulated degraded oil and the absorbance spectrum of the simulated degraded oil, The step (c) is a step of selecting, as a first explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the first correlation coefficient exceeds a first threshold value, selecting, as a second explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the second correlation coefficient exceeds a second threshold value, and selecting, as the explanatory variable, the absorbances at the wavenumbers or the wavelengths that are common between the first explanatory variable and the second explanatory variable. The method for estimating a deterioration index of insulating oil according to claim 1 or 2.

5. The step (c) includes a step of obtaining a change in a collected oil regression determination coefficient, which is a correlation coefficient between a collected oil degradation index actual value, which is the degradation index obtained for the collected oil, and a collected oil degradation index estimated value, which is the degradation index of the collected oil estimated using the estimation formula, when the first threshold value is changed, and obtaining a change in the collected oil regression determination coefficient when the number of the first explanatory variables is changed, selecting a collected oil maximum regression determination coefficient, which is the collected oil regression determination coefficient that shows a maximum value when the number of the first explanatory variables is equal to or less than a first predetermined number, and using the first threshold value corresponding to the collected oil maximum regression determination coefficient as the first threshold value for selecting the first explanatory variable; and a step of obtaining a change in the collected oil regression determination coefficient when the number of the first explanatory variables is changed, which is equal to or less than a first predetermined number. a step of calculating a change in a simulated degraded oil regression coefficient of determination, which is a correlation coefficient between a simulated degraded oil degradation index actual value, which is the degradation index obtained for the simulated degraded oil, and a simulated degraded oil degradation index estimated value, which is the degradation index of the simulated degraded oil estimated using the estimation formula, when the number of the second explanatory variables is changed; calculating a change in the simulated degraded oil regression coefficient of determination, when the number of the second explanatory variables is changed; selecting a simulated degraded oil maximum regression coefficient of determination, which is the simulated degraded oil regression coefficient of determination that shows a maximum value when the number of the second explanatory variables is equal to or less than a second predetermined number; and using the second threshold value corresponding to the simulated degraded oil maximum regression coefficient of determination as the second threshold value for selecting the second explanatory variable. The method for estimating a deterioration index of insulating oil according to claim 4.

6. The step (a) is a step of obtaining the degradation index and the absorbance spectrum of each of the thermally accelerated deteriorated oil and the degradation indicator component-added oil, which are obtained by artificially deteriorating the insulating oil by heating the new insulating oil in an environment simulating the usage state in the facility, as the simulated deteriorated oil, and the degradation indicator component-added oil, which are obtained by adding a component showing the degradation indicator to the new insulating oil, The step (b) is a step of calculating, instead of the second correlation coefficient, a third correlation coefficient which is a correlation coefficient between the deterioration index of the heat-accelerated deteriorated oil and the absorbance spectrum of the heat-accelerated deteriorated oil, and a fourth correlation coefficient which is a correlation coefficient between the deterioration index of the deterioration indicator component-added oil and the absorbance spectrum of the deterioration indicator component-added oil, The step (c) is a step of selecting, instead of the second explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the third correlation coefficient exceeds a third threshold as a third explanatory variable, selecting a plurality of the absorbances at the wavenumbers or the wavelengths at which the fourth correlation coefficient exceeds a fourth threshold as a fourth explanatory variable, and selecting, as the explanatory variable, the absorbances at the wavenumbers or the wavelengths that are common between the first explanatory variable, the third explanatory variable, and the fourth explanatory variable. The method for estimating a deterioration index of insulating oil according to claim 4.

7. The step (a) is a step of obtaining the degradation index and the absorbance spectrum of each of the thermally accelerated deteriorated oil and the degradation indicator component-added oil, which are obtained by artificially deteriorating the insulating oil by heating the new insulating oil in an environment simulating the usage state in the facility, as the simulated deteriorated oil, and the degradation indicator component-added oil, which are obtained by adding a component showing the degradation indicator to the new insulating oil, The step (b) is a step of calculating, instead of the second correlation coefficient, a third correlation coefficient which is a correlation coefficient between the deterioration index of the heat-accelerated deteriorated oil and the absorbance spectrum of the heat-accelerated deteriorated oil, and a fourth correlation coefficient which is a correlation coefficient between the deterioration index of the deterioration indicator component-added oil and the absorbance spectrum of the deterioration indicator component-added oil, The step (c) is a step of selecting, instead of the second explanatory variable, a plurality of the absorbances at the wavenumbers or the wavelengths at which the third correlation coefficient exceeds a third threshold as a third explanatory variable, selecting a plurality of the absorbances at the wavenumbers or the wavelengths at which the fourth correlation coefficient exceeds a fourth threshold as a fourth explanatory variable, and selecting, as the explanatory variable, the absorbances at the wavenumbers or the wavelengths that are common between the first explanatory variable, the third explanatory variable, and the fourth explanatory variable. The method for estimating a deterioration index of insulating oil according to claim 5.

8. The step (c) includes a step of calculating a change in a heat accelerated degraded oil regression coefficient of determination, which is a correlation coefficient between a heat accelerated degraded oil degradation index actual value, which is the degradation index obtained for the heat accelerated degraded oil, and a heat accelerated degraded oil degradation index estimated value, which is the degradation index of the heat accelerated degraded oil estimated using the estimation formula, when the third threshold value is changed, and calculating a change in the heat accelerated degraded oil regression coefficient of determination when the number of the third explanatory variables is changed, selecting a heat accelerated degraded oil maximum regression coefficient of determination, which is the heat accelerated degraded oil regression coefficient of determination that shows a maximum value when the number of the third explanatory variables is equal to or less than a third predetermined number, and using the third threshold value corresponding to the heat accelerated degraded oil maximum regression coefficient of determination as the third threshold value for selecting the third explanatory variable; and a step of calculating a change in the heat accelerated degraded oil regression coefficient of determination when the number of the third explanatory variables is changed, which is a correlation coefficient between a heat accelerated degraded oil actual value, which is the degradation index obtained for the heat accelerated degraded oil, and an estimated value of the heat accelerated degraded oil degradation index estimated using the estimation formula, when the third threshold value is changed. and a step of calculating a change in a degradation indicator component added oil regression coefficient of determination, which is a correlation coefficient between a degradation indicator component added oil degradation index actual measurement value, which is the degradation index obtained for the degradation indicator component added oil, and a degradation indicator component added oil degradation index estimated value, which is the degradation index of the degradation indicator component added oil estimated using the estimation formula, when the deterioration indicator component added oil is measured, and calculating a change in the degradation indicator component added oil regression coefficient of determination when the number of the fourth explanatory variables is changed, and selecting a degradation indicator component added oil maximum regression coefficient of determination, which is the degradation indicator component added oil regression coefficient of determination that shows a maximum value when the number of the fourth explanatory variables is equal to or less than a fourth predetermined number, and using the fourth threshold value corresponding to the degradation indicator component added oil maximum regression coefficient of determination as the fourth threshold value for selecting the fourth explanatory variable. The method for estimating a deterioration index of insulating oil according to claim 7.

9. The deterioration indicator is furfural content, The step (a) is a step of acquiring the absorbance spectrum using infrared spectroscopy; In the step (c), the wave number is 2800 cm -1 or 2900 cm -1 , and / or 1650 cm -1 or 1750 cm -1 A step of selecting a plurality of the absorbances in the The method for estimating a deterioration index of insulating oil according to claim 1 or 2.

10. In the step (c), the wave number is 2800 cm -1 or 2840 cm -1 , and / or 1700 cm -1 or 1740 cm -1 A step of selecting a plurality of the absorbances in the The method for estimating a deterioration index of insulating oil according to claim 9.

11. In the step (c), the wave number is 2810 cm -1 or 2830 cm -1 , and / or 1705 cm -1 or 1725 cm -1 A step of selecting a plurality of the absorbances in the The method for estimating a deterioration index of insulating oil according to claim 10.

12. An insulating oil deterioration index estimation program for estimating a deterioration index which is an index indicating a degree of deterioration of insulating oil, A degradation index estimating program that causes a computer to execute each step of the degradation index estimating method according to claim 1 or 2.

Images