JP4550537B2 - System and method for estimating residual insulation life of high-pressure rotating machine - Google Patents

Description

  The present invention relates to a high-pressure rotating machine that plays a central role in a large-scale building, for example, a high-pressure motor used for a turbo chiller for air conditioning, a large-capacity supply / exhaust fan, etc., or a high-pressure as a power supply facility for regular or emergency use. The present invention relates to a system and an estimation method for estimating a remaining insulation life at an arbitrary number of years of operation of a generator or the like.

  In general, large-scale buildings and factories are equipped with high-voltage motors that are important for maintaining building functions, such as high-voltage motors as power load equipment such as air conditioning equipment and ventilation equipment, and commercial power supply equipment. A normal high voltage generator connected to the grid is installed, or an emergency high voltage generator is installed as a backup for commercial power.

In these high-pressure rotating machines, the electric field stress is repeatedly applied to the insulators applied to the coils of the rotating machine every time the rotating machine is operated and stopped, and thermal stress and electromagnetic mechanical stress due to operating loss are also applied. Stress is added in combination.
Furthermore, it is considered that so-called complex insulation deterioration proceeds in the insulator of the high-pressure rotating machine by the environmental degradation factors such as moisture absorption and dust acting on these stresses.

When an insulation accident occurs in a high-pressure rotating machine, the range of influence in that case is much larger than that of a low-capacity low-speed rotating machine, and the productivity of the building or factory itself can be significantly reduced.
In addition, there is a case where the high-voltage rotating machine insulation accident triggered a trip of the main electrical room circuit breaker in a building or factory, resulting in a power failure throughout the building. Has been.

  From this point of view, in the maintenance of high-pressure rotating machine equipment, which is an important load facility, before the high-pressure rotating machine (high-voltage motor or high-voltage generator) causes an insulation accident, Management such as renewing the high-pressure rotating machine itself will lead to important maintenance technology.

  In order to determine the progress of insulation deterioration of a high-pressure rotating machine and determine whether or not to continue its use, the most commonly used method has been an insulation resistance value measured during regular inspection (usually 10 MΩ to 10,000 MΩ). The degree is determined to be defective when the level is reduced to a reference value (about 10 MΩ).

  In this case, if the operating age of the high-pressure rotating machine is, for example, about 10 to 15 years, the insulator is washed and dried by overhaul, and in some cases, the impregnation material is re-impregnated. .

Furthermore, in the case where about 20 years have elapsed since the operation, a method for renewing the high-pressure rotating machine itself has been proposed by comprehensively judging the degree of insulation recovery and the life cycle cost of the equipment. ing.
In this case, the lower limit threshold value of the insulation resistance value, which is the criterion for judgment, is the trend of the insulation resistance value of the high-pressure rotating machine with respect to the number of years of operation, the relative dielectric breakdown value [%] of the high-pressure rotating machine, and the time of drying and moisture absorption. This is based on the characteristics of the insulation resistance ratio.

  Specifically, the insulation resistance data of multiple high-voltage rotating machines under investigation were brought back from the site to the factory, the distribution characteristics data of the insulation resistance value at the time of acceptance, and the insulation resistance value measured after washing and drying at the factory Plotting with the characteristic distribution data of Fig. 2 shows that the insulation resistance value by cleaning and drying is remarkable for those with an operation age of around 15 years, but the insulation resistance value decreases to about 10 MΩ after 20 years have passed. It can be seen that there is little recovery of the insulation resistance value after washing and drying.

  For example, as a characteristic diagram, plotting the logarithm of the ratio (= RD / RW) of the insulation resistance value RD at the time of drying and the insulation resistance value RW at the time of moisture absorption (= RD / RW) on the horizontal axis, and plotting the measured dielectric breakdown relative value on the vertical axis, As the insulation resistance value RW at the time of moisture absorption decreases and the log (RD / RW) increases, the dielectric breakdown value decreases. When log (RD / RW) ≈3, the dielectric breakdown level is about 50% of the new product. It is understood that it leads to.

Further, from the viewpoint of a general insulation life, it is generally accepted that the time when the dielectric breakdown value falls to half the new product level (50%) is regarded as the life.
Therefore, when the insulation resistance value RD at the time of drying is 10000 MΩ, the insulation resistance value RW at the time of moisture absorption that reduces the dielectric breakdown value by half is 10 MΩ.
Based on such a technical background, there is a history of setting the lower limit threshold value of the insulation resistance value to “10 MΩ” as a standard for determining pass / fail that has been conventionally employed.

On the other hand, the second pass / fail judgment method using the dielectric loss rate tan δ, which is considered to have high detection accuracy of local insulation deterioration, compared to the insulation pass / fail judgment method of the high-pressure rotating machine using the lower threshold (10 MΩ). Has also been proposed (see Non-Patent Document 1, for example).
In this case, the upper limit threshold value of the dielectric loss rate tan δ, which is a criterion for judgment, is the trend of the base dielectric loss rate tan δ0 of the high-voltage rotating machine with respect to the number of years of operation, and the relative breakdown value [%] of the high-voltage rotating machine and the base This is based on the characteristic of the dielectric loss rate tan δ0.

  The characteristic distribution in this case is that after the base dielectric loss rate tan δ0 data of a plurality of high-voltage rotating machines that were investigated are brought back to the factory from the site, the base dielectric loss rate tan δ0 at the time of acceptance, and after washing and drying in the factory This is obtained by plotting the measured base dielectric loss rate tan δ0 separately.

  From this characteristic diagram, the recovery of the base dielectric loss rate tan δ0 due to cleaning and drying is remarkable for those with an operation age of around 15 years, but the tan δ 0 exceeds 10% after the lapse of 20 years. It can be seen that the deterioration of the base dielectric loss rate tan δ0 is small even after cleaning and drying.

  For example, when the value of the base dielectric loss rate tan δ0 is plotted on the horizontal axis and the measured dielectric breakdown relative value is plotted on the vertical axis, the characteristic curve varies, but the magnitude of the base dielectric loss rate tan δ0 is It can be seen that when the base dielectric loss rate tan δ0 is about 10% to 20%, the breakdown level tends to be less than 50% of the new product as the dielectric breakdown value decreases as the value increases.

In addition, as described above, from the viewpoint of the insulation life, it is generally accepted that the time when the dielectric breakdown value is reduced to half of the new product level is regarded as the life, so that the base dielectric whose dielectric breakdown value is halved The loss rate tan δ0 is 15% on average.
Based on such a technical background, there is a history of setting the upper limit threshold value of the base dielectric loss rate tan δ0 to “15%” as a standard of quality determination that has been conventionally adopted.

  Thus, in determining whether or not overhaul (or renewal) of a conventional high-pressure rotating machine is necessary, a certain threshold value is set for the level of the insulation resistance value or base dielectric loss rate tan δ that can be measured during periodic inspection. Judgment is made in consideration of information whether the value exceeds this threshold and the number of years of operation.

However, since the remaining insulation life that is most requested by facility managers from the standpoint of preventive maintenance is unknown, it is not possible to give a quantitative answer as to how many years should be set for overhaul and renewal. could not.
Therefore, quantitative judgment on the remaining insulation life of the high-pressure rotating machine must be left to the subjectivity of the investigator or equipment manager. Insulation accidents could occur, resulting in significant economic losses and, in some cases, social problems.

IEEJ Technical Report II, Part 267

In the conventional high-pressure rotating machine management method, in addition to the number of years of operation, whether or not the test data of the insulation resistance value measured at the time of annual inspection, for example, indicates a management level (10 MΩ) or higher Then, the quality determination and renewal determination of the high-pressure rotating machine were performed, and the renewal time was determined by subjective judgment of the equipment manager or the like.
However, the original purpose of periodic inspection or inspection is to know the remaining life at the time of inspection and inspection and reflect it in future repairs (overhaul) or renewal plans. Since there is no method for realizing a quantitative evaluation of the remaining insulation life and a system (software) for estimating the life, there is a problem that the remaining insulation life cannot be estimated.
Therefore, there is a problem that the risk of dielectric breakdown accident due to the lack of management and loss of the timing of repair or renewal of the high-pressure rotating machine is high.

  The present invention has been made to solve the above-described problems. Regarding high-voltage rotating machines (high-voltage motors and high-voltage generators) operating in buildings, the present invention is used at the time of periodic inspection or irregular inspection. Based on non-destructive insulation data, high voltage that can estimate the statistical remaining years until dielectric breakdown, minimize the risk of dielectric breakdown, and reflect it in preventive maintenance and planned repair plans It is an object of the present invention to obtain a remaining insulation life estimation system and estimation method for a rotating machine.

The system for estimating the remaining insulation life of a high-pressure rotating machine according to the present invention is a system for estimating the remaining insulation life of a high-pressure rotating machine at an arbitrary number of years of operation. Data including data and breakdown voltage data, insulation data measurement means to obtain the current insulation resistance value or dielectric loss rate for high-voltage rotating machines as insulation data, and statistical multiple regression analysis of the database for each insulation index and time passage Statistical multiple regression analysis means for estimating the slope parameter of the data value with respect to the current level determination means for determining the current level of the insulation data for each of the reliability of the plurality of stages based on the insulation data and the slope parameter, the slope parameter and Remaining insulation life that estimates the remaining insulation life of high-pressure rotating machines based on the current level And the remaining insulation life estimation means reaches a non-destructive insulation data threshold at which the reliability data for each reliability with respect to the future operating years in the future is expected to deteriorate along the slope parameter and lead to destruction. The period until is estimated as the remaining insulation life.

The method for estimating the remaining insulation life of a high-pressure rotating machine according to the present invention is a method for estimating the remaining insulation life of a high-pressure rotating machine at an arbitrary number of years of operation. Obtaining a database containing breakdown insulation data and breakdown voltage data, measuring the current insulation resistance value or dielectric loss rate for high-voltage rotating machines as insulation data, and performing a statistical multiple regression analysis on the database for each insulation index Estimating the slope parameter of the data value over time, determining the current level of insulation data for each of the multiple levels of reliability based on the insulation data and the slope parameter, and based on the slope parameter and the current level Insulation data at each reliability level for future operating years It is obtained by a step of estimating the time to reach the non-destructive insulation data threshold is expected to lead to destruction worse along the parameter as residual insulation life.

  According to the present invention, the remaining insulation life until dielectric breakdown can be quantitatively evaluated based on the non-destructive insulation data measured at an arbitrary time point, so that a timely overhaul or equipment update plan can be obtained. Therefore, it is possible to prevent the occurrence of a sudden insulation accident of the high-pressure rotating machine due to the loss of repair or renewal timing as in the past.

Embodiment 1 FIG.
1 is a block diagram schematically showing a residual insulation life estimation system for a high-pressure rotating machine according to Embodiment 1 of the present invention.
In FIG. 1, the remaining insulation life estimation system related to the high-pressure rotating machine 100 includes a database 1, insulation data measurement means 2, statistical multiple regression analysis means 3, current level determination means 4, and remaining insulation life estimation means. 5 and display means 6.

  The database 1 includes a plurality of non-destructive insulation data and breakdown voltage data that are known in advance with respect to the high-pressure rotating machine 100 (for example, published in the aforementioned Non-Patent Document 1).

The insulation data measuring means 2 is used by an operator at the time of periodic inspection or non-periodic arbitrary inspection, and acquires the current insulation resistance value Ri or dielectric loss rate tan δi related to the high-pressure rotating machine 100 as insulation data .
The display means 6 has input means such as a keyboard, and the obtained insulation data Ri and tan δi are input via the screen of the display means 6 as will be described later by operation of the input means. .

The statistical multiple regression analysis means 3 performs a statistical multiple regression analysis of the data values in the database 1 for each insulation index, and estimates the slope parameter P of the data values with respect to time.
Based on the insulation data Ri, tan δi and the inclination parameter P, the current level determination means 4 determines the current level of the insulation data Ri, tan δi for each reliability in a plurality of stages (50%, 90%, 95%, 99%). To do.

The remaining insulation life estimation means 5 estimates the remaining insulation life of the high-pressure rotating machine 100 based on the current level of the insulation data Ri and tan δi and the inclination parameter P.
Specifically, the remaining insulation life estimation means 5 is the non-destructive insulation data in which the insulation data Ri and tan δi at each reliability with respect to the years of operation in the future are expected to deteriorate along the slope parameter P and lead to destruction. The period until the threshold is reached is estimated as the remaining insulation life.

As will be described later, the display unit 6 switches and displays various menu screens having an operation panel function, and displays the estimation result of the remaining insulation life estimation unit 5 in accordance with an operator operation.
Further, a printer is connected to the display means 6 as necessary.

FIG. 2 is an explanatory view showing a display example of a diagnosis result of the remaining insulation life output to the display means 6 in FIG.
3 and 4 are explanatory diagrams showing the processing operation of the remaining insulation life estimation system of FIG. 1 in a screen flow.

  In FIG. 2, the displayed form 20 includes a form title and menu part 21, a data description part 22 of insulation data of the high-pressure rotating machine 100 to be diagnosed, and an algorithm part (a simulation diagram for estimating the remaining insulation life). ) 23, a description portion 24 of the remaining insulation life (additional finding list) for each reliability, and a description portion 25 of the automatically selected diagnostic findings.

In the menu portion 21, the name of the building, the date of survey, the name of the equipment used for the survey, and the like are described in addition to the specifications and date of manufacture of the high-pressure rotating machine to be surveyed.
In the data description part 22, for example, when the insulation inspection of the high-pressure rotating machine to be investigated is executed with tan δ, the data is described.
The algorithm part 23 shows an algorithm for calculating and estimating the remaining insulation life for each reliability from the value of tan δi at the time of investigation.

In the description 24 of the remaining insulation life, the remaining insulation life for each reliability obtained from the simulation of the algorithm portion 23 is described.
In the additional finding description portion 25, the estimation result of the description portion 24 of the remaining insulation life is patterned, and the diagnostic findings registered in advance are automatically output, and the additional finding is arbitrarily described by the investigator.

3 and 4, the screens 31 to 35, 41 to 44, and 51 to 54 are displayed on the display unit 6 in accordance with an interactive operation with the operator.
On the main menu screen 31, a diagnostic data input screen 32 and diagnostic data read screens 33 and 35 can be selected.

  One diagnostic data read screen 33 is selected by a single process and is interactively related to the single process menu screen 34, and the other diagnostic data read screen 35 is selected by a report output process.

  The single processing menu screen 34 is between the remaining life simulation graph screens 41 and 51, the remaining life estimation table output screens 42 and 52, the observation sentence confirmation screens 43 and 53, and the file (report data) storage screens 44 and 54. The screens 41 to 44 and 51 to 54 can be selected.

The screens 41 to 44 are screens when the dielectric loss rate tan δ is selected from the database, and the screens 51 to 54 are screens when the insulation resistance value R is selected from the database.
In the remaining life simulation graph screens 41 and 51, a characteristic line based on the slope parameter P is displayed, and in the remaining life estimation table output screens 42 and 52, the estimated life is represented by each reliability (50%, 90%, 95). %, 99%).

Next, an operation for obtaining the form 20 shown in FIG. 2 using the remaining insulation life estimation system of FIG. 1 will be described with reference to the explanatory diagrams of FIGS.
3 and 4 showing the life estimation flow, first, for example, when a diagnostic data input menu is selected on the main menu screen 31, a diagnostic data input screen 32 is displayed.
Subsequently, on the diagnostic data input screen 32, the name of the building to be surveyed, the specifications and the date of manufacture of the high-pressure rotating machine to be surveyed, and the like are input in an interactive format.

Finally, on the diagnostic data input screen 32, either a dielectric loss rate tan δ (measured value tan δi) or an insulation resistance value R (measured value Ri) serving as an insulation index is selected as diagnostic data.
Thereafter, although not shown in the drawing, the process returns to the main menu screen 31 again through the process of registering the diagnostic data file.

Here, a procedure related to the screens 41 to 44 when the dielectric loss rate tan δ is selected as the insulation index and the data value is input will be described.
Note that the processing procedure (described later) regarding the screens 51 to 54 when the insulation resistance value R is selected as the insulation index is substantially the same as this case.

Next, when the single processing menu is selected on the main menu screen 31, a diagnostic data read screen 33 is displayed.
Subsequently, when the registered property file is selected from the diagnostic data readout screen 33, the process proceeds to the single processing menu screen 34. Hereinafter, the processing menus on the single processing menu screen 34 are processed in the order of numbers 1 to 6 from the top in the figure.

When the first remaining insulation life simulation graph is selected on the single processing menu screen 34, a remaining life simulation graph screen 41 is displayed.
Subsequently, after checking the remaining life simulation graph screen 41 and returning to the single processing menu screen 34, the check box of the processing menu on the single processing menu screen 34 is filled. The estimated output menu will be selected.

As a result, the remaining life estimation table output screen 42 is displayed, and the data value of the input dielectric loss rate tan δ and the remaining insulation life for each reliability are displayed. The operator confirms these output values.
Thereafter, when returning to the single processing menu screen 34 again, since the check boxes of the two processing menus so far are filled, the process proceeds to the third finding sentence confirmation menu.

At this time, the remaining insulation life estimation means 5 (see FIG. 1) considers the estimated remaining insulation life and the number of years of operation, etc., from among the previously registered finding sentences (for example, 6 patterns), The optimum finding sentence confirmation screen 43 is automatically selected and displayed.
Hereinafter, the operator confirms the finding sentence confirmation screen 43 and, depending on the case, makes corrections as necessary.

  Thereafter, returning to the single processing menu screen 34 and selecting the fourth report printer output, a print preview of the form 20 (see FIG. 2) is displayed. After confirming this, the print processing is executed.

  Finally, the process returns to the single processing menu screen 34, selects the fifth report file storage menu, executes registration and storage of the report on the file storage screen 44, and then performs processing on the single processing menu screen 34. By selecting the last (sixth) end menu of the menu, a series of single processing operations (see FIGS. 3 and 4) is ended.

Next, the life estimation algorithm according to the first embodiment of the present invention when the dielectric loss rate tan δ is selected as the insulation index of the remaining insulation life will be described with reference to the explanatory diagrams of FIGS.
5 and 6 correspond to the processing of the statistical multiple regression analysis means 3, and FIG. 7 corresponds to the processing of the current level determination means 4 and the remaining insulation life estimation means 5.

FIG. 5 is an explanatory diagram showing a database value of the dielectric loss rate tan δ relating to a known high-pressure rotating machine 100 (for example, published in Non-Patent Document 1).
FIG. 6 is an explanatory diagram showing a regression line and an approximate expression regarding the dielectric loss rate tan δ obtained by performing multiple regression analysis on the database value of FIG.
FIG. 7 is an explanatory diagram for simulating the statistical remaining insulation life from the investigation value tan δi of the dielectric loss rate of the high-pressure rotating machine 100.

  In FIG. 5, the horizontal axis represents the number of years of operation, the vertical axis represents the value of the dielectric loss rate tan δ, and among the data values of the dielectric loss rate tan δ that was the basis of the prior art, Each data value of the dielectric loss rate tan δ is shown by a plot (see the mark ◆).

  As shown in FIG. 5, since the value of the dielectric loss rate tan δ varies over a wide range even in the same operation years, the statistical multiple regression analysis means 3 treats these data values statistically and performs multiple regression. By performing the analysis, an analysis result as shown in FIG. 6 is acquired.

Here, the data value (marked by ◆) of the dielectric loss rate tan δ shown in FIG. 5 is divided into five zones depending on the length of operating years, and the dielectric loss rate on the high frequency side of the normal distribution for each zone. Assume that a data value corresponding to the occurrence probability of tan δ is obtained.
For example, by dividing into 5 zones of 10 years to 12.5 years, 12.5 years to 15 years, 15 years to 17.5 years, 17.5 years to 20 years, 20 years to 22.5 years All data can be processed efficiently.

In FIG. 6, the data value (tan δ) indicating the occurrence probability of 50% is plotted with “■ mark in □”, and the data value (tan δ) at each occurrence probability of 10%, 5%, and 1% is Plotted with “black triangle mark in Δ”, “x mark in ○”, and “* mark in □”, respectively.
By performing multiple regression analysis on the plots of each of the five zones for each of these four occurrence probabilities (50%, 10%, 5%, 1%), the multiple regression lines L1 to L4 are expressed as follows: Primary approximation formulas (1) to (4) are obtained.

y = 0.0.806x + 0.7747 (1)
y = 0.8634x + 6.5500 (2)
y = 0.8796x + 8.18772 (3)
y = 0.9100x + 11.2580 (4)

In the above approximate expressions (1) to (4), x indicates the elapsed operation years on the horizontal axis, and y indicates the data value (tan δ [%]) on the vertical axis.
Of the first-order approximation formulas (1) to (4) representing the multiple regression lines L1 to L4, the approximation formula (1) of the straight line L1 corresponds to an occurrence probability of 50%, and the straight lines L2 to L4 The approximate expressions (2) to (4) correspond to the occurrence probabilities of 10%, 5%, and 1%, respectively.
It is experimentally found that the correlation coefficients of the regression lines L1 to L4 obtained by the multiple regression analysis as shown in FIG. 6 show sufficiently good values (for example, 0.7 to 0.9). ing.

  By the way, since the dielectric loss rate tan δi measured at the regular inspection of the high-pressure rotating machine 100 to be investigated at an arbitrary site is the only insulation data acquired at the time of investigation of the high-pressure rotating machine 100, it is statistically Is treated as a data value (tan δ) with an occurrence probability of 50% over the years of operation of the high-pressure rotating machine 100.

  Therefore, as shown in the remaining simulation explanatory diagram of FIG. 7, the reliability curve L50 of 50% passing through the insulation data value (tan δ) measured at the current time point, and 90%, 95%, and 99% of the upper level are shown. Each confidence curve L90, L95 and L99 can be located.

These reliability curves L50, L90, L95, and L99 correspond to the same number of years of operation as the current time point (in this case, 17 years), for example, a regression line L3 with an occurrence probability of 10% in FIG. The 90% reliability curve L90 in FIG. 7 is positioned by considering the positional relationship between the occurrence probability and the regression line L1.
The 95% and 99% reliability curves L95 and L99 at the current time are similarly positioned based on FIG.
In addition, the dependency of these four reliability curves L50, L90, L95, and L99 on the number of years of operation is drawn by employing the slope parameters of the above-described approximate equations (1) to (4).

  In FIG. 7, the data value (tan δ) that causes the dielectric breakdown level of the high-pressure rotating machine 100 to be halved from the level at the time of a new product is indicated by an upper limit threshold TH (corresponding to tan δ = 15%) that has been conventionally employed. Has been.

The remaining insulation life estimation means 5 corresponds to the point where the simulation curves L50 to L99 of the respective reliability (50%, 90%, 95%, 99%) intersect with the upper limit threshold value TH, for example, the elapsed years of operation. From the difference between T50 and the current operation age (17 years), the remaining insulation life of 50% reliability is calculated.
Similarly, the remaining insulation lifetimes of 90%, 95% and 99% are calculated quantitatively from the difference between the operating years T90, T95 and T99 and the current operating years.

  As described above, according to Embodiment 1 of the present invention, the multiple regression analysis was performed on the insulation data values (tan δ) of many high-pressure rotating machines 100 included in the database 1 published in Non-Patent Document 1 described above. By combining the statistical analysis result and the upper / lower threshold TH for quality determination with respect to the insulation data value (tan δ) of the conventional high-pressure rotating machine 100, quantitative residual insulation life estimation that cannot be realized by the conventional technology can be performed. Can be realized. Therefore, the effect that it can contribute to the rational preventive maintenance of the high-pressure rotating machine 100 which is an important facility is obtained.

  That is, the lower limit threshold (the lower limit management level of the insulation resistance value R) known from Non-Patent Document 1 is determined from the non-destructive insulation resistance value Ri or the dielectric loss rate tan δ acquired at the time of periodic inspection or non-periodic inspection. The number of years expected to reach the upper limit management level of the dielectric loss rate tan δ can be statistically estimated.

That is, the statistical multiple regression analysis means 3 statistically analyzes these data based on the database 1 including non-destructive insulation data and valuable breakdown voltages regarding many high-voltage rotating machines 100 published in Non-Patent Document 1 and the like. Multiple regression analysis.
Further, the current level determining means 4 calculates the current level of 50% reliability and 90%, 95%, 99% from the insulation data (insulation resistance value Ri or dielectric loss rate tan δi) measured at an arbitrary time point. The current level for each reliability level is determined.

  As a result, the remaining insulation life estimation means 5 follows the slope parameter P (the slope of the characteristic curve) estimated in the multiple regression analysis for each insulation index at each reliability with respect to the future operation years based on the current level. The remaining insulation life until the threshold value of the non-destructive insulation data (the upper limit threshold value TH for the dielectric loss rate tan δ or the lower limit threshold value for the insulation resistance value R), which is expected to cause breakdown due to deterioration of the insulation data, Obtainable.

Since the remaining insulation life estimation method and life estimation system according to Embodiment 1 of the present invention can quantitatively evaluate the remaining insulation life from non-destructive insulation data at any time point to failure, it is timely. An overhaul or an update plan for the internal equipment of the high-pressure rotating machine 100 can be formulated.
Therefore, it is possible to prevent the risk management from being unsatisfactory, such as when the repair or renewal timing is lost as in the conventional system and a sudden insulation accident of the high-pressure rotating machine 100 occurs.

In the above description, the case where the dielectric loss rate tan δ is selected as the insulation index is taken as an example. However, even when the non-destructive insulation resistance value R is selected as the insulation index, the database 1 and the above-described lower threshold ( 10 MΩ), the remaining insulation life can be estimated as described above.
However, in this case, the slope of the reliability curve used in the residual simulation is negative as in the case of the dielectric loss rate tan δ, as in the graph screen 51 in FIG.

Hereinafter, the processing operation when the insulation resistance value R is selected will be described with reference to FIGS. 1, 3, and 4.
In the above description, the case where the dielectric loss rate tan δ (measured value tan δi) is input as the insulation data of the high-pressure rotating machine 100 in FIGS. 3 and 4 (screen flow of the life estimation system) has been described. By inputting the insulation resistance value R (measured value Ri) investigated for the machine 100, the remaining insulation life can be similarly estimated quantitatively.

That is, after selecting the diagnostic data input menu on the main menu screen 31 in FIG. 3 as described above, on the diagnostic data input screen 32, the name of the building to be surveyed, the specification and year of manufacture of the high-pressure rotating machine 100 to be surveyed. The month or the like is input in an interactive format, and finally, the insulation resistance value R is selected as the insulation index of the diagnostic data and the data value (Ri) is input.
Thereafter, the process returns to the main menu screen 31 again through the process of registering the diagnostic data file.

  Next, as described above, when a single processing menu is selected on the main menu screen 31, a diagnostic data readout screen 33 is displayed. When a registered property file is selected from these, the single processing menu screen 34 is displayed. In the following, the processing menus on the single processing menu screen 34 are processed in order.

First, when the first remaining insulation life simulation graph is selected, a remaining life simulation graph screen 51 (see FIG. 4) when the insulation resistance value R is selected is displayed. After confirming this graph screen, the single processing menu is displayed. Return to screen 34.
At this time, since the check box of the processing menu on the single processing menu screen 34 is filled, the second remaining life estimation output menu is selected next.

As a result, the remaining life estimation output screen 52 is displayed, and the input insulation resistance data value (Ri) and the remaining insulation life for each reliability (50%, 90%, 95%, 99%) are displayed. Check these out.
Thereafter, when returning to the screen of the single processing menu screen 34 again, since the check boxes of the two processing menus so far are filled, the process proceeds to the third finding sentence confirmation menu.

Thus, the optimum finding sentence is automatically selected and displayed on the confirmation screen 53 from the finding sentences (6 patterns) registered in advance in consideration of the estimated remaining insulation life and the elapsed years of operation. So, check this and make corrections if necessary.
Thereafter, as described above, report printer output, report file storage, and the like are performed, and the processing operation when the insulation resistance value R is selected is terminated.

Next, an insulation lifetime estimation algorithm when the insulation index is the insulation resistance value R will be described with reference to the explanatory diagrams of FIGS. 8 and 9.
FIG. 8 shows the insulation resistance value R (database 1) of the high-pressure rotating machine 100 published in Non-Patent Document 1, and FIG. 9 is a regression line related to the insulation resistance value R obtained by multiple regression analysis of the data values shown in FIG. And an approximate expression.

In FIG. 8, as in the case of the dielectric loss rate tan δ (FIG. 5), among the insulation resistance values R in the database 1, each insulation resistance data value at the time of factory acceptance is shown by a plot (♦ mark). Even with the same number of years of operation, it varies widely.
Therefore, also in this case, each insulation resistance data value is subjected to multiple regression analysis by the statistical multiple regression analysis means 3.

In the result of the multiple regression analysis shown in FIG. 9, the insulation resistance data shown in FIG. 8 is divided into 11 zones depending on the length of operation years, and the low-frequency side insulation resistance of the normal distribution is divided for each zone. The insulation resistance value corresponding to the occurrence probability (50%, 10%, 5%, 1%) is required.
For example, all data can be efficiently processed by dividing into 11 zones of 12 to 13 years, 13 to 14 years,..., 22 to 23 years.

In FIG. 9, the insulation resistance value indicating the probability of occurrence of 50% is plotted by “■ in □”, and the insulation resistance value at each probability of occurrence of 10%, 5%, and 1% is “Δ”, respectively. It is plotted with “black triangle mark”, “× mark in ○”, and “* mark in □”.
By performing multiple regression analysis for each of the 11 zone plots for each of these four occurrence probabilities (50%, 10%, 5%, 1%), the following is an exponential function of multiple regression lines L11 to L14: Approximation formulas (11) to (14) are obtained.

y = 21044.2148 × exp (−0.2231x) (11)
y = 355.4772 × exp (−0.1695x) (12)
y = 214.7243 × exp (−0.1543x) (13)
y = 83.4080 × exp (−0.1258x) (14)

In the above approximate expressions (11) to (14), x indicates the elapsed operation time on the horizontal axis, and y indicates the data value (insulation resistance value R [MΩ]) on the vertical axis.
Further, among the approximate expressions (11) to (14) of the multiple regression lines L11 to L14, the approximate expression (11) of the straight line L11 corresponds to an occurrence probability of 50%, and the approximate expression of the straight lines L12 to L14 ( 12) to (14) correspond to the occurrence probabilities of 10%, 5%, and 1%, respectively.
It is experimentally found that the correlation coefficients of the regression lines L11 to L14 obtained by the multiple regression analysis as shown in FIG. 9 show sufficiently good values (for example, 0.7 to 0.8). ing.

  By the way, since the insulation resistance value Ri at the time of the periodic inspection of the high-pressure rotating machine 100 that is the subject of investigation at any site is the only insulation data at the time of investigation of the high-pressure rotating machine 100 that is the subject of investigation, It is treated as an insulation resistance value with a probability of occurrence of 50% over the years of operation of the target high-pressure rotating machine 100.

Accordingly, although not shown here, based on the results of the multiple regression analysis of FIG. 9, a simulation diagram for the insulation resistance value R similar to the residual simulation diagram for the dielectric loss rate tan δ described above (see FIG. 7) is similarly applied. Can be built.
Further, by applying “10 MΩ”, which has been conventionally adopted, as the lower limit threshold value of the insulation resistance value at which the dielectric breakdown level of the high-pressure rotating machine 100 is halved from the new level, the reliability of 50% is achieved. Not only the remaining insulation life but also each remaining insulation life of 90%, 95% and 99% can be calculated quantitatively.

  As described above, according to the present invention, the statistical analysis result obtained by performing multiple regression analysis on the insulation data of the insulation resistance of many high-voltage rotating machines published in Non-Patent Document 1, and the insulation resistance value R of the high-pressure rotating machine 100 In combination with the lower limit threshold (10MΩ) for determining the quality of the product, it is possible to estimate the remaining insulation life quantitatively, which could not be achieved by the prior art, and for the rational preventive maintenance of the high-pressure rotating machine 100, which is an important facility. Can contribute.

It is a block diagram which shows the residual insulation life estimation system of the high voltage | pressure rotary machine which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows the form example of the remaining insulation life diagnosis obtained by the remaining insulation life estimation method which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows the screen flow by the residual insulation life estimation system of the high voltage | pressure rotary machine which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows the screen flow by the residual insulation life estimation system of the high voltage | pressure rotary machine which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows the example of distribution with respect to the operation elapsed years of the dielectric material loss rate (tan-delta value) of the high voltage | pressure rotary machine contained in the database which concerns on Embodiment 1 of this invention. FIG. 6 is an explanatory diagram showing a regression line and an approximate expression regarding a dielectric loss rate (tan δ) obtained by statistically performing multiple regression analysis on the data values of FIG. 5. It is explanatory drawing which shows the simulation process for estimating a statistical residual insulation lifetime from the dielectric material loss rate (tan-delta value) of the investigated high voltage | pressure rotating machine. It is explanatory drawing which shows the example of distribution with respect to the elapsed operation years of the insulation resistance value of the high voltage | pressure rotary machine contained in the database which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows the regression line and approximate expression regarding the insulation resistance value which statistically analyzed the data value of FIG. 8 statistically.

Explanation of symbols

  1 database, 2 insulation data measurement means, 3 statistical multiple regression analysis means, 4 current level determination means, 5 remaining insulation life estimation means, 6 display means, 20 forms, 21 form title and menu part, 22 data description part, 23 Algorithm part, 24 Residual insulation life description part, 25 Additional observation description part, 31 Main menu screen, 32 Diagnostic data input screen, 33, 35 Diagnostic data readout screen, 34 Single processing menu screen, 41, 51 Residual life simulation Graph screen, 42, 52 Remaining life estimation table output screen, 43, 53 Observation text confirmation screen, 44, 54 File (report data) storage screen, L1-L4, L11-L14 Multiple regression line, P slope parameter, Ri insulation Measured resistance value, measured tan δi dielectric loss factor, Upper threshold of H tanδ.

Claims (8)

A system for estimating the remaining insulation life of a high-pressure rotating machine at an arbitrary age of operation,
A database containing a plurality of non-destructive insulation data and breakdown voltage data known in advance for the high-pressure rotating machine;
Insulation data measuring means for obtaining the current insulation resistance value or dielectric loss rate for the high-pressure rotating machine as insulation data;
Statistical multiple regression analysis means for estimating the slope parameter of the data value with respect to the time course by statistical multiple regression analysis of the database for each insulation index,
A current level determining means for determining a current level of the insulation data for each of a plurality of reliability levels based on the insulation data and the slope parameter;
A remaining insulation life estimation means for estimating a remaining insulation life of the high-pressure rotating machine based on the inclination parameter and the current level;
The remaining insulation lifetime estimation means until the insulation data at the respective reliability with respect to the future operating years in the future reaches a non-destructive insulation data threshold that is expected to deteriorate along the slope parameter and lead to destruction. A system for estimating a remaining insulation life of a high-pressure rotating machine, wherein a period is estimated as the remaining insulation life.   2. The high-pressure rotating machine according to claim 1, wherein the current level determination unit determines a current level of the insulation data for each reliability of 50%, 90%, 95%, and 99%. Residual insulation life estimation system.   3. The residual insulation life estimation system for a high-pressure rotating machine according to claim 1, further comprising display means for displaying an estimation result of the remaining insulation life estimation means. The display means displays a first menu screen for selecting a display mode, and displays a plurality of selection screens selected on the first menu screen,
The plurality of selection screens are:
A menu screen for selecting an insulation index of the database;
The system for estimating the remaining insulation life of a high-pressure rotating machine according to claim 3, further comprising: an input screen for inputting the insulation data at a current time point to the current level determining means. A method for estimating the remaining insulation life of a high-pressure rotating machine at an arbitrary age of operation,
Obtaining a database including a plurality of non-destructive insulation data and breakdown voltage data known in advance for the high-pressure rotating machine;
A second step of measuring the current insulation resistance value or dielectric loss rate for the high-pressure rotating machine as insulation data ;
A third step of estimating a slope parameter of a data value with respect to time by performing a statistical multiple regression analysis on the database for each insulation index;
A fourth step of determining a current level of the insulation data for each of a plurality of levels of reliability based on the insulation data and the slope parameter;
Based on the slope parameter and the current level, the insulation data at each reliability for future years of operation reaches a non-destructive insulation data threshold that is expected to degrade along with the slope parameter and lead to failure. And a fifth step of estimating the remaining insulation life as the remaining insulation life. A method for estimating the remaining insulation life of a high-pressure rotating machine. 6. The method according to claim 5 , wherein the current level is determined for each reliability of 50%, 90%, 95%, and 99%. The method for estimating a remaining insulation life of a high-pressure rotating machine according to claim 5 or 6 , further comprising a sixth step of displaying an estimation result of the remaining insulation life following the fifth step. . Before the sixth step,
A seventh step of displaying a first menu screen for selecting a display mode;
An eighth step of displaying a plurality of selection screens selected on the first menu screen,
The eighth step includes
A ninth step of displaying a menu screen for selecting an insulation index of the database;
The tenth step of displaying an input screen for inputting the insulation data at the current time point to the current level determination means, and a method for estimating a remaining insulation life of a high-pressure rotating machine according to claim 7.

Images