JP5792667B2 - Sensor diagnostic device and sensor diagnostic method - Google Patents

Description

  The present invention relates to a sensor diagnostic apparatus and a sensor diagnostic method.

  Equipment status monitoring in a nuclear power plant is performed by sensors (detectors) such as thermometers and pressure gauges attached to the equipment. For this reason, if the sensor drifts and the measured value deviates from the correct value, the state of the device cannot be accurately diagnosed. When any abnormality is recognized in the process value of the device measured by the sensor, it is necessary to distinguish between the device abnormality and the sensor drift in order to determine whether the device should be maintained or the sensor should be maintained. Further, when some abnormality is detected, even if the sensor drift is the cause, if it cannot be distinguished, there is a possibility that the nuclear power plant will be stopped. Therefore, the distinction between device abnormality and sensor drift is an important issue.

  As a method of diagnosing sensor drift during plant operation, a drift distribution is set using a sensor calibration database obtained during periodic inspections, and this drift distribution is used to detect sensor drift and device abnormalities. Some are distinguished (for example, refer to Patent Document 1).

  In addition, as another method for diagnosing sensor drift during plant operation, a step of receiving detector signals supplied from a plurality of detectors provided in the detection target and having a correlation with each other, and depending on the detection target Prepared using the estimation model for estimating the true value, estimating the true value based on the actual measurement value of the detector signal, estimating the drift amount over the full span of the detector, There is a detector calibration support method including a step of outputting an estimated true value, the actual measurement value, the drift characteristic, and a drift amount estimation result (see, for example, Patent Document 2). In this method, the drift amount of the sensor during plant operation is estimated based on a database that records the drift amount of the sensor obtained in the calibration test during regular inspections and statistics such as the average value and standard deviation. .

JP 2011-75373 A JP 2003-207373 A

  In the methods described in Patent Documents 1 and 2 described above, sensor drift during plant operation is evaluated using a sensor calibration database obtained during periodic inspection. For this reason, for example, in the case of a newly installed nuclear power plant that does not have a sensor calibration database, sensor drift evaluation during plant operation may not be performed during the period until the amount of calibration data required for evaluation is accumulated. .

  Patent Document 1 describes that, in the case of a newly installed nuclear power plant, the calibration data of the same sensor provided in another existing nuclear power plant is used. However, since such a sensor has a different use environment, there is a possibility that the tendency of the drift of the sensor to be different. For this reason, it is desirable to update the calibration database if the calibration data of the target sensor is obtained even if the calibration data of the sensor of another nuclear power plant is provisionally used. Such a method of updating the calibration database is not described in Patent Document 1 and Patent Document 2.

  The present invention has been made based on the above-mentioned matters, and its purpose is to evaluate drift distribution using provisional calibration data when sensor calibration data in a new plant or the like is not present or sufficient, The present invention provides a sensor diagnostic apparatus and a sensor diagnostic method capable of distinguishing between sensor drift and device abnormality other than sensor drift by updating the drift distribution using newly obtained calibration data.

  In order to solve the above problems, for example, the configuration described in the claims is adopted. The present application includes a plurality of means for solving the above-described problems. To give an example, during the plant operation, the drift amount of the sensor that measures the process value of the plant is estimated and the sensor is diagnosed. Temporary calibration database which is a sensor diagnostic apparatus and stores temporary calibration data which is calibration data of the same measurement object or other sensor of the same type as the sensor in another plant, and new calibration data obtained by calibration work of the sensor A new calibration database storing data, data extraction means for extracting the number of calibration data necessary for normal distribution test from the temporary calibration database and the new calibration database, and whether the calibration data extracted by the data extraction means follows a normal distribution Normal distribution verification means for verifying whether or not, drift that stores drift distribution data of the sensor When the calibration data is determined to follow a normal distribution by the cloth database and the normal distribution test means, the drift distribution of the sensor is calculated based on the extracted calibration data, and the drift of the sensor in the drift distribution database is calculated. Drift distribution update means for updating distribution data, process value input means for inputting the process value of the plant measured by the sensor, and estimated drift amount of the sensor using the process value input from the process value input means A drift estimator that calculates the estimated drift amount and the drift distribution data of the sensor updated in the drift distribution database, diagnoses the drift of the sensor, and outputs a diagnosis result It is characterized by comprising.

  ADVANTAGE OF THE INVENTION According to this invention, even when the calibration data of the sensor in a new plant etc. do not exist or are not enough, the sensor diagnosis which can distinguish sensor drift and apparatus abnormalities other than sensor drift can be implemented. Further, if calibration data is accumulated, the drift distribution can be updated to perform sensor diagnosis. As a result, it is possible to improve the accuracy of abnormality determination of equipment provided in the plant.

It is a block diagram of the apparatus which shows 1st Embodiment of the sensor diagnostic apparatus of this invention. It is a table | surface figure which shows an example of the calibration data stored in the temporary calibration database or new calibration database shown in FIG. It is a table | surface figure which shows an example of the drift distribution data of the sensor stored in the drift distribution database shown in FIG. It is a table | surface figure which shows an example of the process value of the sensor in the data input device shown in FIG. It is a table | surface figure which shows an example of the estimated drift amount of the sensor in the drift amount estimation means shown in FIG. It is a table | surface figure which shows an example of the upper limit and lower limit of the drift distribution of the sensor in the sensor state diagnostic means shown in FIG. It is a screen figure which shows an example of the display screen in the diagnostic result display apparatus shown in FIG. It is a flowchart figure which shows the content of the update process of the drift distribution in 1st Embodiment of the sensor diagnostic apparatus of this invention. It is a flowchart figure which shows the content of the sensor diagnostic process in 1st Embodiment of the sensor diagnostic apparatus of this invention. It is a flowchart figure which shows the content of the update process of the drift distribution in 2nd Embodiment of the sensor diagnostic apparatus of this invention.

  Embodiments of a sensor diagnostic apparatus and a sensor diagnostic method of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a device showing a first embodiment of a sensor diagnostic apparatus of the present invention.
In FIG. 1, the sensor diagnosis apparatus includes a provisional calibration database 1, a new calibration database 2, a data extraction unit 3, a normal distribution test unit 4, a drift distribution update unit 5, a drift distribution database 6, a sensor 7, and the like. A data input device 8, a drift amount estimating means 9, a sensor state diagnosing means 10, and a diagnostic result display device 11 are provided. Here, the data extraction means 3, the normal distribution verification means 4, the drift distribution update means 5, the drift amount estimation means 9, and the sensor state diagnosis means 10 are executed by an arithmetic means such as a computer, for example. It is. Further, the provisional calibration database 1, the new calibration database 2, the drift distribution database 6, and the data input device 8 may be provided separately from the computer, but may be included in the computer.

  The temporary calibration database 1 stores calibration data for temporarily creating a drift distribution of the diagnosis target sensor. For example, when the water supply flow rate sensor A of the new plant (A) is used as a diagnosis target sensor, the calibration data of the water supply flow rate sensor A of the existing plant (B) is stored because no calibration data exists at the time of new installation. In addition, since the water supply flow rate sensor A of the existing plant (b) is considered to be equivalent to the redundant water supply flow rate sensor B of the existing plant (b), the calibration data of the water supply flow rate sensor B of the existing plant (b) is also provisional. It may be stored in the calibration database 1. Note that the calibration data stored in the temporary calibration database 1 confirms in advance that the drift amount follows a normal distribution.

  The new calibration database 2 stores the calibration data when the calibration data of the diagnosis target sensor exists. For example, when the water supply flow rate sensor A of the new plant (A) is used as a diagnosis target sensor, the calibration data is stored when the calibration data of the water supply flow rate sensor A of the new plant (A) is obtained by the calibration operation. In addition, the feed water flow sensor A of the new plant (b) and the redundant feed water sensor B of the new plant (b) are considered to be equivalent, so the calibration data of the feed water flow sensor B of the new plant (b) is also newly calibrated. It may be stored in the database 2.

  FIG. 2 is a table showing an example of calibration data stored in the temporary calibration database or the new calibration database shown in FIG. In FIG. 2, the calibration data of the feed water flow sensor A describes the reference value added during calibration and the output value before calibration, and the difference between the reference value and the output value before calibration is described as the drift amount. Has been. After recording the drift amount, the sensor is calibrated as necessary. The calibration data is obtained for each calibration work for each sensor, and the calibration date and the calibration interval that is the period from the previous calibration date are described.

  Returning to FIG. 1, the data extraction unit 3 extracts calibration data from the temporary calibration database 1 and the new calibration database 2. In order to confirm whether or not the drift distribution of the sensor is a normal distribution, 30 points of calibration data are required as a guide. In the present embodiment, the data extracting means 3 extracts 30 points in order from the newest calibration date from the temporary calibration data and the new calibration data.

  The normal distribution verification unit 4 verifies whether the drift amount of the sensor follows the normal distribution for the calibration data extracted by the data extraction unit 3. For the normal distribution test, a generally well-known method such as the Kini power test is used. When the calibration data extracted by the data extraction means 3 is not a normal distribution by the verification means, the process returns to the data extraction means 3 and the next calibration data is newly extracted. Specifically, for example, 30 points are extracted from the temporary calibration data and the new calibration data in order from the second most recent calibration date. Thereafter, it is tested again whether or not the drift amount of the sensor follows the normal distribution with respect to the calibration data.

  The drift distribution update means 5 calculates the drift distribution using the calibration data determined to follow the normal distribution by the normal distribution verification means 4 and updates the data in the drift distribution database 6.

  The drift distribution database 6 stores drift distribution data of each sensor. FIG. 3 is a table showing an example of sensor drift distribution data stored in the drift distribution database shown in FIG. In FIG. 3, the drift distribution of the feed water flow rate sensor A is given by the average and standard deviation for each reference value.

  Returning to FIG. 1, the sensor 7 is a sensor (detector) that measures a process value of the plant, such as temperature, pressure, differential pressure, water level, and flow rate.

  The data input device 8 converts the plant process value measured by the sensor 7 into a digital signal and inputs it to the drift amount estimating means 9. FIG. 4 is a table showing an example of sensor process values in the data input device shown in FIG. In FIG. 4, the process value of each sensor (feed water flow sensor A, feed water flow sensor B, main steam flow sensor, condensate flow sensor) at each measurement date and time is shown.

  The drift amount estimation means 9 estimates the drift amount of the sensor 7 using the process value of the sensor 7 input by the data input device 8. Since the drift amount estimation method uses a known method such as a method using a neural network, a detailed description thereof will be omitted. As an outline of the drift amount estimation method, first, a process value at a time when drift of the sensor 7 does not occur, such as one month after the plant start-up, is input from the data input device 8 and learned as normal data. Next, the drift amount of the sensor is estimated by comparing the measured process value with normal data. FIG. 5 is a table showing an example of the estimated drift amount of the sensor in the drift amount estimating means shown in FIG. In FIG. 5, the estimated drift amount at each measurement date and time of the feed water flow rate sensor A is shown.

  Returning to FIG. 1, the sensor state diagnosis means 10 calculates the upper limit value and the lower limit value of the drift distribution from the average and standard deviation of the drift distribution of the target sensor stored in the drift distribution database 6. The sensor state is diagnosed by comparing the upper limit value and the lower limit value with the estimated drift amount obtained by the drift amount estimating means 9. Specifically, if the estimated drift amount is between the upper limit value and the lower limit value of the drift distribution, it is determined that the drift is normally generated and is normal. Conversely, if the estimated drift amount is not between the upper limit value and the lower limit value of the drift distribution, it is determined that there is an abnormality such as an abnormal drift or equipment failure.

  Further, the sensor state diagnosis means 10 outputs the diagnosis result to the diagnosis result display device 11. FIG. 6 is a table showing an example of the upper limit value and the lower limit value of the sensor drift distribution in the sensor state diagnosis means shown in FIG. In FIG. 6, the upper limit value and lower limit value of the drift distribution at each measurement date and time of the feed water flow rate sensor A are shown.

The upper limit value and the lower limit value of the drift distribution are calculated by the following procedure, for example.
Assuming that the average μ0 of the drift amount at the calibration interval t0 and the standard deviation σ0, the average μ and the standard deviation σ at an arbitrary time t1 after calibration are expressed by the following equations (1) and (2) from the reproducibility of the normal distribution. Is calculated according to
μ = μ0 × t1 / t0 (1)
σ = σ0 × (t / t0) ^ 0.5 (2)
If the drift distribution is a 95% confidence interval, the upper limit value of the drift distribution can be obtained by μ + 2σ and the lower limit value can be obtained by μ−2σ. These are calculated for each reference value, and the upper limit value and the lower limit value of the drift distribution in the sensor output value can be obtained by interpolation.

  The diagnosis result display device 11 displays the result of the sensor state diagnosis obtained by the sensor state diagnosis means 10, and displays the diagnosis result on, for example, an operation monitoring display in a plant. FIG. 7 is a screen diagram showing an example of an output screen in the diagnostic result display apparatus shown in FIG. In FIG. 7, the time-series fluctuations of the estimated drift amount, drift distribution upper limit value, and drift distribution lower limit value of the feed water flow sensor A are displayed, and the estimated drift amount is between the upper limit value and the lower limit value of the drift distribution. Therefore, the diagnosis result of the water supply flow rate sensor A is displayed as normal.

  Next, the drift distribution update process of the sensor diagnostic apparatus will be described with reference to FIG. FIG. 8 is a flowchart showing the contents of the drift distribution update process in the first embodiment of the sensor diagnostic apparatus of the present invention. Here, the case of the feed water flow sensor A will be described as an example.

  First, the data extraction means 3 inputs calibration data (step S101). Specifically, the data extraction unit 3 extracts calibration data from the temporary calibration database 1 and the new calibration database 2 and inputs the calibration data to the normal distribution test unit 4. For example, 30 points are input in order from the newest calibration date of the provisional calibration data and new calibration data of the feed water flow rate sensor A. That is, calibration data is input from the new calibration database 2 with priority.

  The normal distribution test means 4 performs a normal distribution test on the input calibration data (step S102). For the normal distribution test, a generally well-known method such as the Kini power test is used.

  The normal distribution test means 4 determines whether or not the result of the test performed in (Step S102) is a normal distribution (Step S103). If the result of the test is a normal distribution, the process proceeds to (Step S104). If the result is not a normal distribution, the process returns to (Step S101).

  Here, when returning to (Step S101), the data extraction means 3 extracts new calibration data in order, extracts 30 points of calibration data including the old calibration data, and inputs the calibration data to the normal distribution verification means 4. To do. Thereafter, (Step S102) and (Step S103) are performed, and these processes are repeated until the result of the test becomes a normal distribution in (Step S103). In other words, until the normal distribution is obtained, the extraction of the calibration data of the feed water flow rate sensor A is repeated, and as a method thereof, the new calibration data is removed in order, and 30 points including the old calibration data are substituted for the removed calibration data. The calibration data is extracted.

  The drift distribution updating means 5 calculates the average and the standard deviation using the calibration data of the feed water flow rate sensor A whose normal distribution is confirmed in (Step S103) (Step S104).

  The drift distribution updating means 5 updates the data in the drift distribution database 6 (step S105). The procedure from (Step S101) to (Step S105) is executed every time calibration data of the feed water flow rate sensor A is accumulated, for example, in a periodic inspection or the like.

  Therefore, for example, in the case of a newly installed plant, the drift distribution data in the initial state of the feed water flow rate sensor A is stored in the drift distribution database 6 based on the temporary calibration data from the temporary calibration database 1. Then, in the initial periodic inspection, when the calibration of the feed water flow rate sensor A is performed, the calibration data is stored in the new calibration database 2.

  Since the data extraction means 3 inputs calibration data to the normal distribution verification means 4 with priority from the new calibration database 2, updated calibration data including calibration data for the first periodic inspection is input. When the updated calibration data has a normal distribution, the drift distribution updating means 5 calculates the average and the standard deviation using the updated calibration data, and updates the data in the drift distribution database 6.

  Next, sensor diagnosis processing of the sensor diagnosis apparatus will be described with reference to FIG. FIG. 9 is a flowchart showing the contents of the sensor diagnosis process in the first embodiment of the sensor diagnosis apparatus of the present invention. Here, the case of the feed water flow sensor A will be described as an example.

  First, the data input device 8 inputs the process value of the plant measured by the feed water flow rate sensor A to the drift amount estimating means 9 (step S201). At the time of plant operation, the measurement signal from the feed water flow sensor A provided in the plant is input to the data input device 8, and the digitally converted process value (feed water flow rate measurement signal) is input to the drift amount estimating means 9. .

  The drift amount estimating means 9 estimates the drift amount of the feed water flow rate sensor A using the process value measured by the feed water flow rate sensor A (step S202). As shown in FIG. 5, the drift amount of the feed water flow sensor A is estimated at each measurement date and time.

  The sensor state diagnosis means 10 calculates the upper limit value and the lower limit value of the drift distribution from the average and standard deviation of the drift distribution of the feed water flow rate sensor A stored in the drift distribution database 6 (step S203).

  The sensor state diagnosis means 10 compares the estimated drift amount estimated in (Step S202) with the upper limit value and lower limit value of the drift distribution calculated in (Step S203) in order to diagnose the state of the feed water flow sensor A ( Step S204).

  By comparing (step S204), the sensor state diagnosis means 10 determines whether or not the estimated drift amount is less than the upper limit value of the drift distribution (step S205). If the estimated drift amount is less than the upper limit value of the drift distribution, the process proceeds to (Step S206). Otherwise, the process proceeds to (Step S208).

  The sensor state diagnosis unit 10 determines whether or not the estimated drift amount exceeds the lower limit value of the drift distribution (step S205). If the estimated drift amount exceeds the lower limit value of the drift distribution, the process proceeds to (Step S207). Otherwise, the process proceeds to (Step S208).

  Since the estimated drift amount is between the upper limit value and the lower limit value of the drift distribution, the sensor state diagnosis means 10 determines that the drift is normally generated, and determines that the feed water flow rate sensor A is normal (step S207).

  Since the estimated drift amount is not between the upper limit value and the lower limit value of the drift distribution, the sensor state diagnosis unit 10 determines that there is an abnormality such as an abnormal drift of the feed water flow rate sensor A or a device failure (step S208).

  The sensor state diagnosis means 10 outputs the result of the sensor state diagnosis to the diagnosis result display device 11 (step S209). The result of the sensor state diagnosis of the feed water flow rate sensor A is shown in FIG. Since the estimated drift amount is between the upper limit value and the lower limit value of the drift distribution, the diagnosis result of the feed water flow rate sensor A is displayed as normal. The procedure from (Step S201) to (Step S209) is executed each time the feed water flow rate sensor A inputs a process value, for example, during plant operation.

  According to the first embodiment of the sensor diagnostic device and sensor diagnostic method of the present invention described above, even if the calibration data of the sensor 7 in a new plant or the like does not exist or is not sufficient, devices other than sensor drift and sensor drift Sensor diagnosis capable of distinguishing from abnormality can be performed. Further, if calibration data of the sensor 7 is accumulated, the sensor diagnosis can be performed by updating the drift distribution. As a result, it is possible to improve the accuracy of abnormality determination of equipment provided in the plant.

  Further, according to the first embodiment of the sensor diagnostic apparatus and sensor diagnostic method of the present invention described above, when the calibration data of the sensor 7 is not present or not sufficient, the drift distribution is calculated using provisional calibration data. By evaluating and updating the drift distribution using newly obtained calibration data, it is possible to perform a sensor diagnosis that can distinguish between drift and device abnormality other than drift.

  The second embodiment of the sensor diagnostic apparatus and sensor diagnostic method of the present invention will be described below with reference to the drawings. FIG. 10 is a flowchart showing the contents of the drift distribution update process in the second embodiment of the sensor diagnostic apparatus of the present invention.

  In the second embodiment of the sensor diagnostic device and sensor diagnostic method of the present invention, the configuration of the sensor diagnostic device is the same as that of the first embodiment, but the content of the update processing of the drift distribution in the sensor diagnostic device is the same. Different.

Differences from the first embodiment in the contents of the drift distribution update processing in the present embodiment will be described. The data extraction means 3 in FIG. 1 extracts calibration data in the following procedure.
(1) All calibration data from the temporary calibration database 1 and the new calibration database 2 are input to the data extraction means 3.
(2) The data extraction means 3 sets a combination of calibration data of the number of data necessary to confirm whether or not the sensor drift distribution is a normal distribution from all the calibration data. For example, when there are 40 calibration data in total, all combinations of patterns for extracting 30 calibration data are set. Note that 30 points is an indication of the number of data necessary to confirm that the distribution is normal, and is not limited to 30 points.
(3) The combination calibration data for each set pattern is sent to the normal distribution test means 4.

  The normal distribution verification unit 4 verifies whether the drift amount of the sensor follows the normal distribution for the calibration data extracted by the data extraction unit 3. For the test of the normal distribution, a chi-square test is performed in which the chi-square value is calculated and the normal distribution is determined if the chi-ni value is equal to or smaller than a certain value.

  Next, the drift distribution update process of the sensor diagnostic apparatus according to the present embodiment will be described with reference to FIG. FIG. 10 is a flowchart showing the contents of the drift distribution update process in the second embodiment of the sensor diagnostic apparatus of the present invention. Here, the case of the feed water flow sensor A will be described as an example.

  First, the data extraction means 3 inputs calibration data (step S301). Specifically, all the calibration data from the temporary calibration database 1 and the new calibration database 2 are input to the data extraction means 3. For example, when the provisional calibration data and the new calibration data of the feed water flow rate sensor A are 40 points in total, all the calibration data are input.

  The data extraction means 3 sets a combination of calibration data (step S302). For example, when the number of data necessary to confirm that the distribution is normal is 30, all combinations of patterns for extracting 30 calibration data from 40 calibration data of the feed water flow rate sensor A are set.

  The normal distribution test means 4 calculates a chi-square value for one of the calibration data combinations set in (Step S302) (Step S303).

  The normal distribution test means 4 determines whether or not the chiini power value is less than a certain value (step S304). If the chiini power value is less than a certain value, the process proceeds to (Step S305), and otherwise, the process returns to (Step S302).

  Here, when returning to (Step S301), the data extraction means 3 extracts another one of the combinations of the calibration data set in (Step S302) and sends it to the normal distribution test means 4. Thereafter, (Step S303) and (Step S304) are performed, and these processes are repeated until the chi-square value becomes less than a certain value in (Step S304).

  The normal distribution test means 4 registers a combination of calibration data whose chiin power value is less than a certain value in (Step S304) as a combination candidate (Step S305).

  The normal distribution test means 4 determines whether or not the chi-in power value has been calculated for all combinations of the calibration data set by the data extraction means 3 (step S306). If the calculation of the chiini power value has been performed for all combinations, the process proceeds to (Step S307), and otherwise, the process returns to (Step S302).

  The normal distribution test means 4 determines the combination candidate having the smallest chiin power value among the combination candidates registered in (Step S305) as calibration data used for the drift distribution update means 5 (Step S307).

  The drift distribution updating means 5 calculates an average and a standard deviation by using a combination of calibration data of the feed water flow rate sensor A whose normal distribution is confirmed with a minimum Kini power value in (Step S307) (Step S308).

  The drift distribution update means 5 updates the drift distribution database 6 (step S309). The procedure from (Step S301) to (Step S309) is executed each time calibration data of the feed water flow rate sensor A is accumulated, for example, in a periodic inspection or the like.

  According to the first embodiment of the sensor diagnosis apparatus and sensor diagnosis method of the present invention described above, the same effects as those of the first embodiment can be obtained.

  In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. Each of the above-described configurations, functions, and the like may be realized by software by interpreting and executing a program that realizes each function by the processor.

DESCRIPTION OF SYMBOLS 1 Temporary calibration database 2 New calibration database 3 Data extraction means 4 Normal distribution test means 5 Drift distribution update means 6 Drift distribution database 7 Sensor 8 Data input device 9 Drift amount estimation means 10 Sensor state diagnosis means 11 Diagnosis result display apparatus

Claims (6)

A sensor diagnostic apparatus for diagnosing the sensor by estimating a drift amount of a sensor that measures a process value of the plant during plant operation,
A provisional calibration database storing provisional calibration data which is calibration data of the same measurement object or other sensor of the same type as the sensor in another plant;
A new calibration database storing new calibration data obtained in the calibration of the sensor;
Data extraction means for extracting the number of calibration data necessary for normal distribution test from the provisional calibration database and the new calibration database;
Normal distribution test means for testing whether the calibration data extracted by the data extraction means follows a normal distribution;
A drift distribution database storing drift distribution data of the sensor;
When it is determined by the normal distribution test means that the calibration data follows a normal distribution, the sensor drift distribution is calculated based on the extracted calibration data, and the sensor drift distribution data in the drift distribution database is calculated. Drift distribution updating means for updating;
Process value input means for inputting the process value of the plant measured by the sensor;
Drift estimating means for calculating an estimated drift amount of the sensor using the process value input from the process value input means;
A drift diagnosis unit that compares the estimated drift amount and the drift distribution data of the sensor updated in the drift distribution database, diagnoses the drift of the sensor, and outputs a diagnosis result; Diagnostic sensor device. In the sensor diagnostic apparatus according to claim 1,
The data extracting means extracts new calibration data again from the temporary calibration database and the new calibration database when the calibration data is determined not to follow a normal distribution by the normal distribution testing means. A characteristic sensor diagnostic device. The sensor diagnostic apparatus according to claim 1 or 2,
The process value of the plant is temperature, pressure, differential pressure, water level, or flow rate. The sensor diagnostic apparatus according to any one of claims 1 to 3,
A sensor diagnostic device, further comprising a diagnostic result display device for displaying a diagnostic result of the drift diagnostic means. A sensor diagnosis method for estimating an estimated drift amount of the sensor by estimating a drift amount of a sensor that measures a process value of the plant during plant operation,
Create drift distribution data as the initial state of the sensor using temporary calibration data that is the calibration data of the same measurement target or other sensor of the same type as the sensor in the other plant stored in the temporary calibration database, and create drift Storing in a distribution database;
Storing new calibration data obtained by calibration of the sensor in a new calibration database;
Using a part of the provisional calibration data and the new calibration data, a normal distribution test means performs a normal distribution test,
When the normal distribution test means determines that a part of the provisional calibration data and the new calibration data follow a normal distribution, the drift distribution update means calculates the drift distribution of the sensor based on these calibration data. Updating the drift distribution data as an initial state stored in the drift distribution database;
Calculating an estimated drift amount of the sensor during operation of the plant by using a process value of the plant measured by the sensor by a drift amount estimating means;
Using the updated drift distribution data of the sensor to evaluate and output an estimated drift amount of the sensor by a drift diagnosis means. In the sensor diagnostic method according to claim 5,
When it is determined by the normal distribution test means that a part of the provisional calibration data and the new calibration data do not follow a normal distribution, a part of the new calibration data is reduced to form new calibration data. The sensor diagnosis method further comprises the step of the normal distribution test means again performing a test of the normal distribution.

Images