JP2015228269A - Abnormality diagnostic device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To diagnose abnormality in a plant with high accuracy.SOLUTION: An abnormality diagnostic device comprises: a calculation unit 13 for calculating a Mahalanobis distance by using a command value that is a value of a variable presenting a plant state input from a plant and that is used for plant operation and a measurement value measured at the plant and a value of a variable extracted from group data associated with groups composed of a plurality of variables presenting the plant state; an abnormality determination unit 14 for determining probability of abnormality in the plant by using the Mahalanobis distance calculated by the calculation unit 13; an extraction unit 15 for extracting, when the abnormality determination unit 14 determines that there is probability of abnormality, a cause variable, that is a candidate for a cause of abnormality, from a plurality of variables and extracting a strong correlation variable that is highly correlated with the extracted cause variable; an operation determination unit 16 for referring to the command value input from the plant and determining whether or not there has been operation for the cause variable and the strong correlation variable; and a diagnosis unit 18 for determining abnormality according to a determination result in the operation determination unit 16.

Description

  The present invention relates to an abnormality diagnosis device for diagnosing plant abnormality.

  In the plant, a plurality of devices are operated under various conditions. In the plant, the state of the plant is measured by a plurality of sensors. An abnormality of the plant can be detected from the plurality of condition values and the measurement values of the plurality of sensors. For example, the upper and lower limit values (threshold values) determined for the measurement values of the sensors by the measurement values of the sensors are set. The mainstream method is to determine that there is an abnormality when the number is exceeded. On the other hand, when the upper and lower limit values are set for each measured value of each sensor, it is necessary to increase the number of upper and lower limit values when the number of sensors used is increased, and there is a problem in that the management of values for abnormality determination becomes complicated.

  On the other hand, in recent years, there is also a technique for diagnosing an abnormality using a pattern recognition technique called MT method (Mahalanobis Taguchi method). The MT method is for diagnosing whether or not the state is “same as usual”, and the Mahalanobis distance of the object to be diagnosed is determined from the Mahalanobis space based on the center of the Mahalanobis space that represents the same state as usual. This is a method for determining an abnormality when going far. Specifically, there is one that divides sensor measurement data into a plurality of groups, generates a unit space for each group, and diagnoses an abnormality using the MT method (see, for example, Patent Document 1).

  The plant may be operated while switching a plurality of condition values depending on the situation. However, when the condition value is changed, the measurement value by the sensor also changes, so that the threshold value to be set needs to be changed. In addition, there are various changes such as changes in condition values due to intentional operations in plants and changes in measured values due to seasonal fluctuations. An abnormality may be diagnosed, and the accuracy of the abnormality diagnosis will be reduced.

  Even when the MT method is used, if the condition value is changed, it is necessary to change the judgment criteria such as switching the unit space to be used. However, all the condition values and measured values input from the plant are used. When the unit space is regularly updated, abnormalities and signs of abnormalities may be included in the unit space to be updated, which reduces the accuracy of abnormality diagnosis.

JP 2010-181188 A

  As described above, it has been difficult to diagnose plant abnormalities with high accuracy.

  In view of the above problems, an object of the present invention is to provide an abnormality diagnosis apparatus that diagnoses an abnormality of a plant with high accuracy.

  In order to achieve the above object, the present invention relates to an abnormality diagnosis apparatus for diagnosing a plant abnormality, which is a command value used for operation of a plant, which is a value of a variable representing a state of the plant input from the plant. The calculation unit that calculates the Mahalanobis distance using the measurement value measured in step 1 and the value of the variable that is stored in the group data storage unit and extracted from the group data related to the group that includes the plurality of variables that represent the state of the plant And an abnormality determination unit that determines the possibility of abnormality in the plant using the Mahalanobis distance calculated by the calculation unit, and when the abnormality determination unit determines that there is a possibility of abnormality, the cause of the abnormality from a plurality of variables The cause variable that is a candidate for the extraction, and the extraction unit that extracts the strong correlation variable highly correlated with the extracted cause variable and input from the plant With reference to the command value includes cause variable and the determining operation determining unit whether the operation on the strongly correlated variables, and a diagnosis unit for determining an abnormality in accordance with the determination result in the operation determination unit.

  Further, the present invention is an abnormality diagnosis apparatus for diagnosing plant abnormality, wherein a command value used for operation of the plant, which is a value of a variable representing a state of the plant input from the plant, and a measured value measured by the plant And a calculation unit that calculates a Mahalanobis distance using a value of a variable that is stored in a group data storage unit and that is extracted from group data related to a group that includes a plurality of variables that represent a plant state, and the calculation unit includes: An abnormality determination unit that determines the possibility of abnormality in the plant using the calculated Mahalanobis distance, and when the abnormality determination unit determines that there is a possibility of abnormality, a cause variable that is a candidate for the cause of abnormality from a plurality of variables Extractor that extracts a strong correlation variable that is highly correlated with the extracted cause variable, a change in the value of the cause variable, and a change in the value of the strong correlation variable Anda determining interlock determining unit interlocking trend of changes in, and a diagnosis unit determines abnormality in accordance with the determination result in the interlocking determination unit.

  The present invention further includes an interlocking determination unit that determines the interlocking of the tendency of change between the change in the value of the cause variable and the change in the value of the strong correlation variable, and the diagnosis unit includes the determination result of the operation determination unit An abnormality is determined according to a determination result in the interlock determination unit.

  According to the present invention, it is possible to improve the accuracy of plant abnormality diagnosis.

It is a block diagram explaining the abnormality diagnosis apparatus concerning an embodiment. It is a data block diagram explaining an example of the accumulation | storage data utilized with an abnormality diagnosis apparatus. It is a flowchart explaining an example of the process in an abnormality diagnosis apparatus. It is a conceptual diagram explaining the abnormality diagnosis using MT method. It is a graph explaining the abnormality diagnosis in an abnormality diagnosis apparatus.

  The abnormality diagnosis apparatus according to the embodiment is an abnormality diagnosis apparatus that diagnoses an abnormality in plant operation. For example, the plant diagnosed by the abnormality diagnosis device is a power plant. The power plant includes a plurality of devices (pumps, valves, etc.), and values for controlling these devices are set as target values. This target value is, for example, the pressure of the pump, opening / closing of a valve, or the like. In addition, the power plant includes a plurality of sensors, and each sensor measures the intake air amount, the exhaust air amount, the power generation amount, and the like.

  Furthermore, in a power plant, various modes can be set, and the target value set for each device and the measured value measured by each sensor differ depending on this mode. For example, in a power plant, it is not always necessary to generate a certain amount of power, and power is generated in accordance with power demand. Therefore, a plurality of modes are defined according to the time zone, day of the week, season, etc., and the target value (set value) of each device differs for each set mode. Moreover, since the target value of each device is different, the measured value of each sensor is different for each set mode.

  As shown in FIG. 1, the abnormality diagnosis device 1 according to the embodiment includes an update unit 11 that updates accumulated data with new variables input from a plant, a generation unit 12 that generates a group used for abnormality diagnosis, A variable included in the group data 22, a calculation unit 13 that calculates a Mahalanobis distance using a newly input variable, and an abnormality that determines a possibility of abnormality by comparing the Mahalanobis distance calculated by the calculation unit 13 with a determination value If the determination unit 14 and the abnormality determination unit 14 determine that there is a possibility of abnormality, the extraction unit 15 extracts a cause variable that is a candidate for the cause of the abnormality and a strongly correlated variable that is highly correlated with the cause variable. An operation determination unit 16 that determines the presence / absence of an operation related to the cause variable and the strong correlation variable, an interlock determination unit 17 that determines an interlock of trends between the change in the cause variable and the change in the strong correlation variable, and an operation determination Part And a diagnosis unit 18 for diagnosing the abnormality by using the determination result with the interlocking determination unit 17 and 6.

  The abnormality diagnosis apparatus 1 is, for example, an information processing apparatus including a central processing unit (CPU) 10 and a storage device 20, and the abnormality diagnosis program P stored in the storage device 20 is read and executed. As illustrated in FIG. 1, an update unit 11, a generation unit 12, a calculation unit 13, an abnormality determination unit 14, an extraction unit 15, an operation determination unit 16, an interlock determination unit 17, and a diagnosis unit 18 are mounted on the CPU 10. Further, in the example shown in FIG. 1, accumulated data 21 and group data 22 are stored in the storage device 20. Further, the abnormality diagnosis device 1 is connected to an input device 2 such as a keyboard, a mouse, an operation button, a touch panel or the like for inputting an operation, and an output device 3 such as a display or a speaker for outputting an abnormality diagnosis process or result. It is connected.

  When the value of each variable is newly input from the plant, the update unit 11 generates a new record including the value of each input variable, adds the generated record, and stores the accumulated data 21 stored in the storage device 20. Update.

  The accumulated data 21 is data obtained by accumulating variable records representing past plant states. The variable of the accumulated data is a condition value for plant operation and a measured value measured in the plant in the case of the condition value. In the example shown in FIG. 2, the accumulated data includes target values (variables 1 and 2) set for each device in the plant, which are condition values for plant operation, and measured values (in this mode and target values) It has a record that associates variables 4, 5), etc. In the example of accumulated data shown in FIG. 2, each record is given a sample number such as time.

  The generation unit 12 reads the accumulated data 21 from the storage device 20 at a predetermined timing for generating a group, extracts a record in which a variable satisfies a predetermined extraction condition, generates a group, and generates an abnormality in the plant for the generated group. Determination values used for diagnosis are set and stored in the storage device 20 as group data 22. This extraction condition is input via the input device 2, for example. The determination value is a value representing the Mahalanobis space obtained for this group, and is input via the input device 2.

  The predetermined timing at which the generation unit 12 generates a group is, for example, a timing at which an operation for newly executing an abnormality diagnosis is input via the input device 2 or an operation for changing the group used for the abnormality diagnosis is input. This is the timing input via the device 2. The generation unit 12 may use, for example, any one of the values of the variable 1 and the variable 2 (target value), the value of the variable 4 and the variable 5 (measured value) as an extraction condition, and a combination of values of a plurality of variables. May be used as the extraction condition.

  The group data 22 is data having a history of variables representing past plant states in the case of a predetermined condition for specifying a group (extraction conditions used in the generation unit 12) and a determination value set for this group. .

  The calculation unit 13 reads the group data 22 from the storage device 20 at a predetermined timing for diagnosing an abnormality, and extracts the value of each variable included in the group data 22. Here, the predetermined timing for diagnosing abnormality is, for example, a regular timing.

  Further, the calculation unit 13 obtains the Mahalanobis distance using the value of each variable extracted from the group data 22 and the new variable input from the plant at a predetermined timing for diagnosing the abnormality, and determines the obtained value as an abnormality. To the unit 14. Specifically, (1) First, the calculation unit 13 obtains an average value and a standard deviation for each variable. (2) Thereafter, the calculation unit 13 standardizes the data using the value of each variable and the average value and standard deviation obtained for each variable, and obtains a normalized value for each variable. (3) Subsequently, the calculation unit 13 obtains a correlation matrix for each variable by using the normalized value obtained for each variable, and obtains an inverse matrix of the correlation matrix. (4) Finally, the calculation unit 13 obtains the Mahalanobis distance using the obtained inverse matrix.

  When the abnormality determination unit 14 inputs the Mahalanobis distance from the calculation unit 13, the abnormality determination unit 14 extracts the determination value included in the group data 22 read from the storage device 20, and the determination that the Mahalanobis distance input from the calculation unit 13 is extracted from the group data 22. Compare with the value to determine the possibility of abnormality in the plant. The abnormality determination unit 14 outputs the determination result to the extraction unit 15 and the output device 3.

  For example, as shown in FIG. 4, when determining an abnormality using the MT method, when the Mahalanobis distance obtained by the calculation unit 13 is smaller than a set determination value, that is, within the Mahalanobis space, the abnormality determination unit No. 14 determines that the current plant state is the same as the normal state, and the plant is operating normally. On the other hand, when the Mahalanobis distance obtained by the calculation unit 13 is greater than the set determination value, that is, when the Mahalanobis space is out of the Mahalanobis space, the abnormality determination unit 14 is in a state where the current plant state is different from the normal state. It is determined that there is an abnormality in the plant.

  When the extraction unit 15 inputs a determination result with an abnormality candidate, the extraction unit 15 extracts a cause variable that is a candidate for an abnormality cause from the values of a plurality of variables, and extracts a strongly correlated variable highly correlated with the cause variable, The extraction result is output to the operation determination unit 16 and the interlock determination unit 17.

  In order to extract a candidate for an abnormality cause, for example, the abnormality diagnosis apparatus 1 has an ideal value of each variable value or an ideal range of each variable value in the group data 22, and the extraction unit 15 uses the ideal value. Or, a cause variable that is a candidate for abnormality may be extracted in comparison with the ideal range.

  Moreover, in order to extract a strong correlation variable, for example, in the abnormality diagnosis apparatus 1, when using the MT method, between variables included as elements in a correlation matrix formed of a group that is a Mahalanobis space (unit space) Using the correlation coefficient representing the strength of the relationship, the extraction unit 15 can extract a variable highly correlated with each variable. In addition, in order to extract a strong correlation function, the abnormality diagnosis apparatus 1 stores correlation data (not shown) having a list to which correlated variables are associated in the storage device 20, and the extraction unit 15 stores the correlation data. Alternatively, a variable associated with the cause variable may be extracted as a strongly correlated variable.

  When the cause variable and the strong correlation variable are input from the extraction unit 15, the operation determination unit 16 refers to the value of the variable input from the plant, determines whether the cause variable and the strong correlation variable are operated, and determines The result is output to the diagnosis unit 18. For example, it is assumed that a measured value that changes according to opening / closing of a certain valve is extracted as a cause variable, and a command value for opening / closing this valve is extracted as a strong correlation variable. At this time, if the valve is not operated, that is, if the command value that is a strongly correlated variable has not changed, it can be predicted that the cause variable has changed due to some abnormality. On the other hand, when the valve is operated, that is, when the command value, which is a strongly correlated variable, has changed, the measured value has also changed due to the valve operation (change in the command value), so it is not abnormal. Can be predicted.

  When the cause variable and the strong correlation variable are input from the extraction unit 15, the interlock determination unit 17 refers to the value of the variable input from the plant and determines whether the value of the cause variable and the value of the correlation variable are interlocked. The determination result is output to the diagnosis unit 18. For example, when a valve and a pump fluctuate in conjunction with each other when operating normally, it is assumed that the measured value of this valve is extracted as a cause variable and the measured value of the pump is extracted as a strongly correlated variable. At this time, if the measured value of the pump is changed in conjunction with the variation of the measured value of the valve, it can be predicted that there is no abnormality. On the other hand, when the pump does not change in conjunction with the change in the measured value of the valve, it can be predicted that the cause variable has changed due to some abnormality. Further, for example, it is assumed that a certain measurement value is extracted as a cause variable, and a command value (setting value) related thereto is extracted as a strong correlation variable. At this time, when the measured value and the set value are linked as shown in FIG. 5A, it is predicted that the measurement value is normal, and when the measured value and the set value are not linked as shown in FIG. 5B. Is expected to be abnormal.

  The diagnosis unit 18 diagnoses an abnormality using the determination result in the operation determination unit 16 and the determination result in the interlock determination unit 17, and outputs the diagnosis result to the output device 3. Specifically, when it is determined in the operation determination unit 16 that an operation related to the cause variable or the strong correlation variable is performed, and the value of the cause variable and the value of the strong correlation variable are linked in the interlock determination unit 17. If determined, the diagnosis unit 18 diagnoses that the plant is operating normally. On the other hand, when the operation determination unit 16 determines that an operation related to the cause variable or the strong correlation variable is not performed, or the interlock determination unit 17 determines that the value of the cause variable and the value of the strong correlation variable are not interlocked. When the diagnosis is made, the diagnosis unit 18 diagnoses that an abnormality has occurred in the plant.

  In addition, the abnormality diagnosis apparatus 1 may be comprised from several information processing apparatus, for example, only the update part 11 may be contained in the information processing apparatus different from the other process parts 12-18. Further, only a part of the data stored in the storage device 20 may be stored in an external storage device.

  Furthermore, in the above description, the type of group data 22 stored in the storage device 20 is not mentioned, but the storage device 20 may store group data 22 for a plurality of groups. In that case, the calculation unit 13 reads the group data 22 of the group to be diagnosed from the storage device 20 and uses it.

  Next, a process for diagnosing an abnormality in the abnormality diagnosis apparatus 1 will be described using the flowchart shown in FIG.

  At the timing when the generation unit 12 generates the group data 22, the accumulated data 21 is read from the storage device 20 to extract the record value satisfying the extraction condition, and the determination value is set to generate the group data 22 and store it. It memorize | stores in the apparatus 20 (S1).

  Thereafter, the calculation unit 13 calculates the Mahalanobis distance using the value of the variable input from the plant and the value of the variable included in the group data 22 (S2). Subsequently, the abnormality determination unit 14 determines the possibility of abnormality by comparing the Mahalanobis distance calculated by the calculation unit 13 with a determination value included in the group data 22 (S3).

  When it is determined that there is a possibility of abnormality (YES in S3), the extraction unit 15 extracts a cause variable that is a candidate variable of the cause of the abnormality (S4). Further, the extraction unit 15 extracts a strong correlation variable having a high correlation with the extracted cause variable (S5).

  Subsequently, the operation determination unit 16 determines whether or not an operation related to the cause variable or the strong correlation variable has been performed (S6). The interlock determination unit 17 determines whether or not the change in the value of the cause variable and the tendency of the change in the value of the strong correlation variable are interlocked (S7).

  If the abnormality determination unit 14 determines that there is no possibility of abnormality (NO in S3), the abnormality determination unit 14 determines that the plant is operating normally, and outputs the determination result to the output device 3. The process returns to step S2 (S8). When it is determined that the operation is determined by the operation determination unit 16 and the value of each variable is determined to be interlocked by the interlock determination unit 17 (YES in S6 and YES in S7), the diagnosis unit 18 indicates that the plant is normal. It diagnoses that it is driving, outputs the diagnosis result to the output device 3, and returns to the process of step S2 (S8).

  On the other hand, when the operation determining unit 16 determines that the operation is not performed (NO in S6) or when the interlock determining unit 17 determines that the operation is not interlocked (NO in S7), the diagnosis unit 18 Diagnose that an abnormality has occurred in the plant, output the diagnosis result to the output device 3, and return to the process of step S2 (S9).

  In the above description, the abnormality diagnosis device 1 includes the operation determination unit 16 and the interlock determination unit 17, and the diagnosis unit 18 diagnoses an abnormality using the determination results of the operation determination unit 16 and the interlock determination unit 17. Explained as what to do. However, an abnormality may be diagnosed with the determination result of either the operation determination unit 16 or the interlock determination unit 17.

  3 does not show that the generation of the group data 22 is repeated, the group data 22 may be generated repeatedly at a predetermined timing such as periodically.

  As described above, in the abnormality diagnosis device according to the present invention, after determining the possibility of abnormality, the abnormality is diagnosed using the value of each variable, change in each value, etc. It is possible to diagnose plant abnormalities.

  As mentioned above, although this invention was demonstrated in detail using embodiment, this invention is not limited to embodiment described in this specification. The scope of the present invention is determined by the description of the claims and the scope equivalent to the description of the claims.

1 ... Abnormality diagnosis device 10 ... CPU
DESCRIPTION OF SYMBOLS 11 ... Update part 12 ... Generation | occurrence | production part 13 ... Calculation part 14 ... Abnormality determination part 15 ... Extraction part 16 ... Operation determination part 17 ... Interlocking determination part 18 ... Diagnosis part 20 ... Storage device (group data storage part)
21 ... Accumulated data 22 ... Group data P ... Abnormality diagnosis program 2 ... Input device 3 ... Output device

Claims (1)

An abnormality diagnosis device for diagnosing plant abnormality,
A command value used for operation of the plant, which is a value of a variable representing the state of the plant input from the plant, and a measured value measured in the plant, and a plurality of variables that are stored in the group data storage unit and represent the state of the plant A calculation unit for calculating the Mahalanobis distance using the value of the variable extracted from the group data regarding the group to be configured;
An abnormality determination unit that determines the possibility of abnormality in the plant using the Mahalanobis distance calculated by the calculation unit;
When the abnormality determination unit determines that there is a possibility of abnormality, an extraction unit that extracts a cause variable that is a candidate for the cause of abnormality from a plurality of variables and extracts a strongly correlated variable highly correlated with the extracted cause variable When,
An operation determination unit that refers to a command value input from the plant and determines whether or not there is an operation for a cause variable and a strong correlation variable,
A diagnosis unit that determines an abnormality according to a determination result in the operation determination unit;
An abnormality diagnosis apparatus comprising: