Plant monitoring device, plant monitoring method, and plant monitoring program

Abstract

This plant monitoring device for monitoring a plant is provided with: a measured data acquisition unit which, for each prescribed period, acquires measured data of multiple variables indicating the state of the plant; a comparison unit which compares a diagnostic threshold value, which is for diagnosing anomalies in the plant, and a deviation indicator value, which indicates the deviation between measured data and a reference dataset that relates to the aforementioned multiple variables; a reference data updating unit which updates the reference dataset; and an operation mode switching unit which switches the operation mode of the plant monitoring device between a monitoring mode, in which plant monitoring is performed on the basis of the comparison results by the comparison unit and a learning mode, in which at least the measured data for which the deviation indicator value is greater than the diagnostic threshold value is imported into the reference dataset by the reference data update unit. When the deviation indicator value is greater than the diagnostic threshold value and a learning mode transition condition is satisfied during monitoring-mode operation of the plant monitoring device, the operation mode switching unit switches the operation mode from the monitoring mode to the learning mode.

Description

Technical Field

[0001] The present disclosure relates to a plant monitoring device, a plant monitoring method, and a plant monitoring program. This application claims priority based on Japanese Patent Application No. 2021-125038 filed with the Japan Patent Office on July 30, 2021, the content of which is incorporated herein by reference.

Background Art

[0002] Abnormal diagnosis of a plant may be performed based on a deviation index value indicating a deviation between a reference data set (reference data set) of variables (such as state quantities that can be acquired by sensors) indicating the state of the plant and measurement data for the variables.

[0003] Patent Document 1 describes inputting measurement data of a process acquired in a plant into a process monitoring model, calculating a statistic that is a statistical error index or a statistical dispersion index, and determining whether the process is normal or abnormal based on the magnitude of the deviation of the statistic from a preset normal process state.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] By the way, when an abnormality occurs in a plant, the deviation index value indicating the deviation between the above-mentioned reference data set and the measurement data becomes larger than when the plant is normal. Therefore, abnormal diagnosis of the plant is possible based on comparison between the deviation index value and a threshold value.

[0006] On the one hand, even when there is actually no abnormality in the plant, due to changes in the operating state of the plant or the like, the measured value by the sensor may show a value different from before. Even in such a case, it is expected that the deviation index value calculated from the measurement data will increase, but the operation of the plant can continue. Alternatively, even if an abnormality occurs in the plant equipment and the deviation index value calculated from the measurement data increases, there may be cases where it can be determined based on past experience or the like that the operation of the plant can continue.

[0007] When the operation of the plant is continued when the above-mentioned deviation index value is large in this way, even if another abnormality occurs in the plant later, the magnitude of the deviation index value hardly changes, and it may become difficult to detect the abnormality of the plant.

[0008] In view of the above circumstances, at least one embodiment of the present invention aims to provide a plant monitoring device, a plant monitoring method, and a plant monitoring program capable of robust abnormality diagnosis.

Means for Solving the Problems

[0009] The plant monitoring device according to at least one embodiment of the present invention is a plant monitoring device for monitoring a plant, a measurement data acquisition unit configured to acquire measurement data of a plurality of variables indicating the state of the plant at regular intervals; a comparison unit configured to compare a deviation index value indicating the deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant; a reference data update unit configured to update the reference data set; An operation mode switching unit configured to switch the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result by the comparison unit and a learning mode for incorporating the measurement data into the reference data set at least when the deviation index value is greater than the diagnostic threshold by the reference data updating unit, The operation mode switching unit is configured to switch the operation mode from the monitoring mode to the learning mode when the deviation index value is greater than the diagnostic threshold and the learning mode transition condition is satisfied during the operation of the plant monitoring device in the monitoring mode.

[0010] Further, a plant monitoring method according to at least one embodiment of the present invention is A plant monitoring method using a plant monitoring device for monitoring a plant, A step of acquiring measurement data of a plurality of variables indicating the state of the plant at regular intervals, A comparison step of comparing a deviation index value indicating a deviation between the measurement data and a reference data set which is a set of reference data related to the plurality of variables with a diagnostic threshold for diagnosing an abnormality of the plant, A step of updating the reference data set, An operation mode switching step of switching the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result in the comparison step and a learning mode for incorporating the measurement data into the reference data set at least when the deviation index value is greater than the diagnostic threshold, In the operation mode switching step, when the deviation index value is greater than the diagnostic threshold and the learning mode transition condition is satisfied during the operation of the plant monitoring device in the monitoring mode, the operation mode is switched from the monitoring mode to the learning mode.

[0011] Further, a plant monitoring program according to at least one embodiment of the present invention is A plant monitoring program for operating a plant monitoring device for monitoring a plant, Cause a computer to acquire measurement data of a plurality of variables indicating the state of the plant at regular intervals, compare a deviation index value indicating a deviation between the measurement data and a reference data set that is a set of reference data related to the plurality of variables with a diagnostic threshold value for diagnosing an abnormality of the plant, update the reference data set, and switch the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result in the comparison procedure and a learning mode for incorporating the measurement data into the reference data set at least when the deviation index value is greater than the diagnostic threshold value. The program for causing the computer to execute the procedure is In the procedure for switching the operation mode, when the deviation index value is greater than the diagnostic threshold value and the learning mode transition condition is satisfied during the operation of the plant monitoring device in the monitoring mode, the operation mode is switched from the monitoring mode to the learning mode.

Advantages of the Invention

[0012] According to at least one embodiment of the present invention, there is provided a plant monitoring device, a plant monitoring method, and a plant monitoring program capable of performing robust abnormality diagnosis.

Brief Description of the Drawings

[0013]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

[0014] Hereinafter, some embodiments of the present invention will be described with reference to the accompanying drawings. However, the dimensions, materials, shapes, relative arrangements, etc. of the components described as embodiments or shown in the drawings are not intended to limit the scope of the present invention, but are merely illustrative examples.

[0015] (Configuration example of the plant to be monitored) FIG. 1 is a schematic configuration diagram of an example of a plant to which a plant monitoring device, a plant monitoring method, or a plant monitoring program according to some embodiments is applied. The plant 1 shown in FIG. 1 is a gas turbine combined cycle (GTCC) plant (combined cycle plant) including a gas turbine facility 2 (gas turbine), a heat recovery steam generator (HRSG) 18 (boiler), and a steam turbine facility 12 (steam turbine).

[0016] The gas turbine facility 2 includes a compressor 4 for compressing air, a combustor 6 for burning fuel together with the compressed air from the compressor 4, and a turbine 8 configured to be driven by the combustion gas generated in the combustor 6. A generator 10 is connected to the rotor of the turbine 8, and the generator 10 is rotationally driven by the turbine 8. The combustion gas that has completed its work in the turbine 8 is discharged from the turbine 8 as exhaust gas.

[0017] The exhaust heat recovery boiler 18 is configured to generate steam by the heat of the exhaust gas from the gas turbine facility 2. The exhaust heat recovery boiler 18 has an exhaust duct into which the exhaust gas from the gas turbine facility 2 is introduced, and a heat exchanger provided so as to pass through the inside of the exhaust duct. Condensate from the condenser 20 of the steam turbine facility 12 described later is introduced into the heat exchanger, and in this heat exchanger, steam is generated by heat exchange between the condensate and the exhaust gas flowing through the above-described exhaust duct. Note that the exhaust gas that has flowed through the exhaust duct of the exhaust heat recovery boiler 18 and passed through the heat exchanger may be discharged from a chimney (not shown) or the like.

[0018] The steam turbine facility 12 shown in FIG. 1 includes a turbine 14 configured to be driven by the steam from the exhaust heat recovery boiler 18. A generator 16 is connected to the rotor of the turbine 14, and the generator 16 is rotationally driven by the turbine 14. The steam that has completed work in the turbine 14 is guided to the condenser 20 and condensed, and then returns to the exhaust heat recovery boiler 18 and is heated again by heat exchange with the exhaust gas.

[0019] In some embodiments, the plant to be monitored may be a combined cycle plant as described above. In some embodiments, the plant to be monitored may be a plant including either a gas turbine or a steam turbine.

[0020] The plant 1 is provided with a measurement unit 50 (see FIG. 2) for measuring a plurality of variables indicating the state of the plant. The measurement unit 50 may include a plurality of sensors configured to measure the plurality of variables indicating the state of the plant 1 respectively.

[0021] When Plant 1 includes a gas turbine, the measurement unit 50 may include a sensor configured to measure any one of the rotor rotation speed of the gas turbine, the temperature of each stage blade passage, the average temperature of the blade passage, the turbine inlet pressure, the turbine outlet pressure, the generator output, and the differential pressure of the intake filter as a variable indicating the state of the plant. When Plant 1 is a combined cycle plant including a gas turbine and a steam turbine, the measurement unit 50 may include a sensor configured to measure the pressure of the exhaust duct of the waste heat recovery boiler 18 (the exhaust duct into which the exhaust gas from the gas turbine facility 2 is introduced).

[0022] (Configuration of Plant Monitoring Device) FIG. 2 is a schematic configuration diagram of a plant monitoring device according to an embodiment. The plant monitoring device 30 shown in FIG. 2 is configured to monitor the plant based on the measured values of a plurality of variables indicating the state of the plant measured by the measurement unit 50.

[0023] As shown in FIG. 2, the plant monitoring device 30 according to an embodiment includes a measurement data acquisition unit 32, a reference data acquisition unit 34, a deviation index value calculation unit 36, a comparison unit 38, a reference data update unit 40, and an operation mode switching unit 42. The plant monitoring device 30 may include an alarm output unit 44.

[0024] The plant monitoring device 30 includes a computer having a processor (such as a CPU), a main storage device (memory device; such as a RAM), an auxiliary storage device, and an interface. The plant monitoring device 30 is adapted to receive signals from the measurement unit 50, the input device 46 (such as a keyboard or a mouse), or the storage unit 48 via the interface. The processor is configured to process the signals received in this way. Further, the processor is configured to process a program developed in the main storage device. Thereby, the functions of the above-described respective functional units (the measurement data acquisition unit 32, etc.) are realized.

[0025] The processing content in the plant monitoring device 30 is implemented as a program executed by a processor. The program may be stored, for example, in an auxiliary storage device. When the program is executed, these programs are expanded in the main storage device. The processor is configured to read the program from the main storage device and execute the instructions included in the program.

[0026] Note that the storage unit 48 may include the main storage device or the auxiliary storage device of the computer constituting the plant monitoring device 30, or the storage unit 48 may include a remote storage device connected to the computer via a network.

[0027] The measurement data acquisition unit 32 is configured to acquire measurement data (a data set of a plurality of variables) of a plurality of variables (V1, V2,..., Vn) indicating the state of the plant at each specified period (at each time t1, t2,... of a specified interval). The measurement data acquisition unit 32 may acquire, as the above-described measurement data, a representative value (for example, an average value) of the measured values of the above-described variables in a specified period based on each time (times t1, t2,...). The measurement data acquisition unit 32 may be configured to store the measurement data acquired for each specified period in the storage unit 48.

[0028] When the plant to be monitored includes a gas turbine, the plurality of variables (V1, V2,..., Vn) indicating the state of the plant may include any one of the rotor rotation speed of the gas turbine, the temperature of each stage blade path, the average temperature of the blade path, the turbine inlet pressure, the turbine outlet pressure, the generator output, and the differential pressure of the intake filter. When the plant to be monitored is a combined cycle plant including a gas turbine and a steam turbine, the plurality of variables (V1, V2,..., Vn) indicating the state of the plant may include the pressure of the exhaust duct of the heat recovery boiler 18.

[0029] The reference data acquisition unit 34 acquires a reference data set, which is a set of reference data (data sets of a plurality of variables) related to the above-described plurality of variables. The reference data set is a set of data indicating a reference state of the plant, which is compared with the measurement data of the evaluation target (diagnosis target) in the abnormality diagnosis, and is constituted by, for example, measurement data acquired in the past. The reference data acquisition unit 34 may be configured to acquire the reference data set stored in the storage unit 48.

[0030] The deviation index value calculation unit 36 is configured to calculate a deviation index value indicating the deviation between the reference data set acquired by the reference data acquisition unit 34 and the measurement data acquired by the measurement data acquisition unit 32. The deviation index value indicates the degree of deviation from the reference state of the plant for the measurement data of the evaluation target, and the abnormality diagnosis and monitoring of the plant are performed based on the deviation index value.

[0031] In the case of abnormality diagnosis using the MT method (Mahalanobis-Taguchi method), the deviation index value calculation unit 36 calculates the Mahalanobis distance (MD value), which is the distance from the center of the unit space of the measurement data of the evaluation target, as the deviation index value based on the unit space constituted by the reference data set. Note that if the MD value is small, the target data is likely to be normal, and if the MD value is large, the target data is likely to be abnormal.

[0032] The comparison unit 38 is configured to compare the deviation index value calculated by the deviation index value calculation unit 36 with a diagnosis threshold for performing abnormality diagnosis of the plant. The diagnosis threshold may be set in advance. Further, the diagnosis threshold may be stored in the storage unit 48, and the comparison unit 38 may be configured to read the diagnosis threshold from the storage unit 48.

[0033] Note that the abnormality diagnosis of the plant by the plant monitoring device 30 may be performed based on the result of the comparison unit 38. For example, when the deviation index value calculated from the measurement data of the evaluation target is greater than the diagnosis threshold, the plant monitoring device 30 can determine that an abnormality has occurred in the plant or there is a possibility thereof.

[0034] The reference data update unit 40 is configured to update the reference data set according to the operation mode (monitoring mode or learning mode) of the plant monitoring device described later. Updating the reference data set means incorporating the measurement data of a plurality of newly acquired variables into the reference data set (that is, treating them as reference data constituting the reference data set). The reference data update unit 40 may be configured to determine whether to incorporate each of the measurement data acquired at each time (t1, t2, …) into the reference data set.

[0035] When the reference data update unit 40 incorporates the measurement data of a plurality of variables into the reference data set one by one, it may be configured to remove one of the reference data (for example, past measurement data) constituting the reference data set from the reference data set.

[0036] The operation mode switching unit 42 is configured to switch the operation mode of the plant monitoring device 30 between the monitoring mode and the learning mode. Here, the monitoring mode is an operation mode for monitoring the plant based on the comparison result by the above-described comparison unit 38. The learning mode is an operation mode in which the reference data update unit 40 incorporates at least the measurement data when the above-described deviation index value is greater than the diagnostic threshold into the reference data set.

[0037] Note that in the monitoring mode, the reference data update unit 40 may be configured to incorporate the measurement data when the deviation index value is less than or equal to the diagnostic threshold into the reference data set.

[0038] During the operation of the plant monitoring device 30 in the monitoring mode, when the deviation index value calculated by the deviation index value calculation unit 36 is greater than the diagnostic threshold and a predetermined learning mode transition condition is satisfied, the operation mode switching unit 42 is configured to switch the operation mode of the plant monitoring device 30 from the monitoring mode to the learning mode.

[0039] The plant monitoring device 30 may include an alarm output unit 44 configured to output an alarm when the deviation index value is greater than the diagnostic threshold as a result of comparing the deviation index value with the diagnostic threshold in the comparison unit 38.

[0040] Further, the deviation index value calculated by the deviation index value calculation unit 36, the comparison result of the deviation index value and the diagnostic threshold by the comparison unit 38 and / or the abnormality diagnosis result based on the comparison result, or the alarm output by the alarm output unit 44 may be displayed on the display unit 52 (such as a display).

[0041] Generally, when an abnormality occurs in a plant, the deviation index value indicating the deviation between the reference data set and the measurement data becomes larger than when the plant is normal. Therefore, it is possible to diagnose the abnormality of the plant based on the comparison between the deviation index value and the threshold value.

[0042] On the other hand, even when there is actually no abnormality in the plant, the measured value by the sensor may show a value different from before due to a change in the operating state of the plant or the like. In such a case as well, it is expected that the deviation index value calculated from the measurement data will become larger.

[0043] If the operation of the plant is continued when the above-mentioned deviation index value becomes large in this way, even if another abnormality occurs in the plant later, the magnitude of the deviation index value hardly changes, and it may become difficult to detect the abnormality of the plant.

[0044] This will be described with reference to FIG. 7. Here, FIGS. 7 and 8 are diagrams schematically showing a unit space (reference data set) created based on a plurality of variables indicating the state of a plant, and an MD value (deviation index value) calculated from the unit space and measurement data. In FIG. 7, the region R1 indicated by the dashed line is a set of points where the MD value calculated based on the unit space is equal to the diagnostic threshold value, and the measurement data (e.g., D1) within the range of the region R1 is data evaluated as normal. In FIGS. 7 and 8, for simplicity, a unit space based on two variables (variables measured by sensor A and sensor B) is schematically shown.

[0045] Assuming that the measurement data when the measured value of sensor A deviates from the reference data set is D2, the MD value (the length of arrow MD2) calculated for the measurement data D2 is outside the range of the region R1 and is greater than the diagnostic threshold value (see FIG. 7). However, if the cause of the deviation of the measured value of sensor A from the reference data set is not an abnormality in the plant, it should not be determined that an abnormality has occurred in the plant just because MD2 is greater than the diagnostic threshold value.

[0046] Here, even if it is found as a result of the investigation that no abnormality has occurred in the plant, if the measured value of another sensor (here, sensor B) deviates from the reference data set later, the MD value (the length of arrow MD3) calculated for the measurement data D3 at this time does not change significantly from MD2 (see FIG. 7). For this reason, for example, if the diagnostic threshold value is changed according to MD2, even if an abnormality occurs in the measured value of sensor B, there is a possibility that the abnormality in the plant cannot be detected.

[0047] In this regard, in the plant monitoring device 30 having the above-described configuration, even when the deviation index value is greater than the diagnostic threshold value during operation in the monitoring mode, if a predetermined learning mode transition condition is satisfied, the operation mode of the plant monitoring device 30 is switched to the learning mode, and measurement data when the deviation index value is greater than the diagnostic threshold value is incorporated into the reference data set. Here, the case where the predetermined learning mode transition condition is satisfied means, for example, when an abnormality is determined based on the deviation index value but it can be determined that the deviation index value has become large due to factors other than an abnormality in the plant equipment, or when it can be determined that operation can continue even if there is an abnormality in the plant equipment. Thus, by incorporating measurement data when the deviation index value is greater than the diagnostic threshold value into the reference data set during operation in the learning mode, the reference data set is redefined.

[0048] In the example described with reference to FIG. 7, by operating the plant monitoring device 30 in the learning mode, the measurement data D2 can be incorporated into the reference data set. FIG. 8 shows a unit space composed of the reference data set redefined in this way. That is, in FIG. 8, the region R2 indicated by the broken line is a set of points where the MD value calculated based on the unit space composed of the redefined reference data set (the reference data set including the measurement data D2) is equal to the diagnostic threshold value. Note that the unit space based on the redefined reference data set (see FIG. 8) has a larger standard deviation (variation) than the unit space before redefinition (see FIG. 7). By calculating the MD value using the unit space based on the redefined reference data set (see FIG. 8), actually, the MD value of the measurement data D2 when there is no abnormality in the plant is below the diagnostic threshold value (within the range of the region R2), and the MD value of the measurement data D3 when an abnormality occurs in the plant (or the sensor value) becomes greater than the diagnostic threshold value (outside the range of the region R2).

[0049] In this way, by performing abnormality diagnosis of the plant based on the redefined reference data set, an abnormality of the plant can be appropriately detected. Therefore, according to the plant monitoring device 30 having the above-described configuration, robust abnormality diagnosis is possible.

[0050] (Plant Monitoring Flow) Hereinafter, the plant monitoring method according to several embodiments will be described in more detail. In the following, the case of monitoring the above-described plant 1 using the above-described plant monitoring device 30 will be described. However, in some embodiments, the plant monitoring method may be executed using other devices, or a part of the procedures described below may be performed manually. Further, in the following, the plant monitoring method using the MT method will be described. However, the same description can be applied to the case of monitoring a plant using other statistical methods (statistical process control (SPC), or multi-variate statistical process control (MSPC), etc.). ) The same description can be applied to the case of monitoring a plant using the above methods.

[0051] FIG. 3 is a flowchart of a plant monitoring method according to several embodiments. FIGS. 4 and 5 are diagrams for explaining the plant monitoring method according to several embodiments, and are graphs showing the time change of the calculated deviation index value (specifically, the MD value; vertical axis).

[0052] Among the steps of the flowchart shown in FIG. 3, steps S2 to S11 are procedures when the plant monitoring device 30 is operating in the monitoring mode, and steps S12 to S14 are procedures when the plant monitoring device 30 is operating in the learning mode.

[0053] In a plant monitoring method according to an embodiment, during the operation of the plant monitoring device 30 in the monitoring mode, first, the measurement data acquisition unit 32 acquires measurement data (data sets of a plurality of variables) of a plurality of variables (V1, V2,..., Vn) indicating the state of the plant at regular intervals (at times t1, t2,... of a regular interval) (S2).

[0054] Also, the reference data acquisition unit 34 acquires a reference data set, which is a set of reference data (data sets of a plurality of variables) related to the above-described plurality of variables, from, for example, the storage unit 48 (S4).

[0055] Next, the deviation index value calculation unit 36 calculates a deviation index value indicating the deviation between the reference data set acquired in step S4 and the measurement data acquired in step S2 (S6). In the present embodiment, in step S6, as the above-described deviation index value, a Mahalanobis distance (MD value), which is the distance from the center of the unit space of the measurement data to be evaluated based on the unit space constituted by the reference data set, is calculated.

[0056] Next, the comparison unit 38 compares the MD value (deviation index value) calculated in step S6 with a diagnostic threshold Th_A (see FIGS. 4 and 5) for diagnosing an abnormality of the plant 1 (S8). Note that the plant monitoring device 30 may perform an abnormality diagnosis or monitoring of the plant 1 based on the comparison result between the MD value (deviation index value) and the diagnostic threshold Th_A in step S8.

[0057] In step S8, when the MD value (deviation index value) is less than or equal to the diagnostic threshold Th_A (No in S8), it is determined that no abnormality has occurred in the plant 1, so the process returns to step S2 and the operation in the monitoring mode is continued. In this case, the reference data update unit 40 may incorporate the measurement data (the measurement data acquired in step S2) for which the MD value (deviation index value) was calculated in step S6 into the reference data set and update the reference data set.

[0058] On the other hand, in step S8, when the MD value (deviation index value) is greater than the diagnostic threshold Th_A (Yes in S8), the operation mode switching unit 42 determines whether or not the learning mode transition condition is satisfied (S10).

[0059] When the learning mode transition condition is satisfied in step S10 (Yes in S10; at time t1 in FIG. 4 or time t11 in FIG. 5), the operation mode switching unit 42 switches the operation mode of the plant monitoring device 30 from the monitoring mode to the learning mode. That is, the process proceeds to subsequent step S12, and the reference data update unit 40 incorporates at least the measurement data when the MD value (deviation index value) is greater than the diagnostic threshold into the reference data set (S12).

[0060] On the other hand, when the learning mode transition condition is not satisfied in step S10 (No in S10), the process returns to step S2 and continues to operate in the monitoring mode. In this case, since the MD value (deviation index value) is greater than the diagnostic threshold Th_A and there may be an abnormality in the plant 1, an alarm may be output by the alarm output unit 44 (S11).

[0061] The above learning mode transition condition may include that it is determined that the MD value (deviation index value) is greater than the diagnostic threshold Th_A due to factors other than the abnormality of the plant 1.

[0062] <Example 1 of learning mode transition condition> For example, the learning mode transition condition may include that it is determined that the operation of the plant 1 is in a state where it can continue based on a preset determination condition. The preset determination condition may be stored in the storage unit 48.

[0063] Whether the above operation of the plant 1 can continue may be determined based on the list shown in FIG. 6. Here, FIG. 6 is a diagram showing an example of a list of determination conditions for determining whether the operation of the plant 1 can continue. In the list shown in FIG. 6, determination conditions (determination condition A, determination condition B,...) are set for each row. In each column of the list in FIG. 6, individual conditions (condition 1, condition 2,...) for each determination condition are set, and a monitoring mode return condition described later is set. In the list in FIG. 6, if all of the individual conditions set in the row of determination condition A are satisfied, it can be determined that the determination condition A is satisfied. Also, if any one of the determination conditions (determination condition A, determination condition B,...) set in each row is satisfied, it can be determined that the operation of the plant 1 can continue.

[0064] More specifically, in the case of the plant 1 including a gas turbine, the above determination condition may be the determination condition A (the first line of the list in FIG. 6), which includes that the main factor for the MD value (deviation index value) being larger than the diagnostic threshold Th_A is the increase in the differential pressure of the intake filter of the gas turbine (condition 1 and condition 2).

[0065] Note that condition 1 of determination condition A shown in FIG. 6 is that the SN ratio (SN ratio of the desired maximum characteristic in the MT method) of the upstream intake filter differential pressure (sensor value) provided in the gas turbine is the largest among the sensor values to be measured, and condition 2 is that the upstream intake filter differential pressure (sensor value) has increased compared to the normal time (when the MD value (deviation index value) was below the diagnostic threshold Th_A). Further, condition 3 is that the differential pressure of the downstream intake filter provided downstream of the upstream intake filter in the gas turbine has decreased compared to the normal time. Note that the fact that the SN ratio of a certain sensor value is the largest among the sensor values to be measured can be judged that the said sensor value is the main factor increasing the MD value.

[0066] In the plant 1 including a gas turbine, during rainy days, the intake filter of the gas turbine (the upstream intake filter when a plurality of intake filters are provided) gets wet with rain, and the differential pressure before and after the intake filter tends to increase. Therefore, when the above determination condition A is satisfied, it can be determined that the operation of the plant 1 can be continued because the MD value (deviation index value) is larger than the diagnostic threshold Th_A due to rainy days.

[0067] Alternatively, when the plant 1 is a combined cycle plant including a gas turbine and a steam turbine, the above determination condition may be the determination condition B (the second line of the list in FIG. 6), which includes that the main factor for the MD value (deviation index value) being larger than the diagnostic threshold Th_A is the decrease in the pressure of the exhaust duct constituting the heat recovery boiler 18 (condition 1 and condition 2).

[0068] Note that Condition 1 of determination condition B shown in FIG. 6 is that the SN ratio (SN ratio of the desired maximum characteristic in the MT method) of the pressure (sensor value) of the exhaust duct constituting the exhaust heat recovery boiler 18 is the largest among the sensor values of the measurement target, and Condition 2 is that the pressure of the exhaust duct has decreased compared to the normal time (when the MD value (deviation index value) is equal to or less than the diagnostic threshold Th_A).

[0069] In a combined cycle power plant including a gas turbine and a steam turbine, when switching from a combined cycle operation in which the exhaust gas of the gas turbine is fed to a boiler to a simple cycle operation in which the exhaust gas of the gas turbine is discharged to the outside without being fed to the boiler, the pressure of the exhaust duct constituting the boiler decreases. Therefore, when the above-described determination condition B is satisfied, it can be determined that the operation of the plant 1 can be continued because the MD value (deviation index value) has become larger than the diagnostic threshold Th_A due to the switching of the operation mode of the combined cycle power plant.

[0070] <Example 2 of learning mode transition condition> Alternatively, the learning mode transition condition may include that the number of times of alarm output by the alarm output unit 44 (i.e., the number of executions of step S11) exceeds a specified value.

[0071] The fact that alarms are repeatedly output because the MD value (deviation index value) is larger than the diagnostic threshold Th_A may indicate that the operator or the like has determined that the operation of the plant 1 can be continued. Therefore, when the number of times of alarm output by the alarm output unit 44 exceeds a specified value, it can be determined that the operation of the plant can be continued even if the MD value (deviation index value) is larger than the diagnostic threshold Th_A.

[0072] <Example 3 of learning mode transition condition> Alternatively, the learning mode transition condition may include that the plant monitoring device 30 has received a command to switch the operation mode of the plant monitoring device 30 from the monitoring mode to the learning mode. Note that the command may be input from, for example, an input device 46 or the like by an operator or the like.

[0073] Even when the MD value (deviation index value) is greater than the diagnostic threshold Th_A, there may be a case where it is determined by an operator or the like that the operation of the plant 1 can be continued. Therefore, when the plant monitoring device 30 receives a command to switch the operation mode to the learning mode, the operation mode of the plant monitoring device 30 may be switched to the learning mode.

[0074] As already described, in step S12, among the data acquired during the operation in the learning mode, at least the measurement data when the MD value (deviation index value) calculated in step S6 is greater than the diagnostic threshold Th_A is incorporated into the reference data set.

[0075] In step S12, among the measurement data acquired during the operation in the learning mode, the measurement data when the MD value (deviation index value) calculated in step S6 is greater than the learning threshold Th_B (see FIGS. 4 and 5) may be incorporated into the reference data set. Here, the learning threshold Th_B may be a value smaller than the diagnostic threshold Th_A. The learning threshold Th_B may be, for example, about 1 / 2 of the diagnostic threshold Th_A.

[0076] In this way, by setting the learning threshold Th_B to a value smaller than the diagnostic threshold Th_A, even when the MD value (deviation index value) fluctuates near the diagnostic threshold Th_A during the operation in the learning mode, the measurement data when the MD value is smaller than the diagnostic threshold Th_A can be incorporated into and held in the reference data set. Thereby, false warnings can be suppressed and the plant can be operated stably.

[0077] In step S12, the frequency of incorporating measurement data into the reference data set during operation in the learning mode may be higher than the frequency of incorporating measurement data into the reference data set during operation in the monitoring mode (when the answer is No in step S8).

[0078] For example, in step S12, all of the measurement data obtained during operation in the learning mode when the above-mentioned MD value (deviation index value) is greater than the learning threshold Th_B may be incorporated into the reference data set. On the other hand, during operation in the monitoring mode, for example, measurement data may be incorporated into the reference data set at a ratio of one out of several tens to several hundreds of measurement data points.

[0079] In this way, in the learning mode, by incorporating measurement data into the reference data at a higher frequency than in the monitoring mode, it is possible to more reliably reflect a very small number of abnormal data (measurement data when the MD value (deviation index value) is greater than the diagnostic threshold Th_A) in the reference data set while calculating the MD value (deviation index value) with respect to the total number of measurement data. Set During operation of the plant monitoring device in the learning mode, the operation mode switching unit 42 determines whether the conditions for returning from the learning mode to the monitoring mode (monitoring mode return conditions) are satisfied (S14). If the monitoring mode return conditions are not satisfied in step S14 (No in S14), the process returns to step S12 and the operation in the learning mode continues. On the other hand, if the monitoring mode return conditions are satisfied in step S14 (Yes in S14), the operation mode switching unit 42 switches the operation mode of the plant monitoring device 30 from the learning mode to the monitoring mode. That is, the process returns to step S2 and shifts to the operation in the monitoring mode.

[0080] During operation of the plant monitoring device in the learning mode, the operation mode switching unit 42 determines whether the conditions for returning from the learning mode to the monitoring mode (monitoring mode return conditions) are satisfied (S14). If the monitoring mode return conditions are not satisfied in step S14 (No in S14), the process returns to step S12 and the operation in the learning mode continues. On the other hand, if the monitoring mode return conditions are satisfied in step S14 (Yes in S14), the operation mode switching unit 42 switches the operation mode of the plant monitoring device 30 from the learning mode to the monitoring mode. That is, the process returns to step S2 and shifts to the operation in the monitoring mode. to return During operation of the plant monitoring device in the learning mode, the operation mode switching unit 42 determines whether the conditions for returning from the learning mode to the monitoring mode (monitoring mode return conditions) are satisfied (S14). If the monitoring mode return conditions are not satisfied in step S14 (No in S14), the process returns to step S12 and the operation in the learning mode continues. On the other hand, if the monitoring mode return conditions are satisfied in step S14 (Yes in S14), the operation mode switching unit 42 switches the operation mode of the plant monitoring device 30 from the learning mode to the monitoring mode. That is, the process returns to step S2 and shifts to the operation in the monitoring mode.

[0081] The above monitoring mode return condition may include that a specified time has elapsed since the operation mode of the plant monitoring device 30 was switched from the monitoring mode to the learning mode. For example, in the example shown in FIG. 4, at time t2 when a specified time T1 has elapsed since time t1 when the operation mode of the plant monitoring device 30 was switched from the monitoring mode to the learning mode, it is determined that the above monitoring mode return condition is satisfied, and the operation mode of the plant monitoring device 30 is switched from the learning mode to the monitoring mode. Further, as the monitoring mode return condition corresponding to determination condition B (learning mode transition condition) in FIG. 6, rule One hour is set as the specified time. Note that the length of the above specified time may be determined according to the content of the learning mode transition condition, and may be, for example, a length between several minutes and several hours.

[0082] When measurement data when the MD value (deviation index value) is larger than the diagnostic threshold Th_A in the learning mode is taken into the reference data set, the newly calculated deviation index value gradually becomes smaller as time passes. In this regard, in the above-described embodiment, when a specified time (T1) has elapsed since the time (time t1) when the operation mode of the plant monitoring device 30 was switched to the learning mode, the operation mode is switched from the learning mode to the monitoring mode. Therefore, by setting the length of the specified time so that the MD value (deviation index value) calculated at the time of elapse of the aforementioned specified time (T1) sufficiently falls below the diagnostic threshold Th_A, it is possible to appropriately detect an abnormality of the plant 1 after returning to the monitoring mode. Therefore, robust abnormality diagnosis becomes possible.

[0083] Alternatively, the above monitoring mode return condition may include that the calculated MD value (deviation index value) has become equal to or less than the diagnostic threshold Th_A. For example, in the example shown in FIG. 5, at time t12 after time t11 when the operation mode of the plant monitoring device 30 was switched from the monitoring mode to the learning mode, it is determined that the above monitoring mode return condition is satisfied, and the operation mode of the plant monitoring device 30 is switched from the learning mode to the monitoring mode. Further, as the monitoring mode return condition corresponding to determination condition A (learning mode transition condition) in FIG. 6, MD valueA condition that it is below the diagnostic threshold is set.

[0084] In the learning mode, if measurement data when the MD value (deviation index value) is greater than the diagnostic threshold Th_A is incorporated into the reference data set, the newly calculated deviation index value gradually decreases over time. In this regard, in the above-described embodiment then During the operation of the plant monitoring device 30 in the learning mode, when the MD value (deviation index value) becomes less than or equal to the diagnostic threshold Th_A, the operation mode is switched from the learning mode to the monitoring mode. Therefore, after returning to the monitoring mode, an abnormality of the plant 1 can be appropriately detected. Thus, robust abnormality diagnosis becomes possible.

[0085] In some embodiments, during the operation in the monitoring mode or the learning mode, when the reference data update unit 40 incorporates one measurement data of a plurality of variables into the reference data set (in the case of No in step S8 or step S12), one of the reference data (for example, past measurement data) constituting the reference data set may be removed from the reference data set. Here, among the reference data included in the reference data set, the measurement data incorporated during the operation in the learning mode (step S12) may be preferentially removed rather than the measurement data incorporated during the operation in the monitoring mode (in the case of No in step S8).

[0086] As described above, when one new measurement data is incorporated into the reference data set, one measurement data is removed from the reference data set. Therefore, it is possible to maintain the calculation load for calculating the MD value (deviation index value) without increasing it, and the reference data set can be made to correspond to the latest state of the plant.

[0087] In addition, changes in the operating state of the plant that can cause an increase in the MD value (deviation index value) may return to their original state after some time has passed. For example, an event where the differential pressure of the intake filter of a gas turbine increases due to rain is considered to return to the original differential pressure when the rain stops. Therefore, as described above, among the reference data included in the reference data set, by preferentially removing the measurement data captured during operation in the learning mode rather than the measurement data captured during operation in the monitoring mode, it becomes easier to redefine the reference data set that reflects the current operating state of the plant. For this reason, the abnormality diagnosis of Plant 1 can be made more appropriate.

[0088] The content described in each of the above embodiments is understood as follows, for example.

[0089] (1) The plant monitoring device (30) according to at least one embodiment of the present invention is a plant monitoring device for monitoring a plant (1), a measurement data acquisition unit (32) configured to acquire measurement data of a plurality of variables indicating the state of the plant at regular intervals, a comparison unit (38) configured to compare a deviation index value indicating a deviation between the reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant, a reference data update unit (40) configured to update the reference data set, and an operation mode switching unit (42) configured to switch the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result by the comparison unit and a learning mode in which the measurement data, at least when the deviation index value is greater than the diagnostic threshold value, is incorporated into the reference data set by the reference data update unit, wherein the operation mode switching unit is configured to switch the operation mode from the monitoring mode to the learning mode when the deviation index value is greater than the diagnostic threshold value and the learning mode transition condition is satisfied during operation of the plant monitoring device in the monitoring mode.

[0090] In the configuration of (1) above, during the operation in the monitoring mode of the plant monitoring device, even when the deviation index value is greater than the diagnostic threshold value, if a predetermined learning mode transition condition is satisfied, the operation mode of the monitoring device is switched to the learning mode, and the measurement data when the deviation index value is greater than the diagnostic threshold value is incorporated into the reference data set. Here, the case where the predetermined learning mode transition condition is satisfied means, for example, when an abnormality is determined based on the deviation index value but it can be determined that the deviation index value has become large due to factors other than the abnormality of the plant equipment, or when it can be determined that the operation can continue even if there is an abnormality in the plant equipment. Thus, by incorporating the measurement data when the deviation index value is greater than the diagnostic threshold value into the reference data set during the operation in the learning mode, the reference data set is redefined. Therefore, by performing an abnormality diagnosis of the plant based on the redefined reference data set, the abnormality of the plant can be appropriately detected. Therefore, according to the configuration of (1) above, robust abnormality diagnosis becomes possible.

[0091] (2) In some embodiments, in the configuration of (1) above, The learning mode transition condition includes that it is determined that the deviation index value is greater than the diagnostic threshold value due to factors other than the abnormality of the plant, or that it is determined that the operation of the plant can continue.

[0092] In the configuration of (2) above, during the operation in the monitoring mode of the plant monitoring device, when it is determined that the deviation index value is greater than the diagnostic threshold value due to factors other than the abnormality of the plant (for example, changes in the operating state of the plant, etc.), or when it is determined that the operation of the plant can continue even if there is an abnormality in the plant, the operation mode of the plant monitoring device is switched to the learning mode. Therefore, by performing an abnormality diagnosis of the plant based on the reference data set redefined in the learning mode, the abnormality of the plant can be appropriately detected. Therefore, robust abnormality diagnosis becomes possible.

[0093] (3) In some embodiments, in the configuration of (1) or (2) above, The learning mode transition condition includes that it is determined that the plant can continue to operate based on a preset determination condition.

[0094] According to the configuration of (3) above, during the operation of the plant monitoring device in the monitoring mode, when the deviation index value is greater than the diagnostic threshold, but it is determined that the plant can continue to operate based on a preset determination condition, the operation mode of the plant monitoring device is switched to the learning mode. Therefore, by performing abnormal diagnosis of the plant based on the reference data set redefined in the learning mode, abnormalities of the plant can be appropriately detected. Thus, robust abnormal diagnosis becomes possible.

[0095] (4) In some embodiments, in the configuration of (3) above, The plant includes a gas turbine (for example, the gas turbine facility 2 described above), The determination condition includes that the main cause of the deviation index value being greater than the diagnostic threshold is an increase in the differential pressure of the intake filter of the gas turbine.

[0096] In a plant including a gas turbine, during rainy weather, the intake filter of the gas turbine gets wet with rain, and the differential pressure across the intake filter tends to increase. According to the configuration of (4) above, when the determination condition including that the main cause of the deviation index value being greater than the diagnostic threshold is an increase in the differential pressure of the intake filter of the gas turbine is satisfied, since it is rainy weather and the deviation index value is greater than the diagnostic threshold, it is determined that the plant can continue to operate, and the operation mode of the plant monitoring device is switched to the learning mode. Therefore, by performing abnormal diagnosis of the plant based on the reference data set redefined in the learning mode, abnormalities of the plant can be appropriately detected. Thus, robust abnormal diagnosis becomes possible.

[0097] (5) In some embodiments, in the configuration of (3) above, The plant is a combined cycle plant including a gas turbine (for example, the gas turbine facility 2 described above) and a steam turbine (for example, the steam turbine facility 12 described above), The determination condition includes that the main factor for the deviation index value being greater than the diagnostic threshold is a decrease in the pressure of an exhaust duct configured to form a boiler (for example, the exhaust heat recovery boiler 18 described above) for generating steam supplied to the steam turbine, and through which exhaust gas from the gas turbine is supplied.

[0098] In a combined cycle plant including a gas turbine and a steam turbine, when switching from a combined cycle operation in which exhaust gas from the gas turbine is fed to a boiler to a simple cycle operation in which the exhaust gas from the gas turbine is discharged to the outside without being fed to the boiler, the pressure of the exhaust duct constituting the boiler decreases. According to the configuration of (5) above, when the determination condition including that the main factor for the deviation index value being greater than the diagnostic threshold is the decrease in the pressure of the above exhaust duct is satisfied, it is determined that the plant can continue to operate because the deviation index value becomes greater than the diagnostic threshold due to the switching of the operation mode of the combined cycle plant, and the operation mode of the plant monitoring device is switched to the learning mode. Therefore, by performing an abnormality diagnosis of the plant based on the reference data set redefined in the learning mode, the abnormality of the plant can be appropriately detected. Thus, a robust abnormality diagnosis becomes possible.

[0099] (6) In some embodiments, in the configuration of (1) or (2) above, the plant monitoring device includes an alarm output unit (44) configured to output an alarm when the deviation index value is greater than the diagnostic threshold during operation in the monitoring mode of the plant monitoring device, and the learning mode transition condition includes that the number of times of alarm output by the alarm output unit exceeds a specified value.

[0100] The fact that an alarm is repeatedly output because the deviation index value is greater than the diagnostic threshold may indicate that the operator or the like has determined that the plant can continue to operate. According to the configuration of (6) above, during the operation of the plant monitoring device in the monitoring mode, when the deviation index value is greater than the diagnostic threshold and the number of times the alarm is output by the alarm output unit exceeds the specified value, it is determined that the plant can continue to operate, and the operation mode of the plant monitoring device is switched to the learning mode. Therefore, by performing an abnormality diagnosis of the plant based on the reference data set redefined in the learning mode, the abnormality of the plant can be appropriately detected. Thus, a robust abnormality diagnosis becomes possible.

[0101] (7) In some embodiments, in the configuration of (1) above, The learning mode transition condition includes that the plant monitoring device has received a command to switch the operation mode to the learning mode.

[0102] Even when the deviation index value is greater than the diagnostic threshold, there may be a case where the operator or the like determines that the plant can continue to operate. According to the configuration of (7) above, when the plant monitoring device receives a command to switch the operation mode input by the operator or the like to the learning mode, the operation mode of the plant monitoring device is switched to the learning mode. Therefore, by performing an abnormality diagnosis of the plant based on the reference data set redefined in the learning mode, the abnormality of the plant can be appropriately detected. Thus, a robust abnormality diagnosis becomes possible.

[0103] (8) In some embodiments, in any of the configurations of (1) to (7) above, During the operation of the plant monitoring device in the learning mode, the operation mode switching unit is configured to switch the operation mode from the learning mode to the monitoring mode when a specified time has elapsed since the operation mode was switched from the monitoring mode to the learning mode.

[0104] When measurement data when the deviation index value is greater than the diagnostic threshold in the learning mode is incorporated into the reference data set, the newly calculated deviation index value gradually decreases over time. In this regard, according to the configuration of the above (8), when a specified time has elapsed since the operation mode of the plant monitoring device was switched to the learning mode, the operation mode is switched from the learning mode to the monitoring mode. Therefore, by setting the length of the specified time so that the deviation index value calculated when the specified time has elapsed sufficiently falls below the diagnostic threshold, it is possible to appropriately detect an abnormality in the plant after returning to the monitoring mode. Thus, robust abnormality diagnosis becomes possible.

[0105] (9) In some embodiments, in any of the configurations of the above (1) to (7), The operation mode switching unit is configured to switch the operation mode from the learning mode to the monitoring mode when the deviation index value becomes equal to or less than the diagnostic threshold during operation of the plant monitoring device in the learning mode.

[0106] When measurement data when the deviation index value is greater than the diagnostic threshold in the learning mode is incorporated into the reference data set, the newly calculated deviation index value gradually decreases over time. In this regard, according to the configuration of the above (9), during operation of the plant monitoring device in the learning mode, when the deviation index value becomes equal to or less than the diagnostic threshold, the operation mode is switched from the learning mode to the monitoring mode. Therefore, after returning to the monitoring mode, it is possible to appropriately detect an abnormality in the plant. Thus, robust abnormality diagnosis becomes possible.

[0107] (10) In some embodiments, in any of the configurations of the above (1) to (9), The reference data update unit is configured to incorporate the measurement data when the deviation index value is equal to or less than the diagnostic threshold during operation operation in the monitoring mode into the reference data set.

[0108] According to the configuration of the above (10), during operation in the monitoring mode operationAmong them, when the deviation index value is less than or equal to the diagnostic threshold, the measurement data is incorporated into the reference data set, and the learning mode by of operation Among them, when the deviation index value is greater than the diagnostic threshold, the measurement data is incorporated into the reference data set. That is, the measurement data when the plant is normal mainly in the monitoring mode is incorporated into the reference data set, and the measurement data when the deviation index value is greater than the diagnostic threshold in the learning mode, so that the reference data is redefined. Therefore, by performing abnormal diagnosis of the plant based on the reference data set redefined in this way, the abnormality of the plant can be appropriately detected. Thus, robust abnormal diagnosis becomes possible.

[0109] (11) In some embodiments, in any of the configurations (1) to (10) above, the reference data update unit is configured to incorporate the measurement data into the reference data set at a higher frequency than the monitoring mode in the learning mode.

[0110] According to the configuration of (11) above, in the learning mode, the measurement data is incorporated into the reference data set at a higher frequency than the monitoring mode. Therefore, a very small number of abnormal data (measurement data when the deviation index value is greater than the diagnostic threshold) with respect to the total number of measurement data can be more reliably reflected in the reference data set while calculating the deviation index value. Thus, by performing abnormal diagnosis of the plant based on the deviation index value calculated in this way, robust abnormal diagnosis of the plant becomes possible.

[0111] (12) In some embodiments, in any of the configurations (1) to (11) above, When the reference data update unit incorporates one piece of the measurement data into the reference data set, it is configured to remove one piece of the reference data included in the reference data set from the reference data set, and among the reference data included in the reference data set, the measurement data acquired during operation in the learning mode is preferentially removed rather than the measurement data acquired during operation in the monitoring mode.

[0112] According to the configuration of (12) above, when one piece of new measurement data is incorporated into the reference data set, one piece of measurement data is removed from the reference data set, so that the calculation load for calculating the deviation index value can be maintained without increasing, and the reference data set can be made to correspond to the state of the latest plant. Also, among the reference data included in the reference data set, the measurement data acquired during operation in the learning mode is preferentially removed rather than the measurement data acquired during operation in the monitoring mode, so that it becomes easier to redefine the reference data set reflecting the current operating state of the plant. For this reason, the abnormality diagnosis of the plant can be made more appropriate.

[0113] (13) In some embodiments, in any of the configurations of (1) to (12) above, The deviation index value is the Mahalanobis distance for the measurement data calculated based on the unit space constituted by the reference data set.

[0114] According to the configuration of (13) above, the abnormality diagnosis of the plant can be made appropriate based on the Mahalanobis distance indicating the divergence between the unit space constituted by the reference data set and the measurement data.

[0115] (14) The plant monitoring method according to at least one embodiment of the present invention is A plant monitoring method using a plant monitoring device (30) for monitoring a plant, A step (S2) of acquiring measurement data of a plurality of variables indicating the state of the plant at regular intervals, A comparison step (S8) of comparing a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant; A step (S12) of updating the reference data set; An operation mode switching step (S10) of switching an operation mode of the plant monitoring device between a monitoring mode of monitoring the plant based on a comparison result in the comparison step and a learning mode of incorporating the measurement data into the reference data set at least when the deviation index value is greater than the diagnostic threshold value; In the operation mode switching step, during operation in the monitoring mode of the plant monitoring device, when the deviation index value is greater than the diagnostic threshold value and a learning mode transition condition is satisfied, the operation mode is switched from the monitoring mode to the learning mode.

[0116] In the method of (14) above, even when the deviation index value is greater than the diagnostic threshold value during operation in the monitoring mode of the plant monitoring device, if a predetermined learning mode transition condition is satisfied, the operation mode of the monitoring device is switched to the learning mode, and the measurement data when the deviation index value is greater than the diagnostic threshold value is incorporated into the reference data set. Here, the case where a predetermined learning mode transition condition is satisfied means, for example, when an abnormality is determined based on the deviation index value, but it can be determined that the deviation index value has become large due to factors other than an abnormality of the plant equipment, or when it can be determined that operation can continue even if there is an abnormality in the plant equipment. Thus, by incorporating the measurement data when the deviation index value is greater than the diagnostic threshold value into the reference data set during operation in the learning mode, the reference data set is redefined. Therefore, by diagnosing an abnormality of the plant based on the redefined reference data set, an abnormality of the plant can be appropriately detected. Therefore, according to the method of (14) above, robust abnormality diagnosis becomes possible.

[0117] (15) A plant monitoring program according to at least one embodiment of the present invention is A plant monitoring program for operating a plant monitoring device (30) for monitoring a plant, causing a computer to perform a procedure of acquiring measurement data of a plurality of variables indicating the state of the plant for each specified period, perform a comparison procedure of comparing a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant, perform a procedure of updating the reference data set, and perform a procedure of switching the operation mode of the plant monitoring device between a monitoring mode of monitoring the plant based on the comparison result in the comparison procedure and a learning mode of incorporating the measurement data into the reference data set at least when the deviation index value is greater than the diagnostic threshold value, wherein in the procedure of switching the operation mode, when the deviation index value is greater than the diagnostic threshold value and a learning mode transition condition is satisfied during operation of the plant monitoring device in the monitoring mode, the operation mode is switched from the monitoring mode to the learning mode.

[0118] In the program of (15) above, during the operation in the monitoring mode of the plant monitoring device, even when the deviation index value is greater than the diagnostic threshold value, if a predetermined learning mode transition condition is satisfied, the operation mode of the monitoring device is switched to the learning mode, and the measurement data when the deviation index value is greater than the diagnostic threshold value is incorporated into the reference data set. Here, the case where the predetermined learning mode transition condition is satisfied means, for example, when an abnormality is determined based on the deviation index value but it can be determined that the deviation index value has increased due to factors other than the abnormality of the plant equipment, or when it can be determined that the operation can continue even if there is an abnormality in the plant equipment. In this way, by incorporating the measurement data when the deviation index value is greater than the diagnostic threshold value during the operation in the learning mode into the reference data set, the reference data set corresponding to the operating state of the plant is redefined. Therefore, by performing an abnormality diagnosis of the plant based on the redefined reference data set, the abnormality of the plant can be appropriately detected. Thus, according to the program of (15) above, a robust abnormality diagnosis becomes possible.

[0119] As described above, the embodiments of the present invention have been explained, but the present invention is not limited to the above-described embodiments, and also includes forms obtained by modifying the above-described embodiments and forms obtained by appropriately combining these forms.

[0120] In this specification, expressions indicating relative or absolute arrangements such as "in a certain direction", "along a certain direction", "parallel", "orthogonal", "center", "concentric", or "coaxial" not only strictly represent such arrangements, but also represent states of relative displacement with tolerances, or angles and distances to the extent that the same function can be obtained. For example, expressions indicating that things such as "identical", "equal", and "homogeneous" are in an equal state not only strictly represent an equal state, but also represent states in which there are tolerances or differences to the extent that the same function can be obtained. Also, in this specification, expressions indicating shapes such as a rectangular shape or a cylindrical shape not only represent shapes such as a rectangular shape or a cylindrical shape in a geometrically strict sense, but also represent shapes including concavo-convex portions, chamfered portions, etc. within a range where the same effect can be obtained. Also, in this specification, the expressions "comprises", "includes", or "has" for a component are not exclusive expressions that exclude the existence of other components.

Explanation of Signs

[0121] 1 Plant 2 Gas Turbine Facility 4 Compressor 6 Combustor 8 Turbine 10 Generator 12 Steam Turbine Facility 14 Turbine 16 Generator 18 Exhaust Heat Recovery Boiler 20 Condenser 30 Plant Monitoring Device 32 Measurement Data Acquisition Unit 34 Reference Data Acquisition Unit 36 Deviation Index Value Calculation Unit 38 Comparison Unit 40 Reference Data Update Unit 42 Operation Mode Switching Unit 44 Alarm Output Unit 46 Input Device 48 Storage Unit 50 Measurement Unit 52 Display Unit

Claims

1. A plant monitoring device for monitoring a plant, comprising: a measurement data acquisition unit configured to acquire measurement data of a plurality of variables indicating the state of the plant at regular intervals; a comparison unit configured to compare a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant; a reference data update unit configured to update the reference data set; an operation mode switching unit configured to switch the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on a comparison result by the comparison unit, and a learning mode in which, when at least the deviation index value is greater than the diagnostic threshold value, the measurement data at that time is incorporated into the reference data set by the reference data update unit; the operation mode switching unit is configured to switch the operation mode from the monitoring mode to the learning mode when the deviation index value is greater than the diagnostic threshold value and a learning mode transition condition is satisfied during the operation of the plant monitoring device in the monitoring mode; the reference data update unit: during operation in the monitoring mode, is configured to incorporate into the reference data set the measurement data among the newly acquired measurement data during operation in the monitoring mode when the deviation index value is less than or equal to the diagnostic threshold value; and during operation in the learning mode, is configured to incorporate into the reference data set at least the measurement data among the newly acquired measurement data during operation in the learning mode when the deviation index value is greater than the diagnostic threshold value Plant monitoring device.

2. The learning mode transition condition includes that it is determined that the deviation index value is greater than the diagnostic threshold value due to factors other than an abnormality of the plant, or that it is determined that the operation of the plant can continue. The plant monitoring device according to claim 1.

3. The learning mode transition condition includes that it is determined that the operation of the plant is in a state where it can continue based on a preset determination condition. The plant monitoring device according to claim 1 or 2.

4. The plant includes a gas turbine, The determination condition includes that the main cause of the deviation index value being greater than the diagnostic threshold value is an increase in the differential pressure of the intake filter of the gas turbine. The plant monitoring device according to claim 3.

5. The plant is a combined cycle plant including a gas turbine and a steam turbine, The determination condition includes that the main factor for the deviation index value being larger than the diagnostic threshold value is the pressure drop in the exhaust duct configured to generate steam supplied to the steam turbine and to which the exhaust gas from the gas turbine is supplied. The plant monitoring device according to claim 3.

6. During operation in the monitoring mode of the plant monitoring device, it includes an alarm output unit configured to output an alarm when the deviation index value is larger than the diagnostic threshold value, The learning mode transition condition includes that the number of times of alarm output by the alarm output unit exceeds a specified value. The plant monitoring device according to claim 1 or 2.

7. The learning mode transition condition includes that the plant monitoring device receives a command for switching the operation mode to the learning mode. The plant monitoring device according to claim 1.

8. During operation in the learning mode of the plant monitoring device, the operation mode switching unit is configured to switch the operation mode from the learning mode to the monitoring mode when a specified time has elapsed since the operation mode was switched from the monitoring mode to the learning mode. The plant monitoring device according to claim 1 or 2.

9. A plant monitoring device for monitoring a plant, a measurement data acquisition unit configured to acquire measurement data of a plurality of variables indicating the state of the plant at regular intervals, a comparison unit configured to compare a deviation index value indicating the deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant, a reference data update unit configured to update the reference data set, and an operation mode switching unit configured to switch the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result by the comparison unit and a learning mode in which the measurement data, at least when the deviation index value is larger than the diagnostic threshold value, is incorporated into the reference data set by the reference data update unit. The operation mode switching unit is configured to switch the operation mode from the monitoring mode to the learning mode when, during the operation of the plant monitoring device in the monitoring mode, the deviation index value is greater than the diagnostic threshold value and the learning mode transition condition is satisfied. The operation mode switching unit is configured to switch the operation mode from the learning mode to the monitoring mode when the deviation index value becomes equal to or less than the diagnostic threshold value during the operation of the plant monitoring device in the learning mode. Plant monitoring device.

10. The reference data updating unit is configured to incorporate the measurement data into the reference data set at a higher frequency in the learning mode than in the monitoring mode. The plant monitoring device according to claim 1 or 2.

11. When the reference data updating unit incorporates one piece of the measurement data into the reference data set, it is configured to remove one piece of the reference data included in the reference data set, and to preferentially remove the measurement data incorporated during the operation in the learning mode rather than the measurement data incorporated during the operation in the monitoring mode among the reference data included in the reference data set. The plant monitoring device according to claim 1 or 2.

12. The deviation index value is the Mahalanobis distance for the measurement data calculated based on the unit space constituted by the reference data set. The plant monitoring device according to claim 1 or 2.

13. A plant monitoring method executed by a computer, comprising: a step of acquiring measurement data of a plurality of variables indicating the state of a plant at regular intervals; a comparison step of comparing a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant; a step of updating the reference data set; a monitoring step of monitoring the plant based on the comparison result in the comparison step; a learning step of incorporating at least the measurement data when the deviation index value is greater than the diagnostic threshold value into the reference data set; a switching step of switching between the execution steps between the monitoring step and the monitoring step. In the switching step, during the execution of the monitoring step, when the deviation index value is greater than the diagnostic threshold and the learning mode transition condition is satisfied, the step to be executed is switched from the monitoring step to the learning step. In the step of updating the reference data, during the execution of the monitoring step, among the measurement data newly acquired during the execution of the monitoring step, when the deviation index value is less than or equal to the diagnostic threshold, the measurement data is incorporated into the reference data set, and during the execution of the learning step, among the measurement data newly acquired during the execution of the learning step, at least the measurement data when the deviation index value is greater than the diagnostic threshold is incorporated into the reference data set Plant monitoring method.

14. A plant monitoring method executed by a computer, comprising: a step of acquiring measurement data of a plurality of variables indicating the state of a plant at regular intervals; a comparison step of comparing a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold for diagnosing an abnormality of the plant; a step of updating the reference data set; a monitoring step of monitoring the plant based on the comparison result in the comparison step; a learning step of incorporating at least the measurement data when the deviation index value is greater than the diagnostic threshold into the reference data set; a switching step of switching the step to be executed between the monitoring step and the monitoring step, and comprising: In the switching step, during the execution of the monitoring step, when the deviation index value is greater than the diagnostic threshold and the learning mode transition condition is satisfied, the step to be executed is switched from the monitoring step to the learning step. In the switching step, during the execution of the learning step, when the deviation index value becomes less than or equal to the diagnostic threshold, the step to be executed is switched from the learning step to the monitoring step. Plant monitoring method.

15. A plant monitoring program for operating a plant monitoring device for monitoring a plant, comprising: a computer with a procedure for acquiring measurement data of a plurality of variables indicating the state of the plant at regular intervals; A comparison procedure for comparing a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant, A procedure for updating the reference data set, A program for causing the plant monitoring device to execute a procedure for switching the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result in the comparison procedure and a learning mode for incorporating at least the measurement data when the deviation index value is greater than the diagnostic threshold value into the reference data set, In the procedure for switching the operation mode, during the operation of the plant monitoring device in the monitoring mode, when the deviation index value is greater than the diagnostic threshold value and the learning mode transition condition is satisfied, the operation mode is switched from the monitoring mode to the learning mode, In the procedure for updating the reference data, During the operation of the plant monitoring device in the monitoring mode, among the measurement data newly acquired during the operation in the monitoring mode, the measurement data when the deviation index value is less than or equal to the diagnostic threshold value is incorporated into the reference data set, and During the operation of the plant monitoring device in the learning mode, among the measurement data newly acquired during the operation in the learning mode, at least the measurement data when the deviation index value is greater than the diagnostic threshold value is incorporated into the reference data set Plant monitoring program.

16. A plant monitoring program for operating a plant monitoring device for monitoring a plant, Causing a computer to A procedure for acquiring measurement data of a plurality of variables indicating the state of the plant at regular intervals, A comparison procedure for comparing a deviation index value indicating a deviation between a reference data set, which is a set of reference data related to the plurality of variables, and the measurement data, with a diagnostic threshold value for diagnosing an abnormality of the plant, A procedure for updating the reference data set, A program for causing the plant monitoring device to execute a procedure for switching the operation mode of the plant monitoring device between a monitoring mode for monitoring the plant based on the comparison result in the comparison procedure and a learning mode for incorporating at least the measurement data when the deviation index value is greater than the diagnostic threshold value into the reference data set, In the procedure for switching the operation mode, during the operation of the plant monitoring device in the monitoring mode, when the deviation index value is greater than the diagnostic threshold value and the learning mode transition condition is satisfied, the operation mode is switched from the monitoring mode to the learning mode. In the procedure for switching the operation mode, during the operation of the plant monitoring device in the learning mode, if the deviation index value becomes less than or equal to the diagnostic threshold value, the operation mode is switched from the learning mode to the monitoring mode. Plant monitoring program.

Images