JP5944206B2 - Abnormality monitoring device and abnormality monitoring method for circulating fluidized bed boiler - Google Patents

Description

  The present invention relates to an abnormality monitoring device and an abnormality monitoring method for a circulating fluidized bed boiler.

  Conventionally, a monitoring system described in Patent Document 1 below is known. In this monitoring system, the Mahalanobis distance of the monitoring target data is calculated, and whether or not there is an abnormality in the monitoring target data is determined based on the Mahalanobis distance. When it is determined that there is an abnormality in the monitoring target data, the monitoring target data having a high contribution rate to the desired characteristics of the Mahalanobis distance is extracted as an abnormal signal from the plurality of monitoring target data.

JP 2011-90382 A

  Some monitoring systems for process systems such as electric power plants use multivariate analysis or data mining to determine the presence or absence of abnormality in monitored data. According to such a determination means, it is possible to determine the presence or absence of an abnormality due to a known phenomenon, but for example, the presence or absence of an abnormality due to a phenomenon with a very low probability of occurrence, such as blasting, has never occurred. It is difficult to determine whether there is an abnormality caused by an unknown phenomenon.

  In a system using the Mahalanobis Taguchi system (hereinafter also referred to as “MTS”) method, first, a unit space is created based on data acquired from a normally operating system. Next, the distance from the center of the unit space to the monitoring target data is calculated as the Mahalanobis distance. And the presence or absence of abnormality is determined based on this Mahalanobis distance. According to this MTS, it is possible to determine the presence / absence of an abnormality caused by a phenomenon having a very low probability of occurrence and the presence / absence of an abnormality caused by an unknown phenomenon that has never occurred. Furthermore, according to MTS, it is also possible to extract monitoring target data having a high contribution that is the degree of influence on the Mahalanobis distance from a plurality of monitoring target data.

  However, in a process system such as a power plant, the monitoring target data that can be acquired is limited, and a water supply / steam system and a fuel / exhaust gas system are combined in a complex manner. Therefore, when an abnormality monitoring system using MTS is applied to a process system, a phenomenon that causes an abnormality may not be reliably identified from an abnormality signal.

  Then, an object of this invention is to provide the abnormality monitoring apparatus of the circulating fluidized bed boiler which can pinpoint the cause of driving | operation abnormality more reliably.

  The abnormality monitoring apparatus of the present invention is an abnormality monitoring apparatus for a circulating fluidized bed boiler, and is an operation data acquiring means for acquiring operation data composed of sensor data of the circulating fluidized bed boiler, and a circulating flow based on the Mahalanobis distance of the operating data. When it is determined that there is an operation abnormality by an abnormality determination unit that determines the presence or absence of operation abnormality of the bed boiler, an imaging data acquisition unit that acquires imaging data of a predetermined imaging part of the circulating fluidized bed boiler, and an abnormality determination unit Data recording means for storing sensor data and imaging data of an imaging region related to the sensor data in synchronization.

  Further, the abnormality monitoring method of the present invention is an abnormality monitoring method for a circulating fluidized bed boiler, based on an operation data obtaining step for obtaining operation data composed of sensor data of the circulating fluidized bed boiler, and a Mahalanobis distance of the operation data. It is determined that there is an abnormality in the abnormality determination step for determining whether there is an abnormality in the operation of the circulating fluidized bed boiler, an imaging data acquisition step for acquiring imaging data of a predetermined imaging region of the circulating fluidized bed boiler, and an abnormality determination step. A data recording step of storing sensor data and imaging data of an imaging region related to the sensor data in synchronization.

  In this abnormality monitoring device and abnormality monitoring method, the abnormality determination means determines whether or not there is an abnormality in the operation of the circulating fluidized bed boiler based on the Mahalanobis distance of the operation data with respect to the unit space. Therefore, even if an abnormality caused by an unknown phenomenon occurs, it can be determined that there is an operation abnormality in the circulating fluidized bed boiler that is in operation. Moreover, the imaging data acquisition means acquires the imaging data of the predetermined imaging part of the circulating fluidized bed boiler. Since the data recording means synchronously stores the sensor data and the imaging data of the imaging region related to the sensor data, it is possible to analyze the cause of the driving abnormality based on the sensor data and the imaging data. Become. Therefore, the cause of the operational abnormality can be identified more reliably.

  Further, in the abnormality monitoring device of the present invention, when the abnormality determination unit determines that there is an operation abnormality in the circulating fluidized bed boiler, the contribution degree indicating the degree of influence of each sensor data on the Mahalanobis distance is determined for each sensor data. The data recording means calculates and saves the high contribution sensor data selected based on the contribution degree among the sensor data and the imaging data of the imaging part associated with the high contribution sensor data in synchronization. Also good. In this case, by calculating the contribution to the Mahalanobis distance, sensor data including the possibility of causing the driving abnormality is extracted, so that the cause of the driving abnormality can be identified more quickly. Furthermore, since the data recording means stores the high contribution sensor data and the imaging data of the imaging region associated with the sensor data in synchronization with each other, the driving abnormality is based on the high contribution sensor data and the imaging data. It becomes possible to analyze the cause of this. Therefore, the cause of the operational abnormality can be identified more reliably and quickly.

  In addition, when the abnormality determining unit determines that there is an operation abnormality in the circulating fluidized bed boiler, it may further include a data display unit that displays the operation data and the imaging data in association with each other. In this case, the cause of the driving abnormality can be analyzed based on the sensor data and the imaging data at the place where it is determined that there is the driving abnormality.

  The abnormality monitoring device of the present invention may further include unit space creating means for creating a unit space, and the unit space may be created based on operation data during normal operation of the circulating fluidized bed boiler.

  According to the abnormality monitoring device and abnormality monitoring method for a circulating fluidized bed boiler according to the present invention, the cause of the operation abnormality can be identified more reliably.

It is a figure which shows the structure of the abnormality monitoring apparatus of this embodiment. It is a figure which shows the example of the operation data acquisition means arrange | positioned at a circulating fluidized bed boiler. It is a figure which shows the example of the imaging data acquisition means arrange | positioned at a circulating fluidized bed boiler. It is a figure which shows an example of the imaging region linked | related with sensor data. It is a figure which shows the hardware which comprises a part of abnormality monitoring apparatus. It is a figure which shows the process of creating unit space. It is a figure which shows the process of monitoring the presence or absence of the driving | running abnormality of a circulating fluidized bed boiler. It is a figure which shows the process of calculating the contribution with respect to Mahalanobis distance.

  Hereinafter, embodiments of an abnormality monitoring apparatus and an abnormality monitoring method according to the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted.

  As shown in FIG. 1, the abnormality monitoring apparatus 1 of the present embodiment monitors the operating state of a circulating fluidized bed boiler 2 (hereinafter also simply referred to as “boiler”). First, the boiler 2 will be described. As shown in FIGS. 2 to 4, the boiler 2 is a circulating fluidized bed boiler of an external circulation type (Circulating Fluidized Bed type). The boiler 2 includes a fluidized bed furnace 3 having a vertically long cylindrical shape. A middle portion of the furnace 3 is provided with a fuel inlet 3a for introducing fuel and a gas outlet 3b for discharging combustion gas at the upper portion. The fuel supplied to the furnace 3 from the fuel input device 5 is input into the furnace 3 through the fuel input port 3a.

  A cyclone 7 that functions as a solid-gas separator is connected to the gas outlet 3 b of the furnace 3. The discharge port 7a of the cyclone 7 is connected to a downstream gas processing system via a gas line. A return line 9 called a downcomer extends downward from the bottom outlet of the cyclone 7, and the lower end of the return line 9 is connected to the intermediate side surface of the furnace 3.

  In the furnace 3, the combustion / flowing air introduced from the lower air supply line 3 c causes the solid matter including the fuel input from the fuel input port 3 a to flow, and the fuel flows to about 800 to 900 ° C. while flowing. Burn with. A combustion gas generated in the furnace 3 is introduced into the cyclone 7 with accompanying solid particles. The cyclone 7 separates solid particles and gas by a centrifugal separation action, returns the solid particles separated via the return line 9 to the furnace 3, and removes the combustion gas from which the solid particles have been removed from the discharge port 7 a to the gas line. To the subsequent gas processing system.

  In this furnace 3, a solid material called “in-furnace bed material” is generated and collected at the bottom, and the bed material is sintered and melted and solidified by the concentration of impurities (low melting point materials, etc.) in the in-furnace bed material, or It is necessary to suppress malfunctions caused by incombustible impurities. For this reason, in the furnace 3, the in-furnace bed material is discharged | emitted regularly outside from the discharge port 3d of the bottom part. The discharged bed material is put into the furnace 3 again after removing unsuitable materials such as metal on a circulation line (not shown).

  The gas treatment system includes a gas heat exchange device 13 connected to the discharge port 7a of the cyclone 7 via a gas line, and a bag filter connected to the discharge port 13a of the gas heat exchange device 13 via a gas line. (Dust collector) 15. The gas heat exchanger 13 is provided with a boiler tube 13b that allows water to flow across the exhaust gas flow path. When the high-temperature exhaust gas sent from the cyclone 7 comes into contact with the boiler tube 13b, the heat of the exhaust gas is recovered in the water in the tube, and the generated high-temperature steam is sent to the turbine for power generation through the boiler tube 13b. The bag filter 15 removes fine particles such as fly ash that are still accompanying the combustible gas. The clean gas discharged from the discharge port 15a of the bag filter 15 is discharged from the chimney 19 via the gas line and the pump 17 to the outside.

  Next, the abnormality monitoring device 1 will be described. As shown in FIG. 1, the abnormality monitoring device 1 determines whether or not there is an operation abnormality of the boiler 2 based on operation data of the boiler 2. Next, when the abnormality monitoring apparatus 1 determines that the boiler 2 has an operation abnormality, the abnormality monitoring apparatus 1 extracts high contribution sensor data that is sensor data including a possibility of causing the operation abnormality. Then, the abnormality monitoring apparatus 1 records sensor data and imaging data obtained by imaging a part of the boiler associated with the sensor data in synchronization.

  The abnormality monitoring device 1 determines whether or not there is an abnormality in the operation of the boiler 2 based on a method of MTS (Mahalanobis taguchi System: Mahalanobis Taguchi System). A feature of MTS for pattern recognition is that a unit space is defined and the distance from the center to the target data is obtained as the Mahalanobis distance. The unit space is the center or reference for pattern recognition. The Mahalanobis distance, which is the distance between the center of the existing data group and the target data, is obtained using the correlation between all variables. Details of the Mahalanobis distance will be described later.

  The abnormality monitoring device 1 includes an operation data acquisition unit 21 that acquires operation data, an imaging data acquisition unit 22 that acquires imaging data, and a data processing unit 23 that processes the operation data and imaging data as functional components. I have. Hereinafter, the function of each component will be described.

  The operation data acquisition unit 21 acquires operation data indicating the operation state of the boiler 2. As will be described later, the temperature sensor 24, the pressure sensor 26, and the flow rate sensor 27 included in the operation data acquisition unit 21 are installed in each part of the boiler 2. The operation data refers to a data group composed of sensor data of these sensors. The operation data includes data such as temperature, pressure, and flow rate of each part of the boiler 2 and indicates the operation state of the boiler 2. The operation data acquired by the operation data acquisition unit 21 is output to the data processing unit 23.

  An arrangement example of the temperature sensor 24, the pressure sensor 26, and the flow sensor 27 will be described. In the example described below, the temperature sensors 24a to 24d and 24f to 24k belong to the temperature sensor 24 shown in FIG. 1, and the pressure sensors 26a to 26d and 26f to 26k belong to the pressure sensor 26 shown in FIG. The sensors 27e, 27h, 27j belong to the flow sensor 27 shown in FIG.

  As shown in FIG. 2, a temperature sensor 24 a and a pressure sensor 26 a are disposed in the vicinity of the discharge port 3 d of the furnace 3. The temperature sensor 24a acquires temperature data of the lower part of the furnace 3, and the pressure sensor 26a acquires pressure data of the lower part of the furnace 3. A temperature sensor 24 b and a pressure sensor 26 b are disposed in the middle part of the furnace 3. The temperature sensor 24 b acquires temperature data of the intermediate part of the furnace 3, and the pressure sensor 26 b acquires pressure data of the intermediate part of the furnace 3. A temperature sensor 24 c and a pressure sensor 26 c are disposed on the inner upper surface 3 e of the furnace 3. The temperature sensor 24 c acquires temperature data of the upper part of the furnace 3, and the pressure sensor 26 c acquires pressure data of the upper part of the furnace 3. A temperature sensor 24d and a pressure sensor 26d are arranged at the gas outlet 3b. The temperature sensor 24d acquires exhaust gas temperature data at the gas outlet 3b, and the pressure sensor 26d acquires exhaust gas pressure data at the gas outlet 3b.

  A flow rate sensor 27e is disposed in the air supply line 3c. The flow rate sensor 27e acquires flow rate data of air flowing through the air supply line 3c.

  A temperature sensor 24f and a pressure sensor 26f are disposed in the vicinity of the heat exchange part 13c of the gas heat exchange device 13. The temperature sensor 24f acquires the temperature data of the exhaust gas in the gas heat exchange device 13, and the pressure sensor 26f acquires the pressure data of the exhaust gas in the gas heat exchange device 13. A temperature sensor 24g and a pressure sensor 26g are disposed in the vicinity of the discharge port 13a of the gas heat exchange device 13. The temperature sensor 24g acquires the temperature data of the exhaust gas discharged from the discharge port 13a, and the pressure sensor 26g acquires the pressure data of the exhaust gas discharged from the discharge port 13a.

  A temperature sensor 24h and a flow rate sensor 27h are disposed on the upstream side of the heat exchange unit 13c of the gas heat exchange device 13. The temperature sensor 24h acquires the temperature data of the water in the tube before recovering the heat of the exhaust gas, and the flow sensor 27h acquires the flow data of the water in the tube before recovering the heat of the exhaust gas. A temperature sensor 24j, a pressure sensor 26j, and a flow rate sensor 27j are disposed on the downstream side of the heat exchange portion 13c of the boiler tube 13b. The temperature sensor 24j acquires the temperature data of the water vapor in the tube, the pressure sensor 26j acquires the pressure data of the water vapor in the tube, and the flow sensor 27j acquires the flow data of the water vapor in the tube.

  A temperature sensor 24k and a pressure sensor 26k are disposed near the outlet 15a of the bag filter 15. The temperature sensor 24k acquires exhaust gas temperature data near the exhaust port 15a, and the pressure sensor 26k acquires exhaust gas pressure data near the exhaust port 15a.

  As shown in FIG. 1, the imaging data acquisition unit 22 acquires imaging data of a predetermined imaging part of the boiler 2. As will be described later, the plurality of cameras 28 included in the imaging data acquisition unit 22 are installed in each part of the boiler 2. The imaging data is image data of a predetermined imaging part of the boiler 2 captured by these cameras 28. The imaging data may be a still image or a moving image. The imaging data acquired by the imaging data acquisition unit 22 is output to the data processing unit 23.

  An example of a predetermined imaging part and an arrangement example of a camera for imaging the imaging part will be described. In the example described below, a plurality of cameras 28a to 28e belong to the camera 28 shown in FIG. As shown in FIG. 3, the predetermined imaging part includes a part 16 a near the discharge port 3 d of the furnace 3 and a part 16 b on the side wall surface of the furnace 3. Furthermore, the predetermined imaging part includes a part 16c of the heat exchange part 13c of the gas heat exchange device 13, a part 16d of the boiler tube 13b upstream of the heat exchange part 13c, and a boiler tube downstream of the heat exchange part 13c. There is a portion 16e of 13b. A camera 28 a for imaging the region 16 a is disposed near the fuel inlet 3 a of the furnace 3. The camera 28b for imaging the part 16b is disposed on the inner wall of the furnace 3 facing the side wall surface to be imaged. The camera 28c for imaging the part 16c is disposed on the inner wall of the gas heat exchange device 13 facing the heat exchange unit 13c. The camera 28d for imaging the part 16d is disposed in the vicinity of the part 16d. The camera 28e for imaging the part 16e is disposed in the vicinity of the part 16e.

  Here, an example of imaging data associated with sensor data will be described. As shown in FIG. 4, the camera 28c for imaging the heat exchange unit 13c of the gas heat exchange device 13 includes exhaust gas temperature data of a temperature sensor 24f provided in the vicinity of the heat exchange unit 13c of the gas heat exchange device 13, and It is associated with the exhaust gas pressure data of the pressure sensor 26f. That is, when an abnormality appears in the exhaust gas temperature data acquired by the temperature sensor 24f and / or the exhaust gas pressure data acquired by the pressure sensor 26f, there is a possibility that there is an abnormality in the heat exchange unit 13c that is a water supply / steam pipe. high. Therefore, the image data of the heat exchange unit 13c imaged by the camera 28c is associated with the exhaust gas temperature data acquired by the temperature sensor 24f and the exhaust gas pressure data acquired by the pressure sensor 26f.

  As shown in FIG. 1, the data processing unit 23 processes the sensor data input from the operation data acquisition unit 21 to determine whether the boiler 2 is operating abnormally. When the data processing unit 23 determines that there is a driving abnormality, the data processing unit 23 records the imaging data input from the imaging data acquisition unit 22 and the sensor data in synchronization. When the data processing unit 23 determines that there is a driving abnormality, the data processing unit 23 outputs a command for displaying the data on the display unit 31 to be described later to the display unit 31 in synchronization with the imaging data and the sensor data. Furthermore, the data processing unit 23 acquires operation data during normal operation of the boiler 2 in advance, and creates a unit space used for determination of operation abnormality by the MTS method described later based on the operation data. The data processing unit 23 is realized by using, for example, a computer 100 shown in FIG.

  As shown in FIGS. 1 and 5, the computer 100 is an example of hardware that constitutes the data processing unit 23 of the present embodiment. The computer 100 includes a CPU and various data processing devices such as a server device that performs processing and control by software and a personal computer. The computer 100 includes a CPU 41, a RAM 42 and a ROM 43 that are main storage devices, an input device 44 such as a keyboard and a mouse that are input devices, an output device 46 such as a display, a communication module 47 that is a data transmission / reception device such as a network card, a hard disk, and the like. The auxiliary storage device 48 is configured as a computer system. The input device 44 constitutes an input unit 29 described later, and the output device 46 constitutes a display unit 31 described later (see FIG. 1). The functional components shown in FIG. 1 include an input device 44, an output device 46, and a communication module under the control of the CPU 41 by reading predetermined computer software on hardware such as the CPU 41 and the RAM 42 shown in FIG. 47 is operated, and data is read and written in the RAM 42 and the auxiliary storage device 48.

  As shown in FIG. 1, sensor data is input from the operation data acquisition unit 21 and imaging data is input from the imaging data acquisition unit 22 to the data processing unit 23. When the data processing unit 23 determines that there is a driving abnormality by determining whether or not there is a driving abnormality, a unit space generating unit 32 that creates a unit space, a temporary recording unit 33 that temporarily records driving data for a predetermined period, and Includes an abnormality determination unit 34 that calculates a contribution degree, and an operation data synchronization recording unit 36 that records sensor data and imaging data in synchronization.

  The unit space creating unit 32 collects sensor data during normal operation and creates a unit space. The unit space creation unit 32 calculates and records a matrix used for calculating the Mahalanobis distance based on the unit space. The unit space creation unit 32 is referred to when the abnormality determination unit 34 calculates the Mahalanobis distance, and outputs a matrix based on the unit space to the abnormality determination unit 34.

  The temporary recording unit 33 has a function of recording the operation data output from the operation data acquisition unit 21 and the imaging data output from the imaging data acquisition unit 22 for a predetermined period. When the abnormality determination unit 34 determines that there is an operation abnormality, the temporary recording unit 33 synchronizes the sensor data and imaging data for several seconds to several minutes before and after the abnormality occurrence point, and records operation data synchronously. To the unit 36.

  The abnormality determination unit 34 calculates the Mahalanobis distance based on the sensor data output from the temporary recording unit 33 and the matrix based on the unit space output from the unit space creation unit 32. And the abnormality determination part 34 compares the threshold value previously recorded in the abnormality determination part 34, and Mahalanobis distance, and when it exceeds a threshold value, it determines with the boiler 2 having driving | operation abnormality. When it is determined that there is a driving abnormality, the abnormality determination unit 34 calculates a contribution indicating the degree of influence on the Mahalanobis distance for each sensor data, and extracts high contribution sensor data. Then, the abnormality determination unit 34 outputs a command to the temporary recording unit 33 so as to record the high contribution sensor data, the imaging data associated with the high contribution sensor data, and the Mahalanobis distance in synchronization. To do. Further, the abnormality determination unit 34 outputs the high contribution sensor data and the imaging data to the display unit 31 so as to display the high contribution sensor data and the imaging data.

  The operation data synchronization recording unit 36 records the sensor data output from the temporary recording unit 33, the imaging data, and the Mahalanobis distance of the operation data with respect to the unit space in synchronization. The data recorded in the operation data synchronization recording unit 36 is used as analysis data when analyzing the cause of the operation abnormality. The operation data synchronization recording unit 36 may compress and record each data. Further, the driving data synchronization recording unit 36 may record the high contribution sensor data and the imaging data associated with the high contribution sensor data, and the driving data acquisition unit 21 and the imaging data acquisition unit 22 may record the driving data synchronization recording unit 36. All acquired data may be recorded.

  The data processing unit 23 may include an input unit 29 and a display unit 31 as necessary. The input unit 29 includes an input device such as a keyboard and a mouse. A threshold necessary for data processing is input from the input unit 29. The display unit 31 includes an image display device such as a CRT display or a liquid crystal display. The display unit 31 displays that a driving abnormality has occurred, and also displays the high contribution sensor data, imaging data, and the like output from the abnormality determination unit 34. Note that the display unit 31 may further include a bell, an audio speaker, an illumination lamp, an LED, and the like in order to reliably notify that a driving abnormality has occurred, thereby notifying the occurrence of a driving abnormality. .

  Next, an abnormality monitoring method based on the MTS technique using the abnormality monitoring apparatus 1 will be described. The abnormality monitoring method includes a step S10 for creating the unit space shown in FIG. 6, a step S20 for monitoring the operating state shown in FIG. 7, and a step S30 for analyzing the factors shown in FIG.

<Unit space creation step S10>
With reference to FIG. 6, step S10 for creating a unit space will be described. This step S10 is mainly performed by the unit space creation unit 32 of the data processing unit 23. Since the purpose of the abnormality monitoring method of the present embodiment is to determine whether or not the boiler 2 is operating abnormally, the unit space of the present embodiment is a set of operation data acquired from the boiler 2 that is operating normally. In step S10, to create a matrix A required for calculating the Mahalanobis distance D ² from the unit space is a population of the operation data. This step S10 is executed before the step S20 for monitoring operation abnormality is started. Further, step S10 may be executed every predetermined period, for example, every year, or may be executed at an arbitrary time. That is, it is only necessary that the matrix A is created by executing the step S10 before executing the step S20 for monitoring the operating state.

Operation data is acquired in process S11. This operation data is operation data of the boiler 2 during normal operation of the boiler 2. In step S11, operation data is acquired while an operator confirms that there is no operation abnormality in the boiler 2 in operation. Next, in step S12, the acquired operation data is added to the unit space. Subsequently, in step S13, it is determined whether the number of data points accumulated in the unit space satisfies the number of data points necessary for creating the matrix A. More specifically, in step S13, if it contains the k sensor data x ₁ ~x _k on the operating data, the data points n of the operating data is equal to or greater than or equal to k is determined. When the number of data points n is less than k (NO), the process returns to step S11, and further operation data is acquired. When the number of data points n is greater than or equal to k (YES), the matrix A is created in step S14.

次に、相関行列Ｒを求める。式（２）に示すように、相関行列は、ｋ×ｋの正方行列である。相関行列Ｒの各成分ｒ_ｉｊは、規準化したセンサデータＸ_ｉと、規準化したセンサデータＸ_ｊとの相関係数である。

次に、式（３）に示される相関行列Ｒの逆行列Ａを求める。

Step S14 for creating the matrix A will be described in detail. The matrix A is an inverse matrix A = R ⁻¹ of the correlation matrix R whose components are correlation coefficients between the sensor data x _{1 to} x _k . First, to normalize the sensor data _x 1 ~x _k. Normalized sensor data _x 1 ~x _k is performed by equation (1). In the normalized sensor data X _{1 to} X _k , the average value is zero and the standard deviation is 1.

Next, a correlation matrix R is obtained. As shown in Expression (2), the correlation matrix is a k × k square matrix. Each component r _ij of the correlation matrix R is a correlation coefficient between the normalized sensor data X _i and the normalized sensor data X _j .

Next, an inverse matrix A of the correlation matrix R shown in Expression (3) is obtained.

Matrix A is created in step S14, is recorded in the unit space creating unit 32, when calculating the Mahalanobis distance D ² which will be described later, it is referred to by the abnormality determination unit 34.

<Abnormality monitoring step S20>
With reference to FIG. 7, abnormality monitoring process S20 which monitors the driving | running state of the boiler 2 is demonstrated. This step S20 is mainly performed by the abnormality determination unit 34 of the data processing unit 23. First, in step S21 (operation data acquisition step, imaging data acquisition step), temperature sensors 24a to 24d, 24f to 24k, pressure sensors 26a to 26d, 26f to 26k, and flow rate sensors 27e and 27h arranged in each part of the boiler 2. , 27j to obtain operation data. In addition to the acquisition of the operation data, the imaging data is acquired by the cameras 28a to 28e. The acquired operation data and imaging data are recorded in the temporary recording unit 33. Then, the operation data recorded in the temporary recording unit 33 is transmitted to the abnormality determination unit 34. Next, in step S21, it calculates the Mahalanobis distance ^{D 2} of the operating data.

本実施形態では、変数であるセンサデータｘ_１〜ｘ_ｋは１つではなく、例えばｋ個ある。式（４）を、ｋ個の変数を有する式に変形する。式（４）で示されたマハラノビス距離Ｄ^２を式（５）に示すように変形する。

この式の第１項及び第３項をベクトル形式に書き換えると共に、第２項を分散共分散行列に書き換える。

ここで、変数iと変数jの共分散s_ijは、相関係数r_ijにより式（７）のように示される。

したがって、マハラノビス距離Ｄ^２を相関係数で表すと式（８）により示される。

さらに、式（８）のセンサデータｘ_１〜ｘ_ｋを規準化したセンサデータＸ_１〜Ｘ_ｋ（式（９）参照）で表すと、式（８）は式（１０）のように示される。

式（１０）を変数の数であるｋで除して、マハラノビス距離Ｄ^２を式（１１）のように表す。

工程Ｓ２２では、各センサデータｘ_１〜ｘ_ｋを規準化したセンサデータＸ_１〜Ｘ_ｋに変換する。そして、規準化したセンサデータＸ_１〜Ｘ_ｋと行列Ａとを式（１１）に適用してマハラノビス距離Ｄ^２を算出する。 Here will be described the Mahalanobis distance ^{D 2.} Mahalanobis distance D ² is the variation is used as a reference for measuring the distance from the center of the distribution is arbitrary. For example, the Mahalanobis distance D ² variable is one is shown in formula (4).

In the present embodiment, the sensor data x _{1 to} x _k that are variables are not one, but, for example, there are k pieces. Equation (4) is transformed into an equation having k variables. Deforming the Mahalanobis distance D ² which is represented by the formula (4) as shown in equation (5).

The first term and the third term of this equation are rewritten to the vector format, and the second term is rewritten to the variance-covariance matrix.

Here, the covariance s _ij of the variable i and the variable j is expressed by the correlation coefficient r _ij as shown in Expression (7).

Thus, to represent the Mahalanobis distance D ² by the correlation coefficient represented by the formula (8).

Furthermore, when the sensor data X _{1 to} X _k (refer to the formula (9)) _obtained by normalizing the sensor data x _{1 to} x _k of the formula (8) is expressed, the formula (8) is expressed as the formula (10). .

Divided by k is the number of the formula (10) a variable representing the Mahalanobis distance D ² by the equation (11).

In step S22, it converts each sensor data _x 1 ~x _k to sensor data _X 1 to X _k was normalized. Then, the normalized sensor data X _{1 to} X _k and the matrix A are applied to the equation (11) to calculate the Mahalanobis distance D ² .

Subsequently, in step S23 (abnormality determining step), and compares the Mahalanobis distance ^{D 2} and the threshold. The threshold is input from the input unit 29 and recorded in advance in the abnormality determination unit 34, for example. The threshold value is set in advance to a desired value. If the Mahalanobis distance D ² is less than the threshold value (NO), it is determined that there is no abnormal operation process proceeds to step S28.

On the other hand, if the Mahalanobis distance D ² is equal to or greater than the threshold (YES), it is determined that there is abnormal operation proceeds to step S25. Then, in step S25, the extracted set of sensor data or sensor data that can cause abnormal operation in the sensor data x ₁ ~x _k included in the operation data.

<Factor analysis step S30>
With reference to FIG. 8, factor analysis process S30 which analyzes the factor of driving | operation abnormality is demonstrated. In factor analysis step S30, the abnormality monitoring in the step S23 in step S20 if the Mahalanobis distance ^{D 2} exceeds the threshold value, contributing to that the sensor data _x 1 ~x _k or sensor data _x 1 ~ to increase the Mahalanobis distance ^{D 2} to analyze the set of x _k.

In step S31, it acquires the sensor data _x 1 ~x _k to be factor analyzed. Here, the sensor data _x 1 ~x _k, at step S23, a sensor data _x 1 ~x _k constituting a Mahalanobis distance greater than a threshold.

Next, in step S32, an orthogonal table is created. An orthogonal table compresses many combinations rationally and systematically. An orthogonal table generally used in the factor analysis of MTS is a second-level orthogonal table. Table 1 is an example of the orthogonal table in the case constituted operating data from seven sensor data x ₁ ~x _7.

Subsequently, in step S33, the SN ratio of the desired size characteristic is calculated. Sensor data not belonging to the unit space, i.e., the Mahalanobis distance D ² of the sensor data that causes abnormal operation is greater is desirable. In this way, a characteristic that is more desirable as it is larger is called a desired characteristic. Moreover, since there are a plurality of sensor data, there is generally variation. The signal-to-noise ratio is a quantification of desirability including this variation. This SN ratio is calculated by the equation (12).

The procedure for calculating the SN ratio in step S33 will be described in detail. Reference is made to the first column of the number of data in the orthogonal table of Table 1. The first column, all of the sensor data x ₁ ~x ₇ is the first level. Therefore, Mahalanobis distance _D ^{1 2} using all of the sensor data _x 1 ~x _7. Subsequently, reference is made to the second column of the orthogonal table of Table 1. In the second column, the sensor data x _{1 to} x ₃ are at the first level, and the sensor data x _{4 to} x ₇ are at the second level. Therefore, Mahalanobis distance _D ^{2 2} by using the sensor data _x 1 ~x _3. Hereinafter, the Mahalanobis distances D ₃ ^{2 to} D ₈ ² corresponding to the respective columns of the third to eighth columns of the orthogonal table of Table 1 are obtained by the same procedure.

Then, determine the SN ratio for each sensor data _x 1 ~x _7. Here, the first level SN ratio and the second level SN ratio are obtained for each of the sensor data x _{1 to} x ₇ . Here, the S / N ratio of the first level is an S / N ratio calculated using the Mahalanobis distance in the number of samples that is the first level among the Mahalanobis distances D ₁ ^{2 to} D ₈ ² . Further, the SN ratio of the second level of the Mahalanobis distance D ₁ ² to D ₈ ^2, a SN ratio calculated by using the Mahalanobis distance in number of samples is the second level.

For example, if the sensor data _{x 1,} the number of samples is the first level is 1-4. Therefore, the SN ratio of the first level is calculated using the Mahalanobis distances D ₁ ^{2 to} D ₄ ² . Moreover, the number of samples which are the 2nd level is 5-8. Therefore, to calculate the SN ratio of the second level by using the Mahalanobis distance _D ⁵ 2 _{to D} ^{8 2.} The first level SN ratio η _A1 of the sensor data x ₁ is calculated by the equation (13). Further, the second level SN ratio η _A2 of the sensor data x ₁ is calculated by the equation (14).

In step S34, the contribution is obtained. When attention is paid to the sensor data x _1, and SN ratio eta _A1 of the first level, when the difference between the SN ratio eta _A2 of the second level is large in the positive direction, the sensor data x ₁ is the cause of abnormal operation It has a possibility of becoming. The difference (η _A1 −η _A2 ) between the SN ratio η _A1 of the first level and the SN ratio η _A2 of the second level is called a contribution. Hereinafter, to calculate the contribution of the sensor data _x 2 ~x ₈ in the same procedure.

Through the above steps S31 to S34, the contribution of each sensor data _x 1 ~x ₇ is calculated. Those having the highest contribution or exceeding a preset contribution threshold are extracted as high contribution sensor data or a set of high contribution sensor data.

  Referring to FIG. 7 again, in step S26 (data recording step), sensor data and imaging data are recorded synchronously. This step S26 is mainly executed by the operation data synchronization recording unit 36. Information of the sensor data extracted as the high contribution sensor data in step S25 is output from the abnormality determination unit 34 to the temporary recording unit 33. The temporary recording unit 33 synchronizes the sensor data extracted as the high contribution sensor data based on the information output from the abnormality determination unit 34 with the imaging data associated with the sensor data, and synchronizes the driving data. The data is output to the recording unit 36. The operation data synchronization recording unit 36 records the sensor data and the imaging data output from the temporary recording unit 33.

  In step S27, the display unit 31 displays that an abnormal operation has occurred. This step S27 is mainly executed by the display unit 31. Information indicating that a driving abnormality has occurred is output from the abnormality determination unit 34 to the display unit 31, and the display unit 31 notifies that a driving abnormality has occurred based on this information. Further, the sensor data extracted as the high contribution sensor data and the imaging data of the camera associated with the sensor data are output from the abnormality determination unit 34 to the display unit 31, and these are displayed on the display unit 31. Is done.

  In step S28, it is determined whether or not to end the abnormality monitoring step S20. When it is determined that the abnormality monitoring step S20 is not ended (NO), that is, when the abnormality monitoring operation is continued, the process proceeds to step S21 and the operation from step S21 is performed again. If it is determined that the abnormality monitoring step S20 is to be ended (YES), the abnormality monitoring step S20 is ended.

  As described above, in the abnormality monitoring device and abnormality monitoring method for the circulating fluidized bed boiler according to the present embodiment, the abnormality determination unit 34 causes the boiler in step S23 to be based on the Mahalanobis distance of the operation data with respect to the unit space calculated in step S22. 2 is judged whether or not there is an abnormal operation. Therefore, even when an operation abnormality caused by an unknown phenomenon occurs, it can be determined that the boiler 2 in operation has an operation abnormality. Further, the camera 28 of the imaging data acquisition unit 22 acquires the imaging data of the predetermined imaging parts 16 a to 16 e of the boiler 2. Then, the operation data synchronization recording unit 36 stores the sensor data having a high contribution in step S26 and the imaging data of the imaging region related to the sensor data in synchronization. Therefore, it is possible to analyze the cause of the abnormal operation based on the sensor data and the imaging data. Therefore, the cause of the operational abnormality can be identified more reliably.

  Further, in the abnormality monitoring device and abnormality monitoring method for the circulating fluidized bed boiler of the present embodiment, the contribution to the Mahalanobis distance is calculated in step S30. Therefore, sensor data including the possibility of causing the operation abnormality is extracted, so that the cause of the operation abnormality can be identified more quickly. Further, since the high-contribution sensor data and the imaging data of the imaging region associated with the sensor data are associated with each other and stored in synchronization in step S26, the high-contribution sensor data and the imaging data are used. Therefore, it is possible to analyze the cause of abnormal operation. Therefore, the cause of the operational abnormality can be identified more reliably and quickly.

  Further, the abnormality monitoring apparatus 1 for the circulating fluidized bed boiler of the present embodiment includes a display unit 31. Therefore, the cause of the driving abnormality can be analyzed based on the sensor data and the imaging data at the place where it is determined that there is the driving abnormality.

  In addition, embodiment mentioned above shows an example of the abnormality monitoring apparatus 1 and the abnormality monitoring method. The abnormality monitoring apparatus 1 and the abnormality monitoring method according to the present invention are not limited to the above-described embodiment, and the abnormality monitoring apparatus 1 and the abnormality monitoring according to the above-described embodiment are within the scope not changing the gist described in the claims. The method may be modified or applied to others.

Data other than temperature data, pressure data, and flow rate data may be used as the sensor data. For example, the concentration of a desired substance contained in the exhaust gas such as oxygen concentration, NO _X concentration, CO concentration, SO ₂ concentration, or the amount of fuel input from the fuel input port 3a may be used. Moreover, the place where the temperature sensor 24, the pressure sensor 26, the flow sensor 27, and the camera 28 are arranged may be arranged at a place different from the above-described place.

DESCRIPTION OF SYMBOLS 1 ... Abnormality monitoring apparatus, 2 ... Circulating fluidized bed boiler, 21 ... Operation data acquisition part, 22 ... Imaging data acquisition part, 23 ... Data processing part, 24 ... Temperature sensor, 26 ... Pressure sensor, 27 ... Flow rate sensor, 28 ... camera, 29 ... input unit, 31 ... display unit, 32 ... unit space creating unit, 33 ... temporary storage unit, 34 ... abnormality determining unit, 36 ... operation data synchronization recording unit, 100 ... computer, D 2 ^... Mahalanobis distance, x ₁ ~x k _... sensor data, S10 ... unit space creating step, S20 ... abnormality monitoring step, S30 ... factor analysis step.

Claims (5)

An abnormality monitoring device for a circulating fluidized bed boiler,
Operation data acquisition means for acquiring operation data consisting of a plurality of sensor data of the circulating fluidized bed boiler;
An abnormality determining means for determining the presence or absence of an operation abnormality of the circulating fluidized bed boiler based on the Mahalanobis distance of the operation data;
In the circulating fluidized bed boiler , imaging data acquisition means for acquiring imaging data of an imaging region related to each of the sensor data ;
When the abnormality determining unit determines that there is an operation abnormality, the sensor data having a large contribution to the operation abnormality in the plurality of sensor data and the imaging related to the sensor data having a large contribution to the operation abnormality An abnormality monitoring device for a circulating fluidized bed boiler, comprising data recording means for storing the imaging data of a part in synchronization. The abnormality determining means includes
When it is determined that there is an abnormal operation in the circulating fluidized bed boiler,
A contribution indicating the degree of influence each sensor data has on the Mahalanobis distance is calculated for each sensor data,
The data recording means is
The high contribution sensor data selected based on the contribution degree among the sensor data and the imaging data of the imaging region associated with the high contribution sensor data are stored in synchronization with each other. An abnormality monitoring device for a circulating fluidized bed boiler as described.   The circulation according to claim 1 or 2, further comprising data display means for displaying the operation data and the imaging data in association with each other when the abnormality determination means determines that there is an operation abnormality in the circulating fluidized bed boiler. Abnormality monitoring device for fluidized bed boiler.   The abnormality monitoring apparatus for a circulating fluidized bed boiler according to any one of claims 1 to 3, further comprising unit space creating means for creating a unit space based on the operation data during normal operation of the circulating fluidized bed boiler. . An abnormality monitoring method for a circulating fluidized bed boiler,
An operation data acquisition step of acquiring operation data consisting of a plurality of sensor data of the circulating fluidized bed boiler,
An abnormality determination step for determining presence or absence of an operation abnormality of the circulating fluidized bed boiler based on the Mahalanobis distance of the operation data;
In the circulating fluidized bed boiler , an imaging data acquisition step of acquiring imaging data of an imaging site related to each of the sensor data ;
When it is determined in the abnormality determination step that there is an operation abnormality, the sensor data having a large contribution to the operation abnormality in the plurality of sensor data and the imaging related to the sensor data having a large contribution to the operation abnormality A method for monitoring an abnormality of a circulating fluidized bed boiler, comprising: a data recording step of storing the imaging data of a part in synchronization.

Images