Precursor Detection Device and Precursor Detection Method

Abstract

To detect a sudden change vibration well in advance of the occurrence of the sudden change vibration.SOLUTION: A sign detection device includes a plurality of sensors, a data acquisition unit, a computation unit, and a detection unit. The plurality of sensors are arranged respectively at a plurality of positions of a detection target to measure physical quantities at their respective positions. The data acquisition unit acquires time-series change data of physical quantities from the plurality of sensors. The computation unit computes dynamic network information representing temporal changes in a complex network structure, including a parameter indicating a correlation between the physical quantities at two arbitrary positions among the plurality of positions, based on the time-series change data. The detection unit detects a sign of sudden change vibration of the detection target based on the dynamic network information.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a sign detection device and a sign detection method for detecting a sign of sudden vibration. [Background technology]

[0002] In machines such as gas turbines, steam turbines, engines, boilers, aircraft, and compressors, combustion vibration and shaft vibration can occur in the combustor, compressor, blades, etc. Among these vibrations, unstable vibrations that tend to change suddenly (sudden change vibrations) reach a limit cycle in a short time after an increase in vibration occurs. Reaching the limit cycle can lead to tripping or place a large burden on the equipment.

[0003] Therefore, it is desirable to avoid such sudden vibrations at an early stage. However, since the increase in vibration before the limit cycle is reached is short, there are cases where the sudden vibration cannot be avoided by control after the increase in vibration is detected. In order to avoid sudden vibrations, it is necessary to detect the signs of sudden vibrations well before their occurrence.

[0004] In recent years, detection techniques have been proposed that aim to detect sudden oscillations in advance. For example, Patent Document 1 discloses a device that detects combustion oscillations using a value related to the pressure in a combustor of a gas turbine. This device is configured to acquire a value related to the pressure in the combustor of the gas turbine, calculate network entropy, and detect the occurrence of combustion oscillations when the network entropy falls below a threshold value. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2018-80621 A Summary of the Invention [Problem to be solved by the invention]

[0006] As a result of intensive studies by the inventors of the present application, it has been found that the correlation of physical quantities (for example, pressure) at a plurality of positions is important in detecting the sign of sudden vibration. By using a parameter indicating such a correlation, it becomes possible to detect the sign of sudden vibration.

[0007] However, even if time-series variation data of a physical quantity (a value related to the pressure in the combustor) at one position is acquired and the network entropy is calculated as in Patent Document 1, since the correlation of the physical quantities at a plurality of positions is not considered, it is difficult to detect the sign sufficiently prior to the occurrence of sudden vibration.

[0008] The present disclosure has been made in view of the above circumstances, and an object thereof is to provide a sign detection device and a sign detection method capable of detecting sudden vibration sufficiently prior to the occurrence of sudden vibration.

Means for Solving the Problems

[0009] The sign detection device according to the present disclosure is a plurality of sensors respectively arranged at a plurality of positions in a detection object and configured to measure physical quantities at each position, a data acquisition unit for acquiring time-series variation data of the physical quantity from the plurality of sensors, an arithmetic unit for calculating dynamic network information representing a time change of a complex network structure including a parameter indicating the correlation between the physical quantities at any two positions among the plurality of positions based on the time-series variation data, a detection unit for detecting a sign of sudden vibration of the detection object based on the dynamic network information, and includes.

[0010] The sign detection method according to the present disclosure is a step in which sensors respectively arranged at a plurality of positions in a detection object measure physical quantities at each position, a step of acquiring time-series variation data of the physical quantity from the plurality of sensors, Calculating dynamic network information representing the temporal change of a complex network structure including a parameter indicating the correlation between the physical quantities at any two positions among the plurality of positions based on the time series variation data; Detecting a sign of a sudden change in vibration of the object to be detected based on the dynamic network information; and comprising.

Advantages of the Invention

[0011] According to the present disclosure, it is possible to provide a sign detection device and a sign detection method capable of detecting a sign of a sudden change in vibration sufficiently prior to the occurrence of the sudden change in vibration.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Best Mode for Carrying Out the Invention

[0013] Hereinafter, several embodiments 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 invention thereto, but are merely illustrative examples.

[0014] (Precursor Detection Device) Hereinafter, a precursor detection device 300 according to an embodiment will be described. FIG. 1 is a block diagram showing the configuration of the precursor detection device 300 according to an embodiment.

[0015] As shown in FIG. 1, the precursor detection device 300 includes a plurality of sensors 200 and an arithmetic processing device 100 configured to execute arithmetic processing for detecting precursors of sudden vibrations. The sensor 200 is a sensor configured to measure a physical quantity in a detection target object.

[0016] The plurality of sensors 200 are respectively arranged at a plurality of positions in the detection target object and measure physical quantities at each position. The physical quantity measured by the sensor 200 is, for example, any one or more of pressure, strain, acceleration, velocity, and displacement. Note that the physical quantity measured by the sensor 200 is not limited to these physical quantities. The physical quantity measured by the sensor 200 may be a physical quantity having a high correlation with the occurrence of combustion vibration.

[0017] The arithmetic processing device 100 is, for example, a computer including a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), etc. In the arithmetic processing device 100, the processor (CPU) executes a program stored in the memory (RAM or ROM) to realize various functions described later.

[0018] The functional configuration of the arithmetic processing unit 100 will be described below. As shown in FIG. 1, the arithmetic processing unit 100 functions as a data acquisition unit 110, an arithmetic unit 120, a detection unit 130, and an output unit 140.

[0019] The data acquisition unit 110 is configured to acquire time-series variation data of physical quantities from a plurality of sensors 200. The time-series variation data is measurement data sampled at a plurality of timings (for example, 100 or more timings) in a unit time in the immediate past (for example, 1 second).

[0020] The arithmetic unit 120 is configured to calculate dynamic network information based on the time-series variation data of the physical quantities acquired by the data acquisition unit 110. The dynamic network information represents the time change of a complex network structure including parameters indicating correlation. The parameter indicating correlation is a parameter indicating the correlation between physical quantities at any two positions among the plurality of positions where the sensors 200 are arranged.

[0021] The detection unit 130 is configured to detect a sign of sudden change vibration of the object to be detected based on the dynamic network information calculated by the arithmetic unit 120. The detection result by the detection unit 130 is output from the output unit 140 in an arbitrary manner. The output mode from the output unit 140 may be, for example, image data indicating the vibration state of the object to be detected, or audio data (for example, audio for notifying the sign of sudden change vibration) to an audio output device such as a speaker.

[0022] Further, the output unit 140 may be configured to output a predetermined signal when the detection unit 130 detects a sign of sudden change vibration. The predetermined signal is a signal effective for avoiding sudden change vibration, such as a stop signal for stopping the operation of the object to be detected, an output control signal for reducing the output of the object to be detected, or a notification signal for notifying the user that it is a sign of sudden change vibration.

[0023] The output unit 140 may be configured to output information related to maintenance estimated from the parameters indicating the correlation. The information related to maintenance is, for example, information on parts to be replaced, recommended replacement timing, the presence or absence of a malfunction, etc. Such image data, predetermined signals, and information related to maintenance are generated based on, for example, the calculation results of the calculation unit 120 and the detection results of the detection unit 130.

[0024] (Example of detection target and sensor placement) An example of the arrangement of the detection object and the sensor 200 according to one embodiment will be described below. Fig. 2 is a schematic diagram showing an example of the arrangement of the sensor 200 of the sign detection device 300 according to one embodiment. This figure shows a cross section along a direction perpendicular to the turbine shaft of the gas turbine 20. Fig. 3 is a schematic cross-sectional view showing an example of the arrangement of the sensor 200 of the sign detection device 300 according to one embodiment. This figure shows a cross section along the turbine shaft of the gas turbine 20.

[0025] In one embodiment, the detection target of the sign detection device 300 may be, for example, the gas turbine 20 shown in Fig. 2 and Fig. 3. Note that the detection target may not be the gas turbine 20, but may be, for example, a machine such as a steam turbine, an engine, a boiler, an aircraft, or a compressor.

[0026] As shown in Fig. 2 and Fig. 3, the gas turbine 20 includes a compressor 7, a combustor 8, a stator vane 4, and a rotor blade 6. As shown in Fig. 2, the combustor 8 includes eight can-type combustors. In Fig. 2, the eight can-type combustors are numbered #1 to #8 according to their arrangement positions. As shown in Fig. 3, each of the eight can-type combustors has a fuel nozzle 9, an inner cylinder 2, and a transition piece 3. The sensor 200 is a pressure sensor for measuring the pressure inside the combustor 8. The sensor 200 is disposed in each of the eight transition pieces 3.

[0027] Note that in this example, the sensor 200 is arranged in the tail pipe 3 of the combustor 8 of the gas turbine 20. However, the arrangement of the sensor 200 is not limited to such an example. The sensor 200 may be arranged at a position where the vibration mode can be observed, and depending on the type of the object to be detected, it may be arranged on a compressor, a blade, a bearing, or the like.

[0028] (Specific examples of dynamic network information) Hereinafter, specific examples of the dynamic network information that is the operation target of the operation unit 120 will be described.

[0029] First, the correlation relationship of the physical quantities corresponding to the plurality of sensor 200 positions acquired by the data acquisition unit 110 will be described. According to the inventor of the present application, the correlation relationship between physical quantities can be interpreted as an undirected weighted complex network having each of the plurality of positions as a node. For example, the correlation relationship of physical quantities can be represented as an adjacency matrix A (A is represented by bold type indicating a vector. The same applies hereinafter).

[0030] As shown by the following formula (1), the adjacency matrix A is defined as an n×n square matrix. In the adjacency matrix A, any matrix element w ij indicates the correlation between the i-th physical quantity and the j-th physical quantity. n corresponds to the number of physical quantities (that is, the number of sensors 200). TIFF0007691699000001.tif20170

[0031] In the adjacency matrix A, the diagonal component w 11、 w 22、··· w nn has a value of zero, and the matrix elements other than the diagonal components are values indicating the magnitude of the correlation coefficient (the matrix element in the i-th row and j-th column is the absolute value of the correlation coefficient C ij ). That is, the matrix element indicating the relationship between the physical quantities at different positions is the absolute value of the correlation coefficient C ij , and the matrix element indicating the relationship between the physical quantities at the same position is zero. Also, basically the same correlation coefficient is obtained even when the row and column numbers are exchanged. For example, w 24 and w 42 have the same value.

[0032] In the arrangement example of the sensor 200 shown in FIG. 2, since it has eight nodes numbered #1 to #8, the adjacency matrix A becomes an 8×8 square matrix. For example, the correlation between the physical quantities measured in the can-shaped combustor #2 and the can-shaped combustor #4 is the matrix element w 24 That is, the correlation coefficient C 24 is the absolute value at.

[0033] The parameter indicating the correlation may be the correlation coefficient C ij indicating the correlation of the fluctuations of the physical quantities at each position. The correlation coefficient C ij is represented by, for example, the following formula (2). Note that formula (2) shows an example when the physical quantity is pressure, but the physical quantity may be a physical quantity other than pressure. TIFF0007691699000002.tif23170

[0034] Here, N is the number of samplings per unit time (for example, 100 or more) (for example, 1 second). p i (t) represents the instantaneous value of the pressure at the i-th position, and p j (t) represents the instantaneous value of the pressure at the j-th position. P i is the time average value of p i (t) over the unit time, and P j is the time average value of p j (t) over the unit time. Note that the instantaneous value or the time average value of the pressure fluctuation amount may be applied instead of the instantaneous value or the time average value of the pressure.

[0035] The correlation coefficient C ij becomes a value close to 1 or -1 when there is a correlation between the physical quantities at two positions, and becomes a value close to 0 when there is no correlation. Also, the absolute value of the correlation coefficient C ij is a value within the range of 0 or more and 1 or less. Therefore, the strength of the correlation can be determined from the absolute value of the correlation coefficient C ij . Note that the correlation coefficient C ij is not limited to the calculated value shown in formula (2), and can be appropriately changed within a range that does not impair the essential meaning.

[0036] Here, FIG. 4 is a diagram showing an adjacency matrix A based on the physical quantities acquired by the data acquisition unit 110 at each time. At time t 1 ~t N If the adjacency matrix A based on the physical quantities acquired by the data acquisition unit 110 at time t is defined as A T then the adjacency matrices A corresponding to each of the times T - 2, T - 1, T, T + 1, T + 2 adjacent to time T are A T-2 A T-1 A T A T+1 A T+2 respectively. In addition, in FIG. 4, the interval between adjacent times (for example, time T and time T + 1) corresponds to the sampling time t N respectively.

[0037] The calculation unit 120 converts the adjacency matrix A at each time into a vector X. FIG. 5 is a diagram showing an example of the conversion calculation from the adjacency matrix A to the vector X. The vector X is a vector whose components are the matrix elements indicating the relationship between physical quantities at different positions among the matrix elements of the adjacency matrix A, and its dimensionality is represented by the following equation (3). As described above, the correlation coefficient C ij contained in the adjacency matrix A has an absolute value as a component and has a dimension represented by the following equation (3) using the number of nodes n of the network. As described above, the adjacency matrix A has diagonal components w 11、 w 22、··· w nn whose values are zero, and the matrix elements other than the diagonal components basically have the same correlation coefficient even when the row and column numbers are exchanged. Therefore, as shown by the gradation in FIG. 5, the vector X can be obtained by arranging the matrix elements on one side of the diagonal components w 11、 w 22、··· w nn as components. TIFF0007691699000003.tif9170

[0038] Such conversion from the adjacency matrix A to the vector X is performed for the adjacency matrices corresponding to each time ···, A T-2 A T-1 A T A T+1 A T+2are respectively performed on ···. Hereinafter, for the adjacency matrix ··· A T-2 A T-1 A T A T+1 A T+2 the vectors X corresponding to ··· are appropriately referred to as ··· X T-2 X T-1 X T X T+1 X T+2 and so on.

[0039] Subsequently, the arithmetic unit 120 constructs a high-dimensional vector G represented by the following equation (4) by combining a plurality of adjacent vectors X. k is the number of combined vectors X and represents the past score referred to in the dynamic network information. TIFF0007691699000004.tif9170

[0040] Such a high-dimensional vector G is sequentially calculated as time T progresses. FIG. 6 is a diagram showing the high-dimensional vector G calculated as time T progresses. In FIG. 6, for ease of understanding, the case where the combination number is k = 2 is illustrated, and the high-dimensional vector G calculated at time T T is represented by [X T-2 X T-1 X T , and the high-dimensional vector G calculated at time T + 1 T+1 is represented by [X T-1 X T X T+1 , and the high-dimensional vector G calculated at time T + 2 T+2 is represented by [X T X T+1 X T+2 .

[0041] (Specific example of detecting a sign of sudden vibration) The detection unit 130 detects a sign of sudden vibration based on the calculation result of the above-described arithmetic unit 120. Hereinafter, a specific detection method will be described.

[0042] First, the detection unit 130 pre-sets criteria for determining signs of mutation vibration. Such setting of criteria is performed by learning using sample data. Here, FIG. 7 is an example of sample data showing the time change of link strength representing the strength per link of the network calculated based on the physical quantity acquired from the sensor 200. In this example, the link strength changes with a relatively small value during normal operation when no abnormality occurs in the detection target, but shows a tendency to increase rapidly near the time t = 0 when mutation vibration occurs. This indicates that the link strength is effective as a parameter suggesting the possibility of mutation vibration occurring.

[0043] Therefore, the inventor of the present application sets a plurality of ranks for classifying the state of the detection target step by step based on the high-dimensional vector G. In the present embodiment, ranks 1 to 5 are set based on the high-dimensional vector G, and classification criteria are set such that the higher the number, the higher the possibility of mutation vibration occurring. FIG. 7 shows the result of classifying which rank 1 to 5 each data point included in the time-changing high-dimensional vector G belongs to.

[0044] The detection unit 130 identifies clusters for each rank by expanding each data point included in the high-dimensional vector G classified into each rank in the high-dimensional space. FIG. 8 is a diagram showing the data distribution obtained by expanding the sample data of FIG. 7 in the high-dimensional space. In FIG. 8, for the sake of convenience, it is shown as a two-dimensional plane as if the high-dimensional space is shown on the paper surface).

[0045] In FIG. 8, each data point corresponding to the high-dimensional vector G corresponding to typically three ranks 1 to 3 is shown by a different symbol, and the data points corresponding to the high-dimensional vector G belonging to the same rank are grouped within a predetermined range to form clusters corresponding to each rank respectively (Note that ranks 4 to 5 are omitted from the illustration for easier understanding, but the same applies). The detection unit 130 prepares in advance, as criteria, the data points forming such clusters corresponding to each rank.

[0046] Furthermore, such a determination criterion is effective in increasing the number of data points belonging to each cluster and improving the determination accuracy by increasing the number of sample data.

[0047] Subsequently, the detection unit 130 detects a sign of sudden vibration based on the high-dimensional vector G calculated by the calculation unit 120 using the determination criterion set as described above. Such detection of a sign of sudden vibration is performed by identifying to which of the plurality of clusters defined by the determination criterion the data point corresponding to the high-dimensional vector G calculated by the calculation unit 120 belongs. Such classification into clusters is called so-called clustering. For example, there are the Ward method, support vector machine, etc. Here, as an example, the case of using the K-means method will be described.

[0048] In the K-means method, assuming that the number of data included in the high-dimensional vector G is m and the number of clusters set by the determination criterion is K, clustering is performed according to the following procedure. (i) Randomly assign a cluster to each data xi (i = 1, ···, m). (ii) Calculate the center Vj (j = 1, ···, K) {\\(V_{j}(j = 1,\\cdots,k)\\)} of each cluster based on the assigned data. (iii) Determine the distance between xi and each Vj, and reassign xi {\\(x_{i}\\)} to the cluster with the nearest center. (iv) When the assignment of all xi to clusters does not change in the above processing, or when the change amount is less than a preset fixed threshold value, it is determined that convergence has occurred and the processing ends. Note that when the condition in (iv) is not satisfied, Vj is recalculated from the newly assigned clusters and the above processing is repeatedly performed.

[0049] FIG. 9 is an example of the result of clustering the temporal change of the high-dimensional vector G by the detection unit 130. In this example, the rank changes over time, and it is shown that a sign of the mutation oscillation can be detected as a rank increase from a stage sufficiently before the timing when the mutation oscillation actually occurs at t = 0 (that is, these five ranks are considered to capture the combustion state, and for example, ranks 3 and 4 can be interpreted as indicating a transition region, and rank 5 as indicating a combustion oscillation region).

[0050] As described above, the detection unit 130 can suitably detect a sign of a mutation oscillation by analyzing a high-dimensional vector calculated from operation data using a determination criterion constructed by learning using teacher data and determining which cluster it belongs to.

[0051] (Sign Detection Method) Hereinafter, a specific example of the sign detection method will be described with reference to FIG. 10. FIG. 10 is a flowchart showing the procedure of the sign detection method according to an embodiment. Note that in each of the procedures described below, part or all may be executed manually by the user. The sign detection method described below can be appropriately modified for each procedure so as to correspond to the processing executed by the above-described sign detection device 300. In the following description, descriptions overlapping with those of the sign detection device 300 will be omitted.

[0052] As shown in FIG. 10, first, a plurality of sensors 200 respectively arranged at a plurality of positions in the object to be detected measure physical quantities at each position (step S1). Temporal change data of the physical quantities measured by each of the plurality of sensors 200 is acquired from the plurality of sensors 200 (step S2).

[0053] Next, dynamic network information representing the temporal change of a complex network structure including a parameter indicating the correlation between the physical quantities at any two positions among the plurality of positions is calculated from the temporal change data (step S3). Then, based on the dynamic parameter information indicating the correlation calculated in step S3, a sign of a mutation oscillation of the object to be detected is detected (step S4).

[0054] These steps S1 to S4 may be repeatedly executed periodically. Thereby, it is possible to monitor the sign of the mutation vibration. When the sign of the mutation vibration is detected, a predetermined signal (such as a stop signal or a notification signal) described above may be output. Further, the above-described image data may be output and the image may be displayed on a display device or the like.

[0055] The present disclosure is not limited to the above-described embodiments, and includes forms obtained by modifying the above-described embodiments and forms obtained by appropriately combining a plurality of embodiments.

[0056] For example, when the object to be detected is a compressor, a plurality of sensors 200 for measuring pressure may be arranged at a plurality of positions of the compressor. When the object to be detected is an axial flow compressor, a plurality of sensors 200 may be arranged in the circumferential direction at the outlet portion thereof. When the object to be detected is a centrifugal compressor, a plurality of sensors 200 may be arranged in the annular direction. When detecting the sign of the mutation vibration of the blade vibration, a plurality of sensors 200 may be arranged at the root of the blade.

[0057] When detecting the sign of the mutation vibration of the shaft vibration, sensors 200 may be arranged at respective bearing positions that are different.

[0058] When the object to be detected is a steam turbine, a strain gauge may be used as the sensor 200. For example, a plurality of sensors 200 may be arranged at the root of the blades of the steam turbine arranged along the circumferential direction in the same stage.

[0059] When the object to be detected is a rocket engine, there may be only one combustor. However, even in this case, a plurality of sensors 200 may be arranged in the circumferential direction of the outlet portion of the combustor so that the precursor detection device 300 detects a precursor of a sudden change in vibration. When the object to be detected is an aircraft, the method of detecting a precursor of a sudden change in vibration by the precursor detection device 300 may be applied to its engine or its wing. By arranging a plurality of sensors 200 along the circumferential direction of the cross section at the position where combustion vibration occurs in this way, it is possible to detect precursors of sudden changes in vibration of various objects to be detected.

[0060] In addition, within the scope not departing from the gist of the present disclosure, it is possible to appropriately replace the components in the above-described embodiments with well-known components, and the above-described embodiments may also be appropriately combined.

[0061] The content described in each of the above embodiments is understood as follows, for example.

[0062] (1) The precursor detection device (300) according to one aspect A plurality of sensors (200) respectively arranged at a plurality of positions in the object to be detected (for example, the gas turbine 20) and configured to measure physical quantities at each position, A data acquisition unit (110) for acquiring time-series variation data of the physical quantity from the plurality of sensors, An arithmetic unit (120) for calculating dynamic network information representing a time change of a complex network structure including a parameter indicating a correlation between the physical quantities at any two positions among the plurality of positions based on the time-series variation data, A detection unit (130) for detecting a precursor of a sudden change in vibration of the object to be detected based on the dynamic network information, and includes.

[0063] According to the aspect (1) above, based on the dynamic network information representing the temporal change of the complex network structure including the parameter indicating the correlation of the physical quantity between two positions, a sign of the abrupt change of the object to be detected is detected. Therefore, the sign can be detected sufficiently prior to the occurrence of the abrupt vibration.

[0064] (2) In another aspect, in the aspect (1) above, The temporal change of the dynamic network information is calculated as a high-dimensional vector configured based on a plurality of the complex network structures corresponding to different consecutive times.

[0065] According to the aspect (2) above, the temporal change of the dynamic network information used for detecting the sign is calculated as a high-dimensional vector configured based on the dynamic network information corresponding to different consecutive times. By considering the dynamic network information corresponding to different consecutive times in this way, it is possible to suitably detect the sign stage of the abrupt vibration that is difficult to detect with the information corresponding to a single time.

[0066] (3) In another aspect, in the aspect (2) above, The detection unit uses a classification criterion that defines a plurality of clusters preset based on the possibility of occurrence of the abrupt vibration, and based on which cluster the high-dimensional vector is classified into among the plurality of clusters, detects a sign of the abrupt vibration of the object to be detected.

[0067] According to the aspect (3) above, a classification criterion in which a plurality of clusters are defined based on the possibility of occurrence of the abrupt vibration is prepared in advance. The detection unit evaluates the possibility of occurrence of the abrupt vibration based on which cluster the high-dimensional vector calculated from the detected physical quantity is classified into according to such a classification criterion, and enables detection at the sign stage.

[0068] (4) In another aspect, in the aspect (3) above, The classification criteria are set by learning using at least one sample data corresponding to the case where the mutation vibration occurs in the detection target object.

[0069] According to the aspect of (4) above, it is possible to perform accurate sign detection using the classification criteria set with the classification result based on the sample data corresponding to the case where the mutation vibration occurs in the detection target object as the teacher data.

[0070] (5) A sign detection method according to one aspect is a step in which sensors respectively arranged at a plurality of positions in the detection target object measure physical quantities at each position, a step of acquiring time-series variation data of the physical quantities from the plurality of sensors, a step of calculating dynamic network information representing a time change of a complex network structure including a parameter indicating a correlation between the physical quantities at any two positions among the plurality of positions based on the time-series variation data, a step of detecting a sign of mutation vibration of the detection target object based on the dynamic network information, and includes.

[0071] According to the aspect of (5) above, a sign of mutation variation of the detection target object is detected based on the dynamic network information representing the time change of the complex network structure including the parameter indicating the correlation between the physical quantities between two positions. Therefore, the sign can be detected sufficiently prior to the occurrence of the mutation vibration.

Explanation of Signs

[0072] 2 Inner cylinder 3 Tail cylinder 4 Static blade 6 Rotating blade 7 Compressor 8 Combustor 9 Fuel nozzle 20 Gas turbine 100 Arithmetic processing unit 110 Data acquisition unit 120 Arithmetic unit 130 Detection unit 140 Output section 200 Sensors 300 Predictive Detection Device

Claims

1. A plurality of sensors respectively arranged at a plurality of positions on a detection object and configured to measure physical quantities at each position; A data acquisition unit for acquiring time-series variation data of the physical quantity from the plurality of sensors; An arithmetic unit for calculating dynamic network information representing a time change of a complex network structure including a parameter indicating a correlation between the physical quantities at any two of the plurality of positions based on the time-series variation data; A detection unit for detecting a sign of abrupt vibration of the detection object based on the dynamic network information; A sign detection device comprising the above.

2. The sign detection device according to claim 1, wherein the time change of the dynamic network information is calculated as a high-dimensional vector configured based on a plurality of the complex network structures corresponding to different consecutive times.

3. The detection unit uses a classification criterion that defines a plurality of clusters preset based on the possibility of occurrence of the abrupt vibration, and based on which cluster among the plurality of clusters the high-dimensional vector is classified into, detects a sign of abrupt vibration of the detection object. The sign detection device according to claim 2.

4. The sign detection device according to claim 3, wherein the classification criterion is set by learning using at least one sample data corresponding to the case where the abrupt vibration occurs in the detection object.

5. A step of measuring physical quantities at each position by sensors respectively arranged at a plurality of positions on a detection object; A step of acquiring time-series variation data of the physical quantity from the plurality of sensors; A step of calculating dynamic network information representing a time change of a complex network structure including a parameter indicating a correlation between the physical quantities at any two of the plurality of positions based on the time-series variation data; A step of detecting a sign of abrupt vibration of the detection object based on the dynamic network information; A sign detection method comprising the above.

Images