Monitoring device, monitoring method and program

Abstract

To provide a monitoring device which enhances the monitoring precision by an MD value, and which prevents an abnormal determination from being excessively made.SOLUTION: A monitoring device includes: means for calculating an MD value based on the process value of a rotational machine; means for calculating a gap value representing a ratio between a ratio of a difference between the actual measured value of the process value and an estimated value relative to a difference between a threshold for determining an abnormality in the process value and the estimated value of the process value; and means for determining, based on the MD value and on the gap value, the operation state of the rotational machine.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a monitoring device, a monitoring method, and a program. [Background technology]

[0002] The Mahalanobis-Taguchi method (MT method) is known as a method for detecting anomalies in equipment and plants. In the MT method, the Mahalanobis distance (MD value) is calculated from the time series data of multiple variables, and an anomaly is determined when the MD value exceeds a threshold. However, the measurement values ​​of various sensors used to calculate the MD value may contain noise, missing values, or fluctuations, and when the MD value is calculated using such measurement data, the MD value obtained may deviate from the normal range. For example, in the case of rotating machinery such as turbines and compressors, the bearing temperature may rise with fluctuations in the rotation speed, and even if the machinery is not abnormal, the MD value may exceed the threshold and an alarm may be falsely issued.

[0003] Patent Document 1 discloses a method for detecting signs of abnormality not only by comparing a measurement value obtained by a sensor or the like with a threshold value but also by using other measurement values ​​that affect the measurement value. Patent Document 2 discloses a bearing abnormality prediction method for determining that when vibration during rotation of a rolling bearing falls below a threshold value, it is a sign of excessive temperature rise in the rolling bearing. Patent Documents 1 and 2 do not disclose an abnormality detection method that can complement the MT method. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-201683 A [Patent Document 2] Patent Publication No. 2021-018106 Summary of the Invention [Problem to be solved by the invention]

[0005] There is a need for a method to improve the monitoring accuracy of MT method and prevent false positives.

[0006] The present disclosure provides a monitoring device, a monitoring method, and a program that can solve the above problems. [Means for solving the problem]

[0007] The monitoring device disclosed herein has a means for calculating a Mahalanobis distance based on a process value of a rotating machine, a means for calculating a gap value indicating the ratio of the difference between an actual value of the process value and the estimated value of the process value to the difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value, and a means for determining the operating state of the rotating machine based on the Mahalanobis distance and the gap value.

[0008] The monitoring method disclosed herein includes the steps of calculating a Mahalanobis distance based on a process value of a rotating machine, calculating a gap value indicating the ratio of the difference between an actual value of the process value and the estimated value of the process value to the difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value, and determining the operating state of the rotating machine based on the Mahalanobis distance and the gap value.

[0009] The program disclosed herein causes a computer to function as: a means for calculating a Mahalanobis distance based on a process value of a rotating machine; a means for calculating a gap value indicating the ratio of the difference between an actual value of the process value and the estimated value of the process value to the difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value; and a means for determining the operating state of the rotating machine based on the Mahalanobis distance and the gap value. Effect of the Invention

[0010] According to the above-described monitoring device, monitoring method, and program, it is possible to improve the monitoring accuracy by the MT method and prevent erroneous detection. [Brief description of the drawings]

[0011] [Figure 1] 1 is a block diagram showing an example of a monitoring device according to an embodiment; [Diagram 2] FIG. 11 is a diagram showing an example of a transition of an MD value according to the embodiment. [Diagram 3] FIG. 4 is a diagram showing an example of a transition of the rotation speed according to the embodiment. [Figure 4] FIG. 2 is a diagram illustrating an example of monitoring target items according to the embodiment. [Diagram 5] FIG. 4 is a diagram illustrating a gap value according to the embodiment. [Figure 6] 5 is a flowchart illustrating an example of an abnormality determination process according to the embodiment. [Figure 7] FIG. 2 is a schematic diagram illustrating an example of a hardware configuration of a monitoring device according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] <Embodiment> A monitoring device 10 of the present disclosure will now be described with reference to FIGS. FIG. 1 is a block diagram illustrating an example of a monitoring device according to an embodiment. The monitoring device 10 monitors the operating state of a rotating machine such as a compressor, a turbine, an engine, a pump, a generator, etc. As an example, in this embodiment, a case where the rotating machine is a compressor will be described. The compressor 1 and the monitoring device 10 are connected to each other so as to be able to communicate with each other via a wired or wireless network, a signal line, etc. A plurality of various sensors 2 for measuring process values ​​of the compressor 1, such as a thermometer, a pressure gauge, a vibration meter, a tachometer, an ammeter, and a voltmeter, are provided at various locations on the compressor 1, and the measurement values ​​measured by each sensor 2 are transmitted to the monitoring device 10. The monitoring device 10 includes a process value acquisition unit 11, an MD value calculation unit 12, a gap value calculation unit 13, a determination unit 14, an analysis unit 15, an output unit 16, and a storage unit 17.

[0013] The process value acquiring unit 11 acquires the measurement values ​​measured by each sensor 2 of the compressor 1 or values ​​calculated from the measurement values. The measurement values ​​or values ​​calculated from the measurement values ​​are called process values ​​or actual measured values ​​of the process values. The process values ​​may be calculated by another device or may be calculated by the process value acquiring unit 11. The process value acquiring unit 11 acquires the process values ​​and stores them in the memory unit 17.

[0014] Before the start of monitoring, the MD value calculation unit 12 creates a unit space for calculating a Mahalanobis distance (MD value) using the process value of the compressor 1 in a normal state acquired by the process value acquisition unit 11. After the start of monitoring, the MD value calculation unit 12 calculates an MD value used to evaluate the degree of abnormality of the compressor 1 based on the created unit space and the process value acquired by the process value acquisition unit 11. The MD value calculation unit 12 extracts a process value to be evaluated that is the same type as the process value used to create the unit space from the process values ​​acquired by the process value acquisition unit 11, and calculates a Mahalanobis distance (MD value) between the extracted process value and the unit space. The MD value is a value indicating how far the compressor 1 is deviated from the normal state. Since the method of calculating the Mahalanobis distance is known, a description thereof will be omitted in this specification. The MD value calculation unit 12 outputs the calculated MD value to the determination unit 14.

[0015] FIG. 2 shows a graph that shows a schematic transition of the MD value. The vertical axis of the graph in FIG. 2 is the MD value, the horizontal axis is time, and Th1 is a threshold value. Graph 21 shows the MD value in a time series calculated by the MD value calculation unit 12. In a general abnormality monitoring by the Mahalanobis method, when the MD value exceeds the threshold value Th1 at time t1, it is determined that an abnormality or a sign of an abnormality has been detected. However, in a rotating machine such as the compressor 1, such an abnormality detection may be a false determination. FIG. 3 shows a trend of the rotation speed of the compressor 1. The vertical axis of the graph in FIG. 3 is the rotation speed of the compressor 1, the horizontal axis is time, and graph 31 shows the time series rotation speed acquired by the process value acquisition unit 11. If the same position of the horizontal axis in FIG. 2 and FIG. 3 represents the same time, it can be seen that the MD value shows a transition such that it exceeds the threshold value Th1 as the rotation speed increases. If the bearing temperature rises with the increase in the rotation speed, the MD value exceeds the threshold value and an alarm is issued. However, in reality, the operating state of the compressor 1 may be normal. In the past, when the MD value exceeded the threshold value for such an event, a supervisor would first check whether a process value such as the rotation speed had fluctuated, and determine whether the increase in the MD value was due to an abnormality or an increase in the rotation speed. In contrast, in this embodiment, in addition to the MD value, a monitored item is set, and the value of the monitored item is estimated. The estimated value is then used as an estimated value (the predicted value, average value, and performance predicted value described below are collectively referred to as an estimated value or an estimated value of a process value), and a gap value (described later) is calculated using the difference between the estimated value and the actual measured value. When the MD value exceeds the threshold value and the gap value also exceeds the threshold value, it is determined that an abnormality or a sign of an abnormality has been detected.

[0016] The gap value calculation unit 13 calculates the gap values ​​of the monitored items. An example of the monitored items is shown in FIG. 4. Process values ​​such as the journal bearing temperature in FIG. 4 indicate the monitored items, and predicted values, average values, etc. indicate the calculation methods for the estimated values ​​of each process value used to calculate the gap value. The gap value calculation unit 13 calculates the gap values ​​for each monitored item in FIG. 4. For example, the predicted values ​​of each monitored item are used to calculate the gap values ​​of the journal bearing temperature, thrust bearing temperature, and shaft position (axial displacement of the rotating shaft). The gap value calculation unit 13 calculates the gap values ​​of the journal bearing temperature, etc. by the following formula (1). Gap value = (actual value - predicted value) ÷ (conventional threshold - predicted value) (1)

[0017] The predicted value is calculated, for example, by a prediction formula derived from a theoretical formula for calculating the journal bearing temperature or the like, or a prediction model constructed using machine learning or the like. The monitored items include process values ​​that are affected by fluctuations in the rotation speed of the compressor 1, and examples thereof include the journal bearing temperature and the thrust bearing temperature. For example, in the case of the journal bearing temperature or the thrust bearing temperature, the prediction formula for calculating the predicted value may use parameters such as the rotation speed of the compressor 1 and the temperature of the lubricating oil. These parameters are factors that have a strong correlation with the bearing temperature among the parameters constituting the theoretical formula. The actual measured value is, for example, a process value measured by a sensor 2 that measures each bearing temperature. In the case of the journal bearing temperature, the conventional threshold value is a threshold value set to detect an abnormality in the journal bearing temperature.

[0018] For monitored items for which it is difficult to calculate predicted values, the gap value is calculated using the average value at the start of monitoring. Average values ​​are used to calculate the gap values ​​of shaft vibration, steam pressure / temperature, compressor suction temperature, cooling water temperature, gas seal differential pressure, lubricant pressure, and lubricant temperature in Figure 4. Gap value calculation unit 13 calculates the gap value using the following formula (2). Gap value = (actual value - average value) ÷ (conventional threshold value - average value) (2)

[0019] For example, average values ​​of shaft vibration, steam pressure, etc. collected during past operation may be calculated in advance, and the gap value calculation unit 13 may calculate the gap value of each monitored item using the average value of each process value, the actual measured value (process value) acquired by the process value acquisition unit 11 during monitoring, a conventional threshold value, and formula (2). Also, the gap value calculation unit 13 may calculate the average value of each process value acquired by the process value acquisition unit 11 during a predetermined time period from the start of monitoring the compressor 1, such as shaft vibration, steam pressure, steam temperature, compressor suction temperature, cooling water temperature, gas seal differential pressure, lubricating oil pressure, lubricating oil temperature, etc., and calculate the gap value using the calculated average values ​​and formula (2) after the predetermined time period has elapsed.

[0020] In addition, the gap value calculation unit 13 calculates a predicted value (predicted performance value) of the discharge pressure of the compressor 1 from a predetermined performance curve (a curve showing the relationship between the discharge pressure of the compressor 1 and other parameters) or a predetermined theoretical formula, and calculates the gap value using the following formula (3). Gap value = (actual value - predicted performance value) ÷ (conventional threshold - predicted performance value) (3) Similarly, the gap value calculation unit 13 calculates the discharge temperature of the compressor 1 using a predetermined performance curve or theoretical formula, and calculates the gap value of the discharge temperature by formula (3).

[0021] The monitored items exemplified in Fig. 4 may be used to calculate the MD value. In addition to the monitored items in Fig. 4, the rotation speed, compressor flow rate, steam flow rate, compressor suction pressure, etc. may be used to calculate the MD value. These items may be divided into groups, and the MD value may be calculated for each group.

[0022] The judgment unit 14 compares the MD value calculated by the MD value calculation unit 12 and the gap value calculated by the gap value calculation unit 13 with their respective thresholds, and judges that an abnormality should be detected if the MD value exceeds the threshold Th1 and any of the gap values ​​of the monitored items exceeds a predetermined threshold. Specifically, the judgment unit 14 constantly monitors the MD value and the gap value in parallel, and issues an alarm as an abnormality sign detection or abnormality detection if both exceed the thresholds. By adding judgment based on the gap value in addition to the MD value, it is possible to avoid issuing an alarm even if the MD value exceeds the threshold Th1 with an increase in the rotation speed, as long as the gap value is within the threshold.

[0023] The threshold for the gap value can be set in multiple stages. For example, the threshold for progress observation Tha=30%, the threshold for abnormality sign determination Thb=50%, and the threshold for abnormality determination Thc=100% may be set. The purpose of the abnormality sign alarm is to investigate the cause early and take measures early before the abnormality becomes serious, and it is necessary to avoid issuing an alarm for a fluctuation amount that cannot be considered abnormal. Since it is not desirable to issue an alarm too early or too late, for example, the middle (50%) between the predicted value and the conventional threshold may be set as the threshold for abnormality sign detection Thb. For an object whose predicted value is difficult to calculate, such as shaft vibration, the middle between the average value at the beginning of operation and the conventional threshold can be set as the threshold Thb. In addition, the process value acquired until the abnormality sign detection can be used as learning data for analyzing the cause of the abnormality and for detecting the abnormality sign. Therefore, a threshold value Tha for follow-up observation may be set for a fluctuation level lower than the abnormality sign detection, and when this threshold value is exceeded, a cause analysis using FTA (Fault Tree Analysis) or the like may be started for the process value measured by the sensor 2, and the analysis result and the process value may be recorded for storage. If the fluctuation level is too small, the accuracy of the cause analysis decreases, so a value that provides sufficient analysis accuracy may be set for the threshold value Tha.

[0024] Graph 51 in the upper part of FIG. 5 shows the transition of the predicted value and the actual measured value for the monitored item (for example, journal bearing temperature), and graph 52 in the lower part of FIG. 5 shows an example of the transition of the gap value of the monitored item. The vertical axis of graph 51 is the bearing temperature, and the horizontal axis is time. The vertical axis of graph 52 is the temperature difference between the predicted value and the actual measured value, and the horizontal axis is time. For graph 51 in the upper part of FIG. 5, threshold value 51a shows the threshold value set for the journal bearing temperature (the conventional threshold value of formula (1)), and graphs 51b and 51c show the predicted value and the actual measured value of the journal bearing temperature, respectively. As shown in formula (1), the gap value (%) of the journal bearing temperature is calculated as a fraction with the difference between graphs 51b and 51c as the numerator and the difference between graphs 51a and 51b as the denominator. For graph 52 in the lower part of FIG. 5, graphs 52a and 52b show the denominator and numerator of formula (1), respectively. Also, graph 52c shows the gap value calculated by formula (1). The numerator represents the deviation of the actual measurement value from the predicted value, and the denominator represents the tolerance between the predicted value and the threshold. Normalizing by the tolerance makes it possible to evaluate the deviation of the actual measurement value from the predicted value in a relative manner. The MD value quantifies the deviation from the standard deviation of each signal, and it is difficult to explain its relationship with the actual amount of fluctuation, but by introducing the gap value (a normalized value of the distance between the actual measurement value and the conventional threshold 51a), it is possible to more specifically evaluate the state of the monitored object.

[0025] For graph 52 at the bottom of Fig. 5, when the gap value (graph 52c) exceeds threshold Tha at time t1 (the MD value exceeds threshold Th1; similarly below), the judgment unit 14 judges that follow-up observation is necessary and instructs the analysis unit 15 to perform cause analysis and record the analysis results, etc. When the gap value becomes equal to or less than threshold Tha at time t2, the judgment unit 14 judges that follow-up observation is unnecessary and instructs the analysis unit 15 to stop recording data. When the gap value exceeds threshold Tha again at time t3, the judgment unit 14 instructs the analysis unit 15 to record data.

[0026] When the gap value exceeds the threshold value Thb at time t4, the judgment unit 14 judges that a sign of an abnormality has been detected and instructs the output unit 16 to issue an alarm indicating that a sign of an abnormality has been detected and to prompt the user to start preparing for countermeasures. When the gap value exceeds the threshold value Thc at time t5, the judgment unit 14 judges that an abnormality has been detected and instructs the output unit 16 to issue an alarm indicating the detection of an abnormality.

[0027] The analysis unit 15 performs an analysis of the cause of the abnormality using FTA or the like when a sign of an abnormality is detected. The analysis unit 15 also records the various process values ​​acquired by the process value acquisition unit 11 and the analysis results using FTA or the like in the storage unit 17. The recorded data is used as learning data for analyzing the cause of an abnormality and detecting signs of an abnormality.

[0028] The output unit 16 outputs various information related to the monitoring of the compressor 1 to a display device or other devices. For example, the output unit 16 issues the determination result by the determination unit 14 or an alarm. The storage unit 17 stores various information, such as the process values ​​acquired by the process value acquisition unit 11, various thresholds, a prediction formula or a prediction model used to calculate a predicted value, and a unit space used to calculate an MD value.

[0029] (operation) Next, the flow of the abnormality determination process will be described with reference to FIG. FIG. 6 is a flowchart illustrating an example of the abnormality determination process according to the embodiment. The monitoring device 10 repeatedly performs the following process at a predetermined control period. First, the process value acquisition unit 11 acquires a process value (step S11). Next, the MD value calculation unit 12 calculates an MD value (step S12). The MD value calculation unit 12 outputs the calculated MD value to the determination unit 14. Next, the gap value calculation unit 13 calculates a gap value for each monitored item using equations (1) to (3) (step S13). The gap value calculation unit 13 outputs the calculated gap value to the determination unit 14. There is no restriction on the order in which steps S12 and S13 are performed. Steps S13 and S12 may be performed in that order, or steps S12 and S13 may be performed simultaneously in parallel.

[0030] Next, the determination unit 14 determines whether there is an abnormality or the like based on the MD value and the gap value (steps S14, S15, S17, S20). If the MD value is equal to or smaller than the threshold value Th1 (step S14; No), the process proceeds to step S23.

[0031] If the MD value is greater than the threshold value Th1 and any of the gap values ​​is greater than the threshold value Thc (step S14; Yes, step S15; Yes), the judgment unit 14 judges that an abnormality has been detected and instructs the output unit 16 to issue an alarm. The output unit 16 issues an alarm notifying the detection of the abnormality (step S16). For example, the output unit 16 may display the alarm on a display device connected to the monitoring device 10, or may turn on a lamp, sound a buzzer, or transmit an alarm of the abnormality detection to another terminal device, etc.

[0032] When the MD value is greater than the threshold Th1, all the gap values ​​are equal to or less than the threshold Thc, and any of the gap values ​​is greater than the threshold Thb (step S14; Yes, step S15; No, step S17; Yes), the determination unit 14 determines that a sign of an abnormality has been detected, and instructs the output unit 16 to issue an alarm. The output unit 16 issues an alarm to prompt the user to detect the sign of an abnormality and start preparing countermeasures (step S18). The determination unit 14 also instructs the analysis unit 15 to analyze the cause of the abnormality. The analysis unit 15 executes FTA or the like for the monitored items related to the gap values ​​that exceed the threshold Thb, and analyzes the cause of the abnormality (step S19). For example, the storage unit 17 stores a fault tree for each monitored item, which is related to the value of the monitored item indicating an abnormality, and the analysis unit 15 analyzes the cause of the gap value exceeding the threshold Thb based on this fault tree. The analysis unit 15 displays the result of the analysis of the abnormality by FTA or the like on a display device via the output unit 16. The process of step S19 may also be executed when an abnormality is detected.

[0033] If the MD value is greater than the threshold Th1, all the gap values ​​are equal to or less than the thresholds Thc and Thb, and any of the gap values ​​is greater than the threshold Tha (step S14; Yes, step S15; No, step S17; No, step S20; Yes), the determination unit 14 determines that follow-up observation is necessary and instructs the analysis unit 15 to analyze the cause of the gap value exceeding the threshold Tha and record the analysis result. The analysis unit 15 executes FTA or the like for the monitored items related to the gap value exceeding the threshold Tha, and analyzes the cause of the abnormality (step S21). The analysis unit 15 records and saves the result of the cause analysis and the measurement values ​​used in the cause analysis in the memory unit 17 (step S22).

[0034] If the MD value is greater than the threshold value Th1 and all the gap values ​​are equal to or less than the threshold value Tha (step S14; Yes, step S15; No, step S17; No, step S20; No), the judgment unit 14 judges that the operating state of the compressor 1 is not abnormal. In this case, if an alarm is being issued, the judgment unit 14 stops the alarm and proceeds to step S24 (step S23). For example, if an abnormality or a sign of an abnormality is detected, the judgment unit 14 instructs the output unit 16 to stop issuing the alarm. The output unit 16 stops issuing the alarm. For example, if it is determined that follow-up observation is necessary, the judgment unit 14 instructs the analysis unit 15 to stop the cause analysis and recording of the analysis results. The analysis unit 15 stops the cause analysis and recording.

[0035] Next, the monitoring device 10 judges whether or not to end the monitoring of the compressor 1 (step S24). For example, when a stop command is input by the user, the monitoring device 10 judges to end the monitoring (step S24; Yes), and ends the processing of the flowchart in Fig. 6. If the monitoring is not to be ended (step S24; No), the processing from step S11 is repeatedly executed.

[0036] (effect) As described above, according to this embodiment, the abnormality detection or abnormality sign detection is performed only when the MD value exceeds the threshold value Th1 and the gap value exceeds the threshold value, so that the accuracy of abnormality detection by the MT method can be improved. For example, even if only the MD value exceeds the threshold value Th1 due to an increase in the rotation speed during a normal operating state in which the gap value is equal to or less than the threshold value, erroneous alarm is avoided. In addition, by using not only the gap value but also the MD value, highly accurate abnormality detection is possible. Specifically, by confirming that the operating state of the compressor 1 deviates from the normal range by the MD value exceeding the threshold value Th1, and only when an abnormality in a specific process value is confirmed based on this as a prerequisite, the abnormality detection or abnormality sign detection is performed. As a result, even if the operating state of the compressor 1 is normal, it is possible to prevent erroneous abnormality detection even in a situation in which the prediction accuracy of the predicted value cannot be obtained for some reason and the deviation between the actual value and the predicted value (the numerator of formula (1)) becomes large.

[0037] In addition, by paying attention to the gap value of the process value that is affected by the rotation speed of the compressor 1, such as the journal bearing temperature and the thrust bearing temperature, it is possible to determine whether the increase in the MD value is merely a numerical value affected by the increase in the rotation speed (for example, MD value > threshold value Th1 and gap value of the journal bearing temperature, etc. <= threshold value Thb or threshold value Thc) or whether there is a real abnormality (for example, MD value > threshold value Th1 and gap value of the journal bearing temperature, etc. > threshold value Thb or threshold value Thc), thereby reducing the burden on the monitor.

[0038] In the above embodiment, the compressor 1 is taken as an example, but the object to be monitored is not limited to the compressor, and may be other rotating machines. The monitored items exemplified in FIG. 4 are merely an example, and are not limited thereto. It is not necessary to determine the gap values ​​of all the monitored items exemplified in FIG. 4, etc., and any required monitored items can be selected. In addition, in step S15 in FIG. 6, the determination condition is that "any gap value exceeds the threshold value Thc," but the determination condition may be that the gap values ​​of multiple monitored items exceed the threshold value Thc. In addition, the threshold values ​​Tha=30%, Thb=50%, and Thc=100% are taken as examples, and other values ​​may be applied to each threshold value. In addition, in the embodiment, the threshold value for the gap value is set to three stages, but only one or two stages may be set, or four or more stages may be set. In addition, a threshold value having a different value may be set for each monitored item.

[0039] 7 is a schematic block diagram showing the hardware configuration of a controller according to an embodiment. A computer 900 includes a CPU 901, a main memory device 902, an auxiliary memory device 903, an input / output interface 904, and a communication interface 905. The above-mentioned monitoring device 10 is implemented in a computer 900. Each of the above-mentioned functions is stored in the auxiliary storage device 903 in the form of a program. The CPU 901 reads the program from the auxiliary storage device 903, loads it in the main storage device 902, and executes the above-mentioned processing in accordance with the program. The CPU 901 also reserves a storage area in the main storage device 902 in accordance with the program. The CPU 901 also reserves a storage area in the auxiliary storage device 903 for storing data being processed in accordance with the program. The computer 900 may include a GPU (Graphic Processing Unit), a microprocessor, etc. instead of / in addition to the CPU 901.

[0040] In another embodiment, the computer 900 may include a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) instead of or in addition to the above configuration. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). In this case, some or all of the functions realized by the CPU 901 may be realized by the integrated circuit.

[0041] Examples of the auxiliary storage device 903 include a hard disk drive (HDD), a solid state drive (SSD), a magnetic disk, a magneto-optical disk, a compact disc read only memory (CD-ROM), a digital versatile disc read only memory (DVD-ROM), a semiconductor memory, etc. The auxiliary storage device 903 may be an internal medium directly connected to the bus of the computer 900, or an external medium connected to the computer 900 via the input / output interface 904 or a communication line.

[0042] A program for realizing all or part of the functions of the monitoring device 10 may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer system and executed to perform processing by each functional unit. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. In addition, if a WWW system is used, the term "computer system" also includes a homepage providing environment (or display environment). In addition, the term "computer-readable recording medium" refers to a portable medium such as a CD, DVD, or USB, or a storage device such as a hard disk built into a computer system. In addition, if the program is distributed to the computer 900 via a communication line, the computer 900 that receives the program may deploy the program in the main storage device 902 and execute the above processing. In addition, the above program may be for realizing part of the above-mentioned functions, or may be capable of realizing the above-mentioned functions in combination with a program already recorded in the computer system.

[0043] As described above, several embodiments according to the present disclosure have been described, but all of these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents as described in the claims, as well as in the scope and gist of the invention.

[0044] <Additional Notes> The monitoring device, the monitoring method, and the program described in each embodiment can be understood, for example, as follows.

[0045] (1) A monitoring device according to a first aspect has a means for calculating a Mahalanobis distance based on a process value of a rotating machine, a means for calculating a gap value indicating the ratio of the difference between an actual value of the process value and the estimated value of the process value to the difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value, and a means for determining the operating state of the rotating machine based on the Mahalanobis distance and the gap value. This makes it possible to determine the operating state not only from the Mahalanobis distance but also from the gap value, improving the monitoring accuracy of the MT method and avoiding false positives such as excessive abnormality detection.

[0046] (2) A monitoring device according to a second aspect is the monitoring device of (1), wherein when the Mahalanobis distance exceeds a first threshold value, the judging means evaluates the degree of abnormality regarding the operating condition depending on the magnitude of the gap value. Various operating conditions can be evaluated depending on the size of the gap value.

[0047] (3) A monitoring device according to a third aspect is the monitoring device of (1) to (2), wherein the judging means judges that a sign of an abnormality has been detected when the Mahalanobis distance exceeds the first threshold value and the gap value exceeds a second threshold value. This makes it possible to detect abnormal signs while avoiding excessive abnormal sign detection.

[0048] (4) A monitoring device according to a fourth aspect is a monitoring device of (1) to (3), wherein the judging means judges that the driving condition requires follow-up observation when the Mahalanobis distance exceeds a first threshold value and the gap value exceeds a third threshold value smaller than the second threshold value. This makes it possible to detect operating conditions that may not be signs of an abnormality but may potentially lead to an abnormality in the future.

[0049] (5) A monitoring device according to a fifth aspect is the monitoring device of (4), further comprising: a means for performing a cause analysis of an abnormality based on the actual measured value when the determining means determines that the operating state requires follow-up observation, and for recording the results of the cause analysis. This makes it possible to accumulate data that is useful for detecting signs of abnormalities.

[0050] (6) A monitoring device according to a sixth aspect is a monitoring device according to any one of (1) to (5), wherein the judging means judges that an abnormality has been detected when the MD value exceeds the first threshold value and the gap value exceeds a fourth threshold value that is greater than the second threshold value. This makes it possible to detect abnormalities while avoiding excessive abnormality detection.

[0051] (7) A monitoring device according to a seventh aspect is a monitoring device of (1) to (6), wherein the means for calculating the gap value calculates the gap value using the process value that is affected by fluctuations in the rotation speed of the rotating machine. This makes it possible to determine whether an abnormality exists while checking whether an increase in the MD value is due to a fluctuation in the rotation speed of the rotating machine.

[0052] (8) A monitoring device according to an eighth aspect is a monitoring device of (1) to (7), wherein the means for calculating the gap value calculates an estimate of the process value using a predetermined prediction formula having the rotation speed of the rotating machine as a parameter. This makes it possible to calculate a predicted value of the process value that is affected by the rotation speed.

[0053] (9) A monitoring device according to a ninth aspect is a monitoring device of (1) to (8), wherein the means for calculating the gap value calculates an average value of the actual measurement values ​​measured within a predetermined time period from when monitoring of the rotating machine begins as an estimate of the process value. This makes it possible to calculate estimates for process values ​​that are difficult to predict.

[0054] (10) A monitoring method according to a tenth aspect includes the steps of: calculating a Mahalanobis distance based on a process value of a rotating machine; calculating a gap value indicating the ratio of the difference between an actual value of the process value and the estimated value of the process value to the difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value; and determining the operating state of the rotating machine based on the Mahalanobis distance and the gap value.

[0055] (11) A program according to an eleventh aspect causes a computer to function as: means for calculating a Mahalanobis distance based on a process value of a rotating machine; means for calculating a gap value indicating the ratio of the difference between an actual value of the process value and the estimated value of the process value to the difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value; and means for determining the operating state of the rotating machine based on the Mahalanobis distance and the gap value. [Explanation of symbols]

[0056] 1. Compressor 2. Sensor 10...Monitoring device 11 Process value acquisition section 12 MD value calculation section 13. Gap value calculation section 14... Judgment section 15...Analysis Department 16 Output section 17...Storage section 900...Computer 901··CPU 902...Main memory 903...Auxiliary storage device 904 Input / Output Interface 905 Communication Interface

Claims

1. A means for calculating a Mahalanobis distance based on a process value of a rotating machine; a means for calculating a gap value indicating a ratio of a difference between an actual measurement value of the process value and the estimated value of the process value to a difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value; a means for determining an operating state of the rotary machine based on the Mahalanobis distance and the gap value; A monitoring device having the above configuration.

2. When the Mahalanobis distance exceeds a first threshold, the determining means evaluating the degree of abnormality regarding the operating state according to the magnitude of the gap value; The monitoring device according to claim 1.

3. the determining means determines that a sign of an abnormality has been detected when the Mahalanobis distance exceeds a first threshold value and the gap value exceeds a second threshold value. The monitoring device according to claim 1.

4. the determining means determines that the driving state requires follow-up observation when the Mahalanobis distance exceeds the first threshold value and the gap value exceeds a third threshold value that is smaller than the second threshold value. The monitoring device according to claim 3.

5. When the determining means determines that the operating state requires follow-up observation, A means for performing an analysis of the cause of an abnormality based on the actual measurement value and recording the result of the analysis of the cause; 5. The monitoring device of claim 4, further comprising:

6. the determining means determines that an abnormality has been detected when the Mahalanobis distance exceeds the first threshold value and the gap value exceeds a fourth threshold value that is greater than the second threshold value. The monitoring device according to claim 3.

7. the means for calculating the gap value calculates the gap value by using the process value which varies depending on the rotation speed of the rotating machine. The monitoring device according to claim 1 or 2.

8. the means for calculating the gap value calculates the estimated value of the process value by using a predetermined prediction formula having a rotation speed of the rotating machine as a parameter. The monitoring device according to claim 1 or 2.

9. the means for calculating the gap value calculates, as an estimate of the process value, an average value of the actual measurement values ​​measured until a predetermined time has elapsed since the start of monitoring the rotating machine. The monitoring device according to claim 1 or 2.

10. Calculating a Mahalanobis distance based on a process value of a rotating machine; calculating a gap value indicating a ratio of a difference between an actual measured value of the process value and the estimated value of the process value to a difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value; determining an operating state of the rotary machine based on the Mahalanobis distance and the gap value; The monitoring method includes:

11. Computer, A means for calculating a Mahalanobis distance based on a process value of the rotating machine; a means for calculating a gap value indicating a ratio of a difference between an actual measured value of the process value and the estimated value of the process value to a difference between a threshold value for determining an abnormality in the process value and the estimated value of the process value; a means for determining an operating state of the rotary machine based on the Mahalanobis distance and the gap value; A program to function as a

Images