Abnormality diagnosis support method and device of rotary machine

Abstract

To detect a sign of abnormality of a rotary machine.SOLUTION: An abnormality diagnosis support device (100) which supports abnormality diagnosis of a rotary machine which is operated by changing a set rotation speed includes: actual current value statistic amount calculation means (103) for calculating an actual current value statistic amount which is a current value statistic amount within predetermined time after a set rotation speed is changed by using a current value of an electric motor (M) of the rotary machine; calculation current value statistic amount calculation means (104) for calculating a calculation current value statistic amount which is a current value statistic amount corresponding to the change amount of the set rotation speed, based on a regression equation showing correlation between the change amount of the set rotation speed and the current value statistic amount at normal time; and generation means (105) for generating information which supports the abnormality diagnosis of the rotary machine based on a difference between the actual current value statistic amount and the calculation current value statistic amount.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an abnormality diagnosis support method and device for supporting abnormality diagnosis of a rotating machine that is operated by changing a set rotation speed. [Background technology]

[0002] A method has been proposed in which the current value of an electric motor is used to diagnose abnormalities in a rotating machine such as an electric motor or a roller driven by the electric motor. Patent Document 1 discloses that, with regard to equipment diagnosis of roller tables, a drive device that simultaneously controls multiple electric motors that drive each roll of a roller table that transports a metal strip is treated as one unit, and equipment diagnosis is performed by measuring the motor current of the electric motors individually for each drive device, and comparing the deviation between the average value of the measured motor current and the individually measured motor current value, or by comparing the current values ​​of adjacent motors. Furthermore, Patent Document 2 discloses that, with regard to equipment diagnosis of roller tables, a current change detection mechanism is provided which is composed of a start-up state detection unit which captures the current signal of the motor which drives the rolls one by one and detects the starting current and start-up time immediately after operation, and a roll eccentricity detection unit which detects the current value in a stable period and calculates the current oscillation period and amplitude, and by sequentially switching between each detection unit according to the operation signal of the motor, the behavior of the current change in the motor is continuously diagnosed according to the operating state from the start of operation to the stop of operation. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 3982433 [Patent Document 2] Patent No. 3688890 Summary of the Invention [Problem to be solved by the invention]

[0004] In the method disclosed in Patent Document 1, the motor current is measured during constant-speed operation under no-load conditions when there is no metal strip on the roller table, and when the motor is started. However, the current during constant-speed operation under no-load conditions does not fluctuate unless the equipment abnormality becomes large to a certain extent. Therefore, when an abnormality is detected based on the current measured during constant-speed operation under no-load conditions, it is immediately before a breakdown occurs, and it is necessary to suddenly stop operation for equipment repair to avoid trouble during operation. In addition, the current at the start of the motor is affected by disturbances, and for example, the lubrication state of the roll bearings changes with temperature, so the current fluctuates even under normal conditions, which may cause erroneous detection. The method disclosed in Patent Document 2 also measures the current during a stable, unloaded period when there are no objects on the roller table, and the startup current immediately after operation. As with Patent Document 1, this method detects an abnormality in the equipment just before a breakdown occurs, which may require sudden shutdown of operations for equipment repairs, and there is also a risk of false detection due to the effects of external disturbances.

[0005] The present invention has been made in consideration of the above-mentioned points, and an object of the present invention is to make it possible to detect signs of abnormality in a rotating machine. [Means for solving the problem]

[0006] The method for supporting abnormality diagnosis of a rotating machine of the present invention is an abnormality diagnosis support method for supporting abnormality diagnosis of a rotating machine that is operated with a changed set rotational speed, and is characterized by having the steps of: calculating actual current value statistics, which are current value statistics within a predetermined time after the set rotational speed is changed, using the current value of the motor of the rotating machine; calculating calculated current value statistics, which are current value statistics corresponding to the amount of change in the set rotational speed based on a regression equation that shows the correlation between the amount of change in the set rotational speed under normal conditions and the current value statistics; and generating information for supporting abnormality diagnosis of the rotating machine based on the difference between the actual current value statistics and the calculated current value statistics. The abnormality diagnosis support device for a rotating machine of the present invention is an abnormality diagnosis support device that supports abnormality diagnosis of a rotating machine that is operated with a changed set rotational speed, and is characterized in that it comprises an actual current value statistics calculation means that calculates actual current value statistics, which are current value statistics within a predetermined time after the set rotational speed is changed, using a current value of a motor of the rotating machine, a calculated current value statistics calculation means that calculates calculated current value statistics, which are current value statistics corresponding to the amount of change in the set rotational speed, based on a regression equation that shows the correlation between the amount of change in the set rotational speed and the current value statistics under normal conditions, and a generation means that generates information to support abnormality diagnosis of the rotating machine based on the difference between the actual current value statistics calculated by the actual current value statistics calculation means and the calculated current value statistics calculated by the calculated current value statistics calculation means. Effect of the Invention

[0007] According to the present invention, it becomes possible to detect symptoms of an abnormality in a rotating machine. [Brief description of the drawings]

[0008] [Figure 1] FIG. 2 is a diagram for explaining a hot rolling line and a hot run table according to the embodiment. [Diagram 2] 1 is a diagram illustrating a functional configuration of an abnormality diagnosis support device according to an embodiment. [Diagram 3] 4 is a flowchart showing an abnormality diagnosis support process executed by the abnormality diagnosis support device according to the embodiment. [Figure 4] 11A and 11B are diagrams illustrating an example of time-series changes in on / off status, set rotation speed, and current value. [Diagram 5] FIG. 13 is a diagram showing the relationship between the change amount of the set rotation speed and the current value statistics in a normal state. [Figure 6] 11 is a diagram for explaining a difference between actual current value statistics and calculated current value statistics; FIG. [Figure 7] 13 is a diagram showing a probability distribution of a difference between an actual current value statistic and a calculated current value statistic; FIG. [Figure 8] FIG. 13 is a graph showing the transition of the Kullback-Leibler distance. [Figure 9] FIG. 1 is a graph showing the relationship between the number of days before removal and the Kullback-Leibler distance in an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. In this embodiment, a conveying device called a hot run table in a hot rolling line will be described as an example of equipment to which the present invention is applied. FIG. 1 is a diagram for explaining a hot rolling line 1 and a hot run table 3 according to an embodiment. 1(a), a hot rolling line 1 includes, in order from the upstream side of the line, a roughing mill, a finishing mill 2, a hot run table 3, and a winding device 4 (not shown). A steel sheet 5 that has been hot rolled to a predetermined thickness by the roughing mill and the finishing mill 2 (not shown) is transported to the winding device 4 while being cooled by the hot run table 3, and is wound into a coil by the winding device 4.

[0010] As shown in FIG. 1(b), the hot run table 3 is configured to transport the steel sheet 5 by rotating a plurality of transport rolls R. A driving motor (electric motor) M is connected to each transport roll R. In this embodiment, the transport roll R and the motor M are rotating machines. The power output from the motor M is transmitted to the transport roll R via a rotating shaft, thereby rotating the transport roll R. In the example shown in FIG. 1(b), n transport rolls R and n motors M corresponding to the transport rolls R are provided from the upstream side to the downstream side of the line. In this way, the hot run table 3 is provided with n transport units, each of which is composed of one transport roll R and one motor M. Each transport roll R has the same structure. Moreover, each motor M has the same structure and has the same performance.

[0011] In addition, the main device 200 , the drive device 300 , and the abnormality diagnosis support device 100 are connected via a control backbone LAN 400 . The main device 200 sets the rotation speed of each motor M according to the dimensions of the steel plate 5 being transported, for example, at the timing of load on / off. Load on here means the timing when the leading edge of the steel plate 5 being transported reaches above the motor M, and load off means the timing when the trailing edge of the steel plate 5 being transported passes above the motor M. The main device 200 outputs the set rotation speeds of each motor M (called set rotation speeds) to the drive device 300 and the abnormality diagnosis support device 100 via the control backbone LAN 400. The driving device 300 is electrically connected to each motor M, and outputs a command to drive each motor M according to the set rotation speed output from the main device 200. The current value of the motor M detected by the driving device 300 is input to the abnormality diagnosis support device 100 in real time. The abnormality diagnosis support device 100 supports abnormality diagnosis of rotating machines (each motor M and each conveying roll R) operated by changing the set rotation speed. The inventors of the present application have studied a method for detecting a symptom of abnormality in the rotating machine, and have found that the symptom of abnormality in the rotating machine appears in the current value of the motor M immediately after the set rotation speed is changed. This is considered to be because the rotating machine is less susceptible to disturbances such as those at the time of start-up while it is rotating, and because the current fluctuates due to the change in the set rotation speed, i.e., the acceleration / deceleration characteristics of the motor M, and the sign of abnormality is likely to appear. Therefore, in the abnormality diagnosis support device 100, as described in detail below, an abnormality diagnosis support process is executed that utilizes the current value of the motor M after the set rotation speed is changed.

[0012] FIG. 2 is a diagram showing a functional configuration of the abnormality diagnosis support device 100 according to the embodiment. The abnormality diagnosis support device 100 includes a current value acquisition unit 101, a set rotation speed acquisition unit 102, an actual current value statistics calculation unit 103, a calculated current value statistics calculation unit 104, a determination unit 105, and an output unit . The current value acquisition unit 101 acquires and records the current value of the motor M detected by the driving device 300. The set rotation speed acquisition unit 102 acquires and records the rotation speed (set rotation speed) of the motor M that is set by the main device 200.

[0013] The actual current value statistics calculation unit 103 calculates actual current value statistics, which are current value statistics within a predetermined time after the set rotation speed is changed, using the current value acquired by the current value acquisition unit 101. The current value statistics is an integrated average value of the current within the predetermined time. The predetermined time is appropriately set in advance according to the current value statistics calculated from the behavior of the current value change when the set rotation speed is changed. The calculated current value statistics calculation unit 104 calculates the calculated current value statistics, which are current value statistics corresponding to the change amount of the set rotation speed, based on a regression equation that indicates the correlation between the change amount of the set rotation speed and the current value statistics under normal conditions. The regression equation is obtained in advance based on the actual value of the change amount of the set rotation speed under normal conditions and the actual value of the current value statistics.

[0014] The judgment unit 105 judges whether or not an abnormality in the rotating machine has been detected based on the difference between the actual current value statistics calculated by the actual current value statistics calculation unit 103 and the calculated current value statistics calculated by the calculated current value statistics calculation unit 104, and generates information on the judgment result. The output unit 106 outputs information on the determination result generated by the determination unit 105. Output means that the information on the determination result generated by the determination unit 105 is displayed on a display device, transmitted to an external device, or stored in a storage device.

[0015] The abnormality diagnosis support device 100 described above can be realized by, for example, a computer device equipped with a CPU, ROM, RAM, etc., and the functions of each of the units 101 to 106 are realized by the CPU reading out a program stored in, for example, the ROM and executing this program. In addition, in the present embodiment, the abnormality diagnosis support device 100 is configured separately from the main device 200. However, for example, the main device 200 may function as the abnormality diagnosis support device.

[0016] 3 is a flowchart showing an abnormality diagnosis support process executed by the abnormality diagnosis support device 100 according to the embodiment. The abnormality diagnosis support process shown in FIG. In step S1, the current value acquisition unit 101 acquires the current value I of the motor M detected by the driving device 300. x Obtain and record. In step S2, the set rotation speed acquisition unit 102 acquires and records the rotation speed (set rotation speed) V of the motor M that is set by the main device 200. Here, in Fig. 4, the load on / off, the set rotation speed V, and the current value I x An example of the time series change of the load is shown below. At timing t1, the load status changes from ON to OFF, and the set rotation speed V becomes the rotation speed V a to rotational speed V b In addition, the current value I of the motor M changes with the change in the set rotation speed V. x Also, at timing t2, the load status changes from off to on, and then the set rotation speed V changes to the rotation speed V b to rotational speed V c In addition, the current value I of the motor M changes with the change in the set rotation speed V. x changes.

[0017] In step S3, the actual current value statistics calculation unit 103 calculates the current value I x Using this, actual current value statistics, which are current value statistics within a predetermined time t after the set rotation speed V is changed, are calculated. In this embodiment, actual current value statistics, which are current value statistics within a predetermined time t after the set rotation speed V is changed when switching from load-on to load-off, that is, from timing t1 in the example of Fig. 4, are calculated. Here, the predetermined time t is set to 6 [s]. Specifically, the actual current value statistics calculation unit 103 calculates the change amount V of the set rotation speed by using the data acquired in steps S1 and S2. ab The change in the set rotation speed V ab is expressed as equation (1). V ab =V b -V a (1) In addition, the actual current value statistics calculation unit 103 uses the data acquired in steps S1 and S2 to calculate the integrated average current value I s Extract the integrated average current I s is the current value I x The integrated value I´ for t seconds x and is expressed as equation (2). N is the number of steps at a given time t.

[0018]

number

[0019] In this way, the integrated average current I within a given time t after the set rotation speed V is changed s By using this, signs of abnormality in the motor M that appear in the current value after the set rotation speed V is changed can be made apparent, the difference between normality and abnormality can be clearly shown, and signs of abnormality can be easily detected. Note that the maximum current value within a predetermined time t after the set rotation speed V is changed may be used as the current value statistics. In this embodiment, the actual current value statistics, which are current value statistics within a predetermined time t, are calculated in response to the change in the set rotation speed V when switching from load-on to load-off, but the actual current value statistics may be calculated in response to the change in the set rotation speed V when switching from load-off to load-on, for example. However, the current value of the motor M when the steel sheet 5 is loaded is likely to include the influence of disturbances unrelated to the equipment state, such as the torque current for transporting the steel sheet 5. Therefore, it is preferable to calculate the actual current value statistics at a predetermined time t when the steel sheet 5 is not present in response to the change in the set rotation speed V when switching from load-on to load-off.

[0020] In step S4, the calculated current value statistics calculation unit 104 calculates the change amount V of the set rotation speed based on a regression equation that indicates the correlation between the change amount of the set rotation speed under normal conditions and the current value statistics. ab A calculated current value statistic, which is a current value statistic corresponding to The regression equation is determined in advance based on data when the motor M and the transport roll R are replaced with new ones, for example. As shown in FIG. 5, the change amount of the set rotation speed and the current value statistics (integrated average current I s ) is plotted. In order to grasp the relationship between the rotation speed and the current value, data is plotted with the horizontal axis representing the change in the set rotation speed and the vertical axis representing the current value statistics, and a regression equation is calculated. In this embodiment, the regression equation and the coefficient of determination are obtained as shown in equation (3). Regression equation: y = 0.2151x + 7.7762 (3) (x: change in set rotation speed, y: current value statistics) Coefficient of determination: R 2 =0.9977 Based on this regression equation, the change amount V of the set rotation speed obtained in step S3 is ab A calculated current value statistic, which is a current value statistic corresponding to When obtaining the regression equation, the time after the set rotation speed is changed at which the current value statistics are calculated so as to have a large coefficient of determination is searched for, and this time is set as the predetermined time t.

[0021] In step S5, the determination unit 105 calculates the difference between the actual current value statistics calculated in step S3 and the calculated current value statistics calculated in step S4. FIG. 6 is a diagram for explaining the difference between the actual current value statistics and the calculated current value statistics. As shown in FIG. 6(a), the difference between the actual current value statistics and the calculated current value statistics is called the residual. FIG. 6(b) shows a residual plot under normal conditions, and (c) shows a residual plot two weeks before the failure. The horizontal axis of FIG. 6(b) and (c) shows the change amount of the set rotation speed, and the vertical axis shows the residual. It can be seen that the residual plot two weeks before the failure has a biased distribution compared to the residual plot under normal conditions.

[0022] In step S6, the judgment unit 105 compares the probability distribution of the residual plot calculated in step S5 with the probability distribution of the residual plot in a normal state to judge whether an abnormality of the rotating machine has been detected and generates information on the judgment result. In this case, the Kullback-Leibler distance (also called Kullback-Leibler divergence, hereinafter referred to as KL distance) between the two probability distributions is calculated, and it is judged that an abnormality has been detected when the KL distance exceeds a predetermined threshold value. Specifically, FIG. 7 shows a probability distribution using the two residual plots shown in FIG. 6(b) and (c). In this embodiment, kernel density estimation, which is one of the non-parametric methods, is adopted as a method for calculating the distribution. Although it is not necessary to limit the method for estimating the distribution, since probability distributions of various shapes are expected, kernel density estimation that can express them is effective in this diagnosis. The kernel density estimator is expressed as in Equation (4), and the kernel function of this equation is given by the Gaussian kernel (Equation (5)) assuming a normal distribution. It can be seen that the probability distribution two weeks before the failure shifts to the right compared to the normal state.

[0023]

number

[0024] It can be seen from Fig. 7 that there is a difference between the two probability distributions, and in order to express this difference as an index that can be used for anomaly detection, the distance between the two probability distributions was estimated. In this embodiment, the KL distance was used as a method for estimating the distance between the probability distributions. The KL distance is expressed by equation (6). Fig. 8 shows the transition of the KL distance until the day of the failure. There is not much change until one month before the failure, but the KL distance gradually increases, and two weeks before the failure, the KL distance exceeds 0.2. The determination unit 105 determines that an abnormality has been detected when the KL distance exceeds, for example, 0.2. This makes it possible to detect an abnormality in the rotating machine approximately two weeks before a breakdown occurs.

[0025]

number

[0026] In this way, by using the probability distribution of the difference between the actual current value statistics and the calculated current value statistics, the difference between normal and abnormal can be clearly expressed, and signs of abnormality can be detected early. Note that instead of the probability distribution, the mean, variance, standard deviation, etc. may be used. In addition, by using the KL distance, it is possible to accurately detect anomalies from the probability distribution. Note that instead of the KL distance, the Pearson distance, the Jensen-Shannon distance, and the L 2 Distance or the like may also be used.

[0027] In step S7, the output unit 106 outputs information on the determination result generated in step S6, for example, to a user (a factory operator or a maintenance worker) that a symptom of an abnormality in the rotating machine has been detected.

[0028] As described above, by using the current value of the motor M after changing the set rotation speed, it becomes possible to detect signs of abnormality in the rotating machine at an early stage. This makes it possible to continue operation until, for example, a predetermined scheduled maintenance time, without being required to immediately stop operation for equipment repair after detecting signs of abnormality.

[0029] <Example> The results of an embodiment in which the present invention was applied to an actual hot run table will be described. Based on the current value of the motor M that drives each transport roll R of the hot run table, the KL distance was calculated using the above-mentioned procedure, and at the timing of the next regular maintenance in which the KL distance exceeded 0.2, the transport roll R was removed and a visual inspection of the bearings was performed. For comparison, a transport roll R whose KL distance did not exceed 0.2 was also removed and a visual inspection of the bearings was performed. In the visual inspection of this embodiment, the failure mode phases for the transport roll R bearings were defined as follows. In phases 2 to 3, replacement is recommended, and in phase 4, a sudden stop is required. 0: Healthy 1: Grease deteriorates and discolors (brown to black) due to aging and heat / foreign matter gets mixed into the grease 2: Poor lubrication (oil film breakdown) causes wear on the cage 3: Increased cage play leads to increased roller wear 4: Wear of the cage and rollers causes the rollers to skew

[0030] Figure 9(a) shows the relationship between the number of days before removal and the KL distance for transport roll R, where the KL distance exceeded 0.2. Visual inspection of the transport roll R revealed roller wear and cage wear in the bearings. The failure mode evaluation showed that it was in phase 3, and if it had not been replaced, it could have led to a sudden shutdown. The KL distance began to fluctuate gradually two months before removal, and exceeded 0.2 one month before removal, so it can be said that signs of an abnormality were detected. Figure 9(b) shows the relationship between the number of days before removal and the KL distance for transport roll R, whose KL distance does not exceed 0.2. Visual inspection of the transport roll R revealed minor indentations and discoloration in the bearings, but no abnormalities, and the failure mode assessment indicated that the roll was in Phase 1. The KL distance also showed little variation until the day of removal, remaining a relatively low value. As described above, by using the abnormality diagnosis support according to the present invention, it is possible to prevent a sudden shutdown due to an abnormality in a rotating machine.

[0031] In this embodiment, in the abnormality diagnosis support device 100, the judgment unit 105 functions as a generating means as defined in the present invention, judges whether or not an abnormality in the rotating machine has been detected, and provides the judgment result to the user, but is not limited to this. The abnormality diagnosis support device 100 only needs to have a function of generating information to support abnormality diagnosis of the rotating machine, and may generate a characteristic diagram showing the transition of the KL distance as shown in Fig. 8 and provide it to the user, with the final diagnosis of whether or not there is an abnormality in the motor M being left to the user.

[0032] Although the present invention has been described above with reference to the embodiments, the above embodiments are merely illustrative of the specific examples of the present invention, and the technical scope of the present invention should not be interpreted as being limited by these embodiments. In other words, the present invention can be embodied in various forms without departing from its technical concept or main features. The present invention can also be realized by supplying software (program) that realizes the functions of the present invention to a system or device via a network or various storage media, and having the computer of that system or device read and execute the program.

[0033] It can also be written as follows: [1] A method for assisting in the diagnosis of an abnormality of a rotating machine operated by changing a set rotation speed, comprising: calculating actual current value statistics, which are current value statistics within a predetermined time period after a set rotation speed is changed, using a current value of the electric motor of the rotating machine; calculating a calculated current value statistic, which is a current value statistic corresponding to the change amount of the set rotation speed, based on a regression equation showing a correlation between the change amount of the set rotation speed and the current value statistic in a normal state; a step of generating information for supporting abnormality diagnosis of the rotating machine based on a difference between the actual current value statistics and the calculated current value statistics. [2] The method for supporting abnormality diagnosis of a rotating machine according to [1], wherein the current value statistic is an integrated average value of the current within the specified period of time. [3] The method for supporting abnormality diagnosis of a rotating machine described in [1] or [2], characterized in that in the step of generating information to support abnormality diagnosis of the rotating machine, information is generated based on a probability distribution of the difference between the actual current value statistics and the calculated current value statistics and a probability distribution under normal conditions that has been obtained in advance. [4] The method for supporting abnormality diagnosis of a rotating machine described in [3], characterized in that in the step of generating information to support abnormality diagnosis of the rotating machine, a probability distribution of the difference between the actual current value statistics and the calculated current value statistics is compared with the probability distribution under normal conditions to determine whether or not an abnormality in the rotating machine has been detected, and information on the determination result is generated. [5] The method for supporting abnormality diagnosis of a rotating machine described in [3] or [4], characterized in that in the step of generating information to support abnormality diagnosis of the rotating machine, the Kullback-Leibler distance between the two probability distributions is calculated. [6] An abnormality diagnosis support device that supports abnormality diagnosis of a rotating machine that is operated by changing a set rotation speed, an actual current value statistic calculation means for calculating, using a current value of the electric motor of the rotating machine, an actual current value statistic which is a current value statistic within a predetermined time after a set rotation speed is changed; a calculated current value statistic calculation means for calculating a calculated current value statistic, which is a current value statistic corresponding to an amount of change in the set rotational speed, based on a regression equation showing a correlation between an amount of change in the set rotational speed and a current value statistic under normal conditions; a generation means for generating information to support abnormality diagnosis of the rotating machine based on a difference between the actual current value statistics calculated by the actual current value statistics calculation means and the calculated current value statistics calculated by the calculated current value statistics calculation means. [Explanation of symbols]

[0034] M: motor, R: transport roll, 1: hot rolling line, 3: hot run table, 100: abnormality diagnosis support device, 101: current value acquisition unit, 102: set rotation speed acquisition unit, 103: actual current value statistics calculation unit, 104: calculated current value statistics calculation unit, 105: judgment unit, 106: output unit

Claims

1. 1. An abnormality diagnosis support method for supporting abnormality diagnosis of a rotating machine operated by changing a set rotation speed, comprising: calculating actual current value statistics, which are current value statistics within a predetermined time period after a set rotation speed is changed, using a current value of the electric motor of the rotating machine; calculating a calculated current value statistic, which is a current value statistic corresponding to the change amount of the set rotation speed, based on a regression equation showing a correlation between the change amount of the set rotation speed and the current value statistic in a normal state; a step of generating information for supporting abnormality diagnosis of the rotating machine based on a difference between the actual current value statistics and the calculated current value statistics.

2. 2. The method for assisting in diagnosis of abnormality in a rotating machine according to claim 1, wherein the current value statistic is an integrated average value of the current within the predetermined time period.

3. The method for supporting abnormality diagnosis of a rotating machine as described in claim 1 or 2, characterized in that in the step of generating information to support abnormality diagnosis of the rotating machine, information is generated based on a probability distribution of the difference between the actual current value statistics and the calculated current value statistics and a probability distribution under normal conditions that has been obtained in advance.

4. 4. The method for supporting abnormality diagnosis of a rotating machine as described in claim 3, characterized in that in the step of generating information to support abnormality diagnosis of the rotating machine, a probability distribution of a difference between the actual current value statistics and the calculated current value statistics is compared with the probability distribution under normal conditions to determine whether or not an abnormality has been detected in the rotating machine, and information on the determination result is generated.

5. 4. The method for supporting abnormality diagnosis of a rotating machine according to claim 3, wherein in the step of generating information for supporting abnormality diagnosis of the rotating machine, a Kullback-Leibler distance between the two probability distributions is calculated.

6. An abnormality diagnosis support device that supports abnormality diagnosis of a rotating machine that is operated by changing a set rotation speed, an actual current value statistic calculation means for calculating, using a current value of the electric motor of the rotating machine, an actual current value statistic which is a current value statistic within a predetermined time after a set rotation speed is changed; a calculated current value statistic calculation means for calculating a calculated current value statistic, which is a current value statistic corresponding to an amount of change in the set rotational speed, based on a regression equation showing a correlation between an amount of change in the set rotational speed and a current value statistic under normal conditions; a generation means for generating information to support abnormality diagnosis of the rotating machine based on a difference between the actual current value statistics calculated by the actual current value statistics calculation means and the calculated current value statistics calculated by the calculated current value statistics calculation means.

Images