ANOMALYSIS DETECTION DEVICE, ANOMALYSIS DETECTION METHOD, AND PROGRAM

Abstract

To provide an abnormality detection device which can quickly diagnose the abnormality of a machinery plant.SOLUTION: An abnormality detection device for detecting the presence of abnormality in a machinery plant includes a processor that executes: a process of acquiring, on the basis of a detection signal outputted from a first vibration sensor that measures the vibration of machinery plant, a measured value composed of the amplitude and phase of the vibration at a predetermined frequency point; a process of calculating, with respect to a unit space composed of a plurality of measured values acquired at a plurality of past points of time, the Mahalanobis distance of the measured value acquired at a point of time when evaluating the machinery plant; and a process of assessing, when the calculated Mahalanobis distance exceeds a prescribed threshold, that abnormality has occurred in the machinery plant.SELECTED DRAWING: Figure 1

Description

[Technical field]

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

[0002] A method for detecting anomalies using the Mahalanobis-Taguchi (MT) method is known as a technology for diagnosing the soundness of inspection objects such as mechanical equipment. In the MT method, the Mahalanobis distance is calculated using the inverse matrix of the covariance matrix of reference data (e.g., a set of measured values ​​of various characteristic items under normal conditions). If the calculated Mahalanobis distance exceeds a specified threshold, it can be determined that the condition of the inspection object is abnormal.

[0003] For example, Patent Document 1 describes an abnormality detection device that detects abnormal vibrations in a rotating machine based on the results of frequency analysis of a measurement signal y from a vibration sensor. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-144111 A Summary of the Invention [Problem to be solved by the invention]

[0005] Specifically, the result of the frequency analysis is the Fourier coefficient obtained as a result of the discrete Fourier transform. The anomaly detection device described in Patent Document 1 detects anomalies based on the complex Fourier coefficients {a(0Δf)+jb(0Δf), a(1Δf)+jb(1Δf), a(2Δf)+jb(2Δf), ..., a((N-1)Δf)+jb((N-1)Δf)} obtained for frequency points {0Δf, 1Δf, 2Δf, ..., (N-1)Δf} determined from the time interval ΔT of the time history {y(ΔT), y(2ΔT), ..., y(NΔT)} of the measured signal y and the number of data points N to be subjected to the frequency analysis. j is an imaginary unit. For the discrete Fourier transform, a fast Fourier transform (FFT) is used in terms of the calculation speed. Executing this in a control device for a machine equipment poses problems in terms of the amount of calculation processing and the immediacy of detection.

[0006] In FFT, the number of frequency sample points per signal is set to a power of 2, for example 256 points or 1024 points, and one Fourier coefficient is calculated for each set frequency point, so the number of Fourier coefficient calculations also increases to 256 points or 1024 points. In other words, in FFT, the same number of Fourier coefficients as the number of frequency sample points must be calculated.

[0007] If the signal is 1ch, frequency analysis using FFT is possible. However, it can be difficult if there is a 16ch measurement signal, for example. If the number of frequency sample points is 256, then 16ch x 256 points, or 4096 points, of Fourier coefficients will have to be calculated. Even with today's improvements in calculation speed, this is still not something that can be processed by a general control device, and a dedicated calculation device (frequency analysis device) must be installed. This is not desirable from the perspective of product costs.

[0008] Furthermore, even if Fourier coefficients are calculated for 256 frequency points per channel, in many cases only a portion of them is used for vibration diagnosis. The frequency points where abnormalities are likely to occur are known in advance, so it is sufficient to know the Fourier coefficients of the frequency points that belong to that frequency band. Usually, most of the calculated Fourier coefficients are not used. However, when calculating Fourier coefficients, it is not possible to calculate Fourier coefficients by limiting the frequency points. For this reason, in conventional technology, Fourier coefficients are calculated even for frequency points that will never be used, but this is wasteful.

[0009] It's not all waste of calculations. FFT also has issues with immediacy. In FFT, if the time interval of the measurement signal is ΔT [s] and the number of measurement points to be subjected to FFT is N, the frequency resolution, i.e., the interval between frequency points, is calculated as Δf = (ΔTN)-1 [Hz], and the Fourier coefficients are calculated for frequency points {0Δf, 1Δf, 2Δf, ..., (N-1)Δf}. For example, if ΔT = 0.01 sec and the resolution is evaluated at 0.01 Hz, a data length of N = 1024 points is sufficient. At this time, the time to accumulate data becomes an issue. It takes 10.24 seconds to accumulate 1024 points with ΔT = 0.01 sec. If the target of interest is, for example, a 0.1 Hz component (10 second period), it is not a problem if it takes 10 seconds to accumulate. However, if you want to investigate, for example, a 1 Hz frequency component (1 second period), you want to investigate it in 1 second, not 10 seconds. In order to quickly detect abnormalities or symptoms, it is important to make judgments based on short data sets. Therefore, the time it takes to accumulate data is an issue with the conventional FFT technology.

[0010] Furthermore, FFT only reveals the average frequency characteristics of the N accumulated points. For example, suppose an abnormality appears in the 1 Hz component, and the abnormal characteristics appear in one second in the middle of the N accumulated data points. Since FFT calculates the average frequency characteristics over 10 seconds, the abnormal characteristics that appear in a certain second are diluted and diluted by the other 9 seconds of data, making it difficult to detect the abnormality. This dilution of characteristics is also an issue with FFT.

[0011] The present disclosure has been made in consideration of such problems, and provides an abnormality detection device, an abnormality detection method, and a program that can quickly diagnose abnormalities in mechanical equipment. [Means for solving the problem]

[0012] According to one aspect of the present disclosure, an abnormality detection device for detecting whether or not an abnormality exists in mechanical equipment includes a processor that executes the following processes: acquiring measurement values ​​consisting of the amplitude and phase of vibration at predetermined frequency points based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; calculating a Mahalanobis distance of the measurement values ​​acquired at the time when the mechanical equipment is to be evaluated based on a unit space constituted by multiple measurement values ​​acquired at multiple past time points; and determining that an abnormality has occurred in the mechanical equipment if the calculated Mahalanobis distance exceeds a predetermined threshold value.

[0013] Furthermore, according to one aspect of the present disclosure, an abnormality detection device for detecting the presence or absence of an abnormality in mechanical equipment includes a processor that executes the following processes: acquiring measurement values ​​consisting of amplitude and phase of vibration at predetermined frequency points based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; calculating a plurality of components obtained by decomposing a measurement value vector consisting of the measurement values ​​in the direction of an index vector defined as an index of abnormality; and determining that an abnormality has occurred in the mechanical equipment if the absolute value of the calculated components exceeds a predetermined threshold value.

[0014] According to one aspect of the present disclosure, an abnormality detection device for detecting the presence or absence of an abnormality in mechanical equipment includes a process of acquiring a measurement value vector consisting of measurement values ​​measured by a first vibration sensor that measures vibrations of the mechanical equipment, a process of dividing an index vector defined as an index of an abnormality into a real part and an imaginary part using a phase angle of a predetermined frequency point, and a process of calculating inner products of each of the real part and the imaginary part and the measurement value vector, and a process of determining that an abnormality has occurred in the mechanical equipment when an absolute value of a time average of the calculated inner products over one period exceeds a predetermined threshold value. The device includes a processor that executes the above steps.

[0015] Moreover, according to one aspect of the present disclosure, an abnormality detection method for detecting the presence or absence of an abnormality in mechanical equipment includes the steps of acquiring measurement values ​​consisting of the amplitude and phase of vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment, calculating a Mahalanobis distance of the measurement values ​​acquired at the time when the mechanical equipment is evaluated based on a unit space constituted by a plurality of the measurement values ​​acquired at a plurality of past time points, and determining that an abnormality has occurred in the mechanical equipment if the calculated Mahalanobis distance exceeds a predetermined threshold value.

[0016] In addition, according to one aspect of the present disclosure, the program causes a computer of an abnormality detection device that detects whether or not there is an abnormality in mechanical equipment to execute the following steps: acquiring measurement values ​​consisting of the amplitude and phase of the vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; calculating the Mahalanobis distance of the measurement values ​​acquired at the time when the mechanical equipment is evaluated based on a unit space constituted by multiple measurement values ​​acquired at multiple past time points; and determining that an abnormality has occurred in the mechanical equipment if the calculated Mahalanobis distance exceeds a predetermined threshold value. Effect of the Invention

[0017] According to the abnormality detection device, the abnormality detection method, and the program disclosed herein, abnormality diagnosis of vibrations in mechanical equipment can be performed quickly. [Brief description of the drawings]

[0018] [Figure 1] 1 is a diagram illustrating an overall configuration of an anomaly detection system according to a first embodiment of the present disclosure. [Diagram 2] FIG. 4 is a diagram illustrating an example of a process of the abnormality detection device according to the first embodiment of the present disclosure. [Diagram 3]FIG. 11 is a diagram illustrating an overall configuration of an anomaly detection system according to a second embodiment of the present disclosure. [Figure 4] FIG. 11 is a diagram illustrating an example of a process of an abnormality detection device according to a second embodiment of the present disclosure. [Diagram 5] FIG. 11 is a diagram illustrating a functional configuration of an abnormality detection device according to a third embodiment of the present disclosure. [Figure 6] FIG. 13 is a diagram illustrating an example of a process performed by an abnormality detection device according to a third embodiment of the present disclosure. [Figure 7] FIG. 11 is a diagram illustrating a functional configuration of an abnormality detection device according to a fourth embodiment of the present disclosure. [Figure 8] FIG. 13 is a diagram illustrating an example of a process performed by an abnormality detection device according to a fourth embodiment of the present disclosure. [Figure 9] FIG. 13 is a diagram illustrating an example of processing performed by an abnormality detection device according to a fifth embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] <First embodiment> Hereinafter, an abnormality detection device 2 according to a first embodiment of the present disclosure will be described with reference to FIGS.

[0020] (Overall composition) FIG. 1 is a diagram showing an overall configuration of a mechanical installation according to a first embodiment of the present disclosure. As shown in FIG. 1, the abnormality detection system 100 includes a mechanical equipment 10, an abnormality detection device 2, and a control device 3.

[0021] The mechanical equipment 10 is a machine or a group of equipment that is the target of abnormality diagnosis by the abnormality detection device 2. Furthermore, a first vibration sensor 11 (11A to 11D) is attached to each measurement point of the mechanical equipment 10. The first vibration sensor 11 is, for example, an accelerometer, a microphone that measures sound caused by vibration, a pressure gauge that measures pressure caused by vibration, a stress meter that measures stress caused by vibration, or the like. The first vibration sensor 11 outputs a detection signal that measures the vibration of the mechanical equipment 10 to the abnormality detection device 2.

[0022] The abnormality detection device 2 detects the presence or absence of an abnormality in the mechanical equipment 10 based on a detection signal acquired from the first vibration sensor 11 of the mechanical equipment 10. The functional configuration of the abnormality detection device 2 will be described later.

[0023] The control device 3 generates a control signal for controlling the operation of the mechanical equipment 10. For example, when the control device 3 receives a determination result from the abnormality detection device 2 indicating that abnormal vibration is occurring in the mechanical equipment 10, it generates a control signal for stopping the operation of the mechanical equipment 10, a control signal for lowering the output of the mechanical equipment 10, etc.

[0024] (Functional configuration of the abnormality detection device) Next, the functional configuration of the abnormality detection device 2 will be described with reference to FIG. As shown in FIG. 1, the abnormality detection device 2 includes a processor 21, an input / output unit 22, a main storage device 23, and an auxiliary storage device 24.

[0025] The processor 21 controls the overall operation of the anomaly detection device 2. The processor 21 operates according to a predetermined program to fulfill the functions of a measurement value acquisition unit 210, a unit space generation unit 211, a Mahalanobis distance calculation unit 212, a determination unit 213, and a frequency point designation unit 214.

[0026] The measurement value acquisition unit 210 acquires measurement values ​​consisting of the amplitude and phase of vibration at predetermined frequency points based on the detection signal output from the first vibration sensor 11 of the mechanical equipment 10.

[0027] The unit space generating section 211 generates a unit space including a plurality of measurement values ​​acquired at a plurality of past time points. The unit space generating section 211 also stores the generated unit space in the auxiliary storage device 24.

[0028] The Mahalanobis distance calculation unit 212 calculates the Mahalanobis distance of the measurement value acquired at the time when the mechanical equipment 10 is evaluated, with the unit space as a reference.

[0029] The determination unit 213 determines, based on the calculated Mahalanobis distance, whether or not an abnormality has occurred in the mechanical equipment 10. Specifically, when the calculated Mahalanobis distance exceeds a predetermined threshold, the determination unit 213 determines that an abnormality has occurred in the mechanical equipment 10.

[0030] The frequency point designation unit 214 designates a predetermined frequency point to the measurement value acquisition unit 210 .

[0031] The input / output unit 22 receives an input of a detection signal from the first vibration sensor 11 of the mechanical equipment 10. In addition, the input / output unit 22 outputs to the control device 3 the result of the determination made by the abnormality detection device 2 as to the presence or absence of an abnormality.

[0032] The main memory device 23 has a storage area necessary for the operation of the processor 21 .

[0033] The auxiliary storage device 24 is a large-capacity storage device such as a hard disk drive (HDD) or a solid state drive (SSD). In the auxiliary storage device 24, various data acquired or generated by each unit of the abnormality detection device 2 is stored.

[0034] (Processing of abnormality detection device) FIG. 2 is a diagram illustrating an example of a process of the abnormality detection device according to the first embodiment of the present disclosure. The abnormality diagnosis process of the abnormality detection device 2 will be described in detail below with reference to FIG.

[0035] The feature of this embodiment is the means of frequency analysis executed in the anomaly detection device 2. As mentioned above, FFT has the following problems: (a) it requires a processing device due to its heavy computational load, (b) it takes time to accumulate data points, and (c) it dilutes the characteristics of an anomaly. Therefore, in the anomaly detection device 2 according to this embodiment, the measurement value acquisition unit 210 acquires measurement values ​​that represent the frequency characteristics of each detection signal by using the principle equations (1A) and (1B) of the Fourier transform instead of FFT.

[0036]

number

[0037]

number

[0038] Here, n is a predetermined serial number of the detection signal. For example, as shown in FIG. 1, if detection signals are to be acquired from four vibration sensors 11, n is one of {1, 2, 3, 4}, and the period {T 1 ,T 2 ,T 3 ,T 4 Calculate the Fourier coefficients of the frequency components of period {T 1 ,T 2 ,T 3 ,T 4 The values ​​of} may be the same.

[0039] In addition, the detection signal {y 1 ,y 2 ,y 3 ,y 4} may include the same signal obtained by splitting the signal of the same vibration sensor 11. However, the detection signal {y 1 ,y 2 ,y 3 ,y 4} includes multiple identical signals, the period {T 1 ,T 2 ,T 3 ,T 4 It is common to change the value of}. 1 andy 2 are branched signals of the same vibration sensor 11A, T 1 and T 2 In general, y is set to a different value. Unless there is a special situation such as a broken branch wiring, the same vibration sensor should be branched and 1 andy 2 This is because the Fourier coefficients of the same periodic component of are the same, so it is sufficient to calculate only one of them.

[0040] Equation (1A) and equation (1B) are time history data for one period, i.e., tT n Since it can be calculated from the time history data from t to t, the above problems (b) and (c) are solved. Furthermore, since formulas (1A) and (1B) are simple integral formulas, the above problem (a) is also solved. Furthermore, according to formulas (1A) and (1B), the time history data does not need to be at equal intervals. In the FFT, n had to be an integer multiple of the time interval ΔT of the measurement signal. However, this is not necessary in equations (1A) and (1B), and calculations are possible for frequency points that are not integer multiples of ΔT, as shown in the following equation (2). n is expressed by the following equation (2).

[0041]

number

[0042] Here, m n is T n is the quotient of r n is the remainder.

[0043] For example, the developer or operator of the mechanical equipment 10 has prior knowledge of frequency points at which characteristics of each detection signal tend to appear. 1 ,T 2 ,T 3 ,T 4} is set based on such prior knowledge. The frequency point designation unit 214, as shown in FIG. 1 ,y 2 ,y 3 ,y 4}, a predetermined period {T 1 ,T 2 ,T 3 ,T 4} is specified.

[0044] Then, equations (1A) and (1B) are calculated using the following equations (3A) and (3B).

[0045]

number

[0046]

number

[0047] That is, the measurement value acquisition unit 210 acquires the detection signal y n For each of them, the period T designated by the frequency point designation unit 214 is calculated using the above formulas (3A) and (3B). n The Fourier coefficients in {a n ,b n} is calculated. n ,b n} corresponds to a measurement value consisting of the amplitude and phase of vibration at a predetermined frequency point.

[0048] In addition, the Mahalanobis distance calculation unit 212 calculates the measurement value {a n ,b n} and calculates the Mahalanobis distance. The unit space is generated in advance by the unit space generating unit 211 based on a plurality of measurement values ​​(Fourier coefficients) acquired during a period when the mechanical equipment 10 was operating normally. Since the method of generating the unit space and the method of calculating the Mahalanobis distance can be performed using existing technology (for example, the technology described in Patent Document 1), a description thereof will be omitted.

[0049] Next, when the Mahalanobis distance calculated by the Mahalanobis distance calculation unit 212 exceeds a predetermined threshold, the determination unit 213 determines that an abnormality has occurred in the mechanical equipment 10. On the other hand, when the Mahalanobis distance is equal to or less than the threshold, the determination unit 213 determines that no abnormality has occurred in the mechanical equipment 10. In addition, the determination unit 213 outputs the determination result to the control device 3. When the control device 3 receives a determination result indicating the occurrence of an abnormality from the abnormality detection device 2, it performs control such as stopping the operation of the mechanical equipment 10 or reducing the output.

[0050] The abnormality detection device 2 detects the presence or absence of an abnormality in the mechanical equipment 10 by repeatedly executing the above-mentioned process.

[0051] (Action and effect) As described above, the abnormality detection device 2 according to this embodiment detects the detection signal y n A predetermined period T n The measured amplitude and phase of the vibration (Fourier coefficient a n ,b n ) is acquired, and if the Mahalanobis distance of the acquired measurement value exceeds a predetermined threshold, it is determined that an abnormality has occurred in the mechanical equipment 10.

[0052] In this way, the abnormality detection device 2 only needs to calculate the Fourier coefficients at specific frequency points, and therefore the calculation load can be significantly reduced compared to conventional techniques. This enables the abnormality detection device 2 to quickly diagnose abnormalities in the mechanical equipment 10. Furthermore, the abnormality detection device 2 can calculate the Fourier coefficients by itself without the need for a separate frequency analysis device, and therefore the cost of the entire anomaly detection system 100 can be reduced.

[0053] In addition, the abnormality detection device 2 uses the principle equation of Fourier transform to convert one detection signal y n For one period T n Specify the Fourier coefficients {a n ,b n}.

[0054] In this way, the abnormality detection device 2 can perform abnormality diagnosis from the time history data for one cycle, thereby shortening the time required to accumulate data. For example, when it is desired to examine a 1 Hz frequency component (1 second cycle) for a certain detection signal, it is possible to perform diagnosis from the time history data for the past 1 second, without accumulating data for several seconds. Furthermore, since the abnormality detection device 2 can perform diagnosis using a small number of data points, it is possible to prevent the characteristics of an abnormality from becoming diluted. This allows the abnormality detection device 2 to improve its sensitivity to abnormalities.

[0055] <Second embodiment> Next, an abnormality detection device 2 according to a second embodiment of the present disclosure will be described with reference to FIGS. Components common to the first embodiment are given the same reference numerals and detailed description thereof will be omitted.

[0056] (Overall composition) FIG. 3 is a diagram showing an overall configuration of an anomaly detection system according to the second embodiment of the present disclosure. 3, in the abnormality detection system 100 according to this embodiment, the abnormality detection device 2 further acquires a detection signal from a second vibration sensor 14 provided in the mechanical equipment 10. Like the first vibration sensor 11, the second vibration sensor 14 is an accelerometer, a microphone, or the like.

[0057] For example, the mechanical equipment 10 has a power unit 12 and a gear 13. The power unit 12 drives the mechanical equipment 10. The gear 13 changes the rotation speed (frequency) of the power unit 12 and transmits it to the mechanical equipment 10. As shown in Fig. 3, the second vibration sensor 14 measures, for example, the vibration of the power unit 12.

[0058] (Functional configuration of the abnormality detection device) 3, the processor 21 of the abnormality detection device 2 according to this embodiment has a frequency detection unit 215 as a functional unit instead of the frequency point designation unit 214. The frequency detection unit 215 detects a characteristic frequency that indicates a characteristic of the vibration of the mechanical equipment 10.

[0059] The measurement value acquisition section 210 acquires, for each detection signal of the first vibration sensor 11, a Fourier coefficient at a characteristic frequency detected by the frequency detection section 215.

[0060] (Processing of abnormality detection device) FIG. 4 is a diagram illustrating an example of processing performed by the abnormality detection device according to the second embodiment of the present disclosure. The abnormality diagnosis process of the abnormality detection device 2 will be described in detail below with reference to FIG.

[0061] The feature of this embodiment is that the abnormality detection device 2 detects a characteristic frequency of the mechanical equipment 10 to be diagnosed. The characteristic frequency is, for example, the rotation speed of the power unit 12 of the mechanical equipment 10.

[0062] In the conventional FFT, the frequency points targeted for frequency analysis are discrete values ​​that depend on the number of data points N of the detection signal and the time interval ΔT of the measurement. In other words, in the FFT, if the time interval of the detection signal is ΔT [s] and the number of measurement points targeted for the FFT is N, the frequency resolution is Δf = (ΔTN) -1 [Hz], and the Fourier coefficients are calculated for frequency points {0Δf, 1Δf, 2Δf, ..., (N-1)Δf} that are multiples of the frequency resolution. As in the previous explanation, we will use ΔT = 0.01 [s] and N = 1024 as an example. In this case, the frequency resolution Δf is approximately 0.098 Hz. For example, suppose that the characteristic component of the vibration in the measured value appears at 1.025 Hz. Since the frequency points are restricted to multiples of the resolution in FFT, the closest frequency points to 1.025 Hz are 0.98 Hz, which is the 11th frequency point, and 1.07 Hz, which is the 12th frequency point. In such a case, it is easy to substitute the frequency characteristics of the 1.025 Hz frequency points with those of the 0.98 Hz or 1.025 Hz frequency points. However, in such a case, the 1.025 Hz vibration is divided into a 0.98 Hz component and a 1.07 Hz component. As a result, the FFT cannot correctly measure the characteristic frequency of 1.025 Hz.

[0063] In contrast, the anomaly detection device 2 according to the first embodiment has the advantage that the frequency points can be set continuously, independent of the frequency resolution Δf. However, what if the dominant frequency component is not constant at 1.025 Hz, but fluctuates over time, for example, between 0.9 Hz and 1.1 Hz? In actual mechanical equipment 10, the characteristics of many equipment change depending on the operating conditions and aging, so it is valuable to respond to changes in the dominant frequency component.

[0064] To address this, the abnormality detection device 2 according to this embodiment has a frequency detection unit 215 instead of the frequency point designation unit 214 according to the first embodiment. The frequency detection unit 215 performs the process described below to detect a characteristic frequency (characteristic frequency) of the mechanical equipment 10.

[0065] First, the frequency detection unit 215 measures the signal z for detecting characteristic vibrations, and performs a Fourier transform on this signal using equations (4A) and (4B).

[0066]

number

[0067]

number

[0068] Specifically, z may be a vibration of the power unit 12 such as an electric motor. For example, in the same manner as in the first embodiment, if four detection signals are considered, the signal z n is {z 1 ,z 2 ,z 3 ,z 4 The elements of z may be the outputs of separate vibration sensors 14, or may be branched output signals of the same vibration sensor 14.

[0069] Here, θ in equations (4A) and (4B) n Z nIn order to make the phase of the characteristic frequency component of the signal match that of the signal, the signal is adjusted by a PI controller, for example, as shown in equation (5). In equation (5), atan is the inverse function of the tangent function, and atan(d n / c n ) is z n is an approximation of the phase [rad] of the characteristic frequency component of n / c n The approximation error becomes zero when )=0.

[0070]

number

[0071] The frequency detection unit 215 detects z n As a signal representing the characteristic frequency of n Output.

[0072] The measurement value acquisition unit 210 detects ω n Using this, the phase angle θ used in the Fourier transform from equation (6) is n get.

[0073]

number

[0074] In addition, the measurement value acquisition unit 210 n andr n is determined so as to satisfy equation (7) instead of equation (2) in the first embodiment.

[0075]

number

[0076] Then, the measurement value acquisition unit 210 calculates the detection signal y n Fourier coefficient a n and b n Calculate.

[0077]

number

[0078]

number

[0079] Fourier coefficient a n and b n The processing after obtaining is the same as in the first embodiment.

[0080] (Action and effect) As described above, the abnormality detection device 2 according to the present embodiment detects the detection signal z n Based on this, the characteristic frequency ω n , and the detected characteristic frequency ω n The measured amplitude and phase of the vibration (Fourier coefficient a n ,b n ) to get the

[0081] In this way, the abnormality detection device 2 can automatically detect frequency components that change according to the operating conditions and aging deterioration of the mechanical equipment 10. This allows the abnormality detection device 2 to acquire more accurate Fourier coefficients, thereby further improving the accuracy of abnormality detection.

[0082] <Third embodiment> Next, an abnormality detection device 2 according to a third embodiment of the present disclosure will be described with reference to FIGS. Components common to the first and second embodiments are given the same reference numerals and detailed description thereof will be omitted.

[0083] (Functional configuration of the abnormality detection device) FIG. 5 is a diagram illustrating a functional configuration of an abnormality detection device according to the third embodiment of the present disclosure. 5, the processor 21 of the abnormality detection device 2 according to this embodiment further includes an estimation unit 216. The estimation unit 216 estimates a characteristic frequency of each measurement point of the mechanical equipment 10 based on the detection signal acquired from the second vibration sensor 14 and a vibration frequency model of the mechanical equipment 10.

[0084] In the mechanical equipment 10, for example, the rotation speed of the power unit 12 and its integer multiple harmonics appear as characteristic frequencies. In addition to the rotation speed of the power unit 12, other characteristics include natural frequencies determined by the rigidity and mass of the mechanical equipment 10, and resonant frequencies determined by the density and compressibility of the fluid in the piping. These vibrations are not independent at each location on the mechanical equipment 10, but are linked to each other. The most common link is when vibrations occur at the same frequency. Or, an integer multiple of the characteristic frequency of one part may appear in another part.

[0085] The anomaly detection device 2 according to the second embodiment receives multiple Fourier coefficients and detects the presence or absence of an anomaly by comparing the way in which the multiple Fourier coefficients are linked at each time with the way in which they are linked in a predetermined normal state using the scale of Mahalanobis distance. If there is an error in the frequency at which the Fourier coefficients are evaluated, the calculation results of the Fourier coefficients will contain an error, which may make it difficult to correctly detect an anomaly.

[0086] For example, assume that four measurement points are known to have the property of vibrating at the same frequency around 1 Hz, and the frequency detection unit 215 outputs {1.07 Hz, 1.07 Hz, 0.98 Hz, 1.07 Hz}. If the Mahalanobis distance is evaluated using these frequencies, the third Fourier coefficient will have a different frequency from the others and will not match. In such a case, it is important to prioritize each measurement point vibrating at the same frequency, and to perform a Fourier transform on all of them at a common frequency {1.07 Hz, 1.07 Hz, 1.07 Hz, 1.07 Hz}. This embodiment is a technology for doing this.

[0087] (Processing of abnormality detection device) FIG. 6 is a diagram illustrating an example of processing performed by the abnormality detection device according to the third embodiment of the present disclosure. The abnormality diagnosis process of the abnormality detection device 2 will be described in detail below with reference to FIG.

[0088] For example, if there are four detection signals (four measurement points), the interlocking method is expressed by equation (10). 1 ,A 2 ,A 3 ,A 4 are column vectors of size 3×1.

[0089]

number

[0090] For example, one way of linking these measurement points is defined as ω 1 From ω 4 If any one of these is known, the others can be estimated. 1 From ω 2 ,ω 3 ,ω 4 The estimation of the frequency is expressed by the symbol {^ω 1 ,^ω 2 ,^ω 3 ,^ω 4}, the value is calculated by equation (11). Equation (11) is a vibration frequency model that represents the relationship that holds between characteristic frequencies that represent the vibration characteristics of each measurement point of the mechanical equipment 10.

[0091]

number

[0092] As shown in FIG. 5, the frequency detection unit 215 according to the present embodiment detects a detection signal z 1 and obtain the characteristic frequency ω 1 Then, the estimation unit 216 detects the ω 1 Then, using equation (11), the characteristic frequency ω 2 ,ω 3 ,ω4 Estimate.

[0093] In this way, the characteristic frequency ω 2 ,ω 3 ,ω 4 The estimated value of is the characteristic frequency ω 1 This makes it possible to eliminate the error of the frequency detection unit according to the second embodiment. n According to equation (12), the quantities for integral calculation (m n ,r n ) is determined so as to satisfy equation (13).

[0094]

number

[0095]

number

[0096] Then, the measurement value acquisition unit 210 calculates the detection signal y n Fourier coefficient a n and b n Calculate the Fourier coefficient a n and b n The processing after obtaining is the same as in the first and second embodiments.

[0097] (Action and effect) As described above, the abnormality detection device 2 according to the present embodiment uses a vibration frequency model and a feature frequency ω 1 Based on this, the characteristic frequency ω of each of the multiple measurement points of the mechanical equipment 10 is calculated. 2 ,ω 3 ,ω 4 and the measured value consisting of the amplitude and phase of the vibration at the estimated characteristic frequency (Fourier coefficient a n ,b n ) to get the

[0098] In this way, the abnormality detection device 2 can suppress calculation errors in the Fourier coefficients caused by measurement errors of the vibration sensor, etc. This allows the abnormality detection device 2 to further improve the accuracy of abnormality detection.

[0099] <Fourth embodiment> Next, an abnormality detection device 2 according to a fourth embodiment of the present disclosure will be described with reference to FIGS. Components common to the first to third embodiments are given the same reference numerals and detailed description thereof will be omitted.

[0100] The abnormality detection device 2 according to the first to third embodiments described above calculates the Mahalanobis distance of the measurement value consisting of the Fourier coefficients acquired by the measurement value acquisition unit 210, and determines that the mechanical equipment 10 is abnormal when this Mahalanobis distance exceeds a threshold value. However, the method for detecting an abnormality in the mechanical equipment 10 is not limited to the Mahalanobis distance. In this embodiment, a mode will be described in which the abnormality detection device 2 detects the presence or absence of an abnormality based on the components of the measurement value vector instead of the Mahalanobis distance.

[0101] (Functional configuration of the abnormality detection device) FIG. 7 is a diagram illustrating a functional configuration of an abnormality detection device according to the fourth embodiment of the present disclosure. As shown in FIG. 7, in this embodiment, the processor 21 of the abnormality detection device 2 has a component calculation unit 217 instead of the Mahalanobis distance calculation unit 212.

[0102] The component calculation unit 217 calculates a plurality of components obtained by decomposing a measurement value vector made up of the measurement values ​​acquired by the measurement value acquisition unit 210 in the direction of an index vector defined as an index of anomaly. Note that, in this embodiment, the index vector is a singular vector obtained by performing singular value decomposition on the unit space.

[0103] (Processing of abnormality detection device) FIG. 8 is a diagram illustrating an example of processing of the abnormality detection device according to the fourth embodiment of the present disclosure. The abnormality diagnosis process of the abnormality detection device 2 will be described in detail below with reference to FIG.

[0104] The measurement value acquisition unit 210 according to this embodiment performs the same process as in the first to third embodiments, and acquires the Fourier coefficient (measurement value) at a predetermined frequency point of each measurement point. In the example of Fig. 8, the measurement value acquisition unit 210 performs the same process as in the third embodiment.

[0105] The Fourier coefficients acquired by the measurement value acquisition unit 210 are represented as a vector, namely, a Fourier coefficient vector x (measurement value vector). Equation (14) is for the case where the number of Fourier coefficients is M, and the size of the Fourier coefficient vector x is M×1.

[0106]

number

[0107] The unit space generation unit 211 generates a unit space using the values ​​of the Fourier coefficient vector x in the normal state. The unit space is the covariance matrix Q of the Fourier coefficient vector x acquired when the mechanical equipment is in the normal state, and is expressed by the following formulas (15) and (16). Note that X in formula (15) 0 is a matrix in which multiple Fourier coefficient vectors x are arranged in the column direction. L is the number of Fourier coefficient vectors used to create the unit space, and X 0 is a complex matrix of size M × L. X in Eq. (16) 0 * isX 0 Represents the conjugate transpose of .

[0108]

number

[0109]

number

[0110] Here, it is assumed that the unit space generating section 211 has already generated a unit space based on the Fourier coefficient vector x acquired at a past point in time in this manner, and that this unit space has been stored in the auxiliary storage device 24.

[0111] In the first to third embodiments, the abnormality detection device 2 calculates the Mahalanobis distance MD at time t by the Mahalanobis distance calculation unit 212 using the following formula (17).

[0112]

number

[0113] In contrast, in the abnormality detection device 2 according to this embodiment, the component calculation unit 217 calculates the component ρ k (k=1, 2, ..., M). Specifically, the component calculation unit 217 calculates the component ρ k Calculate.

[0114] The unit space can be subjected to singular value decomposition as shown in the following equation (18).

[0115]

number

[0116] Here, M is the number of the first vibration sensors 11. k (k=1,2,…,M) is a singular vector, of size M×1, and generally a column vector with complex elements. σ k (k=1,2,…,M) are singular values ​​and non-negative real numbers.

[0117] The component calculation unit 217 converts the Fourier coefficient vector x into a singular vector u k Component ρ decomposed in the (k=1,2,…,M) direction k (k=1,2,...,M) is calculated using the following equation (19).

[0118]

number

[0119] In addition, the component calculation unit 217 according to this modification converts the Fourier coefficient vector x into a singular vector u k Component ρ decomposed in the direction k The component calculation unit 217 outputs the component ρ k The absolute value of the component ρ may be output to the decision unit 213. k The absolute value of is expressed by the following equation (20).

[0120]

number

[0121] The determination unit 213 converts the Fourier coefficient vector x into a singular vector u k Component ρ decomposed in the direction k The absolute value of and the singular vector u k The singular value σ corresponding to k A comparison process is performed in which each singular vector is individually compared with the square root of the Fourier coefficient vector x obtained when the mechanical equipment 10 is in a normal state. Since the unit space Q is a covariance matrix of the multiple Fourier coefficient vectors x obtained when the mechanical equipment 10 is in a normal state, the singular values ​​σ k is the component ρ when in normal state k Therefore, if the machine equipment 10 is normal at a certain time, ρ k ρ k * The expectation of is the singular value σ k In other words, if we express the expected value as E, we get Equation (21).

[0122]

number

[0123] And, ρ k ρ k * is a component of the Mahalanobis distance, as shown in equation (22).

[0124]

number

[0125] Therefore, the determination of anomalies using the Mahalanobis distance is determined by the sum of the components on the right side of equation (22). Below, we will explain a method of determining anomalies by checking values ​​at the component stage, before taking the sum.

[0126] Based on this premise, the determination unit 213 determines the component ρ k The absolute value of and the singular value σ k The square root of and the singular vector u k It compares each of the values ​​(k=1,2,...,M) and outputs the judgment result.

[0127] For example, the determination unit 213 determines whether the component ρ k The absolute value of is the singular value σ k If the difference is less than the threshold based on the component ρ k is equal to or greater than the threshold, the determination unit 213 outputs a determination result indicating an abnormality. Specifically, the determination unit 213 determines an abnormality if the following equation (23) is satisfied.

[0128]

number

[0129] Here, γ in equation (23) 1k ,γ 2k is the singular vector u k It is a positive constant defined for each k (k=1,2,…,M).

[0130] 7 and 8 show an example in which a component calculation unit 217 is added instead of the Mahalanobis distance calculation unit 212 of the third embodiment, but the present invention is not limited to this. In other embodiments, a component calculation unit 217 may be added instead of the Mahalanobis distance calculation unit 212 of the first or second embodiment.

[0131] (Action and effect) As described above, the abnormality detection device 2 according to this embodiment calculates multiple components ρ by decomposing the measurement value vector (Fourier coefficient vector x) consisting of measurements acquired by the measurement value acquisition unit 210 in the direction of the singular vector u obtained by singular value decomposition of the unit space, and determines that an abnormality has occurred in the mechanical equipment 10 when the absolute value of the calculated ρ exceeds a predetermined threshold value.

[0132] In this way, the abnormality detection device 2 can monitor the presence or absence of an abnormality for each component, rather than the Mahalanobis distance obtained by summing the components, thereby improving sensitivity to abnormalities. Also, as in the first to third embodiments, the abnormality detection device 2 only needs to calculate the Fourier coefficients at the characteristic frequencies of the mechanical equipment 10, making it possible to perform abnormality diagnosis at high speed without a separate frequency analysis device.

[0133] <Fifth embodiment> Next, an abnormality detection device 2 according to a fifth embodiment of the present disclosure will be described with reference to FIG. Components common to the first to fourth embodiments are given the same reference numerals and detailed description thereof will be omitted.

[0134] The abnormality detection devices 2 according to the above-described first to third embodiments calculate the Fourier coefficients for each first vibration sensor 11. That is, the abnormality detection device 2 according to the first embodiment calculates the Fourier coefficients by calculating formulas (3A) and (3B), and the abnormality detection devices 2 according to the second and third embodiments calculate the Fourier coefficients by calculating formulas (8A) and (8B), for each first vibration sensor 11. If this method is followed, the amount of calculation required to calculate the Fourier coefficients increases as the number of sensors (measurement points) increases.

[0135] In contrast, the anomaly detection device 2 according to this embodiment is characterized by using prior knowledge about anomalies to reduce the number of calculations of Fourier coefficients. This has practical advantages, such as reduced computer costs.

[0136] (Processing of abnormality detection device) FIG. 9 is a diagram illustrating an example of processing performed by the abnormality detection device according to the fifth embodiment of the present disclosure. The abnormality diagnosis process of the abnormality detection device 2 will be described in detail below with reference to FIG.

[0137] As shown in FIG. 9, the measurement value acquisition unit 210 according to this embodiment acquires a measurement value vector {y 1 ,y 2 ,…,y M} to get

[0138] The component calculation unit 217 calculates a singular vector u k is divided into a real part and an imaginary part using a phase angle θ of a predetermined frequency point (feature frequency ω), and an inner product of each of the real part and the imaginary part and the measurement value vector is calculated. Note that, although FIG. 9 shows an example in which the feature frequency ω estimated by the frequency detection unit 215 and the estimation unit 216 according to the third embodiment is input to the component calculation unit 217, the present invention is not limited to this. The component calculation unit 217 receives the period T n may be input, or the characteristic frequency ω detected by the frequency detection unit 215 according to the second embodiment may be input. n may be input.

[0139] The abnormality diagnosis process executed by the abnormality detection device 2 according to this embodiment is a further improvement of the process described in the fourth embodiment. In the fourth embodiment, the Fourier coefficient vector x is divided into singular vectors u k Component ρ decomposed in the (k=1,2,…,M) direction k Based on this, the abnormality was judged by equation (23). There are some abnormalities in the machine equipment 10 that occur frequently and some that occur infrequently. The types of abnormality are determined by the singular vector u k (k=1,2,...,M), among the M singular vectors, there are components that are likely to conflict with the judgment formula in formula (23) and those that are not. Then, it is reasonable in cases where the computing power is limited to judge whether or not only the components that are likely to conflict are judged as abnormal at every time step, and the rest are judged as abnormal after thinning them out, such as once every 100 time steps.

[0140] The present embodiment is a technology that enables such anomaly determination regarding a specific singular vector to be performed at high speed. 1 ,y 2 ,…,y M}, M Fourier transforms are required, but this embodiment is characterized in that only one is required. However, as described in the third embodiment, it is assumed that the feature frequencies at which M vibration features appear are related to each other, and the second to Mth feature frequencies can be known from the feature frequency of the first vibration. In the following, for simplicity of explanation, it is assumed that all feature frequencies are the same as shown in equation (24).

[0141]

number

[0142] Therefore, the estimation unit 216 estimates the characteristic frequency of each measurement point from the characteristic frequency ω1 detected by the frequency detection unit 215 using equation (24).

[0143] Then, the anomaly detection device 2 detects the k-th singular vector u k Here, k is one of the values ​​{1, 2, ..., M}.

[0144] The component calculation unit 217 calculates the phase angle θ of the characteristic frequency of each measurement point (first vibration sensor 11) using the formula (12) in the same manner as in the third embodiment. k is divided into real and imaginary parts, and the measured values ​​of each measurement point {y1, y2, ..., y M}. Here, the symbol [u k ] 1 is the singular vector u k It represents the first element of .

[0145]

number

[0146]

number

[0147] Next, the component calculation unit 217 time-averages the inner product over one period as shown in equation (26). Since the characteristic frequencies of the M vibrations are equal to each other as shown in equation (24), the length of one period is T 1 If there are different values ​​among the M vibration characteristic frequencies, the length of one period is taken as the least common multiple of the periods to calculate equation (26).

[0148]

number

[0149] ρ obtained from equation (26) k is applied to the above-mentioned formula (20). Furthermore, the determination unit 213 determines whether or not an abnormality exists by the above-mentioned formula (23).

[0150] In the calculation of equation (26), the past values ​​{α(t-ΔT), α(t-2ΔT), …, α(tm 1 ΔT)} and {β(t−ΔT),β(t−2ΔT),…,β(tm 1 The abnormality detection device 2 according to the fourth embodiment stores the values ​​of M×2m 1 In contrast, the abnormality detection device 2 according to the present embodiment requires a total of 2m 1 past values. Therefore, according to this embodiment, it is possible to further reduce the amount of past value storage and the number of product-sum operations. This makes it possible to incorporate the function of the anomaly detection device 2 into a normal control device and have it executed.

[0151] Note that equation (26) may be replaced with a first-order low-pass filter. This allows the number of past values ​​to be reduced to two. In this case, it takes T 1Therefore, in applications requiring a fast response, the method using the formula (26) described above is more effective.

[0152] (Action and effect) As described above, the anomaly detection device 2 according to this embodiment calculates the singular vector u k is divided into a real part and an imaginary part using the phase angle θ of the characteristic frequency ω, and the inner products of the real part and the imaginary part and the measurement value vector y are calculated, and the time average ρ of the calculated inner products for one period is calculated. k If the absolute value of exceeds a predetermined threshold, it is determined that an abnormality has occurred in the mechanical equipment 10.

[0153] In this way, the abnormality detection device 2 can reduce the number of Fourier transforms per diagnosis to one, thereby significantly reducing the calculation load. Also, the abnormality detection device 2 can reduce the number of past values ​​used in abnormality diagnosis, thereby reducing the storage capacity and the product-sum calculation load.

[0154] <Modification> In the fourth and fifth embodiments described above, the singular vector u k However, if the mode shape of abnormal vibration is known in advance for the mechanical equipment 10, it is possible to immediately determine an abnormality when the mode shape appears. This modification realizes this. Here, an example in which this modification is applied to the abnormality detection device 2 according to the fifth embodiment will be described.

[0155] In the abnormality detection device 2 according to this modification, the component calculation unit 217 calculates the singular vector u k In place of equation (25A) and equation (25B) of the fifth embodiment, a predetermined mode shape vector φ is used as an index vector. That is, equations (27A) and (27B) are used instead of equations (25A) and (25B) of the fifth embodiment. The mode shape vector φ can be obtained by utilizing a known technique such as modal analysis.

[0156]

number

[0157]

number

[0158] Then, the component calculation unit 217 calculates the component ρ of the mode shape vector in equation (28). k Calculate.

[0159]

number

[0160] In the formula (23) in which the determination unit 213 performs the abnormality determination, σ k is the normal α k and β k It is determined by equation (27).

[0161] (Action and effect) As described above, the abnormality detection device 2 according to this modified example uses the mode shape vector φ as an index vector.

[0162] In this way, the anomaly detection device 2 detects the singular vector u k Since the calculation for obtaining the above can be omitted, the calculation load can be further reduced.

[0163] In the first to fifth embodiments and their modifications, the various processes of the abnormality detection device 2 are stored in a computer-readable recording medium in the form of a program, and the computer reads and executes the program to perform the various processes. The computer-readable recording medium refers to a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, etc. The computer program may be distributed to a computer via a communication line, and the computer that receives the program may execute the program.

[0164] The program may be for implementing some of the above-mentioned functions, or may be a so-called differential file (differential program) that can implement the above-mentioned functions in combination with a program already recorded in the computer system.

[0165] In other embodiments, the control device 3 may have each of the functional units included in the abnormality detection device 2 described in the first to fifth embodiments and their modifications.

[0166] As described above, several embodiments of the present invention 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 spirit of the invention. These embodiments and their modifications are included in the spirit and technical scope of the invention.

[0167] <Additional Notes> The abnormality detection device, the abnormality detection method, and the program described in the above-described embodiments can be understood, for example, as follows.

[0168] (1) According to a first aspect, an abnormality detection device (2) for detecting the presence or absence of an abnormality in mechanical equipment (10) includes a processor (21) that executes the following processes: acquiring measurement values ​​(a, b) consisting of the amplitude and phase of vibration at a predetermined frequency point based on a detection signal (y) output from a first vibration sensor (11) that measures vibration of the mechanical equipment (10); calculating a Mahalanobis distance of the measurement values ​​acquired at the time when the mechanical equipment (10) is evaluated based on a unit space constituted by a plurality of measurement values ​​acquired at a plurality of past time points; and determining that an abnormality has occurred in the mechanical equipment (10) if the calculated Mahalanobis distance exceeds a predetermined threshold value.

[0169] In this way, the anomaly detection device only needs to calculate the Fourier coefficients (amplitude and phase) at specific frequency points, which can significantly reduce the calculation load compared to conventional techniques. This enables the anomaly detection device to quickly diagnose abnormalities in mechanical equipment. In addition, since the anomaly detection device can calculate the Fourier coefficients by itself without the need for a separate frequency analyzer, the cost of the entire anomaly detection system can be reduced.

[0170] (2) According to a second aspect, in the abnormality detection device (2) relating to the first aspect, the processor (21) further executes a process of detecting a characteristic frequency (ω) of the mechanical equipment (10) based on a detection signal output from a second vibration sensor (14) that measures vibrations of the mechanical equipment (10), and in a process of acquiring measurement values, sets the detected characteristic frequency ω as a frequency point to acquire measurement values ​​(a, b).

[0171] In this way, the abnormality detection device can automatically detect frequency components that change according to the operating conditions of the mechanical equipment, deterioration over time, etc. This allows the abnormality detection device to obtain more accurate Fourier coefficients, thereby further improving the accuracy of abnormality detection.

[0172] (3) According to a third aspect, in the abnormality detection device (2) according to the second aspect, the processor (21) calculates a vibration frequency model representing a relationship between characteristic frequencies at a plurality of measurement points of the mechanical equipment (10) and a characteristic frequency (ω 1 ) and the characteristic frequency (^ω n ) is further estimated, and in the process of acquiring the measurement values, the characteristic frequency (^ω n ) is set as the frequency point, and measurement values ​​(a, b) corresponding to each of the multiple measurement points are obtained.

[0173] In this way, the abnormality detection device can suppress calculation errors in the Fourier coefficients caused by measurement errors of the vibration sensor, etc. This allows the abnormality detection device to further improve the accuracy of abnormality detection.

[0174] (4) According to a fourth aspect, an abnormality detection device (2) for detecting the presence or absence of an abnormality in mechanical equipment (10) includes a process for acquiring measurement values ​​(a, b) consisting of amplitude and phase of vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor (11) that measures vibration of the mechanical equipment (10), and a process for converting a measurement value vector consisting of the measurement values ​​into an index vector (u) defined as an index of the abnormality. k or φ) k ) and the calculated component (ρ k If the absolute value of the signal strength exceeds a predetermined threshold, the processor determines that an abnormality has occurred in the mechanical equipment (10).

[0175] In this way, the abnormality detection device can monitor the presence or absence of an abnormality for each component, thereby improving sensitivity to abnormalities. Also, as in the first to third aspects, the abnormality detection device only needs to calculate the Fourier coefficients at the characteristic frequencies of the mechanical equipment, making it possible to perform abnormality diagnosis at high speed without a separate frequency analysis device.

[0176] (5) According to a fifth aspect, an abnormality detection device (2) for detecting the presence or absence of an abnormality in a mechanical equipment (10) includes a process of acquiring a measurement value vector (y) consisting of measurement values ​​measured by a first vibration sensor (11) that measures vibrations of the mechanical equipment (10), and a process of acquiring an index vector (u) defined as an index of the abnormality. k or φ) is divided into a real part and an imaginary part using a phase angle (θ) of a predetermined frequency point, and an inner product of each of the real part and the imaginary part and the measurement value vector (y) is calculated. k If the absolute value of the signal strength exceeds a predetermined threshold, the processor determines that an abnormality has occurred in the mechanical equipment (10).

[0177] In this way, the anomaly detection device can reduce the number of Fourier transforms per diagnosis to one, thereby significantly reducing the computational load. Also, the anomaly detection device can reduce the number of past values ​​used in anomaly diagnosis, thereby reducing the storage capacity and the load of product-sum calculations.

[0178] (6) According to a sixth aspect, in the abnormality detection device (2) according to the fourth or fifth aspect, the processor (21) calculates a singular vector (u k ) is used as the index vector.

[0179] In this way, the anomaly detection device can monitor the presence or absence of anomalies for each component, rather than the Mahalanobis distance obtained by summing the components, thereby improving sensitivity to anomalies.

[0180] (7) According to a seventh aspect, in the abnormality detection device (2) according to the fourth or fifth aspect, the processor (21) uses a mode shape vector (φ) specific to the mechanical equipment (10) as an index vector.

[0181] In this way, the abnormality detection device 2 can omit the calculation for obtaining the singular vectors, and therefore the computational load can be further reduced.

[0182] (8) According to the eighth aspect, an abnormality detection method for detecting the presence or absence of an abnormality in mechanical equipment (10) includes the steps of acquiring measurement values ​​(a, b) consisting of the amplitude and phase of vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor (11) that measures vibration of the mechanical equipment (10), calculating a Mahalanobis distance of the measurement value acquired at the time when the mechanical equipment (10) is evaluated based on a unit space constituted by multiple measurement values ​​acquired at multiple past time points, and determining that an abnormality has occurred in the mechanical equipment (10) if the calculated Mahalanobis distance exceeds a predetermined threshold value.

[0183] (9) According to the ninth aspect, the program causes a computer of an abnormality detection device (2) that detects the presence or absence of an abnormality in the mechanical equipment (10) to execute the steps of acquiring measurement values ​​(a, b) consisting of the amplitude and phase of the vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor (11) that measures the vibration of the mechanical equipment (10); calculating a Mahalanobis distance of the measurement values ​​acquired at the time when the mechanical equipment (10) is evaluated based on a unit space constituted by multiple measurement values ​​acquired at multiple past time points; and determining that an abnormality has occurred in the mechanical equipment (10) if the calculated Mahalanobis distance exceeds a predetermined threshold value. [Explanation of symbols]

[0184] 100 Anomaly Detection System 10. Machinery and Equipment 11, 11A, 11B, 11C, 11D First vibration sensor 12 Power plant 13 Gear 14 Second vibration sensor 2. Anomaly detection device 21 Processors 210 Measurement value acquisition unit 211 Unit Space Generation Unit 212 Mahalanobis distance calculation unit 213 Judgment section 214 Frequency point designation section 215 Frequency detection section 216 Estimation Department 217 Component Calculation Unit 22 Input / output section 23 Main memory 24 Auxiliary storage 3. Control device

Claims

1. An abnormality detection device for detecting the presence or absence of an abnormality in mechanical equipment, A process of acquiring measurement values ​​consisting of amplitude and phase of the vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; A process of calculating a Mahalanobis distance of the measurement value acquired at the time when the mechanical equipment is evaluated, based on a unit space configured by the plurality of measurement values ​​acquired at a plurality of past time points; a process of determining that an abnormality has occurred in the mechanical equipment when the calculated Mahalanobis distance exceeds a predetermined threshold value; a processor for executing The processor, Further performing a process of detecting a characteristic frequency of the mechanical equipment based on a detection signal output from a second vibration sensor that measures vibration of the mechanical equipment; Further executing a process of estimating a characteristic frequency of each of the plurality of measurement points on the mechanical equipment based on a vibration frequency model representing a relationship between characteristic frequencies at the plurality of measurement points on the mechanical equipment and the characteristic frequency of one measurement point on the mechanical equipment detected in the process of detecting the characteristic frequency, In the process of acquiring the measurement values, the feature frequency of each of the estimated plurality of measurement points is set to the frequency point, and the measurement values ​​corresponding to each of the plurality of measurement points are acquired. Anomaly detection device.

2. An abnormality detection device for detecting the presence or absence of an abnormality in mechanical equipment, A process of acquiring measurement values ​​consisting of amplitude and phase of the vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; A process of calculating a plurality of components obtained by decomposing a measurement value vector made up of the measurement values ​​in a direction of an index vector defined as an index of an abnormality; a process of determining that an abnormality has occurred in the mechanical equipment when the absolute value of the calculated component exceeds a predetermined threshold value; a processor for executing The processor, Further performing a process of detecting a characteristic frequency of the mechanical equipment based on a detection signal output from a second vibration sensor that measures vibration of the mechanical equipment; Further executing a process of estimating a characteristic frequency of each of the plurality of measurement points on the mechanical equipment based on a vibration frequency model representing a relationship between characteristic frequencies at the plurality of measurement points on the mechanical equipment and the characteristic frequency of one measurement point on the mechanical equipment detected in the process of detecting the characteristic frequency, In the process of acquiring the measurement values, the feature frequency of each of the estimated plurality of measurement points is set to the frequency point, and the measurement values ​​corresponding to each of the plurality of measurement points are acquired. Anomaly detection device.

3. An abnormality detection device for detecting the presence or absence of an abnormality in mechanical equipment, A process of acquiring a measurement value vector consisting of measurement values ​​measured by a first vibration sensor that measures vibrations of the mechanical equipment; A process of dividing an index vector defined as an index of an abnormality into a real part and an imaginary part using a phase angle of a predetermined frequency point, and calculating an inner product of each of the real part and the imaginary part with the measurement value vector; a process of determining that an abnormality has occurred in the mechanical equipment when an absolute value of a time average of the calculated inner product for one period exceeds a predetermined threshold value; a processor for executing Anomaly detection device.

4. The processor uses, as the index vector, a singular vector obtained from a unit space constituted by a plurality of the measurement values ​​acquired at a plurality of past time points. The abnormality detection device according to claim 2 or 3.

5. the processor uses a mode shape vector specific to the machinery as the index vector. The abnormality detection device according to claim 3 .

6. An abnormality detection method for detecting the presence or absence of an abnormality in mechanical equipment, comprising: acquiring measurement values ​​consisting of amplitude and phase of the vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; A step of calculating a Mahalanobis distance of the measurement value acquired at a time point when the mechanical equipment is evaluated, based on a unit space configured by the plurality of measurement values ​​acquired at a plurality of past time points; determining that an abnormality has occurred in the mechanical equipment when the calculated Mahalanobis distance exceeds a predetermined threshold value; Detecting a characteristic frequency of the mechanical equipment based on a detection signal output from a second vibration sensor that measures vibration of the mechanical equipment; A step of estimating a characteristic frequency of each of the plurality of measurement points on the mechanical equipment based on a vibration frequency model representing a relationship between characteristic frequencies at the plurality of measurement points on the mechanical equipment and the characteristic frequency of one measurement point on the mechanical equipment detected in a process of detecting the characteristic frequency; having In the step of acquiring the measurement values, the feature frequency of each of the estimated plurality of measurement points is set to the frequency point, and the measurement values ​​corresponding to each of the plurality of measurement points are acquired. Anomaly detection methods.

7. The computer of the abnormality detection device that detects the presence or absence of abnormalities in the mechanical equipment is acquiring measurement values ​​consisting of amplitude and phase of the vibration at a predetermined frequency point based on a detection signal output from a first vibration sensor that measures vibration of the mechanical equipment; A step of calculating a Mahalanobis distance of the measurement value acquired at a time point when the mechanical equipment is evaluated, based on a unit space configured by the plurality of measurement values ​​acquired at a plurality of past time points; determining that an abnormality has occurred in the mechanical equipment when the calculated Mahalanobis distance exceeds a predetermined threshold value; Detecting a characteristic frequency of the mechanical equipment based on a detection signal output from a second vibration sensor that measures vibration of the mechanical equipment; A step of estimating a characteristic frequency of each of the plurality of measurement points on the mechanical equipment based on a vibration frequency model representing a relationship between characteristic frequencies at the plurality of measurement points on the mechanical equipment and the characteristic frequency of one measurement point on the mechanical equipment detected in a process of detecting the characteristic frequency; Run the command, In the step of acquiring the measurement values, the feature frequency of each of the estimated plurality of measurement points is set to the frequency point, and the measurement values ​​corresponding to each of the plurality of measurement points are acquired. program.

Images