Abnormality diagnosis device and abnormality diagnosis method

Abstract

An abnormality diagnosis device (100), which is for determining whether the tension of a belt (61) for transmitting power from an electric motor (5) to load equipment (7) is abnormal, performs FET analysis of the current of the electric motor (5) which has been continuously detected in a fixed period of time, calculates, from a spectral peak, a specific frequency which occurs due to a rotation frequency of the belt, and determines whether the tension of the belt is abnormal using a change in signal intensity at the specific frequency within the fixed period of time.

Description

Abnormality diagnosis device and abnormality diagnosis method

[0001] The present disclosure relates to an abnormality diagnosis device and an abnormality diagnosis method.

[0002] Some industrial machinery transmits the rotational power of a motor to a shaft via a power transmission mechanism such as a belt. If the belt tension deviates from the appropriate operating conditions, the belt's lifespan will be shortened or the shaft will not be able to be driven normally. Furthermore, if the belt tension decreases while the industrial machinery is operating, problems such as belt slippage and tooth skipping will occur. Therefore, inspection and maintenance work are required to ensure that the belt tension is maintained within the specified range.

[0003] In response to this, a condition determination device has been disclosed that analyzes the frequency characteristics of feedback data obtained from diagnostic operations, detects frequencies at multiple extreme points in the frequency characteristics as resonant frequencies or anti-resonant frequencies, and determines the state of belt tension (see, for example, Patent Document 1).

[0004] Patent No. 7011106

[0005] In the technology disclosed in Patent Document 1, frequencies at which the signal strength reaches a maximum or minimum are considered to be resonant frequencies, and abnormalities are diagnosed by observing the maximum and minimum fluctuations of these frequencies. In Patent Document 1, when frequency conversion is applied to the motor current detected by the sensor, multiple peaks that are maximum values ​​occur due to the power supply frequency and vibration. Because it is difficult to distinguish these peaks from the resonant frequency, when this method is applied to frequency analysis of current, there is a problem in that the resonant frequency cannot be identified.

[0006] The present disclosure discloses a technique for solving the above-mentioned problems, and aims to provide an abnormality diagnosis device and an abnormality diagnosis method that are capable of diagnosing the condition of a belt without relying on the maximum and minimum of the power supply frequency and other frequencies caused by vibration.

[0007] The abnormality diagnosis device disclosed herein is an abnormality diagnosis device that determines an abnormality in tension of a belt, which is a power transmission mechanism that transmits power from an electric motor to load equipment, and includes: a current detector that detects a current of the electric motor; and an abnormality diagnosis unit that determines an abnormality by performing an FFT analysis on the current detected by the current detector and using the extracted spectral peaks, wherein the abnormality diagnosis unit has: a spectral analysis unit that performs an FFT analysis on each of a plurality of current values ​​detected consecutively within a certain period of time to extract spectral peaks; a specific frequency extraction unit that extracts, as a specific frequency, the spectral peak caused by the rotation frequency of the belt from the power supply frequency of the electric motor and the frequency of a sideband wave of the power supply frequency; an amplitude calculation unit that calculates an amplitude-related value including an amplitude from a change in signal strength of the specific frequency within the certain period of time; and an abnormality determination unit that determines an abnormality in tension of the belt based on the calculated amplitude-related value of the signal strength of the specific frequency.

[0008] According to the present disclosure, it is possible to accurately diagnose abnormalities in belt tension.

[0009] 1 is a diagram for explaining a schematic configuration of an abnormality diagnosis device according to embodiment 1. FIG. 2 is a diagram showing an example of the configuration of a control device that drives an electric motor according to embodiment 1. FIG. 3 is a block diagram showing the configuration of an abnormality diagnosis unit according to embodiment 1. FIG. 4 is a diagram showing a spectral waveform that is an example of an analysis result by current FFT under no load. FIG. 5 is a diagram showing a spectral waveform that is an example of an analysis result by current FFT when a belt with appropriate tension is used. FIG. 6 is a diagram showing the time dependency of the signal strength of the −2nd order component of the belt rotation frequency when the belt is at appropriate tension. FIG. 7 is a diagram showing the time dependency of the signal strength of the −2nd order component of the belt rotation frequency when the belt is in a slack state. FIG. 8 is a diagram showing the relationship between the variance of the signal strength of the −2nd order component of the belt rotation frequency and the belt tension. FIG. 9 is a diagram showing the relationship between the signal strength of a specific frequency and the belt tension. FIG. 10 is a diagram showing the relationship between the variance of the signal strength of the −1st order component of the belt rotation frequency and the belt tension. FIG. 11 is a flowchart showing an overall procedure for performing abnormality diagnosis using the abnormality diagnosis device according to embodiment 1. FIG. 12 is a diagram showing another schematic configuration of the abnormality diagnosis device according to embodiment 1. FIG. 13 is a block diagram showing the configuration of another abnormality diagnosis unit according to embodiment 1. FIG. 14 is a diagram showing the schematic configuration of an abnormality diagnosis device according to embodiment 2. FIG. 1 is a diagram showing the time dependency of the signal strength of the −2nd order component of the belt rotation frequency when the belt begins to loosen. FIG. 2 is a diagram for explaining an abnormality diagnosis method using machine learning in an abnormality diagnosis device according to embodiment 2. FIG. 3 is a diagram showing an example of the hardware configuration of the abnormality diagnosis devices according to embodiments 1 and 2. FIG. 4 is a diagram showing another example of the hardware configuration of the abnormality diagnosis devices according to embodiments 1 and 2.

[0010] The present embodiment will be described below with reference to the drawings. In the drawings, the same reference numerals indicate the same or corresponding parts.

[0011] Embodiment 1. An abnormality diagnosis device according to Embodiment 1 will now be described with reference to the drawings. FIG. 1 is a diagram illustrating the schematic configuration of the abnormality diagnosis device according to Embodiment 1. The abnormality diagnosis results by the abnormality diagnosis device 100 are transmitted to a monitoring and control device 200 in, for example, a plant. The monitoring and control device 200 identifies abnormalities in multiple load devices 10 in the plant and reflects the results in maintenance and maintenance plans. The abnormality diagnosis device 100 detects abnormalities in the load devices 10 and performs abnormality diagnosis. The load devices 10 include an electric motor 5, a power transmission mechanism 6 that transmits power from the electric motor 5 to load equipment 7, and the load equipment 7. The abnormality diagnosis device 100 according to Embodiment 1 detects abnormalities in the power transmission mechanism 6 and performs abnormality diagnosis. In the drawing, the electric motor 5 is an example of one of many motors used in, for example, a plant, and is connected to a power supply 2 for driving the electric motor via a power supply line via an electric motor control device 110.

[0012] <Configuration of Abnormality Diagnosis Device 100> The abnormality diagnosis device 100 includes a motor control device 110, an abnormality diagnosis unit 130, an output unit 140, and a current detector 120 connected to one of the three-phase power lines connected to the motor 5. The abnormality diagnosis device 100 may be mounted in a motor control center that manages a large number of electric motors installed in a plant or the like, or may be a motor diagnosis device provided separately from the motor control center. The motor control device 110 may be provided in the abnormality diagnosis device 100, or may be provided independently of the abnormality diagnosis device 100.

[0013] The current detector 120 may be, for example, a clamp-type current transformer (CT) and installed on each phase of the three-phase power line. However, it is sufficient to measure only one of the phases. Furthermore, the installation location of the current detector 120 is not limited as long as it can measure the drive current of the electric motor 5. This indicates that the detection accuracy does not change depending on the measurement location. In this embodiment, an example of detecting a u-phase current will be described. Furthermore, if current waveform acquisition and abnormality diagnosis are performed continuously, memory pressure or an overload of the arithmetic processing circuit due to abnormality diagnosis may occur. If the target equipment is not expected to experience a sudden abnormality, data should not be acquired continuously, but should be acquired appropriately at the timing of diagnosis. During steady-state operation, current data should be acquired once per hour.

[0014] The output unit 140 displays the diagnosis results from the abnormality diagnosis unit 130, or outputs an audible or visual alarm based on the diagnosis results to notify of abnormalities in the motors and power transmission mechanisms. Furthermore, when the monitoring and control device 200 monitors the entire plant and collects abnormal conditions, it transmits them using the communication function.

[0015] 2 is a diagram for explaining the configuration of the motor control device 110. The motor control device 110 is provided with an inverter (power conversion device) 111 and a control unit 112 for driving the inverter 111. For example, if the inverter 111 is configured with semiconductor switching elements, the semiconductor switching elements of the inverter 111 are driven and controlled so as to perform power conversion using a PWM (Pulse Width Modulation) method or the like from a carrier wave and a rectangular wave generated by the control unit 112. Here, the frequency of the carrier wave that is the basis for driving the inverter 111 is the power supply frequency f of the motor. ｓ The electric power converted by the inverter 111 is supplied to the electric motor 5. That is, the electric motor 5 is driven and controlled by the inverter 111.

[0016] The power transmission mechanism 6 is composed of, for example, an electric motor side pulley Pu1 connected to the rotating shaft of the electric motor 5 and a load equipment side pulley Pu2 connected to the drive shaft of the load equipment 7, with a power transmission member, for example, a belt 61, wrapped around them.

[0017] In this embodiment, there are no particular limitations on the type of the electric motor 5 or the load equipment 7. For example, the electric motor may be a single-phase or three-phase induction motor or a synchronous motor, and the load equipment may be a pump or a fan. The power transmission mechanism is configured to include at least a belt 61.

[0018] <Configuration of Abnormality Diagnosis Unit 130> Next, a description will be given of the configuration of the abnormality diagnosis unit 130. Fig. 3 is a block diagram showing the configuration of the abnormality diagnosis unit 130 of the abnormality diagnosis device 100 according to embodiment 1. In Fig. 3, the abnormality diagnosis unit 130 includes a load device setting unit 131, a storage unit 132, and a calculation unit 133.

[0019] The load device setting unit 131 is used to set information about the electric motor 5, information about the power transmission mechanism 6, and information about the load equipment 7. The load device setting unit 131 is used to acquire specifications of the electric motor 5, such as the power supply frequency, the number of poles, and the rated rotation speed, from information on a nameplate attached to the electric motor 5.

[0020] Furthermore, if the power transmission mechanism 6 is a belt, the load device setting unit 131 sets information according to the type of belt, etc. Furthermore, the load device setting unit 131 sets information according to the type of load equipment 7. For example, if the load equipment 7 is a fan, the load device setting unit 131 sets conditions such as the rotation speed of the fan.

[0021] The specification information of the electric motor 5, the information of the power transmission mechanism, and the information of the load equipment 7 are stored in the storage unit 132. The storage unit 132 also stores the drive current of the electric motor 5 acquired by the current detector 120, a threshold value used for abnormality determination (described later), and the like.

[0022] The calculation unit 133 includes a current data division unit 133a, a spectrum analysis unit 133b, a specific frequency extraction unit 133c, an amplitude calculation unit 133d, and an abnormality determination unit 133e.

[0023] The current data division unit 133a divides the current data acquired from the current detector 120 during diagnosis into a predetermined number of measurement times. The measurement time per measurement is determined based on the number of divisions of the current data. In this example, the number of divisions is set to 60, and the measurement time is set to 60 seconds. As a result, the time corresponding to one division is 1 second. In this embodiment, attention is focused on the amplitude of the signal strength of a specific frequency, as described below. Therefore, if the time period of the divided data used for frequency conversion is long, the signal strength of the specific frequency is averaged over the measurement time, and the amplitude of the signal strength is also averaged accordingly, making it undetectable. On the other hand, if the time period of the divided data is too short, the time resolution deteriorates when the frequency is converted. For these reasons, it is necessary to select an appropriate time period for the data used for frequency conversion. One selection method involves calculating the signal strength of the power supply frequency, the rotational frequencies of the electric motor, the load, and the belt, as described below, including harmonics, and setting a value that allows each calculated frequency to be distinguished, in order to prevent interference between these frequencies.

[0024] The spectrum analysis unit 133b performs a current FFT (Fast Fourier Transform) analysis (frequency transformation) using the current data divided by the current data division unit 133a.

[0025] The specific frequency extraction unit 133c detects all spectral peaks from the spectral waveform analyzed by the spectrum analysis unit 133b. The frequency range to be detected is preferably between 0 and 1000 Hz. Next, the belt rotation frequency is calculated, and spectral peaks of specific frequencies that satisfy the belt rotation frequency condition are extracted from the detected spectral peaks.

[0026] The amplitude calculation unit 133d calculates the amplitude of the signal strength of the belt rotation frequency from the time dependency of the signal strength of the spectrum peak that satisfies the conditions of the belt rotation frequency extracted by the specific frequency extraction unit 133c.

[0027] The abnormality determination unit 133e includes an amplitude evaluation unit 133e1, and determines whether or not an abnormality exists by comparing the amplitude value calculated by the amplitude calculation unit 133d with the threshold value, which is stored in advance in the storage unit 132. The determination result is stored in the storage unit 132 and is also output to the output unit 140 of the abnormality diagnosis device 100.

[0028] The abnormality diagnosis device 100 according to the first embodiment will be described in detail using an example configuration in which the load equipment 7 is a fan, the power transmission mechanism 6 is the belt 61 shown in FIG. 2, and the electric motor 5 is a three-phase induction machine. In a load device 10 using a belt as the power transmission mechanism 6, the tension of the belt gradually decreases when the load device 10 is driven for a long period of time, and the belt must be replaced at an appropriate time. Given the increasing demand for automating the timing of replacement, this example of abnormal belt tension is important, and below, spectral peak analysis, abnormality determination, etc. will also be described for abnormal belt tension in the power transmission mechanism 6.

[0029] <Spectral Peak Analysis and Abnormality Determination in Calculation Unit 133> Next, a description will be given of a method for analyzing spectral peaks in the calculation unit 133. As described above, the spectral analysis unit 133b performs a current FFT analysis using the current of the electric motor 5 detected by the current detector 120 and divided by the current data division unit 133a.

[0030] The spectral waveform analyzed by the spectral analysis unit 133b includes spectral peaks attributable to the motor 5, the belt 61, and the fan that constitutes the load device 10.

[0031] Next, the spectral waveform analyzed by the spectrum analysis unit 133b will be described. ｍ Specifically, first, let us consider the provisional rotation frequency f when the slip s is set to 0. ｍ０ Calculate the rotation frequency f ｍ０ is the power frequency f ｓ , is expressed as the following equation (1) using the number of poles p. ｍ０ = f ｓ / p (1) where the actual rotation frequency f ｍis expressed as the following equation (2) using slip s: ｍ = f ｓ (1-s) / p ... (2) From equations (1) and (2), the slip s is s = 1-f ｍ / f ｍ０ ... (3)

[0032] Here, the actual rotation frequency f ｍ is the tentative rotation frequency f ｍ０ near the power supply frequency f ｓ Therefore, the actual rotation frequency f ｍ This actual rotation frequency f ｍ and the tentative rotation frequency f calculated using equation (1) ｍ０ However, this method requires sufficient frequency resolution to calculate the slip, so the data acquisition time must be set accordingly long.

[0033] The specific frequency extraction unit 133c detects all spectral peaks from the spectral waveform analyzed by the spectral analysis unit 133b. ｓ 4 shows an example of a spectral waveform obtained by current FFT of the u-phase current of the electric motor 5 connected to a belt, which is a power transmission mechanism 6, when the electric motor 5 is at no-load torque. ｓ A large peak is observed at , and multiple sidebands are observed around the peak. The frequency of the sidebands is the power supply frequency f ｓ , each rotation frequency is f ｉ When the order of the harmonic is a natural number n, it is expressed as the following equation (4): f = f ｓ ±nf ｉ ...(4)

[0034] As described above, the spectrum includes the spectrum peaks caused by the electric motor 5, the belt 61, and the fan which is the load equipment 7. The rotation frequency f ｍis expressed by the above-mentioned formula (2), and the rotation frequency f ｌ and the rotation frequency f due to the belt 61 of the power transmission mechanism 6 ｂ are expressed as the following equations (5) and (6), respectively. ｌ = f ｓ (1-s) / pr...(5) f ｂ = 2πr ｍ f ｍ / l ｂ = 2πr ｌ f ｌ / l ｂ ...(6) where r: reduction ratio r ｍ : Diameter of motor side pulley Pu1 r ｌ : Diameter of the load equipment side pulley Pu2 l ｂ : Belt length.

[0035] In this way, the rotation frequency f ｍ , the rotation frequency caused by the fan f ｌ and the rotation frequency f due to the belt 61 ｂ can be calculated from the shapes and drive information of the electric motor 5, the fan, and the belt 61. The specific frequency extraction unit 133c extracts a spectrum peak due to the belt rotation frequency from the extracted sideband wave, and calculates its signal intensity.

[0036] In this driving, the slip is a positive value, so the rotation frequency f due to the motor 5 is calculated from equations (2), (5), and (6). ｍ , the rotation frequency caused by the fan f ｌ and the rotation frequency f due to the belt 61 ｂ Each rotation frequency is calculated from equation (4) as f ｉ When the slip becomes smaller, the specific frequency f shifts inward to approach the power supply frequency fs. Based on this, the temporary frequency value f calculated assuming the slip is 0 is ０ Therefore, if the slip takes the maximum value in the inner frequency band within the range of values ​​that the slip can take, the frequency can be considered as a specific frequency.

[0037] In FIG.ｓ The rotation frequency f caused by the belt 61 is ｂ The sidebands of the belt rotation frequency are shown from the first to third orders for both positive and negative. That is, the specific frequency f, which is the frequency of the sidebands of the belt rotation frequency, is f = f ｓ ±nf ｂ (n is 1 to 3), and a spectral peak is shown at a specific frequency f. FIG. 5 focuses on the spectral peak of the -2nd-order component of the belt rotation frequency in FIG. 4, and shows the change over time in the signal strength of the -2nd-order component of the belt rotation frequency when the belt is at an appropriate tension. In FIG. 5, for the -2nd-order spectral peak of the belt rotation frequency indicated by an ellipse, it can be seen that the signal strength of the spectral peak confirmed at measurement time t=10 s becomes smaller at measurement time t=20 s. This signal strength continues to change after t=20 s, suggesting that the spectral peak is oscillating.

[0038] FIG. 6 shows the time dependence of the signal strength of the -2nd-order component of the belt rotation frequency when the belt is at the appropriate tension, and FIG. 7 shows the time dependence of the signal strength of the -2nd-order component of the belt rotation frequency when the belt is in a slack state. From FIG. 6 , it can be seen that when the belt is at the appropriate tension, the signal strength of the -2nd-order component of the belt rotation frequency oscillates between -50 dB and -40 dB over a measurement time of 60 seconds, with an amplitude of approximately 10 dB. On the other hand, from FIG. 7 , it can be seen that when the belt is in a slack state, the signal strength of the -2nd-order component of the belt rotation frequency remains almost constant over a measurement time of 60 seconds, with almost no oscillation. These findings indicate that the amplitude of the time-varying signal strength of the -2nd-order component of the belt rotation frequency changes depending on the tension, and that this amplitude is an effective parameter for diagnosing abnormalities.

[0039] Next, we will explain the principle behind the occurrence of peaks at the rotation frequency and the meaning of the amplitude in the time change of the signal strength of the belt rotation frequency. The principle behind the occurrence of peaks at the rotation frequency position when an abnormality occurs is as follows: When the rotor of the electric motor vibrates, the gap between the rotor and stator changes periodically, causing a change in the magnetic flux density between the rotor and stator. The change in magnetic flux density causes a change in the current, and this change is reflected in the sideband waves. As this periodic change is caused by the rotation of the electric motor and belt, a peak occurs at the rotation frequency position.

[0040] In particular, when an electric motor is connected by a belt, the motor vibrates due to belt tension, resulting in a peak vibration at the rotational frequency caused by the belt. The absence of this peak vibration means that the belt tension, which is the cause of the motor vibration, is not being transmitted. This is consistent with the results of the time dependence of the signal strength of the -2nd order component of the belt rotational frequency when the belt is at the appropriate tension in Figure 7 and the time dependence of the signal strength of the -2nd order component of the belt rotational frequency when the belt is loose in Figure 8. Therefore, belt abnormality diagnosis using the amplitude of the change in signal strength of the belt rotational frequency over time is an effective diagnostic method, particularly for belt tension.

[0041] The amplitude calculation unit 133d uses the signal intensity of the spectrum peak of the specific frequency that satisfies the conditions of the belt rotation frequency extracted by the specific frequency extraction unit 133c, and calculates the amplitude of the signal intensity of the belt rotation frequency from the time dependence of the signal intensity shown in Figures 7 and 8.

[0042] Here, we will explain how to calculate the amplitude used in abnormality diagnosis. (1) Calculating the amplitude: (1-1) Take the difference between the maximum and minimum of the signal strength of interest within the measurement time to calculate the "amplitude." If the amplitude fluctuates greatly within the measurement time, it may be impossible to calculate an accurate amplitude. (1-2) Calculate the average difference between the maximum and minimum of each period of the peak vibration. For example, in Figure 6, there are two periods of waveform, so it is sufficient to average the differences between the maximum and minimum of each period. (2) Calculating the amplitude variation: (2-1) Calculate the variance of the signal strength of interest within the measurement time. (2-2) Calculate the standard deviation of the signal strength of interest within the measurement time. Any calculation method is acceptable as long as it can calculate a quantity equivalent to the amplitude of the signal strength waveform. That is, the "amplitude" itself or a quantity equivalent to the "amplitude" may be used. These quantities related to the "amplitude," including the "amplitude," are called "amplitude-related values" and are used as evaluation quantities for abnormality diagnosis.

[0043] Next, an example will be described in which the "variance of signal strength" in (2-1) above is used as an evaluation quantity for performing abnormality diagnosis. Fig. 8 is a diagram showing the relationship between the variance of the signal strength of the -2nd order component of the belt rotation frequency and the belt tension, and Fig. 9 is a diagram showing the relationship between the signal strength of the -2nd order component of the belt rotation frequency and the belt tension.

[0044] In Figure 8, when the belt tension decreases from the appropriate tension and becomes less than 50%, the variance value of the signal strength of the -2nd order component of the belt rotation frequency drops sharply. On the other hand, no significant change is observed in the signal strength of the -2nd order component of the belt rotation frequency in Figure 9, making it difficult to detect abnormalities using a method that focuses on signal strength. Therefore, using the variance value of the signal strength of the -2nd order component of the belt rotation frequency as in Figure 8 allows for more accurate abnormality diagnosis.

[0045] Next, a method for determining the threshold for determining whether or not an abnormality exists will be described. As shown in FIG. 8 , the threshold can be set in advance so that the region (indicated by the ellipse) where the variance value drops sharply when the belt tension decreases from the appropriate tension can be detected. For example, the variance σ1 of the signal strength when the belt is at the appropriate tension is acquired multiple times, and the average σ1_ave and σ1_std of the acquired variances σ1 are calculated. Using an arbitrary coefficient a that satisfies a > 0, the number of times the variance σ1 is acquired is n, and where n satisfies 1≦n≦m, the nth variance is σ1_n. The threshold can be determined using the following equations (7) to (9).

[0046] threshold=σ1_ave−a·σ1_std (9) The threshold value "threshold" may be stored in the storage unit 132 and used during abnormality diagnosis.

[0047] In Figure 8, the threshold value is indicated by a dashed line. When the belt tension falls below 50% of the appropriate tension, it becomes possible to detect a tension abnormality. The amplitude evaluation unit 133e1 determines that the belt tension is abnormal when the amplitude variance value is less than the threshold value, and stores the result in the storage unit 132. The abnormality diagnosis unit 130 also outputs the diagnosis result to the output unit 140.

[0048] In the above example, an example was described in which "variance of signal strength of the -2nd order component of the belt rotation frequency" was used as an evaluation quantity for abnormality diagnosis. However, an example will now be described in which "variance of signal strength of the -1st order component of the belt rotation frequency" is used as an evaluation quantity using another specific frequency, the -1st order component of the belt rotation frequency.

[0049] Fig. 10 is a diagram showing the relationship between the variance of the signal strength of the -1st-order component of the belt rotation frequency and the belt tension. In Fig. 10, the variance of the signal strength of the -1st-order component of the belt rotation frequency decreases rapidly when the tension reaches about 70% of the appropriate tension. When the variance of the signal strength of the -2nd-order component was used, it was possible to detect a belt abnormality when the tension fell below 50%, so by focusing on the variance of the signal strength of the -1st-order component, it is possible to detect a belt abnormality at an earlier stage.

[0050] By changing the specific frequency to be focused on in this way, it is possible to change the state of belt tension to be detected. Also, if it is unclear which specific frequency has the greatest effect on tension, multiple specific frequencies can be used. Of these frequencies, the frequencies that fall below the threshold are counted, and the more the count increases, the more the tension abnormality is considered to be progressing.

[0051] In this way, when evaluating belt abnormalities based on the variance of signal strengths at multiple specific frequencies, the detected belt tension, i.e., the threshold, changes depending on the specific frequency. Therefore, by determining whether or not each specific frequency is below the threshold, it is possible to grasp the tension state, i.e., the progression of the tension abnormality. Therefore, the amplitude evaluation unit 133e1 may output not only the presence or absence of an abnormality but also the belt tension, which indicates the degree of abnormality. For example, the number of frequencies below the threshold may be counted, and when the count reaches k, the abnormality may be evaluated as level k and output to the output unit 140.

[0052] <Procedure for Abnormality Diagnosis Using Abnormality Diagnosis Device 100> Next, the operation of the abnormality diagnosis device 100 will be described with reference to Fig. 11. Fig. 11 is a flowchart showing the overall process for performing abnormality diagnosis using the abnormality diagnosis device 100 according to the first embodiment.

[0053] First, in step S101, the current detector 120 detects the u-phase current of the electric motor 5, and the current data dividing unit 133a divides the detected current data into a predetermined number of parts within a predetermined measurement time. In step S102, the spectrum analyzing unit 133b performs current FFT analysis on the current as described above.

[0054] In step S103, a spectral peak of a specific frequency that satisfies the condition of the belt rotation frequency is extracted from the analyzed spectral waveform. In step S104, the amplitude of the signal strength of the specific frequency is calculated from the time dependency of the signal strength of the specific frequency.

[0055] In step S105, the amplitude or a quantity related to the amplitude of the signal strength of the specific frequency calculated in step S104 is used as an evaluation quantity to determine an abnormality, and is compared with a preset threshold to evaluate the belt tension. If an abnormality is determined, the amplitude evaluation unit 133e1 outputs the diagnosis result to the output unit 140. If the belt tension is not abnormal in step S105, the process returns to the step S105 and the next diagnosis is performed.

[0056] Furthermore, as described above, when belt tension is evaluated using a plurality of specific frequencies, the belt tension state may be output to the output unit 140 even if it is determined in step S105 that the belt is not abnormal.

[0057] <Using Load Torque> The biggest problem when an abnormality in belt tension occurs is that the transmission efficiency of the load torque decreases as a result of the abnormal tension. For this reason, calculating the load torque when an abnormality in belt tension is detected makes it possible to more precisely determine whether or not the belt needs to be replaced. However, the torque detector required to measure the load torque is expensive, and it is desirable to avoid installing one. However, by calculating the load torque using current and voltage, it is possible to avoid installing a torque detector.

[0058] Fig. 12 is a diagram showing the configuration of an abnormality diagnosis device 100 in which current detectors 120u and 120v for each phase and voltage detectors 125uv and 125vw for detecting line voltages are added to the configuration of Fig. 1 in order to calculate the load torque. In Fig. 12, the u-phase current, the v-phase current, the uv voltage, and the vw voltage are acquired.

[0059] Fig. 13 is a block diagram showing the configuration of the abnormality diagnosis unit 130 for calculating the load torque, and is configured by further including a load torque estimation unit 133f in addition to the calculation unit 133 of Fig. 3. The voltages and currents detected by the current detectors 120u, 120v of each phase and the voltage detectors 125uv, 125vw that detect line voltages are stored in a storage unit 132, and the load torque estimation unit 133f calculates the load torque from these currents and voltages.

[0060] This load torque estimation requires a large amount of calculation because it is calculated by combining two phase currents and two types of line voltages. Therefore, the amount of calculation can be reduced by estimating the load torque when the abnormality determination unit 133e determines that the belt tension is abnormal and there is a high possibility that an abnormality has occurred in the load torque.

[0061] Furthermore, when load torque estimation is performed, the load torque is also output simultaneously with the belt tension abnormality or the level of the belt tension abnormality to the output unit 140. The ratio of the estimated load torque to the target load torque is output, and belt replacement can be determined based on this ratio.

[0062] Furthermore, the variance of signal strength at a specific frequency changes depending on the load torque. Therefore, more accurate diagnosis can be achieved by classifying the diagnosis according to the magnitude of the load torque. In this case, since the current value increases as the load torque increases, the effective value or amplitude of the current may be used as the classification criterion instead of the load torque.

[0063] While the above example shows how to perform spectrum analysis by dividing the current data, there is also a method that does not require dividing. When not dividing, data is acquired multiple times, with the measurement time per acquisition being the same as the time used for analysis, and abnormality diagnosis is performed using the acquired multiple data. While this method has the advantage of not requiring a data division unit, because the data itself is discontinuous, the accuracy of calculating the amplitude of the signal strength at a specific frequency is lower than when continuous data is divided.

[0064] Furthermore, as described above, storing all of the time-series data of the current requires a large amount of memory. In this embodiment, since the diagnosis is performed using the amplitude of the signal strength of a specific frequency, the acquired current waveform may be frequency converted for spectrum analysis, and only the signal strength of the specific frequency required for amplitude calculation may be stored, after which the time-series data of the current within the measurement time may be deleted. This makes it possible to reduce the memory required for data storage.

[0065] As described above, according to the first embodiment, an abnormality diagnosis device for diagnosing an abnormality in tension of a belt that transmits power from an electric motor to a load equipment includes a current detector that detects the current of the electric motor, and an abnormality diagnosis unit that diagnoses an abnormality using spectral peaks extracted by FFT analysis of the detected current, wherein the abnormality diagnosis unit includes a spectrum analysis unit that performs FFT analysis on each of a plurality of current values ​​detected consecutively within a certain period of time to extract spectral peaks, a specific frequency extraction unit that extracts spectral peaks attributable to the rotation frequency of the belt as specific frequencies from the power supply frequency of the electric motor and the frequencies of sidebands relative to the power supply frequency, an amplitude calculation unit that calculates an amplitude-related value including amplitude from changes in signal strength of the specific frequency within the certain period of time, and an abnormality determination unit that diagnoses an abnormality in the belt based on the calculated amplitude-related value. This configuration makes it possible to diagnose the state of the belt without relying on the local maximums and minimums of other frequencies caused by the power supply frequency and vibration, and to accurately diagnose an abnormality in the belt tension.

[0066] Furthermore, since the belt tension abnormality is determined using any one of the amplitude-related values ​​calculated from the difference between the maximum and minimum values ​​of the signal strength of a specific frequency within a certain period of time, the standard deviation of the signal strength of a specific frequency within a certain period of time, and the variance of the signal strength of a specific frequency within a certain period of time as the amplitude-related value, it becomes possible to accurately diagnose the belt tension abnormality. Furthermore, the belt tension abnormality is determined using multiple amplitude-related values ​​calculated from changes in the signal strength of multiple specific frequencies, and the belt tension state is determined based on whether each of the multiple amplitude-related values ​​is less than a predetermined threshold value. This makes it possible to replace the belt at an appropriate timing as the belt tension gradually decreases, leading to planned belt replacement and automation of the timing of belt replacement.

[0067] Furthermore, the abnormality diagnosis device further includes a voltage detector that detects the voltage of the electric motor, and the abnormality diagnosis unit further includes a load torque estimation unit that estimates the load torque using the current detected by the current detector and the voltage detected by the voltage detector. Therefore, the belt condition can be determined taking into account the belt tension abnormality determination and the load torque at that time, and further, it becomes possible to accurately diagnose belt tension abnormalities.

[0068] Embodiment 2 An abnormality diagnosis device according to embodiment 2 will now be described with reference to the drawings. In embodiment 2, the time change in the signal strength of the rotation frequency is regarded as a waveform, and abnormality diagnosis is performed using the feature quantity of the waveform.

[0069] Fig. 14 is a block diagram showing the configuration of the abnormality diagnosing unit 130 of the abnormality diagnosing device 100 according to embodiment 2. The difference from Fig. 3 of embodiment 1 is that the calculation unit 133 further includes a waveform feature amount calculation unit 133g, and the abnormality determination unit 133e includes a waveform feature amount evaluation unit 133e2 instead of the amplitude evaluation unit 133e1.

[0070] In the second embodiment, the change over time in the signal strength of a specific frequency is regarded as a waveform, and the waveform feature amount calculation unit 133g and the abnormality determination unit 133e use a general method for calculating the feature amount and for diagnosing the abnormality using the feature amount. Examples of waveform feature amounts include amplitude, vibration center, and period. Figures 6 and 15 show feature amounts at an appropriate tension and when the belt is starting to loosen. These figures show that these feature amounts change with different belt tensions. Another diagnostic method using feature amounts is machine learning. The machine learning used may be any method, such as principal component analysis or neural network.

[0071] FIG. 16 is a diagram illustrating a method for diagnosing belt abnormalities using a neural network. In the diagram, a neural network is constructed using the feature quantities of amplitude, vibration center, and period as input and tension as output. In FIG. 16, three types of feature quantities and four patterns of belt tension are used, but these numbers can be arbitrarily specified. By using this method, it is possible to quantitatively estimate tension compared to when only amplitude is used.

[0072] These feature quantities can also be treated as feature quantities in the time domain and feature quantities in the frequency domain. The feature quantity in the time domain is, for example, a quantity obtained by averaging each feature quantity over time, and an example of this is p ｔ１ ~p ｔ１２ Here, the characters used in equations (10) to (21) are calculated using equations (22) and (23).

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086] Furthermore, the feature quantity in the frequency domain can be obtained by inputting the time dependency of the signal strength of the rotation frequency into the spectrum analysis unit 133b and converting it into the frequency domain. The feature quantity calculated from the result of the frequency conversion is used to perform abnormality diagnosis.

[0087] In the second embodiment, feature calculation and machine learning are required, and therefore the amount of calculation increases compared to the first embodiment. Therefore, in order to diagnose only a decrease in belt tension, the first embodiment can handle the situation with a smaller amount of calculation.

[0088] As described above, the second embodiment achieves the same effects as the first embodiment. Furthermore, the abnormality diagnosis unit includes a waveform feature calculation unit that calculates a feature value by regarding a change in signal strength at a specific frequency within a certain period of time as a waveform, and an abnormality determination unit that determines whether a belt tension abnormality exists based on the feature value calculated thereby, thereby enabling a more accurate diagnosis of a belt tension abnormality. Furthermore, the abnormality determination unit compares the feature value with the feature value that has been machine-learned to determine whether a belt tension abnormality exists, thereby enabling a quantitative evaluation of the belt tension.

[0089] In the first and second embodiments described above, the abnormality diagnosis device 100 includes a processor 1000 and a storage device 2000, as shown in FIG. 17 , which illustrates an example of hardware. Although the storage device is not shown, it includes a volatile storage device such as a random access memory and a non-volatile auxiliary storage device such as a flash memory. Alternatively, a hard disk auxiliary storage device may be used instead of the flash memory. The processor 1000 executes a program input from the storage device 2000. In this case, the program is input from the auxiliary storage device to the processor 1000 via the volatile storage device. The processor 1000 may output data such as calculation results to the volatile storage device of the storage device 2000, or may store the data in the auxiliary storage device via the volatile storage device.

[0090] 18 , the system may further include a communication device 3000. For example, if a monitoring control device 200 that collectively monitors the load devices 10 is present in a plant where the load devices 10 are installed, the abnormality diagnosis results determined by the abnormality diagnosis device 100 and the like can be transmitted from the output unit 140 to the monitoring control device 200.

[0091] Although the abnormality diagnosis device 100 has been described as having the hardware configuration shown in FIG. 17 or FIG. 18, the motor control device 110 and the abnormality diagnosis unit 130 may each have the hardware configuration shown in FIG. 17 or FIG. 18.

[0092] Although various exemplary embodiments and examples are described in this disclosure, the various features, aspects, and functions described in one or more embodiments are not limited to the application of a particular embodiment, but may be applied to the embodiments alone or in various combinations. Therefore, countless variations not illustrated are anticipated within the scope of the technology disclosed herein. For example, this includes cases where at least one component is modified, added, or omitted, or where at least one component is extracted and combined with components of another embodiment.

[0093] 2: Power supply, 5: Electric motor, 6: Power transmission mechanism, 61: Belt, 7: Load equipment, 10: Load device, 100: Abnormality diagnosis device, 110: Motor control device, 111: Inverter, 112: Control unit, 120, 120u, 120v: Current detector, 125uv, 125vw: Voltage detector, 130: Abnormality diagnosis unit, 131: Load device setting unit, 132: Storage unit, 133: Calculation unit, 133a: Current data division unit, 133b: Spectral analysis unit, 133c: Specific frequency extraction unit, 133d: Amplitude calculation unit, 133e: Abnormality determination unit, 133e1: Amplitude evaluation unit, 133e2: Waveform feature amount evaluation unit, 133f: Load torque estimation unit, 133g: Waveform feature amount calculation unit, 140: Output unit, 200: Monitoring and control device, 1000: Processor, 2000: Storage device, 3000: Communication device.

Claims

1. An abnormality diagnosis device for determining abnormalities in the tension of a belt, which is a power transmission mechanism that transmits power from an electric motor to a load equipment, comprising: a current detector that detects the current of the electric motor; and an abnormality diagnosis unit that determines abnormalities using a spectral peak extracted by FFT analysis of the current detected by the current detector, wherein the abnormality diagnosis unit has: a spectral analysis unit that performs FFT analysis on each of a number of current values ​​detected continuously within a fixed time period to extract a spectral peak; a specific frequency extraction unit that extracts, as a specific frequency, the spectral peak caused by the rotation frequency of the belt from the power supply frequency of the electric motor and the frequency of a sideband wave for the power supply frequency; an amplitude calculation unit that calculates an amplitude-related value including an amplitude from a change in signal strength of the specific frequency within the fixed time period; and an abnormality determination unit that determines abnormalities in the tension of the belt based on the calculated amplitude-related value of the signal strength of the specific frequency.

2. The abnormality diagnosis device according to claim 1, wherein the abnormality determination unit determines an abnormality in tension of the belt using, as the amplitude-related value, any one of an amplitude calculated from the difference between the maximum and minimum values ​​of the signal strength of the specific frequency within the fixed time period, a standard deviation of the signal strength of the specific frequency within the fixed time period, and a variance of the signal strength of the specific frequency within the fixed time period.

3. The abnormality diagnosis device according to claim 1 or 2, wherein the abnormality determination unit determines that the belt has an abnormality in tension when the amplitude-related value is less than a preset threshold value.

4. The abnormality diagnosis device according to claim 1 or 2, wherein the abnormality determination unit determines abnormality in tension of the belt using the amplitude-related value calculated from changes in signal strength at a plurality of the specific frequencies.

5. The abnormality diagnosis device according to claim 4, further comprising: a preset threshold value for each of the plurality of amplitude-related values ​​calculated from the change in signal strength of the plurality of specific frequencies; and the abnormality determination unit determines whether each of the plurality of amplitude-related values ​​is less than the preset threshold value, thereby determining the tension state of the belt.

6. An abnormality diagnosis device as described in any one of claims 1 to 5, further comprising a voltage detector that detects the voltage of the electric motor, and the abnormality diagnosis unit further comprises a load torque estimation unit that estimates a load torque using the current detected by the current detector and the voltage detected by the voltage detector.

7. The abnormality diagnosis device according to claim 6, wherein the load torque estimation section estimates the load torque when the abnormality determination section determines that the tension of the belt is abnormal.

8. An abnormality diagnosis device for determining an abnormality in the tension of a belt, which is a power transmission mechanism that transmits power from an electric motor to a load equipment, comprising: a current detector that detects the current of the electric motor; and an abnormality diagnosis unit that determines an abnormality using a spectral peak extracted by FFT analysis of the current detected by the current detector, wherein the abnormality diagnosis unit has: a spectral analysis unit that performs FFT analysis on each of a plurality of current values ​​detected continuously within a fixed time period to extract a spectral peak; a specific frequency extraction unit that extracts, as a specific frequency, the spectral peak caused by the rotation frequency of the belt from the power supply frequency of the electric motor and the frequency of a sideband wave for the power supply frequency; a waveform feature calculation unit that calculates a feature amount by regarding a change in signal strength of the specific frequency within the fixed time period as a waveform; and an abnormality determination unit that determines an abnormality in the tension of the belt based on the calculated feature amount.

9. The abnormality diagnosis device according to claim 8, wherein the characteristic amount is a frequency of the waveform.

10. An abnormality diagnosis device as described in claim 8 or 9, wherein the abnormality determination unit compares the feature amount with the feature amount that has been machine-learned to determine whether or not there is an abnormality in the tension of the belt.

11. An abnormality diagnosis method for determining an abnormality in the tension of a belt, which is a power transmission mechanism that transmits power from an electric motor to a load equipment, comprising the steps of: detecting the current of the electric motor; performing FFT analysis on each of a number of current values ​​detected consecutively within a fixed time period to extract a spectral peak; calculating the spectral peak caused by the rotation frequency of the belt as a specific frequency from the power supply frequency of the electric motor and the frequency of a sideband wave relative to the power supply frequency; calculating an amplitude-related value including an amplitude from a change in signal strength of the specific frequency within the fixed time period; and determining an abnormality in the tension of the belt based on the calculated amplitude-related value of the signal strength of the specific frequency.

12. The abnormality diagnosis method according to claim 11, wherein in the step of determining abnormal tension in the belt, the tension abnormality in the belt is determined using, as the amplitude-related value, any one of an amplitude calculated from the difference between the maximum and minimum values ​​of the signal strength of the specific frequency within the fixed time period, a standard deviation of the signal strength of the specific frequency within the fixed time period, and a variance of the signal strength of the specific frequency within the fixed time period.

13. The abnormality diagnosis method according to claim 11 or 12, wherein in the step of determining whether the tension of the belt is abnormal, it is determined that the tension of the belt is abnormal when the amplitude-related value is less than a preset threshold value.

14. The abnormality diagnosis method according to claim 11 or 12, wherein in the step of determining whether or not the tension of the belt is abnormal, the tension of the belt is determined using the amplitude-related value calculated from the change in signal strength of a plurality of the specific frequencies.

15. The abnormality diagnosis method according to claim 14, wherein in the step of determining whether the belt tension is abnormal, it is determined whether each of the multiple amplitude-related values ​​calculated from the changes in signal strength of the multiple specific frequencies is less than a preset threshold value, thereby determining the tension state of the belt.

16. The abnormality diagnosis method according to any one of claims 11 to 15, further comprising a step of estimating a load torque of the motor, and in determining whether or not the tension of the belt is abnormal, the load torque is further used to determine the tension state of the belt.

17. An abnormality diagnosis method for determining an abnormality in the tension of a belt, which is a power transmission mechanism that transmits power from an electric motor to a load equipment, comprising the steps of: detecting the current of the electric motor; extracting a spectral peak by performing FFT analysis on each of a number of current values ​​detected consecutively within a fixed time period; calculating the spectral peak caused by the rotation frequency of the belt as a specific frequency from the power supply frequency of the electric motor and the frequency of a sideband wave relative to the power supply frequency; calculating a feature amount by regarding the change in signal strength of the specific frequency within the fixed time period as a waveform; and determining an abnormality in the tension of the belt based on the calculated feature amount.

18. The abnormality diagnosis method according to claim 17, wherein the characteristic amount is a frequency of the waveform.

Images