Abnormality detection device, air conditioner, and abnormality detection method

Abstract

The purpose of an abnormality detection device (1) according to the present disclosure is to detect an incipient turn-to-turn short. The abnormality detection device (1) is characterized by comprising: a frequency analysis unit (12) that performs frequency analysis on physical-quantity data of a rotating electrical machine (2); a mechanical-angle rotational frequency harmonic component extraction unit (13) that extracts a specific-order harmonic component of a mechanical-angle rotational frequency from among components having been subjected to the frequency analysis; and an abnormality determination unit (15) that compares the extracted specific-order harmonic component with a threshold value (I threshold) and that determines the presence or absence of a turn-to-turn short circuit in the rotating electrical machine (2).

Description

Abnormality detection device, air conditioner, and abnormality detection method

[0001] The present disclosure relates to an abnormality detection device, an air conditioner, and an abnormality detection method.

[0002] A degradation and abnormality diagnosis method has been disclosed for detecting layer shorts in rotating electrical machines, which focuses on the electrical angle and its harmonic components of the physical quantity measured for the rotating electrical machine, and for example, performs frequency analysis on the measured current to detect abnormalities at the magnitude of components five or seven times the electrical angle (see, for example, Patent Document 1).

[0003] JP 2022-101344 A (paragraphs 0061 to 0066)

[0004] However, since insulation breakdown accelerates as the heat generation due to the short-circuit current increases when a layer short occurs, it is necessary to detect layer shorts at an early stage. However, the degradation abnormality diagnosis method described above is unable to detect the abnormality at an early stage and is therefore insufficient to prevent the occurrence of abnormalities.

[0005] The present disclosure is intended to solve the above problem and aims to detect a layer short circuit in the early stages.

[0006] The abnormality detection device disclosed herein is characterized by comprising a frequency analysis unit that performs frequency analysis on physical quantity data of a rotating electric machine, a mechanical angle rotation frequency multiple component extraction unit that extracts specific multiple components of the mechanical angle rotation frequency from the frequency-analyzed components, and an abnormality determination unit that compares the extracted specific multiple components with a threshold value and determines whether or not there is a layer short in the rotating electric machine.

[0007] The anomaly detection method disclosed herein is characterized by including a frequency analysis step of frequency-analyzing physical quantity data of a rotating electric machine, a mechanical angle rotation frequency multiple component extraction step of extracting specific multiple components of the mechanical angle rotation frequency from the frequency-analyzed components, and an anomaly determination step of comparing the extracted specific multiple components with a threshold value to determine whether or not there is a layer short in the rotating electric machine.

[0008] According to the abnormality detection device or abnormality detection method disclosed herein, the presence or absence of an abnormality can be detected based on the mechanical angle of the rotating electric machine and the signal strength of its harmonic components, making it possible to detect layer shorts at an early stage.

[0009] FIG. 1 is a block diagram showing a basic configuration of an anomaly detection device according to a first embodiment. FIG. 2 is a block diagram showing a configuration of an air conditioner including a rotating electric machine that is a target for detecting a layer short circuit by the anomaly detection device or an anomaly detection method according to the first embodiment. FIG. 3 is a block diagram showing a configuration of an anomaly detection device according to the first embodiment and a drive device for a rotating electric machine in which the anomaly detection device is incorporated. FIG. 4 is a waveform diagram showing an example of a frequency spectrum waveform of a U-phase current. FIG. 5 is a flowchart showing the operation of the anomaly detection device or an anomaly detection method according to the first embodiment. FIG. 6 is a waveform diagram showing a frequency spectrum waveform of a U-phase current when the failure rate is 0% and when the failure rate is 5%. FIG. 7 is a block diagram showing a hardware configuration of the anomaly detection device according to the first embodiment. FIG. 8 is a block diagram showing a configuration of an anomaly detection device according to a second embodiment and a drive device for a rotating electric machine in which the anomaly detection device is incorporated. FIG. 9 is a flowchart showing the operation of the anomaly detection device or an anomaly detection method according to the second embodiment. FIG. 10 is a waveform diagram showing a frequency spectrum waveform of a q-axis current when the failure rate is 0% and when the failure rate is 5%.

[0010] 1 to 6 are diagrams for explaining the configuration and operation of an abnormality detection device or an abnormality detection method according to a first embodiment, in which Fig. 1 is a block diagram showing the basic configuration of the abnormality detection device according to the first embodiment together with a rotating electric machine or the like that is a target for detecting a rare short, Fig. 2 is a block diagram showing the configuration of an air conditioner that includes a rotating electric machine that is a target for detecting a rare short, and Fig. 3 is a block diagram showing the configuration of the abnormality detection device and a drive device for the rotating electric machine in which the abnormality detection device is incorporated.

[0011] FIG. 4 is a waveform diagram showing an example of the frequency spectrum waveform of the U-phase current among the phase currents flowing through a rotating electric machine, FIG. 5 is a flowchart showing the operation of the abnormality detection device or the abnormality detection method, and FIG. 6 is a waveform diagram showing the frequency spectrum waveform of the U-phase current when the failure rate is 0% and when it is 5%.

[0012] <Basic Configuration> As shown in Fig. 1, the abnormality detection device 1 of the present disclosure is for detecting a layer short (inter-layer short (between windings)) occurring in a rotating electric machine 2, also known as an electric motor, based on measured values ​​from a physical quantity sensor 9 that measures the physical quantity of the rotating electric machine 2. To this end, the device includes a frequency analysis unit 12 that performs frequency analysis of the measured values, a mechanical angle rotation frequency multiple order component extraction unit 13 that extracts multiple order components of the mechanical angle rotation frequency from the analyzed values, and an abnormality determination unit 15 that determines the presence or absence of an abnormality (layer short) by comparing the extracted multiple order components with a threshold value set based on the characteristics of normal values. Furthermore, the device includes an abnormality notification unit 16 that notifies that a layer short has been detected, and a threshold value storage unit 18 that stores the threshold value.

[0013] The rotating electric machine 2 is used in connection with a load device 3. The rotating electric machine 2 and the load device 3 may be directly connected, or a power transmission mechanism may be interposed between them. Power transmission mechanisms include belts, gears, and couplings. Examples of the load device 3 include pumps and fans. The object to be measured (attached) by the physical quantity sensor 9 that measures the "physical quantity of the rotating electric machine" is not limited to the rotating electric machine 2, but may also be a drive unit 8 that drives the rotating electric machine 2, or a load device 3 that operates using the power of the rotating electric machine 2.

[0014] <Example of Load Device> Detection of a layer short circuit in the abnormality detection device 1 of the present disclosure will be described using an example in which a compressor 3C of an air conditioner 30 is used as the load device 3, as shown in Figure 2. The air conditioner 30 includes the compressor 3C that compresses a refrigerant, a condenser 4 that cools and condenses the compressed and temperature-raised refrigerant by heat exchange with outside air, an expansion valve 5 that decompresses the condensed high-pressure refrigerant, and an evaporator 6 that cools the air by heat exchange with the refrigerant whose temperature has been reduced by decompression. The air conditioner 30 operates as a refrigeration cycle device, and the refrigerant circulates through the compressor 3C, condenser 4, expansion valve 5, and evaporator 6 in this order.

[0015] The rotating electric machine 2 is connected to a compressor 3C that compresses gas refrigerant in a refrigeration cycle, and compresses and discharges the gas refrigerant drawn in using the mechanical driving force of the rotating electric machine 2. The condenser 4 condenses the gas refrigerant discharged from the compressor 3C and discharges liquid refrigerant. The expansion valve 5 controls its opening to expand and reduce the pressure of the refrigerant from the condenser 4. The evaporator 6 evaporates the liquid refrigerant discharged from the expansion valve 5 and discharges gaseous refrigerant.

[0016] 3 will be used to explain an example in which the abnormality detection device 1 is incorporated into a drive device 8 of a rotating electric machine 2, assuming an air conditioner 30. The drive device 8 is connected to the rotating electric machine 2 and supplies drive power to the rotating electric machine 2. The drive device 8 includes a converter 83 that converts AC voltage into DC voltage of a desired voltage, an inverter 82 that converts the DC voltage into AC voltage of a desired frequency and outputs it to the rotating electric machine, a control unit 81 that controls the operation of the inverter 82, and the abnormality detection device 1.

[0017] The converter 83 receives AC current from a power supply (not shown), converts the AC current to DC current of a desired voltage, and outputs the DC current to the inverter 82. The power supply has a frequency of, for example, 50 Hz or 60 Hz. The inverter 82 includes an inverter main circuit including a plurality of switching elements. The inverter 82 receives a PWM (Pulse Width Modulation) signal from the control unit 81 and switches the switching elements on and off, thereby outputting a three-phase (UVW phase) current to the rotating electric machine 2 included in the compressor 3C. A U-phase current IU output to the rotating electric machine 2 is measured by a current sensor 9C, which is a physical quantity sensor 9, and output to the control unit 81.

[0018] The control unit 81 includes a phase current calculation unit 811 that calculates phase currents to be output to the rotating electric machine 2 in order to output PWM signals to the inverter 82 for vector control. The control unit 81 also includes a Clarke transformation unit 812 that performs a Clarke transformation to convert the calculated time domain components of the three-phase system into two components of an orthogonal stationary coordinate system, and a Park transformation unit 813 that performs a Park transformation to convert the two components into an orthogonal rotating reference coordinate system (dq).

[0019] Furthermore, the circuit is provided with a voltage command value calculation unit 814 that calculates a voltage command value, which is a DC signal, from the AC current waveform and voltage waveform based on Clarke transformation and Park transformation, an output voltage vector calculation unit 815 that calculates an output voltage vector based on the voltage command value, and a PWM signal generation unit 816 that generates a PWM signal based on the output voltage vector.

[0020] The phase current calculation unit 811 calculates the U-phase current I output from the current sensor 9C. U , and V-phase current I V The W-phase current I W The phase current calculation unit 811 calculates each phase current (I U , I V , I W ) to the Clarke transform unit 812. U , I V , I W ) changes with a change in the rotation angle θ (mechanical angle) of the rotating electrical machine 2 (strictly speaking, the rotor).

[0021] In the following description, the rotation angle θ is described as a value measured by an angle sensor (not shown), but the angle sensor is not an essential component, and the rotation angle θ may be calculated by other methods. For example, as is done in position sensorless control, the rotation angle θ may be calculated by measuring the phase current (I U , I V , I W The rotation angle θ may be calculated from the torque and the voltage command value. When detecting an abnormality using torque, a torque detector may be added, but it is also possible to calculate the torque using the measured phase current.

[0022] <Abnormality Detection Method> A layer short detection method using an acquired physical quantity will be described. In this embodiment, layer short detection using current as the physical quantity will be described. Detection using current requires only installing a sensor on the power cable, eliminating the need for an additional sensor, which is advantageous in terms of cost. However, layer short detection is also possible using a physical quantity other than current. Examples of physical quantities other than current include voltage, torque, rotation angle, and rotation speed.

[0023] Here, as a prerequisite for detecting a layer short circuit, as shown in FIG.U 5 shows an example of a frequency spectrum waveform of the rotating electric machine 2. When the stator current of the rotating electric machine 2 is frequency-analyzed, current fluctuations due to fault symptoms occur periodically, and sideband waves, which are characteristic spectrum peaks, appear at different positions depending on the type of abnormality near both sides of the power supply frequency (360 Hz in the figure) output by the drive unit 8. A method for detecting a layer short circuit by comparing the strength of these sideband waves with a normal value will be described with reference to the flowchart in FIG.

[0024] In the anomaly detection method according to the first embodiment, a threshold setting mode is used in which a threshold for determining whether or not an anomaly has occurred is set based on the normal value described above, and an anomaly determination mode is used in which the presence or absence of an anomaly is determined based on the set threshold. In addition, whether or not to set a threshold and perform an anomaly determination is determined depending on the operating mode, such as the operating status and operating environment of the air conditioner 30.

[0025] First, the operating condition setting unit 19 determines whether or not operating conditions that serve as criteria for determining whether or not to set a threshold value and perform an abnormality determination have been set (step S100). If the operating conditions have been set (Yes), the process proceeds to a mode switching step (step S120) in which the process switches between a mode for setting a threshold value (threshold setting mode) and a mode for determining whether or not an abnormality has occurred (abnormality determination mode). On the other hand, if the operating conditions have not been set (No), the process sets the operating conditions (step S110) and then proceeds to step S120.

[0026] The operating conditions may be set, for example, by an operator's input / output operations via an interface (not shown). In this case, for example, data on operating conditions according to the specifications of the rotating electric machine 2 and the specifications of the load device 3 may be stored, and the operating conditions may be selected from displayed candidates. Alternatively, the operating conditions may be automatically selected by inputting the model of the rotating electric machine 2 and the model of the load device 3.

[0027] In the mode switching step (step S120), the mode switching unit 14 switches the current mode between the threshold setting mode and the abnormality determination mode, and then proceeds to step S200, where it is determined whether the operating conditions are met. Whether to switch to the threshold setting mode or the abnormality determination mode can be determined based on the operating time of the rotating electric machine 2 and the load device 3, or the time since the abnormality detection device 1 was installed. For example, if it is at the start of operation or during the so-called early stage of operation up to a certain time after installation, it is assumed that data under normal conditions can be acquired, so it is determined to be the threshold setting mode and a threshold can be set.

[0028] Alternatively, it can be determined whether a threshold has already been set, such as whether a threshold has been stored in the threshold storage unit 18, which will be described in a later step, and if a threshold has not been stored, the mode can be switched to the threshold setting mode, and if a threshold has been stored, the mode can be switched to the abnormality determination mode. However, if it is necessary to set a threshold for each of the above-mentioned operating conditions, the mode can be switched for each operating condition.

[0029] In step S200, the operating condition determination unit 11 determines whether the current operating conditions match the set operating conditions, and if the conditions match (Yes), proceeds to a process (step S210 and subsequent steps) in which calculations for abnormality detection, such as threshold setting or abnormality determination, are performed.

[0030] Note that steps S100 to S110 and S200 executed by the operating condition setting unit 19 and the operating condition determination unit 11 are not essential steps. However, since abnormality detection focuses on the signal strength of a specific rotational frequency as described below, detection accuracy is improved by detecting abnormalities under constant speed operating conditions. Also, as described below, when a layer short occurs, the signal strength of a specific rotational frequency tends to increase, but the amount of increase differs depending on the load and rotational speed.

[0031] For this reason, it is desirable to set the load and rotation speed for abnormality detection in advance and perform abnormality detection under those conditions. When setting the rotation speed, it is possible to improve the accuracy of abnormality detection by selecting conditions that do not match with peaks caused by noise. For example, when driven by an inverter, there is noise caused by DC voltage pulsation and asynchronous operation of PWM.

[0032] The frequency of the DC voltage pulsation is f D , the frequency of noise caused by asynchronous operation of PWM is f A It is known that, when f is set, peaks occur at the frequencies shown in equations (1) and (2). D = |mf0±nf ac | (1) f A = mGCD(f0, f s , f c ) (2) In the formulas (1) and (2), m and n are natural numbers, f0 is the driving frequency of the power supply, and f ac is the AC fundamental component contained in the DC voltage, f s is the sampling frequency, f c denotes the carrier frequency, and GCD denotes the greatest common divisor.

[0033] If the 1f component of the mechanical angle rotation frequency (described later) coincides with these frequencies, noise-induced peaks will be superimposed, reducing detection accuracy. Therefore, when specifying operating conditions for anomaly detection, it is desirable to specify operating conditions that avoid these frequencies. Furthermore, if the motor is not driven under these operating conditions during actual operation, subsequent steps of anomaly detection will not begin.

[0034] Therefore, it is necessary to set the operating conditions so that the frequency of the abnormality detection is equal to or greater than the frequency T1 (for example, T1: 1 hour) at which the abnormality detection is desired to be performed. The operating conditions that satisfy these conditions are stored in the operating condition determination unit 11.

[0035] If it is difficult to set conditions based on load and rotation speed, a control pattern such as an operating pattern or environmental conditions may be set as operating conditions. For example, for compressor 3C of air conditioner 30, detection conditions may be set based on the seasons (summer, winter, etc.) in which the air conditioner 30 operates, or on weather (external environment) such as temperature, humidity, and air pressure, or, if there are multiple operating modes, on each operating mode. Even with settings that do not directly specify load and rotation speed, detection accuracy is improved compared to when no settings are made. If this operating condition setting step is omitted, detection accuracy will decrease, but there is no need to perform detection for each operating condition, which has the advantages of simplifying the process and reducing memory.

[0036] If the operating conditions are met ("Yes" in step S200), regardless of the mode, the frequency analysis unit 12 acquires data on the current flowing between the drive unit 8 and the rotating electrical machine 2 as a physical quantity, the data being measured by the current sensor 9C (step S210). The acquired current data is time-series data on the current.

[0037] As explained in FIG. 4, the frequency of interest is near the power supply frequency, and information on the signal strength in the high frequency region is not required. Therefore, a sampling frequency of several times the power supply frequency is sufficient, but sampling is generally performed at a frequency of 10 times or more the power supply frequency.

[0038] Furthermore, the current is not limited to the value measured by the current sensor 9C. For example, it is also possible to use either a command value used for control or a state estimate value of each physical quantity. Furthermore, as described above, the abnormality detection of the present disclosure focuses on a specific frequency component obtained by frequency analysis, so it is desirable to obtain data at a constant rotation speed.

[0039] Then, frequency analysis is performed for each time-series data, and the current is converted into the frequency domain (step S220). Known examples of frequency analysis include current FFT (Fast Fourier Transform) analysis. As described with reference to FIG. 3, the drive device 8 for the rotating electric machine 2 is assumed to be connected to an inverter 82, and noise is superimposed on the converted results.

[0040] Therefore, as described below, it is believed possible to reduce the influence of noise by performing anomaly detection using multiple analysis results (step S240). However, if sufficient accuracy is not achieved even using multiple analysis results, an analysis method for removing noise can be introduced in step S220. An example of an analysis method for removing noise is a method in which compressed sensing is applied to sparse current-frequency characteristics and filtering is performed to emphasize only the characteristic components.

[0041] Next, from the frequency analysis results, the mechanical angle rotational frequency multiple component extraction unit 13 extracts spectral peaks that occur in the rotational frequency band, which is the mechanical angle rotational frequency multiple component, and calculates the strength of the extracted signal (signal intensity) (step S230).

[0042] A layer short is a phenomenon in which resistance within or between phases is short-circuited. When a layer short occurs, an imbalance occurs between inductance and resistance, which in turn causes an imbalance in magnetic flux. Eccentricity, a mechanical abnormality, is an example of an imbalance between inductance, resistance, and magnetic flux. When eccentricity occurs, it is known that the current at the mechanical angular rotation frequency and its harmonic components changes, and there are known examples of detecting eccentricity by focusing on changes in the current at the mechanical angular rotation frequency and its harmonic components.

[0043] Since the change in inductance is a common physical phenomenon, it is assumed that a change in the current signal strength of the mechanical angle rotation frequency can also occur due to a layer short, which is an electrical abnormality, and layer short detection is performed by focusing on the current signal strength of the mechanical angle rotation frequency. p , and information p on the number of pole pairs is input in advance, the 1f component of the mechanical angle rotation frequency of the rotating electrical machine 2 can be calculated.

[0044] The 1f component of the mechanical angle rotation frequency of the phase current is f p / p or 5f p In step S230, the mechanical angle rotation frequency multiple order component extractor 13 calculates the signal strength of this frequency, and stores the calculation result in a memory (not shown).

[0045] Here, in order to remove the influence of noise, the processes from step S200 onwards are repeated until data is acquired a certain number of times or more set for each mode ("No" in step S240). This determination may be made by the mechanical angle rotation frequency multiple order component extraction unit 13 or the mode switching unit 14. If data has been acquired the required number of times ("Yes" in step S240), the process proceeds to the mode determination process (step S300).

[0046] In the mode determination step, in accordance with the mode switched in step S120, if the threshold setting mode is selected, the threshold setting unit 17 sets a threshold (step S400). Here, the threshold is set based on the signal strength extracted in step S230. The signal strength is determined by repeating step S240, and is calculated based on the 1f component f of the mechanical angle rotation frequency in the phase current of the rotating electrical machine 2. p / p or 5f p The current value at / p is acquired the number of times set for the threshold setting mode.

[0047] The reason for acquiring multiple times is to set a threshold that takes into account the influence of variations such as noise. The number of times is the number required to calculate the variations, and is set to 20 as an example. When 20 acquisitions are completed, the average Inormal average and variance σ of the current values ​​of these 20 acquisitions are acquired and stored in the threshold storage unit 18. The threshold Ithreshold is calculated using the stored Inormal average, variance σ, and a real number k greater than 0, as shown in equation (3). Note that the value of k is generally 2 to 3. Ithreshold = Inormal average + kσ (3)

[0048] By calculating the threshold Ithreshold including the variance σ in this way, it is possible to reduce the influence of operational variations. There is also a method for estimating the level of rare shorts by setting multiple patterns for the value of k. For example, k is set to k<k<k. i Using Ithreshold i is calculated as in equation (4).i = Inormal average + k i σ (4)

[0049] In this way, multiple thresholds Ithreshold 1 can be set, for example, caution is required when Ithreshold 1 is reached, confirmation is required at the next inspection when Ithreshold 2 is reached, and immediate stop when Ithreshold 3 is reached. i By setting this, you can understand the progression (severity) of the rare short.

[0050] Furthermore, if multiple operating conditions are set in step S110, it is necessary to set a threshold value Ithreshold for each set operating condition. Furthermore, the acquired current data is deleted when the threshold value Ithreshold is stored in the threshold value storage unit 18. This reduces memory load.

[0051] In addition, whether it is an abnormality determination or a threshold setting, it is necessary to perform frequency analysis, calculate a specific signal strength, and store it (steps S220 to S230), which incurs calculation costs. Furthermore, since the abnormality detection method disclosed herein enables early detection of a rare short, it is sufficient to perform abnormality detection at a timing of about the frequency T1 at which abnormality detection is desired. Specifically, in step S200 in the abnormality determination mode described below, if the operating conditions are met and T1 time has passed since the previous abnormality detection, a "Yes" determination is made and abnormality detection is started.

[0052] In the abnormality determination mode, similarly to the threshold determination mode, if the operating conditions are met in step S200, the processes of steps S210 to S240 are repeated. Then, the signal strength extracted in step S230 is compared with the threshold value Ithreshold or Ithreshold set in the threshold setting mode and stored in the threshold value storage unit 18. i and judge whether or not there is an abnormality (step S500).

[0053] In step S500, the multiple signal intensities stored in the abnormality determination unit 15 by repeating step S240 are compared. For example, if the number of stored intensities is 20, the number is maintained by deleting the oldest signal intensity acquired when the number exceeds 20. Of the 20 signal intensities, the number that exceeds a threshold is counted by the abnormality determination count unit 151. When this count exceeds a set value, an abnormality is determined.

[0054] If the set value is 70 to 90% of the total number, then for 20, the set value would be 14 to 18. The larger the set value, the more counts are required to determine an abnormality, reducing the possibility of false detection, but the detection of layer shorts will be slower, so the set value can be adjusted appropriately depending on the actual situation.

[0055] If it is determined in step S500 that an abnormality does not exist ("No" in step S510), the process returns to step S200. On the other hand, if it is determined that an abnormality exists ("Yes" in step S510), the abnormality notification unit 16 notifies the user that an abnormality exists (step S520). Methods for notifying the user of an abnormality include issuing an alarm or displaying an alarm.

[0056] Figure 6 shows an example of the frequency spectrum results of the phase current when the failure rate is 5% (dashed line) as the time of failure and when the failure rate is 0% as the time of normality. As shown in Figure 6, the frequency spectrum analysis results of the phase current show that there is a larger difference between the normal and time of failure in the 1f components (120 Hz, 600 Hz) of the mechanical rotation frequency (240 Hz) compared to the harmonic components of the electrical rotation frequency (360 Hz).

[0057] Next, we will explain how to calculate the 1f component of the mechanical angle rotation frequency. If the mechanical angle rotation frequency is fr and the power supply frequency is fs, the 1f component of the mechanical angle rotation frequency occurs at fs±fr as a sideband of the power supply frequency. If the number of poles is P, then fr = 2fs / P. In the motor considered here, the electrical angle rotation frequency is 360 and the number of poles is 3, so fr = 240, and fs±fr is 120 Hz and 600 Hz. This means that the 1f component of the mechanical angle rotation frequency is more sensitive than the electrical angle rotation frequency, and by performing abnormality detection that focuses on this component, early detection of layer shorts is possible.

[0058] In other words, by checking the difference from normal times in the signal strength of the mechanical angle of the rotating electrical machine 2 and its harmonic component (1f component), it is possible to detect an abnormality at an early stage. In this case, by focusing on the frequency of the mechanical angle 1f component of the rotating electrical machine 2, which is affected by changes in inductance that affect the principle of a layer short, it is possible to reliably detect a layer short at an early stage.

[0059] It should be noted that, with regard to the anomaly detection device 1 of the present disclosure, the entire anomaly detection device 1, or the part that performs the calculations, can be configured as a single piece of hardware 100 including a processor 101 and a storage device 102, as shown in FIG. 7 . Although not shown, the storage device 102 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 included instead of the flash memory. The processor 101 executes a program input from the storage device 102. In this case, the program is input to the processor 101 from the auxiliary storage device via the volatile storage device. The processor 101 may output data such as calculation results to the volatile storage device of the storage device 102, or may store the data in the auxiliary storage device via the volatile storage device.

[0060] Embodiment 2 In the first embodiment, an example was described in which the 1f component of the mechanical angle rotational frequency of a rotating electric machine is calculated based on the phase currents applied to the rotating electric machine. In the second embodiment, an example will be described in which the 1f component of the mechanical angle rotational frequency of a rotating electric machine is calculated based on the d-axis current and the q-axis current.

[0061] 8 to 10 are diagrams for explaining the configuration and operation of an anomaly detection device or an anomaly detection method according to the second embodiment, with Fig. 8 being a block diagram showing the configuration of an anomaly detection device according to the second embodiment and a drive device for a rotating electric machine incorporating the anomaly detection device, Fig. 9 being a flowchart showing the operation of the anomaly detection device or the anomaly detection method, and Fig. 10 being a waveform diagram showing the frequency spectrum waveforms of the q-axis current when the failure rate is 0% and when it is 5%. Note that the second embodiment is the same as the first embodiment except that the q-axis current is used instead of the current value, and therefore explanations of similar parts will be omitted and Figs. 1 and 2 used in the first embodiment will be used.

[0062] As shown in Fig. 8, the components of the abnormality detection device 1 according to the second embodiment are the same as those of the abnormality detection device 1 described in Fig. 1, except that the input signal is not a current value from the current sensor 9C but a q-axis current from the park conversion unit 813 of the drive device 8. Other configurations and operations, such as switching between the threshold setting mode and the abnormality determination mode, are the same as those described in the first embodiment, and the differences from the first embodiment will be mainly described below.

[0063] In the second embodiment, as shown in the flowchart of FIG. 9 , a d-axis current to q-axis current conversion step (step S215) is added between the current data acquisition step (step S210) and the current frequency domain conversion step (step S220) in the first embodiment. In the drive device 8, as described above, the phase current calculation unit 811 calculates the phase current from the current value obtained from the current sensor 9C, and the Clarke transformation unit 812 and the Park transformation unit 813 convert the phase current into the d-axis current and the q-axis current. Therefore, step S215 can be executed by the Clarke transformation unit 812 and the Park transformation unit 813 in the drive device 8. Therefore, in this example, no new calculation unit is added for the abnormality detection device 1, but it may be installed within the abnormality detection device.

[0064] Then, in step S220, the frequency analysis unit 12 performs frequency analysis on each of the converted q-axis current data. Then, the mechanical angle rotation frequency multiple order component extraction unit 13 extracts the signal intensity of the spectrum peak occurring in the mechanical angle rotation frequency band of the frequency analysis result in step S220. Note that in the dq coordinate system, 2f p / p corresponds to the 1f component of the mechanical angle rotation frequency. Although the second embodiment describes the q-axis current, the d-axis current may also be used.

[0065] Figure 10 shows an example of the frequency spectrum results for the q-axis current when the failure rate is 5% (dashed line) as the fault state and when the failure rate is 0% as the normal state. As shown in Figure 10, the frequency spectrum analysis results for the phase current show that the 1f component of the mechanical rotational frequency (240 Hz) differs more significantly between normal and fault states than the harmonic components of the electrical rotational frequency (360 Hz). This shows that even when using the dq-axis current, the 1f component of the mechanical rotational frequency is more sensitive than the electrical rotational frequency, and by focusing on this component in abnormality detection, it is possible to detect layer shorts early.

[0066] 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 modifications not illustrated are contemplated within the scope of the technology disclosed in this specification. 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.

[0067] For example, in the threshold setting mode, an example has been shown in which the rotating electrical machine 2 actually measures data when it is normal and sets the threshold, but this is not limiting. For example, if the normal value for each model is known in advance, that value may be input, or the corresponding data may be read by selecting the model. In this case, the threshold setting mode may be omitted.

[0068] As described above, the abnormality detection device 1 disclosed herein is configured to include a frequency analysis unit 12 that performs frequency analysis on the physical quantity data of the rotating electric machine 2, a mechanical angle rotation frequency multiple order component extraction unit 13 that extracts specific multiple order components of the mechanical angle rotation frequency from the frequency-analyzed components, and an abnormality determination unit 15 that compares the extracted specific multiple order component with a threshold value Ithreshold to determine whether or not a rare short circuit exists in the rotating electric machine 2. This makes it possible to detect a rare short circuit at an early stage.

[0069] At this time, the mechanical angle rotation frequency multiple order component extractor 13 extracts the frequency-analyzed component from the phase current (I U , I V , I W ), at least one of the first and fifth order components (f p / p,5f p / p) is extracted, and when the frequency-analyzed component is the d-axis current or the q-axis current, the second-order component (2f p / p), the sensitivity is increased, and it becomes possible to reliably detect rare shorts in the early stages.

[0070] The frequency analysis unit 12 can also detect a layer short circuit by performing frequency analysis on at least one of the voltage, current, torque, and rotation speed as physical quantity data.

[0071] The frequency analysis unit 12 can also detect a layer short circuit by using, as physical quantity data, any of the measured values ​​from the sensors, the command values ​​used for drive control of the rotating electric machine 2, and the estimated state values ​​of the physical quantities.

[0072] In particular, by providing an operating condition determination unit 11 that determines whether or not to determine whether or not there is a rare short circuit depending on the power supply frequency that drives the rotating motor 2, so that the frequency extracted by the mechanical angle rotation frequency multiple component extraction unit 13 does not match a peak caused by noise, erroneous detection due to the influence of noise can be prevented and rare short circuits can be detected with high accuracy.

[0073] Furthermore, if the threshold value Ithreshold is set according to the combination of the rotation speed and load factor of the rotating electrical machine 2, accurate determination suited to the operating conditions becomes possible.

[0074] If the rotating electric machine 2 is provided with a threshold setting unit 17 that sets the threshold value Ithreshold based on data obtained from the extracted multiple order components when the rotating electric machine 2 is normal (such as at the beginning of operation), an accurate judgment criterion according to the specifications of the rotating electric machine 2 can be obtained.

[0075] Furthermore, the air conditioner 30 of the present disclosure is equipped with the abnormality detection device 1 described above and the rotating electric machine 2 that drives the compressor 3C, so that it is possible to detect a layer short circuit at an early stage, improving reliability.

[0076] In this case, if the threshold value is set according to at least one of the season, the external environment, and the operating pattern of the air conditioning, it becomes possible to make a more accurate judgment according to the situation.

[0077] Moreover, the abnormality detection method disclosed herein is configured to include a frequency analysis step (step S220) of frequency-analyzing physical quantity data of the rotating electric machine 2, a mechanical angle rotation frequency multiple order component extraction step (step S230) of extracting specific multiple order components of the mechanical angle rotation frequency from the frequency-analyzed components, and an abnormality determination step (step S500) of comparing the extracted specific multiple order component with a threshold value Ithreshold to determine whether or not a rare short circuit has occurred in the rotating electric machine 2. This makes it possible to detect a rare short circuit at an early stage.

[0078] Various aspects of the present disclosure are summarized below as appendices.

[0079] (Supplementary Note 1) An abnormality detection device comprising: a frequency analysis unit that performs frequency analysis on physical quantity data of a rotating electric machine; a mechanical angle rotation frequency multiple order component extraction unit that extracts specific multiple order components of a mechanical angle rotation frequency from the frequency-analyzed components; and an abnormality determination unit that compares the extracted specific multiple order components with a threshold value and determines whether or not there is a layer short in the rotating electric machine.

[0080] (Supplementary Note 2) The abnormality detection device according to Supplementary Note 1, wherein the mechanical angle rotation frequency multiple component extraction unit extracts at least one of a first-order component and a fifth-order component as the specific multiple component when the frequency-analyzed component is a phase current, and extracts a second-order component as the specific multiple component when the frequency-analyzed component is a d-axis current or a q-axis current.

[0081] (Supplementary Note 3) The abnormality detection device according to Supplementary Note 1, wherein the frequency analysis unit performs frequency analysis on at least one of voltage, current, torque, and rotation speed as the physical quantity data.

[0082] (Supplementary Note 4) The abnormality detection device according to Supplementary Note 3, wherein the frequency analysis unit uses, as the physical quantity data, any one of a measurement value from a sensor, a command value used for drive control of the rotating electric machine, and a state estimate value of a physical quantity.

[0083] (Appendix 5) An abnormality detection device according to any one of Appendices 1 to 4, characterized in that it includes an operating condition determination unit that determines whether or not to determine the presence or absence of the layer short circuit depending on the power supply frequency that drives the rotating electric machine, so that the frequency extracted by the mechanical angle rotation frequency multiple order component extraction unit does not match a peak resulting from noise.

[0084] (Supplementary Note 6) The abnormality detection device according to any one of Supplementary Notes 1 to 5, wherein the threshold value is set in accordance with a combination of a rotation speed and a load factor of the rotating electric machine.

[0085] (Appendix 7) The abnormality detection device according to any one of Appendices 1 to 6, further comprising a threshold setting unit that sets the threshold based on data obtained from the extracted multiple order components when the rotating electric machine is normal.

[0086] (Supplementary Note 8) An air conditioner comprising: the abnormality detection device according to any one of Supplementary Notes 1 to 7; and the rotating electric machine that drives a compressor.

[0087] (Supplementary Note 9) The air conditioner according to Supplementary Note 8, wherein the threshold value is set according to at least one of the season, the external environment, and an operating pattern of the air conditioning.

[0088] (Supplementary Note 10) An abnormality detection method comprising: a frequency analysis step of frequency-analyzing physical quantity data of a rotating electric machine; a mechanical angle rotation frequency multiple component extraction step of extracting specific multiple components of a mechanical angle rotation frequency from the frequency-analyzed components; and an abnormality determination step of comparing the extracted specific multiple components with a threshold value to determine whether or not there is a layer short in the rotating electric machine.

[0089] 1: Abnormality detection device, 11: Operating condition determination unit, 12: Frequency analysis unit, 13: Mechanical angle rotation frequency multiple order component extraction unit, 14: Mode switching unit, 15: Abnormality determination unit, 16: Abnormality notification unit, 17: Threshold setting unit, 18: Threshold memory unit, 2: Rotating electric machine, 3: Load equipment, 30: Air conditioner, 3C: Compressor, 9: Physical quantity sensor, Ithreshold: Threshold value.

Claims

1. Frequency analysis unit that performs frequency analysis on physical quantity data of rotating electric machines. A mechanical angular rotation frequency multiple-order component extraction unit that extracts specific multiple-order components of the mechanical angular rotation frequency from the frequency-analyzed components, and An abnormality determination unit compares the extracted specific multiple-order component with a threshold to determine whether or not there is a rare short circuit in the rotating electric machine. An anomaly detection device characterized by being equipped with the following features.

2. The aforementioned mechanical angular rotation frequency multiple component extraction unit is, If the frequency-analyzed component is a phase current, at least one of the first-order and fifth-order components is extracted as the specific multiple-order component. If the frequency-analyzed component is a d-axis current or a q-axis current, a second-order component is extracted as the specific multiple-order component. An anomaly detection device according to claim 1.

3. The anomaly detection device according to claim 1, characterized in that the frequency analysis unit performs frequency analysis on at least one of the following physical quantity data: voltage, current, torque, and rotational speed.

4. The abnormality detection device according to claim 3, characterized in that the frequency analysis unit uses one of the following as the physical quantity data: a measurement value from a sensor, a command value used for drive control of the rotating electric machine, or a state estimate of the physical quantity.

5. An abnormality detection device according to any one of claims 1 to 4, further comprising an operating condition determination unit that determines whether or not to perform the determination of the presence or absence of the rare short circuit according to the power supply frequency driving the rotating electric machine, so that the frequency extracted by the machine angular rotation frequency multiple component extraction unit does not match a peak originating from noise.

6. The abnormality detection device according to any one of claims 1 to 4, characterized in that the threshold is set according to the combination of the rotational speed and load factor of the rotating electric machine.

7. An anomaly detection device according to any one of claims 1 to 4, characterized in that it includes a threshold setting unit that sets the threshold based on data obtained when the rotating electric machine is operating normally, from among the extracted multiple-order components.

8. An anomaly detection device according to any one of claims 1 to 4, and The rotating electric machine that drives the compressor, An air conditioner characterized by having the following features.

9. The air conditioner according to claim 8, characterized in that the threshold is set according to at least one of the season, the external environment, and the operating pattern in the air conditioning.

10. Frequency analysis step for frequency analysis of physical quantity data of a rotating electric machine. A step of extracting a mechanical angular rotation frequency multiple component from the frequency-analyzed components, and An abnormality determination step involves comparing the extracted specific multiple-order component with a threshold to determine whether or not there is a rare short circuit in the rotating electric machine. An anomaly detection method characterized by including