control device

Abstract

The present invention provides a control device that suppresses the fact that the calculation accuracy of Fourier coefficients is determined by the sampling frequency of the AD converter. [Solution] The AD converter 38 converts an analog signal into digital data at a sampling frequency that conforms to the requirements of the motor's synchronous PWM processing. The digital data output by the AD converter 38 is input to the DFT circuit 36. The DFT circuit 36 ​​is equipped with a decimation circuit 54. The decimation circuit 54 inputs the decimated time-series data sampled at the sampling frequency to the FFT 56. The FFT 56 calculates data related to Fourier coefficients using the input data.

Description

Technical Field

[0001] This disclosure relates to a control device.

Background Art

[0002] Patent Document 1 describes a control device that executes a process of correcting a target value of a motor rotation speed according to a frequency fluctuation component obtained by Fourier-transforming the motor rotation speed. This process aims to suppress fluctuations in the output torque of the motor.

[0003] On the other hand, synchronous PWM processing is well-known as a control of the control amount of a motor.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] When executing synchronous PWM processing, it is necessary to sample input data at a sampling frequency that is an integer multiple of the electrical angular frequency. On the other hand, when adjusting the input data used in Fourier transform to the sampling frequency of synchronous PWM processing, its resolution is limited by the sampling frequency and may not achieve the desired resolution.

Means for Solving the Problems

[0006] A control device in the first aspect that solves the above problem is a control device that controls a control variable of a control object, comprising a control circuit, an AD converter that converts an analog signal into digital data, a decimation circuit, and a Fourier transform circuit, wherein the control circuit is configured to perform a process to control the control variable based on first digital data which is the digital data converted by the AD converter from a predetermined analog signal, the Fourier transform circuit is configured to output coefficient data which is data indicating Fourier coefficients based on time-series data of the conversion data to the control circuit, the conversion data is second digital data which is the digital data converted by the AD converter, or data calculated from the second digital data, the second digital data is data synchronized with the first digital data, and the decimation circuit is a circuit that selects the conversion data to be used to calculate the coefficient data by decimating the time-series data.

[0007] In the above configuration, the decimation circuit decimates the conversion data, thereby changing the sampling frequency of the conversion data used to calculate the coefficient data relative to the sampling frequency of the AD converter. This prevents the calculation accuracy of the coefficient data from being limited by the sampling frequency of the AD converter.

[0008] The control device in the second aspect is the control device described in the first aspect, wherein the sampling frequency of the conversion data is equal to the sampling frequency of the first digital data input to the control circuit. In the above configuration, conversion data sampled at the same frequency as the sampling frequency of the digital data input to the control circuit is input to the decimation circuit. By decimating this conversion data, coefficient data can be calculated from the time-series data of the conversion data sampled at a sampling frequency different from the sampling frequency of the digital data input to the control circuit.

[0009] The control device described in the third aspect is the control device described in the second aspect, wherein the controlled object is a rotating electric machine, the output voltage of an inverter is applied to the terminals of the rotating electric machine, the AD converter is configured to convert the predetermined analog signal into first digital data at a frequency that is an integer multiple of the electrical angular frequency of the rotating electric machine, the control circuit is configured to perform synchronous PWM processing, and the synchronous PWM processing is a process that sets the switching pattern of the inverter based on the first digital data as an input variable each time the predetermined analog signal is converted into first digital data at a frequency that is an integer multiple of the frequency.

[0010] In the above configuration, even though the conversion data is sampled at a sampling frequency equal to the sampling frequency defined by the synchronous PWM processing, the sampling frequency of the conversion data used to calculate the coefficient data can be changed by downsampling that conversion data.

[0011] The control device according to the fourth aspect is the control device according to any one of the first to third aspects, wherein the decimation circuit includes a first decimation circuit and a second decimation circuit, the Fourier transform circuit includes a first Fourier transform circuit and a second Fourier transform circuit, the intervals at which the first decimation circuit and the second decimation circuit decimate the conversion data are different from each other, the first Fourier transform circuit is configured to calculate the coefficient data using the conversion data decimated by the first decimation circuit, and the second Fourier transform circuit is configured to calculate the coefficient data using the conversion data decimated by the second decimation circuit.

[0012] In the above configuration, since it includes a first decimation circuit and a second decimation circuit, the first Fourier transform circuit and the second Fourier transform circuit can calculate coefficient data based on conversion data sampled at different sampling frequencies. Therefore, even if the frequencies for which coefficient data is to be obtained are different, the coefficient data can be calculated with high accuracy.

[0013] The control device in the fifth aspect is the control device described in any one of the first to fourth aspects, wherein the E-transform circuit is configured to generate the coefficient data by fast Fourier transform. In the above configuration, by employing the Fast Fourier Transform algorithm, the computational load on the Fourier Transform circuit caused by the calculation of coefficient data can be reduced.

[0014] The control device according to the sixth aspect is the control device according to any one of the first to fifth aspects, wherein the control circuit is configured to input frequency identification data, which is data for identifying the frequency for which Fourier coefficients are to be obtained, to the Fourier transform circuit, and the decimation circuit is configured to select the conversion data to be used for calculating the coefficient data by decimating the conversion data according to the frequency identification data.

[0015] In the above configuration, the Fourier transform circuit thins out the transformation data according to the frequency-specific data, allowing the ratio between the frequency for which the Fourier coefficients are to be calculated and the sampling frequency to be changed. The control device described in the seventh aspect is a control device described in any one of the above three to six aspects (excluding those not included in the third aspect), wherein the rotating electric machine is a motor that drives a compressor, the output voltage of an inverter is applied to the terminals of the motor, the Fourier transform circuit is configured to calculate coefficient data corresponding to a rational multiple of the mechanical angular frequency of the motor using the conversion data thinned by the thinning circuit, the control circuit is configured to perform an output voltage adjustment process, and the output voltage adjustment process is configured to operate the output voltage of the inverter based on the coefficient data as an input variable.

[0016] The torque applied by the compressor to the motor fluctuates according to the rotation angle of the compressor. Therefore, the load torque applied by the compressor to the motor is a rational number multiple of the motor's mechanical angle. In evaluating the effect of this load torque, the amplitude of the fluctuation component with a frequency that is an integer multiple of the fluctuation period of the load torque applied by the compressor to the motor is useful information. However, if conversion data sampled at a sampling frequency determined by the requirements of synchronous PWM processing is used, there is a risk that the coefficient data corresponding to the frequency of that fluctuation component cannot be calculated with high accuracy. Therefore, in the above configuration, by downsampling the conversion data, the coefficient data related to the Fourier coefficient of this fluctuation component can be calculated with the desired resolution. Then, by manipulating the output voltage of the inverter based on this coefficient data, torque fluctuations can be suppressed.

[0017] The control device of the eighth aspect is configured such that the output voltage of an inverter is applied to the terminals of the rotating electric machine, the control device described in any one of the above aspects 3 to 7 (excluding those not included in the third aspect), the Fourier transform circuit is configured to calculate coefficient data corresponding to a frequency 3n times the electrical angular frequency of the rotating electric machine (where n is a natural number of 1 or more) using the conversion data thinned by the thinning circuit, the control circuit is configured to perform an output voltage adjustment process, and the output voltage adjustment process is configured to operate the output voltage of the inverter based on the coefficient data as an input variable.

[0018] Spatial harmonics in rotating electric machines tend to exhibit prominent frequency components at 3n times the electrical angular frequency. However, using conversion data sampled at a sampling frequency determined by the requirements of synchronous PWM processing may result in inaccurate calculation of coefficient data corresponding to 3n times the electrical angular frequency. Therefore, in the above configuration, the above coefficient data can be calculated with the desired resolution by downsampling the conversion data. By then manipulating the inverter output voltage based on the coefficient data of the frequency component at 3n times the electrical angular frequency, the reduction in controllability of the controllable variable of the rotating electric machine caused by spatial harmonics can be suppressed.

[0019] The control device according to the ninth aspect is the control device described in any one of the above aspects 3 to 8 (excluding those that do not include the matters of the third aspect), wherein the control circuit is configured to execute a diagnosis process, and the Fourier transform circuit is configured to calculate the coefficient data corresponding to the frequency that is a rational multiple of the mechanical angular frequency of the rotating electrical machine by using the conversion data thinned out by the thinning-out circuit, and the diagnosis process is a process of diagnosing the presence or absence of an abnormality in the control target based on the coefficient data as an input variable.

[0020] Since the frequency characteristic of the physical quantity caused by the operation of the rotating electrical machine tends to be a frequency that is a rational multiple of the rotation speed, the value of the characteristic frequency component tends to differ depending on the presence or absence of an abnormality in the rotating electrical machine or the device driven by the rotating electrical machine. On the other hand, if the conversion data sampled at the sampling frequency determined by the requirement of the synchronous PWM process is used, there is a possibility that the coefficient data corresponding to the above characteristic frequency cannot be calculated with high accuracy. Therefore, in the above configuration, by thinning out the conversion data, the coefficient data can be calculated with a desired resolution. As a result, the diagnosis of the presence or absence of an abnormality can be executed with high accuracy.

Brief Description of the Drawings

[0021] [Figure 1] It is a block diagram showing the configuration of the air conditioning system according to the first embodiment. [Figure 2] It is a block diagram showing the process executed by the CPU of the control device according to the same embodiment. [Figure 3] It is a block diagram showing the configuration of the DFT circuit of the control device according to the same embodiment. [Figure 4] It is a flowchart showing the procedure of the process executed by the CPU according to the same embodiment. [Figure 5] It is a flowchart showing the procedure of the process executed by the DFT circuit according to the same embodiment. [Figure 6] It is a block diagram showing the process executed by the CPU of the control device according to the second embodiment. [Figure 7]It is a block diagram showing the configuration of the DFT circuit of the control device according to the third embodiment. [Figure 8] It is a flowchart showing the procedure of the process executed by the CPU according to the fourth embodiment. [Figure 9] It is a flowchart showing the procedure of the process executed by the CPU according to the embodiment.

Mode for Carrying Out the Invention

[0022] <First Embodiment> Hereinafter, the first embodiment will be described with reference to the drawings. 「Premise Configuration」 Fig. 1 shows the configuration of the air conditioning system according to the present embodiment.

[0023] The air conditioner 10 receives power supply from the system power supply 20 as an example. The air conditioner 10 includes a compressor 12. The rotating shaft 14a of the motor 14 is mechanically connected to the rotating shaft 12a of the compressor 12. The motor 14 is a three-phase synchronous motor as an example. The motor 14 may be, for example, an embedded magnet synchronous motor. The output voltage of the inverter 16 is applied to each terminal of the motor 14. The inverter 16 is a circuit that converts the voltage of a DC voltage source into an AC voltage and applies it to the motor 14.

[0024] The inverter 16 includes three sets of series-connected bodies of the switching element S#p of the upper arm and the switching element S#n of the lower arm as "#=u, v, w". Freewheel diodes D#p and D#n are connected in antiparallel to each of the switching elements S#p and the switching element S#n.

[0025] Power of the system power supply 20 is supplied to the input terminals of the inverter 16 via the AC-DC conversion circuit 18. A smoothing capacitor 17 is provided between the AC-DC conversion circuit 18 and the inverter 16.

[0026] The control device 30 includes a CPU 32, memory 34, DFT circuit 36, and AD converter 38. The CPU 32 controls the control amount of the motor 14, which is the object controlled by the control device 30, by executing a program stored in the memory 34. The DFT circuit 36 ​​and AD converter 38 are hardware processing circuits.

[0027] The CPU 32 refers to the mechanical angle θm of the motor 14 detected by the rotation angle sensor 40 in order to control the controlled quantity. The CPU 32 also refers to the vibration detection signal O of the compressor 12 detected by the vibration sensor 42. The CPU 32 also refers to the current I flowing through the motor 14 detected by the current sensor 44. The current sensor 44 is configured, for example, with a shunt resistor provided on the negative DC bus of the inverter 16 and a voltage sensor that detects the voltage drop across the shunt resistor. In other words, the current I in this embodiment is the current flowing through the negative DC bus. The CPU 32 also refers to the input voltage Vin of the inverter 16 detected by the voltage sensor 46.

[0028] "CPU32 processing" Figure 2 shows a portion of the processing performed by the CPU 32. The processing shown in Figure 2 is achieved by the CPU 32 repeatedly executing a program stored in memory 34, for example, at a predetermined cycle.

[0029] The target angular velocity setting process M10 is the process of setting the target angular velocity ω*0, which is the target value of the rate of change of the machine angle θm. The correction process M12 is the process of substituting the value obtained by correcting the target angular velocity ω*0 by the correction amount Δωc, which will be described later, into the target angular velocity ω*. The angular velocity calculation process M14 is the process of calculating the angular velocity ωm based on the machine angle θm, which has been converted into digital data by the AD converter 38, as the input variable. The deviation calculation process M16 is the process of calculating the deviation Δω, which is the value obtained by subtracting the angular velocity ωm from the target angular velocity ω*. The current command value setting process M18 is the process of calculating the current command value id* for the d axis and the current command value iq* for the q axis based on the deviation Δω as the input variable.

[0030] The electrical angle calculation process M19 is a process that assigns the remainder obtained by dividing the value obtained by multiplying the mechanical angle θm, which has been converted into digital data by the AD converter 38, by the pole pair number p, by 360, to the electrical angle θe. The two-phase conversion process M20 is a process that calculates the d-axis current id and the q-axis current iq based on the input variables: the current I, which has been converted into digital data by the AD converter 38, the switching pattern, and the electrical angle θe. The switching pattern is determined by the on and off states of the switching elements S#p and S#n of the inverter 16. The switching pattern determines the voltage vector of the output voltage of the inverter 16. The deviation calculation process M22 is a process that calculates the q-axis current deviation, which is the value obtained by subtracting the q-axis current iq from the q-axis current command value iq*. The deviation calculation process M24 is a process that calculates the d-axis current deviation, which is the value obtained by subtracting the d-axis current id from the d-axis current command value id*.

[0031] The voltage command value setting process M26 is a process that sets the output voltage command value of the inverter 16 based on the deviation output of the deviation calculation processes M22 and M24, which are input variables, and the q-axis current command value iq* and the d-axis current command value id*. As an example, the voltage command value setting process M26 includes a process to calculate the q-axis voltage command value and the d-axis voltage command value based on the sum of the output values ​​of the proportional element and the differential element, each of which is an input variable of the pair of deviations. Furthermore, the voltage command value setting process M26 includes a process to correct the above voltage command value by non-interference control based on the q-axis current command value iq* and the d-axis current command value id*, which are input variables. The voltage command value setting process M26 may also include a process to correct the above voltage command value for induced voltage compensation based on the rate of change of the electrical angle θe, which is an input variable.

[0032] Voltage command value setting process M26 is a process that outputs amplitude Va and phase δ as variables indicating the output voltage of inverter 16. The operation signal generation process M28 calculates the operation signals g#p and g#n for the switching elements S#p and S#n based on the amplitude Va and phase δ as input variables. In this embodiment, as an example, the operation signals g#p and g#n are calculated by triangular wave comparison PWM processing. In particular, in this embodiment, the operation signals g#p and g#n are calculated by so-called synchronous PWM processing, which sets the carrier frequency to an integer multiple of the electrical angular frequency of the motor 14. In other words, the switching pattern of the inverter 16 is set by synchronous PWM processing.

[0033] Figure 2 illustrates a case where half a period of the U-phase voltage command value vu*, determined by amplitude Va and phase δ, corresponds to two periods of a triangular-shaped carrier. The PWM signal gu in Figure 2 is determined by comparing the magnitudes of the voltage command value vu* and the carrier. Dead time compensation is applied to the PWM signal gu to generate the final operation signals gup and gun.

[0034] The vibration suppression process M30 is a process that calculates a correction amount Δωc to suppress the vibration of the compressor 12. This vibration is synchronized with the change in the rotation angle of the rotating shaft 12a of the compressor 12. The input variable for the vibration suppression process M30 is the Fourier amplitude Ift. The Fourier amplitude Ift is calculated by the DFT circuit 36.

[0035] "Regarding DFT Circuit 36" Figure 3 shows the configuration of the DFT circuit 36. The electrical angle calculation unit 50 is a circuit that calculates the electrical angle θe by multiplying the mechanical angle θm, which has been converted into digital data by the AD converter 38, by the pole pair number p and finding the remainder when divided by 360. The q-axis current calculation unit 52 is a process that calculates the q-axis current iq based on the current I, which has been converted into digital data by the AD converter 38 as an input variable. The q-axis current calculation unit 52 refers to the switching pattern of the inverter 16 and the electrical angle θe in order to calculate the q-axis current iq.

[0036] The decimation circuit 54 is a circuit that selects the q-axis current iq to be used in the FFT 56 by decimating the time-series data of the q-axis current iq calculated by the q-axis current calculation unit 52. FFT56 is a circuit that calculates the Fourier amplitude Ift by performing a fast Fourier transform on the q-axis current selected by the decimation circuit 54. As an example, FFT56 calculates the Fourier amplitude Ift using the Cooley-Tukey type FFT algorithm.

[0037] Register 58 stores the target frequency ft, which is the frequency at which the Fourier amplitude Ift is to be calculated, and the required resolution Δf for the Fourier amplitude Ift. The sampling frequency setting unit 60 is a circuit that sets the sampling frequency of the AD converter 38.

[0038] "Settings for Register 58" Figure 4 shows the procedure for setting register 58. The series of processes shown in Figure 4 are realized by the CPU 32 repeatedly executing the program stored in memory 34, for example, at a predetermined cycle. In the following, the step number of each process is represented by a number preceded by "S".

[0039] In the series of processes shown in Figure 4, the CPU 32 first obtains the angular velocity ωm (S10). Then, the CPU 32 multiplies the angular velocity ωm by the coefficient Kt and assigns the result to the target frequency ft (S12). The target frequency ft is the frequency for which the Fourier amplitude Ift is to be determined. The coefficient Kt is a coefficient that converts the angular velocity ωm to the frequency of the load torque caused by the compressor 12. The coefficient Kt is, for example, an integer of 1 or more.

[0040] Next, the CPU 32 sets the resolution Δf (S14). Then, the CPU 32 outputs the target frequency ft and the resolution Δf to the DFT circuit 36 ​​(S16). When the CPU 32 completes the process in S16, it terminates the series of processes shown in Figure 4.

[0041] "Operation of DFT circuit 36" Figure 5 shows the operation of the DFT circuit 36 ​​according to the setting of the register 58. The sampling frequency setting unit 60 first multiplies the electrical angular frequency fe by a value of 2 raised to the power of "n" and assigns this value to the PWM frequency fpwm (S20). "n" is an integer greater than or equal to 1. Then, the sampling frequency setting unit 60 outputs the PWM frequency fpwm to the AD converter 38 (S22).

[0042] Meanwhile, the thinning circuit 54 searches for a value for the thinning interval M that satisfies the following conditions A to D (S24). Condition A: This condition states that the number of operations N required to calculate the Fourier amplitude Ift is 2 to the power of "t". "t" is an integer greater than or equal to 1. This condition is based on the fact that, in general, the number of operations in the FFT algorithm is a power of 2.

[0043] Condition B: This condition states that the value obtained by dividing the sampling frequency fs by the number of operations N is equal to the resolution Δf. This condition is necessary to satisfy the set resolution Δf. Condition C: The value obtained by dividing the PWM frequency fpwm by the decimation interval M is equal to the sampling frequency fs.

[0044] Condition D: This condition states that the specified variable k, which determines the target frequency ft, is equal to the value obtained by dividing the target frequency ft by the resolution Δf. Here, the specified variable k is a variable that appears in the independent variable of the exponential function in the discrete Fourier transform, "2·π·j·k·m / (2^t)". Here, "2^t" indicates 2 to the power of t. The variable m is a variable used to add the values ​​of the exponential function from 0 to 2^t-1. The imaginary unit is written as "j".

[0045] The decimation circuit 54 sets the decimation interval M obtained by the process in S24 as the actual decimation interval (S26). The decimation circuit 54 also outputs the value of the specified variable k obtained by the process in S24 to the FFT 56 (S28).

[0046] Furthermore, once the S28 process is complete, the process shown in Figure 5, which is executed periodically, is temporarily terminated. "The operation and effects of this embodiment" The CPU 32 controls the angular velocity of the motor 14 using synchronous PWM processing. Therefore, the CPU 32 samples input variables for the synchronous PWM processing, such as the current I, at a PWM frequency fpwm that is an integer multiple of the electrical angular frequency of the motor 14.

[0047] On the other hand, the DFT circuit 36 ​​shares the AD converter 38 with the CPU 32. Therefore, the output of the AD converter 38 input to the DFT circuit 36 ​​is updated with a period that is the reciprocal of the PWM frequency fpwm.

[0048] When FFT56 calculates the Fourier amplitude Ift using digital data obtained by sampling, where the PWM frequency fpwm is the sampling frequency, it may be difficult to achieve the desired resolution.

[0049] Therefore, the data that has been thinned by the thinning circuit 54 is input to the FFT56. This allows the sampling frequency of the digital data input to the FFT56 to be changed relative to the PWM frequency fpwm. As a result, the FFT56 can calculate the Fourier amplitude Ift with the desired resolution.

[0050] For example, if the angular velocity ωm is "35Hz", the number of pole pairs p is "3", and the variable n in S20 is "8", then the PWM frequency fpwm is set to "105 × 256 = 26880Hz". The CPU32 then updates the switching pattern based on 256 sampling values ​​of current I per electrical angular period.

[0051] On the other hand, in this case, for example, if both the target frequency ft and the resolution Δf are "35", the decimation interval M by the decimation circuit 54 is set to "1". This allows the FFT 56 to use all of the current I input to the DFT circuit 36 ​​to calculate the Fourier amplitude Ift with the desired resolution. In other words, without decimation by the decimation circuit 54, the FFT 56 can use all of the current I input to the DFT circuit 36 ​​to calculate the Fourier amplitude Ift with the desired resolution. However, the number of calculations N in this case is 256.

[0052] However, for example, if the angular velocity ωm and target frequency ft are "30Hz", the number of pole pairs p is "3", the variable n in S20 is "8", and the resolution Δf is "10", then without decimation, the number of calculations required to satisfy the resolution Δf requirement would be "9·256". This does not satisfy the requirement of FFT56 that the number of calculations N be a power of 2.

[0053] According to the embodiment described above, the following effects and benefits can be obtained. (1-1) The sampling frequency of the current I input to the DFT circuit 36 ​​is equal to the sampling frequency of the current I used by the CPU 32 to control the motor 14. Therefore, the sampling frequency of the current input to the DFT circuit 36 ​​is completely determined by the requirements of the synchronous PWM processing. For this reason, the use of the decimation circuit 54 is particularly valuable.

[0054] (1-2) The DFT circuit 36 ​​is equipped with an FFT 56 that calculates the Fourier amplitude Ift using a fast Fourier transform algorithm. Therefore, the Fourier amplitude Ift can be calculated efficiently.

[0055] (1-3) The CPU 32 inputs the target frequency ft and resolution Δf into register 58, and the decimation circuit 54 sets the decimation interval M. This allows the CPU 32 to calculate the Fourier amplitude Ift of the target frequency ft so as to satisfy the resolution Δf.

[0056] (1-4) The torque applied by the compressor 12 to the motor 14 fluctuates according to the rotation angle of the compressor 12. Therefore, the load torque applied by the compressor 12 to the motor 14 is a rational number multiple of the mechanical angle θm of the motor 14. In evaluating the effect of this load torque, the amplitude of the fluctuation component at a frequency that is an integer multiple of the fluctuation period of the load torque applied by the compressor 12 to the motor 14 is useful information. Therefore, the CPU 32 set the target frequency ft to a rational number multiple of the electrical angular frequency. Then, the target angular velocity ω*0 was corrected according to the Fourier amplitude Ift corresponding to the target frequency ft. This makes it possible to suppress torque fluctuations caused by fluctuations in the load torque of the compressor 12.

[0057] <Second Embodiment> The second embodiment will be described below, focusing on the differences from the first embodiment, with reference to the drawings.

[0058] Figure 6 shows the processes executed by CPU 32. For convenience, the same reference numerals are used in Figure 6 for processes corresponding to those shown in Figure 2. As shown in Figure 6, the deviation calculation process M16 in this embodiment inputs the deviation between the target angular velocity ω* set by the target angular velocity setting process M10 and the angular velocity ωm to the current command value setting process M18.

[0059] Phase correction process M30a calculates a correction amount Δδ to correct the phase δ of the output voltage 16 in order to suppress torque ripple caused by spatial harmonics. Correction process M12a is a process that inputs the value obtained by correcting the phase δ set by voltage command value setting process M26 with the correction amount Δδ to the operation signal generation process M28.

[0060] Phase correction process M30a calculates a correction amount Δδ based on the Fourier amplitude Vft as an input variable. The target frequency ft, which corresponds to the Fourier amplitude Vft, is a multiple of 3 of the angular velocity ωm. More specifically, the target frequency ft is a multiple of 3 of the electrical angular frequency of the motor 14. This setting is based on the fact that the spatial harmonics of the motor 14 tend to have prominent frequency components at 3n times the electrical angular frequency.

[0061] In this embodiment, the input voltage Vin is converted into digital data by the AD converter 38 and input to the DFT circuit 36. The AD converter 38 also sequentially converts analog signals other than the current I into digital data at the sampling period of the current I.

[0062] Here, for example, let's consider the case where the angular velocity ωm is "23Hz", the number of pole pairs p is "4", the target frequency ft is "23Hz", and the resolution is "1Hz". In this case, the electrical angular frequency is "92Hz". The sampling frequency of the AD converter 38 is set to an integer multiple of "92Hz", for example, "92 × 128Hz". Then, the decimation circuit 54 is set to a decimation interval M of "92". This makes it possible to achieve a resolution of "1Hz" with a calculation count N of "128".

[0063] <Third Embodiment> The third embodiment will be described below, focusing on the differences from the first embodiment, with reference to the drawings.

[0064] Figure 7 shows the configuration of the DFT circuit 36 ​​according to this embodiment. Note that, for convenience, the same reference numerals are used for the circuits corresponding to those shown in Figure 3 in Figure 7. As shown in Figure 7, this embodiment includes two decimation circuits 54(1) and 54(2) and two FFTs 56(1) and 56(2). In this embodiment, the detection signal O, which has been converted into digital data by the AD converter 38, is input to the DFT circuit 36. The AD converter 38 also converts the detection signal O into digital data at a sampling period equal to the current I.

[0065] Figure 8 shows the procedure for setting register 58. The series of processes shown in Figure 8 are realized by the CPU 32 repeatedly executing the program stored in memory 34, for example, at a predetermined cycle. In Figure 8, for convenience, the processes corresponding to the processes shown in Figure 4 are assigned the same step numbers.

[0066] In the series of processes shown in Figure 8, when the CPU 32 completes the process in S10, it substitutes the value obtained by multiplying the angular velocity ωm by the coefficient Kt(1) into the target frequency ft(1), and also substitutes the value obtained by multiplying the angular velocity ωm by the coefficient Kt(2) into the target frequency ft(2) (S12a). The coefficients Kt(1) and Kt(2) are coefficients that convert the angular velocity ωm to a frequency that becomes significant due to an abnormality in the compressor 12. The coefficients Kt(1) and Kt(2) are rational numbers.

[0067] The CPU 32 sets the resolution Δf(1) for the target frequency ft(1) and the resolution Δf(2) for the target frequency ft(2) (S14a). Then, the CPU 32 outputs the target frequencies ft(1) and ft(2) and the resolutions Δf(1) and Δf(2) to the DFT circuit 36 ​​(S16a).

[0068] Furthermore, when CPU32 completes the processing of S16a, it temporarily terminates the series of processes shown in Figure 8. Figure 9 shows the procedure for diagnosing whether or not there is an abnormality in the compressor 12 based on the detection signal O. The process shown in Figure 9 is realized by the CPU 32 repeatedly executing a program stored in memory 34, for example, at a predetermined period.

[0069] In the series of processes shown in Figure 9, the CPU 32 first obtains the Fourier amplitude Oft(1) related to the target frequency ft(1) calculated by FFT56 (S30). Then, the CPU 32 determines whether the Fourier amplitude Oft(1) is greater than or equal to the threshold Oft1th (S32). This process determines whether or not a first abnormality has occurred in the compressor 12.

[0070] If the CPU 32 determines that the Fourier amplitude Oft(1) is greater than or equal to the threshold Oft1th (S32: YES), it determines that there is a first abnormality in the compressor 12 (S34). When the CPU 32 completes the process in S34, or when it makes a negative determination in the process in S32, it obtains the Fourier amplitude Oft(2) related to the target frequency ft(2) calculated by FFT56 (S36). Then, the CPU 32 determines whether the Fourier amplitude Oft(2) is greater than or equal to the threshold Oft2th (S38). This process determines whether a second abnormality has occurred in the compressor 12.

[0071] If the CPU 32 determines that the Fourier amplitude Oft(2) is greater than or equal to the threshold Oft2th (S38: YES), it determines that there is a second abnormality in the compressor 12 (S40). Furthermore, when CPU32 completes the process in S40, or when it makes a negative determination in the process in S38, it temporarily terminates the series of processes shown in Figure 9.

[0072] For example, let's consider the case where the angular velocity ωm is "60Hz", the number of pole pairs p is "3", the target frequencies ft(1) and ft(2) are "60Hz" and "90Hz" respectively, the resolutions Δf(1) and Δf(2) are "2Hz" and "3Hz" respectively, and the variable n of S20 is "7". In this case, the PWM frequency fpwm becomes 180 × 128 = 23040Hz. The decimation circuit 54(1) is set to a decimation interval M of 90. This allows us to calculate the Fourier amplitude Oft(1) with a resolution Δf(1) of 2Hz and a target frequency ft(1) of 20Hz. The decimation circuit 54(2) is set to a decimation interval M of 60. This allows us to calculate the Fourier amplitude Oft(3) with a resolution Δf(2) of 3Hz and a target frequency ft(2) of 30Hz.

[0073] According to the embodiment described above, in addition to the effects similar to those of (1-1) to (1-3) of the first embodiment, the following further actions and effects can be obtained. (3-1) The characteristic frequencies of the physical quantities resulting from the operation of the motor 14 tend to be rational multiples of the electrical angular frequency. Therefore, the values ​​of these characteristic frequency components tend to differ depending on whether there is a malfunction in the motor 14 or the compressor 12 driven by the motor 14. So the CPU 32 sets the target frequencies ft(1) and ft(2) to these frequencies. This allows the CPU 32 to diagnose the presence or absence of a malfunction based on the Fourier amplitudes Oft(1) and Oft(2) as input variables.

[0074] (3-2) Multiple decimation circuits 54 and FFT 56 are provided. This allows for the simultaneous calculation of Fourier amplitudes Oft(1) and Oft(2) at different target frequencies ft(1) and ft(2).

[0075] <Correspondence> The correspondence between the matters in the above embodiment and the matters described in the "Means for Solving the Problems" section is as follows. Below, the correspondence is shown for each number of the viewpoint described in the "Means for Solving the Problems" section. [1,2,5] The control circuit corresponds to the CPU 32. The AD converter corresponds to the AD converter 38. The decimation circuit corresponds to the decimation circuits 54, 54(1), and 54(2). The Fourier transform circuit corresponds to the FFT 56, 56(1), and 56(2). The first digital data, the second digital data, and the conversion data correspond to the current I, current I, and q-axis current iq in the first embodiment, respectively. The first digital data, the second digital data, and the conversion data correspond to the current I, input voltage Vin, and input voltage Vin in the second embodiment, respectively. The first digital data, the second digital data, and the conversion data correspond to the current I, detection signal O, and detection signal O in the third embodiment, respectively. [3] The rotating electric machine corresponds to the motor 14. [4] The first and second decimation circuits correspond to decimation circuits 54(1) and 54(2). The first and second Fourier transform circuits correspond to FFTs 56(1) and 56(2). [6] The frequency identification data corresponds to the data indicating the target frequency ft. [7] The output voltage adjustment process corresponds to the process shown in Figure 2. [8] The output voltage adjustment process corresponds to the process shown in Figure 6. [9] The diagnostic process corresponds to the process shown in Figure 9.

[0076] <Other Embodiments> Furthermore, this embodiment can be implemented with the following modifications. This embodiment and the following modifications can be combined with each other to the extent that they do not contradict each other technically.

[0077] "Regarding the sampling frequency setting unit 60" In the above embodiment, the resolution Δf of the coefficient data was not specifically mentioned when setting the PWM frequency fpwm by the sampling frequency setting unit 60, but the resolution Δf may be taken into consideration. Specifically, for example, the DFT circuit 36 ​​may perform a process to simultaneously determine the decimation interval M, the PWM frequency fpwm, and the number of calculations N. This can more reliably suppress the situation in which it becomes impossible to set coefficient data that satisfies the resolution Δf despite performing the decimation process.

[0078] It is not essential that the sampling frequency setting unit 60 be packaged together with the FFT56, etc. It is not essential that the control device 30 includes a sampling frequency setting unit 60 as a hardware processing circuit. For example, the processing of S20 and S22 may be performed by the CPU 32.

[0079] "About thinning circuits" The number of decimation circuits provided by the control device 30 is not limited to one or two. For example, it may be three or more.

[0080] It is not essential that the decimation circuit be packaged together with the Fourier transform circuit. It is not mandatory for the decimation circuit to perform the processes S24-S28. For example, if the target frequency Ft is predetermined to be unique, the decimation circuit may simply be a circuit that performs decimation according to an invariant decimation interval M.

[0081] "About Fourier transform circuits" The number of decimation circuits and the number of Fourier transform circuits are not necessarily the same. For example, one decimation circuit may be accompanied by multiple Fourier transform circuits. Here, if the sampling timing of the first data in the time series data of the transformation data used by different Fourier transform circuits to calculate the coefficient data is different, the update cycle of the coefficient data can be shortened.

[0082] • It is not essential that the Fast Fourier Transform algorithm is a Cooley-Tukey type FFT algorithm. It is not essential that the Fourier transform circuit is a circuit that calculates coefficient data based on the Fast Fourier Transform algorithm.

[0083] "About DFT circuits" It is not essential that the DFT circuit 36 ​​includes a q-axis current calculation unit 52. Furthermore, it is not essential that the q-axis current iq input to the decimation circuit 54 is a value calculated by a hardware processing circuit. For example, the q-axis current iq input to the decimation circuit 54 may be a value calculated by the CPU 32. "Regarding the data for conversion" The conversion data for the motor 14's current is not limited to the q-axis current iq. The conversion data for the motor 14's current may be, for example, the d-axis current id. The conversion data for the motor 14's current may be, for example, the current I. Alternatively, for example, the conversion data for the motor 14's current may be the output line current of the inverter 16. Alternatively, for example, the conversion data for the motor 14's current may be the α-phase current or β-phase current obtained by two-phase conversion of the output line current of the inverter 16. The conversion data for the motor 14's current may be, for example, the M-axis current or T-axis current obtained by coordinate transformation based on the angle of the primary magnetic flux of the motor 14. Note that the coordinate transformation used to generate the conversion data for the motor 14's current is not limited to those exemplified above. The conversion data for the motor 14's current does not necessarily have to be data for a current that fluctuates moment by moment. The conversion data for the motor 14's current may be, for example, a quantity proportional to the amplitude of the output line current, α-phase current, or β-phase current, or the square of the same quantity. Alternatively, for example, the conversion data for the motor 14's current may be a quantity relating to the difference between the maximum and minimum values ​​of the q-axis current or d-axis current. The conversion data for the motor 14 voltage is not limited to the input voltage Vin. For example, the conversion data for the motor 14 voltage may be the output line voltage of the inverter 16. Alternatively, for example, the conversion data for the motor 14 voltage may be the d-axis voltage or the q-axis voltage. Note that the coordinate transformation used to generate the conversion data for the motor 14 voltage is not limited to a coordinate transformation to the dq axes. The conversion data for the motor 14 voltage does not necessarily have to be voltage data that fluctuates moment by moment. The conversion data for the motor 14 voltage may be a quantity proportional to the amplitude of the output line voltage, or the square of the same quantity. For example, the conversion data for the motor 14 voltage may be a quantity relating to the difference between the maximum and minimum values ​​of the d-axis voltage or the q-axis voltage. The conversion data may be, for example, the instantaneous power of motor 14. The conversion data related to power may be, for example, instantaneous imaginary power, apparent power, active power, or reactive power. The conversion data may be, for example, armature flux linkage. The conversion data related to magnetic flux may be, for example, the d-axis flux or the q-axis flux. The conversion data may be physical quantities estimated using a physical model based on sensor detection values, such as magnetic pole position or motor 14 torque. The analog signal used to generate the digital data for conversion may be, for example, a magnetic pole position signal obtained from an encoder or the like. Alternatively, the analog signal used to generate the digital data for conversion may be a sound signal obtained from a sound sensor that senses the sound around the motor 14. Alternatively, the analog signal used to generate the digital data for conversion may be a temperature signal obtained from a temperature sensor that senses the temperature around the motor 14. "Regarding coefficient data" • It is not necessary for the coefficient data to consist solely of data relating to the amplitude of the target frequency component. The coefficient data may, for example, consist of data relating to both the amplitude and phase of the target frequency component.

[0084] "Regarding countermeasures for load torque pulsation" It is not mandatory that the manipulated variable for the load torque suppression process be the correction amount Δωc of the target angular velocity ω*. The manipulated variable for the load torque suppression process may be, for example, one of the two values ​​of the output voltage of the inverter 16: amplitude Va and phase δ.

[0085] The Fourier coefficients used as input for the processing to address load torque pulsation are not limited to Fourier coefficients for a single frequency. The Fourier coefficients used as input for the processing to address load torque pulsation may be Fourier coefficients for multiple different frequencies.

[0086] It is not mandatory that the input variable for FFT56 used to calculate the Fourier amplitude Ift, which is the input for the processing to address load torque pulsation, be the q-axis current iq. The input variable for FFT56 may be, for example, the U-phase current. Alternatively, the input variable for FFT56 may be, for example, the current I.

[0087] "Regarding the suppression of torque ripple caused by spatial harmonics" • It is not essential that the manipulated variable for suppressing torque ripple caused by spatial harmonics is the phase δ. The manipulated variable for suppressing torque ripple caused by spatial harmonics may be, for example, the amplitude Va of the output voltage of the inverter 16.

[0088] The input variable for the torque ripple suppression process caused by spatial harmonics does not necessarily have to be a Fourier coefficient for a single harmonic. For example, the input variable for the torque ripple suppression process caused by spatial harmonics may be a Fourier coefficient for multiple different harmonics.

[0089] "Regarding output voltage adjustment processing" The current feedback processing for the dq axis is not limited to processing using PD control. For example, the current feedback processing for the dq axis may use PID control. Furthermore, the current feedback processing for the dq axis is not limited to classical control, but may also use model predictive control, for example.

[0090] • It is not mandatory for the output voltage adjustment process to include current feedback processing for the dq axis. For example, the output voltage adjustment process may be related to direct torque control.

[0091] "Regarding the control amount of motor 14" It is not essential that the controlled variable of motor 14 is angular velocity ωm. The controlled variable of motor 14 may be, for example, the torque of motor 14.

[0092] "About Rotating Electric Machines" The rotating electric machine is not limited to the motor 14 that drives the compressor 12. For example, it could be a motor mounted on a circulator.

[0093] "About motors" The motor is not limited to a synchronous motor with embedded magnets; for example, it could be an induction motor. Alternatively, the motor could be a brushed DC motor.

[0094] "Regarding software processing circuits" • The software processing circuit does not necessarily have to be a CPU. For example, the software processing circuit may include both a CPU and a GPU.

[0095] "Regarding control devices" The control device is not limited to the motor 14 that drives the compressor 12.

[0096] Although embodiments have been described above, it should be understood that various modifications to the form and details are possible without departing from the spirit and scope of the claims. [Explanation of Symbols]

[0097] 10...Air conditioner 12... Compressor 14…motor 30...Control device

Claims

1. A control device that controls the control amount of the controlled object, It comprises a control circuit (32), an A / D converter (38) that converts analog signals to digital data, a decimation circuit (54), and a Fourier transform circuit (56), The control circuit (32) is configured to perform a process to control the control amount based on the first digital data, which is the digital data obtained by the AD converter (38) from a predetermined analog signal. The Fourier transform circuit (56) is configured to output coefficient data, which is data indicating the Fourier coefficients, to the control circuit (32) based on the time-series data of the transformation data. The conversion data is a second digital data which is the digital data converted by the AD converter (38), or data calculated from the second digital data. The second digital data is data that is synchronized with the first digital data, The aforementioned decimation circuit (54) is a control device that selects the conversion data to be used for calculating the coefficient data by decimating the time-series data.

2. The control device according to claim 1, wherein the sampling frequency of the conversion data is equal to the sampling frequency of the first digital data input to the control circuit (32).

3. The controlled object is a rotating electric machine (14), The terminals of the rotating electric machine (14) are configured to receive the output voltage of the inverter (16). The AD converter (38) is configured to convert the predetermined analog signal into first digital data at a frequency that is an integer multiple of the electrical angular frequency of the rotating electric machine (14). The control circuit (32) is configured to perform synchronous PWM processing, The control device according to claim 2, wherein the synchronous PWM processing is a process of setting the switching pattern of the inverter (16) based on the first digital data as an input variable each time the predetermined analog signal is converted into first digital data at an integer multiple frequency.

4. The thinning circuit (54) includes a first thinning circuit (54(1)) and a second thinning circuit (54(2)), The Fourier transform circuit (56) includes a first Fourier transform circuit (56(1)) and a second Fourier transform circuit (56(2)), The first decimation circuit (54(1)) and the second decimation circuit (54(2)) have different intervals for decimating the conversion data. The first Fourier transform circuit (56(1)) is configured to calculate the coefficient data using the transformation data that has been thinned out by the first thinning circuit (54(1)), The control device according to claim 1, wherein the second Fourier transform circuit (56(2)) is configured to calculate the coefficient data using the transformation data that has been thinned by the second thinning circuit (54(2)).

5. The control device according to claim 1, wherein the Fourier transform circuit (56) is configured to generate the coefficient data by fast Fourier transform.

6. The control circuit (32) is configured to input frequency identification data, which is data for identifying the frequency for which the Fourier coefficients are to be determined, to the Fourier transform circuit (56). The control device according to claim 1, wherein the decimation circuit (54) is configured to select the conversion data to be used for calculating the coefficient data by decimating the conversion data according to the frequency identification data.

7. The aforementioned rotating electric machine (14) is a motor that drives the compressor, The terminals of the motor are configured to have the output voltage of the inverter (16) applied to them. The Fourier transform circuit (56) is configured to use the transformation data thinned by the thinning circuit (54) to calculate coefficient data corresponding to frequencies that are rational multiples of the mechanical angular frequency of the motor. The control circuit (32) is configured to perform output voltage adjustment processing, The control device according to claim 3, wherein the output voltage adjustment process is configured to operate the output voltage of the inverter (16) based on the coefficient data as an input variable.

8. The terminals of the rotating electric machine (14) are configured to have the output voltage of the inverter (16) applied to them. The Fourier transform circuit (56) is configured to use the transformation data thinned by the thinning circuit (54) to calculate coefficient data corresponding to a frequency 3n times the electrical angular frequency of the rotating electric machine (14) (where n is a natural number of 1 or more). The control circuit (32) is configured to perform output voltage adjustment processing, The control device according to claim 3, wherein the output voltage adjustment process is configured to operate the output voltage of the inverter (16) based on the coefficient data as an input variable.

9. The control circuit (32) is configured to perform diagnostic processing, The Fourier transform circuit (56) is configured to use the transformation data thinned by the thinning circuit (54) to calculate the coefficient data corresponding to a rational multiple of the mechanical angular frequency of the rotating electric machine (14). The control device according to claim 3, wherein the diagnostic process is a process of diagnosing whether or not there is an abnormality in the controlled object based on the coefficient data as an input variable.

Images