AC motor drive unit, compressor drive unit, and refrigeration cycle unit

Abstract

The drive unit (101) for the AC motor includes: an adaptive observation unit (1) that adaptively estimates the angular velocity of the rotor (2a) of the AC motor (2); a speed control unit (3) that determines a first torque command that makes the angular velocity command consistent with the average value of the estimated angular velocity; a phase lead calculation unit (6) that calculates the phase lead of the transfer function from the actual angular velocity to the model deviation based on the disturbance frequency; a vibration suppression control unit (5) that determines a second torque command that suppresses the speed pulsation of the AC motor (2) based on the frequency of the load torque pulsation, the model deviation, and the phase lead; and a torque control unit (4) that controls the torque of the AC motor (2) based on the first torque command and the second torque command.

Description

Technical Field

[0001] This disclosure relates to an AC motor drive unit (hereinafter referred to as "drive unit") that drives an AC motor such as an induction motor or a synchronous motor, a compressor drive unit, and a refrigeration cycle device having a compressor drive unit. Background Technology

[0002] In the control of AC motors, when there are pulsations in the load torque or the torque generated by the AC motor, the speed of the AC motor also pulsates. When the speed of the AC motor pulsates, vibrations are also generated in the device equipped with the AC motor, which can sometimes cause problems in terms of noise generation and mechanical strength. In order to address these problems, research has been conducted on control methods to reduce torque pulsation and speed pulsation.

[0003] For example, Patent Document 1 discloses a method for reducing torque and speed ripple in a sensorless manner, in order to reduce costs or to make it applicable to devices where sensors are difficult to install.

[0004] In typical sensorless control, the upper limit of the velocity estimation response is several hundred rad / s. Therefore, under typical sensorless control, the response to high-frequency pulsations is insufficient, making it difficult to accurately estimate pulsations. As a technology to address this problem, Patent Document 1 discloses the following technique: setting up two angular velocity estimation units connected in parallel, thereby improving the velocity estimation response at any frequency to accurately estimate velocity pulsations and suppress high-frequency vibrations.

[0005] Furthermore, Patent Document 2 and Non-Patent Document 1 described below are publicly known documents relating to the technology disclosed herein. The contents of these documents will be referenced in the "Detailed Description" section.

[0006] Existing technical documents

[0007] Patent documents

[0008] Patent Document 1: International Publication No. 2017 / 212794

[0009] Patent Document 2: Japanese Patent Application Publication No. 1-308184

[0010] Non-patent literature

[0011] Non-Patent Literature 1: Kanehara and Koyama: "Speed ​​Sensorless Vector Control Method for Induction Motors Including Low-Speed / Regenerative Regions", Journal of the Institute of Electrical Engineers of Japan, Vol. 120, No. 2, pp. 223-229, 2000 (Kanehara and Koyama: "Speed ​​Sensitive Vickers Control Method for Induction Motors Including Low-Speed ​​and Regenerative Regions", Journal of the Institute of Electrical Engineers of Japan, Vol. 120, No. 2, pp. 223-229, 2000) Summary of the Invention

[0012] The technical problem that the invention aims to solve

[0013] In the case of driving an AC motor using a sensorless control method, any change to the method of position estimation or speed estimation can have a wide-ranging impact. Therefore, if the position estimation calculation unit or speed estimation calculation unit is changed, it is necessary to conduct a detailed investigation to determine if any adverse conditions occur in the drive unit as a result of the change.

[0014] When adding the method described in Patent Document 1 as an extended function to an existing AC motor drive device, the more successful the marketing of the drive device, the faster the man-hours for reliability evaluation accompanying the function addition increase. Therefore, it is sometimes difficult to implement the technology of Patent Document 1. Especially for general-purpose drive devices, implementing the technology of Patent Document 1 is extremely difficult.

[0015] This disclosure is made in view of the above-mentioned problems, and aims to provide an AC motor drive device that can easily implement extended functions to add to existing functions under sensorless control of the AC motor.

[0016] Technical solutions for solving technical problems

[0017] To solve the aforementioned technical problems and achieve the objectives, the AC motor drive device disclosed herein includes an adaptive observation unit, a speed control unit, a phase lead calculation unit, a vibration suppression control unit, and a torque control unit. The adaptive observation unit adaptively estimates the angular velocity of the rotor of the AC motor driving a mechanical device with periodic load torque pulsations. The speed control unit determines a first torque command that makes the angular velocity command consistent with the average value of the estimated angular velocity. The phase lead calculation unit calculates the phase lead based on the disturbance frequency, which is a transfer function of the model deviation (an internal quantity of the adaptive observation unit) from the actual angular velocity. The vibration suppression control unit determines a second torque command that suppresses the speed pulsations of the AC motor based on the frequency of the load torque pulsations, the model deviation, and the phase lead. The torque control unit controls the torque of the AC motor based on the first and second torque commands.

[0018] Invention Effects

[0019] According to the AC motor drive device disclosed herein, it is possible to easily implement extended functions to add to existing functions. Attached Figure Description

[0020] Figure 1 A block diagram illustrating the structure of the AC motor drive device according to Embodiment 1.

[0021] Figure 2 To show Figure 1 The diagram shows the relationship between the model deviation of the drive device and the speed estimation error.

[0022] Figure 3 To show Figure 1 The diagram shows an example of a Bode plot of the transfer function of the drive device from the actual rotational angular velocity to the model deviation.

[0023] Figure 4 For illustrative purposes Figure 1 A vector diagram illustrating the necessity of phase lead in the drive unit.

[0024] Figure 5 This is the first figure used in the operation description of the drive device in Embodiment 1.

[0025] Figure 6 The second figure is used in the description of the operating wave of the drive device in Embodiment 1.

[0026] Figure 7 To show Figure 1 A block diagram of a first variation of the structure of Embodiment 1 shown.

[0027] Figure 8 To show Figure 1 A block diagram of a second variation of the structure of Embodiment 1 shown.

[0028] Figure 9 This is a hardware structure diagram of the AC motor drive device according to Embodiment 1.

[0029] Figure 10 A block diagram illustrating the structure of the AC motor drive device according to Embodiment 2.

[0030] Figure 11 This is a diagram illustrating an example of the waveform of the load torque of a rotary compressor, which is an example of a load device in Embodiment 2.

[0031] Figure 12 A block diagram illustrating the structure of the AC motor drive device according to Embodiment 3.

[0032] Figure 13 A block diagram illustrating the structure of the AC motor drive device of Embodiment 4.

[0033] Figure 14 To show Figure 13 A cross-sectional view showing the general internal structure of a refrigerant compressor as the driving object.

[0034] Figure 15 To show Figure 14 A cross-sectional view of the internal structure of the compression section of a refrigerant compressor.

[0035] Figure 16 A diagram illustrating the structure of the refrigeration cycle apparatus of Embodiment 5.

[0036] Figure Labels

[0037] 1: Adaptive observation unit; 2: AC motor; 2a: Rotor; 2-1: Rotor; 2-2: Stator; 2-3: Coil; 2-4: Clearance; 3: Speed ​​control unit; 4, 4a: Torque control unit; 5, 5a, 5b, 5d, 5e, 5f, 5g: Vibration suppression control unit; 6, 6d, 6e, 6f, 6g: Phase lead calculation unit; 7: Gain calculation unit; 11: Model deviation calculation unit; 12: Angular velocity estimation unit; 20: Refrigerant compressor; 51, 51a, 51b: Speed ​​pulsation calculator; 52, 53, 52a, 53a, 52b, 53b: I controller; 54, 54a, 54b: AC restorer; 01, 101a, 101b, 101c: Drive unit; 201: Shaft; 202: Compression unit; 203: Suction pipe; 204: Discharge pipe; 205: Piston; 206: Suction port; 207: Discharge port; 208: Discharge valve; 209: Spring; 210: Vane; 211: Sealed container; 212: Cylinder; 213: Compression chamber; 300: Refrigeration cycle unit; 301: Condenser; 302: Liquid receiver; 303: Expansion valve; 304: Evaporator; 305: Piping; 306: Refrigeration cycle loop; 801: Voltage application unit; 802: Current detection unit; 901: Processor; 902: Memory. Detailed Implementation

[0038] The AC motor drive device, compressor drive device, and refrigeration cycle device of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0039] Implementation method 1.

[0040] Figure 1 A block diagram illustrating the structure of the AC motor drive device 101 according to Embodiment 1. Figure 1 The drive device 101 of Embodiment 1 shown is a drive device that estimates the angular velocity ω of the rotor 2a of the AC motor 2. r And so that the estimated angular velocity, i.e. the estimated angular velocity ω, is obtained.＾ r With angular velocity command ω * r Drive the AC motor 2 in a consistent manner.

[0041] The drive unit 101 includes an adaptive observation unit 1, a speed control unit 3, a torque control unit 4, a vibration suppression control unit 5, and a phase lead calculation unit 6.

[0042] The adaptive observation unit 1 is a component for adaptively estimating the angular velocity of the AC motor 2. Specifically, based on the principle of an adaptive observer, the adaptive observation unit 1 uses the voltage vector applied to the AC motor 2 and the current vector flowing through the AC motor 2 to estimate the angular velocity ω of the AC motor 2. r The AC motor 2 is the power source for a mechanical device (not shown). The mechanical device provides the AC motor 2 with periodic load torque pulsations synchronized with the rotation of the AC motor 2.

[0043] The adaptive observation unit 1 includes a model deviation calculation unit 11 and an angular velocity estimation unit 12. The model deviation calculation unit 11 is based on the voltage vector, current vector, and estimated angular velocity ω. ＾ r The model bias ε is calculated. The model bias ε is an internal quantity of the adaptive observation unit 1. The angular velocity estimation unit 12 calculates the estimated angular velocity ω based on the model bias ε. ＾ r .

[0044] Inside the model deviation calculation unit 11, calculations are performed to deduce the state quantities of the AC motor 2 based on its state equation. Examples of state quantities are current and magnetic flux. Furthermore, it is assumed in this document that the AC motor 2 is a typical embedded magnet type synchronous motor, but it is not limited to this. Any other AC motor can be used, as long as the same state equation as described below can be established in the model deviation calculation unit 11. Examples of other AC motors include surface magnet type synchronous motors and induction motors. Additionally, for ease of explanation, each motor is described as a three-phase motor, but it is not limited to this. Each motor can be a motor with other phase numbers, such as a two-phase motor or a five-phase motor.

[0045] In the case of sensorless driving of a synchronous motor with an embedded magnet, the state equation used in the adaptive observation unit 1 is expressed as shown in equation (1) below. In addition, the output equation used in the adaptive observation unit 1 is expressed as shown in equation (2) below.

[0046] [Mathematical Expression 1]

[0047]

[0048] [Mathematical Expression 2]

[0049]

[0050] In equations (1) and (2) above, L d L q R represents the inductance along the d-axis and q-axis. a This represents the armature resistance. ω r ω represents the electrical angular velocity of AC motor 2, and ω1 represents the primary angular frequency. d Represents the d-axis voltage, v q This represents the q-axis voltage. d Represents the d-axis current, i q φ represents the q-axis current. ds φ represents the d-axis stator flux. qs φ represents the q-axis stator flux. dr This represents the rotor flux along the d-axis. h 11 ~h 32 This indicates the observer gain. The symbol "^" represents the estimated value.

[0051] In addition, the primary angular frequency ω1 appearing in the above equation (1) is given as in the following equation (3).

[0052] [Mathematical Expression 3]

[0053]

[0054] In equation (3) above, h 41 and h 42 Compared with the aforementioned h 11 ~h 32 Similarly, the observer gain is represented.

[0055] Furthermore, Equation (1) above shows an example of selecting the stator's dq-axis flux and the rotor's dq-axis flux as state quantities, but other physical quantities can also be selected as state quantities. For example, Equation (1) above can be modified to use current instead of stator flux for calculation. Alternatively, extended induced voltage can be used instead of rotor flux for calculation. In addition, other coordinate systems can be used instead of the dq coordinate system. For example, the stationary αβ coordinate system can be used instead of the dq coordinate system.

[0056] Since the above equation (1) contains the estimated angular velocity ω ＾ r Therefore, in estimating the angular velocity ω ＾ r With actual angular velocity ω r In cases of inconsistency, errors arise in the current estimation. Here, the model deviation ε is defined as follows (4). The adaptive observation unit 1 uses the angular velocity estimation unit 12 to adjust the estimated angular velocity ω.＾ r The value of is used to make the model bias ε zero.

[0057] [Mathematical Expression 4]

[0058]

[0059] As specific structural examples of the angular velocity estimation unit 12, known examples include those using a proportional-integral (PI) controller and those connecting the PI controller and integrator in series. Additionally, as described in Patent Document 1, other known examples include those using a proportional-integral (PI) controller. Figure 1 An example of a second angular velocity estimation unit being provided in parallel with the angular velocity estimation unit 12. On the other hand, the drive device 101 disclosed herein is a device for solving the problem of increased working hours in reliability evaluation, which is a technical issue of Patent Document 1, and does not adopt the method disclosed in Patent Document 1, namely the structure of providing angular velocity estimation units in parallel.

[0060] The adaptive observation unit 1, based on the above equation (3), uses the estimated magnetic flux vector, estimated current vector, and estimated angular velocity ω. ＾ r The primary angular frequency is calculated. Furthermore, the adaptive observation unit 1 estimates the magnetic pole position, which serves as the rotor position, by integrating the primary angular frequency.

[0061] The method described above for estimating the angular velocity and rotor position of AC motor 2 is commonly referred to as an "adaptive observer". Especially when the state variable in equation (1) is magnetic flux, the adaptive observer is called an "adaptive flux observer". The adaptive flux observer excels in its robustness to variations in the number of linkage fluxes and in producing steady-state speed estimation errors. Therefore, the adaptive flux observer is considered by those skilled in the art to be a high-performance speed estimation method.

[0062] In the adaptive observation unit 1, the observer gain h is adjusted. 11 ~h 42 The control gain of the angular velocity estimation unit 12 allows the response speed of the flux estimation and velocity estimation to be specified to arbitrary values. This is one of the most important properties in observer theory. Details will be explained later, and this property is used for vibration suppression control in the AC motor drive device of this disclosure.

[0063] Next, the operation of the speed control unit 3 and the torque control unit 4 will be explained. The speed control unit 3 is based on the angular velocity command ω. * r and estimated angular velocity ω ＾ r Calculate the first torque command τ *1. Specifically, the speed control unit 3 determines to set the angular velocity command ω. * r With the estimated angular velocity ω ＾ r The first torque command τ is consistent with the average value. * 1.

[0064] For the first torque command τ * The operation of 1 allows for speed control using a standard Proportional-Integral-Differential (PID) controller. However, other types of controllers besides PID can be used as long as the desired control performance can be obtained. Furthermore, PID controllers can be used in conjunction with other types of controllers. For example, a feedforward controller and a PID controller can be connected in parallel to form a two-degree-of-freedom control system.

[0065] The torque control unit 4 includes a dq-axis current control unit (not shown), a coordinate transformation unit (not shown), and... Figure 1 Voltage application unit not shown. Torque control unit 4 is based on the first torque command τ. * 1 and second torque command τ * 2. Determine the dq-axis current command. Specifically, implement control to make the output torque of AC motor 2 match the first torque command τ. * 1 and second torque command τ * The sum of 2 is consistent. Furthermore, the second torque command τ * The method for determining 2 will be explained later. The dq-axis current control unit (not shown) adjusts the voltage vector to match the dq-axis current command with the dq-axis current, and the voltage vector is applied to the AC motor 2 using a voltage application unit (not shown). Since the voltage and current of the AC motor 2 are AC signals, they are appropriately converted into DC signals on the dq coordinates by a coordinate transformation unit (not shown) for control.

[0066] To control the torque of AC motor 2 to the desired value, it is known that controlling the dq-axis current is preferred. However, it goes without saying that control can also be performed using coordinate systems other than the dq coordinate system. In the case of a typical embedded magnet type synchronous motor, the motor torque τ m It is determined by the following formula (5).

[0067] [Mathematical Expression 5]

[0068] τ m =P m Φ a i q +P m (L d -L q )id i q ……(5)

[0069] In equation (5) above, P m φ represents the number of pole pairs of AC motor 2. a This represents the magnetic flux linkage number along the dq axis.

[0070] To output the desired torque, the dq-axis current only needs to be determined based on the above equation (5). Here, the second term on the right-hand side of the above equation (5) represents the reluctance torque. Therefore, when the reluctance torque can be ignored, the motor torque τ m With q-axis current i q This is a proportional relationship. Therefore, if the q-axis current i is increased in response to the torque command... q Increasing or decreasing the torque will allow you to output the desired torque.

[0071] For controlling the dq-axis current, a conventional PI controller can be used. However, to eliminate interference between the dq axes, a decoupling controller is usually used in conjunction. Therefore, it is a preferred implementation to use a controller that includes a decoupling controller to determine the voltage vector that makes the dq-axis current command consistent with the dq-axis current.

[0072] The voltage application unit applies voltage to the AC motor 2 and drives it based on the result of dq-axis current control. The voltage application unit can be conceived as a typical 2-level inverter, but is not limited to this. Any circuit structure is possible as long as it is a device capable of applying the desired voltage. The voltage application unit can be, for example, a multilevel inverter or a matrix converter.

[0073] Next, the vibration suppression control unit 5 and the phase lead calculation unit 6, which are the main parts of the drive device 101 of this disclosure, will be described. In addition, before describing the vibration suppression control of this disclosure, conventional vibration suppression control techniques will be described.

[0074] As mentioned earlier, the AC motor 2 is a power source for certain mechanical devices. Many of these devices have periodic load torque pulsations that are synchronized with the rotation of the AC motor 2, and these load torque pulsations are applied to the AC motor 2. Due to these load torque pulsations, speed pulsations are generated in the AC motor 2. Therefore, vibration or noise problems may sometimes occur.

[0075] In addition, torque fluctuations generated in AC motor 2 can sometimes cause vibration or noise. From the controller's perspective, these torque fluctuations are disturbances similar to the load torque pulsations applied to AC motor 2. Therefore, speed pulsations can sometimes occur due to torque fluctuations generated in AC motor 2. Various causes can be considered regarding the periodic torque fluctuations generated in AC motor 2, such as harmonic distortion of the magnet flux and gain imbalance of the current sensor.

[0076] If the control response and disturbance suppression response of the speed control unit 3 are sufficiently high relative to the frequency of the speed pulsation, the speed pulsation will not be too large. On the other hand, the control response and disturbance suppression response of the speed control unit 3 usually have an upper limit. In order to suppress high-frequency speed pulsation using only the speed control unit 3, the control gain needs to be set very large, but if the control gain is too large, the control system will become unstable.

[0077] Against this backdrop, various vibration suppression control methods have been studied for a long time in order to properly suppress high-frequency velocity pulsations. Well-known methods include those described in Patent Document 2 above, which utilize Fourier series expansion and integral control.

[0078] When a position sensor is available and the disturbance frequency is known, implementing vibration suppression control is not particularly difficult. However, implementing vibration suppression control without a position sensor is quite challenging. This is because, as mentioned earlier, the upper limit of the estimated velocity response in a typical sensorless control system is only about several hundred rad / s, making it difficult to accurately estimate high-frequency velocity fluctuations. The technology disclosed to solve this problem is the one described in Patent Document 1. In Patent Document 1, the angular velocity estimation unit is improved, successfully achieving accurate estimation of high-frequency velocity fluctuations. However, there are situations where the technology described in Patent Document 1 is difficult to apply.

[0079] Typically, when the internal computational processing of a drive unit is altered, a detailed investigation is required to determine if the drive unit experiences any adverse effects as a result of the change. However, in sensorless control of an AC motor, changes to the angular velocity estimation unit can significantly alter performance, thus the impact of such changes can be extremely wide-ranging. When the method of Patent Document 1 is added as an extension to existing functions, the more successful the marketing of the drive unit, the more rapidly the time spent on reliability evaluation accompanying the added extension functions increases. In such cases, implementing the technology of Patent Document 1 becomes extremely difficult.

[0080] In fact, according to the prior art described in Patent Document 1, although high-frequency speed pulsation under sensorless control can be suppressed, it cannot sufficiently suppress the increase in operation time for reliability evaluation associated with the additional extended functions. In order to minimize the operation time for reliability evaluation associated with the additional extended functions, it is important not to change the structure of the adaptive observation unit 1, which serves as the angular velocity estimation unit.

[0081] The drive device 101 disclosed herein was developed based on the above-described situation, and uses an adaptive observation unit 1 with a conventional structure to achieve high-frequency vibration suppression control. Specifically, in the drive device 101 of this disclosure, the vibration suppression control unit 5 determines a second torque command τ to suppress the speed pulsation of the AC motor 2 based on two pieces of information: the model deviation of the adaptive observation unit 1 and the phase lead output by the phase lead calculation unit 6. * 2. Furthermore, the phase lead calculation unit 6 is based on the perturbation frequency f. d The phase lead is calculated. The phase lead calculation unit 6 is the focus of the drive device 101 of this disclosure, and will be described in more detail later.

[0082] In implementation method 1, the disturbance frequency f d It is considered known. The perturbation frequency f can be determined using any method. d For example, in a system that generates a disturbance at a predetermined frequency, the disturbance frequency f can be given in advance. d As a constant. Furthermore, in applications such as compressors that generate disturbances corresponding to the rotational frequency, the rotational frequency can be used as the disturbance frequency f. d The rotation frequency mentioned here can be obtained using a rotational position sensor or a speed sensor. Alternatively, in the case of sensorless control common to all embodiments herein, the estimated angular velocity ω can be used. ＾ r The rotational frequency can be determined using this method. Alternatively, torque meters, accelerometers, or vibration sensors can be used to detect or estimate the frequency of torque pulsations, which can then be used as the disturbance frequency f. d .

[0083] How to acquire pulsation information such as position, velocity, acceleration, and torque varies depending on the literature. In most cases of conventional vibration suppression control technologies represented by Patent Documents 1 and 2, the implicit premise is that this pulsation information can be acquired without phase delay. On the other hand, in this paper, we consider the case where the acquired data contains some phase delay. If the phase delay is known in advance, the effects of the phase delay can be avoided by advancing the phase of the acquired data by a corresponding amount. In this regard, the phase advance calculation unit 6 in this disclosure calculates how much the phase of the acquired data should be advanced.

[0084] Next, refer to the attached diagram. Figures 2 to 6 The structure and operation of the main parts of the drive device 101 in Embodiment 1 will be described. Figure 2 To show Figure 1 A block diagram showing the relationship between the model deviation ε of the drive device 101 and the speed estimation error Δω. Figure 3 To show Figure 1The drive device 101 shown is driven from the actual angular velocity ω r The transfer function G to the model bias ε ε An example of a Bode plot of (s). Figure 4 For the purpose of illustration Figure 1 A vector diagram illustrating the necessity of phase lead in the drive unit 101. Figure 5 This is the first figure used in the operation description of the drive device 101 in Embodiment 1. Figure 6 The second figure is used in the description of the operating wave of the drive device 101 in Embodiment 1.

[0085] First, we will explain what kind of phase delay is included in the pulsation data that can be acquired under sensorless control. As mentioned earlier, the adaptive observation unit 1 estimates the angular velocity ω of the AC motor 2 based on the voltage vector and current vector. r However, since there is an upper limit to the estimated response, the estimated angular velocity ω... ＾ r The high-frequency velocity fluctuation information is lost. On the other hand, to obtain the high-frequency velocity fluctuation information from the internal data of the adaptive observation unit 1, it is only necessary to focus on the model bias ε. However, the model bias ε is related to the true angular velocity ω of the AC motor 2. r In comparison, there is a phase delay. Regarding the model bias ε and the velocity estimation error Δω=ω r -ω ＾ r The relationship is well known in the aforementioned non-patent literature 1, etc.

[0086] Figure 2 The relationship between model deviation ε and velocity estimation error Δω is illustrated in the block diagram. Figure 2 In the middle, G ω (s) represents the transfer function of the angular velocity estimation unit 12. As mentioned above, known examples of the specific structure of the angular velocity estimation unit 12 include instances using a PI controller and instances where the PI controller is connected in series with an integrator. Furthermore, in Figure 2 In the middle, G iq (s) represents the value used based on the rotor flux φ of the d-axis. dr The product of the velocity estimation error Δω is φ dr ·Δω Estimated q-axis current estimation error Δi q The transfer function. Here, it is known that by appropriately determining the observer gain h... 11 ~h 42 Transfer function G iq (s) can be expressed as a first-order low-pass filter as shown in equation (6).

[0087] [Mathematical Expression 6]

[0088]

[0089] In the above equation (6), A x T represents the gain coefficient of the low-pass filter. x Let represent the time constant of the low-pass filter, and s represent the Laplace transform operator. Furthermore, by appropriately adjusting the observer gain h... 11 ~h 42 And able to change the time constant T x .

[0090] exist Figure 2 In the middle, from the true angular velocity ω r The transfer function G to the model bias ε ε (s) is represented by the following formula (7).

[0091] [Mathematical Expression 7]

[0092]

[0093] Here, the transfer function G iq (s), G ω (s) is known to the control designer. Therefore, the transfer function G ε The properties of (s) can be calculated. Figure 3 The diagram shows the representation of the aforementioned transfer function G. ε An example of a Bode plot of the characteristics of (s). Figure 3 The horizontal axis shows ω ac The estimated velocity response for the observer. The estimated velocity response is the ratio of the true angular velocity ω... r Estimated angular velocity ω when a step change occurs ＾ r The value is obtained by quantifying the following velocity. Specifically, it is the reciprocal of the time constant of the step response. This indicates that the larger the value of the estimated velocity response, i.e., the faster the estimated velocity response, the better the control performance of the sensorless control system. However, as mentioned earlier, the limit of the estimated velocity response is at most about several hundred [rad / s].

[0094] according to Figure 3 For example, the transfer function G ε The gain of (s) to the estimated velocity response ω ac Up to this point, the rate increases by approximately 40 dB / decade (decibels per decade), slightly exceeding the estimated speed response ω. ac It reaches its peak near the point of [decade], and then decreases by approximately -20 [dB / decade]. On the other hand, regarding the transfer function G... ε The phase characteristics of (s) exhibit a phase change of approximately 270 degrees when comparing the minimum and maximum frequencies. This is especially evident in the estimated velocity response ω.ac Near the 1[decade] ([10 times the frequency]), the phase change is significant, and when the frequency changes to 10 times, the phase change is more than 135 degrees.

[0095] When we want to use model bias ε to replace the true angular velocity ω r When used as input for vibration suppression control, the aforementioned phase change becomes a critical issue affecting the success or failure of the control. This is because, typically when the motor torque τ... m With load torque τ L When the phase difference between the phases exceeds ±60 degrees, vibration suppression control will have the opposite effect.

[0096] It is easy to imagine that vibration suppression control would fail unless the phase of the vibration information is accurately captured. Therefore, in the drive device 101 of Embodiment 1, a phase lead calculation unit 6 is provided. The phase lead calculation unit 6 is based on the disturbance frequency f. d Calculate the transfer function G ε (s) is the phase lead amount. Specifically, the phase lead amount calculation unit 6 is configured to calculate the phase lead amount based on the disturbance frequency f. d Calculate the phase lead and correct the transfer function G. ε Phase changes in (s) are used for vibration suppression control.

[0097] Next, the reason for needing the phase lead calculation unit 6 will be explained. First, let the load torque τ... L It can be expressed using trigonometric functions as shown in equation (8).

[0098] [Mathematical Expression 8]

[0099] τ L =A L ·cos(2πf d t)+B L ·sin(2πf d t)……(8)

[0100] In equation (8) above, f d Let t represent the perturbation frequency and t represent time. A L B represents the amplitude of the cosine component of the load torque ripple. L This represents the amplitude of the sinusoidal component of the load torque pulsation.

[0101] Similarly, let the motor torque τ m It can also be expressed using the trigonometric functions shown in equation (9).

[0102] [Mathematical Expression 9]

[0103] τ m =A m ·cos(2πfd t)+B m ·sin(2πf d t)……(9)

[0104] In the above equation (9), A m B represents the amplitude of the cosine component of the motor torque pulsation. m This represents the amplitude of the sinusoidal component of the motor torque pulsation.

[0105] When the load torque τ is defined as described above L and motor torque τ m At that time, it was possible to Figure 4 As shown, the cosine and sine components are used to plot the curves on both axes.

[0106] In vibration suppression control, the load torque τ is reduced. L With motor torque τ m Consistent control is desired. Here, the load torque τ is considered. L The absolute value of the amplitude and the motor torque τ m The case where the absolute values ​​of the amplitudes are equal, that is, the case where the relationship of the following equation (10) holds.

[0107] [Mathematical Expression 10]

[0108]

[0109] The relationship in equation (10) holds true and the motor torque τ m With load torque τ L When the phase difference between them is 60 degrees, the load torque τ L With motor torque τ m The difference between these two torques is the torque difference τ. m -τ L The relationship is as follows Figure 4 The equilateral triangle shown. That is, the torque difference τ m -τ L The absolute value of the load torque τ L The absolute values ​​of their amplitudes are equal. Therefore, at the motor torque τ m With load torque τ L When the phase difference between them exceeds 60 degrees, the torque difference τ m -τ L The absolute value is greater than the load torque τ L The absolute value of the amplitude. According to Figure 4 The vector image is an equilateral triangle, which is self-evident.

[0110] Furthermore, even if vibration suppression control is implemented to reduce the motor torque τ m Pulsation cannot reduce the torque difference τm -τ L If the absolute value of τ decreases, then vibration suppression control is best not implemented in terms of power efficiency. Therefore, when the motor torque τ m With load torque τ L If the phase difference between the phases exceeds ±60 degrees, it can be said that the vibration suppression control has failed.

[0111] As can be seen from the above explanation, in vibration suppression control, the motor torque τ m With load torque τ L The phase difference between them is an extremely important factor.

[0112] The phase of the model bias ε depends on the perturbation frequency f. d And it changes significantly. Therefore, when using model bias ε instead of the true angular velocity ω r When used as input for vibration suppression control, without considering phase changes, the disturbance frequency f that enables successful vibration suppression control is... d The range is extremely narrow.

[0113] Therefore, the phase lead calculation unit 6 in embodiment 1 will calculate the perturbation frequency f d As input, calculate Figure 3 The transfer function G is shown in the Bode plot. ε Phase change of (s). In the transfer function G ε The phase of (s) is represented by ∠G ε (jω d In the case of ), ∠G ε (jω d ) signifies the phase in the delay direction. Furthermore, j represents the imaginary unit, and ω d This represents the angular frequency of the disturbance. Additionally, ω... d =2πf d .

[0114] From the model bias ε, the true angular velocity ω r It appears the phase has advanced. Therefore, in this paper, ∠G... ε (jω d This is called the "phase delay," which will make ∠G ε (jω d -∠G is obtained by reversing the sign. ε (jω d This is called the "phase lead".

[0115] If we know the phase lead - ∠G ε (jω dIf the phase of the model deviation ε is advanced by a corresponding amount, vibration suppression control can be performed over a wide disturbance frequency range. Furthermore, the structure of the vibration suppression control unit 5 can be arbitrary. Here, as an example, a method with improved technology described in Patent Document 2 will be explained.

[0116] Return to Figure 1 The vibration suppression control unit 5 includes a speed pulsation calculator 51, an integral controller 52 and 53, and an AC restorer 54.

[0117] The velocity fluctuation calculator 51, based on the principle of Fourier series expansion, DC-converts and extracts specific frequency components contained in the model deviation ε. The specific frequency refers to the aforementioned disturbance frequency f. d However, in this paper, a more extended definition is used. Typically, load torque ripples have several frequency components. In this paper, any frequency among them is referred to as a specific frequency component. The cosine coefficient E is output from the speed ripple arithmetic unit 51. c and the sine coefficient E s These coefficients represent specific frequency components after DC conversion.

[0118] At this point, the model bias ε and the disturbance frequency f are used based on the following equations (11) and (12). d and phase lead - ∠G ε (jω d To calculate the cosine coefficient E of the model bias ε. c and the sine coefficient E of the model bias ε s .

[0119] [Mathematical Expression 11]

[0120]

[0121] [Mathematical Expression 12]

[0122]

[0123] In equations (11) and (12) above, t represents time. Additionally, T... d This represents the period of the disturbance, T. d The perturbation frequency f d The reciprocal of T. d =1 / f d .

[0124] In equations (11) and (12) above, the transfer function G is considered by taking the phase change of the detector signal as the Fourier series expansion rather than the phase of the model bias ε. εThe phase change in (s). Of course, the phase of the model deviation ε can be directly manipulated before the Fourier series expansion to replace the phase change of the detected signal. However, the method of changing the phase of the detected signal is simpler to calculate than the method of directly manipulating the phase of the model deviation ε.

[0125] The cosine coefficient E of the controller 52 for the model deviation ε c Integrate and calculate the pulsating component τ of the motor torque as shown in equation (13). c Additionally, the sine coefficient E of the I controller 53 for the model deviation ε s Integrate and calculate the pulsating component τ of the motor torque as shown in equation (14). s .

[0126] [Mathematical Expression 13]

[0127]

[0128] [Mathematical Expression 14]

[0129]

[0130] Above, here K I Integral gain K represents the integral gain. I For consideration Figure 3 The value is determined by the gain value in the Bode plot, the equation of motion of the mechanical system, etc. Furthermore, although designing the control gain becomes more complex, a PI controller or a PID controller can be used instead of an I controller (52, 53).

[0131] Due to the cosine coefficient E c and the sine coefficient E s Since it is a direct current, the motor torque τ m pulsating component τ c τ s It is also a direct current. To suppress vibration, the motor torque τ needs to be reduced. m Pulsation. Therefore, it is necessary to control the motor torque τ. m pulsating component τ c τ s Resume communication.

[0132] AC restorer 54 based on motor torque τ m pulsating component τ c τ s and disturbance frequency f d The second torque command τ is calculated as follows (15). * 2.

[0133] [Mathematical Expression 15]

[0134]

[0135] The torque pulsation and the velocity pulsation are 90 degrees out of phase. Therefore, in the above equation (15), the sine wave and the cosine wave are shifted by 90 degrees and multiplied by the pulsation component τ respectively. c τ s .

[0136] In the AC recovery according to the above equation (15), the phase lead -∠G is not used. ε (jω d Therefore, a phase difference is generated between the trigonometric functions of the detected signal and the recovered signal. Using this phase difference component, consider the transfer function G. ε The phase change of (s) can determine even if the perturbation frequency f d The second torque command τ, which can appropriately suppress vibration even when changes occur, is also effective. * 2. Additionally, by using the second torque command τ * 2. Make the motor torque τ m The pulsation mechanism enables vibration suppression control across a wide frequency range, even without position sensor control.

[0137] Figure 5 An example of the operating waveform of the drive device 101 in Embodiment 1 is shown. Additionally, Figure 6 The diagram shows the drive device 101 of Embodiment 1 and its connection to... Figure 5 Examples of different working waveforms. Figure 5 and Figure 6 In the middle, the horizontal axis represents time, and the vertical axis of the upper waveform represents angular velocity ω. r The magnitude of the waveform is shown on the lower part, where the vertical axis represents the magnitude of the torque. Figure 5 and Figure 6 The lower part of each section, with solid lines representing the load torque τ L The dashed line represents the motor torque τ. m .in addition, Figure 5 and Figure 6 All horizontal axes in the upper and lower parts are recorded using the same scale.

[0138] Figure 5 This is an example of the working waveform during high-speed rotation. Figure 5 An example is the disturbance frequency f of load torque ripple. d With angular velocity ω r The situation is that the increase is proportional.

[0139] Figure 5 (a) is an example of the operating results without vibration suppression control. In this example, the goal is to drive AC motor 2 at a constant angular velocity, but the load torque τ LThere are non-sinusoidal periodic pulsations. The aforementioned speed control unit 3 has the function of controlling the angular velocity ω. r Follow the angular velocity command ω * r The effect of load torque τ L The frequency is much higher than the speed control response. Therefore, in the motor torque τ m Only slight pulsations were observed, and the load torque τ L With motor torque τ m The amplitude difference is very large. Therefore, in Figure 5 In (a), due to the influence of load torque pulsation, the angular velocity ω r It produced a large pulsation.

[0140] Figure 5 (b) is shown for comparison, with the results of vibration suppression control performed without phase lead calculation. Additionally, Figure 5 (c) shows the working results of vibration suppression control in Implementation 1, which uses phase lead calculation. Figure 5 (b) and Figure 5 Of the two (c), vibration suppression control was successful, and the angular velocity ω r The pulsation is less than Figure 5 (a). The motor torque τ at this time. m It is roughly consistent with the fundamental component of the load torque pulsation.

[0141] In addition, Figure 5 In the example, vibration suppression control is implemented targeting the most dominant frequency component of the load torque waveform. Furthermore, the input to the vibration suppression control is set to the model bias ε, rather than the true angular velocity ω. r Or estimate the angular velocity ω ＾ r .

[0142] exist Figure 5 (b) and Figure 5 (c) Of these two, vibration suppression control is successful because the phase lead is extremely small at the disturbance frequency. Even without calculating the phase lead, vibration suppression is successful when the phase lead is extremely small, but this is only a phenomenon within a narrow frequency range.

[0143] on the other hand, Figure 6 This is an example of the waveform during low-speed rotation. For ease of explanation, Figure 6 The scale of the vertical axis of the upper waveform and Figure 5 There are some differences. Furthermore, the scale of the vertical axis of the lower waveform is different from... Figure 5 same.

[0144] Figure 6(a) is an example of the operating results without vibration suppression control. In this example, although the intention is to drive the AC motor 2 at a constant angular velocity, the load torque ripples cause vibration at angular velocity ω. r Large pulsations were observed. This phenomenon is related to... Figure 5 The situation in (a) is the same. Furthermore, with... Figure 5 The difference between (a) and (b) is that in Figure 6 In (a) the motor torque τ m This also produces pulsations, due to the low frequency of load torque pulsations. Figure 6 In (a), the load torque τ L Phase and motor torque τ m The phase inconsistency cannot properly suppress vibration.

[0145] Figure 6 (b) is shown for comparison, demonstrating the results of vibration suppression control without phase lead calculation. Additionally, Figure 6 (c) shows the working results of vibration suppression control in Implementation 1, which uses phase lead calculation. Figure 5 The difference between (b) and (b) is that... Figure 6 Vibration suppression control failed in (b). A point that should be specifically noted is the angular velocity ω. r The pulsation is greater than Figure 6 (a). In Figure 6 In (b), the motor torque τ m and load torque τ L Both the amplitude and phase deviated significantly, causing vibration suppression control to fail. On the other hand, after performing phase lead calculations... Figure 6 In (c), vibration suppression control is successful, enabling the angular velocity ω to be controlled. r The pulsation is less than Figure 6 (a)

[0146] Like this, according to Figure 5 and Figure 6 The results clearly show that when using the model deviation ε of the adaptive observation unit 1 to implement vibration suppression control, vibration suppression over a wide frequency range can be achieved by considering the phase lead.

[0147] Next, the differences between the method of Embodiment 1 and the prior art will be explained. First, both the method of Embodiment 1 and the method described in Patent Document 2 have components equivalent to the speed pulsation calculator 51, I controllers 52 and 53, and AC restorer 54 described above. On the other hand, the following two points are significant differences between the method of Embodiment 1 and the method described in Patent Document 2.

[0148] In the operation and processing according to the above formulas (11) and (12),

[0149] • Use model bias ε instead of true angular velocity ω r .

[0150] • The phase change and phase lead of the sine and cosine waves of the detector signal, as a Fourier series expansion, are related to the phase change and phase lead – ∠G. ε (jω d The corresponding amount.

[0151] The true angular velocity ω under sensorless control r It is unobservable. Furthermore, using the estimated angular velocity ω^ r Instead of the true angular velocity ω r At that time, high-frequency vibration suppression could not be performed due to the upper limit of the estimated velocity response. However, when using model bias ε instead of the true angular velocity ω... r When used as the input for vibration suppression control, unless the aforementioned phase lead amount -∠G is applied... ε (jω d Correction is required; otherwise, it cannot cope with the disturbance frequency f. d The changes.

[0152] Next, the differences between the method of Embodiment 1 and the method described in Patent Document 1 will be explained. Comparing Embodiment 1 and Patent Document 1, the most obvious difference is the angular velocity estimation unit 12. In the method described in Patent Document 1, the angular velocity estimation unit 12 is significantly different from existing methods. In existing AC motor drive devices, when the angular velocity estimation unit 12 is significantly modified, it is necessary to conduct a detailed investigation into whether any adverse conditions occur in the drive device due to this modification. When adding the method of Patent Document 1 as an extension function to existing functions, the more successful the drive device becomes, the faster the reliability evaluation work time associated with the added function increases. Therefore, there are many situations where it is difficult to implement the technology of Patent Document 1. In response, Embodiment 1 proposes a method for vibration suppression control based on the model deviation ε generated by the internal processing of the adaptive observation unit 1 without modifying the angular velocity estimation unit 12. Using this method, it is easy to add high-performance vibration suppression control without position sensors as an extension function to drive devices with successful sales.

[0153] Furthermore, as a secondary effect, the method of Embodiment 1 reduces the computational load of vibration suppression control compared to the method described in Patent Document 1. The controller described in Patent Document 1 is a structure that performs vibration suppression control after high-precision acceleration estimation. Therefore, Patent Document 1 requires four PI controllers. In contrast, in Embodiment 1, only two PI controllers are used, simplifying the structure. Although the calculation of phase lead compensation is increased, the number of trigonometric function calculations is reduced, making it a more easily implemented technique.

[0154] Furthermore, it will be self-evident to those skilled in the art that the mathematical formulas or flowcharts described herein can be appropriately modified. For example, they can be transformed as follows: Figure 7 That's how it's structured. Figure 7 To show Figure 1 A block diagram of a first variation of the structure shown in Embodiment 1. In Figure 7 In Figure 1 In the structure shown, the vibration suppression control unit 5 is replaced with the vibration suppression control unit 5a. In the vibration suppression control unit 5a, the speed pulsation calculator 51 is replaced with the speed pulsation calculator 51a, the I controllers 52 and 53 are replaced with the I controllers 52a and 53a respectively, and the AC restorer 54 is replaced with the AC restorer 54a.

[0155] Figure 1 The velocity pulsation calculator 51 shown uses the above equations (11) and (12) to calculate the cosine coefficient E of the model deviation ε. c and the sine coefficient E s .on the other hand, Figure 7 The velocity pulsation calculator 51a shown uses the following equations (16) and (17) to calculate the cosine coefficient E of the model deviation ε. c and the sine coefficient E s '.

[0156] [Mathematical Expression 16]

[0157]

[0158] [Mathematical Expression 17]

[0159]

[0160] in addition, Figure 1 The I controllers 52 and 53 shown above use equations (13) and (14) respectively to calculate the motor torque τ. m pulsating component τ c τ s .on the other hand, Figure 7 The I controllers 52a and 53a shown use the following equations (18) and (19) respectively to calculate the motor torque τ. mThe pulsating component τ' c 、τ' s .

[0161] [Mathematical Expression 18]

[0162]

[0163] [Mathematical Expression 19]

[0164]

[0165] in addition, Figure 1 The AC restorer 54 shown uses the above equation (15) to calculate the second torque command τ. * 2. On the other hand, Figure 7 The AC restorer 54a shown uses the following equation (20) to calculate the second torque command τ. * 2.

[0166] [Mathematical Expression 20]

[0167]

[0168] As explained above Figure 1 The vibration suppression control unit 5 shown uses a velocity pulsation calculator 51 to perform calculations using the phase lead amount, while Figure 7 The vibration suppression control unit 5a of the first modified example shown uses an AC restorer 54a to perform the calculation of the phase lead.

[0169] in addition, Figure 8 To show Figure 1 A block diagram of a second variation of the structure shown in Embodiment 1. Figure 8 In Figure 1 In the structure shown, the vibration suppression control unit 5 is replaced with the vibration suppression control unit 5b. In the vibration suppression control unit 5b, the speed pulsation calculator 51 is replaced with the speed pulsation calculator 51b, the I controllers 52 and 53 are replaced with the I controllers 52b and 53b respectively, and the AC restorer 54 is replaced with the AC restorer 54b.

[0170] Figure 1 The velocity pulsation calculator 51 shown uses the above equations (11) and (12) to calculate the cosine coefficient E of the model deviation ε. c and the sine coefficient E s .on the other hand, Figure 8 The velocity pulsation calculator 51b shown uses the following equations (21) and (22) to calculate the cosine coefficient E of the model deviation ε. c "and the sine coefficient E s ″.

[0171] [Mathematical Expression 21]

[0172]

[0173] [Mathematical Expression 22]

[0174]

[0175] Furthermore, K in the above equations (21) and (22) is any real number greater than 0 and less than 1.

[0176] in addition, Figure 1 The I controllers 52 and 53 shown above use equations (13) and (14) respectively to calculate the motor torque τ. m pulsating component τ c τ s .on the other hand, Figure 8 The I controllers 52b and 53b shown use the following equations (23) and (24) respectively to calculate the motor torque τ. m pulsating component τ″ c 、τ″ s .

[0177] [Mathematical Expression 23]

[0178]

[0179] [Mathematical Expression 24]

[0180]

[0181] in addition, Figure 1 The AC restorer 54 shown uses the above equation (15) to calculate the second torque command τ. * 2. On the other hand, Figure 8 The AC restorer 54b shown uses the following equation (25) to calculate the second torque command τ. * 2.

[0182] [Mathematical Expression 25]

[0183]

[0184] As explained above Figure 1 The vibration suppression control unit 5 shown uses only the velocity pulsation calculator 51 to perform the calculation of the phase lead amount. Figure 7 The vibration suppression control unit 5a in the first modified example shown uses only the AC restorer 54a to perform the calculation of the phase lead. In contrast, Figure 8 The vibration suppression control unit 5b shown in the second modified example uses the velocity pulsation calculator 51b and the AC restorer 54b to perform the calculation of the phase lead amount separately.

[0185] Furthermore, detailed explanations are omitted; the block diagrams and mathematical formulas shown in the first and second variations are as follows: Figure 1 And the equivalent transformations of equations (11) to (15). Therefore, whether the calculations are performed using equations (16) to (20) or using equations (21) to (25), the performance of vibration suppression control is the same as that of the calculations using equations (11) to (15).

[0186] Figure 9 This is a hardware structure diagram of the driving device 101 according to Embodiment 1. Figure 1 , Figure 7 and Figure 8 The record is omitted in the text, and Figure 9 The diagram shows a voltage application unit 801 and a current detection unit 802. The voltage application unit 801 is a voltage application unit that applies voltage to an AC motor 2. An example of a voltage application unit is a power converter such as a three-phase PWM inverter. The voltage application unit 801 operates according to a voltage command vector, applying voltage to the AC motor 2. In this document, the result of vectorizing the applied voltage applied by the voltage application unit 801 is referred to as the voltage vector. In the case where the voltage application unit 801 is a power converter, the power converter involves switching. Therefore, there is always an instantaneous difference between the voltage command vector and the voltage vector, but they are almost equivalent in terms of average value.

[0187] The drive unit 101 includes a processor 901. A voltage vector is input to the drive unit 101. The voltage vector is used for speed estimation calculations, but a voltage command vector calculated internally by the processor 901 can also be used instead of the voltage vector for speed estimation calculations. Additionally, a current vector is generated by the current detection unit 802 and input to the drive unit 101. The current vector is vector information about the alternating current flowing through the alternating current motor 2. An example of the current vector is the detected value of the dq-axis current, obtained by converting the alternating current detected by the current detection unit 802 into a value on the dq-axis.

[0188] The drive device 101 includes a memory 902. The memory 902 includes a volatile storage device (not shown), represented by random access memory, and a non-volatile auxiliary storage device (not shown), represented by flash memory. Alternatively, the memory 902 may include a hardware auxiliary storage device instead of the volatile or non-volatile auxiliary storage device. The processor 901 executes a program input from the memory 902. Since the memory 902 includes both auxiliary and volatile storage devices, the program is input to the processor 901 from the auxiliary storage device via the volatile storage device. Furthermore, the processor 901 can either output the result of the calculation to the volatile storage device of the memory 902, or save the data to the auxiliary storage device via the volatile storage device.

[0189] Various configurations have been explored for the voltage application unit 801 and the current detection unit 802, and essentially any configuration can be used. The voltage application unit 801 and the current detection unit 802 can be housed within the drive device 101. Furthermore, the drive device 101 can include a voltage detection unit that detects the voltage vector output by the voltage application unit 801. In this case, the voltage application unit 801 can be configured to send a command value of the voltage vector to the processor 901, and a value related to the voltage detected by the voltage detection unit can be sent to the processor 901. Similarly, the current detection unit 802 can be configured to send the detected value to the processor 901.

[0190] The processor 901 calculates the model deviation ε based on the current vector and voltage vector of the AC motor 2 using the aforementioned model deviation calculation unit 11. Based on the model deviation ε, the processor 901 performs calculations in the aforementioned phase lead calculation unit 6 and vibration suppression control unit 5 to determine the voltage command vector. By performing control calculations in this way, speed pulsations caused by periodic disturbances can be appropriately suppressed over a wide frequency range.

[0191] As explained above, in the AC motor drive device according to Embodiment 1, the adaptive observation unit adaptively estimates the angular velocity of the rotor of the AC motor driving the mechanical device with periodic load torque pulsations. The speed control unit determines a first torque command that makes the angular velocity command consistent with the average value of the estimated angular velocity. The phase lead calculation unit calculates the phase lead amount from the actual angular velocity to the transfer function of the model deviation, which is an internal quantity of the adaptive observation unit, based on the disturbance frequency. The vibration suppression control unit determines a second torque command that suppresses the speed pulsations of the AC motor based on the frequency of the load torque pulsations, the model deviation, and the phase lead amount. The torque control unit controls the torque of the AC motor based on the first torque command and the second torque command. Accordingly, vibration suppression control can be implemented on a general drive device without modifying the adaptive observer, which is the core part of the control. Therefore, it is possible to easily implement extended functions that add to existing functions. In addition, the speed pulsations of the AC motor can be appropriately suppressed regardless of the frequency.

[0192] Furthermore, the drive device for the AC motor in Embodiment 1 can use the conventional speed estimation method as is. Therefore, the "rapid increase in working time due to changes in the speed estimation method" that is a technical problem in the method of Patent Document 1 can be prevented.

[0193] The vibration suppression control unit in Embodiment 1 can be configured into three parts: a velocity pulsation calculator, an integral controller, and an AC recovery unit. The velocity pulsation calculator extracts specific frequency components contained in the model deviation into cosine and sine components. Two integral controllers perform integral control to make both the cosine and sine components zero. The AC recovery unit restores the output of the integral controllers to an AC signal. When performing these processes, it is sufficient that at least one of the velocity pulsation calculator and the AC recovery unit is configured to perform calculations that take into account phase lead.

[0194] Implementation method 2.

[0195] Figure 10 This is a block diagram illustrating the structure of the AC motor drive device 101a according to Embodiment 2. Figure 10 In the context of the drive device 101a in Embodiment 2, Figure 1 In the structure of the drive device 101 of Embodiment 1 shown, the torque control unit 4 is replaced by the torque control unit 4a. Furthermore, the vibration suppression control unit 5 is replaced by the first vibration suppression control unit 5d, and the phase lead calculation unit 6 is replaced by the first phase lead calculation unit 6d. Further, the drive device 101a of Embodiment 2 has second vibration suppression control units 5e to Nth vibration suppression control units 5f, and second phase lead calculation units 6e to Nth phase lead calculation units 6f. Here, N is an integer greater than or equal to 2. That is, Embodiment 1 has a structure with one vibration suppression control unit and one phase lead calculation unit, while Embodiment 2 has a structure with multiple vibration suppression control units and multiple phase lead calculation units. Furthermore, regarding other structures, [the following is a continuation of the previous paragraph, but the translation is incomplete]. Figure 1 For identical or equivalent components, the same reference numerals shall be added, and repeated descriptions shall be omitted.

[0196] Typically, the angular velocity pulsation characteristics of an AC motor vary depending on the application or the connected load. Therefore, considering a rotary compressor as an example, we can examine cases where the connected load exhibits periodic torque variations.

[0197] Figure 11 This is a diagram illustrating an example of the waveform of the load torque of a rotary compressor, which is an example of the load device in Embodiment 2. The horizontal axis represents the rotation angle, and the vertical axis represents the load torque. Here, the number of compression chambers of the rotary compressor is set to k. The rotation angle from 0 to 360 degrees is one cycle of the mechanical angle, i.e., the mechanical angle cycle.

[0198] First, in the case of only one compression chamber, i.e., k=1, such as Figure 11As shown by the solid line, the load torque oscillates significantly during the mechanical angular period. Although the load torque waveform also contains second and third harmonics, the first harmonic oscillation is the largest. Therefore, when applying the structure of Embodiment 1, if the disturbance frequency f... d If set as the first frequency of the mechanical angular frequency, then the largest angular velocity pulsation can be suppressed.

[0199] In embodiment 2, multiple vibration suppression control units are arranged in parallel. Therefore, it is also possible to simultaneously suppress speed pulsations caused by secondary and tertiary torque variations included in the load torque characteristics. Figure 10 In the example, the first disturbance frequency f will be input to the first vibration suppression control unit 5d and the first phase lead calculation unit 6d. d1 The frequency is set to the first frequency of the mechanical angular frequency. Then, the second disturbance frequency f will be input to the second vibration suppression control unit 5e and the second phase lead calculation unit 6e. d2 Let the frequency be the second frequency of the mechanical angular frequency. Then, the Nth disturbance frequency f, which will be input to the Nth vibration suppression control unit 5f and the Nth phase lead calculation unit 6f, will be... dN The frequency is set to the third of the mechanical angular frequency. The calculations performed by each vibration suppression control unit and the phase lead calculation unit are the same as those described in Embodiment 1. The torque control unit 4a is based on the first torque command τ output from the speed control unit 3. * 1 and the second to N+1 torque commands (τ) output from the first vibration suppression control unit 5d to the Nth vibration suppression control unit 5f respectively. * 2. τ * 3、……、τ * N+1 ) to drive AC motor 2. Specifically, control is implemented to give the first to N+1 torque commands (τ) * 1. τ * 2. τ * 3、……、τ * N+1 The sum of these values ​​is consistent with the output torque of AC motor 2. This allows for the complete suppression of primary, secondary, and tertiary speed pulsations.

[0200] The same approach can be taken into account for cases with 2 or 3 compression chambers, i.e., k=2 or k=3. Increasing the number of compression chambers leads to greater structural complexity and higher cost, but... Figure 11 The waveform shown has small pulsations. Specifically, when k=2, the second harmonic component of the mechanical angular frequency becomes larger, and when k=3, the third harmonic component becomes larger.

[0201] In the case of, for example, k=2, such as Figure 11As shown, the secondary vibration of the mechanical angular period is dominant. Therefore, the disturbance frequency f input to the first vibration suppression control unit 5d and the first phase lead calculation unit 6d is... d1 The second frequency component is set as the mechanical angular frequency. Then, if there are vibrations with frequency components exceeding the second frequency that you want to suppress, simply treat them as the second disturbance frequency f. d2 The values ​​can be input to the second vibration suppression control unit 5e and the second phase lead calculation unit 6e, respectively.

[0202] Additionally, in cases such as k=3, as Figure 11 As shown, the three vibrations of the mechanical angular period are dominant. Therefore, the disturbance frequency f input to the first vibration suppression control unit 5d and the first phase lead calculation unit 6d is... d1 The third frequency component is set as the mechanical angular frequency. Then, if there are vibrations with frequency components higher than the third that you want to suppress, simply treat them as the second disturbance frequency f. d2 The values ​​can be input to the second vibration suppression control unit 5e and the second phase lead calculation unit 6e, respectively.

[0203] As explained above, the AC motor drive device according to Embodiment 2 includes multiple vibration suppression control units and phase lead calculation units, with the phase lead calculation unit being provided in proportion to the number of vibration suppression control units. The multiple vibration suppression control units are configured to calculate torque commands based on different specific high-frequency components, and the torque control unit is configured to operate based on the torque commands output from each of the multiple vibration suppression control units. Therefore, multiple vibration frequencies can be suppressed simultaneously. As a result, angular velocity pulsations included in the angular velocity of the AC motor can be suppressed even less.

[0204] Implementation method 3.

[0205] Figure 12 This is a block diagram illustrating the structure of the AC motor drive device 101b according to Embodiment 3. Figure 12 In the context of the drive device 101b in embodiment 3, Figure 1 In the structure of the drive device 101 of Embodiment 1 shown, the vibration suppression control unit 5 is replaced by a vibration suppression control unit 5g, and the phase lead calculation unit 6 is replaced by a phase lead calculation unit 6g. Furthermore, a gain calculation unit 7 is provided in the drive device 101b of Embodiment 3. In addition, in Figure 12 The present invention uses the structure of the drive device 101 of Embodiment 1, but is not limited thereto. The drive device 101b can also be constructed using the drive device 101a described in Embodiment 2. In addition, the structure and function of the drive devices 101 and 101a are as described above, and the description here is omitted.

[0206] exist Figure 12 In the middle, the gain calculation unit 7 has an integral gain K that enables the vibration suppression control shown by the aforementioned equations (13), (14), etc. I With disturbance frequency f d The function of dynamically changing accordingly. The gain calculation unit 7 can specify any value for the time required from the start of vibration suppression control until vibration convergence.

[0207] As mentioned above Figure 3 As shown, from the true angular velocity ω r The transfer function G to the model bias ε ε The gain characteristic of (s) depends on the perturbation frequency f d And it changes significantly. Without considering this, the integral gain K of the vibration suppression control is determined... I In this case, the time constant from the start of vibration suppression control until vibration convergence depends on the disturbance frequency f. d Significant changes can sometimes make control adjustments cumbersome. However, by using the gain calculation unit 7, control adjustments can be made easier. Furthermore, the reciprocal of this time constant is usually referred to as the "control response".

[0208] Next, we will use an embedded magnet synchronous motor as an example to illustrate the gain calculation method. In an embedded magnet synchronous motor, neglecting the reluctance torque caused by the d-axis current, the gain is calculated from the q-axis current i q Angular velocity ω as electric angular velocity r The transfer function is represented by the following equation (26). The integral gain K that determines the vibration suppression control... I When doing so, we first need to consider equation (26).

[0209] [Mathematical Expression 26]

[0210]

[0211] In the above equation (26), P m Let J represent the number of pole pairs, J represent the moment of inertia, and φ represent the number of pole pairs. a This represents the magnetic flux linkage number along the dq axis.

[0212] Since the above equation (26) has an integral characteristic of 1 / s, even when suppressing velocity fluctuations of the same amplitude, the disturbance frequency f d Higher values ​​also require greater torque. Therefore, to maintain a constant control response in vibration suppression control, the integral gain K of vibration suppression control must be increased. I With the disturbance angular frequency ω d (=2πf d Proportional.

[0213] Furthermore, unless the true angular velocity ω is considered simultaneously...r The transfer function G to the model bias ε ε The gain characteristic of (s) is required; otherwise, the control response for vibration suppression control cannot be specified. For example... Figure 3 As shown, the frequency transfer function G ε (jω d The gain characteristic of the vibration suppression control exhibits attenuation in both the low-frequency and high-frequency regions. Therefore, the integral gain K of the vibration suppression control... I It is configured to rise in both low and high frequency regions. Furthermore, due to the frequency transfer function G... ε (jω d The absolute value of |G ε (jω d Since | is known, we only need to multiply its reciprocal by the integral gain K. I That's all.

[0214] When these factors are taken into account, it can be seen that in order to specify the control response of the vibration suppression control as an arbitrary value, it is only necessary to calculate the integral gain K of the vibration suppression control using the following equation (27). I That's all.

[0215] [Mathematical Expression 27]

[0216]

[0217] Gain calculation unit 7 calculates integral gain K using the above formula (27). I The vibration suppression control unit 5g provides the integral gain K from the gain calculation unit 7. I and the phase lead amount -∠G provided by the phase lead calculation unit 6g ε (jω d To implement vibration suppression control.

[0218] As explained above, the AC motor drive unit in Embodiment 3 includes a gain calculation unit that calculates the gain characteristics of the transfer function, and the integral controller performs control calculations considering the gain characteristics. Accordingly, the integral gain of the vibration suppression control can be varied as a function of the disturbance angular frequency, and the time constant until vibration convergence can be specified as a desired value. As a result, control adjustment operations become easier.

[0219] Implementation method 4.

[0220] Figure 13 This is a block diagram illustrating the structure of the AC motor drive device 101c according to Embodiment 4. Figure 13 middle, Figure 1The AC motor 2 shown is replaced with a refrigerant compressor 20 equipped with an AC motor 2. To reduce speed pulsation of the refrigerant compressor 20, the drive unit 101c of Embodiment 4 is configured using the drive unit 101 of Embodiment 1 as the compressor drive unit. Furthermore, in Figure 13 The drive device 101 described in Embodiment 1 is used, but it is not limited to this. The drive device 101c can also be constructed using the drive device 101a described in Embodiment 2 or the drive device 101b described in Embodiment 3. Furthermore, the structure and function of the drive devices 101, 101a, and 101b are as described above, and the description here is omitted.

[0221] Next, refer to Figure 14 and Figure 15 The structure of the refrigerant compressor 20 and the load torque of the refrigerant compressor 20 are described in detail. Figure 14 To show Figure 13 A cross-sectional view showing the schematic internal structure of the refrigerant compressor 20, which is the driving object. Additionally, Figure 15 To show Figure 14 This is a cross-sectional view of the internal structure of the compression section 202 of the refrigerant compressor 20 shown. Furthermore, a refrigerant compressor, also known as a rolling piston type, which is a rotary compressor, will be described here, but it is not limited to this. The refrigerant compressor may also be other types of compressors such as scroll compressors.

[0222] The refrigerant compressor 20 includes: a sealed container 211; an AC motor 2 built into the sealed container 211; a shaft 201, one end of which passes through the rotor 2-1 constituting the AC motor 2; a compression section 202, which is passed through by the other end of the shaft 201 and fixed inside the sealed container 211; a suction pipe 203 provided in the sealed container 211; and a discharge pipe 204 provided in the sealed container 211.

[0223] The stator 2-2 of the AC motor 2 is installed and held in a sealed container 211 by shrink fitting, cold fitting, or welding. Power is supplied to the coils 2-3 of the stator 2-2 via wires not shown. The rotor 2-1 is disposed inside the stator 2-2 with a gap 2-4, and is held in a free-rotating state by a bearing not shown via a shaft 201 at the center of the rotor 2-1.

[0224] In a refrigerant compressor 20 configured as described above, an AC motor 2 is driven to compress refrigerant drawn into the compression section 202 via the suction pipe 203. The compressed refrigerant is then discharged through the discharge pipe 204. In refrigerant compressors 20, the AC motor 2 is often immersed in the refrigerant. Due to drastic temperature changes, it is difficult to install a position sensor on the AC motor 2. Therefore, in the refrigerant compressor 20, the AC motor 2 must be driven without a position sensor.

[0225] In addition, such as Figure 15 As shown, the compression section 202 includes: an annular cylinder 212; a piston 205, which is integrally formed with the shaft 201 and rotatably disposed inside the cylinder 212; and a compression chamber 213, which is disposed on the inner periphery of the cylinder 212.

[0226] Cylinder 212 has the same Figure 13 The suction pipe 203 shown connects to the suction port 206 and the discharge port 207 for discharging compressed refrigerant. The suction port 206 and the discharge port 207 are connected to the compression chamber 213. Additionally, the cylinder 212 includes: blades 210 that divide the compression chamber 213 into a low-pressure chamber communicating with the suction pipe 203 and a high-pressure chamber communicating with the discharge port 207; and a spring 209 that applies force to the blades 210.

[0227] Shaft 201 connects AC motor 2 and piston 205. Piston 205 is eccentric, and the volumes on the discharge and suction sides change with the rotation angle. Refrigerant drawn in through suction port 206 is compressed by piston 205. When the pressure in compression chamber 213 increases, discharge valve 208 opens, and refrigerant is discharged from discharge port 207. Simultaneously, refrigerant flows into the suction side. When AC motor 2 rotates continuously, refrigerant is discharged once for every one mechanical angle rotation of piston 205.

[0228] The load torque pulsation of the refrigerant compressor 20 is a periodic disturbance for the AC motor 2, thus causing speed pulsation. In the refrigerant compressor 20, it is generally known that when the speed pulsation is large, the noise and vibration increase.

[0229] However, the frequencies of load torque pulsation and speed pulsation depend on the construction of the refrigerant compressor 20, and are therefore known. The refrigerant compressor 20 of Embodiment 3 is constructed using this fact. Figure 12 The control system shown uses the vibration suppression control unit 5 and the phase lead calculation unit 6 to calculate the second torque command τ that suppresses the specific frequency component of the speed pulsation in the refrigerant compressor 20. * 2.

[0230] The pulsation pattern of the load torque of the refrigerant compressor 20 depends on the mechanical structure of the refrigerant compressor. In some vibration suppression control methods, the pulsation pattern of the load torque is investigated in detail beforehand, and the investigated data is used to perform feedforward vibration suppression control. However, in feedforward vibration suppression control, the prior investigation and control adjustment are extremely cumbersome.

[0231] For general-purpose AC motor drive units, it is required to drive various types of mechanical devices, including refrigerant compressors. Feedforward-type vibration suppression control, which involves cumbersome pre-adjustment, is unsuitable for general-purpose drive units. Therefore, the drive unit 101c of Embodiment 4 is configured using feedback-type vibration suppression control so that vibration can be suppressed even without prior investigation. Furthermore, regarding general-purpose drive units, since their applications span numerous fields, it is difficult to drastically change the structure of the adaptive observation unit as described in Patent Document 1. This is because drastically changing the structure of the adaptive observation unit could potentially lead to undesirable conditions or defects, and for general-purpose drive units, the time required for investigating such effects would be enormous.

[0232] The drive device 101c of Embodiment 4 was developed in this context. The drive device 101c of Embodiment 4 is very useful because it can provide the function of general drive device vibration suppression control without changing the structure of the adaptive observation unit.

[0233] Implementation method 5.

[0234] Figure 16 This diagram illustrates the structure of the refrigeration cycle device 300 according to Embodiment 5. Figure 16 The refrigeration cycle unit 300 shown includes an AC motor drive unit 101c, a refrigerant compressor 20, a condenser 301 connected to the refrigerant compressor 20 via piping 305, a receiver 302 connected to the condenser 301 via piping 305, an expansion valve 303 connected to the receiver 302 via piping 305, and an evaporator 304 connected to the expansion valve 303 via piping 305. The evaporator 304 is connected to the suction pipe 203.

[0235] The refrigerant compressor 20, condenser 301, receiver 302, expansion valve 303, evaporator 304, and suction pipe 203 are connected by piping 305, thus forming a refrigeration cycle loop 306 for refrigerant circulation. In the refrigeration cycle loop 306, the processes of refrigerant evaporation, compression, condensation, and expansion are repeated, and heat is transferred while the refrigerant repeatedly changes from liquid to gas or from gas to liquid.

[0236] The functions of each device constituting the refrigeration cycle unit 300 will be explained. The evaporator 304 evaporates the refrigerant liquid under low pressure, carrying away heat from the surroundings and thus providing a cooling effect. The refrigerant compressor 20 compresses the refrigerant gas into a high-pressure gas to condense the refrigerant. The refrigerant compressor 20 is driven by the drive device 101c of Embodiment 4. The condenser 301 releases heat to condense the high-pressure refrigerant gas into a refrigerant liquid. The expansion valve 303 throttles and expands the refrigerant liquid into a low-pressure liquid to allow the refrigerant to evaporate. The receiver 302 is provided for adjusting the amount of refrigerant circulating; it can be omitted in smaller units.

[0237] Refrigeration cycle systems are typically required to improve noise reduction and reduce costs. For example, in household refrigeration cycle systems, where cost reduction is particularly critical, single-rotary compressors are often used. A single-rotary compressor refers to... Figure 14 and Figure 15 The rotary compressor described herein is a type of compressor with only one compression chamber 213. Rotary compressors have very large load torque pulsations, which often leads to increased vibration and noise. On the other hand, in conventional feedforward control methods, cumbersome control adjustments are required to suppress vibration and noise.

[0238] In embodiment 5, the refrigeration cycle device 300 employs feedback control so that the drive unit 101c automatically reduces speed pulsation to zero. This feedback-type vibration suppression control is highly useful for general-purpose drive units used in various applications. This is because it can suppress vibrations in various refrigerant compressors without prior adjustment.

[0239] As mentioned above, the technology described in Patent Document 1 is difficult to apply to general-purpose drive devices, while the method disclosed herein makes it extremely easy to add vibration suppression control functionality to general-purpose drive devices, which is very useful. Furthermore, according to Embodiment 5, by suppressing speed pulsations through feedback control, it is possible to flexibly respond to manufacturing variations, constant variations in the motor, and changes in the compressor's load conditions. Accordingly, a refrigeration cycle device 300 with high environmental tolerance can be realized.

[0240] Furthermore, the structure shown in the above embodiments is an example that can be combined with other known technologies, can combine embodiments with each other, and can omit or change a part of the structure without departing from the main idea.

Claims

1. A drive device for an AC motor, characterized in that, have: The adaptive observation unit adaptively estimates the angular velocity of the rotor of an AC motor that drives a mechanical device with periodic load torque pulsations. The speed control unit determines a first torque command that makes the angular velocity command consistent with the average value of the estimated angular velocity; The phase lead calculation unit calculates the phase lead from the true angular velocity to the transfer function of the model deviation, which is an internal quantity of the adaptive observation unit, based on the perturbation frequency. The vibration suppression control unit determines a second torque command to suppress the speed pulsation of the AC motor based on the frequency of the load torque pulsation, the model deviation, and the phase lead; and The torque control unit controls the torque of the AC motor based on the first torque command and the second torque command.

2. The AC motor drive device according to claim 1, characterized in that, The vibration suppression control unit includes: The velocity pulsation calculator separates the specific frequency components contained in the model deviation into cosine and sine components. Two integral controllers are used to perform integral control so that the cosine component and the sine component are both zero. as well as The AC recovery unit restores the output of the integral controller to an AC signal. In this process, at least one of the speed pulsation calculator and the AC recovery unit performs a calculation that takes into account the phase lead amount.

3. The AC motor drive device according to claim 2, characterized in that, It includes a gain calculation unit that calculates the gain characteristics of the transfer function. The integral controller takes the gain characteristics into account when performing control operations.

4. The drive device for an AC motor according to any one of claims 1 to 3, characterized in that, There are multiple vibration suppression control units and phase lead calculation units. The phase lead calculation unit is provided corresponding to the number of vibration suppression control units. The multiple vibration suppression control units calculate the second torque command based on different specific high-frequency components. The torque control unit controls the torque of the AC motor based on the first torque command and a plurality of second torque commands output from the plurality of vibration suppression control units.

5. A compressor drive device, wherein the drive device of the AC motor according to any one of claims 1 to 4 is used.

6. A refrigeration cycle apparatus, wherein the compressor drive device of claim 5 is used.

Images