Motor control device and air conditioner

Abstract

A motor control device and an air conditioner. The motor control device includes: a power supply unit; a current detection unit; a speed and electric angle estimation unit; a coordinate conversion unit that obtains an excitation current and a torque current based on a current and an electric angle; a torque current command determination unit that substitutes a predicted torque calculated based on a motion equation of a mechanical system into a torque component current command value calculated based on a torque formula of a vector control coordinate, generates a torque component current command value that makes a difference between an input speed command and an estimated speed close to zero; and a model predictive control unit that applies an evaluation function that evaluates a size of a difference between a torque component current command value and an excitation component current command value input from the outside and a predicted current corresponding to each of the evaluation function, selects a switching mode, and outputs.

Description

Technical Field

[0001] Embodiments of the present invention relate to a motor control device that controls a motor in a sensorless manner, and an air conditioner that uses the motor control device to control a motor to drive a compressor. Background Technology

[0002] For example, when a brushless DC motor is used as the motor of a compressor, the motor's rotational speed and position are sometimes estimated without a position sensor. If there is a difference between the estimated rotational speed and the target rotational speed, the motor's rotational speed is sometimes adjusted to the target rotational speed by changing the current or voltage command.

[0003] Figure 15 This describes the cross-sectional structure of a single-cylinder rotary compressor, in which, for example... Figure 16 As shown, its mechanism generates load variations based on the rotational mechanical angle of the compressor motor. Due to these load variations, uneven rotational speed occurs during one revolution of the motor at its mechanical angle, leading to vibration and noise. If the compressor continues to operate under these uneven motor speed conditions, stress is applied to piping that transports refrigerant, as in an air conditioner, reducing their lifespan. Therefore, torque control is necessary when controlling the motor's rotation to suppress the uneven rotational speed that accompanies load variations.

[0004] As a prior art for torque control for this purpose, Patent Document 1 (Japanese Patent No. 5175887) discloses the following configuration: The load torque generated by the compressor is estimated based on the torque current, the motor constant, and the inertial torque of the motor including the compressor unit. The phase of the periodic variation shown by the load torque is calculated, and a sinusoidal torque correction current is determined based on the load torque phase. The amplitude and phase of the torque correction current are adjusted to reduce the speed variation of the motor. Furthermore, Patent Document 2 (Japanese Unexamined Patent Application Publication No. 2018-93572) discloses a technique for compensating for the torque command value using the load torque estimated by an interference observer and the determined parameters related to the motor drive.

[0005] However, in the technologies of Patent Documents 1 and 2, since proportional-integral control is used in speed control and current control, the current command and voltage command are gradually determined. That is, due to the delay caused by using past information in feedback control, the control responsiveness is low. Furthermore, when the motor position is estimated using a sensorless method, if the estimation accuracy of the motor position and speed decreases, accurate torque control cannot be performed. Therefore, in the event of torque fluctuations within a very short period, speed fluctuations may not be adequately suppressed. Summary of the Invention

[0006] Therefore, a motor control device and an air conditioner equipped with the motor control device are provided, which can quickly follow the target speed when a sensorless motor control method is applied in a system that generates relatively large load variations.

[0007] The motor control device of the embodiment includes: a power supply unit that supplies AC power to a motor driving a load; a current detection unit that detects the current flowing in the windings of the motor; a speed and electrical angle estimation unit that estimates the rotational speed and electrical angle of the motor based on the voltage output by the current supply unit and the current; a coordinate transformation unit that obtains an excitation current and a torque current based on the current and the electrical angle; a torque current command determination unit that generates a torque component current command value such that the difference between the input speed command and the estimated speed is close to zero by substituting a predicted torque calculated based on the motion equation of the mechanical system into a torque component current command value calculated based on the torque formula of the vector control coordinate of the permanent magnet synchronous motor; and a model prediction control unit that selects a switching mode and outputs multiple predicted currents that can be output by the power supply unit and include current change rates determined according to multiple switching modes based on the space voltage vector, by applying an evaluation function that evaluates the difference between the torque component current command value, the externally input excitation component current command value, and their corresponding predicted currents.

[0008] Furthermore, the air conditioner of the embodiment includes a compressor, a motor that drives the compressor, and a motor control device of the embodiment. Attached Figure Description

[0009] Figure 1 This is a functional block diagram showing the configuration of the motor control device in the first embodiment.

[0010] Figure 2 This is a diagram showing the structure of an air conditioner that uses a motor control device.

[0011] Figure 3 It is a diagram that visually represents the evaluation of each switching mode using the evaluation function.

[0012] Figure 4 It is a diagram representing the space voltage vector.

[0013] Figure 5 It is a diagram showing the relationship between each voltage vector and the phase voltage.

[0014] Figure 6 This represents the current I calculated with respect to the d-axis. d (k) and predicted current I d A graph showing the relationship between (k+1).

[0015] Figure 7 It is a graph showing the actual load torque waveform of the compressor and the estimated sinusoidal load torque waveform.

[0016] Figure 8 This is a diagram showing the waveforms of each signal in conventional torque control and torque control in this embodiment.

[0017] Figure 9 This diagram illustrates an example of the selection of the main vector and sub-vector for representing the space voltage vector in the second embodiment.

[0018] Figure 10 This represents the current I calculated with respect to the d-axis. d (k) and predicted current I d A graph showing the relationship between (k+1).

[0019] Figure 11 This is a diagram showing the waveforms of each signal in conventional torque control and torque control in this embodiment.

[0020] Figure 12 This is a flowchart (1) representing the control content.

[0021] Figure 13 This is a flowchart (2) representing the control content.

[0022] Figure 14 This is a flowchart (3) representing the control content.

[0023] Figure 15 This is a diagram showing the cross-sectional structure of a single-cylinder rotary compressor.

[0024] Figure 16 It is a graph that shows the load variation caused by the rotational mechanical angle of the compressor motor. Detailed Implementation

[0025] (First Embodiment)

[0026] Figure 1 This is a functional block diagram showing the configuration of the motor control device. Although not shown, the inverter circuit 1, which serves as the power supply unit, is constructed by using six IGBTs, which are semiconductor switching elements, and bridging them in a three-phase configuration. The output terminals of each phase of the inverter circuit 1 are connected to the terminals of each phase winding of the permanent magnet synchronous motor, i.e., the brushless DC motor 2, which is connected in a star configuration, for example.

[0027] Current detection units 3u, 3v, and 3w are, for example, current transformers installed on the output lines of inverter circuit 1, detecting the currents Iu, Iv, and Iw of each of phases U, V, and W. Alternatively, they can detect the current of any two phases and calculate the current of the remaining phase. The current detection signals from these current detection units 3u, 3v, and 3w are input to coordinate conversion unit 4, where they are converted into digital data by an A / D converter (not shown). Coordinate conversion unit 4 converts the three-phase currents Iu, Iv, and Iw into two-phase currents Iα and Iβ. Based on the rotational phase angle θ estimated by the position estimation unit 5 (which serves as a speed and electrical angle estimation unit), it converts the currents Iα and Iβ in the stationary coordinate system into the d-axis current I in the rotating coordinate system. d q-axis current I q .

[0028] The q-axis current command generation unit 6, which is the torque current command determination unit, uses the command speed ω ref The integral value ω_I of the motor's angular velocity estimated by the position estimation unit 5, the load torque, the d-axis current command, the motor constant, and the sampling period are calculated. Then, the q-axis current command I is generated. qref To make the motor speed ω follow the command speed ω ref This is then output to the model predictive control unit 7. Details about this will be described later. Additionally, the excitation current command I input to the model predictive control unit 7... dref It is usually set to zero, but a negative value is set when performing tasks such as weak magnetic field control.

[0029] Model predictive control unit 7 uses DC voltage V DC Motor speed ω, rotation phase angle θ, and d-axis current I converted by coordinate transformation unit 4. d q-axis current I q The motor constant and sampling period were used to pre-test the d-axis and q-axis currents I of the inverter circuit 1 when it could output eight switching modes. d I q .

[0030] The evaluation function, for example, is used to determine the predicted d-axis current I. d With excitation current command I dref q-axis current I q The q-axis current command I provided by the q-axis current command generation unit 6 qref The model predictive control unit 7 selects the optimal switching mode based on the evaluation results using the evaluation function and outputs it to the inverter circuit 1. The duty cycle command values ​​Du, Dv, and Dw of the three phases based on the selected switching mode are converted into switching signals by the switching signal generation unit 8 and provided to the gates of each IGBT constituting the inverter circuit 1.

[0031] As described above, the components other than the inverter circuit 1 are implemented by software processing executed by a microcomputer, forming a motor control device 10 that performs vector control and model predictive control. For example, although not specifically illustrated, this microcomputer includes input / output ports, a serial communication circuit, an A / D converter for analog signals such as input current detection signals, and a timer for PWM control.

[0032] Figure 2 This describes the configuration of an air conditioner that incorporates a motor control device. The compressor 22, constituting the air conditioner 21, is constructed by housing the compressor unit 23 and the motor 2 within the same sealed iron container 25, with the rotor shaft of the motor 2 connected to the compressor unit 23. Furthermore, the compressor 22, four-way valve 26, indoor heat exchanger 27, pressure reducing device 28, and outdoor heat exchanger 29 are connected in a closed-loop manner via pipes serving as refrigerant passages. The compressor 22 is, for example, a rotary single-cylinder compressor.

[0033] During heating, the four-way valve 26 is in the state shown by the solid line. The high-temperature refrigerant, compressed by the compressor section 23 of the compressor 22, is supplied through the four-way valve 26 to the indoor heat exchanger 27 for condensation. Then, it is depressurized by the pressure reducing device 28, becoming low-temperature refrigerant that flows to the outdoor heat exchanger 29, where it evaporates and returns to the compressor 22. On the other hand, during cooling, the four-way valve 26 switches to the state shown by the dashed line. Therefore, the high-temperature refrigerant, compressed by the compressor section 23 of the compressor 22, is supplied through the four-way valve 26 to the outdoor heat exchanger 29 for condensation. Then, it is depressurized by the pressure reducing device 28, becoming low-temperature refrigerant that flows to the indoor heat exchanger 27, where it evaporates and returns to the compressor 22. Furthermore, the system is configured to supply air to the indoor and outdoor heat exchangers 27 and 29 respectively via fans 30 and 31, effectively facilitating heat exchange between the heat exchangers 27 and 29 and the indoor and outdoor air.

[0034] Next, refer to Figures 3 to 8 The principle behind the operation of this embodiment will be explained. For example... Figure 3 As shown, the model predictive control unit 7 calculates the predicted current as the first step and selects the switching mode based on the evaluation function as the second step. In the first step, the current after one control cycle for each switching mode is predicted and tested using the current change rate. First, based on the voltage equation of the permanent magnet synchronous motor, the current change rate S along the d-axis and q-axis at zero vector can be calculated using the following formula. d0 S q0 .

[0035]

Mathematical Formula 1

[0036]

[0037]

[0038] Here, R: winding resistance [Ω], L d d-axis inductance [H], L q : q-axis inductance [H], I d d-axis current [A], I q q-axis current [A] Linkage flux [Wb], ω: estimated velocity [rad / sec].

[0039] Next, the d-axis and q-axis current change rate S when the effective vector is obtained by using equations (1) and (2) through the following formula. d1 S q1 .

[0040]

Mathematical Formula 2

[0041]

[0042]

[0043] Here, V d d-axis voltage [V], V q : q-axis voltage [V].

[0044] Figure 4 This indicates the trial switching mode. Figure 5 This represents the relationship between the switching mode and the three-phase voltage of the motor. Eight switching modes, V0 to V7, are tested sequentially to generate the corresponding three-phase voltages for each mode. Then, a coordinate transformation is performed on the three-phase voltages using a rotational phase angle θ. From this, the commanded d-axis voltage V is calculated using the following formula. d q-axis voltage V q This refers to the case of relative transformation. The d-axis voltage V d Used for location estimation calculations.

[0045]

Mathematical Expression 3

[0046]

[0047]

[0048] d-axis voltage V d q-axis voltage V q The switching pattern varies depending on the trial, therefore, S is obtained through equations (3) and (4). d1 S q1 The result also varies, with 8 possible outcomes. For example... Figure 6 As shown, when the tested switching mode is set to fully open during the sampling period, the current I used in this test... d(k) and equations (3) and (4) are used to calculate the next predicted current I after one control cycle using the following formula. d (k+1). In this case, the prediction period of the current is different from the sampling period T. s [sec] equal. Equal to S d1 S q1 Similarly, the next predicted current I d (k+1) also has a number of trial power-on modes, namely 8. Furthermore, since the trial results for V0 and V7 are equal, one is selected based on the evaluation function. This will be discussed later.

[0049] I d (k+1)=I d (k)+T s S d1 (7)

[0050] I q (k+1)=I q (k)+T s S q1 (8)

[0051] Next, the selection of the switching mode in step 2 will be explained. The next predicted current I under each switching mode, calculated according to equations (7) and (8), will be used. d (k+1), I q (k+1), solve for the evaluation function g in the following equation.

[0052] g = ag q +bg d (9)

[0053] g q =[{I qref -I q (k+1)}] 2 (10)

[0054] g d =[{I dref -I d (k+1)}] 2 (11)

[0055] Here, a and b are weighting coefficients, and g is a weighting coefficient. q The evaluation formula for the q-axis current, g d : Evaluation formula for d-axis current.

[0056] (10) g q and g of equation (11) dThe value obtained by multiplying the square of the difference between the command value and the next predicted current by weighting coefficients a and b and summing them is the evaluation function g in equation (9). Since the next predicted current is closest to the current command when the evaluation function g is minimized, the best one can be selected from eight types of switching modes. The evaluation function g in equation (9) is shown as an example here, but the evaluation function can also be changed according to the required driving conditions of the inverter circuit 1. For example, in order to reduce the switching losses of the inverter circuit 1, switching mode transition conditions can be added to the evaluation function. Specifically, the switching mode transition can be limited by making the total number of state changes of the three-phase semiconductor switching elements less than 1. When the case of transition from effective vectors V1 to V6 to zero vectors V0 and V7 is explained as an example, the transition is to V0 in the case of V1, V3, and V5, and to V7 in the case of V2, V4, and V6. As the final output voltage of the inverter circuit 1, the selected switching mode is output as a signal with a 100% duty cycle.

[0057] Furthermore, in the actual test, the switching signal determined by model predictive control is updated and output after one control cycle. Then, the current flowing based on this switching signal is detected after two control cycles. Therefore, in the actual test, model predictive control trials the energizing mode after one control cycle to predict the current after two control cycles, and selects the switching mode whose predicted current is closest to the current command.

[0058] Next, the q-axis current command generation unit 6 will be explained. Figure 1 In the middle, the q-axis current command I is input to the model prediction control unit 7. qref and d-axis current command I dref q-axis current command I qref The current is generated by the q-axis current command generation unit 6, but the d-axis current command I... dref In this embodiment, it is zero. In the q-axis current command generation unit 6, in order to reduce the vibration of the compressor 22, the speed command ω is generated. ref The q-axis current command I for torque that differs from the estimated speed ω by zero. qref Perform calculations. q-axis current change command I qref The calculation formula is obtained based on the calculation formulas for predicted speed and predicted torque in model predictive control. The predicted speed ω(k+1) is expressed by the following formula based on the motion equation of the mechanical system.

[0059]

Mathematical Expression 4

[0060]

[0061] Here, J: torque of inertia [kg·m] 2 D: Coefficient of viscous friction [Nm / (rad / sec)], p: Number of pole pairs, Te Output torque [Nm], T l : Load torque [Nm], ω: Angular velocity [rad / sec].

[0062] When targeting the next output torque T e When solving equation (12) with (k+1), the following equation is obtained.

[0063]

Mathematical Expression 5

[0064]

[0065] Furthermore, in the absence of position sensor control, the estimated speed ω(k) in equations (12) and (13) is not used; instead, the integral value ω_I of the estimated speed is used. This is set as the estimated speed (integral) ω_I. In model predictive control, the d-axis voltage V... d q-axis voltage V q The voltage varies significantly for each sampling period; therefore, the result is that the d-axis induced voltage E, which contains position error information, is... d This also varies significantly. Therefore, the estimated speed ω contains a large variation component for each control cycle. Thus, using the estimated speed (integral) ω_I, the q-axis current command I... qref It does not include the variation component for each sampling period. In equation (13), the predicted velocity ω(k+1) is made to approximate the velocity command ω. ref The method determines the next output torque T e (k+1), therefore, equation (13) can be replaced as follows.

[0066]

Mathematical Expression 6

[0067]

[0068] The next output torque shown in equation (14) is the sum of the torque at which the difference between the speed command and the estimated speed (integral) is zero, and the load torque. Next, the predicted torque as the next output torque is expressed by the torque formula of the vector control coordinates of the permanent magnet synchronous motor through the following equation.

[0069] T e (k+1)

[0070] =p{φI q (k+1)+(L d -Lq)I d (k+1)I q (k+1)} (15)

[0071] When targeting I q When solving equation (15) with (k+1), we get the following equation.

[0072]

Mathematical Expression 7

[0073]

[0074] In equation (16), each predicted current I is made to q (k+1), I d (k+1) respectively approach the current command I qref I dref Substituting into equation (14) and replacing equation (16) as shown in the following equation, we can calculate the q-axis current command I for reducing the vibration of compressor 22. qref .

[0075]

Mathematical Expression 8

[0076]

[0077] According to equation (17), the q-axis current command I is calculated based on the torque at which the difference between the speed command and the speed estimate (integral) is zero, and the load torque. qref For the former's torque, the control quantity can be adjusted by multiplying the aforementioned speed difference by the gain. For the latter's load torque T... l For example, using results inferred from prior art such as Patent Document 1.

[0078] like Figure 7 As shown, for the actual load torque of the compressor varying under the operating conditions of the air conditioner, if the basic wave information of the load torque, which is approximately sinusoidal and obtained by using existing technology to estimate the load torque, is used for model predictive control, a model error will occur, resulting in a speed error. In equation (14), the first term on the right-hand side, which is not present in the existing technology, is added for compensation. Therefore, the q-axis current command I is determined according to equation (17). qref This enables the estimated velocity ω to follow the velocity command ω. ref .

[0079] exist Figure 8 The diagram shows simulation results of driving a motor at 30 rpm, with a variable load, and a sampling frequency of 5 kHz equal to the control frequency. In conventional vector and torque control systems using PI control, the actual estimated speed and current gradually follow the command values ​​to determine the final inverter output voltage, resulting in low responsiveness. Furthermore, PWM signal generation uses a carrier comparison method, therefore, the control frequency is equal to the switching frequency.

[0080] On the other hand, the model predictive control of this embodiment, as described above, directly determines the command value using an electrical / mechanical model, thus achieving high response. Maintaining a control frequency of 5 kHz, the signal with a 100% duty cycle, selected as a prediction result, is switched at the required timing. For example, under the current drive conditions, the average switching frequency is reduced to approximately 1 kHz, indicating that switching losses in the power devices can be reduced, thereby improving efficiency.

[0081] As described above, according to this embodiment, the inverter circuit 1 supplies AC power to the motor 2 that drives the compressor 22 of the air conditioner 21, which serves as a load. The current detection unit 3 detects the currents Iu, Iv, and Iw flowing in the windings of the motor 2. The position estimation unit 5 estimates the rotational speed ω and electrical angle θ of the motor 2 based on the voltage output by the inverter circuit 1 and the aforementioned currents. The coordinate transformation unit 4 obtains the excitation current I based on the currents Iu to Iw and the electrical angle θ. d and torque current I q .

[0082] The q-axis current command generation unit 6 calculates the predicted torque T based on the motion equation of the mechanical system. e Substituting (k+1) into the torque component current command value calculated based on the vector control coordinates of motor 2, a speed command ω is generated that makes the input speed command ω... ref The torque component current command value I, whose difference from the estimated speed ω is close to zero. qref The model predictive control unit 7 evaluates the torque component current command value I by applying multiple predicted currents that the inverter circuit 1 can output, each containing the current change rate determined according to multiple switching modes based on the space voltage vector. qref And the excitation component current command value I input from the outside. dref The switching mode is selected and output using evaluation functions that measure the difference between the predicted current and the corresponding current.

[0083] Specifically, the model prediction control unit 7 selects a switching mode from eight switching modes obtained by adding two zero vectors to six space voltage vectors set for every 60 electrical degrees, and outputs the selected switching mode with a duty cycle of 100% relative to the switching cycle.

[0084] Therefore, in an air conditioner 21 that generates relatively large load variations, similar to compressor 22, when motor 2 is controlled in a sensorless manner, the speed ω of motor 2 can be made to quickly follow the target speed ω. ref Furthermore, by outputting the selected switching mode at 100% duty cycle relative to the switching cycle, switching losses can be reduced, thus improving the overall performance of the air conditioner 21.

[0085] Furthermore, the q-axis current command generation unit 6 generates the torque component current command value I. qref The parameters are obtained by using the integral value ω_I of the estimated speed obtained in the process of calculating the estimated value ω of the rotational speed through the position estimation unit 5. Therefore, it is possible to avoid errors in the command value I. qref It includes the variation of current in each sampling period.

[0086] Furthermore, the model predictive control unit 7 can use the torque component current command value I qref With the corresponding predicted current I q The square of the difference between (k+1) and the excitation component current command value I dref With the corresponding predicted current I d The function of summing the squares of the differences between (k+1) is used as the evaluation function g, thus enabling proper evaluation of each switching mode.

[0087] (Second Implementation)

[0088] Hereinafter, the same symbols will be used for parts identical to those in the first embodiment, and descriptions will be omitted; the different parts will be described. In the control of the first embodiment, a voltage vector as a switching mode is selected to reduce the number of switching operations, but the current ripple of the motor 2 increases. Therefore, in the second embodiment, control for reducing current ripple is added. Therefore, in the first and second steps of the model predictive control in the first embodiment, the calculation of the generation time of each of the two voltage vectors is added as a third step.

[0089] When the selected voltage vector is used as the principal vector as in the first embodiment, either of the voltage vectors adjacent to the principal vector is selected as the secondary vector. The generation time of each vector is adjusted, and the final synthesized voltage vector is output through space vector modulation. Thus, in the carrier comparison mode, the inverter circuit 1 can also output any voltage vector.

[0090] Figure 9 This represents the voltage vector output in the second embodiment. In the first embodiment, the current is predicted when only the main vector is set to fully open, i.e., output with a 100% duty cycle. For the switching mode that minimizes the evaluation function g, one of the eight voltage vectors, including the zero vector, is selected. In the second embodiment, the switching mode can be selected from six modes (V1 to V6) excluding the zero vector. In this case, the selected switching mode becomes the main vector, and therefore, the corresponding current change rate S is... d1 S q1 Let S be the value of S. d1_main S q1_main .

[0091] exist Figure 9 For example, when the principal vector with the minimum evaluation function g is V1, the secondary vector is V2 or V6. When predicting the current, the rate of change of current for all switching modes is calculated; therefore, the S corresponding to the secondary vector is... d1 S q1 Let S be the value of S. d1_sub S q1_sub .

[0092] A formula for predicting the current, taking into account the generation time, is defined. By solving the simultaneous equations, the generation times of the principal and secondary vectors can be calculated. By determining the generation times of the principal and secondary vectors, it is possible to decide whether to choose V2 or V6 as the secondary vector. Suppose that the generation time is calculated using the wrong secondary vector, resulting in a negative value. In this case, a different secondary vector is selected, and the generation time is calculated again.

[0093] Next, derive the formula for generating time. First, as... Figure 10 As shown, the formula for predicting the next current, taking into account the generation time of the principal vector, is obtained by the following equation. The next current is predicted based on the generation times of the effective vector and the zero vector, as well as the rate of change of the current.

[0094] I d (k+1)=I d (k)+(T s -T main )S d0 +T main S d1_main

[0095] =I d (k)+T s S d0 +Tmain(S d1_main -S d0 (18)

[0096] I q (k+1)=I q (k)+(T s -T main )S q0 +T main S q1_main

[0097] =I q (k)+T s S q0 +T main (S q1_main -S q0 (19)

[0098] In addition, T mainIt is the generation time of the principal vector [sec].

[0099] If the generation time of the secondary vector and the rate of change of current are added to equations (18) and (19), the following equations are obtained.

[0100] I d (k+1)=I d (k)+T s S d0 +T main (S d1_main -S d0 )+T sub (S d1_sub -S d0 (20)

[0101] I q (k+1)=I q (k)+T sq0 +T main (S q1_main -S q0 )+T sub (S q1_sub -S q0 ) (twenty one)

[0102] In addition, T sub It is the generation time of the sub-vector [sec].

[0103] Eliminate T from equations (20) and (21) sub , for T main Solving for the equation, we get the following formula.

[0104]

Mathematical Expression 9

[0105]

[0106] In equation (22), the next predicted current I along the d-axis and q-axis is given. d (k+1), I q (k+1) Approaching d-axis current command I dref and q-axis current command I qref The generation time of the principal vector is determined by the method, so equation (22) can be replaced as follows.

[0107]

Mathematical Formula 10

[0108]

[0109] By substituting the rate of change of current obtained in step 1 into equation (23), the generation time of the principal vector can be determined.

[0110] Similarly, the generation time of the secondary vector is determined. When for T...sub When solving equation (21), the following equation is obtained.

[0111]

Mathematical Expression 11

[0112]

[0113] In equation (24), the next predicted current I along the q-axis is... q (k+1) Approaching q-axis current command I qref The generation time of the sub-vector is determined by the method, so equation (24) can be replaced as follows.

[0114]

Mathematical Expression 12

[0115]

[0116] T obtained through equation (23) main Substituting into equation (25), the generation time of the sub-vector can be calculated. Based on the above, the generation times of the main vector and sub-vector can be determined, and the duty cycles of the main vector and sub-vector can be calculated by dividing by the control period. Then, as a duty cycle restriction, the following processing is performed. If the calculated duty cycles of the main vector and sub-vector are negative, they are set to 0%. The maximum sum of the duty cycles of the main vector and sub-vector is 100%. If the sum is greater than 100%, the duty cycles of the main vector and sub-vector are corrected by dividing the sum by the sum. With the above restriction processing implemented, the duty cycle is changed. Therefore, the changed duty cycle is multiplied by the control period, and T is recalculated. main and T sub If based on Figure 9 By allocating the duty cycles of the main vector and the sub-vector to the duty cycles of the three phases, a PWM waveform that is closest to the current command after one control cycle can be generated.

[0117] The following is a repeated model predictive control according to the control cycle. In other words, the PWM signal waveform of this time is the result of the previous model predictive control. When the next predicted current after one control cycle is obtained in the first step above, in the case of the second embodiment, unlike equations (7) and (8) of the first embodiment, the previous values ​​of the current change rate and vector generation time are required. Therefore, they are obtained by equations (20) and (21).

[0118] Furthermore, in the second embodiment, the d-axis voltage V used for position estimation calculation... d The final output value differs from that in the trial switch mode, therefore, it needs to be recalculated using the following formula. Here, represents the relative conversion case.

[0119]

Mathematical Expression 13

[0120]

[0121] Figure 11 The simulation results are shown in the figure. By determining the generation time of the main vector and the adjacent secondary vector determined by the evaluation function in a way that makes the predicted current close to the current command value, the inverter can output any voltage vector. Therefore, it is possible to reduce current ripple based on the selection of the optimal switching mode.

[0122] Figures 12 to 14 This is a flowchart based on the premise of selectively executing the control of the first and second implementation methods. First, the rate of change of the current S of the zero vector is calculated using equations (1) and (2). d0 S q0 During the calculation (S1), each switching mode is tested and evaluated in the repeated processing of steps S3 to S9. The output voltages Vu, Vv, and Vw based on the tested switching modes are calculated (S3), and the excitation voltage V is adjusted accordingly. d Torque voltage V q Perform the operation (S4).

[0123] Next, for one of the effective vectors V1 to V6, the rate of change of current S is expressed by equations (3) and (4). d1 S q1 Perform calculation (S5), and use equations (7) and (8) to predict the current I. d (k+1), I q (k+1) is used for calculation (S6). Additionally, when calculating the duty cycles Du, Dv, and Dw of each phase for control in the second embodiment, the current change rate S is pre-stored. d1 S q1 Then, the evaluation function g is calculated using equation (9) (S7), the results are compared, and the minimum value of the evaluation function g is selected (S8). If the result of this calculation is temporarily the minimum (yes), then the corresponding switching mode and V are updated. d and V q Predicted current I d (k+1), I q (k+1)(S9).

[0124] When all switching modes have been tested (S2; Yes), under the control of the first embodiment (S10; Yes), the duty cycle of the main vector is set to 100% for the selected switching mode (S11). Then, when the duty cycles Du, Dv, and Dw of each phase are set according to the selected main vector (S23), under the control of the first embodiment (S24; Yes), the process ends directly. Furthermore, the user can select which of the first or second embodiments of control to perform.

[0125] On the other hand, under the control of the second embodiment (S10; no), one of the two vectors adjacent to the selected principal vector is selected as the secondary vector (S12). Then, the generation time T of the principal vector is calculated using equations (23) and (25), respectively. main The generation time T of the sub-vector sub Perform calculations (S13, S14) to calculate the duty cycles of the main vector and the sub-vector (S15). If the duty cycle of the sub-vector is positive (S16; yes), the selection of the sub-vector is appropriate, and the duty cycle is restricted (S17). The current I is then predicted by recalculating using equations (20) and (21). d (k+1, I) q (k+1)(S18). Then, proceed to step S23.

[0126] If the duty cycle of the sub-vector is negative (S16; no), then the other of the two vectors adjacent to the main vector is selected as the sub-vector (S19). Then, the same process as steps S13 to S15 (S20 to S22) is performed, and the process proceeds to step S17.

[0127] In addition, Figure 14 In the process shown, under the control of the second embodiment (S24; no), the duty cycles Du, Dv, and Dw of each phase are set according to the selected sub-vector (S25), and the excitation voltage V is recalculated by equation (26). d (S26). Voltage V d Used for sensorless control. Additionally, the torque voltage V can also be calculated simultaneously. q But torque voltage V q It was not actually used.

[0128] As described above, according to the second embodiment, when the model prediction control unit 7 selects a switching mode as the main vector from six space voltage vectors set at 60 electrical degrees, it selects one of the two space voltage vectors adjacent to the main vector as the secondary vector, and adjusts the output time of the main vector and the secondary vector respectively during the switching cycle. Therefore, compared with the control in the first embodiment, current ripple can be further reduced.

[0129] (Other implementation methods)

[0130] The evaluation function g is not limited to equations (9) to (11); any appropriate function can be selected.

[0131] The motor's load is not limited to the compressor.

[0132] Several embodiments of the present invention have been described, but these embodiments are merely illustrative and are not intended to limit the scope of the invention. These new embodiments can be implemented in various other ways, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, and are also included in the invention described in the technical solution and its equivalents.

Claims

1. A motor control device, characterized in that, have: The power supply department supplies alternating current to the motors that drive the loads; The current detection unit detects the current flowing in the windings of the motor. The speed and electrical angle estimation unit estimates the rotational speed and electrical angle of the motor based on the voltage output by the power supply unit and the current. The coordinate transformation unit obtains the excitation current and torque current based on the aforementioned current and electrical angle. The torque current command determination unit generates a torque component current command value that makes the difference between the input speed command and the estimated speed close to zero by substituting the predicted torque calculated based on the motion equation of the mechanical system into the torque formula based on the vector control coordinates of the permanent magnet synchronous motor. as well as The model predictive control unit selects and outputs multiple predicted currents that can be output by the power supply unit, which include the rate of change of current determined according to multiple switching modes based on the space voltage vector. It selects and outputs the switching mode by applying an evaluation function that evaluates the difference between the torque component current command value, the externally input excitation component current command value, and their corresponding predicted currents.

2. The motor control device according to claim 1, characterized in that, The aforementioned model predictive control unit selects one of eight switching modes from six space voltage vectors plus two zero vectors set at 60 electrical degrees, and outputs the selected switching mode with a duty cycle of 100% relative to the switching cycle.

3. The motor control device according to claim 1, characterized in that, When the aforementioned model predictive control unit selects a switching mode as the primary vector from six space voltage vectors set at 60 electrical degrees, it selects one of the two space voltage vectors adjacent to the primary vector as the secondary vector. During the switching cycle, adjust the output time of the main vector and the sub-vector respectively.

4. The motor control device according to claim 1, characterized in that, The torque current command determination unit uses the integral value of the estimated speed obtained in the process of obtaining the estimated value of the rotational speed through the speed and electrical angle estimation unit as a parameter for generating the torque component current command value.

5. The motor control device according to claim 1, characterized in that, The evaluation function described above uses a function that adds the square of the difference between the torque component current command value and the corresponding predicted current and the square of the difference between the excitation component current command value and the corresponding predicted current.

6. An air conditioner, characterized in that, have: compressor; The motor drives the compressor; and The motor control device according to claim 1.

Images