A high-precision model predictive current control system and control method for a dual three-phase motor

Abstract

The application discloses a high-precision model prediction current control system and control method of a double three-phase motor, and relates to the technical field of multiphase motor control. The application extends the traditional 12 virtual voltage vector control set, designs 24 virtual voltage vectors with equal amplitude and uniform phase angle under the premise of not losing voltage utilization rate, and improves control precision. A duty cycle calculation method based on minimum error is proposed, which can track the currents of d-axis and q-axis at the same time even in the case of a single effective virtual voltage vector, and guarantees optimal output duty cycle. In addition, the process of traversing all voltage vectors in the prediction control is simplified, and the calculation burden of the algorithm is reduced. The application improves the precision of model prediction control by extending the control set and reducing the calculation error of the duty cycle, reduces 5th and 7th harmonics, and improves torque ripple. Moreover, even under the action of 24 voltage vectors, the calculation amount is still low, and the execution efficiency of the algorithm is improved.

Description

Technical Field

[0001] This invention belongs to the field of predictive control technology for multiphase motors, and particularly relates to a high-precision model predictive current control system and control method for dual three-phase motors. Background Technology

[0002] With the rapid development of high-end fields such as transportation, aerospace, and national defense, the requirements for motor systems, as core components of equipment, are becoming increasingly stringent. Multiphase permanent magnet motors, with their advantages of high power density, high efficiency, and good fault tolerance, have become the preferred choice for advanced motor systems. Among them, the dual three-phase permanent magnet synchronous motor with center-point isolation and two sets of windings connected by a 30° phase shift, has been widely used due to its special structure, which eliminates six torque ripples. Model predictive control strategies, due to their advantages of multivariable control, ease of handling nonlinear constraints, and intuitive implementation, have shown good performance in power converter applications and are gradually demonstrating their significant engineering application value. However, they suffer from drawbacks such as high computational load and high torque ripple. Chinese invention patent "A Model Predictive Torque Control Method for a Dual-Motor Series System with Low Computational Load" (Patent No.: 202110774817.4) discloses a model predictive control method with low computational load. This method only requires calculating the value functions of two voltage vectors, relatively reducing the computational load. However, this method requires calculating the location of the reference voltage vector, thus adding two observers and complicating the system. Chinese invention patent "A Model Predictive Control Method for Reducing Torque Ripple and Flux Flux in a PMSM" (Patent No.: 202210499366.2) discloses a method for reducing torque ripple and flux ripple in a permanent magnet synchronous motor (PMSM). While this method achieves some effect by using multiple voltage vectors within a single cycle to broaden the modulation range, it is computationally complex. When model predictive control algorithms are applied to multiphase motors, the number of candidate voltage vectors increases exponentially, leading to a corresponding increase in computational load. Furthermore, multiphase motors contain harmonic planes, which must be controlled during system operation; otherwise, motor performance will be negatively impacted, and significant losses will occur. Therefore, to improve the application of model predictive control in multiphase motors, research is urgently needed to reduce the computational burden of the algorithm, or to develop technologies that combine this with improvements in torque and flux ripple. Summary of the Invention

[0003] Objective: To address the issues of large torque ripple and heavy computational burden in model predictive control (MDC) of dual three-phase permanent magnet motors, this invention redesigns the control set. The traditional 12 virtual voltage vector control set is expanded to include 24 virtual voltage vectors with equal amplitude and uniform phase angles, improving control accuracy without sacrificing voltage utilization. Furthermore, a duty cycle calculation method based on minimum error is proposed, enabling simultaneous tracking of d-axis and q-axis currents even with a single effective virtual voltage vector, ensuring optimal output duty cycle. Additionally, the process of traversing all voltage vectors in predictive control is simplified, reducing the computational burden of the algorithm. This invention significantly improves the accuracy of model predictive control by expanding the control set and reducing duty cycle calculation errors, reducing 5th and 7th harmonics, and improving torque ripple. Moreover, even with 24 voltage vectors, the computational load remains low, improving the algorithm's execution efficiency.

[0004] Technical Solution: To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A high-precision model predictive current control system for a dual three-phase motor, comprising system hardware and system software (implemented in programming). The system hardware includes a dual three-phase permanent magnet motor, a DC power supply, a PWM module, an inverter, a position sensor, and a current sensor. The system software includes: a synthesized 24 virtual voltage vector module, a speed controller, a coordinate transformation module, a delay module, a prediction module, a duty cycle calculation module, a simplification module, and a value function module.

[0005] The dual three-phase permanent magnet motor consists of two sets of three-phase windings with a spatial phase shift of 30°. The inverter input is connected to a DC power supply, and the inverter signal is connected to a PWM module. The inverter has a six-phase two-level topology, and its output is connected to phases A, B, C, D, E, and F of the dual three-phase motor, responsible for converting the PWM signal into the six-phase sinusoidal AC power required to drive the motor. The position sensor uses a rotary transformer and is coaxially connected to the dual three-phase permanent magnet motor. The current sensor is connected to the inverter and is responsible for sampling the six-phase current of the motor.

[0006] The coordinate transformation module is connected to a current sensor at its input and to a delay compensation module at its output. It is used to convert the six-phase current in the natural coordinate system into the current in the rotating coordinate system, thereby achieving decoupled control.

[0007] The input end of the delay module is connected to the coordinate transformation module, and the output end is connected to the prediction model, in order to make up for the "one-shot delay" problem caused by the sampling of the digital system;

[0008] The input terminal of the prediction model module is connected to the delay compensation module, the 24 virtual voltage vector module and the position sensor, and is responsible for outputting the position of the dq axis current change under the action of different voltage vectors.

[0009] The speed controller is controlled by a PI controller to obtain the q-axis reference current. Its input is the error between the given speed and the actual speed, and its output is the reference value of the q-axis current.

[0010] The duty cycle calculation module is connected to the speed controller and the prediction model module at its input end, and is used to calculate the location of the optimal voltage vector and its duty cycle under the action of each voltage vector.

[0011] The simplified module and the value function module are connected to the duty cycle calculation module at their input ends to reduce the number of algorithm iterations and select the optimal vector and its duty cycle. The PWM module is connected to the value function module at its input end to convert the optimal vector and duty cycle obtained by the software system into the corresponding PWM signal, which is then output to the inverter to complete the modulation and drive the motor to run.

[0012] The present invention discloses a control method for a high-precision model predictive current control system for a dual three-phase motor, the control method comprising the following steps:

[0013] Step 1) Construct 24 virtual voltage vectors;

[0014] Step 2) Optimize the switching sequence of the voltage vector to standardize it;

[0015] Step 3) Obtain the rotational speed and position angle through the position sensor, obtain the six-phase current through the current sensor, and then obtain the current in the rotating coordinate system through the coordinate transformation module;

[0016] Step 4) Derive the prediction model for the dual three-phase permanent magnet motor;

[0017] Step 5) Calculate the duty cycle of the voltage vector action using the minimum error method;

[0018] Step 6) Simplify the traversal and optimization process;

[0019] Step 7) Select the optimal voltage vector and its duty cycle through the value function and output them to the PWM module. The inverter modulates the output of the corresponding voltage vector to complete the entire control.

[0020] Furthermore, the specific steps of step 1) include:

[0021] The dual three-phase permanent magnet motors are configured with neutral point isolation and driven by a six-phase two-level voltage source inverter. Since the upper and lower switching devices of each bridge arm operate in complementary conduction states, each bridge arm has two switching states, and the entire inverter has a total of 2... 6 =64 switching states, and the 64 voltage vectors corresponding to the changeover switches are determined by the following formula:

[0022]

[0023] Where a = e j30° s A ~s F These represent the switching states of each bridge arm, u αβ The voltage vector representing the αβ plane, u xy U represents the voltage vector in the xy plane. dc The DC bus voltage is represented by "1" when the upper bridge arm is on and "0" when the upper bridge arm is off. The basic voltage vectors are numbered in the order of ABC and DEF. The switch state combinations are represented in octal.

[0024] The virtual voltage vector principle requires that the sum of the effects of the vectors on the harmonic plane be zero, and its composition principle is as follows:

[0025]

[0026] Among them, u x i ,u y i D represents the components of the fundamental voltage vector along the x and y axes. i This indicates the duty cycle of each basic voltage vector.

[0027] To ensure voltage utilization, the 12 large vectors and 1 zero vector on the outermost edge of the fundamental plane are selected as the basic voltage vectors for synthesizing the virtual voltage vector. The new virtual voltage vector control set is synthesized using the adjacent three-vector principle, and the synthesis principle is as follows:

[0028]

[0029] Among them, V i Let u represent the i-th virtual voltage vector to be synthesized, where i = 1, 2, 3…24; 1st u 2nd , and u 3rd These represent the first, second, and third basic voltage vectors, respectively; the superscripts “α”, “β”, “x”, and “y” indicate the components of the voltage vector on the corresponding coordinate axes, and D1, D2, D3, and D0 represent the duty cycles of the first, second, and third basic voltage vectors and the zero vector, respectively.

[0030] The magnitude of each basic voltage vector is specified as 0.59Udc, the starting position is 0°, the angle between two adjacent voltage vectors is 15°, and finally 24 virtual voltage vectors are synthesized in the αβ plane, with zero components in the xy plane.

[0031] Furthermore, the specific steps of step 2) include:

[0032] To ensure the synthesized virtual voltage vectors can be implemented in industrial applications, the switching sequence of the 24 synthesized virtual voltage vectors was optimized and standardized, specifically at V2, V6, and V... 10 V 14 V 18 and V 22 The method of combining inner and outer two-layer voltage vectors is used instead of the method of combining adjacent three-vectors. The 24 virtual voltage vectors that are finally synthesized are shown in Table 1.

[0033] Table 1.24 Virtual Voltage Vector Distribution

[0034]

[0035] Among them, u1, ... u0, ... u 11 …u 66 …u 12 …u 64 These represent the corresponding basic voltage vectors.

[0036] Furthermore, the specific steps of step 3) include:

[0037] The position sensor measures the rotor's angular displacement and angular velocity, converts them into electrical signals, and transmits them to the controller. After decoding, the controller obtains the motor's speed and the rotor's position angle information.

[0038] The phase current of the motor sampled by 6 current sensors is denoted as: i A i B i C i D i E and i F The VSD coordinate transformation method is used to transform the variables of the natural coordinate system to the stationary coordinate system. The transformation matrix is ​​as follows:

[0039]

[0040] Among them, i α i β i x i y i o1 and i o2 Represents the currents along the α-axis, β-axis, x-axis, y-axis, o1-axis, and o2-axis of the stationary coordinate system;

[0041] For a dual three-phase permanent magnet motor, only the fundamental component of the αβ subspace participates in the electromechanical energy conversion. To simplify the analysis, the stationary coordinate system is transformed to a synchronous rotating coordinate system, and the transformation matrix is:

[0042]

[0043] Where θ is the rotor position angle, i d and i q These are the currents along the d-axis and q-axis, respectively.

[0044] The coordinate transformation module above calculates the motor current i in the dq rotating coordinate system at time k. dq (k).

[0045] Furthermore, the specific steps of step 4) include:

[0046] In model predictive control systems based on virtual voltage vectors, the harmonic plane can be neglected. Therefore, only the relevant variables of the dual three-phase permanent magnet motor in the fundamental plane need to be considered. Transforming these variables into a rotating coordinate system, the voltage equation of the motor is obtained as follows:

[0047]

[0048] In the formula, u d u q U s Components on the d-axis and q-axis, R s L is the stator resistance. d L q i d and i q Let ψ be the inductance and current along the d and q axes, respectively. f ω is the amplitude of the permanent magnet flux linkage. e Electric angular velocity;

[0049] Discretizing (6) using Euler's forward formula yields:

[0050]

[0051] Wherein, the superscript "k" represents the real-time values ​​of the dq-axis current and voltage at time k; the superscript "k+1" represents the predicted value of the dq-axis current at time k+1; T s To control the cycle;

[0052] To compensate for the "one-step delay" drawback of digital controllers, a two-step prediction method is used for delay compensation. (7) is predicted again to obtain the final prediction model:

[0053]

[0054] The superscript "pre" indicates the final predicted value of the dq axis current;

[0055] The value function is defined as:

[0056]

[0057] The superscript "*" indicates the reference value for the dq-axis current, using i d * =0 control.

[0058] Furthermore, the specific steps of step 5) include:

[0059] Taking the dq coordinate system as a reference, let i be the value of the zero voltage vector. d and i q The predicted value is point A(x1,y1), i under the action of the effective voltage vector. d and i q The predicted value is point B(x2,y2), and the location of the reference current is C(x0,y0). The distance from point C to line AB is the point with the minimum value of the value function, which is also the point with the minimum error. The required voltage vector can be obtained by finding the intersection of the line perpendicular to AB through point C and AB, and then the corresponding duty cycle can be obtained.

[0060] (5) The equation of the line AB:

[0061]

[0062] (6) The equation of the line passing through point C and perpendicular to AB:

[0063]

[0064] (7) Solve the equations (10) and (11) simultaneously to find the coordinates of the point where the two lines intersect:

[0065]

[0066] in, and This represents the predicted value of the dq-axis current after duty cycle correction;

[0067]

[0068] x1=i d k+1 +T s ·[-R s i d k+1 +ω e L q i q k+1 ] / L d

[0069] y1=i q k+1 +T s ·[-R s i q k+1 -ωe L d i d k+1 -ω e ψ f ] / L q

[0070]

[0071]

[0072] (8) The optimal duty cycle of the voltage vector can be determined based on the intersection point:

[0073]

[0074] Furthermore, the specific steps of step 6) include:

[0075] (6) With V4, V 10 V 16 V 22 The αβ plane is divided into four equal regions, named G1, G2, G3, and G4, with the boundary as the boundary. The virtual voltage vector contained in each region is shown in the table below:

[0076] Table 1 Virtual Voltage Vector Partitioning Rules

[0077]

[0078] (7) V1, V7, V 13 and V 19 Substituting into (5), we obtain the values ​​of the value functions under the action of the four voltage vectors: J(V1), J(V7), J(V... 13 ), J(V 19 Select the vector V that minimizes the value function. 1st This allows for the determination of the optimal region.

[0079] (8) Assuming V is determined in the second step 1st If V1 is the optimal region, then G1 is the optimal region. Then, recalculate V in G1. 13 and V 13 Given the value of the value function, select the vector V that minimizes the value function. 2nd Determine the second optimal region;

[0080] (9) Assuming V is determined in the second step 2nd If it is V1, then determine the value of the value function of V1 and the two adjacent voltage vectors V. 24 By selecting the optimal value function for V2, the final voltage vector index can be determined.

[0081] (10) Other cases follow the same pattern, and the combinations of all regions are shown in Table 2;

[0082] Table 2 shows all combinations of optimal voltage vector selection.

[0083]

[0084] After the simplification process described above, the original prediction process required traversing 24 voltage vectors, but now it only requires traversing 8, reducing the computational burden of the algorithm and improving efficiency.

[0085] Furthermore, the specific steps of step 7) include: inputting the 24 virtual voltage vectors VV one by one into the prediction model, selecting the optimal voltage vector and its duty cycle through the simplification module, outputting it to the PWM module, and outputting the corresponding voltage vector through inverter modulation to complete the entire control.

[0086] Beneficial effects of the present invention

[0087] 1) The present invention provides a high-precision model predictive current control system and control method for dual three-phase motors, which increases the number of traditional 12 virtual voltage vectors to 24, thereby expanding the modulation range without sacrificing voltage utilization.

[0088] 2) The duty cycle calculation method proposed in this invention considers both d-axis and q-axis tracking currents, reducing the minimum error, and the calculated value function is also the smallest among all value functions, thus improving control accuracy.

[0089] 3) Under the proposed duty cycle technology, the amplitude of each voltage vector can be flexibly changed, improving control accuracy, reducing the 5th and 7th harmonics, and improving torque pulsation and flux linkage pulsation.

[0090] 4) The simplified vector selection method proposed in this invention reduces the execution time of the model predictive control algorithm, improves the algorithm efficiency, and can be extended to other multiphase motor predictive control systems. Attached Figure Description

[0091] Figure 1 This is a schematic diagram illustrating the control principle of the method in an embodiment of the present invention;

[0092] Figure 2 A topology diagram of a six-phase voltage source inverter using the method of an embodiment of the present invention;

[0093] Figure 3 This is a spatial voltage vector diagram of the present invention; (a) αβ plane; (b) xy plane;

[0094] Figure 4 A schematic diagram of the maximum three-vector virtual voltage vector construction designed for this invention; (a) αβ plane; (b) xy plane;

[0095] Figure 5 The switching sequence diagram designed for this invention; (a) V2 before correction; (b) V2 after correction;

[0096] Figure 6 A schematic diagram of the inner and outer two-layer virtual voltage vector structure designed for this invention; (a) αβ plane; (b) xy plane;

[0097] Figure 7 The 24 virtual voltage vectors designed for this invention;

[0098] Figure 8 This is a schematic diagram of the minimum duty cycle calculation method proposed in this invention;

[0099] Figure 9 The following diagrams illustrate the simplified process of this invention: (a) is a diagram of region division; (b) is a diagram of the optimal region; and (c) is a diagram of the optimal voltage vector.

[0100] Figure 10 The experimental waveform diagram of the traditional 12 virtual voltage vectors under the action of deadbeat duty cycle;

[0101] Figure 11 This is an experimental waveform diagram of the present invention; Detailed Implementation

[0102] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0103] Figure 1The diagram below illustrates the control block principle of this invention. The control system hardware includes: a dual three-phase permanent magnet motor, a DC power supply, a PWM module, an inverter, a position sensor, and a current sensor. The control system software includes: a synthesized 24 virtual voltage vectors, a speed controller, a coordinate transformation module, a delay module, a prediction module, a duty cycle calculation module, and a simplification module. The dual three-phase permanent magnet motor consists of two sets of three-phase windings with a 30° spatial phase shift. The inverter input is connected to a DC power supply, and the signal input is connected to the PWM module of the software system. The inverter has a six-phase two-level topology, and its output is connected to phases A, B, C, D, E, and F of the dual three-phase motor, responsible for converting the PWM signal output by the software system into the six-phase sinusoidal AC current required to drive the motor. The position sensor uses a rotary transformer and is coaxially connected to the motor, transmitting the collected information to the software system. The current sensor is connected to the inverter and is responsible for sampling the six-phase current of the motor. The coordinate transformation module input is connected to the current sensor, and the output is connected to the delay compensation module, which converts the six-phase current in the natural coordinate system into the current in the rotating coordinate system, achieving decoupled control. The delay module input is connected to the coordinate transformation module, and the output is connected to the prediction model to compensate for the sampling bandwidth of the digital system. The method addresses the "one-shot delay" problem. The prediction model module's input is connected to a delay compensation module, a 24-virtual voltage vector module, and a position sensor, responsible for outputting the position of the dq-axis current change under different voltage vectors. The speed controller, controlled by a PI controller, obtains the q-axis reference current; its input is the error between the given speed and the actual speed, and its output is the reference value of the q-axis current. The duty cycle calculation module's input is connected to the speed controller and the prediction model module, used to calculate the position of the optimal voltage vector and its duty cycle under each voltage vector. The simplification module and value function module's input is connected to the duty cycle calculation module, used to reduce the algorithm's iteration count and select the optimal vector and its duty cycle. The PWM module's input is connected to the value function module, converting the optimal vector and duty cycle obtained by the software system into a corresponding PWM signal, outputting it to the inverter to complete modulation, thereby driving the motor. The implementation steps of the method mainly consist of the following steps:

[0104] Step 1: Construct 24 virtual voltage vectors (24 virtual voltage vector modules).

[0105] like Figure 2 As shown, the dual three-phase permanent magnet motor of this invention is configured with neutral point isolation and driven by a six-phase two-level voltage source inverter. Since the upper and lower switching devices of each bridge arm operate in complementary conduction states, each bridge arm has two switching states, and the entire inverter has a total of 2... 6 =64 switching states. The 64 voltage vectors corresponding to the changeover switches are determined by the following formula:

[0106]

[0107] Where a = e j30° s A ~s F Where u represents the switching state of each bridge arm. αβ The voltage vector representing the αβ plane, u xy The voltage vector represents the xy plane, and Udc represents the DC bus voltage. It is now defined that the upper bridge arm is "1" when it is on and "0" when it is off. The basic voltage vectors are numbered in the order ABC and DEF, and the switching states are represented using octal.

[0108] The final voltage vector is as follows Figure 3 As shown, Figure 3 (a) represents the voltage vectors in the αβ plane, responsible for electromechanical energy conversion. The 48 effective voltage vectors are divided into four layers, with amplitudes from the inside out of the layers: 0.173 Udc, 0.333 Udc, 0.471 Udc, and 0.644 Udc. (b) represents the xy plane, the harmonic plane responsible for generating losses. Similarly, the amplitudes from the inside out of the xy plane are: 0.173 Udc, 0.333 Udc, 0.471 Udc, and 0.644 Udc.

[0109] To suppress the voltage vector generated by the inverter in the harmonic plane, this invention proposes a novel virtual voltage vector synthesis method. The virtual voltage vector principle requires that the sum of the effects of the vectors in the harmonic plane be zero; the synthesis principle is as follows:

[0110]

[0111] in, This represents the components of the fundamental voltage vector along the x and y axes. (D) i This indicates the duty cycle of each basic voltage vector.

[0112] To ensure voltage utilization, this invention selects 12 large vectors and 1 zero vector on the outermost edge of the fundamental wave plane as the basic voltage vectors for synthesizing the virtual voltage vector. A new virtual voltage vector control set is synthesized using the principle of adjacent three vectors. For example... Figure 4 As shown, assume the target voltage vector is V3. V3 makes a 30° angle with the α axis, and its three adjacent fundamental voltage vectors are: u 66 u 64 and u 44 According to the principle of virtual voltage vector synthesis, we can conclude that:

[0113]

[0114] D1, D2, D3, and D0 can be obtained from equation (3).

[0115] Similarly, other virtual voltage vectors can be derived using this principle. The synthesis principle is summarized as follows:

[0116]

[0117] Among them, V i Represents the i-th virtual voltage vector to be synthesized (i = 1, 2, 3… 24); u 1st u 2nd , and u 3rd These represent the first, second, and third fundamental voltage vectors, respectively; the superscripts "α", "β", "x", and "y" indicate the components of the voltage vectors on the corresponding coordinate axes. D1, D2, D3, and D0 represent the duty cycles of the first, second, and third fundamental voltage vectors and the zero vector, respectively.

[0118] Step 2: Optimize the switching sequence of the voltage vector to standardize it.

[0119] Figure 5 (a) is a switching sequence diagram synthesized using the adjacent three-vector principle for the virtual voltage vector V2 (with an angle of 15° to the α axis). It can be seen from the diagram that the switching sequence of phase F requires two operations within one cycle, which is detrimental to digital processor implementation in industrial applications. Therefore, local adjustments are needed for the vector at this specific position, V2. The method is as follows: Figure 6 As shown, V2 uses two layers, inner and outer, to synthesize two voltage vectors that are in the same direction in the αβ plane but opposite in the xy plane. The resulting switching sequence is as follows: Figure 5 As shown in (b), it meets industrial requirements. Similarly, in V2, V6, V... 10 V 14 V 18 and V 22 The method of combining inner and outer voltage vectors is used instead of the method of combining adjacent three vectors.

[0120] After the first two steps, the 24 standard virtual voltage vectors synthesized are shown in Table 1.

[0121] Table 1.24 Virtual Voltage Vector Distribution

[0122]

[0123] Its distribution map is as follows Figure 7 As shown.

[0124] The third step is to obtain the rotational speed and position angle through the position sensor, obtain the six-phase current through the current sensor, and then obtain the current in the rotating coordinate system through the coordinate transformation module.

[0125] The position sensor measures the rotor's angular displacement and angular velocity, converts them into electrical signals, and transmits them to the controller. After decoding, the controller obtains the motor's speed and the rotor's position angle information.

[0126] This invention uses six current sensors to sample the phase current of the motor, denoted as: i A i B i C i D i E and i F The VSD coordinate transformation method is used to transform the variables of the natural coordinate system to the stationary coordinate system. The transformation matrix is ​​as follows:

[0127]

[0128] Among them, i α i β i x i y i o1 and i o2 This represents the current along the α-axis, β-axis, x-axis, y-axis, o1-axis, and o2-axis of the stationary coordinate system.

[0129] For a dual three-phase permanent magnet motor, only the fundamental component of the αβ subspace participates in the electromechanical energy conversion. To simplify the analysis, the stationary coordinate system is transformed to a synchronous rotating coordinate system, and the transformation matrix is:

[0130]

[0131] Where θ is the rotor position angle.

[0132] The coordinate transformation module above calculates the motor current i in the dq rotating coordinate system at time k. dq (k).

[0133] Step 4: Derive the prediction model for the dual three-phase permanent magnet motor.

[0134] This invention is a model predictive control system based on virtual voltage vectors, where the harmonic plane can be neglected. Therefore, only the relevant variables of the dual three-phase permanent magnet motor in the fundamental plane need to be considered. Transforming it to a rotating coordinate system, the voltage equation of the motor is obtained as follows:

[0135]

[0136] In the formula, u d u q U s Components on the d-axis and q-axis, R s L is the stator resistance. d L q i dand i q Let ψ be the inductance and current along the d and q axes, respectively. f ω is the amplitude of the permanent magnet flux linkage. e ω is the electric angular velocity.

[0137] Discretizing (7) using Euler's forward formula yields:

[0138]

[0139] Wherein, the superscript "k" represents the real-time values ​​of the dq-axis current and voltage at time k; the superscript "k+1" represents the predicted value of the dq-axis current at time k+1; T s To control the cycle.

[0140] To compensate for the "one-step delay" defect in digital controllers, a two-step prediction method is used for delay compensation (delay module). (8) is predicted again to obtain the final prediction model (prediction module):

[0141]

[0142] The superscript "pre" indicates the final predicted value of the dq axis current.

[0143] The value function is defined as:

[0144]

[0145] Wherein, the superscript "*" indicates the reference value for the dq-axis current, and this invention uses i d * =0 control.

[0146] Step 5: Calculate the duty cycle of the voltage vector action using the minimum error method (duty cycle calculation module).

[0147] Figure 8 This is a schematic diagram for calculating the minimum error duty cycle. Using the dq coordinate system as a reference, let i be the value under the action of the zero voltage vector. d and i q The predicted value is point A(x1,y1), i under the action of the effective voltage vector. d and i q The predicted value is at point B(x2,y2), and the reference current is located at C(x0,y0). Therefore, the blue dashed line represents the range of the voltage vector under duty cycle adjustment.

[0148] Based on the form of the value function J (value function module), the point where the predicted value is located... and the point where the reference current is located The distance between them represents the value of the value function. If duty cycle adjustment is not used, a complete voltage vector is applied within one control cycle, and its value function is represented by the orange line. When using traditional q-axis current deadbeat duty cycle calculation, the ordinate of the target point of the predicted value is i. q * At this point, draw i through point C. d The intersection of the line parallel to the axis and line AB is the prediction point for the deadbeat duty cycle, and its value function is represented by the green line segment. As we know from set theory, the value function of the duty cycle determined by either of these methods is not the minimum. The distance from point C to line AB is the point where the value function value is minimized, and also the point where the error is minimized. Finding the intersection of the line perpendicular to AB passing through point C and AB yields the required voltage vector, and thus the corresponding duty cycle.

[0149] (1) The equation of line AB:

[0150]

[0151] (2) The equation of the line passing through point C and perpendicular to AB:

[0152]

[0153] (3) Solving equations (11) and (12) simultaneously, we can find the coordinates of the point where the two lines intersect:

[0154]

[0155] Among them, among them, and This represents the predicted value of the dq-axis current after duty cycle correction;

[0156]

[0157] x1=i d k+1 +T s ·[-R s i d k+1 +ω e L q i q k+1 ] / L d

[0158] y1=i q k+1 +T s ·[-R s i q k+1 -ω e L d id k+1 -ω e ψ f ] / L q

[0159]

[0160]

[0161] (4) The optimal duty cycle of the voltage vector can be determined based on the location of the intersection point:

[0162]

[0163] Step 6: Simplify the traversal and optimization process.

[0164] (1) As Figure 9 As shown in (a), with V4 and V 10 V 16 V 22 The αβ plane is divided into four equal regions, named G1, G2, G3, and G4, with the boundary as the boundary. The virtual voltage vector contained in each region is shown in the table below:

[0165] Table 1 Virtual Voltage Vector Partitioning Rules

[0166]

[0167] (2) V1, V7, V 13 and V 19 Substituting into (5), we obtain the values ​​of the value functions under the action of the four voltage vectors: J(V1), J(V7), J(V... 13 ), J(V 19 Select the vector V that minimizes the value function. 1st This allows us to determine the optimal region.

[0168] (3) Figure 9 As shown in (b), assuming the V determined in the second step 1st If V is V1, then the optimal region is G1. Recalculate V in G1. 13 and V 13 Given the value of the value function, select the vector V that minimizes the value function. 2nd Then, determine the second optimal region.

[0169] (4) Figure 9 As shown in (c), assuming the V determined in the second step 2nd If it is V1, then determine the value of the value function of V1 and the two adjacent voltage vectors V. 24 By determining the value of V2 and the optimal value function, the final voltage vector index can be determined.

[0170] (5) Other cases follow the same pattern, and the combinations of all regions are shown in Table 2.

[0171] Table 2 shows all combinations of optimal voltage vector selection.

[0172]

[0173]

[0174] After the above simplification process (simplification module), the original prediction process required traversing 24 voltage vectors, but now only 8 need to be traversed, reducing the computational burden of the algorithm and improving efficiency.

[0175] Step 7: Input the 24 virtual voltage vectors VV one by one into the prediction model. Through the simplification module, select the optimal voltage vector and its duty cycle, and output it to the PWM module. Through inverter modulation, output the corresponding voltage vector to complete the entire control.

[0176] Figure 10 The image shows the experimental waveforms of a traditional 12 virtual voltage vector under the effect of deadbeat duty cycle technology, with a THD of 19.8% and a torque ripple of 10.2 Nm. d i q i x and i y The current ripples were 0.86A, 0.35A, 0.89A and 0.71A, respectively. Figure 11 The image shows the experimental waveforms under the action of the method proposed in this invention, with a THD of 7.5% and a torque ripple of 5.29 Nm. d i q i x and i y The current ripples were 0.33A, 0.18A, 0.43A, and 0.46A, respectively. Compared to traditional methods, this invention significantly improves motor performance.

[0177] The above embodiments are only used to illustrate the design concept and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, all equivalent changes or modifications made based on the principles and design ideas disclosed in the present invention are within the protection scope of the present invention.

Claims

1. A high-precision model predictive current control system for dual three-phase motors, characterized in that, It includes system hardware and system software. The system hardware includes dual three-phase permanent magnet motors, DC power supply, PWM module, inverter, position sensor, and current sensor. The system software includes: a synthesized 24 virtual voltage vector module, speed controller, coordinate transformation module, delay module, prediction module, duty cycle calculation module, simplification module, and value function module. The dual three-phase permanent magnet motor consists of two sets of three-phase windings with a spatial phase shift of 30°. The inverter input is connected to a DC power supply, and the inverter signal is connected to a PWM module. The inverter has a six-phase two-level topology, and its output is connected to phases A, B, C, D, E, and F of the dual three-phase motor, responsible for converting the PWM signal into the six-phase sinusoidal AC power required to drive the motor. The position sensor uses a rotary transformer and is coaxially connected to the dual three-phase permanent magnet motor. The current sensor is connected to the inverter and is responsible for sampling the six-phase current of the motor. The coordinate transformation module is connected to a current sensor at its input and to a delay compensation module at its output. It is used to convert the six-phase current in the natural coordinate system into the current in the rotating coordinate system, thereby achieving decoupled control. The input end of the delay module is connected to the coordinate transformation module, and the output end is connected to the prediction model, in order to make up for the "one-shot delay" problem caused by the sampling of the digital system; The prediction module input is connected to the delay compensation module, the 24 virtual voltage vector module and the position sensor, and is responsible for outputting the position of the dq axis current change under the action of different voltage vectors. The speed controller is controlled by a PI controller to obtain the q-axis reference current. Its input is the error between the given speed and the actual speed, and its output is the reference value of the q-axis current. The duty cycle calculation module is connected to the speed controller and the prediction model module at its input end, and is used to calculate the location of the optimal voltage vector and its duty cycle under the action of each voltage vector. The simplified module and the value function module are connected to the duty cycle calculation module at their input ends to reduce the number of algorithm iterations and select the optimal vector and its duty cycle; the PWM module is connected to the value function module at its input end to convert the optimal vector and duty cycle obtained by the software system into the corresponding PWM signal, which is then output to the inverter to complete the modulation and drive the motor to run. In the duty cycle calculation module, the duty cycle of the voltage vector action is calculated using the minimum error method as follows: i d and i q Let i be the current along the d-axis and q-axis, respectively. With the dq coordinate system as a reference, let i be the current under the action of the zero voltage vector. d and i q The predicted value is point A(x1, y1), i under the action of the effective voltage vector. d and i q The predicted value is point B(x2, y2), and the location of the reference current is C(x0, y0). The distance from point C to line AB is the point where the value function value is minimized, which is also the point where the error is minimized. By finding the intersection of the line perpendicular to AB through point C and AB, the required voltage vector can be obtained, and then the corresponding duty cycle can be derived.

2. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 1, characterized in that, The control method includes the following steps: Step 1) Construct 24 virtual voltage vectors; Step 2) Optimize the switching sequence of the voltage vector to standardize it; Step 3) Obtain the rotational speed and position angle through the position sensor, obtain the six-phase current through the current sensor, and then obtain the current in the rotating coordinate system through the coordinate transformation module; Step 4) Derive the prediction model for the dual three-phase permanent magnet motor; Step 5) Calculate the duty cycle of the voltage vector action using the minimum error method; Step 6) Simplify the traversal and optimization process; Step 7) Select the optimal voltage vector and its duty cycle through the value function and output it to the PWM module. The inverter modulates the output of the corresponding voltage vector to complete the entire control.

3. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 2, characterized in that, Step 1) includes the following specific steps: The dual three-phase permanent magnet motors are configured with neutral point isolation and driven by a six-phase two-level voltage source inverter. Since the upper and lower switching devices of each bridge arm operate in complementary conduction states, each bridge arm has two switching states, and the entire inverter has a total of 2... 6 =64 switching states, and the 64 voltage vectors corresponding to the changeover switches are determined by the following formula: (1)； Where a = e j30° s A ~s F These represent the switching states of each bridge arm, u αβ The voltage vector representing the αβ plane, u xy U represents the voltage vector in the xy plane. dc The DC bus voltage is represented by "1" when the upper bridge arm is on and "0" when the upper bridge arm is off. The basic voltage vectors are numbered in the order of ABC and DEF. The switch state combinations are represented in octal. The virtual voltage vector principle requires that the sum of the effects of the vectors on the harmonic plane be zero, and its composition principle is as follows: (2)； in, , D represents the components of the fundamental voltage vector along the x and y axes. i This indicates the duty cycle of each basic voltage vector. To ensure voltage utilization, the 12 large vectors and 1 zero vector on the outermost edge of the fundamental plane are selected as the basic voltage vectors for synthesizing the virtual voltage vector. The new virtual voltage vector control set is synthesized using the adjacent three-vector principle, and the synthesis principle is as follows: (3)； Among them, V i Let u represent the i-th virtual voltage vector to be synthesized, where i = 1, 2, 3 … 24; 1st u 2nd , and u 3rd These represent the first, second, and third basic voltage vectors, respectively; the superscripts "α", "β", "x", and "y" represent the components of the voltage vectors on the corresponding coordinate axes, and D1, D2, D3, and D0 represent the duty cycles of the first, second, and third basic voltage vectors and the zero vector, respectively. The magnitude of each basic voltage vector is specified as 0.59Udc, the starting position is 0°, the angle between two adjacent voltage vectors is 15°, and finally 24 virtual voltage vectors are synthesized in the αβ plane, with zero components in the xy plane.

4. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 3, characterized in that, Step 2) includes the following specific steps: To ensure the synthesized virtual voltage vectors can be implemented in industrial applications, the switching sequence of the 24 synthesized virtual voltage vectors was optimized and standardized, specifically at V2, V6, and V... 10 V 14 V 18 and V 22 The method of combining inner and outer two-layer voltage vectors is used instead of the method of combining adjacent three-vectors. The 24 virtual voltage vectors that are finally synthesized are shown in Table 1. Table 1 shows the distribution of 24 virtual voltage vectors: ； Among them, u1, ... u0, ... u 11 …u 66 …u 12 …u 64 These represent the corresponding 64 basic voltage vectors.

5. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 4, characterized in that, Step 3) includes the following specific steps: The position sensor measures the rotor's angular displacement and angular velocity, converts them into electrical signals, and transmits them to the controller. After decoding, the controller obtains the motor's speed and the rotor's position angle information. The phase current of the motor sampled by 6 current sensors is denoted as: i A i B i C i D i E and i F The VSD coordinate transformation method is used to transform the variables of the natural coordinate system to the stationary coordinate system. The transformation matrix is ​​as follows: (4)； Among them, i α i β i x i y i o1 and i o2 Represents the currents along the α-axis, β-axis, x-axis, y-axis, o1-axis, and o2-axis of the stationary coordinate system; For a dual three-phase permanent magnet motor, only the fundamental component of the αβ subspace participates in the electromechanical energy conversion. To simplify the analysis, the stationary coordinate system is transformed to a synchronous rotating coordinate system, and the transformation matrix is: (5)； Where θ is the rotor position angle, i d and i q These are the currents along the d-axis and q-axis, respectively. The coordinate transformation module above calculates the motor current i in the dq rotating coordinate system at time k. dq (k).

6. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 5, characterized in that, Step 4) includes the following specific steps: In model predictive control systems based on virtual voltage vectors, the harmonic plane can be neglected. Therefore, only the relevant variables of the dual three-phase permanent magnet motor in the fundamental plane need to be considered. Transforming these variables into a rotating coordinate system, the voltage equation of the motor is obtained as follows: (6)； In the formula, u d u q U s Components on the d-axis and q-axis, R s L is the stator resistance. d L q i d and i q These represent the dq-axis inductance and current, respectively. This represents the amplitude of the permanent magnet flux linkage. Electric angular velocity; Discretizing (6) using Euler's forward formula yields: (7)； Wherein, the superscript "k" represents the real-time values ​​of the dq-axis current and voltage at time k; the superscript "k+1" represents the predicted value of the dq-axis current at time k+1; T s To control the cycle; To compensate for the "one-step delay" drawback of digital controllers, a two-step prediction method is used for delay compensation. (7) is predicted again to obtain the final prediction model: (8)； The superscript "pre" indicates the final predicted value of the dq axis current; The value function is defined as: (9)； The superscript "*" indicates the reference value for the dq-axis current, using i d * = 0 control.

7. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 6, characterized in that, Step 5) involves the following steps for calculating the duty cycle of the voltage vector action: (1) The equation of the line AB: (10); (2) The equation of the line passing through point C and perpendicular to AB: (11); (3) Solve equations (10) and (11) simultaneously to find the coordinates of the point where the two lines intersect: (12)； Where ipre d_duty and ipre q_duty represent the predicted values ​​of the dq axis current after duty cycle correction; ； (4) The optimal duty cycle of the voltage vector can be determined based on the location of the intersection point: (13).

8. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 5, characterized in that, Step 6) includes the following specific steps: (1) With V4, V 10 V 16 V 22 The αβ plane is divided into four equal regions, named G1, G2, G3, and G4, with the boundary as the boundary. The virtual voltage vector contained in each region is shown in the table below: Table 1 shows the virtual voltage vector partitioning rules: ； (2) V1, V7, V 13 and V 19 Substituting into (5), we obtain the values ​​of the value functions under the action of the four voltage vectors: J(V1), J(V7), J(V... 13 ), J(V 19 Select the vector V that minimizes the value function. 1st This allows for the determination of the optimal region. (3) Assume that V is determined in the second step 1st If V1 is the optimal region, then G1 is the optimal region. Then, recalculate V in G1. 13 and V 13 Given the value of the value function, select the vector V that minimizes the value function. 2nd Determine the second optimal region; (4) Assume that V is determined in the second step 2nd If it is V1, then determine the value of the value function of V1 and the two adjacent voltage vectors V. 24 By selecting the optimal value function for V2, the final voltage vector index can be determined. (5) Other cases follow the same pattern, and the combinations of all regions are shown in Table 2; Table 2 shows all combinations of optimal voltage vector selection: ； After the simplification process described above, the original prediction process required traversing 24 voltage vectors, but now it only requires traversing 8, reducing the computational burden of the algorithm and improving efficiency.

9. The control method for a high-precision model predictive current control system for a dual three-phase motor according to claim 2, characterized in that, Step 7) includes the following steps: Substitute each of the 24 virtual voltage vectors VV into the prediction model, select the optimal voltage vector and its duty cycle through the simplification module, output it to the PWM module, and output the corresponding voltage vector through inverter modulation to complete the entire control.

Images