Multi-vector control method for open-winding permanent magnet synchronous motor based on three-dimensional space division

Abstract

The application relates to a kind of open-winding permanent magnet synchronous motor multi-vector control methods based on three-dimensional space division, belong to open-winding permanent magnet synchronous motor control technical field, comprising: obtaining the rotor position angle of motor θ And angular velocity ω , calculate the stator current under two-phase static coordinate system i d 、 i q 、 i 0;According to the q-axis current reference value obtained by the speed outer ring controller, the d-axis and 0-axis current reference values are given;The prediction of reference control voltage;Reference control voltage vector belongs to the area judgment and voltage vector selection;The calculation of voltage vector action time;According to the obtained action time of each space voltage vector, the switch pulse sequence of each group of bridge arms is distributed and generated.The application reduces the calculation amount and complexity of system control algorithm, effectively reduces the electromagnetic torque ripple of open-winding permanent magnet synchronous motor, and solves the zero sequence current suppression problem existing in the common DC bus driving system.

Description

Technical Field

[0001] This invention relates to a multi-vector control method for open-winding permanent magnet synchronous motors based on three-dimensional spatial partitioning, belonging to the field of open-winding permanent magnet synchronous motor control technology. Background Technology

[0002] Open-winding permanent magnet synchronous motors (PMSMs) are based on traditional PMSMs with the neutral point open. Six leads are drawn from both ends of the three-phase windings, with three leads connected to an inverter, employing a dual-inverter topology to drive the motor. Based on the power supply structure of the drive system, open-winding PMSM drive systems can be divided into two main categories: common DC bus topology and independent DC bus topology. The common DC bus topology requires only one DC power source, which is more advantageous for increasing motor output power and widening the motor's speed range within limited power supply capacity, and its application in motor drive systems has attracted widespread attention.

[0003] Traditional vector control strategies based on linear regulators such as PI and PR are complex in structure and have poor suppression effect on the zero-sequence current problem in common DC bus drive systems. In recent years, model predictive control has received widespread attention due to its advantages such as simple control concept, fast dynamic response, and strong ability to handle multivariable nonlinear systems. It also has good application potential in solving the multi-constraint problems unique to open-winding motors.

[0004] However, traditional model predictive control suffers from problems such as large electromagnetic torque ripple in motors, high computational complexity of control algorithms, and complex engineering applications. To better suppress electromagnetic torque ripple, scholars both domestically and internationally have proposed multi-vector control strategies such as duty cycle control, two-vector control, and three-vector control. However, traditional multi-vector control requires multiple traversals of voltage vectors, and the open-winding motor drive system itself has a large number of voltage vectors, which inevitably increases the computational workload of the algorithm and brings great difficulties to its implementation.

[0005] Regarding reducing the algorithmic complexity of model predictive control, some literature proposes using the angle information of the reference control voltage vector to filter candidate voltage vectors, reducing the number of effective voltage vectors to 5, which greatly reduces the number of voltage vector traversals. Other scholars have reduced the number of candidate voltage vectors to 3 by combining the zero-sequence voltage with the perpendicular bisector of the space, which also effectively reduces the computational load of the algorithm. However, the above algorithms are mostly implemented using single-vector control, resulting in large ripples in the electromagnetic torque of the motor and poor zero-sequence current suppression, which require further improvement. Summary of the Invention

[0006] The purpose of this invention is to provide a multi-vector control method for open-winding permanent magnet synchronous motors based on three-dimensional spatial partitioning, which can quickly filter effective voltage vectors, reduce the computational load and application complexity of the algorithm, and suppress electromagnetic torque pulsation and zero-sequence current of the motor.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A multi-vector control method for an open-winding permanent magnet synchronous motor based on three-dimensional spatial partitioning includes the following steps:

[0009] Step 1: Obtain the rotor position angle θ and angular velocity ω of the motor, and calculate the stator current i in the two-phase rotating coordinate system. d i q 、i0;

[0010] Step 2: Obtain the q-axis current reference value based on the speed outer loop controller, and provide the d-axis and 0-axis current reference values;

[0011] Step 3: Predict the reference control voltage;

[0012] Step 4: Determine the region to which the reference control voltage vector belongs and select the voltage vector;

[0013] Step 5: Calculation of voltage vector action time;

[0014] Step 6: Based on the obtained spatial voltage vector action time, allocate and generate the switching pulse sequence for each group of bridge arms.

[0015] A further improvement to the technical solution of the present invention is that the specific process of step 1 is as follows:

[0016] The rotor position angle θ of the motor is obtained by an encoder, and the rotor angular velocity ω is obtained by differentiation. The three-phase stator current i is obtained by measuring a current Hall sensor. a i b i c Transform it to a two-phase rotating coordinate system using coordinate transformation to obtain i d i q 、i0.

[0017] A further improvement to the technical solution of the present invention is that the specific process of step 2 is as follows:

[0018] Reference rotation speed ω ref The difference between the angular velocity ω obtained from the encoder is input to the PI controller, and the output is used as the reference value for the q-axis current. The d-axis current reference value i is then used as the reference value for the d-axis current. dref Set to zero, and in order to suppress zero-sequence current, set the 0-axis current reference value i 0ref It is also set to zero.

[0019] A further improvement to the technical solution of the present invention is that the specific process of step 3 is as follows:

[0020] The discrete mathematical model of an open-winding permanent magnet synchronous motor is shown in the following equation:

[0021]

[0022] Among them, u dq0 (k) represents the stator voltage in the dq0 coordinate system at time k, i dq0 (k) and i dq0 (k+1) represent the stator currents in the dq0 coordinate system at times k and k+1, respectively; matrices A, B, and C are shown below:

[0023]

[0024] The stator dq0 axis current i at time k+1 dq0 (k+1) uses the reference current i dq0ref (k) Substitution yields the reference control voltage u. dq0ref The expression for (k) is shown below:

[0025]

[0026] The reference control voltage u in the dq0 coordinate system is obtained through coordinate transformation. dq0ref Transform to a three-phase stationary coordinate system to obtain the reference control voltage vector U in the three-phase stationary coordinate system. aref U bref U cref .

[0027] A further improvement to the technical solution of the present invention is that step 4 specifically comprises the following steps:

[0028] First, a three-dimensional spatial coordinate system is established, with the zero voltage vector as the origin, the A-phase voltage vector as the x-axis, the B-phase voltage vector as the y-axis, and the C-phase voltage vector as the z-axis. The three-dimensional spatial coordinate system is then established, and all voltage vectors are normalized to ensure that the coordinates of all voltage vectors contain only 0 or 1.

[0029] Then, the three-dimensional space region is divided into 8 small cubes of equal volume, centered on the origin and according to the quadrant in which they are located; each small cube is then divided into 6 cones of equal volume. Each cone region consists of only three non-coplanar non-zero space voltage vectors and one zero space voltage vector, so as to facilitate the multi-vector synthesis of the reference voltage vector in the three-dimensional space.

[0030] Finally, determine the cone region to which the reference control voltage vector belongs; based on U aref U bref U cref The size of the cone determines its quadrant, and the plane equations of the six cone regions in the quadrant are used as constraints for determining its region. The plane equations of each cone region are obtained using the point normal method. Assuming that two direction vectors are formed by three points, s1 and s2, the normal vector s(a,b,c) of the plane can be obtained by the cross product of the two vectors s1×s2. Selecting any point (x0,y0,z0) in the plane, the plane equation can be obtained as follows:

[0031]

[0032] After calculating all the plane equations of the six cones, all the plane equations of the region can be used as constraints for that region, uniquely determining the region to which the reference control voltage belongs. Three non-zero space voltage vectors and one zero space voltage vector within this cone region are selected to synthesize the reference control voltage vector U. aref U bref U cref And it will act on the next control cycle.

[0033] A further improvement to the technical solution of the present invention is that step 5 specifically comprises the following steps:

[0034] Based on the three-dimensional spatial coordinate operation rules, the action time of each voltage vector is calculated for synthesizing the reference control voltage. Assume the coordinates of three non-zero spatial voltage vectors within a certain region are (m1, n1, k1), (m2, n2, k2), and (m3, n3, k3), and their action time accounts for a portion of the control period T. s The proportions are d1, d2, and d3, respectively. The coordinates of the reference control voltage are (x, y, z). Therefore, the formula for calculating the duration of the three non-zero space voltage vectors is:

[0035]

[0036] The obtained d1, d2, and d3 are compared with the control period T. s Multiplying these values ​​yields the action times t1, t2, and t3 of the three non-zero space voltage vectors. When the sum of these three action times is less than the control period T... s When this happens, a zero-space voltage vector can be used for filling, and the duration of the zero-space voltage vector is t0 = T. s - (t1+t2+t3).

[0037] Due to the adoption of the above technical solution, the technical effects achieved by this invention are as follows:

[0038] This invention primarily achieves rapid location and selection of multiple effective voltage vectors by dividing the vector region within a three-dimensional spatial coordinate system. Furthermore, combining vector coordinate arithmetic rules, it presents a simpler method for calculating the voltage vector action time, further optimizing the multi-vector model predictive control algorithm. Compared to traditional technologies, the control method proposed in this invention reduces the computational load and complexity of the system control algorithm, effectively reduces the electromagnetic torque ripple of open-winding permanent magnet synchronous motors, and solves the problem of zero-sequence current suppression in common DC bus drive systems. Attached Figure Description

[0039] Figure 1 It is a common DC bus open winding motor system structure;

[0040] Figure 2 It is a three-dimensional spatial coordinate system established with the voltage vector of phase A as the x-axis, the voltage vector of phase B as the y-axis, and the voltage vector of phase C as the z-axis;

[0041] Figure 3 It describes the division structure of the small cubes in the first quadrant and the specific shapes of the six cones formed after the division;

[0042] Figure 4 This is a control block diagram of the entire invention;

[0043] Figure 5 The waveforms of the three-phase current, speed, electromagnetic torque and zero-sequence current of this invention are shown. Detailed Implementation

[0044] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Obviously, the described content is only a part of the invention, not all of it. Based on the content of this invention, other research content obtained by those skilled in the art without innovative effort should fall within the scope of protection of this invention.

[0045] The research content of this invention is a drive system for an open-winding permanent magnet synchronous motor based on a common DC bus topology, the structure of which is as follows: Figure 1 As shown, the open-winding permanent magnet synchronous motor drive system has nonlinear characteristics such as multiple spatial voltage vectors and multiple control constraints.

[0046] A multi-vector model predictive control method for open-winding permanent magnet synchronous motors based on three-dimensional spatial vector partitioning, the principle of which is as follows: Figure 4As shown, the control system of the open-winding permanent magnet synchronous motor includes the following modules: coordinate transformation module, speed outer loop PI controller, reference voltage prediction module, region judgment and voltage vector selection module, action time calculation module, switching pulse generation module, inverter module, encoder, and open-winding permanent magnet synchronous motor. The inverter module supplies power to the open-winding permanent magnet synchronous motor via a common DC bus, and its control is generated by the switching pulse generation module. The encoder collects the rotor position data of the open-winding permanent magnet synchronous motor and sends the information to the coordinate transformation module and the speed outer loop PI controller. There are two coordinate transformation modules: a stator current coordinate transformation module and a reference voltage coordinate transformation module. The stator current coordinate transformation module transforms the current i in the three-phase stationary coordinate system collected by the current sensor... a i b i c Converted to current i in a two-phase stationary coordinate system d i q i0, and send it to the reference voltage prediction module. The reference voltage coordinate transformation transforms the reference voltage prediction value u in the two-phase rotating coordinate system. dref u qref u 0ref Voltage U converted to a three-phase stationary coordinate system aref U bref U cref The data is then sent to the area judgment and voltage vector selection module for voltage vector selection. The action time calculation module calculates the action time of each voltage vector, and finally, the switching pulse generation module generates switching pulses according to the action sequence and action time of each vector to control the inverter and drive the motor.

[0047] First, the rotor position angle θ and angular velocity ω of the motor are obtained through an encoder, and the three-phase stator current i is measured by a Hall sensor. a i b i c First, the three-phase stator current is transformed into a two-phase rotating coordinate system using coordinate transformation; second, the motor angular velocity setpoint ω is... ref The difference between the actual angular velocity ω and the actual angular velocity ω is used to obtain the reference value i of the stator q-axis current through a PI controller. qref The stator d-axis current reference value i dref The reference value for the stator 0-axis current is set to zero, and in order to suppress zero-sequence current, the reference value i is set to zero. 0refIt is also set to zero; then, based on the concept of deadbeat predictive control, a deadbeat current predictive controller is constructed using the discrete mathematical model of the open-winding permanent magnet synchronous motor to predict the reference control voltage; then, the voltage vector in the corresponding region is selected based on the phase-advancing region of the reference voltage and applied to the next control cycle, and the time coordinate of each voltage vector is calculated; finally, a switching pulse is generated based on the action sequence and action time of the voltage vector and applied to the inverter.

[0048] Specifically, the following steps are included:

[0049] Step 1: Obtain the rotor position angle θ and angular velocity ω of the motor, and calculate the stator current i in the two-phase stationary coordinate system. d i q 、i0.

[0050] The rotor position angle θ of the motor is obtained by an encoder, and the rotor angular velocity ω is obtained by differentiation. The three-phase stator current i is obtained by measuring a current Hall sensor. a i b i c Transform it to a two-phase rotating coordinate system using coordinate transformation to obtain i d i q 、i0.

[0051] Step 2: Obtain the q-axis current reference value based on the speed outer loop controller, and provide the d-axis and 0-axis current reference values.

[0052] Reference rotation speed ω ref The difference between the angular velocity ω obtained from the encoder is input to the PI controller, and the output is used as the reference value for the q-axis current. The d-axis current reference value i is then used as the reference value for the d-axis current. dref Set to zero, and in order to suppress zero-sequence current, set the 0-axis current reference value i 0ref It is also set to zero.

[0053] Step 3: Prediction of the reference voltage vector.

[0054] Based on the concept of deadbeat predictive control, a deadbeat current predictive controller is constructed using the discrete mathematical model of an open-winding permanent magnet synchronous motor to predict the reference control voltage.

[0055] The discrete mathematical model of an open-winding permanent magnet synchronous motor is shown in the following equation:

[0056]

[0057] Among them, u dq0 (k) represents the stator voltage in the dq0 coordinate system at time k, i dq0 (k) and i dq0(k+1) represent the stator currents in the dq0 coordinate system at times k and k+1, respectively. Matrices A, B, and C are shown below;

[0058]

[0059] The stator dq0 axis current i at time k+1 dq0 (k+1) uses the reference current i dq0ref (k) Substitution yields the reference control voltage u. dq0ref The expression for (k) is shown below:

[0060]

[0061] The reference control voltage u in the dq0 coordinate system is obtained through coordinate transformation. dq0ref Transform to a three-phase stationary coordinate system to obtain the reference control voltage vector U in the three-phase stationary coordinate system. aref U bref U cref .

[0062] Step 4: Determine the region to which the reference control voltage vector belongs and select the voltage vector.

[0063] First, a three-dimensional spatial coordinate system is established, with the zero voltage vector as the origin, the A-phase voltage vector as the x-axis, the B-phase voltage vector as the y-axis, and the C-phase voltage vector as the z-axis. The three-dimensional spatial coordinate system is then normalized to ensure that the coordinates of all voltage vectors contain only 0 or 1.

[0064] Then, the three-dimensional space is divided into eight small cubes of equal volume, centered on the origin and according to their respective quadrants, such as... Figure 2 As shown. Each small cube is then divided into 6 cones of equal volume. Each cone region consists of exactly three non-coplanar non-zero space voltage vectors and one zero space voltage vector, so as to facilitate the multi-vector synthesis of the reference voltage vector in three-dimensional space. Figure 3 The diagram shows the method for dividing the six cones in the first quadrant.

[0065] Finally, the cone region to which the reference control voltage vector belongs is determined. According to U aref U bref U crefThe magnitude of the cone shape determines its quadrant, and the plane equations of the six cone regions within that quadrant serve as constraints for determining its region affiliation. The plane equations of each cone region can be obtained using the point-normal method. Assuming three points form two direction vectors s1 and s2, the normal vector s(a,b,c) of the plane can be obtained by the cross product of these two vectors, s1×s2. Selecting any point (x0,y0,z0) in the plane, the plane equation can be obtained as follows:

[0066]

[0067] After calculating all the plane equations of the six cones, all the plane equations of the region can be used as constraints for that region, uniquely determining the region to which the reference control voltage belongs. Three non-zero space voltage vectors and one zero space voltage vector within this cone region are selected to synthesize the reference control voltage vector U. aref U bref U cref And it will act on the next control cycle.

[0068] Step 5: Calculation of action time.

[0069] Based on the three-dimensional spatial coordinate calculation rules, the action time of each voltage vector is calculated for synthesizing the reference control voltage. Assume the coordinates of three non-zero spatial voltage vectors within a certain region are (m1, n1, k1), (m2, n2, k2), and (m3, n3, k3), and their action time accounts for a portion of the control period T. s The proportions are d1, d2, and d3, respectively. The coordinates of the reference control voltage are (x, y, z). Therefore, the formula for calculating the duration of the three non-zero space voltage vectors is:

[0070]

[0071] The obtained d1, d2, and d3 are compared with the control period T. s Multiplying these values ​​yields the action times t1, t2, and t3 of the three non-zero space voltage vectors. When the sum of these three action times is less than the control period T... s When this happens, a zero-space voltage vector can be used for filling, and the duration of the zero-space voltage vector is t0 = T. s - (t1+t2+t3).

[0072] Step 6: Generate switching pulses.

[0073] Based on the obtained duration of each spatial voltage vector, the switching pulse sequence of each bridge arm is allocated and generated to control the turn-on and turn-off of the main power switch of the converter, thereby driving the motor to run.

[0074] Figure 5The simulation results show the waveforms of the three-phase current, speed, electromagnetic torque, and zero-sequence current of this invention. The simulation was conducted under the following conditions: a given motor speed of 500 r / min and a given torque of 3 N·m. At 0.4 s, the given torque jumped from 3 N·m to 6 N·m. The simulation demonstrated that the pulsations in electromagnetic torque and zero-sequence current were effectively suppressed, illustrating the feasibility and superiority of the multi-vector model predictive control method for open-winding permanent magnet synchronous motors based on three-dimensional spatial vector partitioning.

[0075] This invention, based on the concept of three-dimensional spatial vector partitioning, provides a multi-vector model predictive control method for open-winding permanent magnet synchronous motors. This method ensures rapid voltage vector selection while effectively reducing electromagnetic torque ripple and suppressing zero-sequence current. Furthermore, by incorporating vector coordinate arithmetic rules, a simpler method for calculating the voltage vector action time is presented, further reducing the computational load and application complexity of the algorithm.

Claims

1. A multi-vector control method for an open-winding permanent magnet synchronous motor based on three-dimensional spatial partitioning, characterized in that, The multi-vector control method for open-winding permanent magnet synchronous motors is based on the concept of deadbeat predictive control. It utilizes the discrete mathematical model of the open-winding permanent magnet synchronous motor to construct a deadbeat current predictive controller for predicting the reference control voltage. The method includes the following steps: Step 1: Obtain the rotor position angle θ and angular velocity ω of the motor, and calculate the stator current i in the two-phase rotating coordinate system. d i q The specific process is as follows: the rotor position angle θ of the motor is obtained through the encoder, the rotor angular velocity ω is obtained after differentiation, and the three-phase stator current i is obtained by measuring the current Hall sensor. a i b i c Transform it to a two-phase rotating coordinate system using coordinate transformation to obtain i d i q 、i0; Step 2: Obtain the q-axis current reference value based on the outer loop speed controller, and provide the d-axis and 0-axis current reference values. The specific process is as follows: Set the reference speed ω... ref The difference between the angular velocity ω obtained from the encoder is input to the PI controller, and the output is used as the reference value for the q-axis current. The d-axis current reference value i is then used as the reference value for the d-axis current. dref Set to zero, and in order to suppress zero-sequence current, set the 0-axis current reference value i 0ref Also set to zero; Step 3: Prediction of the reference control voltage. The reference control voltage u in the dq0 coordinate system is obtained through coordinate transformation. dq0ref Transform to a three-phase stationary coordinate system to obtain the reference control voltage vector U in the three-phase stationary coordinate system. aref U bref U cref ; Step 4: Based on the region determination and voltage vector selection of the control voltage vector, first, establish a three-dimensional spatial coordinate system, select the zero voltage vector as the origin, take the A-phase voltage vector as the x-axis, the B-phase voltage vector as the y-axis, and the C-phase voltage vector as the z-axis to establish a three-dimensional spatial coordinate system, and normalize all voltage vectors to ensure that the coordinates of all voltage vectors contain only 0 or 1. Then, the three-dimensional space region is divided into 8 small cubes of equal volume, centered on the origin and according to the quadrant in which they are located; each small cube is then divided into 6 cones of equal volume. Each cone region consists of only three non-coplanar non-zero space voltage vectors and one zero space voltage vector, so as to facilitate the multi-vector synthesis of the reference voltage vector in the three-dimensional space. Finally, determine the cone region to which the reference control voltage vector belongs; based on U aref U bref U cref The size of the cone determines its quadrant, and the plane equations of the six cone regions in the quadrant are used as constraints for determining its region; the plane equations of each cone region are obtained using the point method. Step 5: Calculation of voltage vector action time. Based on the three-dimensional spatial coordinate operation rules, calculate the action time of each voltage vector to synthesize the reference control voltage. Step 6: Based on the obtained spatial voltage vector action time, allocate and generate the switching pulse sequence for each group of bridge arms.

2. The multi-vector control method for an open-winding permanent magnet synchronous motor based on three-dimensional spatial partitioning according to claim 1, characterized in that, The specific process of step 3 is as follows: The discrete mathematical model of an open-winding permanent magnet synchronous motor is shown in the following equation: Among them, u dq0 (k) represents the stator voltage in the dq0 coordinate system at time k, i dq0 (k) and i dq0 (k+1) represent the stator currents in the dq0 coordinate system at times k and k+1, respectively; matrices A, B, and C are shown below: The stator dq0 axis current i at time k+1 dq0 (k+1) uses the reference current i dq0ref (k) Substitution yields the reference control voltage u. dq0ref The expression for (k) is shown below: 。 3. The multi-vector control method for an open-winding permanent magnet synchronous motor based on three-dimensional spatial partitioning according to claim 1, characterized in that, The specific steps of step 4 are as follows: Suppose that three points form two direction vectors, s1 and s2. Then the normal vector s(a,b,c) of the plane can be obtained by the cross product of the two vectors, s1×s2. Choosing any point (x0,y0,z0) in the plane, the equation of the plane can be obtained as follows: After calculating all the plane equations of the six cones, all the plane equations of the region can be used as the constraints of the region, and the region to which the reference control voltage belongs can be uniquely determined. Three non-zero space voltage vectors and one zero space voltage vector within the cone region are selected to synthesize the reference control voltage vector U. aref U bref U cref And it will act on the next control cycle.

4. The multi-vector control method for an open-winding permanent magnet synchronous motor based on three-dimensional spatial partitioning according to claim 1, characterized in that, The specific steps of step 5 are as follows: Assume the coordinates of three non-zero space voltage vectors within a certain region are (m1, n1, k1), (m2, n2, k2), and (m3, n3, k3), and their duration of action is 1 / 3 of the control period T. s The proportions are d1, d2, and d3, respectively. The coordinates of the reference control voltage are (x, y, z). Therefore, the formula for calculating the duration of the three non-zero space voltage vectors is: The obtained d1, d2, and d3 are compared with the control period T. s Multiplying these values ​​yields the action times t1, t2, and t3 of the three non-zero space voltage vectors. When the sum of these three action times is less than the control period T... s When this happens, a zero-space voltage vector can be used for filling, and the duration of the zero-space voltage vector is t0 = T. s - (t1+t2+t3).

Images