A multi-vector model predictive control method for permanent magnet synchronous motor considering vector screening

Abstract

The present invention discloses a multi-vector model predictive control method for a permanent magnet synchronous motor taking into account vector screening. First, a current reference value is obtained, and a q-axis current reference value i q * The d-axis current reference value i is obtained through the speed regulator. d * Set to 0 to implement i d =0 control; secondly, according to the deadbeat control idea, the current prediction model is used to calculate the expected voltage vector dq axis component v d * With v q * , and calculate the angle of the expected voltage vector in the αβ coordinate system; again, according to the voltage vector angle, select two adjacent active voltage vectors; then, use the error evaluation function to evaluate the selected active voltage vector and zero vector, determine the optimal action time of the two active vectors and the zero vector by the Lagrange multiplier method, and obtain the optimal switching signal of the inverter. The permanent magnet synchronous motor model predictive control method provided by the present invention not only reduces the number of vectors that need to be evaluated, and reduces the calculation burden of the controller to a certain extent, but also acts on two active vectors and two zero vectors in one control cycle, greatly improving the steady-state performance of the system while achieving fixed switching frequency.

Description

Technical Field

[0001] The invention relates to a multi-vector model predictive control method for a permanent magnet synchronous motor taking vector screening into account, and belongs to the field of motor drive and control. Background Art

[0002] Permanent magnet synchronous motors have attracted extensive attention due to their superior performance such as high power density, large starting torque, and high efficiency. The essence of their control lies in torque control. Model predictive control has been successfully applied to the field of permanent magnet synchronous motor control due to its advantages such as simple control concept, fast dynamic response, and easy implementation of multi-objective control. However, since traditional model predictive control only acts on a single vector in each control cycle, it has defects such as poor steady-state performance and non-fixed switching frequency, which leads to deterioration of system performance and inaccurate torque control. In addition, model predictive control needs to evaluate all vectors that the inverter can provide in each control cycle, which will inevitably bring a huge computational burden to the controller. Therefore, studying a model predictive algorithm that can improve the steady-state performance of the system and reduce the computational burden of the controller has broad application prospects. Summary of the invention

[0003] Technical problem: In view of the above-mentioned existing technologies, a multi-vector model predictive control method of a permanent magnet synchronous motor taking into account vector screening is proposed, which can effectively improve the steady-state performance of the system and reduce the computational burden of the controller to a certain extent.

[0004] Technical solution: 1. A multi-vector model predictive control method for a permanent magnet synchronous motor taking into account vector screening, comprising the following steps:

[0005] Step 1: In each sampling period, sample the electron current i s Calculate the motor speed ω in real time based on the motor electrical angle θ m , use Park transformation (1) to calculate the dq axis current i d with i q ;

[0006] Step 2: Obtain the q-axis current reference value i through the speed PI controller (2) q * , and set the d-axis current reference value i d * is 0 to implement i d =0 control;

[0007] Step 3: Calculate the expected voltage vector u through the expected voltage vector calculation module (3) s * , and input to the vector screening module (4);

[0008] Step 4: According to the expected period voltage vector u s* Perform vector screening to determine two adjacent active voltage vectors u1 and u2, and input them and the zero vector u0 into the error evaluation module (5) for evaluation;

[0009] Step 5: Use the error evaluation function to perform error evaluation on u0, u1 and u2, and obtain the corresponding error evaluation function values ​​ε(u0), ε(u1) and ε(u2);

[0010] Step 6: Using the duty cycle calculation module (6) and the error evaluation function values ​​ε(u0), ε(u1) and ε(u2) obtained in step 5, calculate and screen the optimal duty cycle combination (d0 * ,d1 * ,d2 * ), outputs a corresponding switching signal to drive the inverter (7) to work.

[0011] Furthermore, the voltage vector u described in step 3 s * The specific calculation method is:

[0012] According to the deadbeat control principle, the predicted current value of the permanent magnet synchronous motor shown in equation (1) is equal to the current reference value i d * and i q * Equal, as shown in formula (2),

[0013]

[0014] In the formula, i d k+1 with i q k+1 are the predicted values ​​of d-axis and q-axis currents respectively; i d k with i q k are the d-axis and q-axis current sampling values ​​respectively; L d With L q are the d-axis and q-axis components of the winding inductance respectively; R s is the internal resistance of the winding; T s is the sampling period; ω r is the electrical angular velocity; ψ f is the permanent magnet flux; u d k with u q k They are the d-axis and q-axis components of the acting stator voltage respectively.

[0015]

[0016] According to formula (2), the d-axis and q-axis components of the expected voltage vector can be calculated, as shown in formula (3):

[0017]

[0018] In the formula, u d * with u d * are the d-axis and q-axis components of the desired voltage vector, respectively. Further, the inverse Park transform is used to calculate the α-axis and β-axis components of the desired voltage vector, as shown in equation (4):

[0019]

[0020] In the formula, u α * with u β * are the d-axis and q-axis components of the desired voltage vector, respectively, and written in vector form, we have u s * =u α * +ju β * .

[0021] Furthermore, the specific process of vector screening described in step 4 is as follows:

[0022] Calculate the expected voltage vector u using the inverse tangent function s * The angle δ in the αβ plane,

[0023]

[0024] Determine the desired voltage vector u according to δ s * Sector and determine two adjacent active voltage vectors. Specifically, when 0≤δ<π / 3, the expected voltage vector is located in sector I, at which time u1=V1, u2=V2; when π / 3≤δ<2π / 3, the expected voltage vector is located in sector II, at which time u1=V2, u2=V3; when 2π / 3≤δ<π, the expected voltage vector is located in sector III, at which time u1=V3, u2=V4; when π≤δ<4π / 3, the expected voltage vector is located in sector IV, at which time u1=V4, u2=V5; when 4π / 3≤δ<5π / 3, the expected voltage vector is located in sector V, at which time u1=V5, u2=V6; when 5π / 3≤δ<2π, the expected voltage vector is located in sector VI, at which time u1=V6, u2=V1. In addition, u0 is a zero vector, that is, u0=V0 or V7. The above u0, u1 and u2 are used as vectors to be evaluated.

[0025] Furthermore, the error evaluation method described in step 5 is:

[0026] The error evaluation function shown in formula (6) is used to evaluate the screened u0, u1 and u2.

[0027]

[0028] Furthermore, the optimal duty cycle combination (d0 * ,d1 * ,d2 * ) The specific calculation method is:

[0029] Since the sampling frequency of the system is usually high, it can be approximately considered that the error evaluation function value is linearly related to the corresponding vector duty cycle. Based on this, for the case where three vectors u0, u1 and u2 act in one control cycle, a value function is constructed as shown in formula (7):

[0030]

[0031] Where, d i is the vector u i The duty cycle satisfies the constraint condition shown in formula (8).

[0032]

[0033] In fact, (7) reflects that u i Relative to the desired voltage vector u s * The smaller the root mean square error, the closer the effect of the three vectors u0, u1 and u2 is to u s * The effect of this problem is that the problem is to solve the extreme value of (7) under the constraint condition (8), that is, the conditional extreme value problem. The Lagrange multiplier method can effectively solve this problem and obtain the optimal duty cycle combination (d0 * ,d1 * ,d2 * ) as shown in formula (9);

[0034]

[0035] in

[0036]

[0037] Beneficial effects: 1) Two active voltage vectors and zero vector are applied in one control cycle, which improves the steady-state performance of the system;

[0038] 2) The insertion of zero vector can achieve fixed inverter switching frequency, which is beneficial to the design of filter;

[0039] 3) It retains the advantages of fast dynamic response and high tracking accuracy of the traditional model predictive control algorithm;

[0040] 4) A vector screening method based on deadbeat control is adopted, which greatly reduces the calculation burden of the controller. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a control block diagram of the permanent magnet synchronous motor multi-vector model predictive control method taking into account vector screening of the present invention, in which;

[0042] Figure 2 A space vector distribution diagram of a two-level inverter of a multi-vector model predictive control method for a permanent magnet synchronous motor taking vector screening into account according to the present invention;

[0043] Figure 3 The inverter switching frequency waveform diagram of the permanent magnet synchronous motor multi-vector model predictive control method taking vector screening into account according to the present invention;

[0044] Figure 4 The stator current, torque and speed steady-state simulation waveform diagram of the permanent magnet synchronous motor multi-vector model predictive control method taking into account vector screening of the present invention;

[0045] Figure 5 The simulation waveform diagram of the stator current, torque and speed under the condition of variable load torque of the multi-vector model predictive control method of the permanent magnet synchronous motor taking into account vector screening of the present invention;

[0046] Figure 6 The simulation waveform diagram of the stator current, torque and speed under the given variable speed condition of the multi-vector model predictive control method of the permanent magnet synchronous motor taking into account vector screening of the present invention;

[0047] 1. Park converter; 2. PI controller; 3. Expected voltage vector calculation module; 4. Vector screening module; 5. Error evaluation module; 6. Duty cycle calculation module; 7. Two-level inverter. DETAILED DESCRIPTION

[0048] The present invention will be further described in detail below with reference to the accompanying drawings and by way of examples. The following examples are intended to explain the present invention but the present invention is not limited to the following examples.

[0049] like Figure 1As shown, a control system of a permanent magnet synchronous motor taking into account vector screening is composed of a Park transformation module 1, a speed PI controller 2, a desired voltage vector calculation module 3, a vector screening module 4, an error evaluation module 5, a duty cycle calculation module 6, an inverter 7, and a permanent magnet synchronous motor 8. The inverter 7 is powered by a DC power supply, and its output end is connected to the permanent magnet synchronous motor 8; the Park transformation module 1 collects the current and motor electrical angle information of the permanent magnet synchronous motor 8 through an encoder and a current sensor respectively, and the encoder also transmits the motor electrical angle information to the desired voltage vector module 3; the speed PI controller 2 obtains the real-time speed information of the motor and inputs the q-axis current given value to the desired voltage vector calculation module 3; the desired voltage vector calculation module 3 calculates the desired voltage vector according to the current given value and inputs it to the vector screening module 4, and selects three voltage vectors for evaluation by the error evaluation module 5; the error evaluation module 5 inputs the evaluation result to the duty cycle calculation module 6, calculates the optimal duty cycle of the three voltage vectors, and then controls the inverter 7 to control the permanent magnet synchronous motor 8.

[0050] A multi-vector model predictive control method for a permanent magnet synchronous motor taking into account vector screening comprises the following steps:

[0051] Step 1: In each sampling period, sample the electron current i s Calculate the motor speed ω in real time based on the motor electrical angle θ m , use Park transformation to calculate the dq axis current i d with i q .

[0052] Step 2: Obtain the q-axis current reference value i through the speed PI controller q * , and set the d-axis current reference value i d * is 0 to implement i d =0 control.

[0053] Step 3: Calculate the expected voltage vector u through the expected voltage vector calculation module s * , and input to the vector screening module (4). Specifically, according to the deadbeat control principle, let the permanent magnet synchronous motor current prediction value and the current reference value i shown in formula (1) be d * and i q * Equal, as shown in formula (2),

[0054]

[0055] In the formula, i d k+1 with iq k+1 are the predicted values ​​of d-axis and q-axis currents respectively; i d k with i q k are the d-axis and q-axis current sampling values ​​respectively; L d With L q are the d-axis and q-axis components of the winding inductance respectively; R s is the internal resistance of the winding; T s is the sampling period; ω r is the electrical angular velocity; ψ f is the permanent magnet flux; u d k with u q k are the d-axis and q-axis components of the acting stator voltage respectively.

[0056]

[0057] According to formula (2), the d-axis and q-axis components of the expected voltage vector can be calculated, as shown in formula (3):

[0058]

[0059] In the formula, u d * with u d * are the d-axis and q-axis components of the desired voltage vector, respectively. Further, the inverse Park transform is used to calculate the α-axis and β-axis components of the desired voltage vector, as shown in equation (4):

[0060]

[0061] In the formula, u α * with u β * are the d-axis and q-axis components of the desired voltage vector, respectively, and written in vector form, we have u s * =u α * +ju β * .

[0062] Step 4: According to the expected period voltage vector u s * Perform vector screening. Use the inverse tangent function to calculate the expected voltage vector u s * The angle δ in the αβ plane,

[0063]

[0064] Determine the desired voltage vector u according to δ s * sectors and determine two adjacent active voltage vectors. Specifically, Figure 2 As shown, when 0≤δ<π / 3, the expected voltage vector is located in sector I, at which u1=V1, u2=V2; when π / 3≤δ<2π / 3, the expected voltage vector is located in sector II, at which u1=V2, u2=V3; when 2π / 3≤δ<π, the expected voltage vector is located in sector III, at which u1=V3, u2=V4; when π≤δ<4π / 3, the expected voltage vector is located in sector IV, at which u1=V4, u2=V5; when 4π / 3≤δ<5π / 3, the expected voltage vector is located in sector V, at which u1=V5, u2=V6; when 5π / 3≤δ<2π, the expected voltage vector is located in sector VI, at which u1=V6, u2=V1. In addition, u0 is a zero vector, that is, u0=V0 or V7. The above u0, u1 and u2 are used as vectors to be evaluated and input into the error evaluation module (5) for evaluation.

[0065] Step 5: Use the error evaluation function shown in formula (6) to perform error evaluation on u0, u1 and u2, and obtain the corresponding error evaluation function values ​​ε(u0), ε(u1) and ε(u2).

[0066]

[0067] Step 6: Use the duty cycle calculation module and the error evaluation function values ​​ε(u0), ε(u1) and ε(u2) obtained in step 5 to calculate and screen the optimal duty cycle combination (d0 * ,d1 * ,d2 * ), outputs the corresponding switching signal to drive the inverter (7) to work. Specifically, since the sampling frequency of the system is usually high, it can be approximately considered that the error evaluation function value is linearly related to the corresponding vector duty cycle. Based on this, for the case where three vectors u0, u1 and u2 act in one control cycle, a value function as shown in formula (7) is constructed:

[0068]

[0069] Where, d i is the vector u i The duty cycle satisfies the constraint condition shown in formula (8).

[0070]

[0071] In fact, (7) reflects that u i Relative to the desired voltage vector u s *The smaller the root mean square error, the closer the effect of the three vectors u0, u1 and u2 is to u s * The effect of this problem is that the problem is to solve the extreme value of (7) under the constraint condition (8), that is, the conditional extreme value problem. The Lagrange multiplier method can effectively solve this problem and obtain the optimal duty cycle combination (d0 * ,d1 * ,d2 * ) as shown in formula (9).

[0072]

[0073] in

[0074]

[0075] Under the conditions of a load torque of 10 N∑m and a given speed of 300 r / min, the multi-vector model predictive control method for a permanent magnet synchronous motor taking into account vector screening disclosed in the present invention is implemented, and the inverter switching frequency waveform is as follows: Figure 3 As shown in the figure, it is obvious that the switching frequency is stable at 10kHz, achieving a fixed switching frequency. The stator current, electromagnetic torque and speed waveforms are shown in Figure 4 As shown, it can be seen that the stator current has high sinusoidality, the electromagnetic torque pulsation is small and the speed is stable, and the system has good steady-state performance.

[0076] The load torque is suddenly changed from 10N·m to 15N·m in 0.2s, and the speed setting is maintained at 300r / min. The multi-vector model predictive control method of the permanent magnet synchronous motor taking into account vector screening disclosed in the present invention is implemented. The stator current, electromagnetic torque and speed waveforms are as follows: Figure 5 As shown, it can be seen that the system has a rapid dynamic response, the stator current still maintains a high sinusoidal degree after the torque mutation, the electromagnetic torque tracks quickly and the speed quickly reaches 300r / min after the load mutation.

[0077] The speed setting is suddenly changed from 300r / min to 500r / min at 0.4s, and the load torque is maintained at 15N·m. The multi-vector model predictive control method of the permanent magnet synchronous motor taking into account vector screening disclosed in the present invention is implemented. The stator current, electromagnetic torque and speed waveforms are as follows: Figure 6 As shown in the figure, it can be seen that after the speed is suddenly changed, the stator current quickly reaches the saturation value of 20A, and the electromagnetic torque is kept at the maximum value, so that the speed shows a linear upward trend. After the speed reaches 500r / min, the electromagnetic torque drops rapidly to 15N·m, and the system continues to operate under the new steady-state condition.

[0078] The above contents described in this specification are merely examples of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, as long as they do not deviate from the contents of the present specification or exceed the scope defined in the claims, they shall all fall within the protection scope of the present invention.

Claims

1. A multi-vector model predictive control method for a permanent magnet synchronous motor taking into account vector screening, characterized in that: The steps include: Step 1: In each sampling period, sample the electron current i s Calculate the motor speed ω in real time based on the motor electrical angle θ m , using Park transformation (1) to calculate the dq axis current i d with i q ; Step 2: Obtain the q-axis current reference value i through the speed PI controller (2) q * , and set the d-axis current reference value i d * is 0 to implement i d =0 control; Step 3: Calculate the expected voltage vector u through the expected voltage vector calculation module (3) s * , and input to the vector screening module (4); Among them, the voltage vector u in step 3 is s * The specific calculation method is: According to the deadbeat control principle, the predicted current value of the permanent magnet synchronous motor shown in equation (1) is equal to the current reference value i d * and i q * Equal, as shown in formula (2), In the formula, i d k+1 with i q k+1 are the predicted values ​​of d-axis and q-axis currents respectively; i d k with i q k are the d-axis and q-axis current sampling values ​​respectively; L d With L q are the d-axis and q-axis components of the winding inductance respectively; R s is the internal resistance of the winding; T s is the sampling period; ω r is the electrical angular velocity; ψ f is the permanent magnet flux; u d k with u q k are the d-axis and q-axis components of the acting stator voltage respectively; According to formula (2), the d-axis and q-axis components of the expected voltage vector can be calculated, as shown in formula (3): In the formula, u d * with u d * are the d-axis and q-axis components of the desired voltage vector, respectively; further, the α-axis and β-axis components of the desired voltage vector are calculated using the inverse Park transform, as shown in formula (4): In the formula, u α * with u β * are the d-axis and q-axis components of the desired voltage vector, respectively, and written in vector form, we have u s * =u α * +ju β * Step 4: According to the expected period voltage vector u s * Perform vector screening to determine two adjacent active voltage vectors u1 and u2, and input them and the zero vector u0 into the error evaluation module (5) for evaluation; The specific process of vector screening described in step 4 is as follows: Calculate the expected voltage vector u using the inverse tangent function s * The angle δ in the αβ plane, Determine the desired voltage vector u according to δ s * Sector and determine two adjacent active voltage vectors; specifically, when 0≤δ<π / 3, the expected voltage vector is located in sector I, at which time u1=V1, u2=V2; when π / 3≤δ<2π / 3, the expected voltage vector is located in sector II, at which time u1=V2, u2=V3; when 2π / 3≤δ<π, the expected voltage vector is located in sector III, at which time u1=V3, u2=V4; when π≤δ<4π / 3, the expected voltage vector is located in sector IV, at which time u1=V4, u2=V5; when 4π / 3≤δ<5π / 3, the expected voltage vector is located in sector V, at which time u1=V5, u2=V6; when 5π / 3≤δ<2π, the expected voltage vector is located in sector VI, at which time u1=V6, u2=V1; in addition, u0 is a zero vector, that is, u0=V0 or V7; the above u0, u1 and u2 are used as vectors to be evaluated; Step 5: Use the error evaluation function to perform error evaluation on u0, u1 and u2, and obtain the corresponding error evaluation function values ​​ε(u0), ε(u1) and ε(u2); The error evaluation method described in step 5 is: The error evaluation function shown in formula (6) is used to evaluate the screened u0, u1 and u2. Step 6: Using the duty cycle calculation module (6) and the error evaluation function values ​​ε(u0), ε(u1) and ε(u2) obtained in step 5, calculate and screen the optimal duty cycle combination (d0 * ,d1 * ,d2 * ), outputs a corresponding switching signal to drive the inverter (7) to work; Among them, the optimal duty cycle combination (d0 * ,d1 * ,d2 * ) The specific calculation method is: Since the sampling frequency of the system is usually high, it can be approximately considered that the error evaluation function value is linearly related to the corresponding vector duty cycle; based on this, for the case where three vectors u0, u1 and u2 act in one control cycle, a value function is constructed as shown in formula (7): Where, d i is the vector u i The duty cycle satisfies the constraint conditions shown in formula (8); In fact, formula (7) reflects u i Relative to the desired voltage vector u s * The smaller the root mean square error, the closer the effect of the three vectors u0, u1 and u2 is to u s * The effect of this problem is converted into solving the extreme value of (7) under the constraint condition (8), that is, the conditional extreme value problem. The Lagrange multiplier method can effectively solve this problem and obtain the optimal duty cycle combination (d0 * ,d1 * ,d2 * ) as shown in formula (9); in

Images