A Fast Dual-Vector Model-Free Predictive Control Method for Permanent Magnet Synchronous Motors

Abstract

The present invention discloses a fast dual-vector model-free predictive control method for a permanent magnet synchronous motor. The method is applied to the control system of a permanent magnet synchronous motor. The state equation of the permanent magnet synchronous motor is obtained by introducing the model-free control concept. At the same time, in order to simplify the tuning, the unknown term F in the equation is designed based on an extended state observer. The fast vector selection method is then combined with the dual-vector predictive control method to form a fast dual-vector model-free predictive control, which achieves algorithm simplification while improving system performance. The method of the present invention has a simple structure, a small amount of calculation, and is easy to implement. When applied to a PMSM control system, it has the technical advantages of fast system dynamic response and strong robustness, while also having the advantages of reducing motor torque pulsation and reducing algorithm time consumption.

Description

Technical Field

[0001] The present invention relates to a permanent magnet synchronous motor control method, in particular to a permanent magnet synchronous motor fast dual-vector model-free current prediction control method. Background Art

[0002] Permanent-Magnet Synchronous Motor (PMSM) has the advantages of small size, high efficiency, and strong overload capacity. It has received extensive attention and research in the fields of small and medium power, high precision, wide speed regulation range servo control systems and high-performance professional drives.

[0003] Currently, the commonly used algorithms for PMSM drive systems are vector control and direct torque control. Vector control decomposes the stator current into the excitation current component and the torque component, controlling them separately. This provides good steady-state operation, but its dynamic performance is easily affected by the current inner loop. Direct torque control treats the motor and inverter as a single entity, directly controlling the motor's electromagnetic torque and flux linkage. This avoids the complex coordinate transformation and current loop required by vector control, resulting in better robustness and response speed, but often exhibiting significant torque ripple. In recent years, model predictive control (MPC) has gained increasing attention in the fields of power electronics and motor drives due to its significant advantages in handling complex constrained optimization of nonlinear systems. MPC was developed in the 1960s, and with the rapid development of microcontrollers, its application in power electronics has become a research hotspot in recent years. MPC technology, applied to permanent magnet synchronous motor control systems, offers high-speed dynamic response and excellent steady-state performance. Research on MPC in PMSM control systems holds significant application value.

[0004] However, traditional MPC relies heavily on the accuracy of machine parameters. These parameters can vary depending on the operating point and environment. For example, temperature fluctuations can affect the motor's stator resistance and inductance. Failure to account for these parameter variations can significantly degrade MPC performance, potentially leading to increased noise during motor operation and steady-state current errors. Summary of the Invention

[0005] To address the need for improvements in the aforementioned prior art, the present invention provides a fast dual-vector model-free current predictive control method for PMSMs. This control method utilizes only the system inputs and outputs, disregarding motor parameters. It simultaneously operates two voltage vectors within a single control cycle and employs a fast vector selection method to position the reference voltage vector. Compared to traditional control methods, this method offers improved robustness against parameter variations while maintaining both steady-state control accuracy and algorithmic speed.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is: a PMSM fast dual-vector model-free current prediction control method, specifically comprising:

[0007] Step 1: Based on the input and output of the PMSM system, a current hyperlocal model of the PMSM drive system is established, and then a model-free current prediction equation of the PMSM system is designed based on the hyperlocal model.

[0008] Step 2: Combine the fast vector selection method with the dual vector predictive control method, and derive the reference voltage vector U through the model-free prediction equation described in step 1. ref Then, the sector is located by the fast vector selection method, the first voltage vector Vopt1 is selected by the value function, the double vector combination of the first voltage vector Vopt1 and the seven effective voltage vectors and the synthetic vector combination under the corresponding action time are calculated, the second voltage vector Vopt2 is selected, and finally the first voltage vector V opt1 and the second voltage vector V opt2 The corresponding action times t1 and t2 are output to the inverter controlled by the pulse generator to complete the prediction.

[0009] Furthermore, the step 1 specifically includes: establishing a current hyperlocal model of the PMSM drive system based on the input and output of the PMSM system, wherein the unknown term F is designed based on the extended state observer, and then using the first-order Euler discretization method to establish a model-free current prediction equation of the PMSM.

[0010] Furthermore, the step 2 specifically includes:

[0011] a. Combine the fast vector selection method with the dual vector predictive control method and calculate the reference voltage vector U by formula (4) ref ;

[0012] b. Assume that the reference voltage vector U ref Component u on the α and β axes α and u β , define the variable u ref1 、u ref2 、u ref3 , the reference voltage vector U ref Divide into 6 sectors, each sector contains a center vector and two zero vectors, define four variables M1, A1, B1, C1, and the calibration rules are:

[0013] ①If u ref1 >0, then A1=1, otherwise A1=0;

[0014] ②If u ref2>0, then B1=1, otherwise B1=0;

[0015] ③If u ref3 >0, then C1=1, otherwise C1=0.

[0016] Let the middle value M1 = A1 + 2B1 + 4C1. According to the division of the sector number S, the corresponding relationship between M1 and the quadrilateral large sector number S is obtained to quickly locate the sector;

[0017] c. According to the value function (6), select the first voltage vector V opt1 .

[0018] d. The first voltage vector V opt1 The 7 effective voltage vectors are combined respectively, and the q-axis current is calculated in a deadbeat manner. The action time of the two voltage vectors is calculated using formula (10). Formulas (7) and (8) are substituted into the prediction equation (2). The second voltage vector V is selected through the value function formula (6). opt2 .

[0019] e. The first voltage vector V opt1 and the second voltage vector V opt2 The corresponding action times t1 and t2 are output to the inverter controlled by the pulse generator to complete the model-free fast dual-vector prediction of the permanent magnet synchronous motor.

[0020] Compared with the existing technology, the above technical solution conceived by the present invention has the following advantages over the existing technology:

[0021] (1) Traditional predictive control methods rely heavily on the accuracy of machine parameters. When these parameters are inaccurate, problems such as increased noise during motor operation and steady-state errors in current may occur. The model-free predictive control method for permanent magnet synchronous motors proposed in this patent uses only the system input and output, without considering motor parameters, and thus has better anti-interference performance.

[0022] (2) This method combines fast vector selection and dual vector predictive control, which improves the steady-state control accuracy of the system while optimizing the algorithm time consumption and improving the control performance of the system.

[0023] (3) This method has a simple structure, small computational complexity, and is easy to implement. When used in a PMSM control system, the motor has a fast dynamic response speed, small stator current ripple and distortion, low switching frequency, and excellent dynamic and steady-state performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a structural diagram of a fast dual-vector model-free predictive control method for a permanent magnet synchronous motor according to the present invention;

[0025] Figure 2 It is the basic voltage vector distribution diagram of the two-level inverter output;

[0026] Figure 3 It is a quick vector selection schematic;

[0027] Figure 4 It is a control flow chart of the present invention;

[0028] Figure 5 This is a schematic diagram of the main circuit topology of the present invention;

[0029] Figure 6 Schematic diagram of the speed change curve of the present invention;

[0030] Figure 7 It is a schematic diagram of the electromagnetic torque change curve of the present invention. DETAILED DESCRIPTION

[0031] In order to make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent based on the claims and the following specific embodiments. It should be noted that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0032] Figure 1 This is a schematic diagram of a fast dual-vector model-free predictive control method for a permanent magnet synchronous motor disclosed in an embodiment of the present invention, using i d =0 control mode, the speed outer loop uses the PI regulator, and the current inner loop adopts the proposed fast dual-vector model-free predictive control method. The specific strategy is as follows:

[0033] First, based on the input and output of the PMSM system, a super-local model of the PMSM system is established, which is expressed as

[0034]

[0035] Where: α d , α q Respectively represent the voltage coefficients of the PMSM stator d and q axes; F d 、F q The unknown term F includes the known parts of the system, unmodeled dynamics, and unknown parts such as parameter uncertainty.

[0036] The model-free current prediction equation of the PMSM system can be obtained by discretizing the above equation using the first-order Euler discretization method:

[0037]

[0038] Where: Respectively represent F d ,F q The estimated value at time k, u d * (k),u q * (k) represents the stator d and q axis voltage reference value, which is the switching state S output by the inverter at the current k moment. α (k),S b (k),S c (k) and 3 / 2 coordinate transformation.

[0039] and The calculation formula is designed based on the extended state observer, and the design equation is as follows:

[0040]

[0041] Where: Represents i s and the estimated value of F.

[0042] Figure 2 This is the basic voltage vector distribution of the two-level inverter output. From the figure, we can see that there are 8 voltage vectors to be selected. The sector division of fast vector selection is as follows: Figure 3 As shown in the figure, a fast vector selection method is used to locate the reference voltage vector to reduce the complexity and calculation amount of the control algorithm. The fast vector selection method is as follows:

[0043] The stator current prediction formula obtained based on the super-local model of PMSM is given by formula (2). If we want to achieve the stator current command at the next sampling moment, we need to set i s (k+1)=i s * Substituting into equation (2), the optimal voltage vector u output at the current moment is s (k) will become u s * (k).

[0044]

[0045] Assume that the reference voltage vector U ref The components on the α and β axes are u α and u β , define u ref1 、u ref2 、u ref3 The three variables are:

[0046]

[0047] like Figure 3 As shown, the reference voltage vector U ref It is divided into 6 sectors, each sector contains a center vector and two zero vectors.

[0048] Define four variables: M1, A1, B1, and C1. The calibration rules are as follows:

[0049] ④If u ref1 >0, then A1=1, otherwise A1=0;

[0050] ⑤If u ref2 >0, then B1=1, otherwise B1=0;

[0051] ⑥If u ref3 >0, then C1=1, otherwise C1=0.

[0052] Let the middle value M1=A1+2B1+4C1, according to Figure 3 The division of the medium sector number S, according to the above calibration rules, can be obtained M1 and the corresponding relationship between the quadrilateral large sector number S

[0053] <![CDATA[Median value M1]]> 1 3 2 6 4 5 Large sector number S 1 2 3 4 5 6

[0054] Finally, the dual vector predictive control method is used to improve the steady-state accuracy of the controlled system. According to the previous fast vector selection strategy, when positioning U ref After the sector is located, there are only the center vector and the zero vector in the sector. When selecting the vector, the calculation amount can be greatly simplified. The value function is defined as follows

[0055]

[0056] When selecting the first optimal voltage vector, the cost function corresponding to the center voltage vector and the zero vector is compared to select the first optimal voltage vector, which is determined as the first voltage vector V opt1 The idea of dual vector is to select the first voltage vector V opt1 Based on this, another vector selection is performed to determine the second voltage vector V opt2 , in the second voltage vector V opt2 When selecting, the first voltage vector V opt1 and 7 effective voltage vectors are combined respectively, and the action time of the two voltage vectors in each combination is pre-assigned, so that u in equation (2) is predicted d and u q It can be written as follows

[0057] u q =u qopt1+(T s -t opt1 )u qj (7)

[0058] u d =u dopt1 t opt1 +(T s -t opt1 )u dj (8)

[0059] Where, t opt1 Refers to the first voltage vector V opt1 The action time of u dopt1 and u qopt1 They are the first voltage vector V opt1 The corresponding stator voltage DC and quadrature axis voltage components; u dj and u qj The j-th voltage vector V j In this way, seven groups of voltage vectors and action time combinations are obtained. Finally, the second optimal voltage vector is selected through the value function formula (6), which is determined as the second voltage vector V opt2 , and finally complete the prediction.

[0060] The action time of the two vectors is calculated by the q-axis current deadbeat method, that is, in one sampling period, the action time of the two vectors is distributed so that i q At time k+1, it reaches the given value, that is,

[0061]

[0062] According to the above formula, the action time of the first and second voltage vectors can be obtained

[0063]

[0064] Where k1 and k2 are the first and second voltage vectors selected respectively. q According to the relationship between the stator current and the quadrature and direct axes in the dq coordinate system, k1 and k2 can be expressed as:

[0065]

[0066]

[0067] like Figure 4 : A flow chart of a method for fast dual-vector model-free predictive control of a permanent magnet synchronous motor is shown, and the method comprises the following steps:

[0068] 1.1. Based on the input and output of the PMSM system, the current superlocal model of the PMSM drive system (Equation (1)) is established, where the unknown term F is designed based on the extended state observer;

[0069] 1.2. The model-free current prediction equation (2) of PMSM is established using the first-order Euler discretization method.

[0070] 2.1. Combine the fast vector selection method with the dual vector predictive control method and calculate the reference voltage vector U by formula (4): ref ;

[0071] 2.2, Assume that the reference voltage vector U ref Component u on the α and β axes α and u β , define the variable u ref1 、u ref2 、u ref3 , the reference voltage vector U ref Divide into 6 sectors, each sector contains a center vector and two zero vectors, define four variables M1, A1, B1, C1, and the calibration rules are:

[0072] ⑦If u ref1 >0, then A1=1, otherwise A1=0;

[0073] ⑧If u ref2 >0, then B1=1, otherwise B1=0;

[0074] ⑨If u ref3 >0, then C1=1, otherwise C1=0.

[0075] Let the middle value M1 = A1 + 2B1 + 4C1. According to the division of the sector number S, the corresponding relationship between M1 and the quadrilateral large sector number S is obtained to quickly locate the sector;

[0076] 2.3. According to the value function (6), select the first voltage vector V opt1 .

[0077] 2.4. The first voltage vector V opt1 The 7 effective voltage vectors are combined respectively, and the q-axis current is calculated in a deadbeat manner. The action time of the two voltage vectors is calculated using formula (10). Formulas (7) and (8) are substituted into the prediction equation (2). The second voltage vector V is selected through the value function formula (6). opt2 .

[0078] 2.5. The first voltage vector V opt1 and the second voltage vector V opt2The corresponding action times t1 and t2 are output to the inverter controlled by the pulse generator to complete the model-free fast dual-vector prediction of the permanent magnet synchronous motor.

[0079] Figure 5 This is the main circuit topology of the fast dual-vector model-free predictive control method for a permanent magnet synchronous motor according to the present invention. As can be seen from the figure, the entire main circuit has six switching transistors, which can generate eight sets of basic voltage vectors, of which seven sets are effective voltage vectors.

[0080] Figure 6 This is a schematic diagram of the speed change curve of the present invention. It can be seen from the effect diagram that when the motor rises from 0 to the reference speed, the response speed is relatively fast. When the load torque is suddenly added at t=0.3s, the motor quickly recovers to the given reference speed, and has good dynamic performance.

[0081] Figure 7 Schematic diagram of the electromagnetic torque change curve of the present invention. It can be seen from the schematic diagram that the electromagnetic torque has low fluctuation in steady state and good steady-state performance.

[0082] It will be easily understood by those skilled in the art that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A fast dual-vector model-free current predictive control method for a permanent magnet synchronous motor, characterized by: The following steps are involved: Step 1: Based on the input and output of the PMSM system, a current hyperlocal model of the PMSM drive system is established, and then a model-free current prediction equation of the PMSM system is designed based on the hyperlocal model; Step 2: Combine the fast vector selection method with the dual vector prediction control method, and derive the reference voltage vector U through the model-free current prediction equation described in step 1. ref Then, the sector is located by the fast vector selection method, and the first voltage vector V is selected by the value function. opt1 , calculate the first voltage vector V opt1 Combined with the double vector combination of 7 effective voltage vectors and the synthetic vector combination under the corresponding action time, the second voltage vector V opt2 , and finally the first voltage vector V opt1 and the second voltage vector V opt2 The corresponding action times t1 and t2 are output to the inverter controlled by the pulse generator to complete the prediction.

2. The method for fast dual-vector model-free current prediction control of a permanent magnet synchronous motor according to claim 1, characterized in that: The step 1 specifically includes: According to the input and output of the PMSM system, a current hyperlocal model of the PMSM drive system is established: ； Where: α d , α q Respectively represent the voltage coefficients of the PMSM stator d and q axes; F d 、F q It is the unknown term that includes the known parts of the system, the unmodeled dynamics, and the unknown parts such as parameter uncertainty; The unknown term F is designed based on the extended state observer, and then the model-free current prediction equation of the PMSM is established using the first-order Euler discretization method: ； Where: , Respectively represent F d ,F q The estimated value at time k, u d * (k),u q * (k) represents the stator d and q axis voltage reference value, which is the switching state S output by the inverter at the current k moment. α (k), S b (k), S c (k) and 3 / 2 coordinate transformation.

3. The method for fast dual-vector model-free current prediction control of a permanent magnet synchronous motor according to claim 1, characterized in that: The second step specifically includes: a. Combine the fast vector selection method with the dual vector predictive control method, and use the formula Calculate the reference voltage vector U ref ; b. Assume that the reference voltage vector U ref Component u on the α and β axes α and u β , define the variable u ref1 、u ref2 、u ref3 , the reference voltage vector U ref Divide into 6 sectors, each sector contains a center vector and two zero vectors, define four variables M1, A1, B1, C1, and the calibration rules are: ①If u ref1 >0, then A1=1, otherwise A1=0; ②If u ref2 >0, then B1=1, otherwise B1=0; ③If u ref3 >0, then C1=1, otherwise C1=0; Let the middle value M1 = A1 + 2B1 + 4C1. According to the division of the sector number S, the corresponding relationship between M1 and the quadrilateral large sector number S is obtained to quickly locate the sector; c. Select the first voltage vector V according to the value function opt1 , the value function expression: ； d. The first voltage vector V opt1 Combined with the seven effective voltage vectors, the action time of the two voltage vectors is calculated in a deadbeat manner using the q-axis current. The calculation formula is as follows: ； Where k1 and k2 are the values of the first and second voltage vectors selected, respectively. q The slope of will u d ,u q Substitute into the model-free current prediction equation and select the second voltage vector V through the cost function opt2 ;in ； Where, t opt1 Refers to the first voltage vector V opt1 The action time of u dopt1 and u qopt1 They are the first voltage vector V opt1 The corresponding stator voltage direct and quadrature axis voltage components; u dj and u qj The j-th voltage vector V j The corresponding direct and quadrature axis components; e. The first voltage vector V opt1 and the second voltage vector V opt2 The corresponding action times t1 and t2 are output to the inverter controlled by the pulse generator to complete the model-free fast dual-vector prediction of the permanent magnet synchronous motor.

Images