A dead-time-considered double-vector model predictive current control method for permanent magnet synchronous motor
By introducing a current prediction model with dead zone effect and a dual-vector screening strategy into permanent magnet synchronous motors, combined with deadbeat control, the problems of computational complexity and current distortion in traditional control strategies are solved, and efficient current control is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGDONG OCEAN UNIVERSITY
- Filing Date
- 2026-06-03
- Publication Date
- 2026-07-21
Smart Images
Figure CN122437452A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of motor drive control technology, and in particular to a dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead zone. Background Technology
[0002] As an advanced propulsion method in the field of marine engineering, marine electric propulsion technology has become a global research hotspot. The control performance of the propulsion motor directly determines the efficiency, stability, and quietness of the entire ship's power system, making it a key technology in this field. Ship propulsion mainly uses induction motors and surface-mounted permanent magnet synchronous motors (PMSMs), with PMSMs being particularly favored due to their high power density and high efficiency.
[0003] Permanent magnet synchronous motor (PMSM) drive systems are widely used in marine electric propulsion systems, and their control performance directly determines the reliability, efficiency, and quietness of the propulsion system. Model predictive current control (FCS-MPCC) has become mainstream due to its fast response and strong constraint handling capabilities. However, existing technologies still have significant limitations: (1) Although the traditional single-vector MPC has a simple structure, it only acts on one vector in a single cycle, resulting in large steady-state ripple and the switching frequency is not fixed, which easily generates noise.
[0004] (2) Multi-vector MPC improves ripple by combining vectors, but usually requires multiple traversals or complex optimizations, resulting in a heavy computational burden and making it difficult to meet the requirements of high-frequency control.
[0005] (3) Dead zone effect is ignored: Most solutions ignore or simply deal with the inverter dead zone without considering the nonlinear voltage distortion caused by current polarity.
[0006] (4) Although deadbeat control has a very fast response, it is sensitive to parameters and it is difficult to directly integrate nonlinear factors such as dead zone in analytical calculation.
[0007] In summary, existing technologies struggle to simultaneously achieve low computational complexity, high control accuracy, and dead-zone compensation. There is an urgent need for an optimization strategy that can explicitly compensate for dead-zone voltage and rapidly solve for the dual-vector action time using analytical methods, in order to improve the overall performance of ship PMSMs. Summary of the Invention
[0008] The purpose of this invention is to provide a dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead zone, thereby solving the problems mentioned in the background art.
[0009] To achieve the above objectives, this invention provides a dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead time, comprising the following steps: S1. Establish and sample the discrete prediction model of PMSM: Establish a discrete mathematical model of PMSM, combine motor parameters with the current feedback state, derive the current prediction equation, and use it to calculate the predicted current and current change slope under the action of each base voltage vector. S2. Classification strategy based on cost function and torque current: In each control cycle, based on the PMSM discrete mathematical model, all basic voltage vectors are traversed, and the predicted current of the d-axis and q-axis at the next moment is calculated through the current prediction equation. According to the direction of the q-axis prediction error, the voltage vectors are divided into overshoot vector group and undershoot vector group. S3. Dual-vector screening strategy based on classification strategy: Based on the overshoot vector group and undershoot vector group divided by S2, calculate the cost function under the action of each vector, and screen out the overshoot vector and undershoot vector with the smallest cost function in the group as the optimal dual-vector candidate combination. S4. Dead Zone Determination and Deadbeat Time Synthesis: Based on the deadbeat control principle, and combining the corrected dead zone slope and the optimal vector slope, the precise action time of the two optimal vectors within the control cycle is calculated through the control equation. Under the premise of explicitly reducing the current distortion caused by the dead zone effect, the final switching control pulse is output.
[0010] Therefore, the present invention employs the above-mentioned method for predictive current control of permanent magnet synchronous motors using a dual-vector model that considers dead time, and has the following beneficial effects: (1) Reduce current ripple and improve steady-state control accuracy: This invention uses a dual-vector screening strategy based on overshoot / undershoot grouping to ensure that the two voltage vectors involved in the synthesis are opposite in the torque current adjustment direction, and selects the optimal vector combination under the constraint of cost function, thereby effectively suppressing current fluctuations, reducing steady-state current and torque ripple, and improving the steady-state operation accuracy of the motor.
[0011] (2) Suppress dead zone distortion and improve current waveform quality: This invention establishes an inverter dead zone modeling method based on stator current polarity, which explicitly introduces the influence of dead zone effect on current change into the prediction and control process, effectively reducing distortion and low-order harmonics near the current zero crossing, and significantly improving the waveform quality of stator current.
[0012] (3) Achieving fast dynamic response and deadbeat current tracking: Under the premise of considering the influence of dead zone effect on the effective action time of voltage vector, this invention combines the principle of dual vector synthesis and deadbeat control to perform analytical calculation on the action time of voltage vector, so that the torque current can accurately track its reference value within a single control cycle, thus realizing fast dynamic response and deadbeat tracking performance of current.
[0013] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0014] Figure 1 This is a flowchart illustrating an embodiment of a dual-vector model predictive current control method for permanent magnet synchronous motors considering dead zones according to the present invention. Figure 2 This is a control block diagram of an embodiment of a dual-vector model predictive current control method for permanent magnet synchronous motors considering dead zone according to the present invention; Figure 3 This invention provides a dead-time-free implementation of a dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead time. q Schematic diagram of shaft current synthesis; Figure 4 This is a diagram illustrating the clamping effect in an embodiment of the dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead zones, according to the present invention. Detailed Implementation
[0015] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0016] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0017] Example Please see Figures 1-4This invention provides a dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead time, addressing the technical problems of traditional technologies: (1) Current waveform distortion and model mismatch caused by dead time effect: Traditional model predictive control often ignores the inverter dead time or compensates for it as an error part, and does not establish an accurate mathematical model that includes the dynamic change of dead time voltage with current polarity. This will lead to a mismatch between the effective voltage action time in the prediction equation and the actual physical system, causing problems such as stator current waveform distortion, affecting the smooth operation and quiet performance of ship propulsion motors; (2) Excessive steady-state ripple of single-vector control: Traditional finite control set model predictive control (FCS-MPC) applies only a single voltage vector in one control cycle, resulting in large steady-state ripple of current and torque, which is difficult to meet the stringent requirements of modern ship electric propulsion systems for high precision and low vibration control; (3) Computational complexity and optimization burden of traditional dual-vector control: Existing dual-vector or multi-vector model predictive control usually relies on cost functions to traverse and optimize a large number of vector combinations and duty cycles, or uses complex iterative algorithms to solve. This method involves a huge amount of computation and it is difficult to achieve high switching frequency control with limited controller computing power; (4) It is difficult to balance dynamic response and steady-state accuracy: Although the traditional deadbeat control has a very fast dynamic response, it is sensitive to parameters and has difficulty in handling nonlinear factors such as dead zones; while the traditional MPC has good robustness, but has a large steady-state error.
[0018] Based on this, the present invention proposes a dual-vector model predictive current control method for permanent magnet synchronous motors considering dead zone. The method mainly consists of the following three parts, which are briefly described as follows: (1) Construction of current prediction model based on dead zone effect: Establish a discretized mathematical model of PMSM, introduce a dead zone voltage determination mechanism based on current polarity, and calculate the actual effective voltage vector within the dead zone time in real time. From this, an accurate current prediction equation containing the "dead zone current slope" correction term is derived to reduce the influence of inverter nonlinearity on prediction accuracy. (2) Dual-vector screening strategy based on cost function classification: Traverse the basic voltage vector, first divide the vector into "overshoot vector group" and "undershoot vector group" according to the increase or decrease requirements of torque current (q axis); then calculate the comprehensive cost function containing d-axis and q-axis error weights within each group, and screen out the globally optimal overshoot vector and undershoot vector in parallel. (3) Time calculation and synthesis considering the dead-time zero-beat principle: Based on the dead-time zero-beat control principle, the corrected dead-time slope and the effective slope of the selected optimal vector are substituted into the control equation. Under the premise of explicitly deducting the influence of the dead-time voltage, the precise action time of the two optimal vectors within the control cycle is directly calculated analytically, and the final switching control pulse is output. This method specifically includes the following steps: S1. Establish the PMSM discrete prediction model and sampling.
[0019] A discretized mathematical model of a permanent magnet synchronous motor in a synchronous rotating coordinate system is established, utilizing motor parameters (stator resistance, inductance, flux linkage) and the current input feedback state. The current prediction equation is derived and used to calculate the predicted current and current change slope under the action of each base voltage vector. The specific steps are as follows: The flux linkage equation, voltage equation, and torque equation of the two-phase rotating coordinate system dq axis of a three-phase permanent magnet synchronous motor are shown below: ; Will and The solution is: ; Combining the two equations into matrix form, we obtain the system dynamics representation of a three-phase PMSM considering parameter mismatch and model uncertainty in a synchronously rotating coordinate system, as follows: ; in: ; in, for The rate of change of shaft stator current with respect to time for The rate of change of shaft stator current with respect to time The dq-axis stator current vector. These represent the continuous-time state matrix, input matrix, and additional terms matrix, respectively. for shaft and Shaft stator input voltage, and They are respectively shaft and Shaft stator current, and They are respectively shaft and Stator input voltage on the shaft, ω is the electric angular velocity of the rotor. and They represent shaft and The nominal stator inductance on the shaft, It is a permanent magnet flux chain. It is the stator resistor.
[0020] The system dynamics representation is discretized using a first-order forward Euler discretization, with a sampling period of . ,get dq shaft current kThe prediction equation at time +1 is as follows: ; in: ; Using a discrete model of the motor, a one-step lead prediction of the current is performed to obtain the final current prediction equation: ; in: ; in, for The predicted current vector at time t, The discrete state matrix, For discrete input matrices, for At time, the candidate voltage vector is Stator input voltage vector in coordinate system This is an additional term that includes the permanent magnet flux linkage term and the rotational speed coupling term.
[0021] In this embodiment, the PMSM is powered by a two-level inverter, where the eight voltage vector switching states are possible: ; in: ; G represents the gate signal of the inverter: ; in, This is the DC bus voltage. The rotor electrical angle.
[0022] S2. Classification strategy based on cost function and q-axis current.
[0023] Within each control cycle, based on the PMSM discrete prediction model established in S1, all base voltage vectors of the inverter are iterated one by one. Each base voltage vector is then substituted into the current prediction equation to calculate its corresponding value at the next sampling time. d shaft and q Predict the shaft current value. Then, based on the torque current ( q Based on the prediction error direction of the axis, the vector is divided into "overshoot vector group" and "undershoot vector group". The specific steps are as follows: Substitute the eight vectors into the current prediction formula in sequence: ; Then based on q Axis prediction errors are classified as follows: ; Then, based on the sign of the error, it is divided into overshoot vector groups. With undershoot vector group : ; in, Candidate voltage vector corresponding Axis prediction error for Shaft reference current, Indicates the first Candidate voltage vectors.
[0024] S3. A dual-vector screening strategy based on classification.
[0025] In the overshoot and undershoot vector groups divided by S2, the cost function under the action of each vector is calculated. The overshoot vector and the undershoot vector with the smallest cost function within the group are selected as the optimal candidate combination. If all vectors are in the overshoot or undershoot vector group, and the traversal results show that all base voltage vectors are assigned to the overshoot or undershoot vector group, i.e., a one-sided saturation situation occurs, then the dual-vector synthesis condition cannot be achieved. In this case, the optimal switching vector is directly selected based on the cost function and applied, which transforms into traditional model predictive control and outputs a switching pulse. The specific steps are as follows: Find, respectively, the cost function in the overshoot vector set and the undershoot vector set. The smallest vector is used as the optimal overshoot vector. and optimal undershoot vector : Cost function: ; Optimal vector selection: ; If unilateral saturation occurs, it reverts to traditional model predictive current control: At this time, the saturated vector group is Select the optimal vector under the cost function filtering. And applied to the final current prediction equation, we get: .
[0026] S4. Dead zone determination and dead time synthesis.
[0027] After successfully selecting the optimal overshoot and undershoot vectors in S3, the impact of the inverter dead-time effect on the prediction accuracy of the actual output voltage and current is further considered. During motor operation, when the inverter arm switches between states, a dead time is typically set during commutation to prevent shoot-through of the power switching devices on the same arm. During the dead time, both the upper and lower power switching devices of that phase arm are in the off state. However, since the motor winding current cannot change abruptly, the phase current will continue to flow through the anti-parallel diode or freewheeling path, clamping the output point of that phase arm to the positive or negative terminal of the DC bus. Therefore, the actual output state during the dead time is not entirely equivalent to the ideal switching command, but is determined by both the direction of the phase current and the commutation direction of the arm.
[0028] When phase current When the phase arm is in the dead time, it is equivalent to the freewheeling state of the lower arm; when the phase current... During the dead time, the phase arm is equivalent to the upper arm freewheeling state. Therefore, under different commutation directions and phase current polarities, the dead time effect manifests as commutation lag or equivalent to immediate commutation, as shown in Table 1.
[0029] Table 1 Dead Zone Vector Determination Table Based on Current Polarity
[0030] Previous actual switching state Optimal overshoot vector Optimal undershoot vector Specifically as follows: ; in, , , These represent the switching states of the three-phase bridge arms A, B, and C, with values of 0 or 1.
[0031] Within a control cycle, the dual-vector action process can be represented as: the previous actual switching state First switch to the optimal overshoot vector Then, by the optimal overshoot vector Switch to optimal undershoot vector Therefore, there are two dead-time action segments within a control cycle: the first dead-time action segment corresponds to... The commutation process, the second dead zone action segment corresponds to The commutation process.
[0032] For ease of representation, the following definition is provided: ; in, This is the previous actual switching state. This is the first optimal vector switching state. This is the second optimal vector switching state.
[0033] For the One dead zone action segment, , No. The commutation indicator variable of the phase bridge arm is defined as: ; when When, it indicates the first Phase bridge arm in the first Commutation occurs in the dead zone segment; when When, it indicates the first No phase commutation occurred in the dead zone of the bridge arm.
[0034] According to Table 1 and the phase current polarity, the first Within the dead zone segment, the first The equivalent switching state of the phase bridge arm can be expressed as: ; in, Indicates the first Within the dead zone segment, the first The equivalent switching state of the phase bridge arm; Indicates the first Phase current, .
[0035] Therefore, the equivalent switching states within the two dead-zone operating segments are as follows: ; ; in, and These represent the actual effective switching states within the first and second dead-zone action segments, respectively.
[0036] The two dead-zone action segments correspond to The shaft dead zone voltage vector is ; ; in, , Within the first dead zone action segment The d-axis and q-axis components of the dead zone voltage; , The second dead zone action segment The d-axis and q-axis components of the dead zone voltage.
[0037] Similarly, the optimal overshoot vector and the optimal undershoot vector correspond to The shaft voltages are respectively: ; ; in, , The optimal overshoot vectors are respectively at Voltage components along the d-axis and q-axis in the coordinate system; , The optimal undershoot vector is at Voltage components along the d-axis and q-axis in the coordinate system.
[0038] According to permanent magnet synchronous motors The differential equation for current in the coordinate system can be defined as follows: the slope of the fundamental current, determined by the motor's own state and excluding the input voltage term, is: ; The current slopes corresponding to the two dead-zone action segments are: ; ; The current slope corresponding to the optimal overshoot vector is: ; The current slope corresponding to the optimal undershoot vector is: ; Let the single dead time be... Each of the two dead-time segments occupies one dead-time period. Let the actual action times of the optimal overshoot vector and the optimal undershoot vector after deducting the dead time be respectively... and Then consider the dead time being occupied separately. The axis current prediction equation is: ; Since the dead time occupies a fixed period within the control cycle, the actual action time of the two normal voltage vectors satisfies: ; make ,in To determine the effective action time that can be allocated to the two optimal voltage vectors after removing the two dead times, we have: ; Based on the principle of deadbeat control, it is required that The shaft current tracks the reference value at the next sampling time: ; Will q Substituting the error-free condition into the axis prediction equation, we get: ; Will Substituting into the above formula, we get: ; The actual duration of the optimal undershoot vector is: ; Finally, based on the calculated duration of action... , , , The corresponding dual-vector modulation switching pulse signals are generated in the order of the first dead zone action segment, the optimal overshoot vector action segment, the second dead zone action segment, and the optimal undershoot vector action segment, and then output to the inverter, thereby realizing the dead-time occupancy and dead-time voltage influence of the permanent magnet synchronous motor dual-vector model prediction deadbeat current control.
[0039] Therefore, this invention adopts the above-mentioned dual-vector model predictive current control method for permanent magnet synchronous motors that considers dead zone. By integrating dead zone accurate modeling based on current polarity, grouped parallel screening strategy based on cost function, and deadbeat duty cycle analytical calculation, it aims to reduce current distortion caused by dead zone effect, solve the local optimum problem in traditional strategies, and significantly reduce computational burden. It provides a highly efficient control scheme for marine PMSM propulsion systems that combines high dynamic response, global steady-state accuracy, and low harmonic characteristics.
[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for predictive current control of a permanent magnet synchronous motor using a dual-vector model considering dead time, characterized in that, Includes the following steps: S1. Establish and sample the discrete prediction model of PMSM: Establish a discrete mathematical model of PMSM, combine motor parameters with the current feedback state, derive the current prediction equation, and use it to calculate the predicted current and current change slope under the action of each base voltage vector. S2. Classification strategy based on cost function and torque current: Within each control cycle, based on the PMSM discrete mathematical model, all base voltage vectors are traversed, and the next time step is calculated using the current prediction equation. d shaft and q Shaft predicts current, based on q Based on the direction of axis prediction error, the voltage vector is divided into overshoot vector group and undershoot vector group; S3. Dual-vector screening strategy based on classification strategy: Based on the overshoot vector group and undershoot vector group divided by S2, calculate the cost function under the action of each vector, and screen out the overshoot vector and undershoot vector with the smallest cost function in the group as the optimal dual-vector candidate combination. S4. Dead Zone Determination and Deadbeat Time Synthesis: Based on the deadbeat control principle, and combining the corrected dead zone slope and the optimal vector slope, the precise action time of the two optimal vectors within the control cycle is calculated through the control equation. Under the premise of explicitly reducing the current distortion caused by the dead zone effect, the final switching control pulse is output.
2. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 1, characterized in that, The specific steps of S1 are as follows: S11. Calculate the two-phase rotating coordinate system of a three-phase permanent magnet synchronous motor. d axis and q The equations for magnetic flux linkage and voltage on the shaft, and the torque equation; S12. Construct the system dynamic equations of a three-phase PMSM under synchronous rotating coordinate system, considering parameter mismatch and model uncertainty; S13. The system dynamic equations are discretized using the first-order forward Euler method to obtain the current prediction equations for the d-axis and q-axis. S14. Use the PMSM discrete mathematical model to predict the current in advance and obtain the final current prediction equation.
3. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 2, characterized in that, The specific formula for the current prediction equation in S14 is as follows: ； in, for The predicted current vector at time t, The discrete state matrix, For discrete input matrices, for At time, the candidate voltage vector is Stator input voltage vector in coordinate system This is an additional term that includes the permanent magnet flux linkage term and the rotational speed coupling term.
4. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 2, characterized in that, The specific steps of S2 are as follows: S21. Substitute each of the eight voltage vectors of the PMSM into the current prediction equation; S22. Calculate and determine the direction of the q-axis prediction error; S23. Based on the sign of the q-axis prediction error, each voltage vector is divided into an overshoot vector group and an undershoot vector group.
5. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 4, characterized in that, The specific steps of S3 are as follows: S31. Select the vectors that minimize the cost function from the overshoot vector group and the undershoot vector group respectively, and use them as the optimal overshoot vector and the optimal undershoot vector: S32. If unilateral saturation occurs, switch to traditional model predictive current control, select the optimal cost function vector from the saturation vector group and output it.
6. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 5, characterized in that, The specific steps of S31 are as follows: S311. Introduce the cost function, which is as follows: ； in, Let cost function be This is the reference value for the d-axis current. For the d-axis predicted current under the action of candidate voltage vectors, This is the reference value for the q-axis current. Predict the q-axis current under the action of the candidate voltage vector; S312. Optimal vector selection, as detailed below: ； in, To be the optimal overshoot vector, For overshoot vector group, To be the optimal undershoot vector, For undershoot vector groups, For candidate voltage vectors, This represents the voltage vector that minimizes the cost function.
7. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 5, characterized in that, The specific steps of S4 are as follows: S41. Based on saturation determination, select the optimal overshoot vector and the optimal undershoot vector, and start the dual-vector modulation process; S42. Identify the commutation process of the two bridge arms within the control cycle, and combine the polarity of each phase current to obtain the equivalent switching state and dead zone voltage vector of the two dead zone action segments according to the judgment rules. S43, According to the motor d shaft and q The current equation of the shaft is used to calculate the slope of the base current. The dead zone voltage and the two optimal vector voltages are superimposed to obtain the current slope under the corresponding operating conditions. S44. Split the control cycle, quantify the duration of the two dead time segments, and calculate the remaining total effective duration; S45. By combining the superposition of the dead zone vector and the two optimal voltage vectors, the current variation law of the entire process is clarified. S46, based on the principle of zero-delay control, with q With the shaft current tracking reference value as the objective, the actual action time of the two optimal vectors is calculated, and finally, the switching pulse output is generated according to the timing sequence.
8. The method for predictive current control of a permanent magnet synchronous motor considering dead zone using a dual-vector model according to claim 7, characterized in that, The specific formulas for the actual action time of the two optimal vectors of S46 are as follows: ； ； in, t 1. t 2 represents the actual action time of the optimal overshoot vector and the optimal undershoot vector after deducting the dead zone, respectively. for, T s The sampling period is To sample the q-axis current, For single dead time, , These represent the q-axis current slopes under the optimal vector dead-zone effect for the overshoot and undershoot groups, respectively. , These are the q-axis current slopes under the optimal vector action of the overshoot and undershoot groups, respectively.