Permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization

Abstract

The application relates to the technical field of permanent magnet synchronous motor control, and discloses a permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization, which comprises the following steps: S1, collecting three-phase current, rotor position electric angle and rotating speed information of a three-phase permanent magnet synchronous motor, and converting the current to a dq coordinate system through coordinate transformation; S2, predicting dq axis current of a next control period according to dq axis voltage vectors of a previous control period; S3, generating a q axis current reference value by using a rotating speed controller based on a reference rotating speed of the motor, and setting a d axis current reference value to zero. Through voltage vector preselection and cost function optimization, the number of candidate voltage vectors is reduced from 37 to 4, the calculation complexity is significantly reduced, and the control real-time performance is improved. Meanwhile, the steady-state performance is improved, the current ripple and torque fluctuation are reduced, the fast dynamic response and low algorithm requirement are combined, and an efficient solution is provided for high-precision control of the permanent magnet synchronous motor.

Description

Technical Field

[0001] This invention relates to the field of permanent magnet synchronous motor control technology, specifically a speed regulation method for permanent magnet motors based on current prediction and discrete space vector optimization. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in precision control due to their high power density and efficiency. However, achieving rapid and precise control of their speed, torque, and current remains a technical challenge. While traditional field-oriented control can achieve good static performance through coordinate decoupling, its reliance on PI controllers results in slow dynamic response and requires significant time to fine-tune parameters to adapt to complex environments. Direct torque control (DTC) achieves rapid response through simple switching meters, but in high-precision applications, torque ripple and uncontrollable switching frequencies significantly limit its performance.

[0003] In recent years, model predictive current control (MMDC) has attracted attention as an emerging control strategy due to its fast response and simple structure. However, this method suffers from large current ripple and poor steady-state performance because only one voltage vector is applied during the control cycle. This is particularly problematic under high-precision control requirements, where traditional MMDC struggles to meet the demands. To address this issue, researchers have proposed a virtual voltage vector synthesis technique, which expands the voltage vector set through discrete space vector modulation, thereby effectively reducing current ripple and torque pulsation.

[0004] While discrete space vector modulation (DLP) significantly improves steady-state performance, it comes at the cost of a substantial increase in computational complexity. Since the number of virtual voltage vectors far exceeds the number of real voltage vectors, it is necessary to calculate the predicted current values ​​for each candidate voltage in each control cycle and select the optimal voltage vector using a cost function. This dramatic increase in computation leads to a greater dependence of the controller on hardware computing power, resulting in a significant decrease in real-time performance and cost-effectiveness. Especially on devices with low computing power, the engineering application of this method becomes difficult, becoming a bottleneck for the further development of model predictive control.

[0005] To address the aforementioned problems, there is an urgent need for a motor control method that can maintain superior steady-state performance while significantly reducing computational burden. This invention proposes a permanent magnet motor speed control method based on current prediction and discrete space vector optimization. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a speed control method for permanent magnet motors based on current prediction and discrete space vector optimization, which solves the problems of poor steady-state performance and high computational complexity in traditional model predictive current control.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a speed control method for permanent magnet motors based on current prediction and discrete space vector optimization, comprising the following steps:

[0008] S1. Collect the three-phase current, rotor position electrical angle and speed information of the three-phase permanent magnet synchronous motor, and transform the current to the dq coordinate system through coordinate transformation;

[0009] S2. Based on the dq-axis voltage vector of the previous control cycle, predict the dq-axis current of the next control cycle;

[0010] S3. Based on the motor's reference speed, generate the q-axis current reference value using the speed controller, and set the d-axis current reference value to zero;

[0011] S4. Calculate the dq-axis reference voltage based on the reference current, and convert the reference voltage to a reference voltage in a stationary coordinate system;

[0012] S5. Select the sector and quadrant where the candidate voltage vector is located based on the reference voltage;

[0013] S6. Calculate the predicted current value under the action of candidate voltage vectors, and select the optimal voltage vector through the cost function;

[0014] S7. Output the inverter switching signal corresponding to the optimal voltage vector and apply it to the inverter.

[0015] Preferably, step S1 specifically includes the following steps:

[0016] S1.1 Acquire the three-phase current signal of the three-phase permanent magnet synchronous motor;

[0017] S1.2. Obtain rotor position electrical angle and speed signals through motor sensors;

[0018] S1.3. Using coordinate transformation, the collected three-phase current is converted from the three-phase stationary coordinate system to the dq-axis coordinate system, and the d-axis current and q-axis current are output.

[0019] Preferably, step S2 uses a one-time delay compensation method to predict the dq-axis current, and the one-time delay compensation method includes the following steps:

[0020] S2.1. Based on the dq-axis voltage vector of the previous control cycle and the motor parameters of the current cycle, predict the dq-axis current of the next control cycle.

[0021] S2.2. Compensate for the d-axis and q-axis current values ​​of the current cycle using the stator internal resistance, inductance, electric angular velocity and permanent magnet flux linkage of the motor.

[0022] S2.3 The corrected dq-axis current is further predicted using the inductance, voltage vector, and motor flux linkage model to obtain the d-axis current and q-axis current for the next control cycle.

[0023] Preferably, step S3 specifically includes the following steps:

[0024] S3.1, Collect the deviation between the current actual motor speed and the reference speed;

[0025] S3.2 Calculate the integral and proportional values ​​of the speed error using the speed controller;

[0026] S3.3. Generate the q-axis current reference value based on the calculation results of the rotational speed error;

[0027] S3.4 Set the d-axis current reference value to zero to achieve decoupled control of motor torque and flux linkage.

[0028] Preferably, in step S4, the dq-axis reference voltage is calculated in the following manner:

[0029] S4.1. Based on the stator resistance, d-axis inductance, and q-axis inductance of the motor, and combined with the dq-axis reference current and electric angular velocity, calculate the dq-axis reference voltage.

[0030] S4.2. Using the inverse coordinate transformation method, the dq-axis reference voltage is converted into the α-axis reference voltage and β-axis reference voltage in the stationary coordinate system.

[0031] Preferably, the selection of candidate voltage vectors in step S5 includes the following steps:

[0032] S5.1 Determine the sector where the candidate voltage vector is located based on the magnitude and direction of the reference voltages along the α-axis and β-axis in the stationary coordinate system;

[0033] S5.2. Based on the combination of the virtual voltage vector and the actual voltage vector of the sector, determine the candidate quadrant;

[0034] S5.3 Candidate voltage vectors include zero voltage vectors, active voltage vectors, and virtual voltage vectors synthesized from two or more real voltage vectors.

[0035] Preferably, in step S6, when calculating the predicted current value under the action of the candidate voltage vector, the following formula is used:

[0036] d-axis current prediction:

[0037]

[0038] q-axis current prediction:

[0039]

[0040] in, and T represents the dq-axis component of the candidate voltage vector. s The sampling period is R, where R is the stator internal resistance of the motor, and L is L. d L q For inductance, ω e Let ψ be the electric angular velocity. f For permanent magnet flux linkage, i d (k+1) and i q (k+1) represent the d-axis and q-axis current values ​​for the (k+1)th control cycle, respectively.

[0041] Preferably, the cost function for step S6 is as follows:

[0042]

[0043] in, and This is the reference value for the dq axis current. and This represents the predicted dq-axis current under the action of the candidate voltage vector.

[0044] This invention provides a method for speed control of permanent magnet motors based on current prediction and discrete space vector optimization.

[0045] It has the following beneficial effects:

[0046] 1. This invention employs a model predictive control (MMC) technique based on discrete space vector modulation (DSP). By pre-selecting discrete voltage vectors and filtering for the optimal voltage vector, it achieves a significant improvement in steady-state performance. Compared to existing techniques that directly calculate a large number of virtual voltage vectors, the discrete voltage vector pre-selection method proposed in this invention reduces the number of candidate voltage vectors from 37 to 4, effectively reducing computational complexity and solving the problem of excessive computational burden in traditional model predictive control.

[0047] 2. This invention significantly optimizes current control performance by combining deadbeat prediction with voltage vector partitioning, achieving a balance between steady-state and dynamic performance. Compared with existing technologies that use a single voltage vector, resulting in large steady-state current ripple, this invention ensures smooth output current and small torque fluctuation by selecting candidate voltage vectors through a cost function, overcoming the shortcomings of poor steady-state performance of traditional methods while retaining the advantages of fast dynamic response. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the permanent magnet synchronous motor drive topology of the present invention;

[0049] Figure 2This is a schematic diagram of the real voltage vector space corresponding to the two-level inverter of the present invention;

[0050] Figure 3 A schematic diagram of the control block diagram for the low-computing-power model predictive current control method for permanent magnet synchronous motors based on discrete space vector modulation provided by the present invention;

[0051] Figure 4 A schematic diagram of the control flow of the low-computing-power model predictive current control method for permanent magnet synchronous motors based on discrete space vector modulation provided by the present invention;

[0052] Figure 5 This is a schematic diagram of the virtual voltage space of the present invention;

[0053] Figure 6 This is the quadrant distribution of sector I in this invention;

[0054] Figure 7 This is a schematic diagram of the simulation test results under steady-state conditions of the present invention;

[0055] Figure 8 The figure shows the simulation test results under dynamic conditions of the present invention. Detailed Implementation

[0056] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please see the appendix Figure 1 -Appendix Figure 8 This invention provides a speed control method for permanent magnet motors based on current prediction and discrete space vector optimization, comprising the following steps:

[0058] S1. Collect the three-phase current, rotor position electrical angle and speed information of the three-phase permanent magnet synchronous motor, and transform the current to the dq coordinate system through coordinate transformation;

[0059] In the method of this invention, the acquisition of operating parameters and the transformation of current coordinates of the three-phase permanent magnet synchronous motor are the foundation and key steps of the entire control process. By acquiring three-phase current, rotor position electrical angle, and speed information in real time, and transforming the current from the three-phase stationary coordinate system to the dq coordinate system, necessary initial data can be provided for subsequent current prediction and voltage vector selection.

[0060] Typically, operating parameters such as three-phase current, rotor position electrical angle, and rotational speed are the basic input data for the control system. By acquiring these parameters in real time and converting them to a rotating coordinate system dq, decoupled current control can be achieved, providing support for subsequent optimization.

[0061] In this embodiment, the specific implementation method is as follows:

[0062] First, the three-phase current of the three-phase permanent magnet synchronous motor is collected and denoted as i. a i b and i c These current components are instantaneous currents in the motor stator windings, which are acquired in real time by current sensors (such as Hall sensors or current transformers).

[0063] Simultaneously, the electrical angle of the rotor position is obtained using a position sensor (such as a rotary transformer or encoder), denoted as θ. e This electrical angle is a key input for subsequent coordinate transformations, used to convert the current from a three-phase stationary coordinate system to a rotating coordinate system.

[0064] In addition, the motor speed n can be obtained by directly measuring with a speed sensor or by differential calculation of the rotor position electrical angle signal. In the absence of a speed sensor, a method based on back electromotive force estimation can also be used to calculate the speed n.

[0065] Specifically, the collected three-phase current i a i b i c This refers to the current component in the stationary coordinate system. To achieve subsequent decoupling control, the current components in the three-phase stationary coordinate system need to be transformed into the rotating coordinate system dq. This step requires introducing the rotor position electrical angle θ. e The coordinate transformation is completed using the following formula:

[0066] The converted d-axis current:

[0067]

[0068] The converted q-axis current:

[0069]

[0070] Where: i a i b i c It is a three-phase current component; θ e It is the rotor position electrical angle; i d and i q These are the d-axis current and q-axis current in the rotating coordinate system, respectively.

[0071] It should be noted that in the rotating coordinate system, the d-axis current i d This is usually related to the flux linkage control of the motor, while the q-axis current i q This is related to torque control. The transformation achieved through the above formula enables complete decoupling of the current, providing accurate basic data for subsequent current prediction.

[0072] Typically, the acquired current, position, and speed signals are affected by noise and interference. In some embodiments, these signals can be filtered. For example, for θ e The acquisition signals of n are filtered using a low-pass filter to reduce the impact of mechanical vibration and electromagnetic interference on measurement accuracy.

[0073] Furthermore, in actual control systems, if there is a delay in the data acquisition by sensors, interpolation prediction can be used to reduce the impact of the delay on the acquired data, thereby further improving control accuracy.

[0074] As an alternative, in the absence of a speed sensor, the motor speed n can be estimated using the back electromotive force (EMF). This estimation method derives the back EMF from the voltage and current relationship of the motor stator windings, and then calculates the electrical angular velocity ω using known motor parameters. e And the mechanical rotation speed n.

[0075] To ensure the clarity of the formula, the parameters in the formula are defined as follows:

[0076] i a i b i c Current components in a three-phase stationary coordinate system;

[0077] i d i q : Current component in the rotating coordinate system dq;

[0078] θ e : Rotor position electrical angle, describing the angular position of the rotor magnetic poles relative to the stator reference coordinate system;

[0079] n: Mechanical rotational speed, in rpm;

[0080] ω e Electric angular velocity, measured in radians per second, can be expressed by the formula ω. e =2πn / 60.

[0081] In some embodiments, dynamic error correction can also be incorporated during current acquisition and coordinate transformation. For example, the acquired signal can be corrected online based on known motor parameters (such as inductance and resistance) and system states (such as motor speed and current) to compensate for sensor errors. This method can further improve the accuracy of the data acquisition process.d and i q The calculation accuracy.

[0082] As an alternative, a distributed acquisition method can be used, which sends the three-phase current and position signals to multiple parallel processing modules respectively, improving data processing efficiency. This method is particularly suitable for motor control scenarios with high dynamic response.

[0083] Through the above steps, this invention accurately transforms the current components in the three-phase stationary coordinate system to the rotating coordinate system dq, achieving decoupling and preliminary processing of the current components, laying the foundation for subsequent current prediction, cost function calculation, and voltage vector selection. The entire implementation process closely integrates real-time acquisition and mathematical transformation, ensuring control accuracy and reliability.

[0084] S2. Based on the dq-axis voltage vector of the previous control cycle, predict the dq-axis current of the next control cycle;

[0085] In the implementation of this invention, current prediction is one of the core steps of motor control. By using a mathematical model based on the dq-axis voltage vector and motor parameters of the previous control cycle, the dq-axis current of the next control cycle is predicted, providing accurate reference data for subsequent current control and voltage vector selection.

[0086] Generally, the dq-axis current of a permanent magnet synchronous motor is affected by factors such as the voltage vector of the previous cycle, motor parameters (such as inductance and resistance), and rotor electrical angular velocity. By employing a one-beat delay compensation method, the current for the next control cycle can be accurately predicted within the current cycle.

[0087] In this embodiment, the process of predicting the dq-axis current includes the following:

[0088] First, based on the d-axis voltage u of the previous control cycle d (k-1) and q-axis voltage u q (k-1), combined with the dq-axis current i of the current period d (k) and i q (k) Calculate the predicted dq-axis current i for the next control cycle using the mathematical model of the permanent magnet synchronous motor. d (k+1) and i q (k+1).

[0089] The prediction formula is as follows:

[0090] d-axis current prediction formula:

[0091]

[0092] q-axis current prediction formula:

[0093]

[0094] Wherein: T s The sampling period represents the time interval between each cycle of the control system; L d L q ω represents the d-axis and q-axis inductances, reflecting the inductance characteristics of the motor along different axes; R is the stator internal resistance of the motor, used to describe the resistance characteristics of the motor windings; e The rotor's electric angular velocity is related to its rotational speed n and the number of pole pairs P as follows: ψ f The term "permanent magnet flux linkage" refers to the flux linkage generated by the rotor's permanent magnets; u d (k-1), u q (k-1) represents the d-axis and q-axis voltage vector components of the previous control cycle.

[0095] In the above formula, the d-axis current i d The prediction of (k+1) is affected by the following factors:

[0096] d-axis voltage u d (k-1): Provides the main driving force that directly affects the change of d-axis current.

[0097] The coupling effect ω between q-axis current and rotor electric angular velocity e L q i q (k): Reflects the indirect influence of q-axis inductance, current, and electric angular velocity on d-axis current.

[0098] Stator resistance - Ri d (k): reflects the current consumption caused by the resistance of the motor stator winding.

[0099] For the q-axis current i q (k+1), its variation is affected by the following factors:

[0100] q-axis voltage u q (k-1): is the main driving force of the q-axis current.

[0101] The coupling effect between permanent magnet flux linkage and d-axis current -ω e (L d i d (k)+ψ f This reflects the influence of the rotor permanent magnet flux linkage and its coupling with the d-axis current on the q-axis current.

[0102] Stator resistance - Ri q (k): Reflects the energy loss of the resistor to the q-axis current.

[0103] Using the above formula, the predicted dq-axis current i for the next control cycle can be calculated in real time.d (k+1) and i q (k+1). These predictions provide the basis for subsequent cost function calculations and optimal voltage vector selection.

[0104] In some embodiments, changes in parameters in the prediction formula can be corrected online. For example, to address temperature drift of the stator resistance R, the motor operating temperature can be detected in real time and the value of R can be dynamically adjusted, thereby improving prediction accuracy.

[0105] Furthermore, at higher rotational speeds, the electric angular velocity ω e Changes in current have a significant impact on current prediction. Therefore, the reliability of the prediction results can be further improved by adding a high-precision speed estimation module.

[0106] Through the above steps, this invention achieves dq-axis current prediction based on the voltage vector of the previous control cycle. By introducing motor parameters and a control model, this method effectively improves the real-time performance and accuracy of current prediction, providing accurate input for subsequent control operations.

[0107] S3. Based on the motor's reference speed, generate the q-axis current reference value using the speed controller, and set the d-axis current reference value to zero;

[0108] In this invention, to achieve precise speed control of the motor, it is necessary to compare the reference speed with the actual speed and generate a corresponding q-axis current reference value through a speed controller. At the same time, the d-axis current reference value Set to zero to achieve decoupled control of motor flux and torque.

[0109] Generally, motor speed control is achieved through closed-loop feedback. The core of the speed controller lies in calculating the speed error and generating a suitable current reference value based on the error, thereby adjusting the motor torque to achieve the desired speed.

[0110] In this embodiment, step S3 is implemented in the following ways:

[0111] First, the current actual rotational speed n is collected and compared with the set reference rotational speed n. ref Compare and calculate the speed error Δn:

[0112] Δn=n ref -n

[0113] Where, n ref Let n be the target speed and n be the actual speed.

[0114] Next, the speed error Δn is input into the proportional-integral (PI) controller, which generates a q-axis current reference value based on the speed error. The calculation formula for a PI controller is:

[0115]

[0116] Where: K p K is the proportional gain of the PI controller, used to adjust the system's response speed. i ∫Δndt is the integral coefficient of the PI controller, used to eliminate steady-state error; ∫Δndt is the integral value of the speed error.

[0117] Specifically, the function of a PI controller is:

[0118] When the speed error is large, the proportional part K p Δn provides rapid response capabilities;

[0119] When the speed error persists, the integral part K i ∫Δndt provides correction capability, ensuring that the speed error approaches zero.

[0120] Generated q-axis current reference value It is directly used for subsequent current control and optimization calculations, reflecting the torque component required by the motor.

[0121] In this embodiment, to achieve decoupled control of the current, the d-axis current reference value is... It is set to zero. This is because in a permanent magnet synchronous motor, the d-axis current is mainly related to the flux linkage, and the permanent magnet already provides sufficient flux linkage, so there is no need to supplement it through the d-axis current. Therefore, it is set to zero. It simplifies control and concentrates all control precision on the q-axis current.

[0122] Rotational speed acquisition and feedback:

[0123] Rotational speed n is usually acquired by a speed sensor or obtained by differential calculation using the angle signal from a position sensor (such as a resolver or encoder).

[0124] The accuracy of speed feedback directly affects the calculation of the q-axis current reference value. Therefore, the acquired signal can be filtered to eliminate noise interference.

[0125] Setting PI controller parameters:

[0126] K p and K i The selection of the proportional coefficient K needs to be determined based on the dynamic response requirements of the motor. Generally, the proportional coefficient K... p When the integral coefficient K is large, the system response speed is faster, but oscillations may be introduced; i When the error is large, the steady-state error decreases, but the system response slows down.

[0127] In some embodiments, K can be dynamically adjusted using an adaptive adjustment algorithm. p and K i To balance speed and stability.

[0128] Setting the d-axis current:

[0129] The flux linkage of a permanent magnet synchronous motor is provided by the rotor permanent magnets, so the d-axis current can be set to zero. This setting simplifies the control model and eliminates the coupling between the d-axis and q-axis, avoiding complex decoupling calculations.

[0130] Steady-state and dynamic responses:

[0131] During steady-state operation, the speed error Δn is small, and the PI controller output... It is mainly determined by the integral part, in order to maintain stable operation.

[0132] During dynamic operation (such as changes in speed or sudden load changes), the PI controller can respond quickly and output a large value through its proportional section. It meets dynamic performance requirements.

[0133] In some embodiments, a feedforward control loop can be added to the PI controller. By introducing a feedforward term based on the motor's mathematical model, the dynamic response capability can be further improved. For example, the required q-axis current can be directly estimated based on load changes and added to the output of the PI controller, thereby accelerating the response speed.

[0134] Furthermore, in scenarios without speed sensors, the rotational speed n can be calculated using an estimation method based on back electromotive force. This method indirectly derives the rotational speed by utilizing the voltage and current relationship of the motor stator windings and combining this with known motor parameters.

[0135] Through the above steps, this invention generates a q-axis current reference value based on the reference rotational speed and the speed controller, and sets the d-axis current reference value to zero, thus completing the initial current setting. This process achieves decoupled current control, laying the foundation for subsequent current prediction and optimization calculations.

[0136] S4. Calculate the dq-axis reference voltage based on the reference current, and convert the reference voltage to a reference voltage in a stationary coordinate system;

[0137] In this invention, the dq-axis reference voltage is calculated using a method based on a deadbeat model prediction. and Subsequently, the calculated dq-axis reference voltage is converted into a reference voltage in the stationary plane coordinate system (α-β) through an inverse Park transformation. and Simultaneously, in accordance with the rules of discrete space vector modulation, candidate voltage vectors are pre-selected, including determining candidate sectors and candidate quadrants.

[0138] Generally, discrete space vector modulation methods can effectively improve the accuracy and efficiency of motor control. By calculating the reference voltage and pre-selecting the space vector, candidate voltage vectors are determined, control performance is optimized, and accurate input data is provided for subsequent cost function calculations.

[0139] In this embodiment, the specific implementation method is as follows:

[0140] In one possible implementation, based on the dq-axis reference current... and Combined with the predicted dq-axis current value i for the current period d (k+1),i q (k+1), calculate the dq-axis reference voltage using the following formula:

[0141] d-axis reference voltage:

[0142]

[0143] q-axis reference voltage:

[0144]

[0145] Where: L d L q These are the d-axis and q-axis inductances, respectively; T s Sampling period; R is the stator resistance of the motor; ω e ψ is the rotor's electric angular velocity. f It is a permanent magnet flux linkage.

[0146] The core of the formula is based on the deadbeat control method, which calculates the voltage vector by compensating for the error between the predicted and reference values ​​of the dq axis current.

[0147] Next, the calculated dq-axis reference voltage is obtained through inverse Park coordinate transformation. Converted to static plane reference voltage and The conversion formula is as follows:

[0148] Stationary plane reference voltage:

[0149]

[0150] Where: θ e The rotor position electrical angle is a key input for the inverse Park transform.

[0151] By using the inverse Park transformation, the reference voltage in the rotating coordinate system is mapped to the stationary coordinate system, providing input for subsequent voltage vector pre-selection.

[0152] Obtained through calculation and Determine the sector containing the candidate voltage vector. The discrete voltage vector is defined by the following formula:

[0153]

[0154] Where: V x V y V z for Figure 2 The real voltage vector of the two-level inverter is represented in the figure; x, y, z are adjacent voltage vectors that satisfy x, y, z ∈ {0, 1, ..., 7}.

[0155] Discrete voltage vectors can be divided into three categories:

[0156] When V x V y V z When two zero-voltage vectors are included, the synthesized virtual vector is the smaller virtual vector;

[0157] When V x V y V z When a zero-voltage vector is included, the synthesized virtual vector is a medium virtual vector;

[0158] When V x V y V z When the zero voltage vector is not included, the synthesized virtual vector is a large virtual vector.

[0159] according to and The size of the reference voltage vector and the candidate sectors can be determined using the following table:

[0160]

[0161] For example, when and At that time, the candidate sector is Figure 5 Sector I, or SI, is shown in the diagram below. Figure 6 As shown.

[0162] S5. Select the sector and quadrant where the candidate voltage vector is located based on the reference voltage;

[0163] After determining the candidate sectors, by comparison The virtual vectors in the candidate sectors that are not in the same direction as the real vectors are used to select the quadrants in which the candidate voltages are located, i.e., the candidate quadrants. The voltage vectors in the candidate quadrants are the candidate voltage vectors. The method for selecting candidate quadrants is shown in the table below:

[0164]

[0165]

[0166] For example, when the candidate sector is sector I, when and At that time, the candidate quadrant is Figure 5 Quadrant ① in the middle, i.e. The candidate voltage is V 110 V 120 V 122 V 222 .

[0167] This step calculates the dq-axis reference voltage using the deadbeat-free model prediction method. Combined with the inverse Park transform and discrete space vector modulation rules, it completes the reference voltage conversion and pre-selection of candidate voltage vectors, providing a foundation for subsequent optimized control. The entire method ensures control accuracy while improving operational efficiency.

[0168] S6. Calculate the predicted current value under the action of candidate voltage vectors, and select the optimal voltage vector through the cost function;

[0169] In this invention, by applying each candidate voltage vector V n The function is to predict the current, and by combining the deviation between the reference current value and the predicted value, a cost function is calculated to select the optimal voltage vector V. opt This step fully utilizes the mathematical model of the permanent magnet synchronous motor and the distribution characteristics of the candidate voltage vectors, which not only ensures the dynamic response performance of the motor but also improves its steady-state performance, significantly reducing computational complexity.

[0170] In one possible implementation, based on the candidate voltage vector V obtained in step 5 n Predict the next value of the dq-axis current. and The prediction formula is as follows:

[0171] d-axis current prediction formula:

[0172]

[0173] q-axis current prediction formula:

[0174]

[0175] Explanation of formula parameters:

[0176] i d (k+1),i q (k+1) represents the dq-axis current of the current prediction period; Candidate voltage vector V n The dq-axis current in the next cycle under the action; Candidate voltage vector V n Components in the dq coordinate system; L d ,L q ω represents the d-axis and q-axis inductance; e ψ is the rotor's electric angular velocity. f R is the flux linkage of the permanent magnet; T is the internal resistance of the stator; s The sampling period.

[0177] In the implementation of this invention, by using the candidate voltage vector V n dq components Substituting into the above formula, we predict and calculate the dq-axis current value at the next moment under the influence of each candidate voltage vector.

[0178] After obtaining the predicted current value under the action of each candidate voltage vector, the effect of each candidate vector is evaluated by the cost function g, and finally the voltage vector V with the minimum cost function value is selected. opt .

[0179] Cost function formula:

[0180]

[0181] Explanation of formula parameters:

[0182] This is the reference current value for the dq axis; Candidate voltage vector V n The predicted dq-axis current value under the action; g is the cost function value, used to evaluate the control effect of the candidate voltage vector.

[0183] Selection method:

[0184] Traverse all candidate voltage vectors V n Calculate the corresponding cost function value for each one;

[0185] Compare the cost function values ​​g of all candidate voltage vectors, and select the voltage vector V with the smallest cost function value. opt .

[0186] Assuming the reference current value is The predicted current value for the current period is i d (k+1)=0.4、iq (k+1) = 0.7. The candidate voltage vector is known to be V. 110 V 120 V 122 V 222 The components in the dq coordinate system are respectively

[0187] According to the formula, for each candidate voltage vector V n calculate

[0188] The cost function value g for each candidate voltage vector is calculated using the cost function formula.

[0189] Compare the g values ​​of all candidate voltage vectors. For example:

[0190] g V110 =0.02;

[0191] g V120 =0.015;

[0192] g V122 =0.03;

[0193] g V222 =0.025;

[0194] Select g V120 =0.015 is the minimum value, corresponding to the optimal voltage vector V opt =V 120 .

[0195] The effectiveness of the method of this invention was verified through simulation tests using the MATLAB / Simulink platform. The test scenarios included two conditions: steady-state operation and dynamic speed variation.

[0196] Steady-state operation:

[0197] With the rotational speed set at 500 r / min, the results show that the three-phase current waveform of the method proposed in this invention is smoother, and the total harmonic distortion (THD) of the phase current is 9.2%, which is much lower than the 26.2% of the traditional model predictive control method.

[0198] The ripple of the dq axis current is significantly reduced, and the motor torque fluctuation is effectively suppressed.

[0199] Dynamic changes:

[0200] When the rotational speed changes abruptly from 500 r / min to 1000 r / min, the method of the present invention completes the speed adjustment within 50 ms;

[0201] Simulation results show that the method of the present invention not only inherits the excellent dynamic performance of the traditional method, but also improves the steady-state operation performance.

[0202] S7. Output the inverter switching signal corresponding to the optimal voltage vector and apply it to the inverter.

[0203] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization, characterized in that, The method comprises the following steps: S1, collect three-phase current, rotor position electrical angle and speed information of three-phase permanent magnet synchronous motor, and convert the current to coordinate system through coordinate transformation; S2, according to the previous control period axis voltage vector, predicting the next control period axis current; S3, based on the reference speed of the motor, the speed controller generates a q-axis current reference value, and sets the d-axis current reference value to zero; S4, calculating from the reference current axis reference voltage, and converting the reference voltage into a reference voltage in a stationary coordinate system; In the S4, the following is calculated Axis reference voltage: S4.

1. Calculate the d-axis and q-axis reference currents based on the stator resistance, d-axis inductance and q-axis inductance of the electric machine, combined with the shaft reference current and the electrical angular velocity the shaft reference voltage; S4.2, using the inverse coordinate transformation method, the The axis reference voltage is converted into an α-axis reference voltage and a β-axis reference voltage in the stationary coordinate system. S5, based on the reference voltage, the sector and the quadrant where the candidate voltage vector is located are selected; The selection of the candidate voltage vector in the S5 step comprises the following steps: S5.1, according to the size and direction of the α-axis and β-axis reference voltage in the stationary coordinate system, the sector where the candidate voltage vector is located is determined; S5.2, based on the combination of the virtual voltage vector and the actual voltage vector in the sector, the candidate quadrant is determined; S5.3, the candidate voltage vector includes a zero voltage vector, an active voltage vector, and a virtual voltage vector synthesized by two or more real voltage vectors; S6, calculate the current prediction value under the action of the candidate voltage vector, and select the optimal voltage vector through the cost function; S7, output the inverter switch signal corresponding to the optimal voltage vector and act on the inverter.

2. The permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization according to claim 1, characterized in that, The S1 step specifically comprises the following steps: S1.1, collect three-phase current signals of the three-phase permanent magnet synchronous motor; S1.2, obtain the rotor position electrical angle and speed signal through the motor sensor; S1.3, using coordinate transformation, the collected three-phase current is converted from three-phase stationary coordinate system to axis coordinate system, and output d-axis current and q-axis current.

3. The permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization according to claim 1, characterized in that, The S2 step employs a one-beat delay compensation method for prediction Shaft current, the one-beat delay compensation method includes the following steps: S2.1, based on the motor parameters of the previous control period and the stator voltage vector of the previous control period, predicting the stator voltage vector of the next control period the motor parameters of the current period, predicting the stator voltage vector of the next control period the stator current; S2.2, compensate the d-axis and q-axis current values of the current period by using the motor stator resistance, inductance, electrical angular velocity and permanent magnet flux linkage; S2.3, corrected The shaft current is further predicted by the inductance, voltage vector and motor flux linkage model to obtain the d-axis current and q-axis current of the next control period.

4. The permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization according to claim 1, characterized in that, The S3 step specifically comprises the following steps: S3.1, collect the deviation of the current motor actual speed and the reference speed; S3.2, calculate the integral value and proportional value of the speed error through the speed controller; S3.3, generate a q-axis current reference value according to the calculation result of the speed error; S3.4, set the d-axis current reference value to zero to realize the decoupling control of the motor torque and flux linkage.

5. The permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization according to claim 1, characterized in that, In the S6 step, when calculating the current prediction value under the action of the candidate voltage vector, the following formula is used: d-axis current prediction value: ； q-axis current prediction value: ； wherein, and is the axis component of the candidate voltage vector, is the axis component of the candidate voltage vector, is the sampling period, R is the motor stator resistance, , is the inductance, is the electrical angular velocity, is the permanent magnet flux linkage, and are the d and q axis current values of the k+1 control period, respectively.

6. The permanent magnet motor speed regulation method based on current prediction and discrete space vector optimization according to claim 1, characterized in that, The cost function of the S6 step is as follows: ； in, and for Shaft current reference value, and Under the action of candidate voltage vector Predicted shaft current value.

Images