Vector-variable model predictive control method

By adjusting the voltage vector combination and action time and combining cost function evaluation, the current harmonics and switching frequency are optimized, and the problems of large current harmonics and unfixed switching frequency in traditional control methods are solved, achieving better steady-state performance and low Switching frequency.

CN119030385AInactive Publication Date: 2024-11-26NORTH CHINA ELECTRIC POWER UNIV
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411119913.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Traditional finite state set model predictive control has problems such as large current harmonics and unfixed switching frequency in AC motor control, making it difficult to maintain low switching frequency while improving steady-state performance.

Method used

By adjusting the voltage vector combination and the corresponding vector action time, considering different numbers of voltage vectors and switch jumps during adjacent control cycles, two cost functions are designed to evaluate and determine the optimal solution, and the candidates are optimized to reduce current harmonics. and switching frequency.

Benefits of technology

It effectively improves steady-state performance, maintains a constant low switching frequency, significantly reduces current harmonics and switching frequency, and improves control accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119030385A_ABST
    Figure CN119030385A_ABST
Patent Text Reader

Abstract

The invention provides a variable vector model predictive control method, which comprises the following steps of: A, calculating a reference voltage vector based on a hyper-local model according to a current deadbeat control principle; b, calculating action time of three adjacent vectors for synthesizing the reference voltage vector according to a volt-second balance principle; c, by adjusting a voltage vector combination and corresponding vector action time, considering different numbers of voltage vectors and switch jumps during adjacent control cycles, and obtaining an optimization candidate scheme; and step D, determining an optimal scheme by evaluating the two cost functions, and taking the optimal scheme as the output of the inverter in the next control period. According to the control method provided by the invention, one / two / three / four voltage vectors are considered in one control period, and the problems of high current harmonic, high sampling rate and switching frequency change caused by the fact that traditional finite control set model prediction control only applies one voltage vector in one control period are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of model predictive control, and more specifically, to a variable vector model predictive control method. Background Art

[0002] In the field of AC motor control, model predictive control is known as a new generation of control method after vector control and direct torque control due to its advantages such as simple principle, fast dynamic response, multivariable control and easy handling of nonlinear constraints. Traditional finite state set model predictive control only applies one voltage vector in each control cycle, resulting in higher current harmonics or relying on high sampling rate, and the switching frequency is not fixed.

[0003] Multi-step model predictive control can improve control accuracy and reduce switching frequency, but the amount of calculation increases exponentially with the prediction step. In order to reduce the complexity and amount of calculation of the system, multi-step predictive control based on sphere decoding was proposed, which can greatly reduce the online optimization range, but it is still relatively complex.

[0004] Applying more than one voltage vector in one control cycle can also improve steady-state performance. Dual-vector model predictive control selects the first vector based on the position of the reference voltage vector, and then selects the second vector from two adjacent valid vectors and a zero vector. Model predictive control based on space vector modulation first calculates the reference voltage vector based on current error minimization, and then obtains the final output pulse sequence through space vector modulation. Although multi-vector model predictive control can effectively improve the steady-state performance of the system and reduce torque ripple and current harmonics, it usually comes at the cost of higher switching frequency.

[0005] Therefore, it is necessary to design a control method that can solve the above problems and maintain a constant and low switching frequency while obtaining better steady-state performance. Summary of the invention

[0006] The present invention aims to solve the problem of large current harmonics and unstable switching frequency in the traditional finite control set model predictive control. The present invention provides a variable vector model predictive control method for reducing current harmonics and controlling switching frequency.

[0007] In the present invention, different numbers of voltage vectors and switch jumps between adjacent control cycles are considered, and by adjusting the voltage vector combination and the corresponding vector action time, an optimized candidate vector sequence is obtained in an efficient manner, and two cost functions are designed for evaluation to determine the optimal solution. The control method provided by the present invention considers one / two / three / four voltage vectors in one control cycle, solving the problems of high current harmonics, high sampling rate and switching frequency changes caused by the traditional finite control set model predictive control applying only one voltage vector in one control cycle.

[0008] The present disclosure provides a variable vector model predictive control method, comprising the following steps:

[0009] Step A: According to the current deadbeat control principle, the reference voltage vector is calculated based on the super-local model;

[0010] Step B: According to the volt-second balance principle, the action time of three adjacent vectors used to synthesize the reference voltage vector is calculated;

[0011] Step C: By adjusting the voltage vector combination and the corresponding vector action time, considering different numbers of voltage vectors and switch jumps between adjacent control cycles, an optimized candidate solution is obtained;

[0012] Step D: By evaluating the two cost functions, the optimal solution is determined as the output of the inverter in the next control cycle.

[0013] In some embodiments, step A comprises:

[0014] In the two-phase stationary αβ coordinate system, the mathematical model of the permanent magnet synchronous motor is:

[0015]

[0016] Where E = jω r ψ1 represents the back electromotive force; ω r is the rotor electrical angular velocity; u s ,i s , R s , L, ψ f and θ r They represent stator voltage vector, stator current vector, stator resistance, stator inductance, permanent magnet flux and rotor position respectively; j represents an imaginary unit.

[0017] The current state equation is rewritten as α represents F stands for The stator current at the next sampling instant can be predicted:

[0018]

[0019] in represents the stator current at the kth sampling moment; represents the predicted stator current at the (k+1)th sampling moment; represents the kth control period, i.e., the sampling period; represents the voltage vector applied in the kth control cycle; α k and F k It can be calculated based on the voltage and current of the past two control cycles:

[0020]

[0021] in, represents the stator current at the (k-1)th sampling moment; represents the stator current at the (k-2)th sampling moment; represents the (k-1)th control cycle; represents the (k-2)th control cycle; represents the voltage vector applied in the (k-1)th control cycle; represents the voltage vector applied in the (k-2)th control cycle.

[0022] In actual digital control systems, the program calculation time cannot be ignored, which will lead to a delay between the current sampling moment and the voltage vector action moment. In order to compensate for the adverse effects of this delay, the current value at time (k+1) is estimated first, and then the current value at time (k+2) is predicted with time (k+1) as the starting point:

[0023]

[0024] represents the k+1th control cycle; represents the voltage vector applied in the k+1th control cycle; due to the short control cycle, it can be considered that α k+1 =α k , F k+1 =F k ; The reference current at time (k+2) is calculated as represents the rotor electrical angle at time k, represents the d-axis reference current, represents the q-axis reference current. equal The reference voltage vector of the (k+1)th control cycle is expressed as:

[0025]

[0026] In some embodiments, step B comprises:

[0027] According to the volt-second balance principle, the reference voltage vector is synthesized by three adjacent voltage vectors in one control cycle. When the reference voltage vector is located in the first sector of the space voltage vector diagram, the voltage vectors u1, u2 and u0 (u7) can be synthesized.

[0028]

[0029] T1+T2+T0=T sc

[0030] Among them, T1, T2 and T0 are the action times of u1, u2 and u0 respectively.

[0031] In some embodiments, step C comprises:

[0032] Since the voltage vector at the end of the kth control cycle Probably not One of the three adjacent voltage vectors, according to the volt-second balance principle, It can also be composed of [u6 u7 u2] or [u6 u1 u2], [u3 u0 u1] or [u3 u2 u1];

[0033] Considering the switching transitions between adjacent control cycles, the candidate solutions are determined as:

[0034] 1) Apply a voltage vector in one control cycle:

[0035] The voltage vector acting on the (k+1)th control cycle is the voltage vector at the end of the kth control cycle

[0036] 2) Two voltage vectors are applied in one control cycle:

[0037] The first vector is The second vector is selected from the voltage vectors that only add one switch jump according to the principle of minimizing the current error; according to the current prediction formula and the reference voltage vector calculation formula, minimizing the current error at time (k+2) is equivalent to minimizing the voltage vector error;

[0038] when When u0 is u0, the voltage vectors that only produce one switch jump after u0 are u1, u3 and u5. Obviously, among the voltage vector combinations [u0 u1], [u0 u3] and [u0 u5], [u0 u1] can minimize the voltage error:

[0039]

[0040] Among them, T′0 and T′1 represent the action time of u0 and u1, which are obtained by reconstructing T0 and T1 so that u ε01 minimum; the voltage error equation is reformulated as:

[0041]

[0042] Where k is the proportional factor of the excess action time T2; u ε01 The square of the absolute value is expressed as:

[0043]

[0044] It can be seen that when k is equal to 0.5, |u ε01 | Get the minimum value; therefore, T′0 = T0 + 0.5T2 and T′1 = T1 + 0.5T2;

[0045] 3) Three voltage vectors are applied in one control cycle:

[0046] when When it is u0, u7, u1 or u2, It is composed of u0(u7), u1 and u2; when When u6, the vector combination [u6 u7 u2] or [u6 u1 u2] is applied. Similarly, when When it is u3, the vector combination [u3 u0 u1] or [u3 u2 u1] is applied;

[0047] 4) Four voltage vectors are applied in one control cycle:

[0048] To further improve the steady-state performance, four voltage vectors are acted in one control cycle to not only eliminate the current error but also reduce the RMS value of the harmonic current; in addition to the traditional voltage vector combinations [u0 u1 u2 u7] and [u7 u2 u1 u0], [u0 u1u2 u1], [u1 u2 u1 u0], [u7 u2 u1 u2] and [u2 u1 u2 u7] are introduced.

[0049] In some embodiments, step D comprises:

[0050] After obtaining the optimization candidate solutions, the best solution is determined by evaluating two cost functions; for the first three candidate solutions, the cost function J1 is used for preliminary evaluation:

[0051]

[0052] Among them, N sw is the number of switch transitions in the (k+1)th control cycle, and λ is the weight coefficient;

[0053] If the solution that minimizes J1 is the third candidate, the second cost function J2 is used to further evaluate it and the remaining candidate sequences:

[0054] J2=I h

[0055] Among them, I h It represents the RMS value of harmonic current within a control cycle. The solution that minimizes it is the optimal solution.

[0056] The present invention has the following characteristics and advantages:

[0057] (1) The proposed optimization candidate scheme considers different numbers of voltage vectors within a control cycle and switch transitions between adjacent control cycles.

[0058] (2) By using the reference voltage vector synthesis method based on traditional space vector modulation and non-nearest triangle, the optimization candidate scheme can be determined simply and efficiently.

[0059] (3) It can effectively improve steady-state performance.

[0060] (4) The switching frequency is controllable and fixed. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is the hardware structure diagram of the two-level inverter-fed permanent magnet synchronous motor system;

[0062] Figure 2 It is a variable vector model predictive control block diagram for reducing current harmonics and controlling switching frequency;

[0063] Figure 3 It is the steady-state experimental waveform of the finite state set model predictive control under rated speed and rated load;

[0064] Figure 4 It is a steady-state experimental waveform of model predictive control based on space vector modulation at rated speed and rated load;

[0065] Figure 5 It is the steady-state experimental waveform of the proposed variable vector model predictive control under rated speed and rated load;

[0066] Figure 6 It is a bar chart comparing the total harmonic distortion (THD) of the three control methods, where, in each group of 3 columns, the left column corresponds to the variable vector model predictive control method, the middle column corresponds to the finite state set model predictive control method, and the right column corresponds to the model predictive control method based on space vector modulation.

[0067] Figure 7 It is a schematic flow chart of the variable vector model predictive control method for reducing current harmonics and controlling switching frequency of the present invention. DETAILED DESCRIPTION

[0068] The following embodiments may enable those skilled in the art to more fully understand the present disclosure, but are not intended to limit the present disclosure in any way.

[0069] refer to Figure 7 The present invention provides a variable vector model predictive control method for reducing current harmonics and controlling switching frequency, the steps of which include:

[0070] Step 1: According to the current deadbeat control principle, the reference voltage vector is calculated based on the hyperlocal model;

[0071] Step 2: According to the volt-second balance principle, calculate the action time of three adjacent vectors used to synthesize the reference voltage vector;

[0072] Step 3: By adjusting the voltage vector combination and the corresponding vector action time, considering different numbers of voltage vectors and switch jumps between adjacent control cycles, an optimized candidate solution is obtained;

[0073] Step 4: By evaluating the two cost functions, the optimal solution is determined as the output of the inverter in the next control cycle.

[0074] Figure 1 The hardware circuit structure diagram of the present invention includes a three-phase voltage source, a three-phase diode rectifier bridge, a DC side capacitor, a permanent magnet synchronous motor, a voltage and current sampling circuit, a digital signal processing (DSP) controller and a drive circuit. The voltage and current sampling circuit uses a voltage Hall sensor and a current Hall sensor to respectively collect the DC side voltage and the a and b phase currents of the permanent magnet synchronous motor. The sampling signal enters the DSP controller and is converted into a digital signal after passing through the signal conditioning circuit. The DSP controller completes the calculation of the variable vector model predictive control method of the permanent magnet synchronous motor proposed in the embodiment of the present invention, outputs six switch pulses, and then obtains the final drive signal of the six switch tubes of the inverter after passing through the drive circuit.

[0075] Figure 2 The control principle block diagram of the present invention is shown in FIG. Figure 1 The following steps are implemented on the DSP:

[0076] Step 1: According to the current deadbeat control principle, the reference voltage vector is calculated based on the hyperlocal model.

[0077] In the two-phase stationary αβ coordinate system, the mathematical model of the permanent magnet synchronous motor is:

[0078]

[0079] Where E = jω r ψ1 represents the back electromotive force; ω r is the rotor electrical angular velocity; u s ,i s , R s , L, ψ f and θ r They represent stator voltage vector, stator current vector, stator resistance, stator inductance, permanent magnet flux and rotor position respectively; j represents an imaginary unit.

[0080] The current state equation can be rewritten as α represents F stands for The stator current at the next sampling instant can be predicted:

[0081]

[0082] in represents the stator current at the kth sampling moment; represents the predicted stator current at the (k+1)th sampling moment; represents the kth control period, i.e., the sampling period; represents the voltage vector applied in the kth control cycle; α k and F k It can be calculated based on the voltage and current of the past two control cycles:

[0083]

[0084] in, represents the stator current at the (k-1)th sampling moment; represents the stator current at the (k-2)th sampling moment; represents the (k-1)th control cycle; represents the (k-2)th control cycle; represents the voltage vector applied in the (k-1)th control cycle; represents the voltage vector applied in the (k-2)th control cycle.

[0085] In actual digital control systems, the program calculation time cannot be ignored, which will lead to a delay between the current sampling moment and the voltage vector action moment. In order to compensate for the adverse effects of this delay, the current value at time (k+1) is estimated first, and then the current value at time (k+2) is predicted with time (k+1) as the starting point:

[0086]

[0087] represents the k+1th control cycle; represents the voltage vector applied in the k+1th control cycle; due to the short control cycle, it can be considered that α k+1 =α k , F k+1 =F k ; The reference current at time (k+2) is calculated as represents the rotor electrical angle at time k, represents the d-axis reference current, represents the q-axis reference current. equal The reference voltage vector of the (k+1)th control cycle can be expressed as:

[0088]

[0089] Step 2: According to the volt-second balance principle, calculate the action time of the three adjacent vectors used to synthesize the reference voltage vector.

[0090] According to the volt-second balance principle, the reference voltage vector can be synthesized by three adjacent voltage vectors in one control cycle. For example, when the reference voltage vector is located in the first sector of the space voltage vector diagram, the voltage vectors u1, u2 and u0 (u7) can be synthesized

[0091]

[0092] T1+T2+T0=T sc

[0093] Among them, T1, T2 and T0 are the action time of u1, u2 and u0 (u7) respectively, T sc Indicates the control period.

[0094] Step 3: By adjusting the voltage vector combination and the corresponding vector action time, considering different numbers of voltage vectors and switch jumps between adjacent control cycles, the optimization candidate solution is obtained.

[0095] Since the voltage vector at the end of the kth control cycle Probably not One of the three adjacent voltage vectors, according to the volt-second balance principle, It can also be composed of [u6 u7 u2] or [u6 u1 u2], [u3 u0 u1] or [u3 u2 u1].

[0096] Considering the switching transitions between adjacent control cycles, the candidate solutions can be determined as:

[0097] 1) Apply a voltage vector in one control cycle:

[0098] The voltage vector acting on the (k+1)th control cycle is the voltage vector at the end of the kth control cycle

[0099] 2) Two voltage vectors are applied in one control cycle:

[0100] The first vector is The second vector is selected from the voltage vectors that only add one switch jump according to the principle of minimizing the current error. According to the current prediction formula and the reference voltage vector calculation formula, minimizing the current error at time (k+2) is equivalent to minimizing the voltage vector error.

[0101] For example, when When u0 is u0, the voltage vectors that only produce one switch jump after u0 are u1, u3 and u5. Obviously, among the voltage vector combinations [u0 u1], [u0 u3] and [u0 u5], [u0 u1] can minimize the voltage error:

[0102]

[0103] Among them, T′0 and T′1 represent the action time of u0 and u1, which can be obtained by reconstructing T0 and T1 so that u ε01 Minimum. The voltage error equation can be reformulated as:

[0104]

[0105] Among them, k is the proportional factor of the excess action time T2. ε01 The square of the absolute value is expressed as:

[0106]

[0107] It can be seen that when k is equal to 0.5, |u ε01 | Obtain the minimum value. Therefore, T'0 = T0 + 0.5T2 and T'1 = T1 + 0.5T2.

[0108] 3) Three voltage vectors are applied in one control cycle:

[0109] when When it is u0, u7, u1 or u2, It is composed of u0(u7), u1 and u2. When u6, the vector combination [u6 u7 u2] or [u6 u1 u2] is applied. Similarly, when When u3, the vector combination [u3 u0 u1] or [u3 u2 u1] is applied.

[0110] 4) Four voltage vectors are applied in one control cycle:

[0111] To further improve the steady-state performance, applying four voltage vectors in one control cycle can not only eliminate the current error but also reduce the harmonic current RMS value. In addition to the traditional voltage vector combinations [u0 u1 u2 u7] and [u7 u2 u1u0], [u0 u1 u2 u1], [u1 u2 u1 u0], [u7 u2 u1 u2] and [u2 u1 u2 u7] are introduced.

[0112] Step 4: By evaluating the two cost functions, the optimal solution is determined as the output of the inverter in the next control cycle.

[0113] After obtaining the optimization candidate solutions, the best solution is determined by evaluating two cost functions. For the first three candidate solutions, the cost function J1 is used for preliminary evaluation:

[0114]

[0115] Among them, N sw is the number of switch transitions in the (k+1)th control cycle, and λ is the weight coefficient.

[0116] If the solution that minimizes J1 is the third candidate, the second cost function J2 is used to further evaluate it and the remaining candidate sequences:

[0117] J2=I h

[0118] Among them, I h It represents the RMS value of harmonic current within a control cycle. The solution that minimizes it is the optimal solution.

[0119] The effectiveness of the method proposed in the present invention can be shown by comparing Figure 3 , Figure 4 and Figure 5 The experimental results and observations shown Figure 6 The THD comparison chart shown is obtained. Figure 3 , Figure 4 and Figure 5 The experimental results are all obtained at the same switching frequency. The waveforms from top to bottom are the motor speed, A phase current, switching frequency and space current vector trajectory. It can be seen that the variable vector model predictive control can achieve better steady-state performance. Figure 6 The current THD of the three methods with rated load at different speeds is compared. It can be seen from the figure that compared with the finite state set model predictive control, the variable vector model predictive control shows lower current THD, especially below 65% of the rated speed; the model predictive control based on space vector modulation and the proposed method have similar current THD at medium and low speeds, but the THD of the proposed method is significantly reduced at high speeds, and it has better steady-state performance.

[0120] Those skilled in the art should understand that the above embodiments are merely exemplary embodiments and that various changes, substitutions and alterations may be made without departing from the spirit and scope of the present disclosure.

Claims

1. A variable vector model predictive control method, characterized in that: The steps include: Step A: According to the current deadbeat control principle, the reference voltage vector is calculated based on the super-local model; Step B: According to the volt-second balance principle, the action time of three adjacent vectors used to synthesize the reference voltage vector is calculated; Step C: By adjusting the voltage vector combination and the corresponding vector action time, considering different numbers of voltage vectors and switch jumps between adjacent control cycles, an optimized candidate solution is obtained; Step D: By evaluating the two cost functions, the optimal solution is determined as the output of the inverter in the next control cycle.

2. The variable vector model predictive control method according to claim 1, characterized in that: The step A comprises: In the two-phase stationary αβ coordinate system, the mathematical model of the permanent magnet synchronous motor is: Where E = jω r ψ1 represents the back electromotive force; ω r is the rotor electrical angular velocity; u s ,i s , R s , L, ψ f and θ r They represent stator voltage vector, stator current vector, stator resistance, stator inductance, permanent magnet flux and rotor position respectively; j represents an imaginary unit. The current state equation is rewritten as α represents F stands for The stator current at the next sampling instant can be predicted: in represents the stator current at the kth sampling moment; represents the predicted stator current at the (k+1)th sampling moment; represents the kth control period, i.e., the sampling period; represents the voltage vector applied in the kth control cycle; α k and F k It can be calculated based on the voltage and current of the past two control cycles: in, represents the stator current at the (k-1)th sampling moment; represents the stator current at the (k-2)th sampling moment; represents the (k-1)th control cycle; represents the (k-2)th control cycle; represents the voltage vector applied in the (k-1)th control cycle; represents the voltage vector applied in the (k-2)th control cycle; In actual digital control systems, the program calculation time cannot be ignored, which will lead to a delay between the current sampling moment and the voltage vector action moment. In order to compensate for the adverse effects of this delay, the current value at time (k+1) is estimated first, and then the current value at time (k+2) is predicted with time (k+1) as the starting point: represents the k+1th control cycle; represents the voltage vector applied in the k+1th control cycle; due to the short control cycle, it is considered that α k+1 =α k , F k+1 =F k ; The reference current at time (k+2) is calculated as represents the rotor electrical angle at time k, represents the d-axis reference current, represents the q-axis reference current; let equal The reference voltage vector of the (k+1)th control cycle is expressed as:

3. The variable vector model predictive control method according to claim 1, characterized in that: The step B comprises: According to the volt-second balance principle, the reference voltage vector is synthesized by three adjacent voltage vectors in one control cycle. When the reference voltage vector is located in the first sector of the space voltage vector diagram, the voltage vectors u1, u2 and u0 (u7) can be synthesized. T1+T2+T0=T sc Among them, T1, T2 and T0 are the action times of u1, u2 and u0 respectively.

4. The variable vector model predictive control method according to claim 1, characterized in that: The step C comprises: Since the voltage vector at the end of the kth control cycle Probably not One of the three adjacent voltage vectors, according to the volt-second balance principle, It can also be composed of [u6u7u2] or [u6u1u2], [u3u0u1] or [u3u2u1]; Considering the switching transitions between adjacent control cycles, the candidate solutions are determined as: 1) Apply a voltage vector in one control cycle: The voltage vector acting on the (k+1)th control cycle is the voltage vector at the end of the kth control cycle 2) Two voltage vectors are applied in one control cycle: The first vector is The second vector is selected from the voltage vectors that only add one switch jump according to the principle of minimizing the current error; according to the current prediction formula and the reference voltage vector calculation formula, minimizing the current error at time (k+2) is equivalent to minimizing the voltage vector error; when When u0 is u0, the voltage vectors that only produce one switch jump after u0 are u1, u3 and u5; obviously, among the voltage vector combinations [u0u1], [u0u3] and [u0u5], [u0u1] can minimize the voltage error: Among them, T′0 and T′1 represent the action time of u0 and u1, which are obtained by reconstructing T0 and T1 so that u ε01 minimum; the voltage error equation is reformulated as: u ε01 T sc =u1T1+u2T2-u0T’0-u1T’1 =u1T1+u2T2-u0(T0+kT2)-u1(T1+(1-k)T2) =T2(u2+(k-1)u1) Where k is the proportional factor of the excess action time T2; u ε01 The square of the absolute value is expressed as: It can be seen that when k is equal to 0.5, |u ε01 | Get the minimum value; therefore, T′0 = T0 + 0.5T2 and T′1 = T1 + 0.5T2; 3) Three voltage vectors are applied in one control cycle: when When it is u0, u7, u1 or u2, It is composed of u0(u7), u1 and u2; when When u6, the vector combination [u6 u7u2] or [u6 u1 u2] is applied. Similarly, when When it is u3, the vector combination [u3 u0 u1] or [u3 u2 u1] is applied; 4) Four voltage vectors are applied in one control cycle: To further improve the steady-state performance, four voltage vectors are applied in one control cycle to not only eliminate the current error but also reduce the RMS value of the harmonic current; in addition to the traditional voltage vector combinations [u0 u1 u2 u7] and [u7 u2 u1 u0], [u0 u1 u2u1], [u1 u2 u1 u0], [u7 u2 u1 u2] and [u2 u1 u2 u7] are introduced.

5. The variable vector model predictive control method according to claim 1, characterized in that: The step D comprises: After obtaining the optimization candidate solutions, the best solution is determined by evaluating two cost functions; for the first three candidate solutions, the cost function J1 is used for preliminary evaluation: Among them, N sw is the number of switch transitions in the (k+1)th control cycle, and λ is the weight coefficient; If the solution that minimizes J1 is the third candidate, the second cost function J2 is used to further evaluate it and the remaining candidate sequences: J2=I h Among them, I h It represents the RMS value of harmonic current within a control cycle. The solution that minimizes it is the optimal solution.

Citation Information

Patent Citations

  • Six-phase permanent-magnet synchronous motor model predictive control method

    CN105162380A

  • Double-three-level inverter open winding motor model control method

    CN111541411A

  • Dead-beat predictive current control method for permanent magnet synchronous motor

    CN115276501A

  • Vector-variable sequence induction motor model-free predictive current control method

    CN116073713A