Hybrid predictive control method for single-stage multi-port inverter

Abstract

The invention discloses a hybrid predictive control method for a single-stage multi-port inverter, and the method comprises the steps: firstly building an equivalent two-level model of the single-stage multi-port inverter, and building a port power model of the single-stage multi-port inverter based on the equivalent two-level model; filtering the power required by the load through a high-pass filter, solving a low-frequency power component required by the load, then determining a linear power distribution range of the single-stage multi-port inverter, and finally determining linear regulation factors lambda1 and lambda2 according to the low-frequency power component and the linear power distribution range. And voltage vector distribution of the sub-inverter 1 and the sub-inverter 2 is realized through linear regulation factors lambda1 and lambda2, so that power control of the single-stage multi-port inverter is completed.

Description

Technical Field

[0001] The present invention belongs to the technical field of motor control, and more particularly, relates to a hybrid predictive control method for a single-stage multi-port inverter. Background Art

[0002] With the increasing pressure of fossil fuel depletion and environmental issues, electric vehicles are becoming more and more popular in the transportation industry. Among various types of electric vehicles, hybrid electric vehicles, which combine high power density sources with high energy density sources, have become the preferred choice due to their advanced features in terms of initial investment, efficiency, source life, and reliability.

[0003] Among the various power electronics interfaces for hybrid electric vehicles (HEVs), single-stage multiport inverters have attracted considerable interest due to their high power density with minimal passive filtering. Despite these advantages, the elimination of the DC-DC converter reduces the control freedom of single-stage multiport inverters, requiring simultaneous control of the AC-side stator current and DC-side power distribution. The limited control freedom of the multi-objective problem complicates the controller design process. Furthermore, the time-varying input voltage of the single-stage multiport inverter introduced by the direct energy connection results in an asymmetric voltage vector diagram and increases nonlinearity in the controller design. Among various control methods, model predictive control (MPC) has been widely adopted for power converters due to its intuitive design process, high dynamic performance, and ability to incorporate system constraints and nonlinearities by evaluating cost functions for discrete switching states. Although existing MPC schemes can handle the multi-objective control and unbalanced DC port issues in single-stage multiport inverters, the discrete control mechanism introduces significant power ripple on the DC port, which is undesirable in lifetime-sensitive applications. Furthermore, the complex weighting factor adjustment process is undesirable in practical applications. This makes MPC algorithms difficult to adapt to the development needs of new energy vehicles. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a hybrid predictive control method for a single-stage multi-port inverter, which realizes stator current control on the AC side and port power control on the DC side, and the control process is simpler and the power ripple is small.

[0005] To achieve the above object, the present invention provides a hybrid predictive control method for a single-stage multi-port inverter, characterized by comprising the following steps:

[0006] (1) Establish an equivalent two-level model of a single-stage multi-port inverter and record the active voltage vector used for motor drive control;

[0007] (2) Establish a port power model of a single-stage multi-port inverter based on an equivalent two-level model;

[0008] (3) Filter the power required by the load through a high-pass filter to obtain the low-frequency power component required by the load;

[0009] (4) Determine the linear power distribution range of the single-stage multi-port inverter;

[0010] (5) Determine the linear adjustment factors λ1 and λ2 according to the low-frequency power component and the linear power distribution range;

[0011] (6) The voltage vector distribution of sub-inverter 1 and sub-inverter 2 is realized through linear adjustment factors λ1 and λ2, thereby completing the power control of the single-stage multi-port inverter.

[0012] The object of the invention of the present invention is achieved like this:

[0013] The present invention discloses a hybrid predictive control method for a single-stage multi-port inverter. The method first establishes an equivalent two-level model of the single-stage multi-port inverter, and then establishes a port power model of the single-stage multi-port inverter based on the equivalent two-level model. The method then filters the power required by the load through a high-pass filter to obtain the low-frequency power component required by the load, and then determines the linear power distribution range of the single-stage multi-port inverter. Finally, linear adjustment factors λ1 and λ2 are determined based on the low-frequency power component and the linear power distribution range. The voltage vector distribution of sub-inverter 1 and sub-inverter 2 is implemented by the linear adjustment factors λ1 and λ2, thereby completing the power control of the single-stage multi-port inverter.

[0014] At the same time, the hybrid predictive control method for a single-stage multi-port inverter of the present invention also has the following beneficial effects:

[0015] (1) Using a single-stage multi-port motor driver in the motor drive system of a hybrid electric vehicle can improve the power density and energy transmission efficiency of the hybrid electric vehicle.

[0016] (1) It avoids the weight factor adjustment problem existing in traditional model predictive control schemes in hybrid electric vehicles.

[0017] (2) An equivalent two-level SSMPI is established as a single-source inverter, and the stator current regulation of the motor is achieved by directly calculating the optimal voltage vector, which greatly reduces the iteration time and calculation amount of MPC.

[0018] (3) While ensuring stable motor operation, the linear adjustment factor is directly solved based on the linear port power model, which can achieve fast and accurate control of the port power. The control process is simple and the power ripple is small. In addition, the current quality of the system is also significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a control diagram of a hybrid predictive control method based on a single-stage multi-port inverter of the present invention;

[0020] Figure 2 It is the topology diagram of a single-stage multi-port inverter;

[0021] Figure 3 This is a diagram of different current paths in a single-stage multi-port inverter;

[0022] Figure 4 It is a schematic diagram of the three-level model and the equivalent two-level model;

[0023] Figure 5 It is a linear power distribution diagram with switching time constraints; (a) interval I (b) interval II;

[0024] Figure 6 It is the linear power distribution interval diagram of single-stage multi-port inverter;

[0025] Figure 7 The steady-state performance of the hybrid predictive control scheme and the traditional model predictive control scheme under unbalanced DC link; (a) hybrid predictive control scheme (b) traditional model predictive control scheme

[0026] Figure 8 The steady-state performance of the hybrid predictive control scheme and the traditional model predictive control scheme at different port powers; (a) hybrid predictive control scheme (b) traditional model predictive control scheme

[0027] Figure 9 The current quality of the hybrid predictive control scheme and the traditional model predictive control scheme; (a) hybrid predictive control scheme (b) traditional model predictive control scheme

[0028] Figure 10 The dynamic performance of the hybrid predictive control scheme and the traditional model predictive control scheme under load torque step change; (a) hybrid predictive control scheme (b) traditional model predictive control scheme

[0029] Figure 11 The transient performance of the hybrid predictive control scheme and the traditional model predictive control scheme under speed reversal conditions; (a) hybrid predictive control scheme (b) traditional model predictive control scheme. DETAILED DESCRIPTION

[0030] The following describes the specific embodiments of the present invention in conjunction with the accompanying drawings so that those skilled in the art can better understand the present invention. It should be noted that in the following description, when detailed descriptions of known functions and designs may dilute the main content of the present invention, such descriptions will be omitted here.

[0031] Example

[0032] In this embodiment, if Figure 2 As shown, the single-stage multi-port inverter includes two DC power supplies and three bridge arms, each bridge arm includes four switching tubes and two diodes; the two DC power supplies are respectively the first DC power supply and the second DC power supply, and the voltage of the first DC power supply is greater than the voltage of the second DC power supply; the four switching tubes are respectively the first switching tube Sx1, the second switching tube Sx2, the third switching tube Sx3 and the fourth switching tube Sx4, x = a, b, c represents three phases; wherein, the first switching tube is complementary to the third switching tube, that is, when the first switching tube is turned on, the third switching tube is turned off. Similarly, the second switching tube is complementary to the fourth switching tube. The two diodes are respectively the first diode Dx1 and the second diode Dx2. As shown Figure 3 As shown in Table 1, different output states can be obtained by controlling the switches. H, L, and 0 in the conversion state refer to the high level, low level, and zero level corresponding to the switch tube. As shown in Table 1, when the first switch tube Sx1 and the second switch tube Sx2 are turned on at the same time, the output voltage of the system is V H ,like Figure 3 As shown in (a), the power supply of the system is the first DC power supply. Figure 3 As shown in (b), when the first switch tube Sx1 is turned off and the second switch tube Sx2 is turned on at the same time, the output voltage of the system is V L , at this time the system power supply is the second DC power supply. Figure 3 As shown in (c), when the first switch tube Sx1 and the second switch tube Sx2 are turned off at the same time, the output voltage of the system is 0, and the system has no power supply.

[0033] Table 1 Different state diagrams of control switches

[0034]

[0035] like Figure 1 As shown, the present invention provides a hybrid predictive control method for a single-stage multi-port inverter, comprising the following steps:

[0036] S1. Establish an equivalent two-level model of a single-stage multi-port inverter and record the active voltage vector used for motor drive control;

[0037] In this embodiment, if Figure 4 Through equivalent modeling, 27 vectors can be simplified into 6 active voltage vectors {U1~U6} and 2 zero vectors {U0,U7};

[0038] S2. Establish a port power model of a single-stage multi-port inverter;

[0039] S2.1. Collect the three-phase current i on the AC side of the single-stage multi-port inverter x(k), the three-phase current i x After T1 coordinate transformation, the dq axis component i is obtained d (k), i q (k), k is the sampling time;

[0040] The three-phase AC motor obtains the current position angle θ(k) through the position encoding module, and then the speed calculation module calculates the actual speed ω of the three-phase AC motor e (k);

[0041] S2.2, traverse the adjacent active voltage vectors of the equivalent two-level model, superimpose the two adjacent active voltage vectors and perform T1 coordinate transformation to obtain the dq axis voltage component u d 、u q ;

[0042] In this embodiment, if Figure 4 As shown, among the six active voltage vectors, U1 and U2 are a group of adjacent vectors, U3 and U4 are a group of adjacent vectors, and U5 and U6 are a group of adjacent vectors, for a total of three groups of adjacent vectors.

[0043] S2.3, obtain the predicted current under different voltage vectors through a two-step prediction model;

[0044] will u d 、u q As the pressure component u at the current k moment d (k),u q (k), predict the current components of the dq axis at time k+1:

[0045]

[0046] Among them, L d and L q Represent the d-axis and q-axis inductances, ψ f is the rotor flux, R s is the stator resistance, T s is the sampling period;

[0047] Calculate the voltage component u of the dq axis at time k+1 d (k+1),u q (k+1):

[0048]

[0049] Predict the current components of the dq axis at time k+2:

[0050]

[0051] S2.4. Calculation of i d (k+2), iq (k+2) and the given reference value After all three groups of adjacent vectors are processed, the difference is used as the loss value, and the two adjacent active voltage vectors v1 and v2 with the smallest difference are selected as the optimal voltage vectors for motor drive control;

[0052] S2.5. Calculate the current errors of the d-axis and q-axis;

[0053] S2.5.1. Calculate the current slopes of the active voltage vectors v1 and v2 on the d-axis and q-axis at the current moment;

[0054]

[0055] Where, i=1,2, when i=1, it represents voltage vector v1, when i=2, it represents voltage vector v2; s d,i 、s q,i Indicates the current slope of the d-axis and q-axis; s d,i (k), s q,i (k) is the current slope of the dq axis, ω(k) is the angular velocity of the three-phase AC motor;

[0056] S2.5.2. Calculate the current errors of the active voltage vectors v1 and v2 on the d-axis and q-axis;

[0057]

[0058] in, s d,z (k), s q,z (k) is the current slope of the zero vector on the dq axis, t1 and t2 are the durations of the active voltage vectors v1 and v2, respectively. z is the duration of the zero vector;

[0059] S2.6. Calculate the optimal voltage reference vector v * ;

[0060] S2.6.1. Set the duration of active voltage vectors v1, v2 and zero vector to meet the following conditions:

[0061] t1+t2+t z =T s

[0062] S2.6.2. Construct the cost function J of the motor drive control;

[0063]

[0064] S2.6.3. With the goal of minimizing the cost function J, the least squares method is used to solve the cost function J, namely:

[0065]

[0066] By solving the duration t1, t2 of the active voltage vectors v1, v2, the optimal voltage reference vector v is calculated. * ;

[0067]

[0068] S2.7, according to the optimal voltage reference vector v * To analyze the power boundary of the system and establish the port power model of the system;

[0069] S2.7.1. Decouple the single-stage multi-port inverter into two two-level inverters, denoted as sub-inverter 1 and sub-inverter 2, where the voltage reference vector of sub-inverter 1 is denoted as The voltage reference vector of sub-inverter 2 is recorded as

[0070] S2.7.2. Set the switching times of the two sub-inverters and divide the power distribution range of the single-stage multi-port inverter into interval I and interval II, such as Figure 5 As shown, (a) is interval I, (b) is interval II; then the voltage vector reference v * Proportionally allocated to sub-inverter 1 And sub-inverter 2

[0071] In interval I: the switching times of sub-inverter 1 and sub-inverter 2 are set to 1 and 2 respectively, and the required voltage reference vector and for:

[0072]

[0073] Among them, λ1 is the linear adjustment factor;

[0074] The port power model of a single-stage multi-port inverter is established as:

[0075]

[0076] Among them, P max is the maximum output power of the single-stage multi-port inverter, P b is the boundary power between interval I and interval II, is the conjugate of the stator current vector;

[0077] In this embodiment, based on the above port power model, it can be seen that the output power P of the sub-inverter 2 is sub2 for: The maximum output power of the single-stage multi-port inverter in this range is that all power is output from the lower port, and the boundary power is that the voltage vector v2 is entirely provided by sub-inverter 2.

[0078] In interval II: the switching times of sub-inverter 1 and sub-inverter 2 are set to 2 and 1 respectively, and the required voltage vector and for:

[0079]

[0080] The port power model of a single-stage multi-port inverter is established as:

[0081]

[0082] Among them, P min is the minimum output power of the single-stage multi-port inverter

[0083] In this embodiment, based on the above port power model, it can be seen that the minimum output power of the single-stage multi-port inverter is 0, and the boundary power is that the voltage vector v2 is entirely provided by the sub-inverter 2.

[0084] S3. Filter the power required by the load through a high-pass filter to obtain the low-frequency power component required by the load;

[0085] In this embodiment, the load demand includes low-frequency components and high-frequency components. The high-frequency components are fed into the fast-acting power supply, and the rest are fed into the slow-acting power supply. The high-frequency components are signals with a frequency of 500 Hz or higher, because the power electronic devices (such as inverters) in electric vehicles usually operate within this frequency range. The low-frequency components are signals with a frequency of 20 Hz or lower, because these signals usually involve the low-speed operation and starting of electric vehicles. Fast-acting power supply and slow-acting power supply refer to the speed of power supply response. Power supplies with smaller internal resistance, such as supercapacitors, can be classified as fast-acting power supplies, while power supplies with larger internal resistance, such as batteries, can be classified as slow-acting power supplies.

[0086] S4. determining a linear power distribution range of the single-stage multi-port inverter;

[0087] S4.1. When the three-phase AC motor is operating normally, and The angle difference between them is between 0 and 60°, at this time v * The following conditions must be met:

[0088]

[0089] Among them, v H , v LRespectively represent the high-voltage port voltage and the low-voltage port voltage of the single-stage multi-port inverter;

[0090] S4.2. Define power regulation coefficient Among them, P * is the reference total power required by the load, is the reference low-frequency power required by the load;

[0091] S4.3. Determine the distribution range of the power regulation coefficient;

[0092] S4.3.1. Determine the lower limit of the power regulation coefficient η;

[0093] When the voltage vector Reaching the maximum value means satisfying:

[0094]

[0095] At this time, if the single-stage multi-port inverter is powered only by the main power supply, the lower limit of the power regulation coefficient η is:

[0096]

[0097] If the single-stage multi-port inverter is powered by the main power supply and the auxiliary power supply, the lower limit of the power regulation coefficient η is:

[0098]

[0099] S4.3.2. Determine the upper limit of the power regulation coefficient η;

[0100] When the voltage vector Reaching the maximum value means satisfying:

[0101]

[0102] At this time, if the single-stage multi-port inverter is powered only by the auxiliary power supply, the upper limit of the power regulation coefficient η is:

[0103]

[0104] If a single-stage multi-port inverter is powered by a main power supply and an auxiliary power supply, the upper limit of the power regulation coefficient η is:

[0105]

[0106] S4.3.3. In summary, the power regulation coefficient η range is:

[0107]

[0108] In this embodiment, the linear power distribution range of the single-stage multi-port inverter is as follows: Figure 6 As shown in FIG, by using the established port power model and power regulation coefficient to build the linear power distribution range of the system, the linear power distribution range of the single-stage multi-port inverter is constructed.

[0109] S5. Determine linear adjustment factors λ1 and λ2 according to the low-frequency power component and the linear power distribution range;

[0110] S5.1. Based on the reference low-frequency power Calculate the linear adjustment factor λ1;

[0111]

[0112] S5.2. Calculate the low-frequency power component P of the load sub2 ;

[0113]

[0114] S5.3. Determine P sub2 Is the ratio of / P within the range of the power regulation coefficient η? If not, it means that the load power cannot be distributed and the algorithm ends; otherwise, continue to calculate the linear regulation factor λ2:

[0115]

[0116] S6. The voltage vector distribution of sub-inverter 1 and sub-inverter 2 is realized through linear adjustment factors λ1 and λ2, thereby completing the power control of the single-stage multi-port inverter.

[0117] To further illustrate the technical solutions and effects of the present invention, a single-stage multi-port inverter-type three-phase AC motor, shown in Table 2, is used as an example. This motor was fabricated in a laboratory. The voltage of the second DC power supply was set between 125V and 175V to simulate the source voltage variations encountered in real-world applications, while the voltage of the first DC power supply was set to 300V. Two programmable DC sources were used as hybrid energy sources: the first simulated a slow-acting source and the second simulated a fast-acting source, with the fast-acting source and the slow-acting source corresponding to the speed of dynamic response, respectively. A dSPACE MicroLabBox DS1202 microcontroller was used for digital control, while a Xilinx slave device generated the gate signals for each switch in the multi-port inverter, enabling digital-to-analog and analog-to-digital conversion. The single-stage multi-port inverter was implemented using three insulated gate bipolar transistor power modules (Infineon F3L75R07W2E3). A 5000-pulse incremental encoder was used to acquire rotor speed and position. A commercial software driven induction motor, connected to a three-phase AC motor, is used as a load machine, which can be operated in torque mode to generate the required load torque.

[0118] Table 2 Parameters of the preferred single-stage multi-port inverter three-phase AC motor of the present invention

[0119]

[0120]

[0121] Firstly, the proposed hybrid predictive control scheme and traditional model predictive control scheme were tested under different voltage ratios. Figure 7 The control performance of the two control schemes at rated torque and rated speed under different voltage ratios is shown, where the first DC power supply voltage is set to 300V and the second DC power supply voltage is set to 125 to 175V. The experimental results show that both schemes can achieve stator current control and DC side power distribution regulation under different voltage ratios. Figure 7 As shown in (a), when the second DC power supply voltage is set to 125V, the proposed hybrid predictive control scheme shows ripple-free tracking performance, while the traditional model predictive control experiences large port power ripple. Figure 7 As shown in Figure (b), when the second DC power supply voltage is set to 175V, the hybrid predictive control scheme still outperforms the traditional model predictive control scheme. Furthermore, because the distribution control objective can lead to suboptimal control inputs for stator current regulation, stator current distortion in traditional model predictive control is avoided by hybrid predictive control. Experimental verification demonstrates that the present invention is effective under conditions of DC port voltage imbalance and offers significant performance improvements compared to traditional schemes.

[0122] In order to further confirm the effectiveness and advantages of the hybrid predictive control scheme, the port power allocation capabilities of the two control schemes are studied. Figure 8 (a) and Figure 8 (b) represents the lower port power The power distribution capabilities of the two schemes at rated speed and rated load torque are set to 300W and -300W operating conditions. Both schemes can achieve the required power distribution between the DC ports. However, the proposed hybrid predictive control scheme shows a ripple-free power distribution, while the traditional model predictive control scheme shows excessive ripple. The main reason is that due to the decoupled control structure in the hybrid predictive control scheme, the power distribution control is decoupled from the stator current control, while these two control objectives are interconnected through weighting factors in the traditional model predictive control scheme. In addition, by performing a linear comparison of the predicted boundary power in the proposed hybrid predictive control scheme, accurate power distribution can be achieved through vector decomposition, while only discrete power distribution points are available in the traditional model predictive control scheme due to its discrete switching state evaluation mechanism.

[0123] In addition, the Fast Fourier Transform (FFT) analysis of the stator current further confirms these results. Figure 9 As shown in (a), the stator current total harmonic distortion (THD) of the proposed hybrid predictive control scheme is 2.87%, while Figure 9 (b) indicates that the THD of the conventional model predictive control scheme is 7.80%. The hybrid predictive control scheme shows a significant improvement over the conventional model predictive control scheme. In addition, the current harmonics of the proposed hybrid predictive control scheme are concentrated at 20kHz, indicating that a constant switching frequency is achieved. To highlight the benefits of the proposed hybrid predictive control scheme, the quantitative analysis is summarized in Table 1. In terms of computational burden, considering SSMPI as a single-source inverter in the hybrid predictive control scheme, 6 iteration times and cost function evaluation times are required to obtain the optimal adjacent active voltage vector. In this process, i should be predicted at each iteration. d (k+1) and i q (k+1). In addition, P max and P b To achieve the power distribution. Therefore, the total number of predictors in the hybrid predictive control scheme is 14. However, all feasible switching states should be evaluated in the traditional model predictive control scheme. In each iteration, i d (k+1), i q (k+1) and P sub2 Therefore, in the traditional model predictive control scheme, 27 iteration times and cost function evaluation time are required to obtain the optimal voltage vector. In addition, in the traditional model predictive control scheme, the total number of predictor variables is 81. Therefore, the hybrid predictive control scheme performs better in terms of computational burden.

[0124] Table 3 Comparison of computational burden between hybrid predictive control scheme and traditional model predictive control scheme

[0125] plan Number of iterations Number of predictions Number of evaluations Computational burden Traditional MPC 27 81 27 25.69μs HPC 6 14 6 24.16μs

[0126] Finally, the dynamic performance of the system is studied to verify the effectiveness of the hybrid predictive control scheme. Figure 10 The figure shows the changes of external load torque from 4 to 8 Nm and from 8 to 4 Nm under rated speed conditions in two schemes. Figure 10 As shown in (a), the proposed hybrid predictive control scheme shows fast current response and stability when the external load of the system changes, and the output power of the slow power supply remains unchanged. This shows the effectiveness of the proposed scheme under variable load torque. Figure 10 (b) It can be seen that although the traditional model predictive control scheme can also achieve the corresponding function, its excessive ripple will limit its practical application. This once again highlights the superiority of the hybrid predictive control scheme over the traditional model predictive control scheme.

[0127] Figure 11 The speed reversal test of two schemes at rated torque is shown. In the test, the speed reference value is set to 500rpm and then changed step by step to -500rpm. Figure 11 As shown in (a), the rotor speed can be adjusted from forward to reverse by the proposed hybrid predictive control scheme. In addition, when the rotor speed is reversed, the stator current control performance can be guaranteed due to the decoupling control structure, and the output power of the power supply can be tracked at the power reference value after transient adjustment. However, when the rotor speed is reversed by adopting the traditional model predictive control scheme, the performance of the stator current control and the power distribution control affect each other, as shown in Figure 11 (b) Due to the multi-control objective mechanism of the traditional model predictive control scheme, trade-offs are made through weighting factor adjustment, resulting in large ripple in the q-axis current and port power ripple. Therefore, this scheme can operate under dynamic power distribution conditions and achieves better control results than the traditional model predictive control scheme.

[0128] Although the above describes the illustrative specific embodiments of the present invention to facilitate understanding of the present invention by those skilled in the art, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concepts of the present invention are protected.

Claims

1. A hybrid predictive control method for a single-stage multi-port inverter, characterized in that: The following steps are involved: (1) Establish an equivalent two-level model of a single-stage multi-port inverter and record the active voltage vector used for motor drive control; (2) Establish a port power model of a single-stage multi-port inverter based on an equivalent two-level model; (3) Filter the power required by the load through a high-pass filter to obtain the low-frequency power component required by the load; (4) Determine the linear power distribution range of the single-stage multi-port inverter; (5) Determine the linear adjustment factors λ1 and λ2 according to the low-frequency power component and the linear power distribution range; (6) The voltage vector distribution of sub-inverter 1 and sub-inverter 2 is realized through the linear adjustment factors λ1 and λ2, thereby completing the power control of the single-stage multi-port inverter.

2. The hybrid predictive control method for a single-stage multi-port inverter according to claim 1, wherein: The method for establishing the port power model of the single-stage multi-port inverter is as follows: (2.1) Collect the three-phase current i on the AC side of the single-stage multi-port inverter x (k), the three-phase current i x After T1 coordinate transformation, the dq axis component i is obtained d (k), i q (k), k is the sampling time; The three-phase AC motor obtains the current position angle θ(k) through the position encoding module, and then the speed calculation module calculates the actual speed ω of the three-phase AC motor e (k); (2.2) Traverse the adjacent active voltage vectors of the equivalent two-level model, superimpose the two adjacent active voltage vectors and perform T1 coordinate transformation to obtain the dq axis voltage component u d 、u q ; ( 2.3) Obtain the predicted current under different voltage vectors through a two-step prediction model; will u d 、u q As the pressure component u at the current k moment d (k),u q (k), predict the current components of the dq axis at time k+1: Among them, L d and L q Represent the d-axis and q-axis inductances, ψ f is the rotor flux, R s is the stator resistance, T s is the sampling period; Calculate the voltage component u of the dq axis at time k+1 d (k+1),u q (k+1): Predict the current components of the dq axis at time k+2: (2.4), calculate i d (k+2), i q (k+2) and the given reference value The difference between the two adjacent active voltage vectors v1 and v2 with the smallest difference is selected as the optimal voltage vector for motor drive control; (2.5), calculate the current error of d-axis and q-axis; (2.5.1) Calculate the current slopes of the active voltage vectors v1 and v2 on the d-axis and q-axis at the current moment; Where, i=1,2, when i=1, it represents voltage vector v1, when i=2, it represents voltage vector v2; s d,i 、s q,i Indicates the current slope of the d-axis and q-axis; s d,i (k), s q,i (k) is the current slope of the dq axis, ω(k) is the angular velocity of the three-phase AC motor; (2.5.2) Calculate the current errors of the active voltage vectors v1 and v2 on the d-axis and q-axis; in, s d,z (k), s q,z (k) is the current slope of the zero vector on the dq axis, t1 and t2 are the durations of the active voltage vectors v1 and v2, respectively. z is the duration of the zero vector; (2.6), calculate the optimal voltage reference vector v * ; (2.6.1) Set the duration of active voltage vectors v1, v2 and zero vector to meet the following conditions: t1+t2+t z =T s (2.6.2) Construct the cost function J of the motor drive control; (2.6.3) With the goal of minimizing the cost function J, the least squares method is used to solve the cost function J, namely: By solving the duration t1, t2 of the active voltage vectors v1, v2, the optimal voltage reference vector v is calculated. * ; (2.7), according to the optimal voltage reference vector v * To analyze the power boundary of the system and establish the port power model of the system; (2.7.1) Decouple the single-stage multi-port inverter into two two-level inverters, denoted as sub-inverter 1 and sub-inverter 2, where the voltage reference vector of sub-inverter 1 is denoted as The voltage reference vector of sub-inverter 2 is recorded as (2.7.2) Set the switching times of the two sub-inverters, divide the power distribution range of the single-stage multi-port inverter into interval I and interval II, and set the voltage vector reference v in different intervals. * Proportionally allocated to sub-inverter 1 And sub-inverter 2 In interval I: the switching times of sub-inverter 1 and sub-inverter 2 are set to 1 and 2 respectively, and the required voltage reference vector and for: Among them, λ1 is the linear adjustment factor; The port power model of a single-stage multi-port inverter is established as: Among them, P max is the maximum output power of the single-stage multi-port inverter, P b is the boundary power between interval I and interval II, is the conjugate of the stator current vector; In interval II: the switching times of sub-inverter 1 and sub-inverter 2 are set to 2 and 1 respectively, and the required voltage vector and for: The port power model of a single-stage multi-port inverter is established as: Among them, P min is the minimum output power of a single-stage multi-port inverter.

3. The hybrid predictive control method for a single-stage multi-port inverter according to claim 1, wherein: The linear power distribution range of the single-stage multi-port inverter is: (3.1) When the three-phase AC motor operates normally, v * The following conditions must be met: Among them, v H , v L Respectively represent the high-voltage port voltage and the low-voltage port voltage of the single-stage multi-port inverter; (3.2) Define the power regulation coefficient Among them, P * is the reference total power required by the load, is the reference low-frequency power required by the load; (3.3) Determine the distribution range of the power regulation coefficient; (3.3.1) Determine the lower limit of the power regulation coefficient η; When the voltage vector Reaching the maximum value means satisfying: At this time, if the single-stage multi-port inverter is powered only by the main power supply, the lower limit of the power regulation coefficient η is: If the single-stage multi-port inverter is powered by the main power supply and the auxiliary power supply, the lower limit of the power regulation coefficient η is: (3.3.2) Determine the upper limit of the power regulation coefficient η; When the voltage vector Reaching the maximum value means satisfying: At this time, if the single-stage multi-port inverter is powered only by the auxiliary power supply, the upper limit of the power regulation coefficient η is: If a single-stage multi-port inverter is powered by a main power supply and an auxiliary power supply, the upper limit of the power regulation coefficient η is: (3.3.3) In summary, the power regulation coefficient η range is:

4. The hybrid predictive control method for a single-stage multi-port inverter according to claim 1, wherein: The method for determining the linear adjustment factors λ1 and λ2 is: (4.1), according to the reference low-frequency power Calculate the linear adjustment factor λ1; (4.2), calculate the low-frequency power component P of the load sub2 ; (4.3), judge P sub2 Is the ratio of / P within the range of the power regulation coefficient η? If not, it means that the load power cannot be distributed and the algorithm ends; otherwise, continue to calculate the linear regulation factor λ2:

Images