Model predictive torque control method and system for high-speed train lift wing converter

Abstract

The application provides a high-speed train lift wing variable frequency converter model predictive torque control method and system, which comprises the following steps: obtaining real-time motor data of a driving motor in the lift wing at a current sampling time; predicting a target stator flux and a target motor torque required by the driving motor at a next sampling time; dividing a four-level space voltage vector diagram into six vector regions according to six maximum voltage vectors in the four-level space voltage vector diagram; combining the target stator flux and the target motor torque to construct a basic cost function; using the basic cost function to perform rolling optimization on the six maximum voltage vectors; using an advanced cost function to perform rolling optimization on all voltage vectors in a target vector region to obtain an optimal voltage vector in the target vector region; and controlling the motor torque of the driving motor through the optimal voltage vector. The application has the effect of improving the real-time performance and calculation efficiency of the torque control system during torque control.

Description

Technical Field

[0001] This invention belongs to the field of train control technology, specifically relating to a model predictive torque control method and system for a high-speed train lift wing converter. Background Technology

[0002] The high-speed train lifting wing is a novel structure designed to increase train speed and reduce axle mass. It provides lift to the train by controlling the angle of attack. The lifting wing structure consists of a lifting and angle-of-attack conversion section, a bottom rotating section, and a tail retraction section. Its angle-of-attack conversion component is driven by a permanent magnet synchronous motor (PMSM). Multilevel active neutral point clamped (ANPC) inverters have advantages such as low voltage stress on power devices, low current harmonics, and adjustable device losses, and are widely used in industrial drives. Among them, three-level ANPC inverters have been more widely used in traditional industrial drives.

[0003] In practical applications, to achieve precise angle-of-attack control, existing technologies employ model predictive torque control algorithms, thus avoiding complex modulation processes. This algorithm uses a cost function that considers stator flux linkage, electromagnetic torque, and midpoint voltage control. It utilizes this cost function to perform rolling optimization on the voltage vectors corresponding to all switching states of the three-level ANPC inverter, thereby calculating the current optimal voltage vector. This optimal voltage vector is then output to the three-level ANPC inverter to control the drive motor to output the optimal torque.

[0004] However, with the continuous development of industrial technology, the usage rate of four-level ANPC inverters is constantly increasing. Compared with three-level ANPC inverters, four-level ANPC inverters require the same number of components but can provide more voltage levels and lower current harmonics, which is beneficial for the lightweighting of the lift wing control system and improving system reliability. Model predictive control has advantages such as fast response and strong robustness, and can effectively overcome process uncertainties, nonlinearities, and parallelism, and can easily add various constraints. However, compared with two-level or three-level inverters, four-level inverters have more switching states, and correspondingly more voltage vectors. Traditional model predictive torque control algorithms need to perform rolling optimization on all voltage vectors during the calculation process, resulting in a huge computational load for torque control systems, which in turn reduces the real-time performance and computational efficiency of the torque control system. Summary of the Invention

[0005] This invention provides a model predictive torque control method and system for a high-speed train lifting wing converter, in order to solve the problem that the torque control system has a very large computational load during torque control, which leads to a reduction in the real-time performance and computational efficiency of the torque control system.

[0006] In a first aspect, the present invention provides a model predictive torque control method for a high-speed train lift wing converter, the method comprising the following steps:

[0007] Acquire real-time motor data of the drive motor in the lifting wing at the current sampling time. The real-time motor data includes motor stator current, motor speed, and motor inductance.

[0008] Based on the motor stator current, motor speed, and motor inductance, predict the target stator flux and target motor torque required by the drive motor at the next sampling time.

[0009] A four-level space voltage vector diagram is constructed based on the switching state of the inverter in the lifting wing, wherein the inverter is a four-level active midpoint clamped inverter.

[0010] The four-level space voltage vector diagram is divided into six vector regions based on the six maximum voltage vectors in the diagram, with the maximum voltage vectors located at the vertices of the diagram.

[0011] The basic cost function is constructed by combining the target stator flux linkage and the target motor torque;

[0012] The six maximum voltage vectors are optimized by rolling search using the basic cost function. The maximum voltage vector that minimizes the basic cost function is taken as the target voltage vector, and the vector region corresponding to the target voltage vector is marked as the target vector region.

[0013] Collect the midpoint current and DC bus voltage of the inverter at the current sampling time;

[0014] By combining the midpoint current and the DC bus voltage, the basic cost function is adjusted to obtain the adjusted advanced cost function;

[0015] The advanced cost function is used to perform rolling optimization on all voltage vectors in the target vector region to obtain the optimal voltage vector in the target vector region;

[0016] The motor torque of the drive motor is controlled by the optimal voltage vector.

[0017] Optionally, the step of combining the motor stator current, motor speed, and motor inductance to predict the target stator flux and target motor torque required by the drive motor at the next sampling time includes the following steps:

[0018] The motor voltage of the drive motor at the current sampling time is calculated by combining the motor stator current, motor speed, and motor inductance.

[0019] Based on the motor voltage and using the Euler method, predict the stator current of the drive motor at the next sampling time;

[0020] The target stator flux required by the drive motor at the next sampling moment is calculated by combining the predicted stator current of the motor and the motor inductance.

[0021] The target motor torque required by the drive motor at the next sampling moment is calculated by combining the target stator flux linkage and the predicted stator current of the motor.

[0022] Optionally, the expression for the basic cost function is as follows:

[0023]

[0024] In the formula: g1 represents the basic cost function, λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. This represents the target motor torque required by the drive motor at the next sampling time.

[0025] Optionally, adjusting the basic cost function by combining the midpoint current and the DC bus voltage to obtain the adjusted advanced cost function includes the following steps:

[0026] The three DC capacitor voltages in the inverter are calculated based on the midpoint current, and the three predicted DC capacitor voltages of the inverter at the next sampling time are obtained by discretizing the three DC capacitor voltages.

[0027] The nominal value of the midpoint voltage of the inverter is determined based on the DC bus voltage.

[0028] The predicted neutral point voltage of the inverter is calculated by combining the three predicted DC capacitor voltages and the nominal value of the neutral point voltage.

[0029] The base cost function is adjusted using the predicted midpoint voltage to obtain the adjusted advanced cost function.

[0030] Optionally, the expression for the advanced cost function is as follows:

[0031]

[0032] in,

[0033] ΔU C =|U C1 (k+1)-U dc / 3|+|U C2 (k+1)-U dc / 3|+|U C3 (k+1)-U dc / 3|

[0034] In the formula, g2 represents the advanced cost function, λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. λ represents the target motor torque required by the drive motor at the next sampling time. dc The DC capacitor voltage error, ΔU C U represents the predicted midpoint voltage. C1 (k+1), U C2 (k+1), U C3 (k+1) represent the three predicted DC capacitor voltages of the inverter at the next sampling time, U dc U represents the DC bus voltage. dc / 3 represents the nominal value of the midpoint voltage.

[0035] Secondly, the present invention also provides a model predictive torque control system for a high-speed train lift wing converter, the system comprising:

[0036] The motor data acquisition module is used to acquire real-time motor data of the drive motor in the lifting wing at the current sampling time. The real-time motor data includes motor stator current, motor speed and motor inductance.

[0037] The torque prediction module is used to combine the motor stator current, motor speed and motor inductance to predict the target stator flux and target motor torque required by the drive motor at the next sampling time.

[0038] The vector diagram construction module is used to construct a four-level space voltage vector diagram based on the switching state of the inverter in the lifting wing, wherein the inverter is a four-level active midpoint clamped inverter.

[0039] The vector map region division module is used to divide the four-level spatial voltage vector map into six vector regions based on the six maximum voltage vectors in the four-level spatial voltage vector map, wherein the maximum voltage vectors are located at the vertices of the four-level spatial voltage vector map.

[0040] The cost function construction module is used to construct a basic cost function by combining the target stator flux and the target motor torque;

[0041] The region selection module is used to perform rolling optimization on the six maximum voltage vectors using the basic cost function, and to take the maximum voltage vector that makes the basic cost function minimum as the target voltage vector, and to mark the vector region corresponding to the target voltage vector as the target vector region.

[0042] The inverter data acquisition module is used to acquire the midpoint current and DC bus voltage of the inverter at the current sampling time.

[0043] The function adjustment module is used to adjust the basic cost function by combining the midpoint current and the DC bus voltage to obtain the adjusted advanced cost function.

[0044] The vector selection module is used to perform rolling optimization on all voltage vectors in the target vector region using the advanced cost function to obtain the optimal voltage vector in the target vector region.

[0045] A torque control module is used to control the motor torque of the drive motor through the optimal voltage vector.

[0046] Optionally, the torque prediction module includes:

[0047] A voltage calculation unit is used to calculate the motor voltage of the drive motor at the current sampling time by combining the motor stator current, motor speed and motor inductance.

[0048] A current prediction unit is used to predict the stator current of the drive motor at the next sampling time based on the motor voltage and using the Euler method.

[0049] The flux linkage prediction unit is used to calculate the target stator flux linkage required by the drive motor at the next sampling time by combining the predicted stator current of the motor and the motor inductance.

[0050] The torque prediction unit calculates the target motor torque required by the drive motor at the next sampling moment by combining the target stator flux linkage and the predicted stator current of the motor.

[0051] Optionally, the expression for the basic cost function is as follows:

[0052]

[0053] In the formula: g1 represents the basic cost function, λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. This represents the target motor torque required by the drive motor at the next sampling time.

[0054] Optionally, the function adjustment module includes:

[0055] A capacitor voltage prediction unit is used to calculate the three DC capacitor voltages in the inverter based on the midpoint current, and to obtain the three predicted DC capacitor voltages of the inverter at the next sampling time by discretizing the three DC capacitor voltages.

[0056] A calibration unit is used to determine the nominal value of the midpoint voltage of the inverter based on the DC bus voltage.

[0057] The midpoint voltage prediction unit is used to calculate the predicted midpoint voltage of the inverter by combining the three predicted DC capacitor voltages and the nominal value of the midpoint voltage.

[0058] The function adjustment unit is used to adjust the basic cost function using the predicted midpoint voltage to obtain the adjusted advanced cost function.

[0059] Optionally, the expression for the advanced cost function is as follows:

[0060]

[0061] in,

[0062] ΔU C =|U C1 (k+1)-U dc / 3|+|U C2 (k+1)-U dc / 3|+|U C3 (k+1)-U dc / 3|

[0063] In the formula, g2 represents the advanced cost function, λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. λ represents the target motor torque required by the drive motor at the next sampling time. dc The DC capacitor voltage error, ΔU C U represents the predicted midpoint voltage. C1 (k+1), U C2 (k+1), U C3 (k+1) represent the three predicted DC capacitor voltages of the inverter at the next sampling time, U dc U represents the DC bus voltage. dc / 3 represents the nominal value of the midpoint voltage.

[0064] The beneficial effects of this invention are:

[0065] First, the stator current, speed, and inductance of the drive motor in the lifting wing are obtained at the current sampling time. Based on these parameters, the target stator flux linkage and target motor torque required by the drive motor at the next sampling time are predicted. Then, a basic cost function is constructed using these parameters. Since the four-level active midpoint clamped inverter has 64 switching states, the constructed four-level spatial voltage vector diagram is primarily a hexagonal vector diagram. Therefore, the six vertices of the hexagonal vector diagram represent the corresponding maximum voltage vectors. The basic cost function is used to perform rolling optimization on the six maximum voltage vectors to find the target vector region containing the optimal voltage vector. Then, the basic cost function is adjusted to an advanced cost function, and rolling optimization is performed within the target vector region using the advanced cost function to ultimately find the optimal voltage vector. Compared to using a cost function to perform rolling optimization on all voltage vectors, using different cost functions for two progressive rolling optimizations can greatly reduce the computational load of the rolling optimization process, thereby improving the real-time performance and computational efficiency of the torque control system during torque control. Attached Figure Description

[0066] Figure 1 This is a schematic flowchart of the predictive torque control method for a high-speed train lift wing converter model in this invention.

[0067] Figure 2 This is a four-level space voltage vector diagram in this invention.

[0068] Figure 3 This is a schematic diagram of one of the vector regions in the four-level space voltage vector diagram of the present invention.

[0069] Figure 4This is a schematic diagram of the topology of the four-level ANPC inverter in this invention.

[0070] Figure 5 This is a circuit diagram illustrating the effect of the inverter midpoint current on the DC capacitor voltage in this invention. Detailed Implementation

[0071] This invention discloses a model predictive torque control method for a high-speed train lift wing converter.

[0072] Reference Figure 1 The predictive torque control method for high-speed train lift wing converter models specifically includes the following steps:

[0073] S101. Obtain the real-time motor data of the drive motor in the lifting wing at the current sampling time.

[0074] The lifting wing structure consists of an elevation and angle-of-attack conversion section, a bottom rotating section, and a tail retraction section. Its angle-of-attack conversion component is driven by a permanent magnet synchronous motor (PMSM). Therefore, in this embodiment, the drive motor can be a permanent magnet synchronous motor. Real-time motor data of the drive motor is acquired by sensors installed in the lifting wing. Specifically, the real-time motor data includes the motor stator current, motor speed, and motor inductance.

[0075] S102. Combining the motor stator current, motor speed, and motor inductance, predict the target stator flux and target motor torque required to drive the motor at the next sampling time.

[0076] First, the PMSM voltage equation and flux linkage equation in the synchronous rotating coordinate system are constructed. Then, the motor voltage at the current sampling moment is calculated by combining the motor stator current, motor speed and motor inductance. The motor voltage calculation equation is discretized to obtain the stator current prediction equation at the next sampling moment. Then, the target stator flux linkage and target motor torque required by the drive motor at the next sampling moment can be predicted based on the flux linkage equation.

[0077] S103. Construct a four-level space voltage vector diagram based on the switching state of the inverter in the lifting wing. The inverter is a four-level active midpoint clamped inverter.

[0078] For a four-level ANPC inverter, it is assumed that the neutral point voltage (NP) can be controlled to the inverter DC bus voltage U. dc If 1 / 3 is used, then each phase arm of the inverter can generate four voltage levels relative to the midpoint, namely U... dc 2U dc / 3、U dc / 3 and 0. (Refer to...) Figure 2The four-level active neutral clamp inverter generates a total of 64 switching states. In the diagram, "3", "2", "1" and "0" represent the inverter output level U, respectively. dc 2U dc / 3、U dc / 3 and 0, as shown in Table 1.

[0079] Table 1 Switching Status of Four-Level ANPC Inverter

[0080]

[0081]

[0082] The 64 switching states of a four-level active neutral-point clamped inverter correspond to 37 voltage vectors (36 effective voltage vectors and 1 zero voltage vector). As shown in Table 2, the 36 effective voltage vectors can be divided into three categories: Large Voltage Vector (LVV), Medium Voltage Vector (MVV), and Small Voltage Vector (SVV).

[0083] Table 2 Voltage Vector Types

[0084]

[0085] Reference Figure 4 The switching function of each phase of a four-level ANPC inverter can be expressed as:

[0086]

[0087] The AC voltage measurement of the inverter can be expressed using the switching function and the DC bus voltage as follows:

[0088] U xn =S x / 3×U dc

[0089] The inverter output voltage vector can be expressed as:

[0090]

[0091] S104. Divide the four-level space voltage vector diagram into six vector regions based on the six maximum voltage vectors in the four-level space voltage vector diagram.

[0092] Since the four-level active neutral-point clamped inverter has 64 switching states, the constructed four-level space voltage vector diagram is mainly a hexagonal vector diagram. Therefore, the six vertices of the hexagonal vector diagram represent the corresponding maximum voltage vectors. Figure 2For example, Figure 2 The maximum voltage vectors in the range are 033, 003, 303, 300, 330, and 030. (Refer to...) Figure 3 The vector region is divided based on the maximum voltage vector and the voltage vectors adjacent to the maximum voltage vector. Figure 3 What is shown is one of the divided vector regions.

[0093] S105. Construct a basic cost function by combining the target stator flux linkage and the target motor torque.

[0094] Among them, reference Figure 4 , Figure 4 This is a topology diagram of a four-level ANPC inverter. For a four-level ANPC inverter, controlling the midpoint potential is necessary; therefore, the midpoint potential usually needs to be considered in the cost function. When the inverter output level is 2U... dc / 3 and U dc At / 3, the load current flows out from the neutral points O1 and O2 respectively. Therefore, the midpoint current in one sampling period can be expressed as:

[0095]

[0096] Among them, i NP1 and i NP2 Let O1 and O2 represent the output currents at neutral points, respectively. Assuming the DC capacitors have the same capacitance (C1 = C2 = C3), the distribution of the midpoint current between the DC capacitors is as follows: Figure 5 As shown in the figure. The relationship between the midpoint current and the voltages of the three capacitors can be expressed as:

[0097]

[0098] In the formula, T s Representing the sampling period, discretizing the above equation yields:

[0099]

[0100] Reference Figure 5 As shown in Table 2, MVV and SVV have redundant voltage vectors. Although they can produce the same output voltage, their effects on the midpoint differ. For example, voltage vectors 010, 121, and 232 all belong to SVV and can output the same voltage vector, but from... Figure 5 It can be seen that the three switching states produce different midpoint currents. However, in this embodiment, the LVV does not affect the midpoint voltage, so the midpoint voltage can be ignored in the basic cost function.

[0101] The expression for the basic cost function is as follows:

[0102]

[0103] In the formula: g1 represents the basic cost function, λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. This represents the target motor torque required to drive the motor at the next sampling time.

[0104] S106. Use the basic cost function to perform rolling optimization on the 6 maximum voltage vectors, and take the maximum voltage vector that makes the basic cost function minimum as the target voltage vector, and mark the vector region corresponding to the target voltage vector as the target vector region.

[0105] Among them, reference Figure 2 Based on Table 2, using the six LVVs (300, 330, 030, 033, 003, 303) located at the vertices of the hexagon as a finite control set, the target vector region can be determined according to the following formula:

[0106] (2N-3)π / 6≤S N ≤(2N-1)π / 6

[0107] Where N represents the sector number, N = 1, 2, ..., 6.

[0108] For example, if the optimal voltage vector is 300, then choose as follows: Figure 3 The S1 vector region shown has only 13 switching states in each vector region, including 12 effective voltage vector switching states and 1 zero vector switching state. Therefore, when performing rolling optimization on all voltage vectors in the vector region, the overall computational load is reduced, which is beneficial to improving computational efficiency.

[0109] S107. Collect the inverter's midpoint current and DC bus voltage at the current sampling time.

[0110] S108. By combining the midpoint current and DC bus voltage to adjust the basic cost function, the adjusted advanced cost function is obtained.

[0111] Since each vector region includes LVV, MVV, SVV and zero vector, and as can be seen from the detailed description in step S105, MVV and SVV have an impact on the midpoint voltage, the advanced cost function needs to consider the midpoint voltage term. The predicted midpoint voltage can be calculated by the midpoint current and DC bus voltage, thereby adjusting the basic cost function.

[0112] S109. Use the advanced cost function to perform rolling optimization on all voltage vectors in the target vector region to obtain the optimal voltage vector in the target vector region.

[0113] S110. The motor torque of the drive motor is controlled by the optimal voltage vector.

[0114] The optimal voltage vector is input into the inverter, which then outputs the corresponding three-phase current to the drive motor, thereby controlling the motor torque of the starting motor.

[0115] The implementation principle of this method is as follows:

[0116] First, the stator current, speed, and inductance of the drive motor in the lifting wing are obtained at the current sampling time. Based on these parameters, the target stator flux linkage and target motor torque required by the drive motor at the next sampling time are predicted. Then, a basic cost function is constructed using these parameters. Since the four-level active midpoint clamped inverter has 64 switching states, the constructed four-level spatial voltage vector diagram is primarily a hexagonal vector diagram. Therefore, the six vertices of the hexagonal vector diagram represent the corresponding maximum voltage vectors. The basic cost function is used to perform rolling optimization on the six maximum voltage vectors to find the target vector region containing the optimal voltage vector. Then, the basic cost function is adjusted to an advanced cost function, and rolling optimization is performed within the target vector region using the advanced cost function to ultimately find the optimal voltage vector. Compared to using a cost function to perform rolling optimization on all voltage vectors, using different cost functions for two progressive rolling optimizations can greatly reduce the computational load of the rolling optimization process, thereby improving the real-time performance and computational efficiency of the torque control system during torque control.

[0117] In one embodiment, step S102, which combines the motor stator current, motor speed, and motor inductance to predict the target stator flux linkage and target motor torque required to drive the motor at the next sampling time, specifically includes the following steps:

[0118] The motor voltage at the current sampling moment is calculated by combining the motor stator current, motor speed, and motor inductance.

[0119] Based on the motor voltage and using the Euler method, predict the stator current of the drive motor at the next sampling time;

[0120] The target stator flux required by the drive motor at the next sampling time is calculated by combining the predicted stator current and motor inductance.

[0121] The target motor torque required by the drive motor at the next sampling time is calculated by combining the target stator flux linkage and the predicted stator current of the motor.

[0122] In this embodiment, the motor voltage is the stator voltage in the drive motor. The basic motor voltage equation and flux linkage equation of the drive motor in the synchronous rotating coordinate system can be constructed. The basic motor voltage equation is as follows:

[0123]

[0124] In the formula, u d and u q i represents the component of the motor voltage in the synchronous rotating coordinate system. d and i q R represents the component of the motor stator current in the synchronous rotating coordinate system. s ω represents the stator resistance in the drive motor. e Ψ represents the motor speed of the drive motor, or it can be expressed as the angular velocity of the drive motor. d and Ψ q This represents the component of the stator flux linkage in the synchronous rotating coordinate system of the drive motor.

[0125] The specific equations for magnetic flux linkage are as follows:

[0126]

[0127] In the formula, L d and L q Ψ represents the component of the motor inductance in the synchronous rotating coordinate system. f This refers to the magnetic flux linkage of permanent magnets in a drive motor.

[0128] Substituting the flux linkage equation into the basic motor voltage equation yields the final motor voltage equation, as follows:

[0129]

[0130] Therefore, the motor voltage at the current sampling moment is calculated by using the final motor voltage equation and combining it with the motor stator current, motor speed and motor inductance.

[0131] By performing a forward Euler calculation on the final motor voltage equation, the predicted stator current equation is obtained, as follows:

[0132]

[0133] In the formula, T s Let k be the sampling period and k+1 represent the kT-th sampling period, respectively. s With (k+1)T sThe sampling time refers to both the current sampling time and the next sampling time. Therefore, the formula for calculating the target stator flux linkage required by the drive motor at the next sampling time, combining the predicted stator current and motor inductance, is as follows:

[0134]

[0135] In the formula, and This represents the component of the target stator flux linkage in the synchronous rotating coordinate system.

[0136] The formula for calculating the target motor torque required by the drive motor at the next sampling time, combining the target stator flux linkage and the predicted stator current of the motor, is as follows:

[0137]

[0138] In the formula, n represents the target motor torque. p This indicates the number of pole pairs of the drive motor.

[0139] In one implementation, step S108, which involves adjusting the basic cost function by combining the midpoint current and the DC bus voltage to obtain the adjusted advanced cost function, specifically includes the following steps:

[0140] The voltages of the three DC capacitors in the inverter are calculated based on the midpoint current, and the three predicted DC capacitor voltages of the inverter at the next sampling time are obtained by discretizing the three DC capacitor voltages.

[0141] Determine the nominal value of the inverter's midpoint voltage based on the DC bus voltage;

[0142] The predicted neutral point voltage of the inverter is calculated by combining the three predicted DC capacitor voltages and the nominal value of the neutral point voltage.

[0143] By adjusting the basic cost function using the predicted midpoint voltage, the adjusted advanced cost function is obtained.

[0144] In this embodiment, the midpoint current in one sampling period can be expressed as:

[0145]

[0146] Among them, i NP1 and i NP2 Let O1 and O2 represent the output currents at neutral points, respectively. Assuming the DC capacitors have the same capacitance (C1 = C2 = C3), the distribution of the midpoint current between the DC capacitors is as follows: Figure 5 As shown. The relationship between the midpoint current and the voltages of the three DC capacitors can be expressed as:

[0147]

[0148] In the formula, ΔU C1 ΔU C2 ΔU C3 T represents the voltage of three DC capacitors. s Representing the sampling period, discretizing the above equation yields:

[0149]

[0150] In the formula, U C1 (k+1), U C2 (k+1), U C3 (k+1) represents the three predicted DC capacitor voltages.

[0151] The expression for the advanced cost function is as follows:

[0152]

[0153] The formula for calculating the predicted neutral point voltage of the inverter by combining the three predicted DC capacitor voltages and the nominal value of the neutral point voltage is as follows:

[0154] ΔU C =|U C1 (k+1)-U dc / 3|+|U C2 (k+1)-U dc / 3|+|U C3 (k+1)-U dc / 3|

[0155] In the formula, g2 represents the advanced cost function, and λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. λ represents the target motor torque required to drive the motor at the next sampling time. dc The DC capacitor voltage error, ΔU C U represents the predicted midpoint voltage. C1 (k+1), U C2 (k+1), U C3 (k+1) represent the three predicted DC capacitor voltages of the inverter at the next sampling time, U dc U represents the DC bus voltage. dc / 3 represents the nominal value of the midpoint voltage.

[0156] This invention also discloses a model predictive torque control system for a high-speed train lift wing converter, the system specifically comprising:

[0157] The motor data acquisition module is used to acquire real-time motor data of the drive motor in the lifting wing at the current sampling time. The real-time motor data includes motor stator current, motor speed and motor inductance.

[0158] The torque prediction module is used to combine the motor stator current, motor speed and motor inductance to predict the target stator flux and target motor torque required to drive the motor at the next sampling time.

[0159] The vector diagram construction module is used to construct a four-level space voltage vector diagram based on the switching state of the inverter in the lifting wing. The inverter is a four-level active midpoint clamped inverter.

[0160] The vector map region division module is used to divide the four-level space voltage vector map into six vector regions based on the six maximum voltage vectors in the four-level space voltage vector map. The maximum voltage vectors are located at the vertices of the four-level space voltage vector map.

[0161] The cost function construction module is used to construct the basic cost function by combining the target stator flux and the target motor torque;

[0162] The region selection module is used to perform rolling optimization on the six maximum voltage vectors using the basic cost function. The maximum voltage vector that minimizes the basic cost function is taken as the target voltage vector, and the vector region corresponding to the target voltage vector is marked as the target vector region.

[0163] The inverter data acquisition module is used to acquire the inverter's midpoint current and DC bus voltage at the current sampling moment;

[0164] The function adjustment module is used to adjust the basic cost function by combining the midpoint current and DC bus voltage to obtain the adjusted advanced cost function.

[0165] The vector selection module is used to perform rolling optimization on all voltage vectors in the target vector region using an advanced cost function to obtain the optimal voltage vector in the target vector region.

[0166] The torque control module is used to control the motor torque of the drive motor through optimal voltage vector control.

[0167] In this embodiment, the expression for the basic cost function is as follows:

[0168]

[0169] In the formula: g1 represents the basic cost function, λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. This represents the target motor torque required to drive the motor at the next sampling time.

[0170] In one embodiment, the torque prediction module includes:

[0171] The voltage calculation unit is used to calculate the motor voltage of the drive motor at the current sampling moment by combining the motor stator current, motor speed and motor inductance.

[0172] The current prediction unit is used to predict the stator current of the drive motor at the next sampling time based on the motor voltage and using the Euler method.

[0173] The flux prediction unit is used to calculate the target stator flux required by the drive motor at the next sampling time by combining the predicted stator current and motor inductance.

[0174] The torque prediction unit calculates the target motor torque required by the drive motor at the next sampling moment by combining the target stator flux linkage and the predicted stator current of the motor.

[0175] In one implementation, the function adjustment module includes:

[0176] The capacitor voltage prediction unit is used to calculate the three DC capacitor voltages in the inverter based on the midpoint current, and to obtain the three predicted DC capacitor voltages of the inverter at the next sampling time by discretizing the three DC capacitor voltages.

[0177] The calibration unit is used to determine the nominal value of the inverter's midpoint voltage based on the DC bus voltage.

[0178] The midpoint voltage prediction unit is used to calculate the predicted midpoint voltage of the inverter by combining the three predicted DC capacitor voltages and the nominal midpoint voltage value.

[0179] The function adjustment unit is used to adjust the basic cost function using the predicted midpoint voltage to obtain the adjusted advanced cost function.

[0180] In this embodiment, the expression for the advanced cost function is as follows:

[0181]

[0182] in,

[0183] ΔU C =|U C1(k+1)-U dc / 3|+|U C2 (k+1)-U dc / 3|+|U C3 (k+1)-U dc / 3|

[0184] In the formula, g2 represents the advanced cost function, and λ Ψ Indicates flux linkage error. This represents the stator flux linkage reference value, k represents the current sampling time, and k+1 represents the next sampling time. λ represents the target stator flux required by the drive motor at the next sampling time. T Indicates torque error. Table of motor torque reference values. λ represents the target motor torque required to drive the motor at the next sampling time. dc The DC capacitor voltage error, ΔU C U represents the predicted midpoint voltage. C1 (k+1), U C2 (k+1), U C3 (k+1) represent the three predicted DC capacitor voltages of the inverter at the next sampling time, U dc U represents the DC bus voltage. dc / 3 represents the nominal value of the midpoint voltage.

[0185] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; under the concept of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0186] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A model predictive torque control method for a high-speed train lift wing converter, characterized in that, Includes the following steps: Acquire real-time motor data of the drive motor in the lifting wing at the current sampling time. The real-time motor data includes motor stator current, motor speed, and motor inductance. Based on the motor stator current, motor speed, and motor inductance, predict the target stator flux and target motor torque required by the drive motor at the next sampling time. A four-level space voltage vector diagram is constructed based on the switching state of the inverter in the lifting wing, wherein the inverter is a four-level active midpoint clamped inverter. The four-level space voltage vector diagram is divided into six vector regions based on the six maximum voltage vectors in the diagram, with the maximum voltage vectors located at the vertices of the diagram. A basic cost function is constructed by combining the target stator flux linkage and the target motor torque. The expression of the basic cost function is as follows: In the formula: This represents the basic cost function. Indicates flux linkage error. Indicates the stator flux linkage reference value. Indicates the current sampling time. Indicates the next sampling time. This indicates the target stator flux required by the drive motor at the next sampling time. Indicates torque error. Table of motor torque reference values. This indicates the target motor torque required by the drive motor at the next sampling time. The six maximum voltage vectors are optimized by rolling search using the basic cost function. The maximum voltage vector that minimizes the basic cost function is taken as the target voltage vector, and the vector region corresponding to the target voltage vector is marked as the target vector region. Collect the midpoint current and DC bus voltage of the inverter at the current sampling time; The basic cost function is adjusted by combining the midpoint current and the DC bus voltage to obtain the adjusted advanced cost function. This includes: calculating the three DC capacitor voltages in the inverter based on the midpoint current, and obtaining three predicted DC capacitor voltages of the inverter at the next sampling time by discretizing the three DC capacitor voltages; determining the nominal value of the inverter's midpoint voltage based on the DC bus voltage; calculating the predicted midpoint voltage of the inverter by combining the three predicted DC capacitor voltages and the nominal midpoint voltage value; and adjusting the basic cost function using the predicted midpoint voltage to obtain the adjusted advanced cost function. The expression of the advanced cost function is as follows: in, In the formula, This represents the advanced cost function. Indicates flux linkage error. Indicates the stator flux linkage reference value. Indicates the current sampling time. Indicates the next sampling time. This indicates the target stator flux required by the drive motor at the next sampling time. Indicates torque error. Table of motor torque reference values. This represents the target motor torque required by the drive motor at the next sampling time. This indicates the DC capacitor voltage error. This represents the predicted midpoint voltage. , , These represent the three predicted DC capacitor voltages of the inverter at the next sampling time. This represents the DC bus voltage. This indicates the nominal value of the midpoint voltage; The advanced cost function is used to perform rolling optimization on all voltage vectors in the target vector region to obtain the optimal voltage vector in the target vector region; The motor torque of the drive motor is controlled by the optimal voltage vector.

2. The high-speed train lift wing converter model predictive torque control method according to claim 1, characterized in that, The step of combining the motor stator current, motor speed, and motor inductance to predict the target stator flux and target motor torque required by the drive motor at the next sampling time includes the following steps: The motor voltage of the drive motor at the current sampling time is calculated by combining the motor stator current, motor speed, and motor inductance. Based on the motor voltage and using the Euler method, predict the stator current of the drive motor at the next sampling time; The target stator flux required by the drive motor at the next sampling moment is calculated by combining the predicted stator current of the motor and the motor inductance. The target motor torque required by the drive motor at the next sampling moment is calculated by combining the target stator flux linkage and the predicted stator current of the motor.

3. A model predictive torque control system for a high-speed train lift wing converter, characterized in that, The system includes: The motor data acquisition module is used to acquire real-time motor data of the drive motor in the lifting wing at the current sampling time. The real-time motor data includes motor stator current, motor speed and motor inductance. The torque prediction module is used to combine the motor stator current, motor speed and motor inductance to predict the target stator flux and target motor torque required by the drive motor at the next sampling time. The vector diagram construction module is used to construct a four-level space voltage vector diagram based on the switching state of the inverter in the lifting wing, wherein the inverter is a four-level active midpoint clamped inverter. The vector map region division module is used to divide the four-level spatial voltage vector map into six vector regions based on the six maximum voltage vectors in the four-level spatial voltage vector map, wherein the maximum voltage vectors are located at the vertices of the four-level spatial voltage vector map. The cost function construction module is used to construct a basic cost function by combining the target stator flux linkage and the target motor torque. The expression of the basic cost function is as follows: In the formula: This represents the basic cost function. Indicates flux linkage error. Indicates the stator flux linkage reference value. Indicates the current sampling time. Indicates the next sampling time. This indicates the target stator flux required by the drive motor at the next sampling time. Indicates torque error. Table of motor torque reference values. This indicates the target motor torque required by the drive motor at the next sampling time. The region selection module is used to perform rolling optimization on the six maximum voltage vectors using the basic cost function, and to take the maximum voltage vector that makes the basic cost function minimum as the target voltage vector, and to mark the vector region corresponding to the target voltage vector as the target vector region. The inverter data acquisition module is used to acquire the midpoint current and DC bus voltage of the inverter at the current sampling time. The function adjustment module is used to adjust the basic cost function by combining the midpoint current and the DC bus voltage to obtain the adjusted advanced cost function. The function adjustment module includes: a capacitor voltage prediction unit, used to calculate the three DC capacitor voltages in the inverter based on the midpoint current, and to obtain three predicted DC capacitor voltages of the inverter at the next sampling time by discretizing the three DC capacitor voltages; a calibration unit, used to determine the nominal value of the midpoint voltage of the inverter based on the DC bus voltage; a midpoint voltage prediction unit, used to calculate the predicted midpoint voltage of the inverter by combining the three predicted DC capacitor voltages and the nominal value of the midpoint voltage; and a function adjustment unit, used to adjust the basic cost function using the predicted midpoint voltage to obtain the adjusted advanced cost function, the expression of which is as follows: in, In the formula, This represents the advanced cost function. Indicates flux linkage error. Indicates the stator flux linkage reference value. Indicates the current sampling time. Indicates the next sampling time. This indicates the target stator flux required by the drive motor at the next sampling time. Indicates torque error. Table of motor torque reference values. This represents the target motor torque required by the drive motor at the next sampling time. This indicates the DC capacitor voltage error. This represents the predicted midpoint voltage. , , These represent the three predicted DC capacitor voltages of the inverter at the next sampling time. This represents the DC bus voltage. This indicates the nominal value of the midpoint voltage; The vector selection module is used to perform rolling optimization on all voltage vectors in the target vector region using the advanced cost function to obtain the optimal voltage vector in the target vector region. A torque control module is used to control the motor torque of the drive motor through the optimal voltage vector.

4. The high-speed train lift wing converter model predictive torque control system according to claim 3, characterized in that, The torque prediction module includes: A voltage calculation unit is used to calculate the motor voltage of the drive motor at the current sampling time by combining the motor stator current, motor speed and motor inductance. A current prediction unit is used to predict the stator current of the drive motor at the next sampling time based on the motor voltage and using the Euler method. The flux linkage prediction unit is used to calculate the target stator flux linkage required by the drive motor at the next sampling time by combining the predicted stator current of the motor and the motor inductance. The torque prediction unit calculates the target motor torque required by the drive motor at the next sampling moment by combining the target stator flux linkage and the predicted stator current of the motor.

Images