Megawatt axial flux motor model prediction control method and system

Abstract

The invention discloses a megawatt axial flux motor model prediction control method and system, and the method comprises the steps: constructing a prediction model according to a mathematical model of an axial flux motor system and a state parameter at a current moment, and introducing an extended state observer to compensate a current error caused by parameter disturbance; designing a minimum cost function through the prediction model, and selecting an optimal voltage vector by adopting a rapid vector selection mode based on the minimum cost function; and taking an error between a reference value preset according to system requirements and a predicted value obtained according to the prediction model as a feedback value, and continuously updating a control strategy according to the feedback value by adopting a rolling optimization mode to realize prediction control of the megawatt axial flux motor model. According to the invention, a cost function of a multi-objective optimization criterion is designed, a fast vector selection algorithm is added to reduce the hardware calculation time, and the control performance of the system is improved by analyzing the influence of different dead zone voltage vectors on the current voltage vector.

Description

Technical Field

[0001] The present invention relates to the technical field of motor predictive control, and in particular to a megawatt-class axial flux motor model predictive control method and system. Background Art

[0002] Axial flux motors are a type of permanent magnet synchronous motors. They occupy an important position in the field of electric motors and are currently widely used in home appliances, automobiles, aerospace and other fields. The air gap of the axial flux motor is planar, and the air gap magnetic field is distributed axially perpendicular to the plane of the stator and rotor. The axial length is short and has a richer topological structure. Compared with traditional motors, axial flux motors have the characteristics of high power density, high efficiency, light weight, less core material, and large size. In addition, axial flux motors can be directly integrated into the system in a modular manner and combined into a multi-disc structure to broaden the motor speed regulation range and increase the motor torque.

[0003] According to the number of stators and rotors, the relative position of the stators and rotors, and the main magnetic circuit, it can be divided into five categories: single stator single rotor structure, double stator single rotor structure, single stator double rotor structure, single stator double rotor structure, and multi-disc structure. The present invention selects a double stator single rotor axial flux motor as the research object. The advantages of the double stator single rotor structure are: (1) The main magnetic flux of the motor directly passes through the rotor permanent magnet along the axial direction, and there is no circumferential path on the rotor, which can improve the power density and reduce the loss of the rotor core. (2) The rotor has a small mass and a small moment of inertia. (3) There are equal and opposite magnetic pulls between the rotor and the stators on both sides, which are balanced with each other and can eliminate the axial magnetic pull. (4) The stator winding is directly connected to the casing through the stator core, and the heat dissipation effect is better.

[0004] On the other hand, due to the axial design, at the same power level, the inductance of the axial flux motor is smaller than that of the radial flux motor, and it is more sensitive to the external environment. This will cause the change in inductance in the general control model to reduce the overall model control accuracy and affect the dynamic and steady-state performance of the motor. In addition, the axial flux motor is also susceptible to torque pulsation, which is particularly critical for high-quality drives that require smooth speed and position regulation. Especially at low speeds, the impact of torque pulsation may dominate and affect speed and position regulation. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a megawatt-class axial flux motor model predictive control method and system to solve the problems of low multi-objective optimization efficiency, large dead zone error, and inductance parameter mismatch of the axial flux motor.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, the present invention provides a megawatt-class axial flux motor model predictive control method, comprising: constructing a predictive model based on the mathematical model of the axial flux motor system and the state parameters at the current moment, and introducing an extended state observer to compensate for the current error caused by parameter disturbance; designing a minimum cost function through the predictive model, and selecting the optimal voltage vector by a fast vector selection method based on the minimum cost function; using the error between a reference value pre-set according to system requirements and a predicted value obtained according to the predictive model as a feedback value, and continuously updating the control strategy according to the feedback value by a rolling optimization method to achieve predictive control of the megawatt-class axial flux motor model.

[0009] As a preferred solution of the megawatt-class axial flux motor model predictive control method of the present invention, wherein: the construction of the prediction model includes:

[0010] Based on the mathematical model of the axial flux motor system and the state parameters at the current moment, the forward Euler discrete equation is used to predict the d-axis current and q-axis current at the next moment. According to the predicted d-axis current and q-axis current at the next moment, the axial flux motor system is compensated for one-beat delay to obtain a current prediction value of one-beat delay.

[0011] As a preferred solution of the megawatt-class axial flux motor model predictive control method of the present invention, the design minimum cost function includes:

[0012] Substitute the voltage vector into the one-beat delay compensation current prediction equation to obtain the current prediction value at time k+1, substitute the current prediction value into the cost function, and use the voltage vector corresponding to the minimum cost function value as the action voltage vector at time k+1, wherein the cost function is designed as:

[0013]

[0014] Among them, K d and K q represents the weight coefficient, i dp represents the predicted d-axis current, i qp represents the predicted current on the q axis, i d represents the d-axis current, i q represents the q-axis current, ω e Indicates motor speed, R s Indicates stator resistance, N s =500 indicates the speed domain dividing point;

[0015] If the motor speed is greater than the speed domain distinction point, the motor runs in the high-speed domain, and the optimization target of the cost function is to minimize power loss; if the motor speed is less than or equal to the speed domain distinction point, the motor runs in the low-speed domain, and the optimization target of the cost function is to minimize torque pulsation, wherein the standard calculation for minimizing torque pulsation is expressed as:

[0016]

[0017] Among them, i dref represents the d-axis reference current, i qref Indicates the q-axis reference current, L d represents the d-axis inductance component, L q represents the q-axis inductance component, ψ f Represents magnetic linkage.

[0018] As a preferred solution of the megawatt-class axial flux motor model predictive control method of the present invention, the selection of the optimal voltage vector includes:

[0019] The reference voltage vector is calculated as:

[0020]

[0021] Among them, u d represents the d-axis voltage, u q represents the q-axis voltage, T represents the system control period;

[0022] The reference voltage vector phase angle is calculated as:

[0023]

[0024] Among them, u α represents the α-axis voltage vector, u β represents the β-axis voltage vector, θ ref represents the reference voltage vector phase angle;

[0025] The voltage vector plane is divided into 12 sectors, each sector includes a non-zero vector and a zero vector, and the determined non-zero vector and zero vector are evaluated according to the cost function, so as to obtain an optimal voltage vector to act on the inverter.

[0026] As a preferred solution of the megawatt-class axial flux motor model predictive control method described in the present invention, wherein: after the optimal voltage vector of the next beat is obtained through the predictive model predictive control, it is compared with the voltage vector acting at the current moment. If the two voltage vectors are consistent, the switching state of the switching device remains unchanged and no dead zone is required. Otherwise, a dead zone needs to be added during the transition of the switching device from the current moment to the next moment. The dead zone voltage time optimization process is:

[0027] The encoder is used to measure the rotor electrical angle θ and calculate the axial flux motor electrical angular velocity ω e ;

[0028] The stator phase current is decomposed into dq axis by coordinate transformation to obtain i d (k) and i q (k);

[0029] For the obtained dq axis current i d (k) and i q (k) Perform fast vector selection to obtain the optimal voltage vector u uesd , the dead zone voltage vector u is determined based on the predicted optimal voltage vector, the current voltage vector and the phase current direction dead ;

[0030] The dead zone voltage vector u is determined by the sector where the reference voltage vector is located. dead Whether the action time is variable, and according to the grouping result, the dead zone voltage vector u is calculated dead The action time can be directly set to the fixed dead time of the inverter 2μs;

[0031] The optimized dead zone is introduced into the inverter switching state to control the inverter operation.

[0032] As a preferred solution of the megawatt-class axial flux motor model predictive control method of the present invention, the current error caused by the parameter disturbance introduced by the extended state observer compensation includes:

[0033] Based on the state equation of the first-order single-input single-output system of the extended state observer, a second-order linear system is constructed by taking x2 as the new system state variable of the observer expansion.

[0034] According to the second-order linear system, an extended state observer of a second-order single-input single-output system is established, which is expressed as:

[0035]

[0036] Among them, e represents the error signal between the actual value and the estimated value, z1 represents the estimated value of the system state variable, y represents the first-order control object, b represents the estimated gain of the control quantity, z2 represents the total disturbance of the controlled object, β1 and β2 represent the parameters of the extended state observer, and fal(ε, α, δ) represents the error nonlinear function, which is expressed as follows:

[0037]

[0038] Among them, δ represents the filter factor, α represents the nonlinear factor, and ε represents the error factor;

[0039] According to the extended state observer of the second-order single-input single-output system, a total disturbance model of the dq axis current is constructed, the total disturbance is selected as an extended state variable of the observer and the change rate of the total disturbance is set to A second-order extended state observer is established for the q-axis current;

[0040] The expression of the second-order extended state observer is discretized to obtain a dq-axis discrete extended state observer.

[0041] As a preferred solution of the megawatt-class axial flux motor model predictive control method of the present invention, the current error caused by the introduced extended state observer compensation parameter disturbance also includes:

[0042] The stability analysis of the dq-axis discrete extended state observer is carried out. Taking the q-axis as an example, the open-loop function G of the extended state observer is o (z) is:

[0043]

[0044] Where z represents a discrete complex frequency variable;

[0045] Closed loop function G c (z) is:

[0046]

[0047] Where β represents the observer gain;

[0048] The characteristic equation is derived as:

[0049] λ 2 +(β2-2)λ+Tβ2-β1+1=0

[0050] Where λ represents the eigenvalue of the characteristic equation;

[0051] By calculating the pole position, the stability condition of the extended state observer is:

[0052]

[0053] In a second aspect, the present invention provides a megawatt-class axial flux motor model predictive control system, comprising:

[0054] A model building unit is used to build a prediction model based on the mathematical model of the axial flux motor system and the state parameters at the current moment, and introduce an extended state observer to compensate for the current error caused by the parameter disturbance;

[0055] An optimal vector selection unit, configured to design a minimum cost function through the prediction model, and select an optimal voltage vector by adopting a fast vector selection method based on the minimum cost function;

[0056] The prediction control unit is used to use the reference value pre-set according to system requirements and the error of the prediction value obtained according to the prediction model as a feedback value, and adopt a rolling optimization method to continuously update the control strategy according to the feedback value to realize the predictive control of the megawatt-level axial flux motor model.

[0057] In a third aspect, the present invention provides a computing device, comprising:

[0058] Memory and processor;

[0059] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method are implemented.

[0060] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which implement the steps of the method when executed by a processor.

[0061] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a model predictive control method and system for a megawatt-class axial flux motor. Aiming at the problems existing in the megawatt-class axial flux motor, a cost function of a multi-objective optimization criterion is designed. In the low-speed domain, the torque pulsation is minimized as the optimization target, and in the high-speed domain, the efficiency loss is minimized as the optimization target. A fast vector selection algorithm is added to reduce the hardware calculation time. Aiming at the problem of prediction model accuracy error caused by dead time, a control algorithm based on dead-time voltage time optimization is proposed by analyzing the influence of different dead-time voltage vectors on the current voltage vector, which can improve the control performance of the system while keeping the switching device frequency unchanged. Aiming at the problem of inductance parameter and flux linkage parameter disturbances that are prone to occur in axial flux motors under special environments, an extended state observer with dq-axis current as input is proposed to compensate for the current error caused by parameter disturbance. At the same time, an improved error nonlinear function is proposed, so that it has smoother characteristics at the segmentation points compared with traditional error nonlinearity. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0063] Figure 1 This is an overall flow chart of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0064] Figure 2 A schematic diagram of a voltage vector sector of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0065] Figure 3 This is an overall control block diagram of model predictive control for fast vector selection of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0066] Figure 4 A schematic diagram of the influence of different dead zone voltage vectors on a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention (taking U1 (1 0 0) as an example);

[0067] Figure 5 A schematic diagram of the influence of the dead zone duration of the megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention (taking U1 (1 0 0) as an example);

[0068] Figure 6 A dead zone voltage time optimization flow chart of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0069] Figure 7 A structural block diagram of a model predictive control method and system for a megawatt-class axial flux motor according to an embodiment of the present invention;

[0070] Figure 8 A block diagram of a robust model predictive control structure of an axial flux motor based on dead time optimization, of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0071] Fig. 9 A schematic diagram of a simulation model of a robust model predictive control of an axial flux motor based on dead time optimization of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0072] Fig.10 A comparative waveform diagram of inductance parameter mismatch at the start-up time of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0073] Fig.11 A comparative waveform diagram of inductance parameter mismatch at the moment of sudden load addition of a megawatt-class axial flux motor model predictive control method and system according to an embodiment of the present invention;

[0074] Fig.12A schematic diagram of the THD analysis waveform of the A-phase current of the conventional model predictive control method at a given speed of 600 rpm for the model predictive control method and system for a megawatt-class axial flux motor according to an embodiment of the present invention;

[0075] Fig.13 This is a schematic diagram of the A-phase current THD analysis waveform of the model predictive control method with dead-zone voltage time optimization added at a given speed of 600 rpm for the megawatt-class axial flux motor model predictive control method and system described in one embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0077] Example 1

[0078] Reference Figures 1 to 8 In one embodiment of the present invention, a megawatt-class axial flux motor model predictive control method and system are provided, including a megawatt-class axial flux motor model predictive control method and a megawatt-class axial flux motor model predictive control system. Figure 1 The specific steps shown include:

[0079] S100: constructing a prediction model according to a mathematical model of the axial flux motor system and state parameters at a current moment;

[0080] It should be noted that according to the dual-stator single-rotor axial flux motor topology, the stators on both sides of the motor share a rotor shaft in the middle, which can be regarded as two independent stator disks driving the same rotor shaft. Therefore, the following mathematical model uses a single-stator single-rotor axial flux motor structure for modeling.

[0081] In the embodiment of the present application, the forward Euler discrete equation is used to predict the d-axis current and q-axis current at the next moment, which is expressed as:

[0082]

[0083] Among them, i d represents the d-axis current, i q represents the q-axis current, T represents the system control period, R s Indicates stator resistance, L d represents the d-axis inductance component, ω eIndicates the motor speed, u d represents the d-axis voltage, L q represents the q-axis inductance component, u q represents the q-axis voltage, ψ f Indicates magnetic linkage;

[0084] Furthermore, the axial flux motor system is subjected to one-beat delay compensation, and the one-beat delay compensation current is derived according to the formula of the predicted d-axis current and q-axis current at the next moment:

[0085]

[0086] Among them, i dc Indicates the d-axis one-beat delay compensation current, i qc Indicates the q-axis one-beat delay compensation current.

[0087] S200: designing a minimum cost function through a prediction model, and selecting an optimal voltage vector by adopting a fast vector selection method based on the minimum cost function;

[0088] In the embodiment of the present application, designing the minimum cost function includes:

[0089] Substitute the eight groups of voltage vectors into the one-beat delay compensation current prediction equation to obtain the current prediction value at time k+1. Substitute the eight groups of current prediction values ​​into the cost function, and use the voltage vector corresponding to the minimum cost function value as the action voltage vector at time k+1. The cost function is designed as:

[0090]

[0091] Among them, K d and K q represents the weight coefficient, i dp represents the predicted d-axis current, i qp represents the predicted q-axis current, N s =500 indicates the speed domain dividing point;

[0092] It should be noted that if the motor speed is greater than the speed domain distinction point, the motor runs in the high-speed domain, and the optimization goal of the cost function is to minimize the power loss; if the motor speed is less than or equal to the speed domain distinction point, the motor runs in the low-speed domain, and the optimization goal of the cost function is to minimize the torque ripple. The standard calculation for minimizing the torque ripple is expressed as:

[0093]

[0094] Among them, i dref represents the d-axis reference current, i dref Represents the q-axis reference current.

[0095] Furthermore, according to the minimum cost function formula, selecting the optimal voltage vector requires enumerating 8 candidate voltage vectors, which will take up a lot of hardware calculation time and have high requirements on hardware performance. Therefore, a fast vector selection method is used to optimize the process of selecting the optimal voltage vector, specifically:

[0096] The reference voltage vector is calculated as:

[0097]

[0098] The reference voltage vector phase angle is calculated as:

[0099]

[0100] Among them, u α represents the α-axis voltage vector, u β represents the β-axis voltage vector, θ ref represents the reference voltage vector phase angle;

[0101] The voltage vector plane is divided into 12 sectors, each sector includes a non-zero vector and a zero vector, such as Figure 2 As shown, the determined non-zero vector and zero vector are evaluated according to the cost function, and then the optimal voltage vector is obtained to act on the inverter. The overall control block diagram of the model predictive control for fast vector selection is shown in Figure 3 shown.

[0102] S300: After the optimal voltage vector of the next beat is obtained through the prediction model prediction control, it is compared with the voltage vector acting at the current moment. If the two voltage vectors are consistent, the switching state of the switching device remains unchanged and no dead zone is required. Otherwise, a dead zone needs to be added during the transition of the switching device from the current moment to the next moment.

[0103] It should be noted that in the actual axial flux motor control system, when the upper and lower bridge arms of the switching device of the inverter are turned on at the same time, the bus side power supply will be directly turned on and short-circuited, which will further cause serious consequences such as damage to the switching device. Therefore, it is necessary to set a dead zone in the software's pwm register configuration to prevent the upper and lower bridge arms from being turned on due to the switching of the switching device when the voltage vector changes. However, the dead zone also brings about an error between the actual inverter voltage vector and the ideal voltage vector. This will affect the dynamic and steady-state performance of the axial flux motor to a certain extent for the model predictive control that relies on the accuracy of the model. Therefore, it is necessary to reduce the influence of the voltage vector and stator current errors caused by the dead zone in the model predictive control of the axial flux motor. This embodiment proposes a model predictive current control strategy based on dead zone voltage time optimization, which improves the dynamic and steady-state performance of the axial flux motor model predictive control while keeping the switching device frequency unchanged.

[0104] In a feasible implementation, taking U1(1 0 0) and U6(1 0 1) as examples, U1(1 0 0) is the voltage vector at the current moment, and U6(1 0 1) is the voltage vector at the next moment. For the A-phase bridge arm and the B-phase bridge arm, when entering the next moment, the switch states of the two-phase bridge arms do not need to be changed, but the switch state of the C-phase bridge arm changes, and a dead zone needs to be introduced. At this time, the upper and lower bridge arms of the C-phase are all turned off, and the diode plays a freewheeling role. If the C-phase current is in the negative direction, the upper bridge arm of the C-phase is turned on, and the dead zone voltage vector is U6(1 0 1). If the C-phase current is in the positive direction, the lower bridge arm of the C-phase is turned on, and the dead zone voltage vector is U1(1 0 0).

[0105] In a feasible implementation manner, similarly, each time the switch state of the three-phase bridge arm changes and a dead zone is introduced, a corresponding dead zone voltage vector will be generated in the dead zone time according to the phase current direction of the bridge arm that currently needs to be turned off. Therefore, it is necessary to pay attention to the impact of different dead zone voltage vectors on the prediction error caused by the voltage vector at the current moment.

[0106] like Figure 4 As shown, taking the current voltage vector as U1 (1 0 0) as an example, the dead zone vectors are U2 (1 1 0) and U6 (1 0 1), where u ref is the reference voltage vector, u used is the selected optimal voltage vector, u dead is the dead zone voltage vector, u r The actual voltage vector synthesized by the selected voltage vector and the dead zone voltage vector, Δu used is the error between the selected optimal voltage vector and the actual voltage vector, Δu r is the error between the reference voltage vector and the actual voltage vector. Figure 4 It can be seen that when the dead zone voltage vector is U2(1 1 0), Δu r Significantly smaller than Δu used This shows that the existence of the dead zone reduces the error and improves the control performance. However, it can be seen that when the dead zone voltage vector is U6 (1 0 1), the existence of the dead zone increases the prediction error.

[0107] It can be seen that most dead zone voltage vectors are beneficial to improving control accuracy. However, when the dead zone voltage vector and the optimal voltage vector are determined, different action times will also produce different errors, such as Figure 5 As shown in the figure, the optimal voltage vector selected at this time is U1 (1 0 0), and the dead zone voltage vector is U2 (1 1 0). It can be seen that after appropriately increasing the action time of the dead zone voltage vector, the error is significantly reduced, which proves that appropriately increasing the dead zone action time is conducive to improving the control accuracy.

[0108] Furthermore, Table 1 groups the dead zone voltage vectors with gain and attenuation for model predictive control. The dead zone duration of the dead zone voltage vector with gain can be varied, while the dead zone duration of the dead zone voltage vector with attenuation is fixed at 2 μs.

[0109] Table 1: Dead zone voltage vector grouping for different sectors;

[0110]

[0111]

[0112] It should be noted that, through the above analysis, it is necessary to consider the reference voltage vector when analyzing the dead zone voltage vector, divide the voltage vector plane into 12 sectors, calculate the reference voltage vector and determine which sector the voltage vector is in. The calculation formula is as follows:

[0113] According to the current deadbeat principle, the expression for the dead zone duration is obtained:

[0114] i sref =i s (k+1)=i s(k) +K used t used +K dead (Tt used )

[0115] The dead zone duration can be expressed as:

[0116]

[0117] Among them, i sref represents the reference current, i s represents the actual current, t used Represents the duration of the selected optimal vector, Tt used Represents the duration of the dead zone. K uesd represents the current slope of the optimal voltage vector used, K dead The current slope representing the dead zone voltage vector is calculated as follows:

[0118]

[0119] Further, such as Figure 6 The figure shows the dead zone voltage time optimization flow chart, and the specific implementation process is:

[0120] The encoder is used to measure the rotor electrical angle θ and calculate the axial flux motor electrical angular velocity ω e ;

[0121] The stator phase current is decomposed into dq axis by coordinate transformation to obtain id (k) and i q (k);

[0122] For the obtained dq axis current i d (k) and i q (k) Perform fast vector selection to obtain the optimal voltage vector u uesd , the dead zone voltage vector u is determined based on the predicted optimal voltage vector, the current voltage vector and the phase current direction dead ;

[0123] The dead zone voltage vector u is determined by the sector where the reference voltage vector is located. dead Whether the action time is variable, and according to the grouping results, calculate the dead zone voltage vector u dead The action time can be directly set to the fixed dead time of the inverter 2μs;

[0124] The optimized dead zone is introduced into the inverter switching state to control the inverter operation.

[0125] S400: introducing an extended state observer to compensate for the current error caused by parameter disturbance;

[0126] It should be noted that the axial flux motor model predictive control is a control algorithm that relies on the accuracy of the system model. Its control block diagram is as follows: Figure 7 When the motor runs in a high temperature and strong magnetic field environment, the inductance, resistance and flux linkage parameters will change. At this time, the error between the predicted value and the theoretical value of the d and q axis current is:

[0127]

[0128] It can be seen from the above formula that when the resistance, inductance, and flux parameters of the motor system are disturbed, the accuracy of the dq-axis current value calculated by the prediction algorithm decreases, and there is a large error with the theoretical current, causing a static error in the dq-axis current, which in turn causes the motor to produce large torque pulsations and speed oscillations. At the same time, it can be seen that the error model is relatively complex, and the error relationship caused by the changes in the three parameters is a nonlinear relationship, among which the inductance parameter has the greatest impact, and the error impact cannot be reduced by simple linear compensation. Therefore, it is necessary to add an observer to compensate for the current error caused by parameter disturbances and improve the robustness of the axial flux motor control system.

[0129] In the embodiment of the present application, the state equation of the first-order single-input single-output system for establishing the extended state observer is:

[0130]

[0131] Based on the state equation of the first-order single-input single-output system of the extended state observer, a second-order linear system is constructed by taking x2 as the new system state variable of the observer expansion, which is expressed as:

[0132]

[0133] Where x1 represents the system state variable, u represents the input signal of the control system, F(t)=f(x1,w(t),t) represents the total disturbance input to the control system, y represents the first-order control object, and b represents the estimated gain of the control quantity;

[0134] The extended state observer of the second-order single-input single-output system is established based on the second-order linear system, which is expressed as:

[0135]

[0136] Where e represents the error signal between the actual value and the estimated value, z1 represents the estimated value of the system state variable, z2 represents the total disturbance of the controlled object, β1 and β2 represent the parameters of the extended state observer, and fal(ε, α, δ) represents the error nonlinear function, which has the nonlinear characteristics of small error and large gain, and large error and small gain, which can significantly improve the performance of the control system;

[0137] It should be noted that the traditional fal(ε,α,δ) formula is:

[0138]

[0139] However, the traditional fal(ε, α, δ) function has the characteristic of being continuous and not smooth at the segmentation point. When δ is small, the derivative of the traditional fal(ε, α, δ) function changes greatly at the segmentation point, which affects the anti-disturbance performance of the extended state observer. To address this problem, this embodiment designs a new fal(ε, α, δ) function, which is expressed as follows:

[0140]

[0141] Among them, δ represents the filtering factor, which is 0.1, α represents the nonlinear factor, which is 0.25, and ε represents the error factor;

[0142] According to the extended state observer of the second-order single-input single-output system, the total disturbance model of the dq axis current is constructed. The total disturbance is selected as an extended state variable of the observer and the change rate of the total disturbance is set to A second-order extended state observer is established for the q-axis current, which is expressed as:

[0143]

[0144] Among them, e q Indicates iq The error signal between the actual value and the observed value;

[0145] Discretize the expression of the second-order extended state observer, and obtain the dq-axis discrete extended state observer equations:

[0146]

[0147]

[0148] in, Indicates i q The estimated value of Indicates i q Perturbation estimate, ε q Indicates i q Error value, ε d Indicates i d Error value, Indicates i d Estimated value, Indicates i d Perturbation estimate, β1=1.3, β2=6000, b=1.

[0149] Furthermore, the stability analysis of the dq-axis discrete extended state observer is carried out. Taking the q-axis as an example, the open-loop function G of the extended state observer is o (z) is:

[0150]

[0151] Where z represents a discrete complex frequency variable;

[0152] Closed loop function G c (z) is:

[0153]

[0154] Where β represents the observer gain;

[0155] The characteristic equation is derived as:

[0156] λ 2 +(β2-2)λ+Tβ2-β1+1=0

[0157] Where λ represents the eigenvalue of the characteristic equation;

[0158] It can be concluded that only two parameters, β1 and β_2, determine the position and bandwidth of the poles. By calculating the pole positions, the stability condition of the extended state observer is:

[0159]

[0160] It should be noted that the extended state observer designed in this embodiment for dq axis disturbance meets the stability design requirements. The block diagram of the robust model predictive control structure of the axial flux motor based on dead time optimization is as follows: Figure 8 shown.

[0161] S500: using the error between the reference value pre-set according to the system requirements and the predicted value obtained according to the prediction model as the feedback value, and using a rolling optimization method to continuously update the control strategy according to the feedback value, so as to realize the predictive control of the megawatt-class axial flux motor model;

[0162] From the above, it can be seen that the present invention provides a megawatt-class axial flux motor model predictive control method and system. In view of the problems existing in megawatt-class axial flux motors, a cost function of a multi-objective optimization criterion is designed. In the low-speed domain, the torque pulsation is minimized as the optimization target, and in the high-speed domain, the efficiency loss is minimized as the optimization target, and a fast vector selection algorithm is added to reduce the hardware calculation time; in view of the problem of prediction model accuracy error caused by dead time, by analyzing the influence of different dead-zone voltage vectors on the current voltage vector, a control algorithm based on dead-zone voltage time optimization is proposed, which can improve the control performance of the system while keeping the switching device frequency unchanged; in view of the problem of inductance parameter and flux linkage parameter disturbances that are prone to occur in axial flux motors under special environments, an extended state observer with dq-axis current as input is proposed to compensate for the current error caused by parameter disturbances, and an improved error nonlinear function is proposed, so that it has a smoother characteristic at the segmentation point compared with the traditional error nonlinearity.

[0163] The above is a schematic scheme of a megawatt-class axial flux motor model predictive control method of this embodiment. It should be noted that the technical solution of the megawatt-class axial flux motor model predictive control system and the technical solution of the megawatt-class axial flux motor model predictive control method mentioned above belong to the same concept. The details of the technical solution of the megawatt-class axial flux motor model predictive control system not described in detail in this embodiment can be found in the description of the technical solution of the megawatt-class axial flux motor model predictive control method mentioned above.

[0164] This embodiment also provides a megawatt-class axial flux motor model predictive control system, including:

[0165] A model building unit is used to build a prediction model based on the mathematical model of the axial flux motor system and the state parameters at the current moment, and introduce an extended state observer to compensate for the current error caused by the parameter disturbance;

[0166] An optimal vector selection unit, used to design a minimum cost function through a prediction model, and select an optimal voltage vector by a fast vector selection method based on the minimum cost function;

[0167] The prediction control unit is used to use the error between the reference value pre-set according to system requirements and the prediction value obtained according to the prediction model as the feedback value, and adopts a rolling optimization method to continuously update the control strategy according to the feedback value to realize the predictive control of the megawatt-level axial flux motor model.

[0168] The above-mentioned unit modules may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above-mentioned modules.

[0169] This embodiment also provides a computing device, which includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a megawatt-class axial flux motor model predictive control method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0170] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method proposed in the above embodiment is implemented.

[0171] The storage medium proposed in this embodiment and the method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0172] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0173] Example 2

[0174] Reference Figures 9 to 13 Based on the previous embodiment, this embodiment provides an application example of a megawatt-class axial flux motor model predictive control method and system to verify and illustrate the technical effects used in this method.

[0175] In this embodiment, a single-stator single-rotor axial flux motor simulation model is established in MTALAB / Simulink to verify the improved control algorithm. Fig. 9 Shown is a schematic diagram of the robust model predictive control simulation model of the axial flux motor based on dead time optimization. A robust model predictive control simulation model of the axial flux motor based on dead time optimization is established in MATLAB / Simulink. Compared with the traditional model predictive control, the dead time calculation and extended state observer modules are added. Next, the feasibility of the two modules is verified respectively.

[0176] ① Motor robustness verification

[0177] In actual operation of the axial flux motor, the inductance parameters will change significantly due to the influence of magnetic field and temperature, and the general range of change is between 0.5 and 1.5. Therefore, when the inductance parameters are set to 0.5L, L, and 1.5L, a comparative analysis is conducted between the traditional model predictive control system and the model predictive control system with the extended state observer. Fig.10 and Fig.11 As shown, the upper figure is the traditional model predictive control, and the lower figure is the model predictive control with the addition of the extended state observer. The motor is set to have a given speed of 600 at time 0, an initial load torque of 30 N·m, and a sudden load of 60 N·m is applied at 0.04 seconds.

[0178] From the comparison waveforms, we can see that when the inductance parameter changes from 0.5L to 1.5L at the start-up time, the speed of the traditional model predictive control changes between -40rpm and 60rpm, and when the load is suddenly added, the speed changes between -35rpm and 25rpm; when the model predictive control with the extended state observer is added, the speed changes between -10rpm and 40rpm at the start-up time, and between -15rpm and 10rpm at the sudden load. Therefore, it can be seen that after adding the extended state observer, the robustness of the system has been significantly improved.

[0179] ② Dead zone voltage time optimization verification

[0180] In order to verify the steady-state performance of the dead-zone voltage time optimization control, the motor speed is set at 600rpm, and the steady-state performance of the A-phase current is verified for the model predictive control system with the dead-zone voltage time optimization. Fig.12 and Fig.13 It can be seen that the THD of phase A current of the model predictive control system with dead-time voltage optimization is 8.36%, which is smaller than the THD of phase A current of the traditional model predictive control system of 19.80%. This is because the dead-time voltage time optimization control algorithm fully considers the availability of the dead zone, and obtains the dead-time action time of the dead-time voltage vector that can reduce the prediction error by calculating, thereby reducing the prediction error and reducing the three-phase current distortion rate.

[0181] Therefore, ① in response to the problems existing in megawatt-class axial flux motors, a cost function of a multi-objective optimization criterion is designed, with the minimum torque pulsation as the optimization target in the low-speed domain and the minimum efficiency loss as the optimization target in the high-speed domain, and a fast vector selection algorithm is added to reduce the hardware calculation time; ② in response to the problem of prediction model accuracy error caused by dead time, a control algorithm based on dead-time voltage time optimization is proposed by analyzing the influence of different dead-time voltage vectors on the current voltage vector, which can improve the control performance of the system while keeping the switching device frequency unchanged; ③ in response to the problem of inductance parameter and flux linkage parameter disturbances that are prone to occur in axial flux motors under special environments, an extended state observer with dq-axis current as input is proposed to compensate for the current error caused by parameter disturbances. At the same time, an improved error nonlinear function is proposed, which makes it have smoother characteristics at the segmentation points compared with the traditional error nonlinearity.

[0182] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0183] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0184] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0185] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0186] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0187] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0188] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A megawatt-class axial flux motor model predictive control method, characterized in that: include: A prediction model is constructed based on the mathematical model of the axial flux motor system and the state parameters at the current moment, and an extended state observer is introduced to compensate for the current error caused by parameter disturbance. Designing a minimum cost function through the prediction model, and selecting an optimal voltage vector by fast vector selection based on the minimum cost function; The reference value pre-set according to system requirements and the error between the predicted value obtained according to the prediction model are used as feedback values, and the control strategy is continuously updated according to the feedback values ​​in a rolling optimization manner to achieve predictive control of the megawatt-level axial flux motor model.

2. The megawatt-class axial flux motor model predictive control method according to claim 1, characterized in that: The constructing of the prediction model comprises: Based on the mathematical model of the axial flux motor system and the state parameters at the current moment, the forward Euler discrete equation is used to predict the d-axis current and q-axis current at the next moment. According to the predicted d-axis current and q-axis current at the next moment, the axial flux motor system is compensated for one-beat delay to obtain a current prediction value of one-beat delay.

3. The megawatt-class axial flux motor model predictive control method according to claim 2, characterized in that: The designed minimum cost function includes: Substitute the voltage vector into the one-beat delay compensation current prediction equation to obtain the current prediction value at time k+1, substitute the current prediction value into the cost function, and use the voltage vector corresponding to the minimum cost function value as the action voltage vector at time k+1, wherein the cost function is designed as: Among them, K d and K q represents the weight coefficient, i dp represents the predicted d-axis current, i qp represents the predicted current on the q axis, i d represents the d-axis current, i q represents the q-axis current, ω e Indicates motor speed, R s Indicates stator resistance, N s =500 indicates the speed domain dividing point; If the motor speed is greater than the speed domain distinction point, the motor runs in the high-speed domain, and the optimization target of the cost function is to minimize power loss; if the motor speed is less than or equal to the speed domain distinction point, the motor runs in the low-speed domain, and the optimization target of the cost function is to minimize torque pulsation, wherein the standard calculation for minimizing torque pulsation is expressed as: Among them, i dref represents the d-axis reference current, i qref Indicates the q-axis reference current, L d represents the d-axis inductance component, L q represents the q-axis inductance component, ψ f Represents magnetic linkage.

4. The megawatt-class axial flux motor model predictive control method according to claim 3, characterized in that: The selecting of the optimal voltage vector comprises: The reference voltage vector is calculated as: Among them, u d represents the d-axis voltage, u q represents the q-axis voltage, T represents the system control period; The reference voltage vector phase angle is calculated as: Among them, u α represents the α-axis voltage vector, u β represents the β-axis voltage vector, θ ref represents the reference voltage vector phase angle; The voltage vector plane is divided into 12 sectors, each sector includes a non-zero vector and a zero vector, and the determined non-zero vector and zero vector are evaluated according to the cost function, so as to obtain an optimal voltage vector to act on the inverter.

5. The megawatt-class axial flux motor model predictive control method according to claim 4, characterized in that: After the prediction model predictive control obtains the optimal voltage vector for the next beat, it is compared with the voltage vector acting at the current moment. If the two voltage vectors are consistent, the switching state of the switching device remains unchanged and no dead zone is required. Otherwise, a dead zone needs to be added during the transition of the switching device from the current moment to the next moment. The dead zone voltage time optimization process is: The encoder is used to measure the rotor electrical angle θ and calculate the axial flux motor electrical angular velocity ω e ; The stator phase current is decomposed into dq axis by coordinate transformation to obtain i d (k) and i q (k); For the obtained dq axis current i d (k) and i q (k) Perform fast vector selection to obtain the optimal voltage vector u uesd , the dead zone voltage vector u is determined based on the predicted optimal voltage vector, the current voltage vector and the phase current direction dead ; The dead zone voltage vector u is determined by the sector where the reference voltage vector is located. dead Whether the action time is variable, and according to the grouping result, the dead zone voltage vector u is calculated dead The action time can be directly set to the fixed dead time of the inverter 2μs; The optimized dead zone is introduced into the inverter switching state to control the inverter operation.

6. The megawatt-class axial flux motor model predictive control method according to claim 1, characterized in that: The current error caused by introducing the extended state observer to compensate for the parameter disturbance includes: Based on the state equation of the first-order single-input single-output system of the extended state observer, a second-order linear system is constructed by taking x2 as the new system state variable of the observer expansion. According to the second-order linear system, an extended state observer of a second-order single-input single-output system is established, which is expressed as: Among them, e represents the error signal between the actual value and the estimated value, z1 represents the estimated value of the system state variable, y represents the first-order control object, b represents the estimated gain of the control quantity, z2 represents the total disturbance of the controlled object, β1 and β2 represent the parameters of the extended state observer, and fal(ε, α, δ) represents the error nonlinear function, which is expressed as follows: Among them, δ represents the filter factor, α represents the nonlinear factor, and ε represents the error factor; According to the extended state observer of the second-order single-input single-output system, a total disturbance model of the dq axis current is constructed, the total disturbance is selected as an extended state variable of the observer and the change rate of the total disturbance is set to A second-order extended state observer is established for the q-axis current; The expression of the second-order extended state observer is discretized to obtain a dq-axis discrete extended state observer.

7. The megawatt-class axial flux motor model predictive control method according to claim 6, characterized in that: The current error caused by introducing the extended state observer to compensate for the parameter disturbance also includes: The stability analysis of the dq-axis discrete extended state observer is carried out. Taking the q-axis as an example, the open-loop function G of the extended state observer is o (z) is: Where z represents a discrete complex frequency variable; Closed loop function G c (z) is: Where β represents the observer gain; The characteristic equation is derived as: l 2 +(β2-2)λ+Tβ2-β1+1=0 Where λ represents the eigenvalue of the characteristic equation; By calculating the pole position, the stability condition of the extended state observer is:

8. A system using the model predictive control method for a megawatt-class axial flux motor as claimed in any one of claims 1 to 7, characterized in that: include: A model building unit is used to build a prediction model based on the mathematical model of the axial flux motor system and the state parameters at the current moment, and introduce an extended state observer to compensate for the current error caused by the parameter disturbance; An optimal vector selection unit, configured to design a minimum cost function through the prediction model, and select an optimal voltage vector by adopting a fast vector selection method based on the minimum cost function; The prediction control unit is used to use the reference value pre-set according to system requirements and the error of the prediction value obtained according to the prediction model as a feedback value, and adopt a rolling optimization method to continuously update the control strategy according to the feedback value to realize the predictive control of the megawatt-level axial flux motor model.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having computer program instructions stored thereon, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images