A fault-tolerant predictive control system and method for dual three-phase permanent magnet motors

Abstract

The present invention discloses a fault-tolerant predictive control system and method for a dual three-phase permanent magnet motor, and relates to the technical field of multiphase motor control. A dimensionality reduction decoupling matrix is ​​used to derive the distribution and prediction model of the voltage vector after a motor fault; according to the voltage vector after the fault, a set of virtual voltage vector control sets with zero harmonic plane components are constructed, and the value function is used for evaluation and optimization; at the same time, the amplitude of the virtual voltage vector is corrected by current deadbeat to expand the modulation range; in addition, a virtual zero vector without fundamental component is designed, and the virtual zero vector and the deadbeat virtual voltage vector together form a closed-loop predictive control system to achieve high-precision control of motor current. After adopting the dimensionality reduction decoupling matrix, the present invention can achieve current decoupling of the fundamental and harmonic planes, which is convenient for controller design, and the closed-loop control system can effectively reduce the 5th and 7th harmonics caused by factors such as non-sinusoidal back electromotive force and inverter nonlinearity, thereby improving system performance.

Description

Technical Field

[0001] The present invention belongs to the technical field of multi-phase motor predictive control, and in particular relates to a fault-tolerant predictive control method for a dual three-phase permanent magnet motor. Background Art

[0002] With the rapid development of high-end fields such as transportation, aerospace, and national defense, the requirements for motor systems as core components of equipment have been further improved. Multiphase permanent magnet motors have the advantages of high power density, high efficiency, and good fault tolerance, and have become the first choice for advanced motor systems. Among them, the dual three-phase permanent magnet synchronous motor with center point isolation and two sets of windings connected with a phase shift of 30° has been widely used due to its special structure, which eliminates the 6th torque pulsation. The model predictive control strategy has good performance in power converter applications due to its advantages of multivariable control, easy handling of nonlinear constraints, and intuitive and easy implementation, and gradually reflects good engineering application value. The harmonic suppression problem has always been a hot topic in the field of dual three-phase motors. The Chinese invention patent "A model predictive current control method and system for dual three-phase permanent magnet synchronous motors" (patent number CN202111517942.3) discloses a predictive control method for dual three-phase motors under the condition of a virtual voltage control set. The Chinese invention patent "A dual three-phase permanent magnet synchronous generator dual space duty cycle model predictive current control method" (patent number CN202111135289.4) discloses a method using harmonic plane closed-loop control to improve the control performance of the motor. However, after a single-phase open-circuit fault occurs in a dual three-phase motor, the harmonic plane and the fundamental plane are coupled, and the control method of the above patent is no longer applicable. Traditional vector control technology controls by modifying the given harmonic current reference value, but does not achieve the purpose of suppressing harmonic current. Moreover, in the field of model predictive control, current research does not involve post-fault harmonic current suppression technology. Therefore, in order to enhance the advantages of the dual three-phase capacitor fault characteristics, it is urgent to develop relevant technologies to improve motor performance after a fault. Summary of the invention

[0003] Purpose of the invention: In view of the coupling problem between harmonic current and fundamental current after a single-phase open-circuit fault in a dual three-phase motor, a fault-tolerant predictive control method for a dual three-phase permanent magnet motor is proposed. First, the distribution of the basic voltage vector after the fault is derived by using a dimensionality reduction decoupling matrix. Secondly, according to the principle of virtual voltage vector synthesis, 12 fundamental plane virtual voltage vectors are synthesized and used as the control set for predictive control; in addition, the concept of a virtual zero vector is proposed, which is characterized in that the fundamental component of the vector is zero, and the harmonic component is not zero. It is only used for harmonic plane regulation and does not participate in electromechanical energy conversion. Finally, in the event of a fault, the dual three-phase motor can still maintain dual-plane closed-loop control of the fundamental plane and the harmonic plane to improve motor performance.

[0004] Technical solution: To achieve the above-mentioned invention object, the technical solution adopted by the present invention is as follows: a fault-tolerant predictive control system for a dual three-phase permanent magnet motor, comprising a speed controller, a current change rate module, a delay compensation module, a deadbeat duty cycle calculation module, a value function module, a harmonic current controller, a compensated duty cycle module, a bridge arm switch state module, a PWM module, a coordinate transformation module, an inverter module, a position sensor and a dual three-phase permanent magnet motor;

[0005] The coordinate transformation module is connected to the inverter module via a current sensor and is used to transform the natural coordinate system variables of each phase into the rotating coordinate system variables;

[0006] The speed controller is controlled by PI to obtain the q-axis reference current, the input of which is the error between the given speed and the actual speed, and the output of which is the reference value of the q-axis current;

[0007] The input end of the delay compensation module is connected to the coordinate transformation module, and the output end is connected to the current change rate calculation module, so as to compensate for the "one-beat delay" problem caused by the sampling of the digital system;

[0008] The input end of the current change rate calculation module is connected to the delay compensation module and the position sensor, and the output end is connected to the deadbeat duty cycle calculation module to obtain the change rate of the dq axis current when different voltage vectors act;

[0009] The deadbeat duty cycle calculation module has an input end connected to the current change rate calculation module, and an output connected to the value function module, and is used to output the duty cycle of each voltage vector acting on the fundamental wave plane;

[0010] The input end of the value function module is connected to the deadbeat duty cycle calculation module, which is used to traverse all fundamental wave candidate voltage vectors and their duty cycles, and input the selected optimal voltage vector and its duty cycle;

[0011] The speed controller, delay compensation module, current change rate calculation module, deadbeat duty cycle calculation module and value function module together constitute a prediction model, which functions to control the variables of the fundamental wave plane;

[0012] The harmonic current controller is controlled by PI to obtain the z-axis reference voltage. Its input is the error between the given z-axis current and the actual z-axis current, and its output is the reference value of the z-axis voltage. The duty cycle that needs to be compensated for the final harmonic plane is calculated by comparing it with the synthesized virtual zero vector, forming a harmonic compensation module.

[0013] The input end of the compensated duty cycle module is connected to the prediction model and the harmonic compensation module, and its function is to integrate the duty cycles required by the fundamental plane and the harmonic plane, and to reasonably distribute them; the output of the compensated duty cycle module is connected to the bridge arm switch state module;

[0014] The input end of the bridge arm switch state module is connected to the compensated duty cycle module, and the output end is connected to the PWM module;

[0015] The PWM module is connected to the inverter module to perform signal modulation and current output, thereby driving the dual three-phase permanent magnet motor to operate.

[0016] A fault-tolerant predictive control method for a dual three-phase permanent magnet motor of the present invention comprises the following steps:

[0017] Step 1) using the dimension reduction decoupling matrix to derive the distribution of the voltage vector after the fault;

[0018] Step 2) constructing a virtual voltage vector (Virtual vector, VV) with zero harmonic component according to the voltage distribution;

[0019] Step 3) constructing a virtual null-vector (VN) with a fundamental component of zero according to the voltage distribution;

[0020] Step 4) using the coordinate transformation module to obtain the control variables required for control;

[0021] Step 5) The fundamental wave plane uses the deadbeat current control to calculate the duty cycle d of VV VV ;

[0022] Step 6) The harmonic plane uses a PI controller to obtain the harmonic reference voltage, and determines the required duty cycle d of VN through the harmonic reference voltage and the amplitude of VN VN ;

[0023] Step 7) Adjust according to the action time of VV and VN to ensure that VV and VN are reasonably arranged within a control cycle;

[0024] Step 8) deriving a single-phase open circuit prediction model of a dual three-phase permanent magnet motor based on a dimensionality reduction decoupling matrix;

[0025] Step 9) Complete the post-fault dual-plane closed-loop control under the new virtual vector control set (VV and VN).

[0026] Further, step 1) specifically includes: deriving the voltage vector distribution of the dual three-phase motor under the single-phase open circuit fault based on the dimension reduction decoupling matrix, taking the F phase open circuit as an example, the remaining phase open circuits can be obtained by phase shifting the corresponding angles,

[0027] Under normal circumstances, the voltage, current and flux of the dual three-phase motor can be decomposed into different planes by the method of vector space decoupling, namely the fundamental wave plane (αβ plane), the harmonic plane (z1z2 plane) and the zero sequence plane (o1o2 plane), and the corresponding relationship is:

[0028] [f α f β f x f y f o1 f o2 ] T =T 6s [f A f B f C f D f E f F ] T (1)

[0029] Among them, f α 、f β 、f x 、f y 、f o1 and f o2 Represent the vectors of voltage, current and flux in the α-axis, β-axis, x-axis, y-axis, o1-axis and o2-axis respectively, and the decoupling matrix T 6s As shown below:

[0030]

[0031] When an open circuit fault occurs on phase F, the dimension reduction decoupling matrix is ​​used to re-decompose the vector space, and the relationship between the voltage vectors can be expressed as follows:

[0032] [u α u β u z u o1 u o2 ] T =T 5s [u A u B u C u D u E ] T (3)

[0033] Among them, u A ~u E are the phase voltages of phases A to F, u α ~u o2 are the voltage components of the corresponding planes, and the dimension reduction decoupling matrix T 5s for:

[0034]

[0035] The dual three-phase motor adopts the center point isolation connection mode, so the zero sequence plane can be ignored, so equation (3) can be transformed into:

[0036]

[0037] Among them, u α ~u z are the voltage components of the corresponding planes, u A 、u B 、u C are the phase voltages of phase A, phase B and phase C, u DE is the potential difference between phase D and phase E;

[0038] The dual three-phase motor is driven by a two-level six-phase voltage source inverter, and the relationship between its phase voltage and switch state is:

[0039]

[0040] Among them, U dc Indicates bus voltage, S A ~S E Represents the switching state of the upper bridge arm switch tube of each phase of the inverter. "1" means that the upper bridge arm is turned on and the lower bridge arm is turned off; "0" means that the upper bridge arm is turned off and the lower bridge arm is turned on. The numbering of the basic voltage vector is based on S A S B S C S D S E The switch state combination is represented in decimal in the order of "10101", and the number is 21, which is represented by v 21 express.

[0041] According to (5) and (6), we can deduce:

[0042]

[0043] It can be deduced that the switching state of the voltage vector after the fault is 2 5 = 32, and its components on the α-axis, β-axis and z-axis are shown in the following table:

[0044] Table 1 Voltage vector distribution under F phase open circuit fault condition

[0045]

[0046] Table 1 shows the voltage vector distribution under the F phase open circuit fault condition, where v 00 、v 01 ,…v31 The meaning is the voltage vector in the 32 switch states after the fault.

[0047] Further, the specific steps of step 2) include: selecting adjacent maximum three vectors as the basic voltage vectors for synthesizing the fundamental plane virtual voltage vector. The synthesis concept of the fundamental plane virtual voltage vector is: the sum of all voltage vectors in the harmonic plane components is equal to zero, which is expressed as follows:

[0048]

[0049] Among them, v 1st , v 2nd , and v 3rd where V represents the first, second and third effective voltage vectors respectively; i is the ith virtual voltage vector; the superscripts “α”, “β” and “z” represent the components of the voltage vector on the α-axis, β-axis and z-axis respectively; D1, D2, D3 are the duty cycles of the first, second and third effective voltage vectors, respectively, and D0 is the duty cycle of the zero vector; D0, D1, D2, D3∈[0, 1];

[0050] Finally, the distribution of the fundamental wave plane virtual voltage vector is shown in Table 2:

[0051] Table 2 Distribution of fundamental wave plane virtual voltage vector

[0052]

[0053]

[0054] Twelve fundamental plane virtual voltage vectors are synthesized, with an amplitude of 0.29Udc and an angle of 30° between the two vectors.

[0055] Further, the specific steps of step 3) include:

[0056] According to Table 1, we can find that the vector v 15 The direction of v2 is opposite to that of v2 on the α axis, but the same on the z axis. Therefore, the action ratio of the two voltage vectors can be reasonably configured so that the sum of their components on the α axis is zero and they remain negative on the z axis. Similarly, v 29 and v 16 , thus synthesizing a positive effective z-axis voltage vector again. After calculation, the synthesis principle is as follows:

[0057]

[0058] Among them, v positive represents the virtual zero vector effective in the positive direction, v negative Indicates a virtual zero vector that is effective in the negative direction;

[0059] The amplitude of the virtual voltage vector of the synthetic harmonic plane is 0.3Udc, and its component in the fundamental wave plane is zero, which is called the virtual zero vector VN. The application of the VN vector does not participate in the electromechanical energy conversion of the fundamental wave plane.

[0060] Further, the specific steps of step 4) include:

[0061] The 5-phase current i of a dual three-phase motor after a single-phase open circuit fault occurs A ,i B ,i C ,i D ,i E The current sensor is directly obtained, and the dimension reduction decoupling matrix is ​​used to transform each variable of the natural coordinate system into the stationary coordinate system. The transformation process is:

[0062] [i α i β i z i o1 i o2 ] T =T 5s [i A i B i C i D i E ] T (10)

[0063] Among them, i α 、i β 、i z 、i o1 、i o2 is the current of α axis, β axis, z axis, o1 axis and o2 axis; decoupling matrix T 5s As shown below:

[0064]

[0065] For dual three-phase permanent magnet motors, only the fundamental wave component of the αβ subspace participates in the electromechanical energy conversion. In order to simplify the analysis, the stationary coordinate system is transformed into the synchronous rotating coordinate system, and its transformation matrix is:

[0066]

[0067] The above coordinate conversion module calculates the current of the motor at time k in the dq rotating coordinate system and Further, the specific steps of step 5) include:

[0068] After the F phase open circuit fault, the voltage equation of the dual three-phase permanent magnet motor in the dq coordinate system is:

[0069]

[0070] Among them, u d 、u q 、u z are the d-axis, q-axis, and z-axis voltages respectively; i d 、i q 、i z are d-axis, q-axis, and z-axis currents respectively; L d , L q , L l are the inductances of the d-axis, q-axis, and z-axis respectively; Rs is the stator resistance; ω e is the electrical angular velocity, ψ f is the permanent magnet flux.

[0071]

[0072] Where θ is the rotor position angle.

[0073] From the voltage equation, we can get: the change rate of the dq axis current under different virtual voltage vectors is as follows (current change rate module):

[0074]

[0075] Taking the q-axis reference current as the prediction target, the deadbeat principle is used to calculate the duty cycle of VV (deadbeat duty cycle calculation module), and the formula is as follows:

[0076]

[0077] in, is the reference value of q-axis current; is the actual value of the q-axis current at time k; s VV When the fundamental wave virtual voltage vector acts q The rate of change; s null_vector When the vector is zero, i q The rate of change of VV is the duty cycle of the fundamental wave virtual voltage vector; d null_vector is the duty cycle of the zero vector; T s is the control period. The solution is:

[0078]

[0079] Further, the specific steps of step 6) include:

[0080] In the harmonic plane, a PI controller is used to input the harmonic reference current and obtain the harmonic reference voltage. (Harmonic current controller) The duty cycle is determined according to the proportional relationship between the amplitude of the reference voltage and the amplitude of VN.

[0081]

[0082] Further, the specific steps of step 7) include:

[0083] Judgement VV The size of VV When <1, VN participates in control, otherwise VN is disabled;

[0084] After the conditions are met, if 0<d VV +d VN ≤1, then d VV and d VN remains unchanged, and the remaining duty cycle is compensated by the zero vector, i.e. d null_vector =1-d VV -d VN ;

[0085] If d VV +d VN >1, then d VN =1-d VV , d null_vector =0.

[0086] Finally determine the allocation method of VV and VN.

[0087] Further, the specific steps of step 8) include:

[0088] The forward Euler method is used to discretize the voltage equation after the fault to obtain the prediction model:

[0089]

[0090] The superscript “k” represents the value of the corresponding variable at time k; the superscript “k+1” represents the value of the corresponding variable at time k+1; T s To control the cycle;

[0091] In order to compensate for the "one-shot delay" characteristic of the digital system, the two-step prediction method adopted by the delay compensation module is used for delay compensation, that is, one more step of prediction. The final prediction model is:

[0092]

[0093] Among them, the superscript "k+2" represents the value of the corresponding variable at time k+2;

[0094] Further, the specific steps of step 9) include:

[0095] After the above steps, the final required voltage vector can be expressed as three parts: VV, VN and V null Its expression needs to be modulated and generated by the bridge arm switch state module. The duty cycle of each bridge arm opening and closing is the sum of the duty cycles of each voltage vector corresponding to the phase, that is,

[0096]

[0097] in, Indicates the final opening duty cycle of the corresponding phase bridge arm; Indicates the duty cycle of each bridge arm of the virtual voltage vector; Represents the duty cycle of each bridge arm of the virtual zero vector; since the bridge arm switch state of the zero vector is selected as "00000", this part is omitted in the above formula.

[0098] Substitute the 12 virtual voltage vectors VV into the prediction model one by one, select the optimal voltage vector and its duty cycle, and inject the modified VN into the harmonic plane to complete the entire control system. So far, the fundamental plane adopts the predictive current control based on the deadbeat duty cycle for closed-loop control, and the harmonic plane adopts the method based on the PI controller for closed-loop control. The two obtain the corresponding d VV and d VN Then it is output to the PWM module to complete the motor control.

[0099] Beneficial effects of the present invention:

[0100] 1) The decoupling fault-tolerant predictive current closed-loop control method for a single-phase open-circuit fault of a dual three-phase permanent magnet motor of the present invention can still ensure the control of the harmonic plane under fault conditions and does not affect the electromechanical energy conversion of the fundamental wave plane, thereby improving the performance of the motor.

[0101] 2) The construction of the virtual voltage vector after the fault eliminates the voltage vector component in the harmonic plane, further improving the control performance.

[0102] 3) The proposed virtual zero vector method can ensure the closed-loop control of the harmonic plane and reduce the harmonic problems caused by dead zone and inverter nonlinearity.

[0103] 4) The voltage vector distribution under single-phase open circuit fault derived by the present invention provides basic work for studying the fault-tolerant control of dual three-phase motors. Subsequent research can be carried out directly on this basis. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 Schematic diagram of the control principle of the method of the embodiment of the present invention

[0105] Figure 2 The topological structure diagram of a six-phase voltage source inverter to which the method according to the embodiment of the present invention is applied is as follows:

[0106] (a) Six-phase two-level voltage source inverter, (b) dual three-phase permanent magnet synchronous motor;

[0107] Figure 3 The spatial voltage vector diagram of the present invention: (a) αβ plane, (b) z plane;

[0108] Figure 4 Schematic diagram of the virtual voltage vector structure designed for the present invention: (a) αβ plane, (b) z plane;

[0109] Figure 5 Virtual voltage vector distribution diagram designed for the present invention

[0110] Figure 6 Schematic diagram of the virtual zero vector structure designed for the present invention: (a) αβ plane, (b) z plane;

[0111] Figure 7 Virtual zero vector distribution diagram designed for the present invention

[0112] Figure 8 Schematic diagram of the distribution of different vectors in one cycle of the present invention

[0113] Fig. 9 The experimental waveform of the traditional harmonic closed-loop control strategy when the F phase is open circuit

[0114] Fig.10 The experimental waveform of the present invention when the F phase is open circuit DETAILED DESCRIPTION

[0115] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0116] Figure 1It is a schematic diagram of the control frame principle of the present invention, which includes a speed controller, a current change rate module, a delay compensation module, a deadbeat duty cycle calculation module, a value function module, a harmonic current controller, a compensated duty cycle module, a bridge arm switch state module, a PWM module, a coordinate transformation module, an inverter module, a position sensor and a dual three-phase permanent magnet motor. The coordinate transformation module is connected to the inverter module through the current sensor, and its function is to convert the natural coordinate system variables of each phase into the rotating coordinate system variables; the speed controller is controlled by PI to obtain the q-axis reference current, and its input end is the error between the given speed and the actual speed, and the output end is the reference value of the q-axis current; the input end of the delay compensation module is connected to the coordinate transformation, and the output end is connected to the current change rate calculation module to compensate for the "one-beat delay" problem caused by digital system sampling; the input end of the current change rate module is connected to the delay compensation module and the position sensor, and the output end is connected to the deadbeat duty cycle calculation module to obtain the change rate of the dq-axis current when different voltage vectors act; the input end of the deadbeat duty cycle calculation module is connected to the current change rate module, and its function is to output the duty cycle of each voltage vector in the fundamental plane; the input end of the value function module is connected to the deadbeat duty cycle calculation module, and its function is to traverse all fundamental candidate voltage vectors and their duty cycle, and input the selected optimal voltage vector and its duty cycle; the speed controller, delay compensation module, current change rate calculation module, deadbeat duty cycle calculation module and value function module together constitute a prediction model, which is used to control the variables of the fundamental wave plane; the harmonic current controller is controlled by PI to obtain the z-axis reference voltage, and its input end is the error between the given z-axis current and the actual z-axis current, and the output end is the reference value of the z-axis voltage, and the duty cycle required to be compensated for the final harmonic plane is calculated by ratioing with the synthesized virtual zero vector; the input end of the compensated duty cycle module is connected to the prediction model and the harmonic compensation module, and its function is to integrate the duty cycles required for the fundamental wave plane and the harmonic plane, and reasonably allocate them; the input end of the bridge arm switch state module is connected to the compensated duty cycle module, and the output end is connected to the PWM module; the PWM module is connected to the inverter module to perform signal modulation and current output, thereby driving the motor to run.

[0117] The implementation steps of the method are mainly divided into the following steps:

[0118] Step 1: Obtain the voltage vector distribution after the fault.

[0119] When an open circuit fault occurs on phase F, the dimension reduction decoupling matrix is ​​used to re-decompose the vector space, and the relationship between the voltage vectors can be expressed as follows:

[0120] [u α u β u z u o1 u o2] T =T 5s [u A u B u C u D u E ] T (1)

[0121] Among them, u A ~u E are the phase voltages of phases A to F, u α ~u o2 are the voltage components of the corresponding planes respectively.

[0122] The dual three-phase motor of the present invention adopts a center point isolation connection mode, so the zero sequence plane can be ignored, and thus equation (1) can be converted into:

[0123]

[0124] Among them, u α ~u z are the voltage components of the corresponding planes, u A 、u B 、u C are the phase voltages of phase A, phase B and phase C respectively. DE is the potential difference between phase D and phase E.

[0125] like Figure 2 As shown, the dual three-phase motor is driven by a two-level six-phase voltage source inverter, and the relationship between its phase voltage and switch state is:

[0126]

[0127] Among them, U dc Indicates bus voltage, S A ~S E Represents the switching state of the upper bridge arm switch tube of each phase of the inverter. "1" means that the upper bridge arm is turned on and the lower bridge arm is turned off; "0" means that the upper bridge arm is turned off and the lower bridge arm is turned on. The numbering of the basic voltage vector is based on S A S B S C S D S E The switch state combination is represented in decimal in the order of "10101", and the number is 21, which is represented by v 21 express.

[0128] According to (2) and (3), we can deduce:

[0129]

[0130] It can be deduced that the switching state of the voltage vector after the fault is 2 5 = 32, and its components on the α-axis, β-axis and z-axis are shown in the following table:

[0131] Table 1 Voltage vector distribution under F phase open circuit fault condition

[0132]

[0133]

[0134] The voltage vector distribution in the αβ subspace and the z subspace is as follows: Figure 3 shown.

[0135] Step 2: Construct a virtual voltage vector.

[0136] Depend on Figure 3 It can be seen that the distribution of the voltage vector on the fundamental plane after the fault is not uniform, so the idea of ​​voltage vector modulation is used to redesign the virtual voltage vector. The adjacent maximum three vectors are selected as the basic voltage vectors for synthesizing the virtual voltage vector on the fundamental plane. Figure 4 As shown, v obj is the target voltage vector to be synthesized, and its three adjacent voltage vectors are v 18 ,v 27 , and v 26 According to the virtual voltage vector principle, it is necessary to ensure that the sum of the components of the three voltage vectors in the harmonic plane is equal to zero, that is:

[0137]

[0138] Then its resultant vector in the αβ plane is expressed as:

[0139]

[0140] Similarly, according to the number and distribution of voltage vectors before the dual three-phase motor fails, 12 virtual voltage vectors are redesigned, and their distribution is as follows: Figure 5 The synthesis principles are summarized as follows:

[0141]

[0142] Among them, v 1st , v 2nd , and v 3rd where V represents the first, second and third effective voltage vectors respectively; iis the i-th virtual voltage vector; the superscripts “α”, “β” and “z” represent the components of the voltage vector on the α-axis, β-axis and z-axis respectively; D1, D2, D3 are the duty cycles of the first, second and third effective voltage vectors respectively, and D0 is the duty cycle of the zero vector; D0, D1, D2, D3∈[0, 1].

[0143] Finally, 12 fundamental plane virtual voltage vectors are synthesized, with an amplitude of 0.29Udc and an angle of 30° between the two vectors.

[0144] Table 2 Distribution of fundamental wave plane virtual voltage vector

[0145]

[0146]

[0147] The synthetic result of the virtual voltage vector is as follows: Figure 5 As shown in the figure, the amplitude of each voltage vector is equal, and the phase angles between the voltage vectors are evenly distributed. This distribution method is conducive to improving the control performance of the motor.

[0148] Step 3: Construct a virtual zero vector.

[0149] like Figure 6 As shown, it can be found that the vector v 15 The direction of v2 is opposite to that of v2 on the α axis, but the same on the z axis. Therefore, the action ratio of the two voltage vectors can be reasonably configured so that the sum of their components on the α axis is zero and they remain negative on the z axis. Similarly, v 29 and v 16 , thus synthesizing a positive effective z-axis voltage vector again. After calculation, the synthesis principle is as follows:

[0150]

[0151] Among them, v positive represents the virtual zero vector effective in the positive direction, v negative Indicates a virtual zero vector that is valid in the negative direction.

[0152] The virtual voltage vector of the synthetic harmonic plane is as follows: Figure 7 As shown, its amplitude is 0.3Udc, and its component in the fundamental wave plane is zero, which is called the virtual zero vector VN. The application of the VN vector does not participate in the electromechanical energy conversion in the fundamental wave plane.

[0153] The above three steps are all offline calculations, and the corresponding control set is generated at the beginning of the algorithm control. It only needs to be used in the subsequent control process.

[0154] Step 4: Get the control variable i dand i q (Coordinate transformation module).

[0155] The voltage vector distribution of the dual three-phase motor under single-phase open circuit fault based on the dimension reduction decoupling matrix is ​​derived. Taking the F phase open circuit as an example, the remaining phase open circuits can be obtained by phase shifting the corresponding angles.

[0156] Under normal circumstances, the voltage, current and flux of the dual three-phase motor can be decomposed into different planes by the method of vector space decoupling, namely the fundamental wave plane (αβ plane), the harmonic plane (z1z2 plane) and the zero sequence plane (o1o2 plane), and the corresponding relationship is:

[0157] [f α f β f x f y f o1 f o2 ] T =T 6s [f A f B f C f D f E f F ] T (9)

[0158] Among them, f α 、f β 、f x 、f y 、f o1 and f o2 Represent the vectors of voltage, current and flux in the α-axis, β-axis, x-axis, y-axis, o1-axis and o2-axis respectively, and the decoupling matrix T 6s As shown below:

[0159]

[0160] The 5-phase current i of a dual three-phase motor after a single-phase open circuit fault occurs A ,i B ,i C ,i D ,i E The current sensor is directly obtained, and the dimension reduction decoupling matrix is ​​used to transform each variable of the natural coordinate system into the stationary coordinate system. The transformation process is:

[0161] [i α i β i z i o1 i o2 ] T =T 5s [iA i B i C i D i E ] T (11)

[0162] Among them, i α 、i β 、i z 、i o1 、i o2 is the current of α axis, β axis, z axis, o1 axis and o2 axis; decoupling matrix T 5s As shown below:

[0163]

[0164] For dual three-phase permanent magnet motors, only the fundamental wave component of the αβ subspace participates in the electromechanical energy conversion. In order to simplify the analysis, the stationary coordinate system is transformed into the synchronous rotating coordinate system, and its transformation matrix is:

[0165]

[0166] The above coordinate conversion module calculates the current of the motor at time k in the dq rotating coordinate system and

[0167] Step 5: Calculate the duty cycle of VV.

[0168] After the F phase open circuit fault, the voltage equation of the dual three-phase permanent magnet motor in the dq coordinate system is:

[0169]

[0170] Among them, u d 、u q 、u z are the d-axis, q-axis, and z-axis voltages respectively; i d 、i q 、i z are d-axis, q-axis, and z-axis currents respectively; L d , L q , L l are the inductances of the d-axis, q-axis, and z-axis respectively; Rs is the stator resistance; ω e is the electrical angular velocity, ψ f is the permanent magnet flux.

[0171]

[0172] Where θ is the rotor position angle.

[0173] From the voltage equation, we can get: the change rate of the dq axis current under different virtual voltage vectors is as follows (current change rate module):

[0174]

[0175] The present invention takes the q-axis reference current as the prediction target and adopts the deadbeat principle to calculate the duty cycle of VV (deadbeat duty cycle calculation module), and the formula is as follows:

[0176]

[0177] in, is the reference value of q-axis current; is the actual value of the q-axis current at time k; s VV When the fundamental wave virtual voltage vector acts q The rate of change; s null_vector When the vector is zero, i q The rate of change of VV is the duty cycle of the fundamental wave virtual voltage vector; d null_vector is the duty cycle of the zero vector; T s is the control period. The solution is:

[0178]

[0179] Step 6: Calculate the duty cycle of VN

[0180] In the harmonic plane, a PI controller is used to input the harmonic reference current and obtain the harmonic reference voltage. (Harmonic current controller) The duty cycle is determined according to the proportional relationship between the amplitude of the reference voltage and the amplitude of VN.

[0181]

[0182] Step 7: Reasonably allocate duty cycle (compensated duty cycle module)

[0183] Judgement VV The size of VV When <1, VN participates in control, otherwise VN is disabled;

[0184] After the conditions are met, if 0<d VV +d VN ≤1, then d VV and d VN remains unchanged, and the remaining duty cycle is compensated by the zero vector, i.e. d null_vector =1-d VV -d VN ;

[0185] If d VV +dVN >1, then d VN =1-d VV , d null_vector =0.

[0186] Finally, the distribution of each voltage vector in a control cycle is as follows: Figure 8 shown.

[0187] Step 8: Fault-tolerant prediction model.

[0188] The forward Euler method is used to discretize the voltage equation after the fault and obtain the prediction model:

[0189]

[0190] Among them, the superscript "k" represents the value of the corresponding variable at time k; the superscript "k+1" represents the value of the corresponding variable at time k+1;

[0191] T s To control the cycle;

[0192] In order to compensate for the "one-shot delay" characteristic of the digital system, the two-step prediction method adopted by the delay compensation module is used for delay compensation, that is, one more step of prediction. The final prediction model is:

[0193]

[0194] Among them, the superscript "k+2" represents the value of the corresponding variable at time k+2;

[0195] The value function used in the prediction model is:

[0196]

[0197] Step 9: Dual-plane closed-loop control.

[0198] After the above steps, the final required voltage vector can be expressed as three parts: VV, VN and V null Its expression needs to be modulated and generated by the bridge arm switch state module. The duty cycle of each bridge arm opening and closing is the sum of the duty cycles of each voltage vector corresponding to the phase, that is,

[0199]

[0200] in, Indicates the final opening duty cycle of the corresponding phase bridge arm; Indicates the duty cycle of each bridge arm of the virtual voltage vector; Represents the duty cycle of each bridge arm of the virtual zero vector; since the bridge arm switch state of the zero vector is selected as "00000", this part is omitted in the above formula.

[0201] Substitute the 12 virtual voltage vectors VV into the prediction model one by one, select the optimal voltage vector and its duty cycle, and inject the modified VN into the harmonic plane to complete the entire control system. So far, the fundamental plane adopts the predictive current control based on the deadbeat duty cycle for closed-loop control, and the harmonic plane adopts the method based on the PI controller for closed-loop control. The two obtain the corresponding d VV and d VN Then it is output to the PWM module to complete the motor control.

[0202] Fig. 9 This is the experimental waveform of the traditional harmonic closed-loop control method under the F phase open circuit fault. It can be seen from the figure that the harmonic current i y It is sinusoidal in shape because the harmonic current and fundamental current are coupled under fault conditions, making i y =-i β , and the torque pulsation reaches 7.21Nm. Fig.10 is the experimental waveform diagram of the present invention, and Fig. 9 In comparison, the steady-state performance of the motor has been improved in all aspects. The detailed parameters are shown in Table 3.

[0203] Table 3 Parameter index comparison

[0204]

[0205] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. A fault-tolerant predictive control system for a dual three-phase permanent magnet motor, characterized in that: It includes a speed controller, a current change rate module, a delay compensation module, a deadbeat duty cycle calculation module, a value function module, a harmonic current controller, a compensated duty cycle module, a bridge arm switch state module, a PWM module, a coordinate transformation module, an inverter module, a position sensor and a dual three-phase permanent magnet motor; The coordinate transformation module is connected to the inverter module via a current sensor and is used to transform the natural coordinate system variables of each phase into the rotating coordinate system variables; The speed controller is controlled by PI to obtain the q-axis reference current, the input of which is the error between the given speed and the actual speed, and the output of which is the reference value of the q-axis current; The input end of the delay compensation module is connected to the coordinate transformation module, and the output end is connected to the current change rate calculation module, so as to compensate for the "one-shot delay" problem caused by the sampling of the digital system; The input end of the current change rate calculation module is connected to the delay compensation module and the position sensor, and the output end is connected to the deadbeat duty cycle calculation module to obtain the change rate of the dq axis current when different voltage vectors act; The deadbeat duty cycle calculation module has an input end connected to the current change rate calculation module, and an output connected to the value function module, and is used to output the duty cycle of each voltage vector acting on the fundamental wave plane; The input end of the value function module is connected to the deadbeat duty cycle calculation module, which is used to traverse all fundamental wave candidate voltage vectors and their duty cycles, and input the selected optimal voltage vector and its duty cycle; The speed controller, delay compensation module, current change rate calculation module, deadbeat duty cycle calculation module and value function module together constitute a prediction model, which functions to control the variables of the fundamental wave plane; The harmonic current controller is controlled by PI to obtain the z-axis reference voltage. Its input is the error between the given z-axis current and the actual z-axis current, and its output is the reference value of the z-axis voltage. The duty cycle that needs to be compensated for the final harmonic plane is calculated by comparing it with the synthesized virtual zero vector, forming a harmonic compensation module. The input end of the compensated duty cycle module is connected to the prediction model and the harmonic compensation module, and its function is to integrate the duty cycles required by the fundamental wave plane and the harmonic plane respectively, and distribute them reasonably; The output of the compensated duty cycle module is connected to the bridge arm switch state module; The input end of the bridge arm switch state module is connected to the compensated duty cycle module, and the output end is connected to the PWM module; The PWM module is connected to the inverter module to perform signal modulation and current output, thereby driving the dual three-phase permanent magnet motor to operate.

2. The control method for a fault-tolerant predictive control system for a dual three-phase permanent magnet motor according to claim 1, characterized in that: The steps include: Step 1) using the dimension reduction decoupling matrix to derive the distribution of the voltage vector after the fault; Step 2) constructing a virtual voltage vector (Virtual vector, VV) with zero harmonic component according to the voltage distribution; Step 3) constructing a virtual null-vector (VN) with a fundamental component of zero according to the voltage distribution; Step 4) using the coordinate transformation module to obtain the control variables required for control; Step 5) The fundamental wave plane uses the deadbeat current control to calculate the duty cycle d of VV VV ; Step 6) The harmonic plane uses a PI controller to obtain the harmonic reference voltage, and determines the required duty cycle d of VN through the harmonic reference voltage and the amplitude of VN VN ; Step 7) Adjust according to the action time of VV and VN to ensure that VV and VN are reasonably arranged within a control cycle; Step 8) deriving a single-phase open circuit prediction model of a dual three-phase permanent magnet motor based on a dimensionality reduction decoupling matrix; Step 9) Complete the post-fault dual-plane closed-loop control under the new virtual vector control set (VV and VN).

3. The method according to claim 2, characterized in that Step 1) specifically includes: deriving the voltage vector distribution of the dual three-phase motor under the single-phase open circuit fault based on the dimension reduction decoupling matrix, taking the F phase open circuit as an example, the remaining phase open circuits can be obtained by phase shifting the corresponding angles, Under normal circumstances, the voltage, current and flux of the dual three-phase motor can be decomposed into different planes by the vector space decoupling method, namely the fundamental wave plane (αβ plane), the harmonic plane (z1z2 plane) and the zero sequence plane (o1o2 plane), and the corresponding relationship is: [f α f β f x f y f o1 f o2 ] T =T 6s [f A f B f C f D f E f F ] T (1) Where, f α 、f β 、f x 、f y 、f o1 and f o2 Represent the vectors of voltage, current and flux in the α-axis, β-axis, x-axis, y-axis, o1-axis and o2-axis respectively, and the decoupling matrix T 6s As shown below: When an open circuit fault occurs on phase F, the dimension reduction decoupling matrix is ​​used to re-decompose the vector space, and the relationship between the voltage vectors can be expressed as follows: [u α u β u z u o1 u o2 ] T =T 5s [u A u B u C u D u E ] T (3) Where u A ~u E are the phase voltages of phases A to F, u α ~u o2 are the voltage components of the corresponding planes, and the dimension reduction decoupling matrix T 5s for: The dual three-phase motor adopts the center point isolation connection mode, so the zero sequence plane can be ignored, and equation (3) can be transformed into: Among them, u α ~u z are the voltage components of the corresponding planes, u A 、u B 、u C are the phase voltages of phase A, phase B and phase C, u DE is the potential difference between phase D and phase E; The dual three-phase motor is driven by a two-level six-phase voltage source inverter, and the relationship between its phase voltage and switch state is: Among them, U dc Indicates bus voltage, S A ~S E Represents the switching state of the upper bridge arm switch tube of each phase of the inverter. "1" means that the upper bridge arm is turned on and the lower bridge arm is turned off; "0" means that the upper bridge arm is turned off and the lower bridge arm is turned on. The numbering of the basic voltage vector is based on S A S B S C S D S E The switch status combination is expressed in decimal. For example, if the switch status is "10101", the number is 21, and v 21 express; According to (5) and (6), we can deduce: It can be deduced that the switching state of the voltage vector after the fault is 2 5 = 32, and its components on the α-axis, β-axis and z-axis are shown in the following table: Table 1 Voltage vector distribution under phase F open circuit fault condition Table 1 shows the voltage vector distribution under the F phase open circuit fault condition, where v 00 、v 01 ,…v 31 The meaning is the voltage vector in the 32 switch states after the fault.

4. The method according to claim 3, characterized in that The specific steps of step 2) include: selecting adjacent maximum three vectors as the basic voltage vectors for synthesizing the fundamental plane virtual voltage vector. The synthesis concept of the fundamental plane virtual voltage vector is: the sum of all voltage vectors in the harmonic plane components is equal to zero, which is expressed as follows: Among them, v 1st , v 2nd , and v 3rd where V represents the first, second and third effective voltage vectors respectively; i is the ith virtual voltage vector; the superscripts "α", "β" and "z" represent the components of the voltage vector on the α-axis, β-axis and z-axis respectively; D1, D2, D3 are the duty cycles of the first, second and third effective voltage vectors, respectively, and D0 is the duty cycle of the zero vector; D0, D1, D2, D3∈[0,1]; Finally, the distribution of the fundamental wave plane virtual voltage vector is shown in Table 2: Table 2 Distribution of fundamental wave plane virtual voltage vector Twelve fundamental plane virtual voltage vectors are synthesized, with an amplitude of 0.29Udc and an angle of 30° between the two vectors.

5. The method according to claim 2, characterized in that: The specific steps of step 3) include: According to Table 1, we can find that the vector v 15 The direction of v2 is opposite to that of v2 on the α axis, but the same on the z axis. Therefore, the action ratio of the two voltage vectors can be reasonably configured so that the sum of their components on the α axis is zero and they remain negative on the z axis. Similarly, v 29 and v 16 , thus synthesizing a positive effective z-axis voltage vector again. After calculation, the synthesis principle is as follows: Among them, v positive represents the virtual zero vector effective in the positive direction, v negative Indicates a virtual zero vector that is effective in the negative direction; The amplitude of the virtual voltage vector of the synthetic harmonic plane is 0.3Udc, and its component in the fundamental wave plane is zero, which is called the virtual zero vector VN. The application of the VN vector does not participate in the electromechanical energy conversion of the fundamental wave plane.

6. The method according to claim 2, characterized in that The specific steps of step 4) include: The 5-phase current i of a dual three-phase motor after a single-phase open circuit fault occurs A ,i B ,i C ,i D ,i E The current sensor is directly obtained, and the dimension reduction decoupling matrix is ​​used to transform each variable of the natural coordinate system into the stationary coordinate system. The transformation process is: [i α i β i z i o1 i o2 ] T =T 5s [i A i B i C i D i E ] T (10) Where i α 、i β 、i z 、i o1 、i o2 is the current of α axis, β axis, z axis, o1 axis and o2 axis; the decoupling matrix T 5s As shown below: For dual three-phase permanent magnet motors, only the fundamental wave component of the αβ subspace participates in the electromechanical energy conversion. In order to simplify the analysis, the stationary coordinate system is transformed into the synchronous rotating coordinate system, and its transformation matrix is: The above coordinate conversion module calculates the current of the motor at time k in the dq rotating coordinate system and 7. The method according to claim 2, characterized in that The specific steps of step 5) include: After the F phase open circuit fault, the voltage equation of the dual three-phase permanent magnet motor in the dq coordinate system is: Among them, u d 、u q 、u z are the d-axis, q-axis, and z-axis voltages respectively; i d 、i q 、i z are d-axis, q-axis, and z-axis currents respectively; L d , L q , L l are the inductances of the d-axis, q-axis, and z-axis respectively; Rs is the stator resistance; ω e is the electrical angular velocity, ψ f is the permanent magnet flux; Where θ is the rotor position angle; From the voltage equation, we can get: The current change rate module is used to obtain the change rate of the dq axis current under different virtual voltage vectors as follows: Taking the q-axis reference current as the prediction target, the deadbeat duty cycle calculation module is used to calculate the duty cycle of VV, and the formula is as follows: in, is the reference value of q-axis current; is the actual value of the q-axis current at time k; s VV When the fundamental wave virtual voltage vector acts q The rate of change; s null_vector When the vector is zero, i q The rate of change of VV is the duty cycle of the fundamental wave virtual voltage vector; d null_vector is the duty cycle of the zero vector; T s is the control period, and we get:

8. The method according to claim 2, characterized in that: The specific steps of step 6) include: In the harmonic plane, a PI controller is used to construct a harmonic current controller, and the harmonic reference current is input to obtain the harmonic reference voltage. The duty cycle is determined according to the proportional relationship between the amplitude of the reference voltage and the amplitude of VN.

9. The method according to claim 2, characterized in that: The specific steps of step 7) include: Judgement VV The size of VV When <1, VN participates in control, otherwise VN is disabled; After the conditions are met, if 0<d VV +d VN ≤1, then d VV and d VN remains unchanged, and the remaining duty cycle is compensated by the zero vector, i.e. d null_vector =1-d VV -d VN ; If d VV +d VN >1, then d VN =1-d VV , d null_vector =0; Finally determine the allocation method of VV and VN.

10. The method according to claim 2, characterized in that The specific steps of step 8) include: The forward Euler method is used to discretize the voltage equation after the fault to obtain the prediction model: The superscript "k" represents the value of the corresponding variable at time k; the superscript "k+1" represents the value of the corresponding variable at time k+1; T s To control the cycle; In order to compensate for the "one-shot delay" characteristic of the digital system, the two-step prediction method adopted by the delay compensation module is used for delay compensation, that is, one more step of prediction. The final prediction model is: Among them, the superscript "k+2" represents the value of the corresponding variable at time k+2; The specific steps of step 9) include: After the above steps, the final required voltage vector can be expressed as three parts: VV, VN and V null , its expression needs to be modulated and generated by the bridge arm switch state module. The duty cycle of each bridge arm opening and closing is the sum of the duty cycles of each voltage vector corresponding to the phase, that is, in, Indicates the final opening duty cycle of the corresponding phase bridge arm; Indicates the duty cycle of each bridge arm of the virtual voltage vector; Represents the duty cycle of each bridge arm of the virtual zero vector; since the bridge arm switch state of the zero vector is selected as "00000", this part is omitted in the above formula; Substitute the 12 virtual voltage vectors VV into the prediction model one by one, select the optimal voltage vector and its duty cycle, and inject the modified VN into the harmonic plane to complete the entire control system. So far, the fundamental plane adopts the predictive current control based on the deadbeat duty cycle for closed-loop control, and the harmonic plane adopts the method based on the PI controller for closed-loop control. The two obtain the corresponding d VV and d VN Then it is output to the PWM module to complete the motor control.

Images