Five-phase brushless direct current motor torque control method

Abstract

The invention discloses a torque control method for a five-phase brushless direct current motor, and relates to the field of motor control. The method comprises the following steps: determining a commutation interval of a rotor according to the position of the rotor, obtaining a predicted torque corresponding to each vector in an alternative vector set of the commutation interval, taking the predicted torque as an input, and determining an optimal vector in the alternative vector set of the commutation interval through a minimum value function; and determining a non-commutation interval in which the rotor is located according to the position of the rotor, obtaining a predicted torque corresponding to each vector in the alternative vector set of the non-commutation interval, taking the predicted torque as an input, and determining an optimal vector in the alternative vector set of the non-commutation interval through a minimum value function. According to the control method, phase current fluctuation can be effectively reduced in a commutation interval and a non-commutation interval, and a better torque ripple suppression effect is achieved.

Description

Technical Field

[0001] The invention relates to the field of motor control, and in particular to a five-phase brushless DC motor torque control method. Background Art

[0002] In some scenarios with high requirements for motor performance and harsh working environments, traditional three-phase brushless DC motors are difficult to be widely used due to their large torque pulsation and difficulty in achieving fault-tolerant control. Multi-phase permanent magnet motors have shown great application potential in the ship electric propulsion and electric vehicle industries due to their high reliability and high torque density. Multi-phase permanent magnet motors have many similarities with three-phase permanent magnet motors. Like other motors, permanent magnet synchronous motors and brushless DC motors can also be produced and used as multi-phase motors. In recent years, multi-phase brushless DC motors have obvious advantages at high power, but their commutation torque fluctuations are still large.

[0003] In order to solve this problem, a variety of control strategies have been developed in the prior art, including vector control, direct torque control and model predictive control. Since the vector control of the five-phase motor can be converted into the DC part of the d1, q1, d3, and q3 axes after the extended Clark and Park transformation, the plane component of the harmonics must be considered in the vector control of the multi-phase motor. Since there may be serious high-order harmonic currents in the stator windings of the multi-phase motor, which affects the stability and dynamic quality of the entire control system, the existing vector control method needs to take into account harmonic current suppression. Direct torque control strategies are mainly divided into two categories. One is the traditional direct torque strategy, which selects the corresponding vector through the switch table and uses the hysteresis loop width as the standard for changing the vector selection. The other is to combine other modulation methods with direct torque, keep the switching frequency unchanged while controlling the torque, and reduce the switching loss of the power tube. The control effect of model predictive control on multi-parameter and nonlinear systems is consistent with the multi-phase motor control strategy. However, the existing multi-phase motor control methods have the advantage that more voltage vectors can be controlled due to the increase in the number of phases, thereby controlling the motor with a higher degree of freedom. However, this also leads to increased interference from multiple harmonics. How to achieve a control method that strikes a balance between the two is an urgent problem to be solved. Summary of the invention

[0004] In view of the above-mentioned defects of the prior art, the present invention provides a five-phase brushless DC motor torque control method, comprising the following steps:

[0005] S1, judging whether it is at the phase switching moment according to the phase current state, if so, executing step S2, if not, executing step S3;

[0006] S2, determining the commutation interval in which the rotor is located according to its position, obtaining the predicted torque corresponding to each vector in the candidate vector set of the commutation interval, taking the predicted torque as input, and determining the optimal vector in the candidate vector set of the commutation interval through a minimum value function;

[0007] S3. Determine the non-commutation interval according to the position of the rotor, obtain the predicted torque corresponding to each vector in the candidate vector set of the non-commutation interval, use the predicted torque as input, and determine the optimal vector in the candidate vector set of the non-commutation interval through a minimum cost function.

[0008] Furthermore, the method for judging whether it is at the commutation moment according to the phase current state includes: taking the change of the Hall signal of the on phase as the starting point of the commutation interval prediction model, and reducing the off phase current to zero as the starting point of the non-commutation interval prediction model.

[0009] Furthermore, the candidate vector set of the commutation interval is: during the commutation process, each phase bridge arm in the inverter undergoes three change stages, and three voltage vectors corresponding to the three change stages.

[0010] Furthermore, the method for obtaining the predicted torque corresponding to each voltage vector in the candidate vector set of the commutation interval includes:

[0011] Under the voltage vector corresponding to the current change stage, the electromagnetic torque is calculated as:

[0012] Among them, o represents the on phase, c represents the off phase, and s represents the non-commutation phase;

[0013] Get the shutdown phase current change rate The rate of change of the conduction phase current

[0014] The torque change rate of the motor in the commutation interval is:

[0015] pass The torque change rate is discretized to obtain the predicted torque T at the next moment. e (k+1).

[0016] Furthermore, the candidate vector set of the non-commutation interval is: four voltage vectors selected from all the switching vectors in each conduction interval based on the switching times and voltage utilization factors during vector switching.

[0017] Furthermore, the method for obtaining the predicted torque corresponding to each voltage vector in the candidate vector set of the non-commutation interval includes:

[0018] The electromagnetic torque under the current voltage vector is calculated as:

[0019] Among them, e f1 、e f2 They represent the reverse electromotive force of the upper bridge arm, e n1 、e n2 They represent the reverse electromotive force of the lower bridge arm, i represents the conduction phase current, and Ω is the mechanical angular velocity of the rotor;

[0020] Get the current change rate of the conducting phase Then we can get the torque change rate

[0021] pass The torque change rate is discretized to obtain the predicted torque T at the next moment. e (k+1).

[0022] Furthermore, the value function g is:

[0023] g=w 1 |T e * -T e (k+1)|+w 2 K;

[0024] Among them, w 1 and w 2 are weight coefficients; T e * is the torque reference value; K is the switching frequency parameter.

[0025] Furthermore, the switching frequency parameter K is:

[0026]

[0027] Among them, T e (k) is the torque at the current moment, T e top is the upper boundary of the torque loop width, T e bot is the lower boundary of the torque loop width.

[0028] Furthermore, the control method includes a fault handling step, and when a fault occurs, the commutation area candidate vector set and the non-commutation area candidate vector set under the fault-tolerant working condition are updated.

[0029] Furthermore, the fault includes a single-phase loss fault.

[0030] Compared with the prior art, the present invention has the following technical effects:

[0031] 1. The torque pulsation is suppressed by selecting the optimal vector in the non-commutation interval and the commutation interval respectively, and the alternative vector table is optimized according to the torque change rate under different vectors and the number of switches between each vector, thereby reducing the amount of calculation.

[0032] 2. The present invention aims at the problem of high switching frequency in finite set model prediction and improves the minimum value function. While suppressing torque pulsation, it can take into account the constraints on the switching frequency and reduce system losses.

[0033] 3. The control method of the present invention can effectively reduce phase current fluctuations in both the commutation interval and the non-commutation interval, and has a better torque pulsation suppression effect.

[0034] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the overall process of a specific embodiment of the present invention;

[0036] Figure 2 is an equivalent circuit diagram of a five-phase brushless DC motor according to a specific embodiment of the present invention;

[0037] Figure 3 It is the equivalent circuit diagram of the commutation process;

[0038] Figure 4 This is an equivalent circuit diagram of AB / DE conduction during the commutation process of a specific embodiment of the present invention;

[0039] Figure 5 This is an equivalent circuit diagram of AB conduction during the commutation process of a specific embodiment of the present invention;

[0040] Figure 6 This is an equivalent circuit diagram of AB / CDE conduction during the commutation process of a specific embodiment of the present invention;

[0041] Figure 7 is an equivalent circuit diagram of a specific embodiment of the present invention in a normal conduction state;

[0042] Figure 8 This is an equivalent circuit diagram of AB / C conduction when not switching phases according to a specific embodiment of the present invention;

[0043] Fig. 9 1 is an equivalent circuit diagram of a conduction state under a zero vector 1100 of a specific embodiment of the present invention;

[0044] Fig.10 1 is an equivalent circuit diagram of a conduction state under zero vector 1000 of a specific embodiment of the present invention;

[0045] Fig.11 It is a schematic diagram of the working principle of vector selection under hysteresis loop according to a specific embodiment of the present invention;

[0046] Fig.12 1 is a schematic diagram of an equivalent circuit of B / DE conduction state under A phase fault in a specific embodiment of the present invention;

[0047] Fig.13 1 is a schematic diagram of an equivalent circuit of a commutation conduction state under a phase A fault in a specific embodiment of the present invention;

[0048] Fig.14 It is a block diagram of a five-phase brushless DC motor fault-tolerant control system based on a new switch table in a specific embodiment of the present invention;

[0049] Fig.15 It is a simulation result of motor load operation at low speed in a specific embodiment of the present invention. Fig.15 a is the phase current, speed and torque waveform under DTC control, Fig.15 b is the phase current, speed and torque waveform under FCS-MPC control;

[0050] Fig.16 is a simulation waveform of a switch tube driving signal at low speed in a specific embodiment of the present invention, Fig.16 a is the switching frequency waveform before optimization, Fig.16 b is the switching frequency waveform after optimization;

[0051] Fig.17 It is a simulation waveform of a motor running under load at high speed in a specific embodiment of the present invention. Fig.17 a is the phase current, speed and torque waveform under DTC control, Fig.17 b is the phase current, speed and torque waveform under FCS-MPC control;

[0052] Fig.18 In a specific embodiment of the present invention, the switching tube driving signal simulation waveform at high speed is shown. Fig.18 a is the switching frequency waveform before optimization, Fig.18 b is the switching frequency waveform after optimization;

[0053] Fig.19 In a specific embodiment of the present invention, the simulation waveform under the sudden change of speed is: Fig.19 a is the phase current, speed and torque waveform under DTC control, Fig.19 b is the phase current, speed and torque waveform under FCS-MPC control;

[0054] Fig. 20 In a specific embodiment of the present invention, the simulation waveform under torque mutation is: Fig. 20 a is the torque and speed waveform under DTC control, Fig. 20b is the torque and speed waveform under FCS-MPC control;

[0055] Fig.21 is a torque variation waveform in a single-phase open circuit state in a specific embodiment of the present invention, Fig.21 a is the torque variation waveform under the non-fault-tolerant control strategy, Fig.21 b is the torque variation waveform under the fault-tolerant control strategy;

[0056] Fig. 22 In a specific embodiment of the present invention, the waveforms of the current changes of the remaining four phases in a single-phase open circuit state are shown in FIG. Fig. 22 a is the torque variation waveform under the non-fault-tolerant control strategy, Fig. 22 b is the torque variation waveform under the fault-tolerant control strategy;

[0057] Fig.23 It is a waveform of single-phase current change before and after an open circuit fault in a specific embodiment of the present invention. Fig.23 a is the current change waveform of phase C and phase E under the non-fault-tolerant control strategy, Fig.23 b is the current change waveform of phase C and phase E under the fault-tolerant control strategy;

[0058] Fig.24 It is a waveform of speed change before and after an open circuit fault in a specific embodiment of the present invention. Fig.24 a is the speed change waveform under the non-fault-tolerant control strategy, Fig.24 b is the speed change waveform under the fault-tolerant control strategy;

[0059] Fig.25 It is the dynamic performance verification result of two control strategies under fault conditions in a specific embodiment of the present invention. Fig.25 a is the torque variation waveform under the non-fault-tolerant control strategy, Fig.25 b is the torque variation waveform under the fault-tolerant control strategy. Fig.25 a is the speed change waveform under the non-fault-tolerant control strategy, Fig.25 b is the speed change waveform under the fault-tolerant control strategy. DETAILED DESCRIPTION

[0060] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0061] like Figure 1As shown, in a specific embodiment, a five-phase brushless DC motor torque control method is provided, and the circuit schematic diagram of the five-phase brushless DC motor is shown in FIG. Figure 2 As shown, the inverter circuit includes five bridge arms, each bridge arm is composed of two switch tubes connected in series, and the mathematical model of the five-phase brushless DC motor is:

[0062] Voltage equivalent equation: R is the equivalent resistance in the stator winding, i x is the phase current flowing through each phase of the stator winding, u x is the voltage at each phase terminal of the stator winding, L is the self-inductance of the winding, M is the mutual inductance between each phase winding, e x It is the reverse electromotive force contained in the interaction between each phase in the stator winding and the rotor permanent magnet.

[0063] Electromagnetic torque equation: T e is the electromagnetic torque generated, Ω is the mechanical angular velocity of the rotor, P e is the electromagnetic power generated in the current state, which is: P e =e A i A +e B i B +e C i C +e D i D +e E i E , E represents the magnitude of the back electromotive force, and I represents the magnitude of the phase current.

[0064] Mechanical equation of motion: J is the moment of inertia, B is the damping coefficient, p is the number of rotor permanent magnet poles in the motor, ω is the electrical angular velocity of the rotor permanent magnet,

[0065] Back electromotive force equation: E = K e ω; K e is the back electromotive force coefficient, is the number of winding turns in the stator, α i is the polar arc coefficient, Φ δ is the unipolar flux of the permanent magnet.

[0066] The steps include:

[0067] S1. Determine whether it is a phase switching moment according to the phase current state. If so, execute step S2; if not, execute step S3.

[0068] The five-phase brushless DC motor in this embodiment adopts a four-four conduction mode. In the non-commutation interval, there will be current flowing through the four-phase stator winding, of which the current of two-phase stator winding is the input phase, that is, the upper bridge arm is turned on, and the current of the other two-phase stator winding is the output phase, that is, the lower bridge arm is turned on, and the remaining one phase is not turned on as a non-excitation phase; when commutating, one conduction phase is disconnected, and the other off phase is turned on, so there will be a phenomenon that the current passes through the five-phase winding at the same time in the circuit. It can be seen that in the non-commutation interval and the commutation interval, the control process of the conduction phase and the off phase is completely different. The present invention establishes torque prediction models and corresponding alternative voltage vector sets in different operating intervals respectively. The accuracy of switching between different prediction models has an important influence on the suppression of torque pulsation. In this embodiment, the change of the Hall signal of the on phase is used as the starting point of the commutation interval prediction model, and the reduction of the off phase current to zero is used as the starting point of the non-commutation interval prediction model, so that different prediction models in different intervals can be accurately switched. The detection conditions of different model switching points are shown in Table 1.

[0069] Table 1: Commutation process parameter changes

[0070]

[0071] S2. Determine the commutation interval according to the position of the rotor, obtain the predicted torque corresponding to each vector in the candidate vector set of the commutation interval, use the predicted torque as input, and determine the optimal vector in the candidate vector set of the commutation interval through a minimum cost function.

[0072] Taking the commutation interval from AB / CD conduction to AB / DE conduction as an example, the equivalent circuit of the commutation process is as follows Figure 3 As shown. Current flows through all the windings of the motor, but in order to ensure successful commutation and the minimum number of switches between different vectors, it is necessary to keep the power tube on the non-commutation side unchanged, and change the output torque of the commutation zone by controlling the switch state of the power tube on the commutation side. There are 2 to the power of 3, that is, 8 combinations of the switch state combination of the power tube on the commutation side. Since the torque change rate generated is the same when the number of power tubes turned on is the same, this embodiment selects three voltage vectors corresponding to three of the conduction states as the candidate vector set.

[0073] Since current flows through the windings of all phases in the five-phase brushless DC motor during the commutation period, the on phase can be defined as o, the off phase as c, and the non-commutation winding as s according to the state of the power tube of each phase winding at this time. At this time, the electromagnetic torque T e for:

[0074] From the above formula, we can see that if the reverse electromotive force of the on phase is equal to that of the off phase, that is, e o =e cIn this case, the current of the open phase and the closed phase have a decisive effect on the torque. When the two cannot offset each other, the output torque will also change accordingly. Therefore, ignoring the phase resistance voltage drop, the voltage equation of each phase winding of the motor during the commutation process is listed, and the change rate of the closed phase current in the stator winding under the commutation state is deduced. The rate of change of the conduction phase current for:

[0075]

[0076] The torque change rate of the motor in the commutation interval for:

[0077]

[0078] pass The torque change rate is discretized to obtain the predicted torque T at the next moment. e (k+1).

[0079] Where: T s is the sampling period, T e (k) is the torque at time k, T e (k+1) is the output torque at time k+1, and E is the back electromotive force amplitude of the conducting phase.

[0080] State 1: Keep the A and B phases open, disconnect the C phase, and open the D and E phases. At this time, the C phase current cannot drop to zero immediately, and the current flows through the diode D5 to form a loop. Ignoring the diode voltage drop, the terminal voltage U co =U dc , the current direction of phase A and phase B winding is inflow, so U ao =U bo =U dc , and the direction of the D-phase and E-phase winding current is outflow, so U do =U eo =0, the inverter circuit conduction state is as follows Figure 4 shown.

[0081] In state 1, the rate of change of the E-phase winding current is:

[0082] The change rate of the C phase winding current is:

[0083] The current change rate of the non-commutated A phase winding is:

[0084] In the commutation interval, since the reverse electromotive force of the on phase and the off phase can be considered to be the same, the torque change rate can be obtained according to the non-commutation A phase winding current change rate and electromagnetic torque formula:

[0085] pass By discretizing the torque change rate, the predicted torque T at the next moment can be obtained. e (k+1) is:

[0086]

[0087] State 2: Under the premise of keeping the AB phase on, another circuit conduction state is that the power tubes of the lower bridge arms of phases C, D and E are all turned off, while the upper bridge arms of phases A and B remain on. At this time, the current flows through T 5 、T 7 and T 9 The body diode of the five-phase brushless DC motor returns to the busbar, and the five-phase brushless DC motor inverter is in the freewheeling state. If the diode voltage drop is ignored, U co =U do =U eo =U dc The equivalent circuit is as follows Figure 5 Shown

[0088] In state 2, the rate of change of the current of phase A winding is:

[0089] Since it is in the freewheeling state at this time, the on-phase current is basically the same as that of phase A, and the off-phase winding current change rate is consistent. Therefore, taking phase E as an example, the phase current change rate can be obtained as:

[0090] The torque change rate in this state is:

[0091] Discretizing the above formula, the predicted torque at the next moment can be obtained as:

[0092] State 3: If the circuit conduction state is changed to keep the upper bridge arms of phases A and B open, and the lower bridge arms of the remaining phases open, the current flows in from phases A and B and flows out from phases C, D, and E. At this time, U ao =U bo =U dc , U co =U do =U ec . Its inverter circuit conduction state Figure 6 shown.

[0093] The change rate of the current in the A phase winding is:

[0094] The E-phase winding current change rate is:

[0095] From this, the torque change rate at the next moment can be obtained as:

[0096] Discretizing the above formula, the predicted torque at the next moment can be obtained as:

[0097]

[0098] The predicted torque obtained under the above three states is used as input, and the optimal vector in the commutation interval candidate vector set is determined through the minimum value function.

[0099] The power tube state in the inverter circuit is defined as 0 or 1, 0 represents the power tube is turned off, 1 represents the power tube is turned on, and through the above method, the candidate vector set in other commutation intervals is obtained, and then the commutation interval vector selection table is constructed as shown in Table 2.

[0100] Table 2 Commutation interval vector selection table

[0101]

[0102] S3. Determine the non-commutation interval according to the position of the rotor, obtain the predicted torque corresponding to each vector in the candidate vector set of the non-commutation interval, use the predicted torque as input, and determine the optimal vector in the candidate vector set of the non-commutation interval through a minimum cost function.

[0103] When the motor is in the non-commutation interval, the output torque change rate will also change under different switching states. Since the five-phase brushless motor has four-phase windings that can be turned on in the non-commutation interval, there are theoretically 16 switching states. However, through experimental analysis, it is found that when the number of power tubes turned on is the same, different vectors have the same effect on torque. For example, there is no difference in the torque change rate between 1011 and 0111 (1 represents on, 0 represents off). Therefore, considering the number of switches and voltage utilization when different vectors are switched, taking AB / CD phase conduction as an example, non-zero vectors 1111 and 1110 and zero vectors 1100 and 1000 are selected as the candidate vector set for the non-commutation interval.

[0104] According to the current direction of each phase winding, the upper bridge arm conduction phase winding is defined as f 1 、f 2 The conducting phase windings of the lower bridge arm are n 1 、n 2 , and the non-conducting phase g, at this time the five-phase brushless DC motor outputs electromagnetic torque T e It can be expressed as:

[0105]

[0106] Where: e f1 、e f2They represent the reverse electromotive force of the upper bridge arm, e n1 、e n2 They represent the reverse electromotive force of the lower bridge arm, i represents the conduction phase current, and Ω is the mechanical angular velocity of the rotor.

[0107] The voltage equation of the conducting phase winding terminal of the five-phase brushless DC motor in the non-commutation interval is:

[0108]

[0109] Because I f1 =i f2 =-i n1 =-i n1 =i, add the equations in the above formula to find the neutral point voltage u N for:

[0110]

[0111] Then we can get the conduction phase current change rate for:

[0112]

[0113] After neglecting the resistance, the phase current change rate can be expressed as:

[0114] Then the torque change rate of the brushless DC motor in the non-commutation range is obtained for:

[0115]

[0116] When the upper bridge arm conduction phase and the lower bridge arm conduction phase are turned on or off at the same time, that is, u f1 =u f2 、u n1 =u n2 、e f1 =e f2 、e n1 =e n2 , then the above formula can be simplified to:

[0117] Under the non-zero vector 1111, the circuit is in the AB / CD phase conduction state, and the circuit schematic diagram is as follows Figure 7 As shown, at this time, the upper bridge arms of phases A and B are turned on at the same time as the lower bridge arms of phases C and D, so the voltage equations of each phase terminal can be listed as:

[0118]

[0119] Due to e A =e B =E, eC =e D =-E, and the current change rate of the stator winding that is turned on at this time is obtained for:

[0120] Then we can get the torque change rate for:

[0121] according to Torque change rate After discretization, the predicted torque can be obtained as:

[0122]

[0123] When the circuit is in the AB / C conduction state under the non-zero vector 1110, the circuit schematic diagram is as follows Figure 8 As shown, the terminal voltage equation at this time is:

[0124]

[0125] Also through By discretizing it, the predicted torque can be obtained as:

[0126]

[0127] In the same non-commutation region, the zero vector (1100) means that all the power tubes of the upper bridge arm or the lower bridge arm are turned off, while the power tubes on the other side remain turned on. At this time, the current flows into the bus from one end, passes through the turned-on power tube and the anti-parallel diode of the turned-off phase power tube, and returns to the bus. The zero vector (1000) means that only one power tube is turned on, and the rest are all turned off. The freewheeling current is freewheeled through the anti-parallel diode. The two conduction states are as follows: Fig. 9 and Fig.10 shown.

[0128] Depend on Fig. 9 It can be seen that when the circuit is in the first conduction mode, that is, the upper bridge arm remains open and the lower bridge arm is turned off, the current forms a loop with the upper bridge arm through the anti-parallel diode. At this time, the voltage equation of the conduction phase terminal is:

[0129]

[0130] Assume that the absolute value of each phase electromotive force is E. Since the voltages at the upper bridge arm conduction phase and the lower bridge arm conduction phase are both the bus voltage U dc , the stator winding current change rate of the conducting phase is obtained as:

[0131] Then we can get the torque change rate for:

[0132] After discretization, the predicted torque can be obtained as:

[0133] Depend on Fig.10 It can be seen that when the circuit is in the second situation, that is, only one phase of the upper bridge arm is turned on and the phase power tube is turned on, and the current forms a loop through other opposite parallel diodes. At this time, the voltage equation of the conducting phase terminal is:

[0134]

[0135] Similarly, the predicted torque at the next moment can be obtained as:

[0136] The predicted torque obtained under the above four states is used as input, and the optimal vector in the commutation interval candidate vector set is determined through the minimum value function.

[0137] The power tube state in the inverter circuit is defined as 0 or 1, 0 represents the power tube is turned off, 1 represents the power tube is turned on. During normal operation, the vector selection table established in other non-commutation intervals is shown in Table 3.

[0138] Table 3 Vector selection table in non-commutation interval

[0139]

[0140]

[0141] Since the five-phase brushless DC motor control system based on the finite set model prediction mainly controls the power tube switch state, switches between different conduction modes to directly control the torque at the next moment, thereby suppressing the torque pulsation. However, since the switching between different conduction states is achieved by changing the power tube switch state, in the scenario of high-power application, too high switching frequency will cause greater losses, which directly affects the overall efficiency of the system.

[0142] In order to solve the above problems, this paper introduces the principle of hysteresis control into the finite set model prediction and constructs a vector selection method with loop width. And by changing the size of the loop width, the system response speed can be quickly adjusted to meet the control strategy requirements of the five-phase brushless DC motor under high power. The vector selection method after adding hysteresis is as follows Fig.11 As shown in the figure, the torque reference value T e * As the center, set the upper limit of the ring width to be T e top , the lower limit of the ring width is T e bot Using linear extrapolation, the current observation value T e(k) and the predicted value T under the action of each candidate vector e (k+1) and a straight line can be obtained that intersects the upper and lower bounds of the hysteresis loop. The slope of the extended straight line is used as the switching frequency parameter, which is:

[0143]

[0144] At this time, if the predicted values ​​of two or more candidate vectors are within the loop width, the selection of the vector at the next moment is no longer based on the minimum torque error as the only criterion, but the time that the vector can be maintained within the loop width is added as another reference. Therefore, the reconstructed value function is:

[0145] g=w 1 |T e * -T e (k+1)|+w 2 K;

[0146] Among them, w 1 and w 2 are weight coefficients respectively; K is the switching frequency parameter. Fig.11 For example, in the figure, S 1 With S 2 All are within the given ring width, although S 1 The error from the given value is smaller, but S 2 It can stay inside the ring width longer, so S will be selected. 2 The system is acted on at the next moment. Through the design and optimization of the minimum value function, the system can reduce the switching frequency and improve the overall efficiency of the system while suppressing the torque pulsation.

[0147] When a single-phase open circuit fault occurs, the four-four conduction mode of the five-phase brushless DC motor is changed to a two-three conduction mode, thereby forming a new rotating voltage vector and magnetomotive force. However, this commutation mode will result in inconsistent motor output torque before and after commutation, and does not completely solve the problem of torque change caused by phase loss. Secondly, since the back electromotive force in the brushless DC motor is a trapezoidal wave, it is difficult to calculate the current when the faulty phase is replaced by the conduction phase in other sectors. In this embodiment, a new switch table is formulated according to the conduction mode under single-phase fault conditions. When a single-phase open circuit fault occurs, the new switch table can be used to exclude the faulty phase, so that the number of conduction phases before and after commutation remains unchanged. Thereby avoiding torque pulsation before and after commutation due to different conduction phases. Since the number of conduction phases and the circuit state before and after commutation change, it is necessary to reselect the alternative vectors for the commutation zone and the non-commutation zone.

[0148] Take the case of a fault in phase A and the BC / DE conduction interval as an example. At this time, the conduction phase changes to B / DE conduction. The circuit conduction state is as follows: Fig.12 At this time, the upper bridge arm of phase B is turned on, and the lower bridge arms of phases D and E are turned on. Similar to the normal state, the torque can also be changed by changing the conduction state of the switch tube, thereby achieving torque pulsation suppression.

[0149] Take the case where the upper tube of phase B is turned on and phases D and E are turned off as an example. At this time, the voltage at each phase terminal of the motor stator winding is:

[0150]

[0151] Assuming the back electromotive force amplitude is E, add the phases in the above formula to get the neutral point voltage of the circuit at this time:

[0152] At this time, the current change rate in the single-phase conducting half bridge is:

[0153] The current change rate in the half-bridge with two phases turned on is:

[0154] The torque change rate at this time is:

[0155] By discretizing it, the predicted torque at the next moment can be obtained as:

[0156] Similarly, when the upper bridge arm of phase B is turned on, and the lower bridge arms of phases D and E continue to be turned on, the voltage equations of each phase terminal are:

[0157]

[0158] At this time, the torque change rate of the motor output is:

[0159] If all the three-phase switches through which the current flows are turned off, the voltage equation of each phase terminal when the circuit is on is:

[0160]

[0161] At this time, the change rate of the motor output torque is:

[0162] From the above, it can be seen that when the new switch table under fault condition is adopted, the finite set model prediction can also predict the torque at the next moment in the non-commutation interval according to the different torque change rates under the action of three different vectors, so as to control the torque in the non-commutation interval and reduce the torque pulsation. The voltage vector selection table in the non-commutation interval under the new switch table is shown in Table 4.

[0163] Table 4. Vector selection table for non-commutation interval under phase A open circuit fault

[0164]

[0165] In the commutation interval, taking the commutation from B / DE to BC / E as an example, the circuit conduction state is no longer five-phase conduction compared to the commutation interval before the fault occurs. Since phase A fails at this time, the number of phases that are simultaneously conducted is reduced to four. At this time, the circuit conduction state is as follows Fig.13 shown.

[0166] When the switch state of phases B and E remains unchanged and phases C and D are turned on at the same time, the voltage equation of each phase in the motor stator winding can be listed as:

[0167]

[0168] Adding each term in the above formula, we can get the neutral point voltage U dc / 2, it can be deduced that the current change rate in the stator winding is:

[0169]

[0170] According to the electromagnetic torque formula of the five-phase brushless DC motor, the change rate of the motor output torque at this time is:

[0171]

[0172] Similarly, when the circuit enters the phase-changing state, when other circuits are in the on state, such as phases B and E remain unchanged, phase C is turned on, and phase D is turned off, the rate of change of the motor output torque is:

[0173] The motor torque change rate when phase C and phase D are all turned off is:

[0174] Discretizing the torque change rates of the above three vectors respectively can obtain the predicted torque at the next moment:

[0175]

[0176] In summary, when a single-phase open circuit fault occurs, first changing the switch table in the normal state to a new switch table under fault-tolerant control can eliminate the increase in fault torque pulsation caused by phase loss, and by combining the finite set model prediction with the new switch table, the torque pulsation can be further suppressed. The voltage vector selection table in the commutation interval under the new switch table is shown in Table 5.

[0177] Table 5 Commutation interval vector selection table under phase A open circuit fault

[0178]

[0179] In order to further verify the above embodiment, a five-phase brushless DC motor torque control system is built in a specific embodiment. Fig.14 As shown. The position Hall signal is captured to calculate the actual speed of the current motor. Then the given speed is subtracted from the actual speed obtained through the position signal, and then the torque that needs to be increased or decreased at the next time is obtained through the incremental PI algorithm as the torque loop given value at the next moment. Determine whether the current state is a fault, and make corresponding changes to the vector selection table and the prediction model according to the current motor operation state. Determine the interval where the current motor rotor is located, and different alternative vectors need to be used according to the commutation area and the non-commutation area. Finally, based on the speed loop given torque and the current output torque of the motor, select the optimal output vector, update the system drive signal at the next moment, select the corresponding switch tube to turn on, and complete the closed loop of the control strategy. According to Fig.14 The control system structure diagram is modeled by modularizing each function, and a control system simulation model of a five-phase brushless DC motor based on finite set model prediction is built. In this simulation model, in order to keep the simulation parameters consistent with the experimental motor parameters, a five-phase brushless DC motor body is built, with a rated voltage of 350V DC, a rated speed of 3000 rpm, a back electromotive force coefficient of 0.531V / (rad / s), a motor stator phase resistance of 0.073Ω, a stator phase inductance of 1.35mH, and a pole pair number of 2. The specific parameters of the motor are shown in Table 6.

[0180] Table 6 Parameters of five-phase brushless DC motor

[0181]

[0182] Assuming the initial parameters of the control system are a given speed of 800r / min and a given load torque of 10N·m, simulation comparisons are performed under direct torque control and finite set model predictive control, and the phase current, speed, and torque waveforms under low speed can be obtained. Fig.15 As shown. In terms of current, the current fluctuation under DTC control is about 0.7A, while the current fluctuation using the FCS-MPC control strategy is about 0.4A, which is smaller in comparison. In terms of speed, although both methods can quickly reach a given speed, the speed fluctuation under DTC is significantly greater than that under FCS-MPC. In terms of torque, the torque pulsation under DTC control is 1.7N·m, while the torque pulsation under FCS-MPC control is 0.8N·m, which has a significant inhibitory effect on torque pulsation. The above analysis proves that the control system designed in this paper has good control performance under low speed and load.

[0183] The simulated waveform of the switch tube drive signal at low speed is as follows Fig.16As shown in the figure, before optimization, in order to maximize the suppression of torque pulsation, the switching frequency under the finite set model predictive control strategy is about 9.5kHz, which also means an increase in switching losses. After optimizing the value function in combination with the hysteresis principle, the switching frequency is about 5.5kHz, which is significantly lower than that. It proves that the optimized value function no longer takes the minimum error as the only criterion, but also pays attention to the constraints on the switching frequency.

[0184] Assuming the initial parameters of the control system are a given speed of 2000r / min and a load torque of 10N·m, when a stable state is obtained, the torque simulation waveforms under the two control strategies are as follows: Fig.17 As shown in the figure, under the DTC strategy, the phase current fluctuation is about 0.6A, while the current fluctuation under the FCS-MPC control strategy is only about 0.3A, and the speed fluctuation is also smaller. The torque pulsation is 1.4N·m under DTC control, and 0.7N·m under FCS-MPC control. The torque pulsation and current fluctuation under the FCS-MPC control strategy are significantly reduced, proving that the five-phase brushless DC motor can still achieve good operating performance at high speed through the control system designed in this article.

[0185] The simulation waveform of the switch tube drive signal at high speed is as follows Fig.18 As shown in the figure, when the motor runs at high speed, when the traditional value function is used as the selection criterion, the switching frequency of the power tube is about 8.5kHz, while the optimized value function can reduce the switching frequency of the power tube to about 5kHz. The optimized value function can significantly reduce the switching frequency.

[0186] The initial parameters of the control system are set to a given speed of 1500 r / min and a load torque of 10 N·m. After reaching steady state, the speed is increased to 2000 r / min at 0.25 s. The simulation results are as follows: Fig.19 As shown in the figure. Let the initial parameters of the control system be a given load torque of 5N·m and a given speed of 1200r / min. After the system stabilizes, the load torque is suddenly increased to 10N·m at 0.5s. The simulation results are shown in the figure. Fig. 20 As shown. Fig.19 and Fig. 20 It can be seen that the speed and torque under the two control strategies can achieve the effect of following the given value, but the torque pulsation under the control strategy based on FCS-MPC is smaller, and the speed and torque can quickly follow after the given value changes. It proves that the FCS-MPC control strategy can not only achieve the effect of torque pulsation suppression, but also has good dynamic performance.

[0187] The initial parameters of the control system are set to a given load torque of 5 N·m and a torque of 1000 r / min. In the initial operation, the given torque and speed are kept constant and the motor is operated in a steady state. At 0.3 s, a fault control signal is issued to simulate the operation of the motor in an open circuit state. The operation of the motor is observed under the fault-tolerant control strategy and the fault-tolerant control strategy adopted in this embodiment. The torque change waveform after the fault occurs is shown in Figure 2. Fig.21 As shown in the figure, without adopting the fault-tolerant control strategy, the five-phase brushless DC motor can still continue to operate in the single-phase open circuit state, but the torque pulsation increases to nearly 2.5N·m compared with before the fault. Compared with the torque pulsation before the fault, the torque pulsation increases significantly. After applying the fault-tolerant control strategy, it can be seen from the figure that the torque pulsation after the fault increases less, about 1.5N·m. And there is no obvious peak in the torque waveform. It proves that the strategy has a good effect on suppressing torque pulsation under open circuit fault. The fault-tolerant control strategy proposed in this paper can achieve stable output of motor torque while performing fault-tolerant control on the motor.

[0188] The simulated current waveforms of the remaining four phases except the fault phase A are as follows: Fig. 22 As shown in the figure. After a fault occurs, the current magnitude of the remaining phases will also change due to the change in the number of conducting phases. Fig. 22 In a, the conduction angle of each phase remains unchanged, and the conduction current of the power tube of the bridge arm on the same side as the fault phase increases to nearly twice that before the fault, while the power tube on the different side remains basically unchanged. Fig. 22 In b, due to the change of the switch table and control strategy, the conduction angle changes. Although the current amplitude will also increase, the fluctuation of the current will decrease. The single-phase current waveform is as follows Fig.23 As shown. After a fault occurs in phase A, phases B and E are located on the same side of the bridge arm as the fault phase and are turned on at the same time as the fault phase, so the current becomes larger when turned on than in the normal state, while phases C and D change less. Taking the current of phases C and E as an example, before a single-phase fault occurs, the conduction angles are the same under the two control strategies, but after an open circuit fault occurs, when a non-fault-tolerant control strategy is adopted, the conduction angles of the remaining phases remain unchanged, but the current fluctuates greatly. After adopting the fault-tolerant control strategy, due to the change in the switch table, the conduction angle of phase C is reduced compared to before the fault. The conduction angle of phase E is larger than before the fault, and due to the improvement of the control strategy, the fluctuation of the current is reduced.

[0189] The speed waveform changes under the two control strategies as follows Fig.24As shown. Before the fault occurs, the motor can run stably at 1000r / min. When it runs to 0.3s, a fault signal is issued, causing phase A to be open-circuited. At this time, the motor is operating in a phase-missing state. Although the speed can still follow the given value under the non-fault-tolerant control strategy, the speed fluctuates greatly, and the speed fluctuation is close to 60r / min. After adopting the fault-tolerant control strategy, although the speed drops slightly when the fault occurs, the speed fluctuation is small after stabilization, and the speed fluctuation is about 25r / min. The motor can continue to maintain stable operation under the fault state. If the motor is always in the phase A open-circuit state, the dynamic performance under the two control strategies is verified, and the results can be obtained as follows Fig.25 As shown. Fig.25 a and Fig.25 In b, the speed is kept constant, the initial torque is given as 5N·m, and the load is increased to 10N·m at 0.3s. Although the motor output torque under the two given torques can achieve the effect of following when the given torque changes, under the non-fault-tolerant control strategy, the torque pulsation increases as the load increases, while the torque pulsation under the fault-tolerant control strategy is relatively small, which is more conducive to the stable operation of the motor. Fig.25 c and Fig.25 It can be further verified in Figure d that when the torque remains unchanged, the given speed is increased from 900r / min to 1500r / min at 0.3s. Although the speeds under both control strategies can follow the given speed mutation, the speed fluctuation under the fault-tolerant control strategy is smaller than that of the non-fault-tolerant control at both low and high speeds.

[0190] Based on the above simulation analysis results, the finite set model-based predictive control strategy of the present invention can effectively solve the problem of excessive torque pulsation under the traditional control strategy, and while achieving good suppression of torque pulsation, it can also reduce the switching frequency and reduce the switching loss, and has good controllability of the motor output torque and stability of system efficiency. By comparing the speed fluctuations under different strategies, it can also further prove the torque pulsation suppression effect of the invented control method on the five-phase brushless DC motor when it is in fault-tolerant operation under a single-phase open circuit state.

[0191] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A five-phase brushless DC motor torque control method, characterized in that: The steps include: S1, judging whether it is at the phase switching moment according to the phase current state, if so, executing step S2, if not, executing step S3; S2, determining the commutation interval in which the rotor is located according to its position, obtaining the predicted torque corresponding to each vector in the candidate vector set of the commutation interval, taking the predicted torque as input, and determining the optimal vector in the candidate vector set of the commutation interval through a minimum value function; S3. Determine the non-commutation interval according to the position of the rotor, obtain the predicted torque corresponding to each vector in the candidate vector set of the non-commutation interval, use the predicted torque as input, and determine the optimal vector in the candidate vector set of the non-commutation interval through a minimum cost function.

2. A five-phase brushless DC motor torque control method according to claim 1, characterized in that: The method for judging whether it is at the commutation moment according to the phase current state includes: taking the change of the Hall signal of the on phase as the starting point of the commutation interval prediction model, and reducing the off phase current to zero as the starting point of the non-commutation interval prediction model.

3. A five-phase brushless DC motor torque control method according to claim 1, characterized in that: The candidate vector set of the commutation interval is: during the commutation process, each phase bridge arm in the inverter undergoes three change stages, and three voltage vectors corresponding to the three change stages.

4. A five-phase brushless DC motor torque control method according to claim 3, characterized in that: The method for obtaining the predicted torque corresponding to each voltage vector in the candidate vector set of the commutation interval includes: Under the voltage vector corresponding to the current change stage, the electromagnetic torque is calculated as: Among them, o represents the on phase, c represents the off phase, and s represents the non-commutation phase; Get the shutdown phase current change rate The rate of change of the conduction phase current The torque change rate of the motor in the commutation interval is: pass The torque change rate is discretized to obtain the predicted torque T at the next moment. e (k+1).

5. A five-phase brushless DC motor torque control method according to claim 1, characterized in that: The candidate vector set of the non-commutation interval is: four voltage vectors selected from all the switching vectors in each conduction interval based on the switching times and voltage utilization factors during vector switching.

6. A five-phase brushless DC motor torque control method according to claim 1, characterized in that: The method for obtaining the predicted torque corresponding to each voltage vector in the candidate vector set of the non-commutation interval includes: The electromagnetic torque under the current voltage vector is calculated as: Among them, e f1 、e f2 They represent the reverse electromotive force of the upper bridge arm, e n1 、e n2 They represent the reverse electromotive force of the lower bridge arm, i represents the conduction phase current, and Ω is the mechanical angular velocity of the rotor; Get the current change rate of the conducting phase Then we can get the torque change rate pass The torque change rate is discretized to obtain the predicted torque T at the next moment. e (k+1).

7. A five-phase brushless DC motor torque control method according to claim 1, characterized in that: The value function g is: Among them, w1 and w2 are weight coefficients respectively; T e * is the torque reference value; K is the switching frequency parameter.

8. A five-phase brushless DC motor torque control method according to claim 7, characterized in that: The switching frequency parameter K is: Among them, T e (k) is the torque at the current moment, T e top is the upper boundary of the torque loop width, T e bot is the lower boundary of the torque loop width.

9. A five-phase brushless DC motor torque control method according to claim 1, characterized in that: The control method comprises a fault processing step, and when a fault occurs, a commutation area candidate vector set and a non-commutation area candidate vector set under a fault-tolerant working condition are updated.

10. A five-phase brushless DC motor torque control method according to claim 9, characterized in that: The fault includes a single-phase loss fault.

Images