阅读说明：本技术 一种故障容错的开关磁阻电机无位置传感器控制方法 (Fault-tolerant switched reluctance motor position-sensorless control method ) 是由 陈昊 崔思航 刘亮 董锋 张珂 巩士磊 阎明 张战 袁利 李祥阳 徐天宝 于 2020-06-04 设计创作，主要内容包括：本发明提出了一种故障容错的开关磁阻电机无位置传感器控制方法,适用于不同控制策略的开关磁阻电机控制。该方法采用在开关磁阻电机非励磁相的退磁和闲置区间注入检测脉冲,通过比较脉冲响应电流的上升时间来估计转子的对齐位置。该方法在开关磁阻电机的各相都注入检测脉冲,对于四相开关磁阻电机而言,一个转子周期内将会检测到四相对齐位置,任何一相的对齐位置都可以用来估计其他时刻的转子位置。因此,当开关磁阻电机出现缺相故障时,只要存在一相正常运行,就可以保证转子位置估计的准确性,从而实现开关磁阻电机的故障容错运行。该方法降低了系统的成本,提高了系统的可靠性,具有良好的工程应用价值。(The invention provides a fault-tolerant switched reluctance motor position sensorless control method which is suitable for controlling switched reluctance motors with different control strategies. The method adopts the technical scheme that detection pulses are injected into a demagnetization and idle interval of a non-excitation phase of the switched reluctance motor, and the alignment position of a rotor is estimated by comparing the rise time of pulse response current. According to the method, detection pulses are injected into each phase of the switched reluctance motor, for the four-phase switched reluctance motor, four-phase alignment positions are detected in one rotor period, and the alignment position of any one phase can be used for estimating the rotor position at other moments. Therefore, when the switched reluctance motor has a phase-lack fault, the accuracy of rotor position estimation can be ensured as long as one phase of normal operation exists, and the fault-tolerant operation of the switched reluctance motor is realized. The method reduces the cost of the system, improves the reliability of the system and has good engineering application value.)

1. A fault-tolerant switched reluctance motor position sensorless control method is characterized by comprising the following steps: detecting pulses are injected into a demagnetization and idle interval of a non-excitation phase of the switched reluctance motor, the detecting pulses can generate detecting current, and the rising time of the detecting current is in direct proportion to the value of phase inductance; the value of the phase inductance is gradually increased before the aligned position of the rotor, and thus the rise time of the detection current is also gradually increased; after the aligned position of the rotor, the value of the phase inductance starts to decrease, and therefore the rise time of the detection current also gradually decreases; by comparing two adjacent rise times of the detected current, the aligned position of the rotor can be determined when the rise time of the detected current starts to decrease.

2. The fault tolerant switched reluctance motor position sensorless control method according to claim 1, wherein: when the switched reluctance motor works in a normal state, detection pulses are injected into each phase, and for a four-phase 8/6 motor, four-phase alignment positions are detected in one rotor period; thus dividing a rotor cycle into 4 sections, RI, RII, RIII and RIV, RI representing the section between the a alignment position and the B alignment position, RII representing the section between the B alignment position and the C alignment position, RIII representing the section between the C alignment position and the D alignment position, and RIV representing the section between the D alignment position and the a alignment position;

in a normal state, the rotor positions in RI, RII, RIII and RIV intervals are calculated by depending on the initial rotor position in the current interval and the average angular speed in the previous interval;

when the switched reluctance motor has single-phase fault, the phase stops injecting detection pulse, and other three phases continue to inject detection pulse to estimate the position of the rotor;

when the switched reluctance motor has two-phase faults, the two phases stop injecting detection pulses, and the other two phases continue injecting the detection pulses to estimate the position of the rotor;

when the three-phase fault occurs to the switched reluctance motor, the three phases stop injecting the detection pulse, and finally, a detection pulse is continuously injected in succession to estimate the position of the rotor;

according to the fault tolerance method, the position-sensorless control method provided by the invention can effectively realize the fault tolerance operation of the switched reluctance motor.

Technical Field

The invention relates to a fault-tolerant switched reluctance motor position sensorless control method, which is particularly suitable for fault-tolerant operation of a switched reluctance motor under a phase-loss fault.

Background

A switched reluctance motor is a self-synchronous motor that requires rotor position information to ensure continuous operation of the motor. The position sensor is an important device for providing rotor position information, and is easy to fail due to interference of severe environments such as moisture, dust and the like, so that the reliability of the switched reluctance motor system is reduced. Therefore, a switched reluctance motor position sensorless control method has been widely studied. The position-sensorless control method of the switched reluctance motor proposed by scholars at home and abroad mainly comprises the following steps: current waveform method, flux linkage method, pulse injection method, inductance gradient method, inductance model method. They reduce the cost of the system and increase the reliability of the system. However, the above various position sensorless control methods are lack of research on fault tolerance methods. Because the switched reluctance motor system is easy to have phase failure when working in a severe environment, the position-sensorless control method provided by the invention cannot meet the fault-tolerant operation of the switched reluctance motor system under the fault condition. Therefore, it is necessary to provide a fault-tolerant switched reluctance motor sensorless control method.

Disclosure of Invention

The invention aims to provide a fault-tolerant control method for a switched reluctance motor sensorless control method, aiming at the problems of the existing control method for the switched reluctance motor sensorless control method.

The position sensorless control method provided by the invention comprises the following steps:

under inductive unsaturated conditions, the voltage equation for the phase winding can be represented by:

in the formula u_m,i_m,L_m,R_mω and m represent phase voltage, phase current, phase inductance, winding impedance, angular velocity of the motor rotor, and phase number of the motor, respectively;

assuming that there are n on periods in total in the detection current injection interval, then at t_n-1And t_nThe phase voltage equation at a time may be expressed as:

since the sense pulse is generated by hysteresis current control and the current loop width is very small, at t_n-1And t_nThe transient current at the moment is approximately equal to the reference current value I of the detection current_refI.e. i_m(t_n-1)≈i_m(t_n)≈I_refThus can obtain

i_m(t_n-1)R_m≈i_m(t_n)R_m(4)

Since the motor operates at low speed, the magnitude of the detected current is small and the inductance gradients at two adjacent moments are approximately equal, the third parts of equations (2) and (3) can be considered equal, i.e.

Considering that the voltage drop of the power switch tube is negligible compared with the direct current bus voltage, the phase voltage at two ends of the winding can be considered to be equal to the bus voltage, so that the voltage has

u_m(t_n-1)＝u_m(t_n)＝U_DC(6)

In the formula of U_DCRepresents the dc bus voltage;

by combining formulae (4), (5) and (6), the subtraction of formula (3) from formula (2) can be used

According to equation (7), the following relationship exists between the rate of change of phase current and the phase inductance:

if L is_m(t_n-1)＜L_m(t_n) Then, then

If L is_m(t_n-1)＞L_m(t_n) Then, then

Suppose that the corresponding current rise times in the last two consecutive on periods are each Δ T_n-1And Δ T_nAnd the loop width of the detected current is Δ i, the amount of change of the current in one on period is 2 Δ i, and thus equation (7) can be expressed as

By dividing both sides of the above formula by 2 Δ i, the formula (10) can be simplified to

According to the above equation, the phase inductance and the rise time of the detection current have the following relationship:

if L is_m(t_n-1)＜L_m(t_n) Then Δ T_n-1＜ΔT_n(12)

If L is_m(t_n-1)＞L_m(t_n) Then Δ T_n-1＞ΔT_n(13)

It can be seen that the rise time of the detection current is proportional to the value of the phase inductance, and the value of the phase inductance is gradually increased before the rotor is aligned, and therefore the rise time of the detection current is also gradually increased, and Δ T is present₁<ΔT₂<…<ΔT_n-1(ii) a After the aligned position of the rotor, the value of the phase inductance starts to decrease, and therefore the rise time of the detection current also gradually decreases, so there is Δ T_n-1>ΔT_n(ii) a In summary, the alignment position of the rotor can be determined by comparing two adjacent rise times of the detection current.

When the aligned position of the rotor of any phase is detected, the timer 2 starts counting until the count value is stored in the register after the aligned position of the rotor of the next phase is detected, and at the same time, the timer is immediately reset and starts to count again. Thus, the time interval Δ T between two successive alignment positions_{2_all}Then is

ΔT_{2_all}＝N_{2_all}×T2 (14)

In the formula N_{2_all}Represents the total count value of timer 1 and T2 represents the count period of timer 2. Thus, the average angular velocity ω between two successive alignment positions can be calculated by:

where Δ θ represents the angular difference between the two aligned positions. The rotor position at other times can then be calculated:

θ＝θ_o+ωN₂×T2 (16)

in the formula [ theta ]_oAnd N₂Indicating the initial rotor position and the real time count of the timer 2, respectively.

When the four-phase 8/6 switched reluctance motor operates in a normal state, a detection pulse is injected into each phase, and the four-phase alignment position is detected in one rotor cycle. One rotor cycle is thus divided into 4 sections, RI, RII, RIII and RIV, respectively, RI representing the section between the a alignment position and the B alignment position, RII representing the section between the B alignment position and the C alignment position, RIII representing the section between the C alignment position and the D alignment position, and RIV representing the section between the D alignment position and the a alignment position.

In a normal state, the rotor positions in RI, RII, RIII and RIV intervals are calculated according to the formula (16), and are mainly calculated according to the initial rotor position in the current interval and the average angular speed in the previous interval;

when the switched reluctance motor has single-phase fault, the phase stops injecting detection pulse, and other three phases continue to inject detection pulse to estimate the position of the rotor;

when the switched reluctance motor has two-phase faults, the two phases stop injecting detection pulses, and the other two phases continue injecting the detection pulses to estimate the position of the rotor;

when the three-phase fault occurs to the switched reluctance motor, the three phases stop injecting the detection pulse, and finally, a detection pulse is continuously injected in succession to estimate the position of the rotor;

in conclusion, the position-sensorless control method provided by the invention can effectively realize fault-tolerant operation of the switched reluctance motor.

Has the advantages that: the invention provides a fault-tolerant position-sensorless control method which does not need extra hardware, complex calculation and priori knowledge of the electromagnetic characteristics of a switched reluctance motor. In addition, it does not produce negative torque and has good fault tolerance.

Drawings

FIG. 1 is a functional schematic of the position sensorless control method of the present invention.

FIG. 2 is a schematic diagram of rotor position estimation during normal operation of the position sensorless control method of the present invention.

FIG. 3 is a schematic diagram of rotor position estimation during a single phase fault condition for the position sensor-less control method of the present invention.

FIG. 4 is a schematic diagram of rotor position estimation during a two-phase fault condition for the position sensorless control method of the present invention.

FIG. 5 is a schematic diagram of rotor position estimation during a three-phase fault condition for the position sensor-less control method of the present invention.

Fig. 6 is an experimental diagram of the position sensorless control method of the present invention operating in a normal state.

Fig. 7 is an experimental diagram of the position sensorless control method of the present invention operating in a single-phase fault condition.

FIG. 8 is an experimental plot of the position sensorless control method of the present invention operating in a two-phase fault condition.

Fig. 9 is an experimental diagram of the position sensorless control method of the present invention operating in a three-phase fault condition.

Detailed Description

The position sensorless control method proposed by the present invention is further described below with reference to the accompanying drawings:

when the position sensorless control method provided by the invention is applied to a current chopping control strategy, the working principle of the position sensorless control strategy provided by the invention is shown in fig. 1.

Under inductive unsaturated conditions, the voltage equation for the phase winding can be represented by:

in the formula u_m,i_m,L_m,R_mAnd ω and m represent phase voltage, phase current, phase inductance, winding impedance, angular velocity of the motor rotor, and number of phases of the motor, respectively.

Assuming that there are n on periods in total in the detection current injection interval, then at t_n-1And t_nThe phase voltage equation at a time may be expressed as:

since the sense pulse is generated by hysteresis current control and the current loop width is very small, at t_n-1And t_nThe transient current at the moment is approximately equal to the reference current value I of the detection current_refI.e. i_m(t_n-1)≈i_m(t_n)≈I_refThus can obtain

i_m(t_n-1)R_m≈i_m(t_n)R_m(4)

Since the motor operates at low speed, the magnitude of the detected current is small and the inductance gradients at two adjacent moments are approximately equal, the third parts of equations (2) and (3) can be considered equal, i.e.

Considering that the voltage drop of the power switch tube is negligible compared with the direct current bus voltage, the phase voltage at two ends of the winding can be considered to be equal to the bus voltage, so that the voltage has

u_m(t_n-1)＝u_m(t_n)＝U_DC(6)

In the formula of U_DCRepresenting the dc bus voltage.

By combining formulae (4), (5) and (6), the subtraction of formula (3) from formula (2) can be used

According to equation (7), the following relationship exists between the rate of change of phase current and the phase inductance:

if L is_m(t_n-1)＜L_m(t_n) Then, then

If L is_m(t_n-1)＞L_m(t_n) Then, then

Suppose that the corresponding current rise times in the last two consecutive on periods are each Δ T_n-1And Δ T_nAnd the loop width of the detected current is Δ i, the amount of change of the current in one on period is 2 Δ i, and thus equation (7) can be expressed as

By dividing both sides of the above formula by 2 Δ i, the formula (10) can be simplified to

According to the above equation, the phase inductance and the rise time of the detection current have the following relationship:

if L is_m(t_n-1)＜L_m(t_n) Then Δ T_n-1＜ΔT_n(12)

If L is_m(t_n-1)＞L_m(t_n) Then Δ T_n-1＞ΔT_n(13)

It can be seen that the rise time of the detection current is proportional to the value of the phase inductance, and the value of the phase inductance is gradually increased before the rotor is aligned, and therefore the rise time of the detection current is also gradually increased, and Δ T is present₁<ΔT₂<…<ΔT_n-1. After the aligned position of the rotor, the value of the phase inductance starts to decrease, thus detecting the electricityThe rise time of the flow is also gradually reduced, so there is a Δ T_n-1>ΔT_n. In summary, the alignment position of the rotor can be determined by comparing two adjacent rise times of the detection current.

The rotor alignment position can be obtained by comparing the rise times of two adjacent currents of the detection pulses, and the rise time of the response current generated by each detection pulse can be measured and counted by the timer 1, and the counting period of the timer 1 is 5 microseconds, which is denoted as t1. Whereas the calculation of the average rotation speed requires a time interval between two successive alignment positions, the timer 2 can be used to measure this time interval, and the counting period of the timer 2 is 10 microseconds, denoted T2. In the pulse injection interval, when the follow current is reduced to the lower limit value of the detection current, the main switch tube of the switched reluctance motor starts to be switched on, the detection current starts to rise, and the timer 1 starts to count, when the detection current rises to the upper limit value, the main switch tube of the switched reluctance motor is immediately switched off, the detection current starts to fall, and the timer 1 stops counting, stores the counting value at the moment into a register, and waits for being compared with the counting value of the next switching-on period. When the detection current drops to the lower limit value again, the main switch tube of the switched reluctance motor is turned on again, the detection current starts to rise, meanwhile, the timer 1 is immediately reset to start counting again, and therefore the second on period is started. The current rise time Δ T per on period can be calculated by:

ΔT＝N_{1_all}×T1 (14)

in the formula N_{1_all}Indicating the total count value of timer 1 during one on period.

Comparing the count value of the current period with the count value of the previous period, if the count value is larger than the count value of the previous period, continuing to inject the detection pulse, and if the count value is smaller than the count value of the previous period, stopping injecting the detection pulse, and indicating that the moment is the alignment position of the rotor, the timer 2 needs to immediately start countingWhen the alignment position of a rotor is detected, the timer 2 stores the count value in the register, and immediately resets and starts to count again. Thus, the time interval Δ T between two successive alignment positions_{2_all}Then is

ΔT_{2_all}＝N_{2_all}×T2 (15)

In the formula N_{2_all}Representing the total count value of the timer 2. Thus, the average angular velocity ω between two successive alignment positions can be calculated by:

where Δ θ represents the angular difference between the two aligned positions. The rotor position at other times can then be calculated:

θ＝θ_o+ωN₂×T2 (17)

in the formula [ theta ]_oAnd N₂Indicating the initial rotor position and the real time count of the timer 2, respectively.

When the four-phase 8/6 switched reluctance motor works in a normal state, each phase of the switched reluctance motor is injected with a detection pulse to detect each alignment position, so that four alignment positions can be detected in one rotor period, and the fault tolerance of the position-sensorless control method is guaranteed. FIG. 2 shows the rotor position estimation principle when the switched reluctance motor is operating in a normal state, where i_a、i_b、i_cAnd i_dAre respectively four-phase current, L_a、L_b、L_cAnd L_dRespectively four-phase inductors. Further, each rotor cycle is divided into 4 sections, RI, RII, RIII, and RIV, where RI represents the section between the a alignment position and the B alignment position, RII represents the section between the B alignment position and the C alignment position, RIII represents the section between the C alignment position and the D alignment position, and RIV represents the section between the D alignment position and the a alignment position. These intervals are mainly for convenience of explaining the rotor position of the switched reluctance motor in normal and fault statesAnd (4) a calculation method. In a normal state, taking the RII interval as an example, the calculation of the rotor position in this interval can refer to equation (17), wherein the initial rotor position is the aligned position of the B phase, i.e. 45 °, and the average angular velocity needs to be the average angular velocity in the RI interval, which can be calculated by equation (16), wherein the rotor position angular difference is 15 degrees, and the total time is the total time recorded by the timer 2 in the RI interval, so that the rotor position at any time in the RII interval can be calculated. Similarly, rotor positions in the RI, RIII and RIV intervals are calculated in the same manner, and the initial rotor positions in the RI, RIII and RIV intervals are 30, 60 and 75, respectively, which represent the alignment of phase A, C, D. It should also be noted that the average angular velocity in each interval is used to calculate the rotor position for the next interval.

When the switched reluctance motor has single-phase fault, the online fault diagnosis method can immediately determine the fault phase, the phase stops injecting detection pulses, and other three phases continue to inject the detection pulses to estimate the position of the rotor. Taking the D phase as an example, when the open circuit fault occurs in the D phase, as shown in fig. 3, the ABC three phases are kept in normal operation, and the rotor positions in the RII and RIII sections are calculated in the same manner as the rotor positions in the normal state, but the rotor position in the RIV section needs to be calculated depending on the initial rotor position in the RIII section and the average angular velocity in the RII section, and the rotor position calculation in the RI section uses the aligned position of the a phase as the initial rotor position, but the used average angular velocity is the average angular velocity in both the RIII and RIV sections.

When a two-phase fault occurs in the switched reluctance motor, taking two phases C and D as an example, fig. 4 is a rotor position calculation schematic diagram when an open circuit fault occurs in C, D two phases, at this time, A, B two phases can estimate a rotor position when the motor is working normally, the rotor position calculation in the RI interval can depend on an initial rotor position in the RI interval and average angular velocities in the RII, RIII, and RIV intervals, and the rotor position in the RII, RIII, and RIV intervals needs to depend on an initial rotor position in the RII interval and an average rotation speed in the RI interval.

When a three-phase fault occurs in the switched reluctance motor, using B, C, D three phases as an example, fig. 5 shows a rotor position calculation schematic diagram when B, C, D three phases have an open circuit fault, at this time, only the a phase can work normally, the estimation of the rotor position can only depend on the aligned position detected by the a phase, the rotor position calculation of the whole rotor period uses the aligned position of the a phase as an initial rotor position, and the used average angular velocity is the average angular velocity of one rotor period.

FIG. 6 is a waveform diagram illustrating an experiment when the position sensorless control method according to the present invention works in a normal state, where θ is_e、θ_rAnd theta_erThe estimated rotor position, the actual rotor position and the rotor position estimation error are respectively the winding current of four phases of the switched reluctance motor. As can be seen from the figure, θ_eAnd theta_rMaintain coincidence, θ_erFluctuating around zero, thus accounting for the accuracy of the rotor position estimate for the position sensorless control method.

Fig. 7, fig. 8 and fig. 9 are experimental waveform diagrams of the position sensorless control method according to the present invention when the method works in a single-phase fault state, a two-phase fault state and a three-phase fault state, respectively, where flag represents a fault occurrence flag. As can be seen from the figure, θ_eAnd theta_rThe method is consistent all the time, so that the accuracy of rotor position estimation can still be ensured under the fault condition by the control method without the position sensor, and the method has good fault tolerance capability.

In conclusion, the fault-tolerant position-sensorless control method provided by the invention ensures the accuracy of rotor position estimation and realizes fault-tolerant operation of the switched reluctance motor.