Sampling Method for Switch Open-Circuit Fault Current of PMSM Driver

Abstract

The present invention relates to a method for sampling the open - circuit fault current of a PMSM driver, which realizes the fault current conducted through the free - wheeling diode after the driver detects an open - circuit of the power switch. First, the equivalent circuit after the open - circuit fault of the power switch is analyzed to obtain the condition for the free - wheeling diode to conduct after the fault; then, the voltage vectors that can achieve the above - mentioned conduction condition are analyzed, and it is pointed out that at the end of the action time of the above - mentioned voltage vector, the fault current reaches the peak value in the current switching period; based on the end of the vector action time, the sampling and holding time required by the analog - to - digital conversion circuit is further considered to determine the current sampling timing; finally, taking the count value of the PWM counter in the controller as a reference, the current sampling is triggered when the count value reaches the sampling timing. This method expands the current sampling ability of the driver, enabling it to detect the fault current unique to the open - circuit fault of the switch, thus laying a foundation for distinguishing this type of fault from other open - circuit faults.

Description

Technical Field

[0001] The present invention belongs to the field of fault diagnosis of PMSM drive systems and relates to a sampling method for the open - circuit fault current of a PMSM driver. Background Art

[0002] In various industrial fields, permanent magnet synchronous motors and their drive systems have been widely used due to their advantages such as fast response, high power density, and high efficiency. With the increasingly critical application scenarios and the improvement of the rated power and voltage levels, their reliability and maintainability have become important indicators in product design, and improving reliability and maintainability has become an important research direction in this field.

[0003] Failures occurring in different parts can cause similar fault phenomena. The most typical one is the open - circuit fault, which is also the most typical fault phenomenon in motor drivers. The open - circuit phenomenon can be caused by connection failures, such as the failure of the connector connection and the damage of the phase winding. Such faults are called open - phase faults. In addition, the failure of the power switch in the driver can also cause an open - circuit phenomenon. For example, the open - circuit phenomenon caused by the failure of the driving chip and the failure of the power switch closing line is called a switch open - circuit fault. After an open - phase fault occurs, the fault - phase current loses the conduction path, while after a switch open - circuit fault, due to the existence of the parallel free - wheeling diode of the power switch, the fault - phase current is not zero.

[0004] However, the switch open - circuit fault and the open - phase fault are still classified as the same type of fault at present, and they cannot be effectively distinguished, which results in the inability to locate whether the open - circuit occurs on the inverter side or the motor side. The reason for classifying the switch open - circuit fault and the open - phase fault as the same type is that the motor driver cannot detect the fault current after the switch open - circuit fault, resulting in the same fault phenomenon for the switch open - circuit fault and the open - phase fault, that is, the fault - phase current is zero. Summary of the Invention

[0005] Technical Problems to be Solved

[0006] In order to avoid the deficiencies of the prior art, the present invention proposes a sampling method for the open - circuit fault current of a PMSM driver, enabling the motor driver to detect the non - zero fault current after the switch open - circuit fault. Based on its fault mechanism, the present invention proposes a sampling method, enabling the motor driver to self - sense the switch open - circuit fault current and providing a diagnostic basis for distinguishing the switch open - circuit fault from the open - phase fault.

[0007] Technical Solutions

[0008] A sampling method for the open - circuit fault current of a PMSM driver is characterized by the following steps:

[0009] Step 1: Based on the equivalent circuit after the power - switch open - circuit fault, the conditions for the conduction of the subsequent free - wheeling diode are obtained;

[0010] Fault state 1: The post-fault conduction condition of the freewheeling diode in the upper half-bridge arm of the faulty phase is as follows:

[0011] u XP = u XN - u PN = V bemfX - u PN > u D ;

[0012] Where: u XP is the voltage across the freewheeling diode in the upper half-bridge arm, V bemfX is the back electromotive force of the faulty phase equal to the voltage u of the faulty phase after the switch is open XN , u PN is the voltage difference from the positive pole of the DC bus to the neutral point, u D is the forward conduction voltage of the diode;

[0013] Fault state 2: The post-fault conduction condition of the freewheeling diode in the lower half-bridge arm of the faulty phase is as follows:

[0014] u OX = u ON - u XN = u ON - V bemfX > u D

[0015] Where: u OX is the voltage across the freewheeling diode in the lower half-bridge arm, u ON is the voltage difference from the negative pole of the DC bus to the neutral point;

[0016] Step 2: Obtain the voltage vector of the conduction condition, i.e., the active vector, and at the end of the voltage vector application time, the fault current reaches the peak value in the current switching cycle;

[0017] When fault state 1 occurs, connect the phase winding without fault to the positive pole of the bus. When V bemfX > 0, the voltage vector that enables the conduction of the freewheeling diode in the upper bridge arm of the faulty phase, i.e., the active vector

[0018] When fault state 2 occurs, connect the phase winding without fault to the negative pole of the bus. When V bemfX < 0, the voltage vector that enables the conduction of the diode in the upper bridge arm of the faulty phase, i.e., the active vector

[0019] Step 3: Subtract the sampling and holding time CNT of the analog-to-digital conversion circuit from the end of the active vector application time delay , to obtain the count value for the trigger sampling timing, which are the count values CNT when the freewheeling diode in the upper bridge arm conductsv7samp , and the count value CNT when the freewheeling diode of the lower arm conducts v0samp ;

[0020] The end of the action time is: find the maximum duty cycle DR in each phase of the current control cycle in the PWM calculation max and the minimum duty cycle DR min , then DR max and the total count value CNT of the switching period ts The product of is the end of the action time;

[0021] The end of the action time is: DR min *CNT ts is the end of the action time;

[0022] Then: The timing of sampling the fault current is CNT v7samp = DR min *CNT ts + CNT delay ;

[0023] The timing of sampling the current after the fault is CNT v0samp = DR max *CNT ts - CNT delay ;

[0024] Step 4: When V bemfX > 0, that is, in fault state 1, when the PWM time base counter is the same as CNT v7samp trigger sampling;

[0025] When V bemfX < 0, that is, in fault state 2, when the PWM time base counter is the same as CNT v0samp trigger sampling.

[0026] In the step 4 mentioned above, triggering the current sampling by comparing the PWM time base counter with CNT v7samp and CNT v0samp is applicable to all digital chips used for motor control.

[0027] Beneficial effects

[0028] A sampling method for the open - circuit fault current of a PMSM driver proposed by the present invention realizes detecting the fault current conducted through the free - wheeling diode after the power switch is detected to be open - circuited by the driver. First, the equivalent circuit after the open - circuit fault of the power switch is analyzed, and the condition for the free - wheeling diode to conduct after the fault is obtained; then, the voltage vectors that can achieve the above - mentioned conduction condition are analyzed, and it is pointed out that at the end of the action time of the above - mentioned voltage vector, the fault current reaches the peak value in the current switching period; based on the end of the vector action time, the sampling - hold time required by the analog - to - digital conversion circuit is further considered, and the current sampling timing is determined; finally, taking the count value of the PWM counter in the controller as a reference, when the count value reaches the sampling timing, the current sampling is triggered. This method expands the current sampling ability of the driver, enabling it to detect the fault current unique to the open - circuit fault of the switch, thus laying a foundation for distinguishing this type of fault from other open - circuit faults.

[0029] The present invention enables the motor drive to detect the fault current after the switch is open - circuited for the first time, providing a diagnostic basis for distinguishing the open - circuit fault of the switch from the open - phase fault;

[0030] The present invention is not only applicable to the open - circuit fault of the switch, but also can detect the healthy current and the open - phase fault current. Therefore, the proposed sampling method can replace the ordinary sampling method. Brief Description of the Drawings

[0031] Figure 1 Schematic diagram of the sampling method for the open - circuit fault current of the PMS driver in the present invention;

[0032] Figure 2 Equivalent circuit diagram of the fault phase after the fault proposed in the present invention;

[0033] Figure 3 Waveform diagram of the experimental results of the traditional sampling method;

[0034] Figure 4 Waveform diagram of the experimental results of the present invention. Detailed Embodiment

[0035] The present invention will be further described in combination with the embodiments and the drawings:

[0036] The described embodiments are some embodiments of the present invention, rather than all embodiments. As Figure 1 Shown in the schematic diagram of the sampling method proposed in the present invention, taking the open - circuit fault of the switch in the A - phase bridge arm of the motor driver as an example, the specific implementation is as follows:

[0037] Step (1) According to the equivalent circuit of the fault phase after the open - circuit fault of the switch, as Figure 2 shown, the condition for the free - wheeling diode to conduct after the fault is obtained:

[0038] The post-fault conduction condition of the freewheeling diode in the upper half-bridge arm of the faulty phase is: the voltage u across the diode AP is greater than the forward conduction voltage u D of the diode. u AP depends on the phase voltage of the faulty phase, i.e., the voltage difference u AN between the faulty phase terminal and the motor neutral point, as well as the voltage difference u PN between the positive pole of the DC bus and the neutral point. Also, since the faulty phase voltage u AN equals the back electromotive force V bemfA of the faulty phase after the switch opens, therefore, the conduction condition of the freewheeling diode in the upper half-bridge arm can be expressed as: u AP = u AN - u PN = V benfA - u PN > u D ;

[0039] The post-fault conduction condition of the freewheeling diode in the lower half-bridge arm of the faulty phase is: the voltage u across the diode OA is greater than the forward conduction voltage u D of the diode. u OA depends on the phase voltage of the faulty phase, i.e., the voltage difference u AN between the faulty phase terminal and the motor neutral point, as well as the voltage difference u ON between the negative pole of the DC bus and the neutral point. Also, since the faulty phase voltage u AN equals the back electromotive force V bemfA of the faulty phase after the switch opens, therefore, the conduction condition of the freewheeling diode in the upper half-bridge arm can be expressed as: u OA = u ON - u AN = u ON - V bemfA > u D ;

[0040] Step (2): Determine the voltage vector that can achieve the above conduction condition, which is called the active vector:

[0041] Step (2.1): To achieve the conduction of the diode in the upper bridge arm of the faulty phase, V bemfA needs to be greater than 0. Also, considering that u PN can ensure u AP > u D always holds only when it is equal to 0. Therefore, it is necessary to connect the phase winding that has not failed to the positive pole of the bus. When V bemfA > 0, only can achieve the conduction of the diode in the upper bridge arm of the faulty phase.

[0042] Step (2.2): To achieve the conduction of the diode in the lower bridge arm of the faulty phase, VbemfA It needs to be less than 0, and considering u ON Only when it is equal to 0 can u be guaranteed OA >u D always holds. Therefore, it is necessary to connect the non-faulty phase winding to the negative bus. At V bemfA <0, only can achieve the conduction of the upper-bridge-arm diode of the faulty phase.

[0043] Step (3) Subtract the sampling and holding time of the analog-to-digital conversion circuit from the end of the active vector action time to obtain the count value of the trigger sampling timing, which are the count values CNT v7samp when the upper-bridge-arm diode conducts, and the count value CNT v0samp when the lower-bridge-arm diode conducts:

[0044] Step (3.1) Find the maximum duty cycle DR max and the minimum duty cycle DR min in each phase during the current control period in the PWM calculation. Then the product of DR max and the total count value CNT ts of the switching period is the end of the action time, and DR min *CNT ts is the end of the action time;

[0045] Step (3.2) Since the PWM uses up and down counting to achieve symmetric modulation, and the sampling and holding time of the digital-to-analog conversion circuit is set as CNT delay , so at the timing of sampling the fault current is CNT v7samp =DR min *CNT ts +CNT delay , and at the timing of sampling the current after the fault is CNT v0samp =DR max *CNT ts -CNT delay ;

[0046] Step (4) When V bemfA >0, trigger sampling when the PWM time-base counter is the same as CNT v7samp ; when V bemfX <0, trigger sampling when the PWM time-base counter is the same as CNT v0samp .

[0047] The overall effect of the proposed fault-tolerant method is as shown in Figure 3 , Figure 4As shown. When using the ordinary current sampling method, the driver cannot correctly sample the fault current after the switch opens, such as Figure 3 shown. After adopting the proposed current sampling method, the driver can completely and correctly detect the envelope and amplitude of the post-fault current, such as Figure 4 shown;

[0048] Experimental conditions: DC bus voltage 200V, number of pole pairs of the motor is 4 pole pairs, L d = L q = 2.4mH, winding resistance 0.306Ω, permanent magnet flux linkage 0.281Wb, operating speed 500rpm, inverter switching frequency 10kHz, control chip is TMS320F28335, IGBT module used in the inverter is FUJI 7MBR25VA120.

Claims

1. A sampling method for the open - circuit fault current of a PMSM driver, characterized in that The steps are as follows: Step 1: Based on the equivalent circuit after the open-circuit fault of the power switch, the conditions for the conduction of the subsequent flow diode after the fault are obtained; Fault state 1: The conditions for the conduction of the freewheeling diode of the upper half-bridge arm of the faulty phase after the fault are: u XP = u XN - u PN = V bemfX - u PN > u D ; Where: u XP is the voltage across the freewheeling diode of the upper half-bridge arm, V bemfX is the counter electromotive force of the faulty phase equal to the voltage of the faulty phase u XN after the switch opens, u PN is the voltage difference from the positive pole of the DC bus to the neutral point, u D is the forward conduction voltage of the diode; Fault state 2: The conditions for the conduction of the freewheeling diode of the lower half-bridge arm of the faulty phase after the fault are: u OX = u ON - u XN = u ON - V bemfX > u D Where: u OX is the voltage across the freewheeling diode of the lower half-bridge arm, and u ON is the voltage difference from the negative pole of the DC bus to the neutral point; Step 2: Obtain the voltage vector, i.e., the activation vector, of the conduction condition, and at the end of the action time of the voltage vector, the fault current reaches the peak value in the current switching cycle; When a fault state 1 occurs, connect the phase winding without a fault to the positive pole of the bus. When V bemfX > 0, the voltage vector that enables the freewheeling diode of the upper bridge arm of the faulty phase to conduct, i.e., the active vector, is activated When the fault state 2 occurs, connect the phase winding without fault to the negative pole of the bus, and at V bemfX When <0, the voltage vector that enables the diode of the upper bridge arm of the faulty phase to conduct, i.e., the active vector, is activated Step 3: Subtract the sampling hold time CNT of the analog-to-digital conversion circuit from the end of the activation vector action time delay , to obtain the count values for the trigger sampling timing, which are the count value CNT when the freewheeling diode of the upper bridge arm conducts v7samp , and the count value CNT when the freewheeling diode of the lower bridge arm conducts v0samp ; The end of the action time is: find the maximum duty cycle DR in each phase of the current control cycle in the PWM calculation max and the minimum duty cycle DR min , then DR max multiplied by the total value CNT of the switching period ts is the end of the action time; The end of the action time is: DR min *CNT ts is the end of the action time; Then: The timing of sampling the fault current is CNT v7samp = DR min * CNT ts + CNT delay ; The timing for sampling the post-fault current is CNT v0samp = DR max * CNT ts - CNT delay ; Step 4: When V bemfX > 0, that is, in fault state 1, sampling is triggered when the PWM time base counter is the same as CNT v7samp ; At V bemfX <When it is less than 0, that is, in fault state 2, when the PWM time base counter and CNT v0samp are the same, sampling is triggered.

2. The sampling method for the open - circuit fault current of the PMSM driver according to claim 1, characterized in that: In the above-mentioned step 4, it is mentioned that the current sampling is triggered by comparing the PWM time-base counter with CNT v7samp and CNT v0samp and is applicable to all digital chips used for motor control.

Images