A motor open-phase detection method and system for electric vehicle drive fault detection

Abstract

The application discloses a motor open-phase detection method and system for electric vehicle drive fault detection, relates to the technical field of motor open-phase detection, and is characterized in that a low-computing-load, high-robustness and two-sampling-one algorithm architecture is adopted, RMS is combined with the energy average in a period, and cannot become zero due to instantaneous zero-crossing, because false positives cannot be caused due to zero-crossing; moreover, the RMS effective value is used instead of the instantaneous value, the influence of zero-crossing points can be eliminated in the physical nature, the complex symmetrical component decomposition is abandoned, and the median reference comparison method is used, so that the sensitivity to unbalance is retained, false judgments caused by current error calculation are avoided, and the calculation amount is greatly reduced. Through a segmented mixed smoothing algorithm, the contradiction between large calculation amount and slow starting of the traditional RMS is solved. A speed self-adaptive threshold and a hysteresis mechanism are introduced, false positives in a low-speed and small-current working condition are solved, and reliable detection in a full-speed range is realized.

Description

Technical Field

[0001] This invention relates to the field of motor phase loss detection technology, and in particular to a motor phase loss detection method and system for detecting faults in electric vehicle drive systems. Background Technology

[0002] Phase loss detection is a core protection function in motor drive systems. If a phase loss is not detected correctly, the motor will operate under an asymmetrical magnetic field, leading to severe torque pulsation, excessive temperature rise, and ultimately burning out the IGBT power module or damaging the motor windings, resulting in significant economic losses. Traditional motor phase loss detection methods typically monitor whether the three-phase current is below a certain fixed threshold or close to zero. Because alternating current has sinusoidal characteristics, it passes through zero twice in each cycle. At the zero-crossing moment, the instantaneous current value of the normal phase is also close to zero, making it impossible to distinguish between zero-crossing and actual phase loss, resulting in false zero-crossing judgments.

[0003] To avoid false zero-crossing detection, existing technologies mainly use timing shielding and imbalance detection methods. Timing shielding detects the zero-crossing region of the current and shields the fault within that region. This method relies on accurate zero-crossing detection, and in variable frequency speed control, the frequency changes rapidly, making it difficult to quickly detect the zero-crossing region. Imbalance detection determines whether the motor is missing a phase by comparing the amplitude differences of the three-phase currents.

[0004] Traditional imbalance detection methods mainly fall into two categories:

[0005] One type calculates the absolute value of the three-phase current in real time and averages it. If the current of one phase is significantly lower than the average value, an alarm is triggered. However, this method is greatly affected by instantaneous value fluctuations, and the absolute value averaging cannot accurately reflect the thermal effect of the current. It is also sensitive to waveform distortion.

[0006] Another method is the symmetrical component method, which involves sampling the three-phase current and then decomposing it into positive-sequence, negative-sequence, and zero-sequence components using Clark / Park transform or complex number operations. Theoretically, the negative-sequence component increases significantly when a phase is missing. Calculating the positive-sequence, negative-sequence, and zero-sequence components can determine if a phase is missing. This method is computationally extremely demanding, requiring numerous trigonometric function operations and multiplications, consuming valuable MCU resources; furthermore, it necessitates precise sampling of the motor's three-phase current.

[0007] It is typically assumed that the three-phase currents are independently and accurately sampled. However, current sampling in actual motor use involves errors. This traditional unbalanced detection method is extremely sensitive to waveform distortion and errors, and current detection errors are easily amplified in calculations, leading to frequent false alarms due to motor phase loss. In particular, actual motor systems often only have two current sensors installed, and the third-phase current is calculated from the currents of the other two phases. The calculated third-phase current accumulates the sampling errors and offsets of the first two phases, increasing the amplification factor of the current detection error and resulting in a higher false alarm rate for motor phase loss. Summary of the Invention

[0008] The purpose of this invention is to address the problem that existing symmetrical component detection methods for motor phase loss require precise sampling and are extremely sensitive to waveform distortion and errors. In actual motor use, current sampling errors and current detection errors are easily amplified in calculations, leading to frequent false alarms due to motor phase loss. This invention provides a motor phase loss detection method and system for detecting faults in electric vehicle drive systems.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0010] A method for detecting motor phase loss in electric vehicle drive systems includes the following steps:

[0011] S1. Initialization: Define the segment length N, accumulator segment_sum, counter segment_count, output value rms_value, and total_segments count;

[0012] S2, Collect the two-phase instantaneous current value I of the motor. a I b and according to the I a I b Calculate and obtain I c ;

[0013] S3. Calculate the instantaneous current value I for each sampling period. a I b and I c The sum of squares is accumulated in segment_sum, and the value of the counter segment_count is incremented by one;

[0014] instantaneous current I a The sum of squares is segment_count[a] = N times Ia 2 sum;

[0015] instantaneous current I b The sum of squares is segment_count[b] = N times Ib 2 sum;

[0016] instantaneous current I c The sum of squares is segment_count[c] = N times Ic 2 sum;

[0017] S4. When the value of the counter segment_count reaches N, calculate the RMS value R of each segment according to formula one. seg The calculation formula is as follows:

[0018] ;

[0019] Where segment-sum is the instantaneous current value I a I b and I c The sum of squares, where N is the segment length and R is the sum of squares. seg For each segment's RMS value;

[0020] S5, Two-level Hybrid Smooth Update:

[0021] During the startup phase, if the total number of completed segments (total_segments) is less than the segment count threshold, the RMS output value (rms_valueR) is calculated according to formula two. out [k], the second calculation formula is:

[0022] ;

[0023] Among them, R out [k] represents the RMS output value rms_value, R seg For each segment, k represents the RMS value, and k is the number of segments.

[0024] In the steady-state phase, when the total number of segments (total_segments) is greater than or equal to the segment count threshold, the system switches to first-order low-pass filtering and calculates the RMS output value (rms_valueR) according to Equation 3. out [k], the calculation formula three is:

[0025] ;

[0026] Where 0.9 represents the historical output weight and 0.1 represents the current input weight;

[0027] S6. Based on the RMS output value rms_value obtained in S5, set the RMS base value I. ref Set the reference coefficient K according to the motor speed. th and counting step up ;

[0028] S7. Based on the RMS output value rms_valueR out [k], RMS baseline value I ref and benchmark coefficient K th To determine if the motor has a phase loss fault:

[0029] S71. If the RMS output value R out [k]<K th ×I ref And C i < Single-phase suspected fault counter threshold, then Ci =Ci+Step up The phase loss fault count increases; if R out [k]>=K th ×I ref C i= Ci-2, the phase loss fault count decreases; where C i It is a single-phase fault counter with a counting range of 0~100;

[0030] S72, If C i If the value is greater than or equal to the maximum single-phase count, then set the corresponding phase fault flag (Flag). i =1, motor set; if C i If the value is less than or equal to the minimum single-phase count, then clear the corresponding phase fault flag. i =0, motor reset; if the single-phase count is the minimum value. <C i If the single-phase count reaches its maximum value, then the Flag will be retained. i If it remains unchanged, the motor enters the hysteresis region.

[0031] The motor phase loss detection method for electric vehicle drive fault detection described in this invention improves the conventional RMS calculation method through S1~S5 into a more refined RMS algorithm that combines instantaneous current segmented square accumulation and two-stage hybrid smooth update with a startup-period arithmetic mean plus a steady-state first-order low-pass filter. It utilizes a segment counting mechanism to achieve automatic switching and smooth update, and in conjunction with the completion of the counting stage calculation in S4, corresponding to the motor startup stage, it ensures that the RMS output value rms_value converges quickly and without deviation from 0 to the true value, avoiding initial values ​​of the IIR filter. The slow rise caused by zero-crossing allows the RMS output value rms_value to replace the traditional instantaneous value, fundamentally eliminating the risk of misjudgment caused by the AC zero-crossing point. Furthermore, combined with the first-order low-pass filtering stage of S4, corresponding to the motor's steady-state stage, it filters out single-stage calculation noise while maintaining a rapid response to load changes, requiring only one multiplication and addition operation, resulting in extremely high efficiency. This avoids errors in calculating the instantaneous current of one phase from the instantaneous currents of the other two phases, or the impact of extreme conditions such as a complete phase disconnection. Furthermore, S6~S7 utilize the RMS reference value I... ref With effective settings and references, more accurate fault and normal fault identification can be performed, thereby reducing the false alarm rate of phase loss faults and improving the accuracy of motor phase loss judgment. At the same time, through the segmented hybrid RMS algorithm of S3~S5, the expensive square root operation frequency is reduced by N times, and the arithmetic mean during the startup period is used to solve the problem of slow convergence of traditional filters, thus achieving a balance between low power consumption and fast response.

[0032] Preferably, in the motor phase loss detection method for electric vehicle drive fault detection according to the present invention, step S6 specifically includes the following steps:

[0033] The median of the three-phase current RMS output values ​​rms_value is taken as the RMS reference value I. med According to the RMS reference value I med The RMS baseline value is calculated according to formula four; formula four is:

[0034] .

[0035] As a preferred embodiment of the present invention, by more accurately calculating and setting the reference value, the median of the three-phase RMS values, rather than the average value, is used as the health reference in the three-phase current imbalance detection. Combined with the calculation method of the maximum lower limit constraint, even if one phase is completely disconnected and the current is zero, the median can resist the influence of the single-phase disconnection to a certain extent and can still be determined by the health values ​​of the remaining two phases, and will not be lowered by the reference value, thereby further ensuring the reliability of the reference.

[0036] Preferably, in the motor phase loss detection method for electric vehicle drive fault detection described in this invention, the RMS reference value I... med Replace with the maximum value; or replace the RMS reference value Imed with the average value after removing the maximum and minimum values.

[0037] Preferably, the motor phase loss detection method for electric vehicle drive fault detection described in this invention sets a reference coefficient K based on the motor speed. th and counting step up The specific steps include:

[0038] If the absolute value of the motor speed is ≥2000 RPM, then K th =0.35, Step up =2;

[0039] If 500 RPM ≤ absolute value of motor speed < 2000 RPM, then K th =0.40, Step up =1;

[0040] If the absolute value of the motor speed is <500 RPM, then K th =0.40, Step up =1.

[0041] As a preferred embodiment of the present invention, by using three absolute values ​​of motor speed, further protecting the strategy of dynamically adjusting the fault judgment threshold ratio and fault accumulation step size according to the motor speed range, and using anti-bouncing logic with different trigger / reset thresholds in conjunction with the bidirectional hysteresis counter segment_count, the reliability of the reference coefficient setting is further improved.

[0042] Preferably, in the motor phase loss detection method for electric vehicle drive fault detection described in this invention, after step S5, a pre-lock check is performed to determine whether the lock-up conditions are met. If the lock-up conditions are met, the lock-up is triggered.

[0043] The interlocking conditions include: the electronic switch is not enabled, a delay is initiated within 150ms before power-on, a third trigger condition, and a fourth trigger condition.

[0044] The process of determining whether the third trigger condition has been met includes the following steps: calculating the motor's electrical frequency f based on the motor speed. e =RPM / 60; where RPM represents the motor speed; the unit is revolutions per minute; if |f e If the frequency is less than 10Hz, the third trigger condition is met.

[0045] The determination of whether the fourth trigger condition has been met includes the following steps: if the RMS output value rms_value of the three-phase current is less than the preset current threshold, then the fourth trigger condition has been met.

[0046] As a preferred embodiment of the present invention, by protecting the phase loss detection triggering mechanism which includes four preconditions such as start-up delay, minimum electrical frequency lockout, and minimum current amplitude lockout, the parameter deviations caused by actual motor operation are further eliminated, the phase loss parameters are identified more accurately, false phase loss alarms are avoided, and phase loss is identified more accurately.

[0047] Preferably, in the motor phase loss detection method for electric vehicle drive fault detection according to the present invention, step S71 further includes the step:

[0048] If the number of faulty phases fault_count=1, then output the single-phase fault code of the corresponding phase;

[0049] If the number of faulty phases fault_count ≥ 2, then enter multiple fault codes;

[0050] If the number of faulty phases fault_count=0, then no fault code is entered.

[0051] As a preferred embodiment of the present invention, by outputting the corresponding fault code according to the number of faulty phases, a more intelligent interactive display of motor phase loss fault is achieved, providing more accurate evaluation data for motor fault solutions, and facilitating the rapid detection and convenient maintenance of motor phase loss faults.

[0052] Preferably, in the motor phase loss detection method for electric vehicle drive fault detection described in this invention, the two-stage hybrid smooth update is replaced by the following steps:

[0053] During the startup phase, set the initial gain value. A large time-varying gain IIR filter; the formula for calculating the time-varying gain IIR filter is:

[0054] ;

[0055] Where alpha[n] represents linear decay or exponential decay. Smooth transition to The function; in the steady state, alpha[n] = 0.1.

[0056] To achieve the objectives of this invention, another technical solution is provided:

[0057] A motor phase loss detection system for detecting faults in an electric vehicle drive system, wherein the user implements the motor phase loss detection method described in this invention, comprising:

[0058] Instantaneous current acquisition module, instantaneous square accumulation calculation module, hybrid smoothing calculation module, parameter setting module, phase loss analysis calculation module, fault output module;

[0059] The instantaneous square accumulation calculation module is used to calculate the instantaneous current value I. a I b and I c The sum of squares, accumulated by segment_sum, and count;

[0060] The parameter setting module is used to initialize and set parameters and thresholds;

[0061] The hybrid smoothing calculation module is used for two-level hybrid smoothing update calculation and obtaining the initial RMS value;

[0062] The phase loss analysis and calculation module is used to determine whether a phase loss fault exists;

[0063] The fault output module is used to output phase loss fault and fault phase data.

[0064] The motor phase loss detection system of this invention achieves energy averaging within a cycle, preventing it from becoming zero due to instantaneous zero crossings, thus avoiding false alarms of phase loss faults caused by zero crossings. It uses the effective RMS value to replace the instantaneous value, eliminating the influence of zero crossings from a physical perspective. It can abandon the traditional complex symmetrical component decomposition and set a more reasonable reference value, which retains the sensitivity to imbalance while avoiding misjudgments caused by current calculation errors. It significantly reduces the amount of calculation, has high calculation efficiency, and outputs results in a block. Based on the reference coefficient set by the motor speed and the counting step size, it participates in the judgment, realizing reliable detection of the motor in the entire speed range.

[0065] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0066] 1. The detection method described above utilizes a segment counting mechanism to achieve automatic switching and smooth updates. Corresponding to the motor start-up phase, it enables the RMS output value rms_value to converge quickly and without deviation from 0 to the true value, avoiding the slow rise caused by the initial value of the IIR filter being 0. This allows the RMS output value rms_value to replace the traditional instantaneous value, fundamentally eliminating the risk of misjudgment caused by the zero crossing of the AC current.

[0067] 2. The aforementioned detection method, corresponding to the steady-state stage of the motor, filters out single-segment calculation noise while maintaining a rapid response to load changes, and the calculation only requires one multiplication and addition, resulting in extremely high efficiency; it avoids errors in the calculation of one phase's instantaneous current from the instantaneous currents of the other two phases, or avoids the influence of extreme conditions such as a complete disconnection of a phase, and further utilizes S6~S7 to pass the RMS reference value I. ref With effective settings and references, more accurate fault and normal judgments can be made, thereby reducing the false alarm rate of phase loss faults and improving the accuracy of motor phase loss judgment; at the same time, through the segmented hybrid RMS algorithm of S3~S5, the expensive square root operation frequency is reduced by N times, and the arithmetic mean during the startup period is used to solve the problem of slow convergence of traditional filters, thus achieving the unity of low power consumption and fast response.

[0068] 3. The motor phase loss detection system achieves energy averaging within the cycle, and will not become zero due to instantaneous zero crossing, thereby avoiding false alarms of phase loss fault caused by zero crossing. It realizes that the effective value of RMS replaces the instantaneous value, which can eliminate the influence of zero crossing from a physical essence.

[0069] 4. The motor phase loss detection system can abandon the traditional complex symmetrical component decomposition and set a more reasonable reference value. It retains the sensitivity to imbalance and avoids misjudgment caused by the error of current calculation. It greatly reduces the amount of calculation, has high calculation efficiency, and outputs results in a block. It uses the reference coefficient set by the motor speed and the counting step size to participate in the judgment, so as to realize reliable detection of the motor in the whole speed range. Attached Figure Description

[0070] Figure 1 This is a schematic flowchart of the motor phase loss detection method of the present invention;

[0071] Figure 2 This is a flowchart illustrating the specific execution process of the present invention, which collects instantaneous current values ​​and implements two-stage hybrid smooth updates of three-phase RMS values.

[0072] Figure 3 This is a schematic diagram of the module structure connection of the motor phase loss detection system of the present invention. Detailed Implementation

[0073] The present invention will now be described in detail with reference to the accompanying drawings.

[0074] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0075] The median refers to the value in the middle of a set of data arranged in ascending order. In this invention, it represents the healthy phase level of a three-phase current and is more resistant to outliers, such as the zero current during a phase loss, than the average value. The segment_count counter refers to a counting mechanism with different trigger and reset thresholds to prevent frequent output state jumps caused by signal fluctuations near critical points.

[0076] The two-sample-one-calculation architecture refers to a three-phase current where two currents are obtained through sampling, and the third-phase current is obtained by the negative sum of the other two phase currents, Ic = -(I a +I b According to Kirchhoff's laws, the third phase current contains the sampling errors of the first two phases.

[0077] First-order low-pass filtering (IIR) refers to a highly efficient digital filtering algorithm whose core function is to allow low-frequency components of a signal to pass smoothly while effectively suppressing high-frequency noise. In this invention, this filter is used for RMS calculation in the steady-state phase. It can filter out random noise generated by a single calculation while maintaining a fast response capability to load changes, achieving a good balance between smoothing effect and response speed.

[0078] Electronic switches, which can utilize IGBT power modules, are high-power electronic switches used to drive motors. They can control the switching of large currents using small current signals. They play a crucial role in managing the transmission and cutoff of high-power electrical energy in high-voltage, high-current applications such as frequency converters and motor drives.

[0079] A phase loss is understood as a fault state in which one or more phases of the three-phase power supply or winding of the motor are disconnected, resulting in the inability of current to flow normally. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0080] Example 1:

[0081] refer to Figure 1 , Figure 2 As shown, this embodiment discloses a motor phase loss detection method for electric vehicle drive fault detection, including the following steps:

[0082] S1. Initialization: Define the segment length N, accumulator segment_sum, counter segment_count, output value rms_value, and total_segments count;

[0083] S2, Collect the two-phase instantaneous current value I of the motor. a I b and according to the I a I b Calculate and obtain I c ;

[0084] S3. Calculate the instantaneous current value I for each sampling period. a I b and I c The sum of squares is accumulated in segment_sum, and the value of the counter segment_count is incremented by one;

[0085] instantaneous current I a The sum of squares is segment_count[a] = N times Ia 2 sum;

[0086] instantaneous current I b The sum of squares is segment_count[b] = N times Ib 2 sum;

[0087] instantaneous current I c The sum of squares is segment_count[c] = N times Ic 2 sum;

[0088] S4. When the value of the counter segment_count reaches N, calculate the RMS value R of each segment according to formula one. seg The calculation formula is as follows:

[0089] ;

[0090] Where segment-sum is the instantaneous current value I a I b and I c The sum of squares, where N is the segment length and R is the sum of squares. seg For each segment's RMS value;

[0091] S5, Two-level Hybrid Smooth Update:

[0092] During the startup phase, if the total number of completed segments (total_segments) is less than the segment count threshold, the RMS output value (rms_valueR) is calculated according to formula two. out [k], the second calculation formula is:

[0093] ;

[0094] Among them, R out[k] represents the RMS output value rms_value, R seg For each segment, k represents the RMS value, and k is the number of segments.

[0095] In the steady-state phase, when the total number of segments (total_segments) is greater than or equal to the segment count threshold, the system switches to first-order low-pass filtering and calculates the RMS output value (rms_valueR) according to Equation 3. out [k], the calculation formula three is:

[0096] ;

[0097] Where 0.9 represents the historical output weight and 0.1 represents the current input weight;

[0098] S6. Based on the RMS output value rms_value obtained in S5, set the RMS base value I. ref Set the reference coefficient K according to the motor speed. th and counting step up ;

[0099] S7. Based on the RMS output value rms_valueR out [k], RMS baseline value I ref and benchmark coefficient K th To determine if the motor has a phase loss fault:

[0100] S71. If the RMS output value R out [k]<K th ×I ref And C i < Single-phase suspected fault counter threshold, then C i =Ci+Step up The phase loss fault count increases; if R out [k]>=K th ×I ref C i= Ci-2, the phase loss fault count decreases; where C i It is a single-phase fault counter with a counting range of 0~100;

[0101] S72, If C i If the value is greater than or equal to the maximum single-phase count, then set the corresponding phase fault flag (Flag). i =1, motor set; if C i If the value is less than or equal to the minimum single-phase count, then clear the corresponding phase fault flag. i =0, motor reset; if the single-phase count is the minimum value. <C i If the single-phase count reaches its maximum value, then the Flag will be retained. i If it remains unchanged, the motor enters the hysteresis region.

[0102] In this invention, RMS is understood as the root mean square value / effective value. RMS refers to the square root of the average of the squares of the instantaneous values ​​of alternating current over one cycle. It represents the thermal equivalence of alternating current to a certain value of direct current. Compared to the absolute average value, RMS better reflects the true energy characteristics of the current and is less sensitive to noise. In this invention, the root mean square value is used to determine the balance of three-phase current; the interruption length N is understood as the duration of sampling, for example, 1000 points correspond to 100ms at a 10kHz sampling rate.

[0103] Specifically, S3 involves calculating the square of each instantaneous current value and accumulating it into the corresponding segment summation. This is achieved through a computer-executed program in a loop. For example:

[0104] segment_sum[3], segment_count[3], rms_value[3] and total_segment[3] are all three-dimensional arrays, corresponding to three-phase currents.

[0105] The RMS reference value I described in this invention ref This can be understood as being able to participate in determining whether the output RMS value rms_value has encountered a critical fault. The RMS output values ​​rms_value of the three-phase currents are calculated separately. The initial RMS value can be obtained through step S5. Based on the initial RMS value, in this preferred embodiment, step S6 specifically includes the step of: taking the median, maximum, or average value (excluding the maximum and minimum values) of the RMS output values ​​rms_value of the three-phase currents as the RMS reference value I. med According to the RMS reference value I med The RMS baseline value is calculated according to formula four; formula four is:

[0106] .

[0107] It should be noted that the RMS value mentioned in this invention is understood to be the value obtained based on the motor phase loss detection algorithm calculated using the root mean square (RMS) method. The simplest and most conventional RMS algorithm is calculated using formula five:

[0108] ;

[0109] Wherein I(t) i ) represents the current value, and N represents the number of current values.

[0110] The method for calculating the initial value of RMS described in this invention, namely the instantaneous square accumulation calculation module + the hybrid smoothing calculation module, has advantages over the ordinary RMS calculation method.

[0111] In this preferred embodiment, based on the calculated initial RMS value, the median of the RMS output values ​​rms_value of the three-phase currents is taken as the RMS reference value I. med As shown in calculation formula six:

[0112] ;

[0113] And a reference benchmark is set according to calculation formula four; the calculation formula four is:

[0114] ;

[0115] Therefore, even if the current is 0 when one phase is completely disconnected, the median can still be determined by the healthy values ​​of the remaining two phases and will not be pulled down, thus ensuring the reliability of the reference.

[0116] More specifically, in another implementation of this embodiment, the RMS reference value I med Replace with the maximum value; or the RMS reference value I. med Replace with the average value after removing the maximum and minimum values.

[0117] It should be noted that the segment threshold mentioned in this invention can be set according to experimental or phase loss fault judgment experience or actual calculated length requirements, and can be matched with motor model and operating requirements. For example, the segment threshold can be set to 30.

[0118] In S3 and S4 of this invention, the segment_count logic does not have to be a linear accumulator segment_count. In this preferred embodiment, a shift register logic that is triggered when the condition is met N times consecutively can be used.

[0119] The two-phase current sampling method described in this invention allows the instantaneous current of the third phase to be calculated using traditional calculation methods, thus making it suitable for most two-phase current sampling motor detection applications.

[0120] In this preferred embodiment, the reference coefficient K is set according to the motor speed. th and counting step up The specific steps include: if the absolute value of the motor speed is ≥2000 RPM, then K th =0.35, Step up =2; Corresponding to the high-speed range of the motor, the signal quality is good, improving sensitivity and speeding up fault confirmation; If 500RPM ≤ absolute value of motor speed < 2000RPM, then K th =0.40, Step up =1; corresponding to the medium speed range of the motor, achieving standard balance; if the absolute value of the motor speed is <500 RPM, then K th=0.40, Step up =1; Corresponding to the low-speed range of the motor, it achieves high low-speed harmonics and relaxes the threshold to prevent false alarms.

[0121] In this preferred embodiment, after step S5, a pre-locking check is performed to determine whether the locking condition is met. If the locking condition is met, the locking is triggered.

[0122] The interlocking conditions include: the electronic switch is not enabled, a delay is initiated within 150ms before power-on, a third trigger condition, and a fourth trigger condition.

[0123] The process of determining whether the third trigger condition has been met includes the following steps: calculating the motor's electrical frequency f based on the motor speed. e =RPM / 60; where RPM represents the motor speed, measured in revolutions per minute; if |f e If the frequency is less than 10Hz, the third trigger condition is met.

[0124] The determination of whether the fourth trigger condition has been met includes the following steps: if the RMS output value rms_value of the three-phase current is less than the preset current threshold, then the fourth trigger condition has been met.

[0125] The determination of whether the fourth trigger condition has been met includes the following steps: if the RMS output value rms_value of the three-phase current is less than the preset current threshold, then the fourth trigger condition has been met.

[0126] The specific implementation steps for the pre-lockout check are as follows: First, enable check: if the IGBT is not enabled, clear all faults; Second, start delay: within 150ms before system power-on, wait for RMS convergence and pause detection; Third, low-frequency lockout: calculate the electrical frequency fe=RPM / 60. If |fe|<10Hz, pause detection, as the 100ms RMS window may not cover the complete cycle, leading to calculation distortion; Fourth, low-current lockout: if all three-phase RMS are less than 0.5A, pause detection to avoid noise dominance; In short, if any of the above lockouts is triggered, maintain the previous state or reset directly, and do not execute the subsequent S6 steps.

[0127] In this embodiment, the maximum single-phase count value in S7 can be understood as a value that can be set based on fault experience or experiments. Different motors are set with corresponding maximum or minimum single-phase count values ​​to achieve the accuracy of the judgment and to maintain a single-phase counter Ci for each phase. In this embodiment, the maximum single-phase count value can be 40 and the minimum single-phase count value can be 5.

[0128] In a preferred embodiment, step S71 further includes the following steps: if the number of faulty phases fault_count=1, then output the single-phase fault code of the corresponding phase; if the number of faulty phases fault_count≥2, then output multiple fault codes; if the number of faulty phases fault_count=0, then output no fault code.

[0129] In this preferred embodiment, the two-stage hybrid smooth update is replaced by the following steps: During the startup phase, setting an initial gain value. A large time-varying gain IIR filter; the formula for calculating the time-varying gain IIR filter is:

[0130] ;

[0131] Where alpha[n] represents linear decay or exponential decay. Smooth transition to The function; in the steady state, alpha[n] = 0.1.

[0132] By appropriately increasing the gain, the RMS output value rms_value may decrease more quickly when a motor phase is actually lost, thus reducing the time required for the counter segment_countCi to accumulate to the threshold. However, due to the increased gain, the RMS output value rms_value may experience larger ripple due to normal current fluctuations. To reduce false alarms, the counting step size Step in step S71 should be adjusted accordingly. up Or relax the benchmark coefficient K th This achieves a balance between sensitivity and stability.

[0133] Example 2:

[0134] Based on Example 1, and referring to Figure 3 As shown, this embodiment discloses a motor phase loss detection system for electric vehicle drive fault detection, which is used to implement the motor phase loss detection method for electric vehicle drive fault detection described in Embodiment 1. The system includes: an instantaneous current acquisition module, an instantaneous square accumulation calculation module, a hybrid smoothing calculation module, a parameter setting module, a phase loss analysis calculation module, and a fault output module.

[0135] The instantaneous square accumulation calculation module is used to calculate the instantaneous current value I. a I b and I c The sum of squares, accumulated by segment_sum, and count;

[0136] The hybrid smoothing calculation module is used for two-level hybrid smoothing update calculation and obtaining the initial RMS value;

[0137] The parameter setting module is used to initialize and set parameters and thresholds;

[0138] The phase loss analysis and calculation module is used to determine whether a phase loss fault exists;

[0139] The fault output module is used to output phase loss fault and fault phase data.

[0140] The detection system of this invention is particularly suitable for cost-sensitive home appliances, power tools, or light electric vehicle drives, which are typically equipped with only two current sensors. The traditional symmetrical component method performs poorly on these devices, while this solution offers greater practicality and improved accuracy.

[0141] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting motor phase loss in electric vehicle drive systems, characterized in that, Including the following steps: S1. Initialization: Define the segment length N, accumulator segment_sum, counter segment_count, output value rms_value, and total_segments count; S2. Collect the two-phase instantaneous current value I of the motor. a I b and according to the I a I b Calculate and obtain I c ; S3. Calculate the instantaneous current value I for each sampling period. a I b and I c The sum of squares is accumulated in segment_sum, and the value of the counter segment_count is incremented by one; instantaneous current I a The sum of squares is segment_count[a] = N times Ia 2 sum; instantaneous current I b The sum of squares is segment_count[b] = N times Ib 2 sum; instantaneous current I c The sum of squares is segment_count[c] = N times Ic 2 sum; S4. When the value of the counter segment_count reaches N, calculate the RMS value R of each segment according to formula one. seg The calculation formula is as follows: ； Where segment-sum is the instantaneous current value I a I b and I c The sum of squares, where N is the segment length and R is the sum of squares. seg For each segment's RMS value; S5, Two-level Hybrid Smooth Update: During the startup phase, if the total number of completed segments (total_segments) is less than the segment count threshold, the RMS output value (rms_valueR) is calculated according to formula two. out [k], the second calculation formula is: ； Among them, R out [k] represents the RMS output value rms_value, R seg For each segment, k represents the RMS value, and k is the number of segments. In the steady-state phase, when the total number of segments (total_segments) is greater than or equal to the segment count threshold, the system switches to first-order low-pass filtering and calculates the RMS output value (rms_valueR) according to Equation 3. out [k], the calculation formula three is: ； Where 0.9 represents the historical output weight and 0.1 represents the current input weight; S6. Based on the RMS output value rms_value obtained in S5, set the RMS base value I. ref Set the reference coefficient K according to the motor speed. th and counting step up ; S7. Based on the RMS output value rms_valueR out [k], RMS baseline value I ref and benchmark coefficient K th To determine if the motor has a phase loss fault: S71. If the RMS output value R out [k]<K th ×I ref And C i < Single-phase suspected fault counter threshold, then C i =Ci+Step up The phase loss fault count increases; if R out [k]>=K th ×I ref C i= Ci-2, the phase loss fault count decreases; where C i It is a single-phase fault counter with a counting range of 0~100; S72, If C i If the value of the single-phase count is greater than or equal to the maximum value, then set the corresponding phase fault flag (Flag). i =1, motor set; if C i If the value is less than or equal to the minimum single-phase count, then clear the corresponding phase fault flag. i =0, motor reset; if the single-phase count is the minimum value. <C i If the single-phase count reaches its maximum value, then the Flag will be retained. i If it remains unchanged, the motor enters the hysteresis region.

2. The motor phase loss detection method for electric vehicle drive fault detection according to claim 1, characterized in that, S6 specifically includes the following steps: The median of the three-phase current RMS output values ​​rms_value is taken as the RMS reference value I. med According to the RMS reference value I med The RMS baseline value is calculated according to formula four; formula four is: 。 3. The motor phase loss detection method for electric vehicle drive fault detection according to claim 2, characterized in that, The RMS reference value I med Replace with the maximum value; or the RMS reference value I. med Replace with the average value after removing the maximum and minimum values.

4. The motor phase loss detection method for electric vehicle drive fault detection according to claim 1, characterized in that, The reference coefficient K is set according to the motor speed. th and counting step up The specific steps include: If the absolute value of the motor speed is ≥2000 RPM, then K th =0.35, Step up =2; If 500 RPM ≤ absolute value of motor speed < 2000 RPM, then K th =0.40, Step up =1; If the absolute value of the motor speed is <500 RPM, then K th =0.40, Step up =1.

5. The motor phase loss detection method for electric vehicle drive fault detection according to claim 1, characterized in that, After S5, a pre-lock check is performed to determine whether the locking condition is met. If the locking condition is met, the locking is triggered. The interlocking conditions include: the electronic switch is not enabled, a delay is initiated within 150ms before power-on, a third trigger condition, and a fourth trigger condition. The process of determining whether the third trigger condition has been met includes the following steps: calculating the motor's electrical frequency f based on the motor speed. e =RPM / 60; where RPM represents the motor speed, measured in revolutions per minute; if |f e If the frequency is less than 10Hz, the third trigger condition is met. The determination of whether the fourth trigger condition has been met includes the following steps: if the RMS output value rms_value of the three-phase current is less than the preset current threshold, then the fourth trigger condition has been met.

6. The motor phase loss detection method for electric vehicle drive fault detection according to claim 1, characterized in that, S71 further includes the following steps: If the number of faulty phases fault_count=1, then output the single-phase fault code of the corresponding phase; If the number of faulty phases fault_count ≥ 2, then enter multiple fault codes; If the number of faulty phases fault_count=0, then no fault code is entered.

7. The motor phase loss detection method for electric vehicle drive fault detection according to claim 1, wherein the two-stage hybrid smooth update is replaced by the following steps: During the startup phase, set the initial gain value. A large time-varying gain IIR filter; the formula for calculating the time-varying gain IIR filter is: ； in, alpha[n] represents linear or exponential decay. Smooth transition to The function; In the steady state phase, alpha[n] = 0.

1.

8. A motor phase loss detection system for detecting faults in an electric vehicle drive system, characterized in that, A motor phase loss detection method for implementing electric vehicle drive fault detection as described in any one of claims 1-7 includes: Instantaneous current acquisition module, instantaneous square accumulation calculation module, hybrid smoothing calculation module, parameter setting module, phase loss analysis calculation module, fault output module; The instantaneous square accumulation calculation module is used to calculate the instantaneous current value I. a I b and I c The sum of squares, accumulated by segment_sum, and count; The hybrid smoothing calculation module is used for two-level hybrid smoothing update calculation and obtaining the initial RMS value; The parameter setting module is used to initialize and set parameters and thresholds; The phase loss analysis and calculation module is used to determine whether a phase loss fault exists; The fault output module is used to output phase loss fault and fault phase data.

Images