A method for monitoring inter-turn insulation degradation based on winding impedance spectrum division

Abstract

The application relates to the technical field of permanent magnet synchronous motor fault monitoring, and discloses a turn-to-turn insulation degradation monitoring method based on winding impedance spectrum division, which establishes a lumped parameter model of a stator winding, analyzes winding impedance spectrum division for insulation state monitoring on the basis, verifies the influence of turn-to-turn insulation degradation on different regions of the winding impedance spectrum, proposes a turn-to-turn insulation monitoring index extracted from a transient current response based on impedance spectrum division, and analyzes a determination method of qualified PWM excitation. The application divides the frequency domain sensitive to turn-to-turn insulation degradation in the impedance spectrum, establishes a direct corresponding relationship between turn-to-turn insulation degradation and the transient current response, and monitors insulation degradation through the transient current response.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of permanent magnet synchronous motor fault monitoring, and in particular to an inter-turn insulation degradation monitoring method based on winding impedance spectrum division. BACKGROUND

[0002] The insulation system of inverter-fed motors (IFMs) is continuously subjected to various aging stresses during operation, gradually degrading, especially with the popularization of wide-bandgap devices in inverters, higher dv / dt and switching frequency greatly accelerate this process. The switching transient response method is a commonly used method for monitoring the stator insulation of inverter motors in recent years. It can detect insulation degradation at an early stage, thereby preventing serious insulation faults.

[0003] However, the relationship between the phase current transient response and the inter-turn insulation degradation in the above method cannot be determined, so the degree of insulation aging of the motor cannot be accurately monitored.

[0004] Due to the weak degradation characteristics of inter-turn insulation, it is difficult to monitor its early degradation, and it is considered to be one of the main research gaps for low-voltage and medium-voltage motors. Although the existing method uses the amplitude change of the frequency band centered on the parallel resonance point of the transient current transient response to detect the inter-turn insulation condition, the different trends of the winding impedance and switching transient current response spectrum sensitive to inter-turn insulation degradation in different ranges are not analyzed in detail. On the other hand, the switching transient response is directly affected by the PWM voltage excitation mode of the inverter, which may dynamically change with the working conditions of the IFMs. SUMMARY

[0005] The purpose of the present application is to solve the above problems, and a winding impedance spectrum division-based inter-turn insulation degradation monitoring method is designed.

[0006] To achieve the above purpose, the technical scheme of the present application is a winding impedance spectrum division-based inter-turn insulation degradation monitoring method, which comprises:

[0007] Establishing a high-frequency stator winding model to determine different impedance spectrum sensitive regions of inter-turn insulation degradation;

[0008] Establishing an analytical relationship between inter-turn insulation degradation and the corresponding sensitive regions of winding impedance and switching transient current response spectrum;

[0009] Extracting inter-turn insulation monitoring indicators based on impedance spectrum division of the transient current response, and obtaining qualified PWM voltage excitation.

[0010] In the high-frequency stator winding model, C g1 and C g2These are the equivalent capacitances between the winding end coil and the stator core, and between the neutral wire and the stator core, respectively; C g1 With C g2 The sum of these is the equivalent capacitance of the stator windings and stator core at low frequencies; C t This is due to the capacitive coupling effect within the stator winding; R e R is the equivalent resistance for losses; L is the winding resistance; d For winding inductance.

[0011] The process for determining the sensitive regions of different impedance spectra is as follows:

[0012] Based on the high-frequency stator winding model, the single-phase winding impedance Z to ground is determined. PG for:

[0013]

[0014] In the formula, Z PG The inter-turn insulation capacitance C is determined by the test frequency and model parameters. t Depending on the configuration, shape, and materials of the insulation system, the capacitance C increases when the inter-turn insulation deteriorates. t It will change, C t It increases with the degree of inter-turn insulation degradation;

[0015] Z PG Amplitude versus C t The partial derivatives are:

[0016]

[0017] in,

[0018]

[0019]

[0020] In the formula, K is always positive;

[0021] Solving equation (2), the signs of the partial derivatives for different frequency bands are obtained as follows:

[0022]

[0023] In equation (5), the boundary frequencies f1 and f2 can be calculated based on (4)-(6) in combination with the model parameters. The sensitive regions of different impedance spectra are determined by dividing them using f1 and f2.

[0024] The extraction process of the inter-turn insulation monitoring indicators is as follows:

[0025] The spectrum of the transient current response is represented as I(g), where g represents the discrete frequency order, and the sampling rate F is used. s and sample size Ns corresponding frequency f is calculated g as shown in the following formula:

[0026]

[0027] The turn-to-turn insulation monitoring index of different impedance spectrum sensitive regions is defined as IIMI1 p,k and IIMI2 p,k

[0028]

[0029] In the formula, p is a p-phase winding, p=A, B, C; K is K th switching event; f1 and f2 are the boundary frequencies of the above-mentioned impedance spectrum division;

[0030] The turn-to-turn insulation monitoring index IIMI p The expression of the turn-to-turn insulation monitoring index IIMI

[0031]

[0032] Wherein, m is the number of switching events.

[0033] The rising edge of the PWM voltage under positive load current or the falling edge of the PWM voltage under negative load current is selected as the qualified PWM voltage excitation.

[0034] Compared with the prior art, the application has the following advantages:

[0035] 1. The application utilizes the high-frequency model of the motor stator winding, theoretically determines different impedance spectrum sensitive regions of turn-to-turn insulation degradation, and establishes an analytical connection between turn-to-turn insulation degradation and winding impedance and corresponding sensitive regions in the switching transient current response spectrum;

[0036] 2. The application determines the qualified PWM voltage excitation of I FMs in normal operation, greatly improves the reliability of the monitoring result;

[0037] 3. The application does not require additional signal injection equipment, is non-invasive, and is easy to apply in industry. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is a principle diagram of the winding impedance spectrum division-based turn-to-turn insulation degradation monitoring method described in the application;

[0039] Figure 2 is a boundary calculation flowchart of the impedance region division described in the application;

[0040] Figure 3 is a high-frequency model diagram of the stator single-phase winding described in the application;

[0041] Figure 4 is the statistical diagram of the rise time (t r ) and the amplitude of the switching transient current response (I amp ) of the PWM voltage under different switching events in one basic period according to the present application;

[0042] Figure 5 is the switching transition diagram of the half-bridge inverter with positive current polarity according to the present application; in the diagram, (a) t < t1, (b) t1 < t < t2, (c) t2 < t < t3, (d) t > t3;

[0043] Figure 6 is the output voltage diagram of the half-bridge inverter when the load current is (a) positive and (b) negative according to the present application;

[0044] Figure 7 is the statistical diagram of the rise time (t r ) and the fall time (t f ) of the PWM voltage under different switching events in one basic period under different load currents according to the present application. DETAILED DESCRIPTION

[0045] The present application will be described in detail below with reference to the accompanying drawings, as shown in Figures 1-7 ;

[0046] The research on the switching transient characteristics mainly focuses on the frequency range of several hundred kHz to several MHz, and thus the modeling of the stator winding only considers its high-frequency characteristics. The lumped parameter single-phase high-frequency winding model of the PMSM (permanent magnet synchronous motor) is shown in Figure 3 , wherein C g1 and C g2 are the equivalent capacitances between the winding end coil and the stator core and between the neutral line and the stator core, respectively; the sum of C g1 and C g2 is the equivalent capacitance of the stator winding and the stator core at low frequencies; C t is the capacitive coupling effect in the stator winding; R e is the loss equivalent resistance; R is the winding resistance; and L d is the winding inductance. R can be ignored because it is much smaller than the winding inductance (ωL d ) at high-frequency bands.

[0047] In order to obtain accurate monitoring indicators, it is necessary to identify different regions in the winding impedance spectrum that are sensitive to interturn insulation degradation. As can be known from the model shown in Figure 3 , the single-phase winding impedance to ground Z PG is

[0048]

[0049] ZPG determined by the test frequency and the model parameters. The inter-turn insulation capacitance C t depends on the configuration, shape and material of the insulation system, and changes when the inter-turn insulation deteriorates, C t t increases with the degree of inter-turn insulation degradation. To illustrate the impact of C t on |Z PG , we derive the partial derivative of Z PG with respect to C t , which is given by

[0050]

[0051] where,

[0052]

[0053] K is always positive, and the sign of the partial derivative determines the trend of |Z t when C PG changes. Solving (2), we further derive that the sign of the partial derivative at different frequency bands is given by

[0054]

[0055] The boundary frequencies f1 and f2 can be calculated according to (4)-(6) combined with the model parameters. Therefore, the single-phase winding-to-ground impedance can be divided into three regions, and the impact of inter-turn insulation degradation on the impedance amplitude is different in different regions. In region 1 with frequency lower than f1 and region 3 with frequency higher than f2, the impedance amplitude decreases with the inter-turn insulation degradation, because the partial derivative of Z PG with respect to C t is negative. Existing research shows that the inter-turn capacitance only starts to affect the winding impedance at relatively high frequency range. Therefore, in region 1, the impact of inter-turn insulation on the impedance can be neglected, because the impedance in this region is mainly determined by the GW insulation (ground wall insulation). On the other hand, in region 2 between f1 and f2, the impedance amplitude increases with the severity of degradation. In region 3 with frequency higher than f2, the impedance amplitude decreases. The procedure of calculating the impedance region division boundary is shown in Figure 2 .

[0056] ​It is verified that the upper frequency f2 of region 2 decreases as the degradation level increases. As the degradation severity increases, the negative ΔZ (impedance change, normalized by the health condition value, denoted as ΔZ) in region 2 increases, but the positive ΔZ increases more, so the overall ΔZ increases. However, if region 2 is used for insulation monitoring, the weakening effect of the increase in degradation level on the monitoring index becomes apparent. In region 3, the impedance amplitude continuously decreases as the degradation severity increases. Therefore, the index established in region 3 can more accurately monitor insulation degradation of different severities.

[0057] For the installed motor, it is more convenient to use current ratio measurement to measure winding impedance from the perspective of condition monitoring. The switching transient current response is determined by the PWM voltage as high-frequency excitation and the stator winding as load, and the impedance change of the winding caused by inter-turn insulation degradation can be extracted from the transient current response. Therefore, the corresponding frequency components in the transient current response can be extracted based on the impedance spectrum division to evaluate the inter-turn insulation condition.

[0058] According to the above analysis, as the degradation severity increases, for the frequencies between f1 and f2 in region 2, the winding impedance amplitude increases, so the amplitude of the corresponding frequency component in the current response will decrease; for the frequencies higher than f2 in region 3, the winding impedance amplitude decreases, and the amplitude of the corresponding frequency component in the current response increases. The frequency spectrum of the transient current response is denoted as I(g), where g represents the discrete frequency order, and the sampling rate F s and the number of samples N s The corresponding frequency f g is calculated as follows:

[0059]

[0060] Based on the two inter-turn insulation monitoring indexes of region 2 and region 3, respectively defined as IIMI1 p,k and IIMI2 p,k

[0061]

[0062] where p is the p-phase winding, p = A, B, C; K is the K th switching event; f1 and f2 are the boundary frequencies of the impedance spectrum division described above. IIMI1 p,k and IIMI2 p,k are the superpositions of the amplitudes of the corresponding frequency components of the current responses in region 2 and region 3, respectively. As the degradation level increases, IIMI1 p,k decreases, IIMI2 p,k increases.

[0063] As discussed above, the degradation degree weakens the monitoring index based on Region 2. Therefore, the change of IIMI1 p,k will be smaller than IIMI2 p,k , which means IIMI2 p,k is more sensitive to the degradation of inter-turn insulation. Therefore, the inter-turn insulation monitoring index IIMI p proposed in this paper is based on IIMI2 p,k , as shown in the following formula:

[0064]

[0065] where m is the number of switching events. The switching frequency of the inverter is usually above several kHz, so a large number of transient current responses can be obtained within a few milliseconds. In formula (9), this averaging can effectively reduce the influence of noise and system uncertainty.

[0066] Since the switching frequency of the inverter is usually several kHz or above, there are a large number of switching events when the variable frequency motor system is running normally, and good monitoring accuracy can be obtained by simply averaging the characteristic parameters. The specific inter-turn insulation state evaluation process is shown in Figure 1 . During the operation of the motor, the high-frequency voltage and current sensors are used to collect the inverter three-phase voltage output and motor current, respectively, and the corresponding transient current responses are extracted according to the PWM voltage excitation of different phases, combined with the impedance region division boundary frequency to extract the inter-turn insulation state evaluation indexes IIMI2 A , IIMI2 B , IIMI2 C of different phase windings, and realize the evaluation of the inter-turn insulation state of the motor winding.

[0067] The PWM voltage excitation mode of the inverter directly affects the switching transient current response. Therefore, in order to ensure the effectiveness of the monitoring results, it is necessary to determine the qualified PWM voltage excitation.

[0068] The high-frequency characteristics of the PWM voltage mainly depend on its rise time. Figure 4 is a statistical diagram of the rise time (t r ) of the PWM voltage and the corresponding switching transient current response amplitude (I amp ) under different switching events in a fundamental period, where t r is the time for the PWM voltage to rise from 10% to 90% of the DC voltage. It can be seen that the smaller t r is, the larger I amp is, because the shorter the rise time, the higher the excitation voltage. In the positive half cycle of the load current, t r has the smallest value, and I ampThe values ​​are the largest. More importantly, they are almost constant, which is of great significance for monitoring the winding insulation condition based on switching transient currents.

[0069] To explain the above phenomenon in detail, studies were conducted on... Figure 5 The diagram shows two IGBTs (S) H and S L The circuit diagram of a half-bridge inverter. The main parasitic component is the gate-emitter capacitor (C). ge Collector-gate capacitance (C) cg ) and collector-emitter capacitance (C ce When current flows out from the midpoint (A) of the half-bridge, the polarity of the output current is defined as positive. Figure 4 The detailed switching transitions when the output current is positive are given, (a)-(c) represent the transitions from the upper switch S. H Switch S L The conduction phase change, (c)-(d) represent the opposite cases.

[0070] Figure 5 The corresponding output voltage is as follows Figure 6 As shown, where V AN This is the voltage between the midpoint A and the negative point N of the DC link. When S H When turned on, the load current flows through S H V AN +V DC At time t1, for S H Apply a shutdown signal to S L The anti-parallel diode applies a reverse voltage. Due to the load current, S... H Output capacitor (loss = C) cg +C ce ) charging, thus making S L Output capacitor (loss = C) cg +C ce Discharge, therefore V AN The current will not drop rapidly. The drop time depends on the instantaneous values ​​of the load current and the output capacitor loss, as shown in equation (10). When the current is fully discharged S... L When the output capacitor is , the voltage V AN When the voltage drops to 0V, current flows through S at point t2. L Anti-parallel diodes, such as Figure 5 As shown in (c). The current will flow directly from S. L The anti-parallel diode is transferred to S H And V AN It will immediately rise from 0V to +V DC In this case, V AN From 0V to +V DCthe rise time of V g is controlled only by the gate resistance (R r ) of the gate driver. Since the voltage rises very fast, the rising edge of VAN is recorded as a step edge, as shown in Figure 6 (a) at t3.

[0071]

[0072] When the load current is negative, according to similar principles, the switching transition shown in Figure 6 (b) can be obtained, in which the rising edge t r is a slope, and the falling edge t f is a step. In summary, when the load current is positive, the falling edge of V AN is affected by the load current and the parasitic capacitance. Conversely, when the current is negative, the rising edge of V AN is affected by the load current and the parasitic capacitance.

[0073] Figure 7 The statistics of the rise time (t r ) and the fall time (t f ) of the PWM voltage at different switching events within a fundamental period under different load currents are shown, where t f is the time for the PWM voltage to drop from 90% to 10% of the DC voltage. When the current is positive, t r is short and is not affected by the load current; when the current is negative, t r is greater than t r when the current is positive, and the greater the absolute value of the negative current, the smaller t r is, and t f shows the opposite trend. Therefore, to ensure the consistency of the PWM excitation characteristics, the rising edge of the PWM voltage under positive load current or the falling edge of the PWM voltage under negative load current can be chosen as the qualified PWM voltage excitation. This way, the impact of the load current on the switching transient response can be minimized.

[0074] It is to be understood that the terms such as first and second, etc., are used herein only to distinguish one from another without necessarily implying any actual relationship or order. Also, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, the statement "includes one of A and B" does not foreclose the existence of unlimited other of A and B.

[0075] The above technical solutions only reflect the preferred technical solutions of the present application, and some changes made by the skilled in the art to some parts thereof also reflect the principles of the present application and are within the scope of protection of the present application.

Claims

1. A method for monitoring inter-turn insulation degradation based on winding impedance spectrum division, characterized in that, include: A high-frequency stator winding model was established to determine the sensitive regions of different impedance spectra in the inter-turn insulation degradation. Establish an analytical relationship between inter-turn insulation degradation and the corresponding sensitive regions in the winding impedance and switch transient current response spectra; The transient current response based on impedance spectrum division is used to extract inter-turn insulation monitoring indicators to evaluate the inter-turn insulation status, and the evaluation results are verified using qualified PWM voltage excitation. In the high-frequency stator winding model, C g1 and C g2 These are the equivalent capacitances between the winding end coil and the stator core, and between the neutral wire and the stator core, respectively; C g1 With C g2 The sum of these is the equivalent capacitance of the stator winding and stator core at low frequencies; R e R is the equivalent resistance of the loss; L is the winding resistance; d For winding inductance; The process for determining the sensitive regions of different impedance spectra is as follows: Based on the high-frequency stator winding model, the single-phase winding impedance Z to ground is determined. PG for: In the formula, Z PG The inter-turn insulation capacitance C is determined by the test frequency and model parameters. t Depending on the configuration, shape, and materials of the insulation system, the capacitance C increases when the inter-turn insulation deteriorates. t It will change, C t It increases with the degree of inter-turn insulation degradation; Z PG Amplitude versus C t The partial derivatives are: in, In the formula, K is always positive; Solving equation (2), the signs of the partial derivatives for different frequency bands are obtained as follows: In equation (5), the boundary frequencies f1 and f2 can be calculated based on (4)-(6) and the model parameters. The sensitive regions of different impedance spectra are determined by f1 and f2. The extraction process of the inter-turn insulation monitoring indicators is as follows: The spectrum of the transient current response is represented as I(g), where g represents the discrete frequency order, and the sampling rate F is used. s and sample size N s Calculate the corresponding frequency f g As shown in the following formula: Inter-turn insulation monitoring indicators for sensitive regions of different impedance spectra are defined as IIMI1. p,k and IIMI2 p,k In the formula, p is the p-phase winding, p = A, B, C; k is K th Switching events; f1 and f2 are the boundary frequencies of the impedance spectrum division mentioned above; Inter-turn insulation monitoring index IIMI p The expression is: Where m is the number of switch events.

2. The method for monitoring inter-turn insulation degradation based on winding impedance spectrum division according to claim 1, characterized in that, The PWM voltage excitation is selected from the rising edge of the PWM voltage under positive load current or the falling edge of the PWM voltage under negative load current as the qualified PWM voltage excitation.

Images