Method for determining a frequency-dependent winding impedance and driving device for carrying out a method of eary

Abstract

The invention relates to a method for determining a frequency-dependent winding impedance, in which, for a different first frequency fn, sinusoidal voltages having a frequency corresponding to the first frequency fn are sequentially applied to at least one of the motor phases in each case in a period of time, a motor current value detected by a current sensor is converted into a digital data stream by a sigma-delta modulator, the data stream: being delivered to a first decimation filter; for a plurality of natural numbers k, each multiplied by a time-dependent sinusoidal signal having a frequency k times the respective first frequency fn, and forming a sliding mean on the resulting products to determine a respective first spectral component; the winding impedance Z is determined from the spectral components of the current and the voltage for each first frequency and k times thereof by multiplying, for a natural number k, a respective time-dependent cosine signal having a frequency of k times of the respective first frequency fn, and forming a sliding mean for the resultant product in order to determine a respective second spectral component, the winding impedance Z being determined from the spectral components of the current and the voltage for each first frequency and k times thereof.

Description

Technical Field

[0001] The present invention relates to a method for determining frequency-dependent winding impedance and a drive device for performing the method according to any one of the preceding claims. Background Technology

[0002] As is well known, the stator windings of an electric motor have winding impedance.

[0003] CN115166515A, as prior art, discloses a method for determining the stator insulation state.

[0004] US6483319B1 discloses a method for performing broadband impedance response to predict stator winding failure.

[0005] WO2016 / 029234A1 discloses a method for detecting the insulation condition of an AC motor.

[0006] WO2022 / 094726A1 discloses the determination and classification of the degree of insulation degradation of the stator winding of an electric motor. Summary of the Invention

[0007] Therefore, the purpose of this invention is to improve the operational safety of motors, and in particular, to enable the application of condition monitoring.

[0008] According to the present invention, this objective is achieved by the method according to claim 1 and the features of the drive device according to claim 11.

[0009] An important feature of this invention for determining the frequency-dependent winding impedance of a three-phase motor fed by a frequency converter or inverter with a current sensor is: Especially during the commissioning / trial operation of three-phase motors, different first frequencies f are required. n The frequency f is sequentially set to the corresponding first frequency according to the time period. n A time-dependent, particularly square-wave / rectangular voltage curve / voltage distribution is applied to at least one of the motor phases. The motor current value, detected by a current sensor, particularly in analog mode, is converted into a digital data stream, specifically a 1-bit data stream, by a Σ-Δ modulator. The data stream is fed to a first decimation filter to form a moving average / smoothed average, which is particularly used as the spectral component at zero frequency. - For each of the multiple natural numbers k, execute the first frequency f corresponding to the frequency. n The corresponding k times time-dependent sinusoidal signals are multiplied, and the moving average of the resulting product is used to determine the corresponding first spectral component. - For the natural number k, execute the first frequency f corresponding to the frequency. n The corresponding k times time-dependent cosine signals are multiplied, and the moving average of the resulting product is used to determine the corresponding second spectral component. For each first frequency and its k-th multiple, the winding impedance Z is determined by quotient of the spectral components of the current and voltage, specifically according to the following formula: Especially among them, U(f) n ) is the voltage at the first frequency f n spectral components at, I(f n ) is the current at the first frequency f n Fourier components at the location.

[0010] The advantage is that a square-wave voltage curve with a first frequency is first applied, and the spectral components of the current are determined for the resulting current curve / current distribution. Therefore, the real and imaginary parts of the Fourier transform of the current at this first frequency are known. Because the spectral components of the applied square-wave voltage curve for this first frequency are known or can be easily determined, where the amplitude of the voltage curve is already known without the need for sensor detection, the impedance can be easily determined by the quotient of the spectral components of the voltage at this first frequency and the Fourier transform of the current at this first frequency. The key here is that the detected current value is first converted into a 1-bit data stream by a Σ-Δ modulator, and then the spectral components are determined by the decimation filter when the product of the 1-bit data stream and a sine or cosine signal with the first frequency is fed to the input of a corresponding simple decimation filter. In a subsequent step, the frequency of the first sine or cosine signal is increased, i.e., increased by a factor of k, and the impedance at that corresponding frequency is determined again. This is done for multiple integer values ​​of k, e.g., k=1, then k=2, then k=3, etc.

[0011] In this way, the impedance value can be determined for each given first frequency value and its integer multiples. The first frequency f used is thus chosen. n The quantity of the natural number k is such that the impedance can be represented by at least 30 points in the frequency domain.

[0012] Therefore, it is possible to monitor whether anti-resonance occurs in the impedance, which may be caused by voltage stress in the stator windings. To this end, the impedance, especially its magnitude, can be monitored, either by detecting unacceptably high deviations from its value curve relative to a given value or relative to a frequency-dependent value, or by determining the local minimum of the impedance in the frequency domain to detect any emerging anti-resonance. Thus, the motor's operating status can be monitored, and the reliability of motor operation can be improved.

[0013] A further advantage of this invention is that the semiconductor switches present in the converter or frequency converter, configured for controlled operation, can provide square-wave voltages for the corresponding phase windings according to the invention. Especially during commissioning, there is sufficient time to sequentially measure each motor phase in time as described above, thereby individually determining the winding impedance of each, particularly all three motor phases, and especially the phase windings. Furthermore, according to the invention, current sensing devices configured for controlled operation in the converter or frequency converter can be used. Therefore, the invention can be implemented without additional hardware costs.

[0014] In an improved approach, only a single decimation filter can be used. The current curve multiplied by the sine or cosine signal is repeatedly fed to this decimation filter to determine the corresponding spectral components at the corresponding first frequency or at a corresponding k-fold multiple of the first frequency. For k=0, even a simple averaging can be performed, thus requiring only one decimation filter overall. The disadvantage is that the discrete data stream on the output side of the moving average can only be provided at relatively long time intervals.

[0015] An important feature of the method for determining the frequency-dependent winding impedance of a three-phase motor fed by a frequency converter or inverter with a current sensor according to claim 2 is: Especially during the commissioning and operation of three-phase motors, For different first frequencies f n The frequency f is sequentially set to the corresponding first frequency according to the time period. n A time-dependent, especially square-wave-shaped voltage curve is applied to at least one of the motor phases accordingly; The motor current value, detected by a current sensor, particularly in an analog manner, is converted into a digital data stream by a Σ-Δ modulator, and this data stream is then transmitted to... -First decimation filter, - A first multiplier, which performs multiplication on a plurality of natural numbers k with a time-dependent sinusoidal signal, the frequency of which is a corresponding first frequency f. n k times, - A second multiplier performs multiplication on a plurality of natural numbers k with a time-dependent cosine signal, the frequency of which is a corresponding first frequency f. n k times; The output signal of the first multiplier is sent to the second decimation filter to determine the corresponding first spectral component. The output signal of the second multiplier is fed to the third decimation filter to determine the corresponding second spectral component; For each first frequency and its k-fold frequencies, the winding impedance Z is determined by quotienting the spectral components of the current and voltage, specifically according to the following formula: Especially among them, U(f) n ) is the voltage at the first frequency f n spectral components at, I(f n ) is the current at the first frequency f n Fourier components at the location; In particular, the corresponding spectral components of the voltage are determined by calculation based on the voltage curve applied by the frequency converter or inverter.

[0016] The advantage is that a square-wave voltage curve with a first frequency is first applied, and the spectral components of the current are determined for the resulting current curve. Therefore, the real and imaginary parts of the Fourier transform of the current at this first frequency are known. Because the spectral components of the square-wave voltage curve at this first frequency are known or can be determined in a simple way—where the amplitude of the voltage curve is known rather than detected by a sensor—the impedance can be easily determined by the quotient of the spectral components of the voltage at this first frequency and the Fourier transform of the current at this first frequency. The key here is that the detected current value is first converted into a 1-bit data stream by a Σ-Δ modulator, and then the spectral components are determined by the decimation filter when the product of the 1-bit data stream and a sine or cosine signal with the first frequency is fed to the input of a corresponding simple decimation filter. In a subsequent step, the frequency of the first sine or cosine signal is increased, specifically by a factor of k, and the impedance at that corresponding frequency is determined again. This is done for multiple integer values ​​of k, for example, k=1, then k=2, then k=3, and so on.

[0017] In this way, the impedance value can be determined for each given first frequency value and its integer multiples. The first frequency f used is thus chosen. n A quantity of integer k such that the impedance can be represented by at least 30 points in the frequency domain.

[0018] Therefore, it is possible to monitor whether anti-resonance occurs in the impedance, which may be caused by voltage stress in the stator windings. To this end, the impedance, especially its magnitude, can be monitored, either by detecting unacceptably high deviations from its value curve relative to a given value or frequency-dependent values, or by determining local minima of the impedance in the frequency domain to detect emerging anti-resonance. Thus, the motor's operating status can be monitored, and the motor's operational reliability can be improved.

[0019] Another advantage of this invention is that the semiconductor switches present in the converter or frequency converter, configured for controlled operation, can provide square-wave voltages for the corresponding phase windings according to the invention. Especially during commissioning, there is sufficient time to sequentially measure each motor phase in time as described, thereby individually determining the winding impedance of each, especially the three motor phases, and particularly the phase windings. Furthermore, according to the invention, the current sensing device configured for controlled operation in the converter or frequency converter can be used, thus enabling the invention to be implemented without additional hardware costs.

[0020] In one improved approach, a single decimation filter can be used. The current curve, multiplied by a sine or cosine signal, is repeatedly fed to this filter to determine the corresponding spectral component at a first frequency or a k-fold of the first frequency. For k=0, even a simple averaging can be performed, thus requiring only one decimation filter overall. The disadvantage is that the discrete data stream output from the moving average can only be provided at relatively long time intervals.

[0021] In an advantageous design, at least one local minimum of the determined frequency-dependent winding impedance is identified, particularly a local minimum of the magnitude of the determined frequency-dependent winding impedance, and it is monitored whether this exceeds an unacceptably high deviation relative to a given value. The advantage is that motor condition monitoring can be implemented simply and without additional hardware components, since the Σ-Δ modulator and decimation filter are already present in the microcontroller of the frequency converter or inverter.

[0022] A method for monitoring a three-phase motor by determining and evaluating the frequency-dependent winding impedance of the motor fed by a frequency converter or inverter with a current sensor, particularly including the above-described method, has the following important feature: Especially during or in conjunction with commissioning and operation, When the motor rotor shaft is stationary or at low speed, especially as a condition monitoring function, in controlled operation of a motor fed by a converter operating in pulse width modulation mode, during the application of a zero vector, the winding impedance Z is determined at least at a single first frequency, and the magnitude of the winding impedance Z is monitored for an unacceptably high deviation relative to a predetermined threshold. In order to determine the winding impedance Z at a single first frequency, a square wave voltage with a frequency of the first frequency is applied to at least one of the motor phases. The motor current value, detected by a current sensor, particularly in an analog manner, is converted into a digital data stream by a Σ-Δ modulator, and this data stream is then transmitted to... - First decimation filter; - The first multiplier performs AND operations at a frequency of a first frequency f.n Multiplication of time-dependent sinusoidal signals; - A second multiplier performs multiplication with a time-dependent cosine signal at a frequency equal to the first frequency; The output signal of the first multiplier is fed to the second decimation filter to determine the first spectral component. The output signal of the second multiplier is fed to the third decimation filter to determine the second spectral component. The winding impedance Z is determined by quotienting the spectral components of the current and the amplitude of the voltage, specifically according to the following formula: , Especially among them, U(f) n ) is the voltage at the first frequency f n spectral components at, I(f n ) is the current at the first frequency f n Fourier components at the location.

[0023] Its advantage is that it can perform status monitoring during operation by re-determining the impedance when the speed is extremely low or zero, especially at least at a single first frequency, thereby enabling the detection of changes and the display or transmission of warnings based on the monitoring results.

[0024] In an advantageous design, each decimation filter has its own sinc² filter, i.e., a second-order sinc filter, with a first decimation rate M1 and downstream downsampling. Its output signal is fed to a sinc filter, i.e., a first-order sinc filter, with a second decimation rate M2 and downstream downsampling. The advantage is that the second-order sinc filter suppresses or at least reduces quantization noise, while the first-order sinc filter determines the moving average of its input signal. Therefore, sinc filters can be used without additional cost, as they are integrated into the converter's signal electronics FPGA or microcontroller.

[0025] In a favorable design, the first decimation rate M1 is between 31 and 65. This has the advantage of suppressing or at least reducing quantization noise.

[0026] In an advantageous design, the second decimation rate M2 is determined such that the first-order Sinc filter, in particular together with the downstream downsampling module, can effectively average the pulse width modulation period duration of the pulse width modulation controlled power semiconductor switches of the frequency converter or inverter over an exact integer multiple, particularly exactly four times. The advantage is that averaging can be performed over the pulse width modulation period, thereby enabling the determination of the moving average.

[0027] In an advantageous design, the voltage spectral components used to determine the winding impedance are obtained from the Fourier series of a voltage, particularly a square-wave voltage. The advantage is that the voltage spectral components are known in advance, as the spectral components of the applied voltage are mathematically given, thus eliminating the need for calculation or measurement to determine them.

[0028] In a favorable design, the corresponding first spectral component is the imaginary part of the current, and the corresponding second spectral component is the real part of the current. The advantage is that the Fourier transform of the current can be easily determined by mixing in sine and cosine signals and integrating using a first-order Sinc filter.

[0029] For drive units with motors fed by inverters or frequency converters, an important feature is that the Σ-Δ modulator, as well as the digital filter and / or the Sinc filter as an on-chip Sinc filter, are arranged in the microcontroller or set in the FPGA of the inverter or frequency converter.

[0030] Its advantage is that implementing the method according to the invention requires no additional cost and can utilize existing frequency converters or inverters. Essentially, only software adaptation is required.

[0031] An important feature of the method for determining the reliability of a drive device is that the drive device has a three-phase motor fed by a converter or frequency converter, wherein the frequency-dependent winding impedance of the three-phase motor is determined according to the method of any one of the preceding claims, wherein at least one local minimum of the frequency-dependent winding impedance is determined and monitored to see if it exceeds the allowable deviation level relative to a given value.

[0032] Its advantage is that by determining the frequency-dependent winding impedance, a local minimum can be determined, thereby enabling the determination of whether anti-resonance occurs.

[0033] Other advantages arise from the dependent claims. The invention is not limited to the combination of features of the claims. For those skilled in the art, there are other meaningful combinations of features of the claims and / or individual claims and / or features of the specification and / or drawings, particularly from the task description and / or the task derived by comparison with the prior art. Attached Figure Description

[0034] The present invention will now be described in detail with reference to the schematic diagrams: Figure 1 A portion of the method according to the invention is illustrated schematically, wherein the sliding mean of the current and its spectral components are determined.

[0035] Figure 2 An analog-to-digital converter with a Σ-Δ modulator 20 and a decimation filter 21 is schematically shown.

[0036] Figure 3 Suggestively illustrating Sinc K The decimation filter has a digital low-pass filter and a downstream downsampling module 31.

[0037] Figure 4 The diagram schematically illustrates the decimation filter structure that performs averaging over the pulse width modulation (PWM) cycle. Detailed Implementation

[0038] As shown in the figure, in the method according to the present invention, the analog current value detected by the current sensor is sent to the Σ-Δ converter 4, the output data stream of which is a digital 1-bit data stream, and is sent to the first decimation filter 1 on the one hand, and to other decimation filters (2, 3) through multipliers (5, 6) on the other hand, to determine the real and imaginary parts of the Fourier components of the current.

[0039] Inside the decimation filters (1, 2, 3), the 1-bit data stream is converted into a parallel output data stream by the Sinc² filter, and its quantization noise is reduced.

[0040] Based on the Fourier components of the current, the current can be expressed as a complex parameter. Therefore, when the frequency converter or converter provides a feedback voltage U, the winding impedance can be determined according to the following formula: , It can also monitor the impedance of the winding, such as its anti-resonance, local minimum, resonant peak or local maximum, to improve operational safety.

[0041] Here, it can be used as U(f) n The voltage U at frequency f is n The spectral components at point n, where n is the point number in the discrete Fourier spectrum.

[0042] The drive unit has a three-phase motor fed by a frequency converter or converter via connecting cables, wherein the motor current is detected by a current sensor. Here, the current flowing in each phase of the motor is detected or determined. The semiconductor switches in the final stage of the frequency converter or converter are driven in a pulse-width modulation (PWM) manner.

[0043] Here, the final stage includes three series circuits connected in parallel, wherein each of the three series circuits respectively includes two semiconductor switches connected in series, particularly MOSFETs or IGBTs, wherein each of the semiconductor switches has a freewheeling diode connected in parallel.

[0044] By driving the semiconductor switch with pulse width modulation (PWM), the frequency and amplitude of the motor voltage can be freely adjusted by setting the PWM ratio accordingly. The PWM switching frequency is, for example, fPWM = 8kHz, and the fundamental oscillation / harmonic associated with the motor is, for example, 50Hz. The motor winding inductance LW smooths the current. Only the fundamental oscillation amplitude produces the desired torque of the motor.

[0045] The frequency converter or inverter operates in current regulation mode, therefore the current is sampled at twice the switching frequency, so that the switching moment of each positive and negative PWM switching edge can be used for regulation. At a switching frequency of 8kHz, the sampling frequency of the regulation loop is f. s =16kHz. According to the sampling theorem, frequencies up to half the sampling frequency can be analyzed using this sampled current signal, which is f in this case. s / 2=8kHz. However, according to the present invention, by using Sinc K The filter also allows higher frequency spectral components to be acquired and evaluated.

[0046] Anti-resonance in the high-frequency characteristics of the motor stator winding indicates that an unspecified voltage load has occurred during operation. This type of voltage load is a major cause of premature winding insulation failure.

[0047] Frequency f m The minimum impedance characteristic at a certain point, known as anti-resonance, indicates series resonance. In the series resonance state, the capacitive reactance of the parasitic capacitance in the series circuit... It compensated for part of the inductive reactance of the winding. The characteristic of series resonance is that the voltage applied to the inductor Lm or capacitor Cm can be significantly greater than the voltage applied to the motor terminals.

[0048] Therefore, by determining the frequency-dependent winding impedance and its local minimum, it is possible to monitor whether the load, especially the load accumulated during operation, deviates from a given value beyond permissible limits, thereby improving operational safety. This is because warning messages can be displayed and / or transmitted before motor failure.

[0049] To determine the frequency-dependent winding impedance, a square-wave symmetrical AC voltage is applied to the first phase of the motor during commissioning. Its Fourier analysis can be expressed as: , Where fPWM is the switching frequency, for example fPWM=8kHz, and TPWM is the pulse width modulation period, for example when the duty cycle is 50%.

[0050] Therefore, the switching frequency of the square wave voltage, for example, 8 kHz, has spectral components at f1=8 kHz, f3=24 kHz, f5=40 kHz, f7=56 kHz, f9=72 kHz, etc.

[0051] For higher frequency spectral components, the latch-up time and edge steepness of the transistor can also be considered. However, the latch-up time and rise time are relatively short, especially when using silicon carbide (SiC) transistors, which are considered particularly critical for windings.

[0052] Therefore, by applying a load to allow current to flow through the first motor phase winding, its spectral components are determined sequentially over time for different frequencies: Figure 1 In the method shown, k=1 is first chosen. Here, the current value multiplied by sin(kωt) is sent to the second decimation filter 2 to determine the imaginary part of the current, i.e., the spectral component. Especially, where n=k.

[0053] Similarly, the current value multiplied by cos(kωt) is fed to the third decimation filter 3 to determine the real part of the current, i.e., the spectral component. Especially, where n=k.

[0054] In order to determine the spectral components for other values ​​of k in the same way, the current curve i(t) used to determine the spectral components when k=1 is stored and reused accordingly when determining with other values ​​of k.

[0055] The total current formed by the combination of such defined spectral components can be expressed as: .

[0056] Therefore, higher frequency spectral components, especially those exceeding 8 kHz, can be determined using the method according to the invention.

[0057] In the next step of the method according to the invention, the angular frequency ω of the square wave voltage curve provided by the frequency converter or converter is changed, and spectral components are determined accordingly for each angular frequency ω and for a plurality of k values.

[0058] By sequentially providing a sufficient number of square-wave voltage curves with different angular frequencies ω from a frequency converter or inverter, and by selecting a sufficient number of k values, a sufficient current Fourier spectrum can be obtained. This allows for the accurate identification of local minima in the frequency spectrum, thereby identifying unacceptable deviations from predetermined values, i.e., monitoring the motor. For example, the number of both should be chosen such that the Fourier spectrum can be represented by 100 to 1000 points, particularly determining the spectral components for 100 to 1000 frequency values. In practice, such determination typically takes 1 to 20 minutes.

[0059] like Figure 2 As shown, the detected analog time-continuous motor phase current is converted into a digital bit stream, particularly a 1-bit data stream, using a Σ-Δ modulator 20. The Σ-Δ modulation frequency is preferably f = 20 MHz. A downstream decimation filter 21 reduces quantization noise and periodically forms sampled data words from the motor phase current, i.e., a parallel digital data stream.

[0060] like Figure 3 As shown, the decimation filter 21 has a digital filter 30, especially Sinc K The filter can be configured by two parameters: the order K, preferably K=2 or K=3; and the decimation rate M, preferably M=128. Higher values ​​of K and M result in a lower cutoff frequency for the filter, but improve the suppression of quantization noise, thereby increasing the effective resolution of the analog-to-digital conversion.

[0061] Preferably, the Σ-Δ modulator 20 is implemented as a second-order modulator in the integrated circuit and has integrated electrical isolation functionality.

[0062] like Figure 3 As shown, in digital filter 30, especially Sinc K Downsampling module 31, especially M-order, can be connected downstream of the filter.

[0063] However, according to the present invention, the first decimation filter 1, the second decimation filter 2, and the third decimation filter 3 are used Figure 4 The structure shown.

[0064] like Figure 4 As shown, the structure here has a first digital filter 40, which is located in the first stage of the decimation filter structure, and a first downsampling module 41 is connected downstream of the first downsampling module 41. A second digital filter 42 (located in the second stage of the decimation filter structure) is connected downstream of the first downsampling module 41, and then a second downsampling module 43 is connected.

[0065] The first digital filter is constructed as a Sinc² filter, i.e., a second-order filter with K=2. Here, the decimation rate M1 of the first digital filter 40 is a value in the range of 32 to 64.

[0066] The second digital filter is used as a mean filter and is implemented as a Sinc¹ filter, i.e., a first-order K=1 filter.

[0067] Here, the decimation rate M2 of the second digital filter 42 is chosen such that the averaging is performed exactly as an integer multiple of the pulse width modulation period duration T, i.e., TI = N. TPWM. This closely approximates the integral of the time curve over its period duration T, and is particularly consistent with the integral in Fourier analysis. Therefore, the second digital filter 42 synchronously forms a moving average over exactly one PWM cycle with the PWM.

[0068] Figure 4 The structure shown is preferably used as a basis Figure 1 The first decimation filter 1 in the implementation scheme also serves as the second decimation filter 2 and the third decimation filter 3. Therefore, all three decimation filters (1, 2, 3) have the same structure.

[0069] Therefore, the moving average of the detected current signal is determined by means of the first decimation filter 1.

[0070] Therefore, a spectral component, particularly the imaginary part of the Fourier component of the current, is determined by means of the second decimation filter 2, especially since the detected current is used as the input signal of the second decimation filter 2 after being multiplied by a sine wave (signal).

[0071] Therefore, another spectral component, especially the real part of the Fourier component of the current, is determined by the third decimation filter 3, especially since the detected current is used as the input signal of the third decimation filter 3 after being multiplied by the cosine (signal).

[0072] Preferably, M2 is chosen such that averaging is performed over multiple complete PWM cycles, particularly exactly four PWM cycles, i.e., TI = 4TPWM. Thus, despite the stronger quantization noise of the Σ-Δ modulator at higher frequencies, accuracy is still improved.

[0073] Therefore, according to the present invention, by selectively changing the pulse width modulation period duration TPWM or the pulse width modulation frequency f... PWM Then, all specified spectral components of the current are measured sequentially.

[0074] In other embodiments of the invention, during controlled operation of the motor, when the rotor shaft of the motor is stationary or rotating at very low speed, during the application of a zero vector in pulse width modulation operation, it may also be in accordance with... Figure 1The spectral components are determined in the manner shown. Therefore, although a densely distributed Fourier spectrum cannot be determined, for different k values, the spectral components at at least one frequency of the square-wave voltage provided by the frequency converter or converter can be determined, thereby allowing monitoring of the motor winding impedance at at least some points in the Fourier spectrum.

[0075] In other embodiments of the invention, the first decimation filter 1 can be omitted by also considering the value of k=0 in the third decimation filter 3, thus enabling the moving average to be performed without a cosine function. In this way, only two decimation filters (2, 3) are required.

[0076] In other embodiments of the invention, each of the decimation filters (1, 2, 3) does not employ the aforementioned Sinc² and Sinc¹ filters with downstream downsampling modules, but instead employs a Sinc³ filter (i.e., a third-order Sinc filter having a first decimation rate M) and a downstream downsampling module, wherein the Sinc³ filter is triggered synchronously with the pulse width modulation frequency, and in particular, M = 128 or 256.

[0077] List of reference numerals

[0078] 1 First Decimation Filter

[0079] 2. Second Decimation Filter

[0080] 3 Third Decimation Filter

[0081] 4 Current sensor with Σ-Δ modulator

[0082] 5. Multipliers

[0083] 6. Multipliers

[0084] 20 Σ-Δ modulator

[0085] 21 Decimation Filter

[0086] 30 digital filters, especially Sinc K filter

[0087] 31 Downsampling Module

[0088] 40 The first digital filter, especially the first stage of the decimation filter structure.

[0089] 41 Downsampling Module

[0090] 42 The second digital filter, especially located in the second stage of the decimation filter structure

[0091] 43 Downsampling Module

[0092] The order of the K-harmonic oscillation

[0093] ω angular frequency

[0094] τ is the integral variable.

[0095] T-cycle duration

[0096] t time

[0097] i Current

[0098] f n frequency

Claims

1. A method for determining the frequency-dependent winding impedance of a three-phase motor fed by a frequency converter or inverter with a current sensor. Its features are, Especially during the commissioning and operation of three-phase motors, For different first frequencies f n A time-dependent, particularly square-wave-shaped voltage curve is sequentially applied to at least one of the motor phase windings at time intervals, the frequency of which is a corresponding first frequency f. n , The motor current value, detected by a current sensor, particularly in analog mode, is converted into a digital data stream, specifically a 1-bit data stream, by a Σ-Δ modulator. - It is fed to the first decimation filter to form a moving average, specifically the moving average is used as the spectral component at frequency zero. -For multiple natural numbers k, each is associated with a first frequency f. n Multiply by a time-dependent sinusoidal signal of k times, and take a moving average of the product to determine the corresponding first spectral component, especially the imaginary part. -For the natural number k, respectively, and with the corresponding first frequency f. n Multiply by a k-fold time-dependent cosine signal and take a moving average of the product to determine the corresponding second spectral component, especially the real part. For each first frequency and its k-fold multiple, the winding impedance Z is determined based on the spectral components of the current and voltage, particularly by quotient, according to the following formula: Especially among them, U(f) n ) is the voltage at the first frequency f n spectral components at, I(f n ) is the current at the first frequency f n Fourier components at the location.

2. A method for determining the frequency-dependent winding impedance of a three-phase motor fed by a frequency converter or inverter with a current sensor. Its features are, Especially during the commissioning and operation of three-phase motors, For different first frequencies f n A time-dependent, particularly square-wave-shaped voltage curve is sequentially applied to at least one of the motor phases, with the frequency of the voltage curve being a corresponding first frequency f. n , The motor current value detected by the current sensor, particularly in an analog manner, is converted into a digital data stream, specifically a 1-bit data stream, by a Σ-Δ modulator. This data stream is then transmitted to... - First decimation filter; - A first multiplier, which performs multiplication on a plurality of natural numbers k with a time-dependent sinusoidal signal, the frequency of which is a corresponding first frequency f. n k times; - A second multiplier performs multiplication on a plurality of natural numbers k with a time-dependent cosine signal, the frequency of which is a corresponding first frequency f. n k times; The output signal of the first multiplier is fed to the second decimation filter to determine the corresponding first spectral component, especially the imaginary part. The output signal of the second multiplier is fed to a third decimation filter to determine the corresponding second spectral component, particularly the real part. For each first frequency and its k-fold frequencies, the winding impedance Z is determined based on the spectral components of the current and voltage, particularly by quotient, according to the following formula: Especially among them, U(f) n ) is the voltage at the first frequency f n spectral components at, I(f n ) is the current at the first frequency f n Fourier components at the point, In particular, the corresponding spectral components of the voltage are determined by calculation based on the voltage curve applied by the frequency converter or the inverter.

3. The method according to any one of the preceding claims, characterized in that, Determine at least one local minimum of the determined frequency-dependent winding impedance, in particular a local minimum of the magnitude of the determined frequency-dependent winding impedance, and monitor whether there is an unacceptably high deviation of the local minimum relative to a predetermined value.

4. A method for monitoring a motor, wherein the motor is monitored by means of determining and evaluating the frequency-dependent winding impedance of a three-phase motor fed by a frequency converter or inverter with a current sensor, particularly comprising the method according to claim 1 or 2. Its features are, Especially during or in conjunction with commissioning and operation, When the motor rotor shaft is stationary or at low speed, especially for condition monitoring, in controlled operation of a motor fed by a converter operating in pulse width modulation mode, during the application of a zero vector, the winding impedance Z is determined at least at a single first frequency, and the winding impedance Z, especially at least the magnitude of the winding impedance Z, is monitored for an unacceptably high deviation relative to a predetermined threshold. In order to determine the winding impedance Z at a single first frequency, a square wave voltage with a frequency of the first frequency is applied to at least one of the motor phases. The motor current value, detected by a current sensor, particularly in an analog manner, is converted into a digital data stream, specifically a 1-bit data stream, by a Σ-Δ modulator. This data stream is then transmitted to... - First decimation filter; - The first multiplier performs AND operations at a frequency of a first frequency f. n Multiplication of time-dependent sinusoidal signals; - A second multiplier that performs multiplication with a time-dependent cosine signal at a frequency of the first frequency; The output signal of the first multiplier is fed to the second decimation filter to determine the first spectral component, particularly the imaginary part. The output signal of the second multiplier is fed to a third decimation filter to determine the second spectral component, particularly the real part. The winding impedance Z is determined based on the spectral components of the current and the amplitude of the voltage, particularly by quotient, according to the following formula: Especially among them, U(f) n ) is the voltage at the first frequency f n spectral components at, I(f n ) is the current at the first frequency f n Fourier components at the location.

5. The method according to any one of the preceding claims, characterized in that, Each decimation filter has its own Sinc² filter, i.e., a second-order Sinc filter, which has a first decimation rate M1 and a downstream downsampling module. The output signal of the Sinc² filter is fed to the Sinc filter, i.e., a first-order Sinc filter, which has a second decimation rate M2 and a downstream downsampling module. or, Each decimation filter has a Sinc³ filter, i.e., a third-order Sinc filter, which has a first decimation rate M and a downstream downsampling module. The Sinc³ filter is triggered synchronously with the pulse width modulation frequency, and specifically, M = 128 or 256.

6. The method according to any one of the preceding claims, characterized in that, The first extraction rate M1 has a value of 31 to 65.

7. The method according to any one of the preceding claims, characterized in that, The second decimation rate M2 is determined such that the first-order Sinc filter, in particular together with the downstream downsampling module, can average the duration of exactly an integer number, especially exactly four, pulse width modulation cycles of the power semiconductor switches controlled by the pulse width modulation of the inverter or frequency converter.

8. The method according to any one of the preceding claims, characterized in that, The voltage spectrum components used to determine the winding impedance are obtained from the Fourier series of voltages, especially those in a square wave shape.

9. The method according to any one of the preceding claims, characterized in that, The corresponding first spectral component is the imaginary part of the current, and the corresponding second spectral component is the real part of the current.

10. The method according to any one of the preceding claims, characterized in that, The first and second multipliers are arranged in the field-programmable gate array of the converter or the frequency converter, wherein digital data streams, especially 1-bit data streams, are also fed to the field-programmable gate array.

11. A drive device having a motor fed by a converter or frequency converter for performing the method according to any one of the preceding claims. Its features are, The Σ-Δ modulator, along with the corresponding decimation filter, especially the digital filter, and / or the Sinc filter as an on-chip Sinc filter, is arranged in the microcontroller or in the field programmable gate array of the inverter or converter.

12. The driving device according to any one of the preceding claims, characterized in that, The data stream output from the output side of the Σ-Δ modulator, especially the 1-bit data stream, is fed to the corresponding multiplier to be multiplied with the sine or cosine signal.

13. A method for determining the reliability of a drive device, particularly according to claim 11 or 12. in, The drive unit includes a three-phase motor powered by a converter or frequency converter. The frequency-dependent winding impedance of the three-phase motor is determined according to the method described in any one of the preceding claims. Its features are, Determine at least one local minimum of the frequency-dependent winding impedance and monitor whether the deviation of the local minimum from a predetermined value exceeds the allowable deviation amount.

14. A driving device for performing the method according to claim 13, characterized in that, The drive unit includes a three-phase motor powered by a converter or frequency converter.

Images