Calibrating current sensors in AC inverters

Abstract

In a method for generating calibration data (10a-c) for an internal current sensor (20a-c) of an inverter (2), the inverter (2) generates alternating current (18a-c) in its output lines (16a-c) during inverter operation (IB), external current sensors (34a-c) outside the inverter (2) measure external measurements (42a-c) of the alternating current (18a-c) through the internal current sensors (20a-c), the uncalibrated internal current sensors (20a-c) measure internal measurements (44a-c) of the alternating current (18a-c), and the calibration data (10a-c) for the internal current sensor (20a-c) are generated based on the measured internal measurements (44a-c) and external measurements (42a-c). A computing unit (52) performs the procedure and contains interfaces (54) for receiving the internal measurement values ​​(44a-c) and the external measurement values ​​(42a-c) and for outputting the calibration data (10a-c). A measuring arrangement (50) contains the computing unit (42) and the external current sensors (34a-c).

Description

[0001] The invention relates to the calibration of an internal current sensor of an inverter, wherein the inverter has an input for a DC electrical power and at least one output line for an AC electrical current generated from the DC power, wherein the internal current sensor is arranged in the output line for detecting the AC current.

[0002] From WO 2013 / 068156 A2, a method is known for calibrating a multi-phase, in particular three-phase, inverter which has one high-voltage switching element and one low-voltage switching element for each of its phases, as well as a current sensor for at least some of the phases. The following steps are proposed: switching off all switching elements, switching on one high-voltage switching element of a first phase and one low-voltage switching element of a second phase, measuring the currents flowing through the first and second phases, calculating an average value from the measured currents, and calibrating the inverter based on this average value.

[0003] The object of the present invention is to propose improvements with regard to the calibration of an internal current sensor of an inverter.

[0004] The problem is solved by a method according to claim 1. Preferred or advantageous embodiments of the invention and of other invention categories will become apparent from the further claims, the following description and the accompanying figures.

[0005] The method is designed for generating, i.e., determining, calibration data. This calibration data pertains to one or more internal current sensors of an inverter. In addition to the current sensors, the inverter has an input through which direct current (DC) is supplied or can be supplied. The inverter contains at least one, and in particular three, output lines. During inverter operation, the inverter generates alternating current (AC) in each output line, specifically three-phase AC in three output lines. The AC is generated from the DC power. Specifically, one phase of three-phase AC is generated in each of the three output lines.

[0006] In its operation, the so-called inverter operation, the inverter generates the alternating currents in the output lines from the supplied direct power and outputs the alternating currents via the output lines.

[0007] Each of the internal current sensors is located in exactly one of the output lines and serves to detect the alternating current currently flowing in that respective output line. Specifically, each of the three output lines has exactly one internal current sensor.

[0008] In this method, the inverter is operated in inverter mode. The alternating current generated by the inverter in each of the output lines flows through the respective internal current sensor. In other words, the internal current sensor can thus detect the alternating current flowing in each output line.

[0009] In each of the lines containing an internal current sensor to be calibrated, the alternating current is also measured by an external current sensor located outside the inverter. Thus, external current sensors outside the inverter measure the alternating current flowing through the respective internal current sensor. These external current sensors are highly accurate, calibrated reference current sensors that provide precise readings within the required measurement accuracy, serving as a reference for generating the calibration data.

[0010] In particular, the external current sensor is connected to the relevant output line (or its continuation outside the inverter) in the same way as the internal current sensor, but located outside the inverter. In other words, the external current sensor carries the same alternating current as the internal current sensor in the relevant output line and detects this current in the same way as the internal current sensor.

[0011] Within the inverter, the alternating current is also measured using the still uncalibrated internal current sensor. More precisely, internal readings of the alternating current are measured. Since these readings are taken from the uncalibrated internal current sensor, they can be inaccurate. For example, the internal current sensor might erroneously (averaging) measure an internal DC component, even though the alternating current does not actually contain such a component. The measured current values ​​(instantaneous values) might also be incorrectly scaled. Therefore, the internal readings can be erroneous, representing a measurement error that must be corrected through calibration by calibrating the internal current sensor, i.e., by assigning it the calibration data.

[0012] The calibration data for the internal current sensor are then determined / calculated / generated based on the measured internal and external measurements.

[0013] According to the invention, the advantage arises that the calibration data can be determined using alternating current during regular inverter operation. Therefore, it is unnecessary to supply direct current for calibrating the internal current sensor. This simplifies, for example, a measurement procedure commonly used in practice as part of a strip-end test, where corresponding direct currents are required in addition to alternating currents. This is explained in more detail below.

[0014] In a preferred embodiment, the inverter generates the alternating current in inverter mode at a predefinable frequency, which is therefore freely selectable or predefinable. In particular, this allows for the generation of calibration data, enabling various measurements (external / internal values) at different frequencies. This allows, for example, averaging the calibration data over several frequencies or determining frequency and / or phase responses of internal current sensors. The predefinable frequency is, in particular, at most 500 Hz or 400 Hz and at least 50 Hz or 75 Hz, and is preferably 100 Hz. The method thus makes it possible to specify the most suitable frequency for generating the calibration data or the internal and external values.

[0015] In a preferred embodiment, a slope and / or a zero offset are generated as calibration data for the internal current sensors. In other words, the zero offset shifts, for example, a measurement characteristic of the internal current sensor in the ordinate direction. Changing the slope alters its inclination to the abscissa. A zero offset and a slope are generally sufficient to adequately calibrate an internal current sensor.

[0016] In a preferred embodiment, an internal DC component of the alternating current is determined from the internal measurements, and the zero offset is calculated based on this determined internal DC component. Alternatively or additionally, an internal amplitude is determined from the internal measurements, and an external amplitude of the alternating current is determined from the external measurements, and the slope is calculated based on these internal and external amplitudes. Generating the corresponding calibration data in this way can be particularly simple and reliable, as explained in more detail below.

[0017] In a preferred embodiment, the external DC component of the alternating current is determined from the external measurements, and a fault condition of the inverter is identified based on this determined DC component. The fault condition of the inverter can be determined particularly easily if the fault-free inverter does not generate any DC components in the alternating currents. If a DC component is nevertheless measured using the (high-precision or sufficiently accurate) external current sensor, this indicates a faulty inverter.

[0018] In a preferred embodiment, the internal measurements from the internal sensor are taken as the internal waveform of the alternating current over time. This internal waveform is measured over a specific internal duration. Alternatively or additionally, the external measurements from the external sensor are taken as the external waveform of the alternating current over a specific external duration. In other words, the waveform of the alternating current within a time window, defined by the internal / external duration, is measured, recorded, and stored for later evaluation. This provides two waveforms of specific durations of alternating current measurements from the internal current sensor and the external current sensor. Using a frequency transformation, the aforementioned internal amplitude in the frequency domain is determined from the internal waveform.Alternatively or additionally, the external amplitude is determined in the frequency domain from the external time course. In other words, the determination of the alternating current amplitudes is carried out in the frequency domain. This yields the well-known advantages of transitioning from time-domain to frequency-domain analyses.

[0019] The time windows for the internal and external durations can coincide, overlap, or be staggered. The only crucial factor is that the same alternating currents are recorded and measured.

[0020] In a preferred embodiment, the internal and external durations are chosen to be of equal length. In other words, the alternating current is detected by both the external and internal sensors within time windows of equal length. The corresponding measured, recorded, and stored internal and external profiles are therefore identical in duration, which simplifies their processing.

[0021] In a preferred embodiment, the internal duration and / or the external duration is selected such that it is an even multiple of a period (inverse of the frequency) of the alternating current. In other words, the internal sensor and / or the external sensor each detect, measure, record, and store a specific number of exactly complete periods of the alternating current. This allows the use of certain frequency transformation methods that are based on the evaluation of one or more periods of periodic alternating signals.

[0022] In a preferred embodiment of this system, the internal and / or external current waveform is measured, recorded, and stored as digital samples of the alternating current. The measured waveform thus constitutes a data set of samples taken during the respective time window, which can be directly processed digitally. In other words, digitally sampled alternating current waveforms are measured and evaluated. This yields the well-known advantages of digital signal processing.

[0023] In a preferred embodiment, the sampling intervals between two samples and / or the period of the alternating current (i.e., its frequency, see above) are selected such that the period is an even multiple of the sampling intervals of the internal and / or external waveforms. In other words, the period of the alternating current is divided into exactly an even number of sampling intervals. This allows signal processing methods to be used for further evaluation of the measured values ​​that require such a division of the period.

[0024] In a preferred embodiment of this method, the internal amplitude is determined from the internal response and / or the external amplitude from the external response using a discrete Fourier transform (DFT) as a frequency transformation. This provides all the usual advantages of using the well-known DFT for the present method.

[0025] In a preferred embodiment, the internal measurement values ​​are transmitted from the inverter. The generation of the calibration data then takes place outside the inverter. Subsequently, after its generation, the determined calibration data is transmitted into the inverter and used there to calibrate the internal current sensor. This offers the advantage that the potentially complex generation of the calibration data does not have to take place within the inverter; therefore, no corresponding computing / evaluation equipment, storage facilities, etc., need to be provided in the inverter. The inverter can thus be kept simple, simple, and cost-effective, yet the calibration method according to the invention can still be used for it.

[0026] In a preferred embodiment, the method is performed as part of an end-of-line (EOL) test on the inverter during its manufacturing process. In other words, the calibration is carried out during, at the end of, or immediately after completion of the inverter. It is common practice to operate the inverter in inverter mode during such an end-of-line test, supplying it with DC power and verifying its functionality while generating AC currents. Without additional effort, and in particular without having to switch to supplying the inverter with DC voltage / current, the calibration data can therefore also be generated during this end-of-line test while the inverter is in inverter mode. This results in savings of effort and time during the end-of-line test.

[0027] The object of the invention is also achieved by a computing unit according to claim 14. This unit is configured to execute the method according to the invention. For this purpose, the computing unit includes a respective interface for receiving internal and external measured values ​​and for outputting the calibration data. These interfaces can be logical, e.g., in the form of program code, or physical. Individual interfaces can also be combined. In particular, a single interface is used. At least one of the interfaces can be implemented as a CAN bus interface.

[0028] The interfaces are primarily for the internal current sensors or the inverter and for the external sensors, in order to receive the internal and external measured values ​​and to send the calibration data. As explained above, a prerequisite for this procedure is that the inverter is operating in inverter mode. The processing unit therefore includes, in particular, an interface configured to put the inverter into inverter mode and, if necessary, to set a frequency for the alternating current.

[0029] The object of the invention is also achieved by a measuring arrangement according to claim 15. This arrangement comprises the computing unit according to the invention and the external current sensors. Optionally, it also includes a current sink or current diversion for receiving the alternating current from the output lines and / or diverting the current from one output line back to another output line, e.g., in the case of three-phase alternating current. The current sink is, in particular, a choke arrangement. The measuring arrangement is thus sufficient to calibrate the inverter according to the method of the invention.

[0030] The invention is based on the following findings, observations, and considerations and further comprises the following preferred embodiments. These embodiments are sometimes referred to simply as "the invention." The embodiments may also include parts or combinations of the embodiments mentioned above, correspond to them, and / or may include previously unmentioned embodiments.

[0031] According to the invention, a calibration of current sensors is achieved.

[0032] The invention is based on the following practical observation: For the DC calibration of current sensors after their manufacture, at the end of a production line / process (end of line, EOL), a 500A DC high-current power supply and high-current relays are required. For calibration, the inverter under test must be disconnected from a previously connected high-voltage supply (high-voltage end-of-line test) and discharged. Then, high-current relays must be switched and the high-current power supply switched on. After calibration, the inverter is reconnected to the high-voltage supply and powered up for further testing. This procedure is time-consuming during EOL testing. Another disadvantage is that a maximum current of 500A can be used for DC calibration. Higher currents damage the power semiconductors in the inverter.For the calibration of the current sensors, it is important to calibrate with a high current in order to achieve an accurate calibration.

[0033] The basic idea of ​​the invention is this: Calibration is performed using alternating current generated by the inverter itself (inverter operation). A command is sent to the inverter to operate at a specific current and frequency (generating alternating current of a particular frequency and current) and hold it for a few seconds. The calibration is then carried out during test operation. The advantages are: The calibration current can be applied at the maximum current the inverter can supply. No high-current power supply or relay is required. The inverter also does not need to be powered down and on again to switch between current and voltage supply. Calibration takes place during a test phase in which the inverter is already running. This saves testing time.

[0034] The motivation for the invention is as follows: The idea for AC calibration (alternating current) stems from the consideration that the AC sensors (internal current sensors) should be calibrated at currents >500A. For DC calibration (direct current), the limit is 500A. The IGBTs or SiC FETs in the inverter cannot tolerate more DC current. Furthermore, the effort required to calibrate with DC at end-of-life (EOL) is high, as an additional high-current source is needed, as well as high-current relays for the targeted short-circuiting of the phases (output lines of the inverter). These additional components are not required for AC calibration. The phase current (alternating current in the relevant output line) generated by the inverter is used for calibration.

[0035] The idea behind the invention is this: The inverter, during operation, sets a specific operating point with the desired peak current (amplitude of the alternating current). This can be, for example, up to 900 A. The values ​​from the current sensors (internal measurements) are recorded and compared with the values ​​from the reference current transformers (external measurements). For projects with peak currents up to 1.8 kA, the advantage of AC calibration is significant and it is the only method for calibrating the current sensors.

[0036] Additionally, AC calibration has the advantage that measurements can be taken at different electrical frequencies (of the generated alternating current). This is particularly useful for ironless current sensors (internal current sensors), which exhibit frequency response, allowing for the measurement of both frequency and phase response.

[0037] In particular, the calibration process is carried out as follows: For the calculation, the phase currents (alternating current in the output line) in the inverter are recorded for a few periods (internal measurements). These values ​​are analyzed using a discrete Fourier transform (DFT) without a window, and the phasor length "A1" of the fundamental frequency (frequency of the generated alternating current) and the DC component "A0" are calculated.

[0038] The same procedure is performed with the three measurement signals (external measured values) of the reference current transformers (external current sensor).

[0039] The DFT has the great advantage that disturbances such as harmonics and noise are very strongly attenuated.

[0040] The pointer lengths "A1" (from internal measurements) of the phase currents, U, V, W (three alternating currents as three-phase current in the three output lines), are compared with the pointer lengths "A1" of the reference current transformers (from external measurements), U, V, W, to determine the calibration values. The calculated gain and offset values ​​are stored as calibration values / calibration data in the inverter.

[0041] One disadvantage of the DFT is that, due to the fixed sampling frequency (4 kHz) of the microcontroller (internal current sensor), the test software (WTSW: factory test software) cannot analyze every arbitrary frequency (of the alternating current). The WTSW is loaded ("flashed") into the inverter to enable "proper" testing during the end-of-line test. Numerous different tests are performed on the inverter's hardware (the inverter itself) under the control of this WTSW. These many tests cannot be performed with the actual (operational) software. The final software for the inverter, and therefore the final version for manufacturing, is loaded into the inverter shortly before delivery to the customer.

[0042] For each sinusoidal waveform of the alternating current, an integer number of samples must always be present (the period is an exact multiple of the sampling interval); otherwise, the calculated pointer length "A1" will have a significant error. The same applies to the current values ​​from the reference current transformer (external current sensor), which has a higher sampling frequency than the microcontroller (internal current sensor). Therefore, the electrical frequency (of the alternating current) must be calculated and set very precisely for calibration.

[0043] For example, a calibration frequency (of the alternating current) of 100 Hz is selected. This can be adjusted depending on the project. Another limitation is the storage depth of the measured values ​​in the microcontroller (i.e., how many samples of the alternating current can be stored as an internal waveform from the internal current sensor in the inverter).

[0044] For example, in an inverter, a sample frequency of 4 kHz (internal current sensor) corresponds to a sampling interval of 250 µs. A measurement frequency (alternating current) of 100 Hz corresponds to an alternating current period of 100 ms. Thus, 10 ms / 250 µs = 40 samples per period can be generated (an even number).

[0045] In the reference current transformer (external current sensor): A sample frequency of 32 kHz corresponds to a sampling interval of 31.25 µs. At the same measurement frequency of 100 Hz, this results in 10 ms / 31.25 µs = 320 samples per period (an even number).

[0046] Therefore, choosing 100 Hz as the calibration frequency is appropriate. Other frequencies can also be chosen, but the number of samples per period must be taken into account.

[0047] Here are some other possible calibration frequencies: Calibration frequency N Sensor 4kHz N Reference current transformer 32kHz [Hz] [Samples / Period] [Samples / Period] 100 40 320 200 20 160 400 10 80 500 8 64 800 5 40 1000 4 32

[0048] Since the WTSW samples the internal current sensors at 4 kHz, only a few points (N=40) are available even at 100 Hz per sine period. Therefore, calibration frequencies above 400 Hz (N=10) are not recommended.

[0049] The following DFT principles apply. The DFT (Discrete Fourier Transform) analyzes alternating current, measured by the reference current transformer (external measurements) or the ADC (analog-to-digital converter) values ​​of the sensors (internal measurements), at a constant and known frequency (of the alternating current). The fundamental frequency A1 (harmonic m=1) is analyzed. It is important that the actual frequency of the measured current (alternating current) exactly matches the assumed frequency (of the DFT). This assumed frequency is defined as the fundamental frequency.

[0050] It should be noted that a sine wave from the measurement signal (alternating current) must be sampled with an integer number of sine waves (complete periods), as described above. As the brief example calculation shows, a measurement frequency of exactly 100 Hz or 200 Hz would be ideal, given the specified sampling frequencies of 4,000 samples / s and 32,000 samples / s (external current sensor), sensor voltage, and reference current.

[0051] The DFT is calculated as follows: X(m)=∑n=0N−1Sample(n)[cos(2πnmN)−j sin(2πnmN)] Description: X(m) Result of the harmonics: m (In the calculation m = 1, 3, 5, 7, 9) Sample(s) Sample values n Index of the sample values. N: 0, 1, 2, ... N-1 N Number of samples fanalyse(m)=m fsN f S Sample Frequency m Harmonic N number of samples

[0052] The result X(1) is complex. The magnitude of the fundamental wave is calculated. |X(1)|=Re{X(1)}2+Im{X(1)}2

[0053] The calculation of the offset (zero offset) is generally described as follows: The offset of the sensors and the reference current transformer is determined during the DFT calculation. Since only an integer number of sine periods and an integer number of samples per sine period are used, the DC component "A0" is calculated as the average of all samples from the used sine periods. The offset of the reference current transformer approaches zero because the currents of all three phases (U, V, W) should not contain any DC component, as measurements also show. If DC components are present, then an error exists and a corresponding response is required (inverter fault condition). The DC component of the reference current transformer "A0" is used as an error indicator and is not used in the offset calculation.

[0054] Each phase has its own gain and offset value for each ADC type.

[0055] The calculation of the offset of a V-ADC (Versatile Analog-to-Digital Converter) is described as follows: The three A0 values ​​of the three phases correspond to the offsets of the sensors (internal current sensor). Since the V-ADC uses the operating voltage Vcc of the sensors as a reference, and the current sensor outputs the signal ratiometrically to Vcc, the influence of Vcc is compensated. Thus, A0 is the offset of the sensor (internal current sensor) and can be directly stored as a calibration value in the inverter. OffsetVADC=A0Sensor;[Digit]

[0056] The calculation of the offset of a DS-ADC (Delta Sigma Analog-to-Digital Converter) is described as follows: The DS-ADC has a CIC filter (Cascaded Integrator Comb Filter) for reducing the very high number of sample values. The CIC filter introduces a multiplicative component into the measurement result (internal measurement value). The gain depends on the CIC filter setting. The WTSW (Wireless Technology Software) sets the CIC filter for high measurement accuracy, and the application software for low phase shift. Therefore, A0 and A1 change depending on the CIC filter setting. To avoid constantly creating new WTSW versions with new CIC filter settings, a CIC coefficient, corresponding to the CIC filter setting, is used during calibration and calculated (divided) by A0. The resulting offset is dimensionless []. OffsetDSADC=A0Sensor[Digit]∗CICKoefficient[VDigit];[V]

[0057] The calculation of the gain (slope) of the V-ADC (internal current sensor) is described as follows: The calculation of GAIN_VADS can be performed directly. The value A1_REF from the reference current transformer after the DFT (external current sensor) is divided by the value A1_Sensor of the internal current sensors. The quotient is the desired gain value in [Digit / A]. This value is stored in the inverter. GAINVADC=A1LEMA1Sensor;[ADigit]

[0058] The calculation of the gain (slope) of the DS-ADC (internal current sensor) is described as follows: As described for the offset, the amplifying effect of the CIC filter must be factored out. To do this, A1_Sensor, after the DFT, is multiplied by the CIC coefficient, A1_(Sensor_Comp). Then, the magnitude of A1_REF (reference current transformer, external current sensor) is divided by A1_(Sensor_Comp) (internal current sensor). The product is the desired gain value in [V / A]. This value is stored in the inverter. GAINDSADC=A1LEMA1Sensor∗CICKoefficient;[AV]

[0059] Current calculations using calibrated values ​​in the application software are performed as follows: The application software calculates the phase currents (alternating currents) from the ADC (internal current sensor) values. Gain and offset values ​​are used for this calculation. Since, as already described, the ADCs use different conversion methods, the appropriate calculation must be applied for each ADC. Each phase has its own gain and offset value for each ADC type, thus 6 x gain and 6 x offset.

[0060] The V-ADC current calculation is performed as follows: With the V-ADC, the current is calculated directly. IPhase=(ValueVADC−OffsetVADC)∗GAINVADC;[A]

[0061] The DS-ADC current calculation is performed as follows: With this ADC, the CIC coefficient must again be considered in the current calculation. The CIC coefficient must be specifically defined in the application software, depending on the CIC filter settings, and is project-specific. IPhase=((ValueDSADC∗CICKoefficient)−OffsetDSADC)∗GAINDSADC;[A]

[0062] The EOL calibration process takes place in several steps: Steps Description 1 Approach operating point for I (alternating current, amplitude, frequency) 2 Reference current transformer: synchronously records current values ​​of the three phases over several sine cycles (external current sensor) 3 LE triggers for recording the V-ADC values ​​(internal current sensor). All three phases are recorded "synchronously" at 500 samples per phase. 4 Read logger data from the LE (internal history) 5 LE triggers for recording the DS-ADC values ​​(internal current sensor). All three phases are recorded "synchronously" at 500 samples per phase. 6 Read logger data from the LE (internal history) 7 V-ADC: Calculation of gain and offset via DFT 8 V-ADC: Storing the three gain and three offset values ​​in the LE, logistics data 9 DS-ADC: Calculation of gain and offset via DFT 10 DS-ADC: Storing the three gain and three offset values ​​in the LE, logistics data

[0063] Further features, effects, and advantages of the invention will become apparent from the following description of a preferred embodiment of the invention and the accompanying figures. These figures are shown in a schematic diagram: Fig. 1 an inverter with a measuring arrangement during a strip end test as a block diagram, Fig. 2 abstract detail II from Fig. 1, Fig. 3 the detail III from Fig.1, Fig. 4 the detail IV from Fig. 1, Fig. 5. A flowchart for a [project / initiative] in [location / context] Fig. 1 procedure performed.

[0064] Fig. Figure 1 shows a symbolic diagram of an inverter 2 at the end of a production line 4, which is only symbolically indicated here. This line is used to manufacture the inverter 2. Currently, the inverter 2 is fully assembled and is therefore located at the end of the production line 6, where the manufactured inverter 2 undergoes various final tests, namely a line-end test 8. As part of this line-end test 8, calibration data 10a-c for the inverter 2 and its internal current sensors 20a-c are generated.

[0065] Inverter 2 includes an input 12 for a DC power supply 14, specifically a two-pole input for a high voltage of 400 V. Inverter 2 also includes three output lines 16a-c. On each of the output lines 16a-c, Inverter 2, in its inverter mode IB, generates an AC current 18a-c, representing one phase U, V, and W of a three-phase current. The nominal current of each AC current 18a-c is 900 A. Inverter 2 generates the AC currents 18a-c at a preset frequency fs of 100 Hz.

[0066] Each of the output lines 16a-c is equipped with an internal current sensor 20a-c. This sensor is used, or rather configured, to detect / measure one of the alternating currents 18a-c. Fig.1 The current sensors 20a-c are symbolically and distributed, namely by a respective point of attack on the respective output line 16a-c and a first analog-to-digital converter 22a (ADC), here a V-ADC, as well as a redundant or additional second analog-to-digital converter 22b, here a DS-ADC.

[0067] Each of the internal current sensors 20a-c is assigned corresponding calibration data 10a-c. This data is stored in a logistics data area 26 of inverter 2. Each of the respective first ADU 22a and second ADU 22b is assigned a slope 28 and a zero offset 30. In total, there are twelve individual calibration values. The assignment is shown in Fig. 1 explained by symbolic arrows and symbolically represented in the manner of a table.

[0068] The slope 28 and zero offset 30 serve to individually adjust or calibrate the current measurement characteristic of the respective first ADU 22a and second ADU 22b of the respective internal current sensor 20a-c with respect to slope / incline / gain and zero offset. After calibration, the measurement data of the internal current sensors 20a-c correspond to the actual current values ​​of the alternating currents 18a-c.

[0069] Within the framework of a procedure for generating the calibration data 10a-c explained below, the calibration data 10a-c in the form of all slopes 20 and zero offsets 30 for the internal current sensors 20a-c or their first ADUs 22a and second ADUs 22b are generated or determined.

[0070] In order to generate the calibration data 10a-c, the respective output line 16a-c of the inverter 2 is routed, extended, or connected to a choke assembly 32 as part of the end-of-line test 8. An external current sensor 34a-c, in this case a reference current transformer, is arranged in or on each of the (extended) output lines 16a-c.

[0071] Each of the external sensors 34a-c is configured to record, measure, or determine the respective external measured values ​​42a-c of the alternating currents 18a-c. In this case, it records and stores the respective external profile 36a-c of the respective alternating current 18a-c over time. This is done by measuring and storing digital samples of the respective alternating current 18a-c. The respective external profile 36a-c has a respective external duration 38a-c. The latter corresponds to the time span over which the temporal profile of the alternating current 18a-c is recorded, or in this case, digitally sampled and stored. The result is therefore digital samples 40 of the respective alternating current 18a-c. It is assumed that the external current sensors 34a-c are calibrated with high precision or represent reference current sensors and that the external measured values ​​42a-c are therefore highly accurate or technically exact.

[0072] The internal current sensors 20a-c also detect the same alternating currents 18a-c as the external current sensors 34a-c, whether simultaneously or at different times.

[0073] In the same manner as explained above, the internal current sensors 20a-c also record internal measured values ​​44a-c of the alternating currents 18a-c. These are also digitally measured, recorded, and stored here as internal waveforms 46a-c of their respective internal durations 48a-c as sampled values ​​40.

[0074] However, the internal measurements 44a-c were taken using the internal current sensors 20a-c, which may still be uncalibrated, and could therefore contain errors. Both analog-to-digital converters 22a and 22b each provide internal measurements 44a-c separately.

[0075] The determination of the calibration data 10a-c, or a corresponding procedure for generating the calibration data 10a-c, is carried out using a measuring arrangement 50. This arrangement comprises the external current sensors 34a-c and a computing unit 52. The internal measured values ​​44a-c, or digital samples 40, as internal profiles 46a-c, are (temporarily) stored in internal memories of the inverter 2, which are not further described and are each symbolically indicated by arrows, assigned to the first ADU 22a and the second ADU 42b.

[0076] The computing unit 52 is connected to the inverter 2 for the exchange of all necessary data / sizes / information (especially internal measurement values ​​44a-c, calibration data 10a-c) via an interface 54, here a CAN bus.

[0077] The generation of calibration data 10a-c is carried out only according to the following procedure, which is executed in the computing unit 52: Fig.Figure 5 shows a flowchart of the procedure for generating the calibration data 10a-c for the internal current sensors 20a-c of the inverter 2. The procedure is carried out as part of the end-of-line test 8 on the inverter 2 during its manufacture.

[0078] In step S1 (and also unchanged in steps S2 and S3), the inverter is operated in inverter mode IB or – if it is not already in this mode – switched to it. In this inverter mode IB, inverter 2 generates the alternating currents 18a-c in the output lines 16a-c from the DC power 14. In steps S1 to S3, inverter 2 generates the alternating currents 18a-c in inverter mode IB with the frequency fs.

[0079] In step S2, external current measurements 42a-c of the alternating currents 18a-c flowing through the internal current sensors 20a-c are taken outside the inverter 2 using the external current sensors 34a-c. The external current measurements 42a-c are each recorded by the external current sensors 34a-c as time-dependent external profiles 36a-c in the form of digital samples 40 over external durations 38a-c.

[0080] In step S3, internal measurements 44a-c of the same alternating currents 18a-c are taken within inverter 2 using the still uncalibrated internal current sensors 20a-c. The internal measurements 44a-c are each recorded by the internal current sensors 20a-c as time-dependent internal profiles 46a-c in the form of digital samples 40 over internal durations 48a-c.

[0081] The internal durations 48a-c and the external durations 38a-c are each chosen to be of equal length and such that they are even multiples of the period Ts=1 / fs of the alternating currents 18a-c. The sampling intervals Ta=1 / fa (associated sampling frequency fa) between two samples 40 and the period of the alternating current (associated frequency fs) are chosen such that the period is an even multiple of the sampling intervals 40 of the internal waveform 46a-c and the external waveform 36a-c.

[0082] In step S4, the internal measured values ​​44a-c from inverter 2 are transmitted to the processing unit 52 via interface 54. The calibration data 10a-c are determined in this unit, i.e., outside of inverter 2. The external measured values ​​42a-c are transmitted from the external sensors 34a-c to the processing unit 52 via further interfaces 54, in this case data interfaces.

[0083] The calibration data 10a-c for the internal current sensors 20a-c (for both the first ADU 22a and the second ADU 22b) are generated based on the measured internal values ​​44a-c and external values ​​42a-c. The slope 28 and the zero offset 30 are generated as calibration data 10a-c for each of the internal current sensors 20a-c and for each of the ADUs 22a and 22b. An internal DC component 72a-c of the AC current 18a-c is determined from the internal values ​​44a-c, and the zero offset 30 is determined based on the internal DC component 72a-c. Furthermore, an internal amplitude 74a-c is determined from each of the internal measurements 44a-c and an external amplitude 78a-c of the alternating current is determined from each of the external measurements 42a-c, and the slope 28 is determined based on the determined internal amplitudes 74a-c and external amplitudes 78a-c.

[0084] From the external measured values ​​42a-c, an external DC component 76a-c of the alternating currents 18a-c is also determined, and a fault condition 82 of the inverter 2 is determined based on the determined external DC components 76a-c.

[0085] Using a frequency transformation, here a discrete Fourier transform (DFT), the internal amplitudes 74a-c in the frequency domain are determined from the temporal internal responses 46a-c and the external amplitudes 78a-c in the frequency domain are determined from the temporal external responses 36a-c.

[0086] The calibration data 10a-c are transferred from the computing unit 52 to the inverter 2 via the interface 54 after their generation.

[0087] Details of the procedure will now be provided based on the Fig. 1-4 or details II, III and IV from Fig. 1 explained.

[0088] Fig. 2 shows detail II from Fig.1, namely symbolically the basic determination of a DC component 60 (representing one of the internal DC components 72a-c or external DC components 76a-c) and an amplitude 62 (representing one of the internal amplitudes 74a-c or external amplitudes 78a-c) of one of the alternating currents 18a-c.

[0089] Each time waveform 64 (representing one of the internal waveforms 46a-c or external waveforms 36a-c) is supplied as a list of sampled values ​​40 to a DC component module 66 and a DFT module 68. Furthermore, the frequency fs of the alternating currents 18a-c, the respective sampling frequency fa at which the time waveform 64 was sampled, and the number P of periods of the alternating current 18a-c contained in the time waveform 64 are supplied to a control module 70. The control module 70 determines unspecified DFT parameters for the DFT module 68 and the DFT performed by it. The control module 70 specifies which harmonic m of the DFT is to be determined and calculates the number of iterations D for the corresponding determinations.

[0090] Fig. 3 and Fig. Figures 4 show in details III and IV that according to the basic principle Fig.2 for first and second ADU 22a,b, the calculations of DC component 60 and amplitude 62 for the external measurements 42a-c and the internal measurements 44a-c are determined as time courses 64.

[0091] In Fig. 3. The DC component 60 is derived from the external measurements 42a-c to form an external DC component 76a-c of the respective AC current 18a-c. In a fault-free inverter 2, this external DC component 76a-c must always be zero (low threshold / noise, etc.). By comparing the actually measured external DC component 76a-c with a maximum threshold (not shown), a fault condition 82 of the inverter 2 can thus be determined.

[0092] The amplitudes 62, each determined as interamplitudes 74a-c based on the internal measurements 44a-c and as external amplitudes 78a-c based on the external measurements 42a-c, are combined in a manner not further explained in order to determine, as part of the calibration data 10a-c, the respective slope for the respective ADU 22a of the respective internal current sensor 20a-c. The respective DC components 60 based on the internal measurements 44a-c are internal DC components 72 ac and directly represent the respective zero offset 30 for the respective ADU 22a of the respective internal current sensor 20a-c.

[0093] Fig. 4 shows detail IV from Fig. 1, namely the determination of the slopes 28 and zero offsets 30 for the second ADU 22b. The determination is carried out fundamentally in accordance with the principles as they are for Fig. 3 were explained. One difference is that the second ADUs 22b each have a CIC parameter 80.

[0094] This is multiplied by the respective DC components 60 (internal DC components 72a-c) and amplitudes 62 (internal amplitudes 74a-c), which are based on the internal measurements 44a-c. This directly yields the respective zero offset 30. The result of the multiplication with the amplitudes 62 (external amplitudes 78a-c) based on the external measurements 42a-c is then calculated.

[0095] In order to determine all calibration data 10a-c in parallel, the data in the Fig. 3 and Fig. The arrangement shown in 4 is present three times each in the computing unit 52. Reference sign 2 Inverter 4 production line 6 Band end 8 Band end test 10a-c Calibration data 12 Entrance 14 DC power 16a-c output line 18a-c alternating current 20a-c Internal current sensor 22a,b Analog-to-Digital Converter 26 Logistics Data Area 28 incline 30 Zero offset 32 Throttle arrangement 34a-c External current sensor 36a-c External route 38a-c External duration 40 sample value 42a-c External measurement value 44a-c Internal measurement value 46a-c Internal course 48a-c Internal duration 50 Measuring setup 52 computing units 54 Interface 60 equal share 62 Amplitude 64 Time course 66 DC component module 68 DFT module 70 Control module 72a-c Internal Equalization 74a-c Internal amplitude 76a-c External Equalization Component 78a-c External amplitude 80 CIC parameters 82 Error condition fs frequency Ts period duration f sampling frequency Ta sampling interval P Number (periods alternating current) m Harmonic D runs (number) IB Inverter operation S1-4 Step

Claims

[1] Method for generating calibration data (10a-c) for an internal current sensor (20a-c) of an inverter (2), wherein the inverter (2) has an input (12) for a direct electrical power (14) and at least one output line (16a-c) for an alternating electrical current (18a-c) generated from the direct electrical power (14), wherein the internal current sensor (20a-c) is arranged in the output line (16a-c) for detecting the alternating current (18a-c), in which: - the inverter (2) is operated in an inverter mode (IB) in which it generates the alternating current (18a-c) in the output line (16a-c) from the direct power (14), - external measured values ​​(42a-c) of the alternating current (18a-c) flowing through the internal current sensor (20a-c) are measured outside the inverter (2) using an external current sensor (34a-c), - internal measured values ​​(44a-c) of the alternating current (18a-c) are measured within the inverter (2) using the still uncalibrated internal current sensor (20a-c), - the calibration data (10a-c) for the internal current sensor (20a-c) are generated based on the measured internal values ​​(44a-c) and external values ​​(42a-c). [2] Method according to claim 1, characterized by that the inverter (2) generates the alternating current (18a-c) in inverter operation (IB) with a predeterminable frequency (fs). [3] Method according to one of the preceding claims, characterized by that a gradient (28) and / or a zero offset (30) for the internal current sensor (20a-c) is generated as calibration data (10a-c). [4] Method according to claim 3, characterized bythat an internal DC component (72a-c) of the alternating current (18a-c) is determined from the internal measured values ​​(44a-c) and the zero offset (30) is determined based on the determined internal DC component (72a-c), and / or an internal amplitude (74a-c) is determined from the internal measured values ​​(44a-c) and an external amplitude (78a-c) of the alternating current (18a-c) is determined from the external measured values ​​(42a-c) and the gradient (28) is determined based on the determined internal amplitude (74a-c) and external amplitude (78a-c). [5] Method according to one of the preceding claims, characterized by , that - an external DC component (76a-c) of the alternating current (18a-c) is determined from the external measured values ​​(42a-c), and - an error state (82) of the inverter (2) is determined based on the determined external DC component (76a-c). [6] Method according to one of the preceding claims, characterized by , that - the internal measured values ​​(44a-c) from the internal current sensor (20a-c) are measured as a temporal internal profile (46a-c) of an internal duration (48a-c) and / or the external measured values ​​(42a-c) from the external current sensor (34a-c) are measured as a temporal external profile (36a-c) of an external duration (38a-c), in each case of the alternating current (18a-c), - with the aid of a frequency transformation, the internal amplitude (74a-c) in the frequency domain is determined from the temporal internal curve (46a-c) and / or the external amplitude (78a-c) in the frequency domain is determined from the temporal external curve (36a-c). [7] Method according to claim 6, characterized by that the internal duration (48a-c) and the external duration (38a-c) are chosen to be the same length. [8] Method according to one of claims 6 to 7, characterized by that the internal duration (48a-c) and / or the external duration (38a-c) is selected such that it is an even multiple of a period duration (Ts) of the alternating current (18a-c). [9] Method according to one of claims 6 to 8, characterized by that the internal curve (46a-c) and / or the external curve (36a-c) are measured in the form of digital samples (40) of the alternating current (18a-c). [10] Method according to claim 9, characterized by that sampling intervals (Ta) between two sampling values ​​(40) and / or the period duration (Ts) of the alternating current (18a-c) are selected such that the period duration (Ts) is an even multiple of the sampling intervals (Ta) of the internal curve (46a-c) and / or the external curve (36a-c). [11] Method according to one of claims 6 to 10, characterized by that the internal amplitude (74a-c) is determined from the internal curve (46a-c) and / or the external amplitude (78a-c) from the external curve (36a-c) using a discrete Fourier transformation as a frequency transformation. [12] Method according to one of the preceding claims, characterized by , that - the internal measured values ​​(44a-c) are transferred from the inverter (2) and - the calibration data (10a-c) are generated outside the inverter (2) and - the calibration data (10a-c) are transferred into the inverter (2) after they have been generated. [13] Method according to one of the preceding claims, characterized by that the method is carried out as part of an end-of-line test (8) on the inverter (2) during its manufacture. [14] Computing unit (52), - which is designed to carry out the method according to one of the preceding claims, - with a respective interface (54) for receiving the internal measured values ​​(44a-c) and the external measured values ​​(42a-c) and for outputting the calibration data (10a-c). [15] Measuring arrangement (50), with the computing unit (52) according to claim 14 and with the external current sensors (34a-c).

Images