Disturbance correction current detection in inverters

Abstract

A method for detecting current flowing through an inverter (IN1, IN1′) or between the inverter (IN1, IN1′) and a multiphase motor (EM) connected thereto. The method involves determining the current (I1–I3, I1′–I3′) flowing in the half-bridge (H1–H3, H1′–H3′) of the inverter (IN1, IN1′), which provides the current as a pulse-width modulated phase current to the motor (EM). This determination can be performed by measuring at least one current (I1–I3, I1′–I3′) at the end of a sampling period (E) and by estimating the current value. The estimation is made for currents that cannot be measured with a specified minimum accuracy due to interference during the sampling period (E). The estimation is based on previously determined current values. The current is determined by estimation when the current (I1–I3, I1′–I3′) flowing in the half-bridge under test (H1) at the end of the acquisition period (E) is less than a predetermined decay time (tA) from a switching edge (SF) of the other half-bridge (H1–H3, H1′–H3′) of the inverter (IN1, IN1′). Otherwise, it is determined by measuring the current in the half-bridge under test (H1). A method for controlling an electric traction drive and a current determination device are also described.

Description

Technical Field

[0001] In an electrically driven vehicle, a traction battery is provided to operate an electric drive system. This drive system includes an inverter that generates current through the rotating magnetic field of a motor within the drive system. Specifically, this current is generated via pulse width modulation (PWM) for controlling the drive system. Background Technology

[0002] Controlling the drive requires knowledge of the current phase currents, i.e., the current flowing in the motor. These currents are directly derived from the current flowing in the inverter, particularly the current flowing in the low-side section of the inverter's half-bridge, which, due to its low potential, is suitable as a location for current measurement using a shunt. Given the high switching speed, it must be considered that the measurement is not performed at a single moment, but within a measurement interval where the current should not fluctuate significantly. In particular, interference signals should not occur during current measurement; that is, no interference signals should fall into the measurement interval, otherwise, interference will lead to inaccuracies in current detection. Summary of the Invention

[0003] Therefore, the task at hand is to explain a scheme that enables high-precision current detection in electric drives.

[0004] This task is addressed through a separate topic. Further details regarding its nature, characteristics, implementation, and advantages are provided in the appendices, specifications, and accompanying drawings.

[0005] It is recommended to consider interference (interference voltage) generated by the switching edges in phase current sensing. Interference voltage occurs due to the switching edges generated during pulse width modulation to produce the phase current. It has been recognized that induced and / or capacitive coupling between different motor windings or phases can be considered to improve accuracy or reduce distortion in current sensing.

[0006] In particular, it has been recognized that crosstalk exists between the phases of the motor (and therefore between the inverter half-bridges connected to it), where interference voltages can be transmitted from one half-bridge or phase to the other half-bridge or phase due to the coupling effect of the neutral point. Measurements in one half-bridge, if superimposed with interference signals due to the coupling of interference signals from the other half-bridge into the half-bridge under test, will be suppressed or not output, and especially not used as actual values ​​in control (control of the motor).

[0007] When current is detected using a Kalman filter, measurements or current values ​​that fall within the interference voltage time interval (generated in another phase, for example, due to a switching edge) are evaluated as erroneous measurements. These current values ​​are subsequently—unlike valid measurements—not used by the Kalman filter to correct estimates within the filter. The time interval during which the interference voltage occurs, and the time interval during which measurements should not be taken in other phases, are derived from the modulation parameters (especially the duty cycle) and the resulting switching moments.

[0008] All phases or half-bridges except one can be equipped with current sensors. The current in the remaining phases is determined using the nodal law (KCL), which is the difference between the total current (e.g., measured at the DC voltage input of the inverter) and the sum of the currents measured in each phase. Generally, current can be detected by measuring the current, for example using a shunt resistor, a Hall sensor, or another sensor, and also by calculation, for example by calculation based on the difference (nodal law (KCL)), by calculating the current value in one winding system (or in the corresponding inverter) based on the current value in another winding system (or in the corresponding inverter), or by applying a Kalman filter. The latter method is particularly useful for detecting a current value that cannot be measured with sufficient accuracy (without interference) due to an interference signal; in this case, the Kalman filter outputs the current value. The motor control uses measured (using current sensors), calculated (based on current in another winding system / another inverter), determined (based on the nodal law (KCL)), or detected (by estimation, e.g., using a Kalman filter) current values ​​as inputs (actual quantities). The motor control can also (temporarily) rely solely on current values ​​in a portion of the windings or half-bridge to avoid the influence of disturbances. In the presence of multiple winding systems, the current values ​​of the phases of different winding systems can be averaged as inputs to the motor control. Preferably, measured current values ​​are used primarily, but in the event of disturbances, values ​​from another winding system can be used (adjusted by phase shifts if necessary). If these values ​​are also disturbed, estimated values ​​can be used, and / or the control may not be updated, i.e., the feed of actual values ​​can be suppressed (thus suppressing the undesirable effects of disturbed values ​​or incorrectly estimated values ​​due to disturbances).

[0009] A method for detecting current flowing in an inverter or between an inverter and a multiphase electric motor is described. Since the current flowing in the inverter also flows in the windings of the motor in the same manner or in combination (and also between the inverter and the motor), no distinction is made below between current flowing inside the inverter, current flowing between the inverter and the motor, and current flowing in the motor or its windings.

[0010] Determine the current. In this context, determination is intended as a higher-level concept, including measurement (using current sensors such as shunts), calculation based on the nodal law (KCL) (as a difference), calculation based on the current of another winding system (e.g., through phase shift), and estimation (e.g., using a Kalman filter or another method for estimation or prediction). Determine the current flowing in the half-bridge or generally the phase of the inverter. Since the motor is connected to the inverter, the phase current is also determined. If the motor is configured in a star or delta configuration, the phase current flowing in the motor is derived directly and easily from the current flowing in the individual phases or half-bridges of the inverter. By determining the current flowing between the inverter and the motor, or the current flowing in the inverter, the current flowing in the windings or phases of the motor is also determined. The inverter provides the motor with pulse-width modulated phase currents, for example, by means of a (multiphase) pulse-width modulation (PWM) actuation element for motor control.

[0011] At least one current is measured during an acquisition period (E). Here, to perform the current measurement, a switch in a corresponding half-bridge equipped with a current sensor is closed. Specifically, the switch downstream of the current sensor in the half-bridge is closed. If the current sensor is connected to a negative supply potential, the switch in the half-bridge connected to the negative supply potential via the current sensor is closed. This generates the current to be measured, but this current needs to be amplified. This amplification is accompanied by a settling time, during which the signal received from the current sensor (and amplified) has not yet stabilized (but carries errors due to the settling process). Therefore, it is necessary to wait for the settling process to decay, or until the settling process can be ignored (i.e., until the error / interference caused by the settling process is below a predetermined tolerance threshold). Therefore, measurement is only performed at the end of or after the settling process. This forms an acquisition period that begins with the settling process (when the switch is closed) and allows sampling or measurement at its end. If the effect occurring in the amplifier is considered as an effect typically involved in the current sensor and its signal processing, the acquisition period can be considered as a time interval that begins with switch operation and ends with sampling. Therefore, the acquisition period is characterized by a start (i.e., the start time when the current to be measured begins to flow), a duration defined by the setup characteristics of the current sensor's signal processing, and measurement at or after the end of the setup characteristics.

[0012] The duration of the acquisition period is predetermined and is derived, for example, from the setup time of a measurement / amplifier circuit required to achieve the desired accuracy. Preferably, at least one current, i.e., the current in at least one half-bridge, is measured in each PWM cycle. Preferably, each PWM cycle has one acquisition period. The acquisition period is preferably located at the beginning of each PWM cycle. The current is preferably measured using a shunt resistor that connects a lower (negative side) section of the half-bridge, or a semiconductor switch located therein, to a negative potential of a supply voltage. This section is also called the low-side section and specifically corresponds to so-called pinpoint measurements.

[0013] During or at the end of the acquisition period, the current of at least one phase is measured. The current can also be estimated. Therefore, it is stipulated that the current value cannot be measured with a predetermined minimum accuracy due to interference during the acquisition period (including the end of the acquisition period). This estimation is particularly based on previously determined current values. The interference is particularly caused by a switching edge of pulse width modulation in a phase connected to or coupled to the phase whose current is to be detected. Therefore, the timing of the interference can be determined. Due to ringing effects, the interference persists for a specific duration after the switching edge. This duration corresponds to a decay time (tA), during which the interference significantly affects the measurement (i.e., causes accuracy worse than a minimum accuracy). After the decay time, measurement can be performed with an accuracy at least corresponding to or better than the minimum accuracy.

[0014] The current flowing in the half-bridge under test during or at the end of the acquisition period is estimated when interference from the other half-bridge significantly affects the current measurement at the half-bridge under test. To avoid interference, estimation is performed when the acquisition period occurs less than a predetermined decay time after a switching edge in the other half-bridge of the inverter. With respect to the interference, the other half-bridge is coupled to the half-bridge whose current is determined (by estimation), either through a common neutral point, or through inductive or capacitive coupling. The decay time can vary depending on the coupling strength. If the coupling strength between the interfering half-bridge and the half-bridge to be determined (whose current is to be determined) is low, the decay time is shorter than in cases of stronger coupling. If the acquisition period occurs less than a predetermined decay time after a switching edge in the other half-bridge of the inverter, estimation must be performed; otherwise, the current of the half-bridge under test is determined by measuring the current in that half-bridge, either by measuring the current in that half-bridge or by measuring the current in other half-bridges and subsequently calculating based on the nodal rule (KCL).

[0015] It can be specified that the half-bridge causing the interference and the half-bridge to be determined are located in the same winding system. In other words, it can be specified that the half-bridge to be tested and the other half-bridge are connected to different windings, which are connected via a common neutral point or via a delta configuration. It can also be specified that the half-bridge to be tested and the other half-bridge belong to different winding systems located in the same stator. In this case, the different winding systems and the windings existing in the different winding systems are coupled to each other via the common stator. This coupling is in particular a magnetic coupling, and may also (additionally) have capacitive coupling.

[0016] If a predetermined decay time following the most recent switching edge has not yet ended, the current can only be determined inaccurately due to interference that occurs and is injected into the half-bridge under test during the decay time. Therefore, in this case, a relevant current for the acquisition period can be estimated. This estimation preferably ignores the current of the half-bridge under test, i.e., by measuring the current that should have been obtained but was incorrectly measured due to interference. For this purpose, an estimation algorithm, such as a Kalman filter, can be used to track the current flow of the relevant half-bridge or winding, for example, using a model (a model of the motor and / or the inverter). If the acquisition period is less than a predetermined decay time following the most recent executed switching edge in the half-bridge, this is evaluated as a measurement failure by the estimation algorithm, i.e., a measurement value determined (in whole or in part) during the decay time is not included in the estimation or model-based prediction. Alternatively, the last (valid, i.e., unaffected by interference) measurement value, a moving average, or the result of the estimation algorithm (Kalman filter) for the relevant acquisition period can be used. Therefore, if the sampling period occurs less than a predetermined decay time after the most recently executed switching edge of all switching edges in the half-bridge of the inverter, the current is determined by estimation. Specifically, it is specified that the current is determined by estimation (through a common neutral point, through a delta connection, or through capacitive or inductive coupling) when the sampling period for current detection in the half-bridge / winding occurs less than the predetermined decay time after the most recently executed switching edge of all switching edges in those half-bridges coupled to the half-bridge to be detected. The decay time to be considered may depend on the degree of coupling between the interfering half-bridge (which executed the switching edge) and the half-bridge whose current is to be detected. If a direct connection exists via a star or delta connection, a longer decay time can be considered than if only capacitive or inductive coupling exists. Preferably, estimation is only performed when the current cannot be obtained by applying the node law (KCL) (the sum of all currents = 0) in the winding system, nor by calculation from a second winding system of the same stator (relative to its phase shift) (without significant interference). If two winding systems are configured, they can be staggered by 180°. Generally, if n winding systems are configured, these winding systems can be uniformly staggered by 360° / n. Due to the (rotational) symmetrical current conduction (in order to utilize multiple winding systems to generate a uniform rotating magnetic field), the current in half the bridge or winding of one winding system can be inferred from the current in half the bridge or winding of another winding system. The fundamental relationship stems from rotational symmetry, i.e., the phase shift between the winding systems.If the phase shift is 180° for a two-winding system, then the current in the first phase or winding of the first winding system corresponds to the negative value of the current in the first phase or winding of the second winding system, the current in the second phase or winding of the first winding system corresponds to the negative value of the current in the second phase or winding of the second winding system, and the current in the third phase or winding of the first winding system corresponds to the negative value of the current in the third phase or winding of the second winding system, and so on. This also applies to connected half-bridges. For two half-bridges, this results in a 180° phase shift, which manifests as a change in polarity in the case of a sinusoidal waveform, i.e., by forming a negative current value. When the number of winding systems differs, corresponding factors are obtained. Generally, the phase shift, from which the current value of another winding system differs from its current value by a phase difference of p, is derived from the cosine of p.

[0017] The acquisition period preferably begins with the PWM cycle (the PWM cycle of the winding system or the half-bridge connected to the relevant winding system), i.e., from the closing of the half-bridge switch connected to the current sensor. The acquisition period can be shifted relative to the start of the PWM cycle, for example, to the end of the decay time after the most recent switching edge. The acquisition period can be shifted so that it does not occur before the end of the decay time to be considered for the half-bridge under test. The acquisition period preferably does not fall into a switching edge. If such a shift results in a switching edge falling into the acquisition period, and its interference is transmitted from the switching half-bridge to the half-bridge under test, the measurement is abandoned (or the measurement result is abandoned—especially in the sense of measurement failure of the estimation algorithm). In this case, the relevant current value can be estimated, especially if it cannot be calculated by applying the nodal rule (KCL) or by calculating from the second winding system of the same stator (relative to its phase shift) (without significant interference). If the duty cycle is variable (i.e., when two consecutive PWM cycles have different duty cycles), the timing of the change in the switching edge of the new PWM cycle must be considered.

[0018] The current flowing through the half-bridge under test during or at the end of the acquisition period is determined by measuring the current flowing through the half-bridge under test (and especially the current flowing through the switch connected to the current sensor). This is done under the following conditions: the decay time does not extend into the acquisition period (i.e., the end of the decay time falls into the acquisition period), and the acquisition period cannot be shifted (delayed) so that no switching edge falls into the acquisition period. In this case, measurement can be performed because the decay time does not overlap with the acquisition period, or the acquisition period can be shifted so that neither the decay time after a switching edge of the previous (first) PWM cycle extends into the acquisition period, nor the switching edge of the immediately following (second) PWM cycle falls into the acquisition period.

[0019] If a half-bridge under test does not have a current sensor, or if the measurement of that half-bridge is interfered with, the current value of the half-bridge under test can be determined by applying the nodal rule (KCL) and measuring in the other half-bridges of the winding system. The current of the half-bridge under test can be calculated as the difference between the total current and the residual summation current. The residual summation current is measured as the current flowing through all the other half-bridges (of the winding system), i.e., as the sum of the currents flowing through the other half-bridges / phases. The total current is the current supplied to the winding system, such as a supply current of the inverter. This total current can also be measured to determine the residual summation current, i.e., the sum of the (measured) currents of the remaining half-bridges. For the measurement of the currents in the other half-bridges and / or the total current, it is equally applicable that the measurement is performed during a sampling period (or at the end of a sampling period), wherein the last switching edge that caused the interference in one of the measurements should be at least earlier than the decay time (before the start of the sampling period).

[0020] As mentioned above, if the measurement is subject to interference, the current can also be used as the current value for estimation. Based on a previously determined current value, the current value belonging to the current can be estimated. A model characterizing the current flow in the inverter and the motor connected thereto can be used in the estimation. The estimation can be performed using a Kalman filter that takes the model into account. Alternatively, estimation can be performed using a (moving) average, or by assuming that a previously measured current value corresponds to the current value to be determined, or by extrapolating previous current values ​​to estimate the current value to be determined.

[0021] As described above, preferably, only undisturbed measurements are used for estimation, and currents affected by interference are ignored or suppressed. When estimating current values, previously (undisturbed) current values ​​are preferably used, and current values ​​that cannot be measured with a predetermined minimum accuracy due to interference during the acquisition period are excluded. This is particularly applicable to interference that arises and is injected into the half-bridge under test due to the switching edges of the PWM control of the half-bridge. For this purpose, it is known that the decay time during which the measurement is interfered with is relevant, and that this decay time begins from the moment the switching edge occurs. The moment the switching edge occurs is determined based on the duty cycle of the PWM phase current, and particularly based on the duty cycle. When shifting / delaying the acquisition period (so that the decay time does not extend into the acquisition period, thus introducing interference), it is preferably ensured that subsequent switching edges do not fall into the acquisition period. Therefore, the acquisition period—if possible—is not so drastically delayed or shifted that the switching edge of the current PWM cycle falls into the acquisition period, and especially does not coincide with the end of the acquisition period in which the measurement is performed. If the time between the last switching edge of the preceding PWM cycle and the first switching edge of the current PWM cycle is less than the minimum duration of the acquisition period, then no measurement is performed for that acquisition period. Instead, it is estimated, for example, or determined based on measurements in another winding system of a phase shift or in the half-bridge connected thereto. Since the occurrence time of the first switching edge of the current PWM cycle is relevant here, the acquisition period is also set taking into account the duty cycle of the current PWM cycle. This duty cycle can deviate from the duty cycle of the previous PWM cycle when it changes. Preferably, the occurrence time of the most recently occurring switching edge is determined based on the largest duty cycle in the PWM phase current. In particular, to shift the acquisition period or to determine whether a measurement can be performed or if the current value must be determined in other ways (estimated, calculated, etc.), the largest duty cycle is used from all half-bridges connected to the same winding system. This avoids interference from the last switching edge of the other half-bridge (the half-bridge with the highest duty cycle) in the measurement of the half-bridge under test, which, due to its lower duty cycle, would have made the acquisition period possible after the decay time of its last switching edge (without considering other half-bridges).

[0022] The method described herein can be used to operate an electric traction drive for a vehicle. A corresponding method for controlling the electric traction drive specifies that it includes a motor and an inverter, the inverter providing the motor with pulse-width modulated phase currents (generally: a PWM drive signal). The motor is multiphase and has one or more winding systems. The control (as a method) specifies space vector control of the motor. This control is implemented using the inverter. If the motor has only one winding system, the inverter can be a single unit. If it is necessary to drive multiple winding systems of the motor, the inverter can be considered a multi-part inverter because it has multiple half-bridge groups, each connected to and operating a specific winding system. The pulse-width modulated phase currents are generated according to the control. The control receives the phase currents as feedback or input quantities (actual quantities). These phase currents are detected according to the method described herein, and are derived, in particular, from the detected currents as described herein. The detected currents can thus undergo space vector transformation to be processed in the control as the space vector transformed quantity.

[0023] Furthermore, a current determination device is described, configured to perform the methods described herein. Specifically, the current determination device is configured to detect current as described herein, particularly by measurement when there is no significant interference due to switching edges, and by calculation and / or estimation as described herein. The current determination device may have an input configured to receive a signal characterizing the current. The input may be a digital input configured to receive a digital signal, or an analog input configured to receive an analog signal; in the latter case, an analog-to-digital converter is connected downstream of the input to convert the signal into digital form. The input may be multi-part, i.e., it may physically have multiple input connections or logically have multiple channels for receiving signals. The current determination device has an output configured to output the determined current, particularly as a signal characterizing the current. The output terminal can be configured as a physical output terminal, but is preferably configured as a storage area that stores values ​​characterizing the current, and these values ​​can be retrieved again from the storage area for further processing, such as to implement a control, for example, as an actual quantity for control. Therefore, the current determination device can be part of a data processing device or control device that performs further functions, preferably based on the current values ​​provided by the current determination device. The control device or the current determination device can be implemented by a programmable circuit, on which software code implements the functionality of the method described herein (running on a processor). Software code that executes the method described herein while running on a processor also implements the method for current determination described herein. Attached Figure Description

[0024] Figure 1 , Figure 2a and Figure 2b Exemplary embodiments used to explain the methods and apparatus described herein. Detailed Implementation

[0025] Figure 1A motor EM with two winding systems WS and WS′ is shown. Each winding system has three windings, labeled 1, 2, 3 and 1′, 2′, and 3′ respectively. Therefore, the motor EM has two three-phase winding systems. Each three-phase winding system WS and WS′ comprises three windings connected to each other via a common neutral point. Thus, each winding system WS and WS′ has three phase terminals, where the phase terminals of winding system WS are labeled U, V, and W, and the phase terminals of the second winding system WS′ are labeled U′, V′, and W′. The first winding system WS is connected to the first inverter IN1 via these phase terminals U, V, and W. The second winding system WS is connected to the second inverter IN1′ via phase terminals U′, V′, and W′. The two inverters IN1 and IN1′ can also be considered as part of a common inverter, especially since they are supplied with the same voltage from the battery AK (at high voltage potentials (positive / negative) HV+ and HV). Power supply between (between).

[0026] Each section or each inverter IN1, IN1′ has a half-bridge H1–H3 or H1′–H3′ for each phase (of the motor). Each half-bridge has two switches 11–23 connected in series, wherein the connection points of each switch are used as the connection points of the phase terminals U, V, W (U′, V′, W′). The first half-bridge H1 is formed by the series connection of switches 11 and 21, the second half-bridge H2 is formed by the series connection of switches 12 and 22, and the half-bridge H3 is formed by the series connection of switches 13 and 23. The half-bridges H1′ to H3′ of the second inverter IN1′ are also formed in the same manner, and are correspondingly connected to the phases U′, V′, W′ of the second winding system WS′ through the connection points within the half-bridges.

[0027] To detect the phase current, where current IU is schematically plotted as the phase U current of the first winding system WS, currents I1, I2, and I3 are detected in half-bridges H1, H2, and H3. The same applies to half-bridges H1′, H2′, and H3′, in which currents I1′, I2′, and I3′ flow. The corresponding currents in each half-bridge H1–H3 and H1′–H3′ are plotted at the pin points, i.e., at a negative potential HV. The position is between the subsequent low-side switches, such as 21, 22, or 23 of the first inverter IN1 or the corresponding low-side switch LS of the second inverter IN1′. The current sensor for detection is preferably located at the current detection location, i.e., at the location indicated by the reference numerals used for current. Based on the currents I1 to I3 and I1′ to I3′ detected at the pin points, the phase current IU can be determined, or the current in windings 1–3 and 1′–3′ can be derived.

[0028] The half-bridges of the inverters IN1 and IN1' are controlled by a signal generated via pulse width modulation (PWM). For this purpose, the half-bridge switches have control inputs symbolically indicated by arrows. It can be seen that, due to the neutral point connection via the connected windings 1–3, as occurs in PWM, a switching edge can be transmitted from one half-bridge H1 to the other half-bridge H2 or H3. Therefore, if half-bridge H1 generates a switching edge, the resulting interference will be transmitted to half-bridges H2 and H3 connected to half-bridge H1 via the relevant winding system. If half-bridge H2 or H3 is the half-bridge under test, it will potentially be affected by the switching edge in half-bridge H1. This will be further examined below to illustrate this point in more detail. Figure 2a and Figure 2b .

[0029] Figure 2a This shows a portion of a PWM signal, where the previous PWM cycle A(k) is visible on the left. 1) A switch begins at point SF, and the subsequent PWM cycle A(k) starts at point PB. Point PB indicates the start of the current or subsequent PWM cycle A(k). Point PB also corresponds to the previous PWM cycle A(k). The end of period 1), especially assuming the current period directly follows the previous period. It can be seen that the switch edge SF is farther from time PB than the switch edge of the immediately following period A(k) is farther from the start point PB. This is due to the variable duty cycle, where the previous period A(k)... 1) has a duty cycle that is less than the current or immediately following duty cycle of period A(k).

[0030] exist Figure 2a The diagram also symbolically illustrates an interference ST generated by the switch along SF. This interference originates in the half-bridge of the first switching operation or in the winding connected thereto, and is transmitted via the neutral point, via inductive and / or capacitive coupling, to other windings in the same winding system or half-bridges connected to the same winding system. Therefore, it should be ensured that interference from the first switching half-bridge does not cause erroneous or interfered measurements in the other half-bridge. According to the method described herein, it is therefore necessary to consider whether the switching half-bridge coupled via the winding system will interfere with the measurements of the other half-bridge.

[0031] Figure 2b For the two phases U and V, the previous period A is shown. U (k 1) A V (k 1) and the immediately following or current cycle A U (k), A VThe PWM signals of (k). A switch in the upper PWM signal (phase U) will generate an interference along SF1 (see...). Figure 2a This interference will cause subsequent acquisition periods (E) of the half-bridge to be affected, and the measured current value will not achieve minimum accuracy. If the switching edge SF1 occurs only in the half-bridge and the half-bridge connected to or coupled to it, the interference will become irrelevant or attenuated after a decay time (tA), and measurements can be performed in the half-bridge or the other half-bridge with minimum accuracy. In this case, the decay time (tA) is calculated from the start of the interference accompanying the switching edge SF1.

[0032] from Figure 2b As can be seen, the switching in the second phase V occurs later along SF2. Therefore, for the decay time (tA) shown in the figure, interference occurs not only in phase V (shown below) but also in the first phase U (shown above) because the interference is transmitted through coupling between the half-bridge or windings of the winding system.

[0033] To measure the current, a sampling period (E) is established during which the current is detected, particularly at its end. During this sampling period (E), for example, an amplifier may build up, therefore measurements cannot be taken during this build-up period, but only at or after the end of the build-up. It is stipulated that the earliest end of the sampling period (E) (i.e., the earliest time used for measurement) is after the end of any interference or after the end of a decay time (tA). The decay time (tA) represents the duration during which an interference can disrupt the measurement (either at the same half-bridge or at the other half-bridge) to such an extent that minimum accuracy cannot be guaranteed.

[0034] exist Figure 2b China and in Figure 2aIn the diagram, a cross indicates a possible moment for current measurement. However, it should be considered that a decay time (tA) must be observed from either the last interference or the last switching edge, whereby a measurement can only be performed at the end of the acquisition period (E) when the decay time (tA) has completely ended. Otherwise, a measurement at the end of the acquisition period (E) will carry interference that persists throughout the entire decay time (tA), and this interference is only considered to have ended thereafter. According to the method described herein, not only the decay time and the switching edge present in one phase are considered, but also the switching edge in one of the switching half-bridges is considered when measuring current in the other half-bridge and determining the acquisition period (E). This avoids the situation where, although a sufficient interval (decay time) is maintained between the current and the last switching edge in the same phase to enable measurement with minimum accuracy, the measurement result can still be distorted by a later switching edge in another phase, whose interference is transmitted to the half-bridge being measured via coupling. Figure 2b This means that even when measuring in phase U (above), the earlier edge SF1 should be considered, but rather the later edge SF2. The acquisition period (E) of the upper phase is defined based on the later edge SF2, where the decay time (tA) begins with the later edge, and the end of the decay time defines the earliest time at which measurements can begin at the end of the acquisition period (E).

[0035] However, it should also be considered that during such a shifted or delayed acquisition period (E), a switching edge of the same phase or another phase or half-bridge may occur, appearing in the immediately following PWM cycle. Since this second switching edge in the immediately following or current PWM cycle (A(k)) will interfere with the measurement at the end of the acquisition period (E), measurements are only performed during the acquisition period (E) if its start is after the end of the decay time and the end of the acquisition period (E) is before the earlier of the switching edges (SF3, SF4) of the immediately following PWM cycle. Here, it is necessary to consider defining a variable duty cycle for the switching edge timing. Furthermore, the inverter's dead time must also be considered if necessary. (See the diagram...) Figure 2b In this context, edge SF4 is the earlier of the two edges SF3 and SF4 that immediately follow the PWM cycle k. Therefore, the acquisition period (E) should end before the occurrence of the switching edge SF4. Since the acquisition period (E) itself has a minimum duration and measurements can only be taken at the end of that minimum duration, it may not be possible to perform interference-free measurements with established current sensor data processing. Consequently, the current will be estimated or otherwise determined instead of measured.

[0036] In summary, the previous PWM cycle k The later of the two switching edges SF1 and SF2, SF2, marks the beginning of the decay time (tA) for all measurements in all half-bridges coupled to the half-bridge executing that later switching edge SF2. This coupling can be achieved, in particular, through a delta connection or a neutral point connection of the windings 1–3 and 1′–3′ connected to it. The end of the decay time (tA) defines the earliest possible start of the acquisition period (E) for all half-bridges connected to the half-bridge executing the switching edge SF2 (= the half-bridge causing the interference). Measurements at the end of the acquisition period (E) are only passed on as (valid) current values ​​if the earliest of all the immediately following switching edges SF3 and SF4 occurs after the acquisition period (E). In other words, if the measurement time does not occur at least in the previous PWM cycle k... After the decay time following the last switching edge of 1 (the last switching edge of all switching edges of all coupled half-bridges), and if the end of the acquisition period (E) is before the earliest of all switching edges of all coupled half-bridges in the current PWM period k, the measurement in the acquisition period (E) is not performed or is discarded.

[0037] Half-bridges are coupled when they are connected to each other via a delta circuit, a neutral point circuit, capacitive coupling, or inductive coupling through their connecting windings. In particular, they are coupled when they belong to the same winding system or are connected to each other via a delta circuit or a neutral point circuit. The capacitive or inductive coupling can be generated by the delta circuit or the neutral point circuit.

[0038] If the measurement is not performed or is discarded, it can be estimated, for example, using a Kalman filter. However, if possible, it is preferable to use the measured current of the corresponding phase of another winding system, for example after changing the polarity of the current (i.e., forming a negative value) and / or after otherwise taking into account the phase shift between the winding systems.

[0039] If this is not possible, for example because in the second winding system, the sampling period also starts too early after the last switching edge (i.e., still during the decay time), or a switching edge of a subsequent cycle will already fall into the sampling period, then the current value can be determined by estimation, for example, through a Kalman filter. When considering the second winding system WS′, the coupling between the half-bridges H1′–H3′ at that point is also considered, i.e., the decay time is relative to the previous PWM cycle k. For the last switching edge among all switching edges of all half-bridges H1′–H3′ of the second inverter, and considering the earliest switching edge among all switching edges of all half-bridges H1′–H3′ of the current period k, or if the earliest switching edge among all half-bridges of the second winding system falls into the acquisition period (E) or coincides with the end of the acquisition period (E) (i.e., the measurement time), then the measurement of the half-bridge of the second inverter IN1′ in the acquisition period (E) will be suppressed or not performed.

Claims

1. A method for detecting current flowing through the interior of an inverter (IN1, IN1′) or between the inverter (IN1, IN1′) and a multiphase motor (EM) connected thereto, comprising: The currents (I1–I3, I1′–I3′) flowing in the half-bridges (H1–H3, H1′–H3′) of the inverter (IN1, IN1′) are determined by measuring at least one current (I1–I3, I1′–I3′) at the end of the acquisition period (E) and estimating the current values ​​of currents that cannot be measured with a specified minimum accuracy at the end of the acquisition period (E) due to interference, based on previously determined current values. Specifically, if the time interval between the end of the acquisition period (E) and the occurrence of a switching edge (SF) in the other half-bridge (H1–H3, H1′–H3′) is less than a predetermined decay time (tA), the current flowing in the half-bridge under test (H1) during the acquisition period (E) is determined by estimation; otherwise, it is determined by measuring the current in the half-bridge under test (H1).

2. The method according to claim 1, wherein, The half-bridge under test and the other half-bridge (H1–H3, H1′–H3′) are respectively connected to different windings (1, 2, 3) that are interconnected via a common neutral point or a delta connection. Alternatively, the half-bridge under test and the other half-bridge (H1–H3, H1′–H3′) belong to different winding systems (WS, WS′) set in the same stator.

3. The method according to claim 1, wherein, The half-bridge under test and the other half-bridge (H1–H3, H1′–H3′) are set in the same winding system (WS), and the current (I1–I3, I1′–I3′) flowing in the half-bridge under test (H1) during the acquisition period (E) is determined by phase shifting the measured current of the half-bridge (H1′) of the other winding system (WS′), wherein the phase shift is performed according to the complementary angle of the phase difference between the two winding systems.

4. The method according to claim 1, 2 or 3, wherein, If the measurement performed at the end of the acquisition period (E) is less than the time interval between the most recently executed switching edge (SF2) of all half-bridges (H1–H3, H1′–H3′) of the inverter (IN1, IN1′) and the measurement is performed at the end of the acquisition period (E), the current is determined by estimation.

5. The method according to any one of the preceding claims, wherein, The current flowing through the half-bridge under test (H1–H3, H1′–H3′) during the acquisition period (E) is determined by measuring the current (I1–I3, I1′–I3′) flowing through the half-bridge under test (H1–H3, H1′–H3′), or by measuring the residual summation current (I2+I3) flowing through all other half-bridges (H2, H3), and calculating the current of the half-bridge under test as the difference between the total current (I1+I2+I3) and the residual summation current (I2+I3).

6. The method according to any one of the preceding claims, wherein, The current values ​​belonging to the current (I1–I3, I1′–I3′) are estimated based on previously determined current values ​​and a model characterizing the current flow in the inverter (IN1, IN1′) and the motor (EM) connected to it.

7. The method according to claim 6, wherein, The current value belonging to the current is estimated using a Kalman filter that takes into account the model.

8. The method according to any one of the preceding claims, wherein, When estimating the current value, the previously (interference-free) current value is used, and those current values ​​that cannot be measured with the specified minimum accuracy due to interference (ST) during the acquisition period (E) are excluded.

9. The method according to any one of the preceding claims, wherein, The occurrence time of the switching edge (SF, SF1, SF2) is determined based on the duty cycle of the pulse width modulation phase current.

10. The method according to claim 9, wherein, The occurrence time of the most recent switching edge (SF, SF1, SF2) is determined based on the maximum duty cycle in the pulse width modulation phase current.

11. A method for controlling an electric traction drive, the electric traction drive comprising a motor (EM) and an inverter (IN1, IN1′), the inverter (IN1, IN1′) being used to provide pulse-width modulated phase currents to the motor (EM), wherein, The control performs space vector control on the motor (EM) via the inverter (IN1, IN1′) and generates the pulse width modulated phase current according to the control. The control is based on the phase current as the actual control value, and the phase current is detected according to the method of claim 1.

12. A current determining device having an input for receiving a signal characterizing a current, the current determining device being configured to perform a method for detecting a current according to any one of claims 1 to 10, and having an output for outputting the determined current.

Images