Method for fault detection of sub-module capacitors of a modular multilevel converter

Abstract

The invention relates to a method for fault detection of a sub-module (1), wherein the sub-module (1) has at least two serially arranged semiconductor switches (11) and a capacitor (12) arranged in parallel to the two serially arranged semiconductor switches (11), wherein by means of the semiconductor switches (11) a zero voltage can be generated at the terminals (13', 13'') of the sub-module (1) in a first switching state. In order to improve the fault detection, it is proposed to detect the voltage (U C ) of the capacitor (12) during the first switching state, wherein by the change in the voltage (U C ) of the capacitor (12) by a factor of two or three of the quantization level of the measurement value acquisition unit (15) for the voltage (U C ) of the capacitor (12) a fault of the capacitor (12) is detected. Furthermore, the invention also relates to a sub-module (1) for a modular multilevel converter (2), wherein the sub-module (1) has at least two serially arranged semiconductor switches (11) and a capacitor (12) arranged in parallel to the two serially arranged semiconductor switches (11); wherein the quantization level is designed such that a fault of the capacitor (12) can be detected when the voltage (U C ) of the capacitor (12) changes by a factor of two or three of the quantization level. The invention also relates to a modular multilevel converter (2) having a plurality of such sub-modules (1).

Description

Technical Field

[0001] This invention relates to a fault detection method for a submodule, wherein the submodule has at least two semiconductor switches arranged in series and a capacitor arranged in parallel with the two series-arranged semiconductor switches, wherein zero voltage can be generated at the terminals of the submodule in a first switching state by means of the semiconductor switches. Furthermore, this invention relates to a submodule for a modular multilevel converter, wherein the submodule has at least two semiconductor switches arranged in series and a capacitor arranged in parallel with the two series-arranged semiconductor switches, wherein zero voltage can be generated at the terminals of the submodule in a first switching state by means of the semiconductor switches. This invention also relates to a modular multilevel converter having a plurality of such submodules. Background Technology

[0002] Modular multilevel converter topologies consist of multiple submodules arranged in series. The most common example is a modular multilevel converter with a half-bridge submodule. In this submodule, each submodule has two semiconductor switches arranged in series. These series-connected semiconductor switches are connected in parallel with a capacitor. The series-connected semiconductor switches are generally not subjected to the same current. Using this submodule, zero voltage or capacitor voltage can be generated at the submodule's terminals. The capacitor can be a single capacitor or can consist of multiple subcapacitors that act as a single (total) capacitor.

[0003] In addition to a switching element that can conduct current in a first current flow direction, a semiconductor switch also has a diode that conducts current in a second current flow direction opposite to the first current flow direction. This diode is also called a freewheeling diode.

[0004] For example, the capacitor or capacitor bank can be designed as a film capacitor. Alternatively, electrolytic capacitors can be considered as an alternative. When such capacitors fail, it is advantageous to reliably detect the fault so that the appropriate fault response of the converter can be initiated. This fault response could, for example, be shutting down the converter or short-circuiting the affected submodule using a suitable switch. Summary of the Invention

[0005] The purpose of this invention is to improve the fault detection of the submodule in terms of capacitor failure.

[0006] This objective is achieved by a fault detection method for a submodule having the features of claim 1. Furthermore, this objective is achieved by a submodule for a modular multilevel converter having the features of claim 4. This objective is also achieved by a modular multilevel converter having multiple such submodules.

[0007] Other advantageous designs of the invention are given in the dependent claims.

[0008] This invention is particularly based on the understanding that faults in a capacitor within a submodule can be detected especially safely and reliably when the voltage of the capacitor (also referred to as the voltage across the capacitor) is evaluated in a first switching state. In the first switching state, zero voltage is applied to the terminals of the submodule. This submodule can be used as a submodule of a modular multilevel converter. For example, the voltage of the capacitor can be measured by a measurement acquisition unit at the capacitor and transmitted to the control unit.

[0009] In this submodule, the first semiconductor switch of the two semiconductor switches is arranged between the terminals of the submodule, and the second semiconductor switch of the two semiconductor switches is arranged between one of the terminals of the submodule and the capacitor. To generate the first switching state, the first semiconductor switch is controlled to be turned on, and the second semiconductor switch is controlled to be turned off. Alternatively, the first switching state can also be achieved when neither semiconductor switch is controlled. This situation is also called pulse blocking of the submodule because no semiconductor switch is in the on state. Depending on the current flow direction, one of the freewheeling diodes is turned on. In this state, the capacitor can only be charged and cannot be discharged due to the current flow direction.

[0010] Therefore, a semiconductor switch has multiple switching states, especially a first switching state (in which the first semiconductor switch is on and the second semiconductor switch is off) and pulse blocking. These switching states enable the first switching state of the submodule. The first switching state is thus formed when the first of the two semiconductor switches is positioned between the terminals of the submodule and is conducting. This may be because the semiconductor switch is actively placed in the conducting state by control, or because a diode arranged in parallel with the switching element of the semiconductor switch in the opposite direction of current flow is conducting.

[0011] In this first switching state, assuming the submodule is fault-free, the current through the capacitor is extremely small. The current is caused solely by, for example, existing equalizing resistors and / or discharge resistors, and by, for example, the power supply of auxiliary components. Examples of auxiliary components include module logic, drive circuitry, or control units. It has been confirmed that this current is virtually independent of the converter's operating state. This current causes a voltage change across the capacitor, which can be easily detected. The capacitor current can be inferred from the change in capacitor voltage. Therefore, if the current caused by the fault current is excessive in the first switching state, a capacitor fault can be reliably detected. This fault can be reliably detected by the change in capacitor voltage.

[0012] Compared to monitoring independent of submodule switching states, the proposed fault detection method allows for the selection of a particularly small observable capacitor voltage change for detection, as this change does not need to account for current generated during operation in other switching states, thus avoiding erroneous responses. This small change value enables reliable fault detection and avoids prolonged current feeding to the fault point. Such feeding, even lasting only a few grid cycles (on the order of approximately 100 ms), can lead to damage such as fire. The proposed method significantly reduces the risk of fire. Depending on the design of the detection threshold, converter fires caused by capacitor faults can also be reliably prevented, thus eliminating the need for corresponding fire detectors. Since the faulty submodule is also reliably detected, the modular multilevel converter composed of such submodules can continue to operate when the relevant submodule is shut down, for example, with the aid of a suitable short-circuit device.

[0013] The proposed fault detection method and apparatus can safely and reliably detect capacitor faults that do not directly cause a significant voltage drop but persist for an extended period, resulting only in voltage variations similar to those observed during normal operation. Alternative methods that do not consider submodule switching states can only detect such faults when the capacitor voltage exhibits a large voltage change (du / dt) exceeding the voltage change observed during normal operation. Therefore, the proposed method prevents prolonged power feeding to the faulty capacitor, thereby significantly reducing or even eliminating subsequent damage from such capacitor faults, such as fires.

[0014] Therefore, the proposed capacitor voltage monitoring reliably prevents energy from being fed into the fault point and limits the damage to the affected capacitor. Feeding less energy into the fault point can prevent or at least significantly reduce the risk of fire.

[0015] Therefore, voltage monitoring can also be applied to submodules in a pulse-blocked state (i.e., a semiconductor switch switching state where both semiconductor switches are off). In this case, considering the current flow to reliably identify the first switching state is advantageous, but not mandatory. Based on the sign of the current flowing through the submodule, either the freewheeling diode of the first semiconductor switch or the freewheeling diode of the second semiconductor switch is on. Accordingly, capacitor voltage changes can only be monitored as described above if the freewheeling diode of the first semiconductor switch conducts current, thus realizing the first switching state of the submodule. Otherwise, current flows through the submodule capacitor and causes a voltage change. However, this voltage change refers to an increase in the capacitor voltage. Typically, the current in all branches of a modular multilevel converter is measured, so the sign of the current flowing through the submodule is known. Here, a branch refers to a series circuit of one or more submodules with an inductor. The proposed method can be applied to other submodule variants besides half-bridge submodules, such as full-bridge modules. Based on its possible switching states, capacitor voltage can be monitored in a switching state where the fault-free capacitor discharges only through self-discharge, leakage current, discharge resistance, etc.

[0016] In this invention, a fault is detected when the capacitor voltage change reaches two or three times the quantization level of the capacitor voltage measurement acquisition unit. During fault-free operation, the voltage change in the first switching state is typically so small that, depending on the quantization of the voltage measurement, it causes an unmeasurable or nearly unmeasurable voltage change between two switching processes of the submodule. However, for the longer-duration fault conditions described above, the change will exceed at least two quantization levels of the voltage acquisition. Therefore, in an advantageous design of the invention, fault detection responds as soon as the change exceeds exactly two quantization levels of the voltage acquisition. Thus, even with a small voltage change, particularly rapid fault detection can be achieved without increasing the accuracy of the voltage acquisition. In this case, fault detection is not only fast but can also be achieved through voltage acquisition designed for operation.

[0017] In particular, it has been proven advantageous to design the quantization level of the measurement acquisition unit in such a way that a capacitor fault can be effectively detected when the capacitor voltage changes by two or three quantization levels in the first switching state. Therefore, a very large quantization level can be selected for the measurement acquisition unit. This makes the measurement acquisition unit particularly economical. Simultaneously, with this economical measurement acquisition unit, faults can be detected reliably and quickly, especially those caused by the unacceptable significant attenuation of the capacitor capacitance. In other words, the minimum allowable capacitance of the capacitor determines the parameter design of the measurement acquisition unit. Therefore, by designing a correspondingly large quantization level, reliable protection against capacitor faults can be achieved with a particularly advantageous measurement acquisition unit.

[0018] In an advantageous embodiment of the invention, capacitor faults are identified by a negative change in capacitor voltage that falls below a preset lower limit. Since the fault manifests as a discharge current in the capacitor, a discharge current exists at the capacitor during a fault, resulting in a voltage drop at the capacitor. Therefore, when performing current measurements, only the measurement value in one current direction needs to be evaluated. Thus, the method is significantly simpler and less computationally intensive, as only one current direction needs to be checked. Furthermore, the method is applicable to both the switching states of semiconductor switches with conducting semiconductor switches and to submodules under pulse blocking, regardless of the specific switching state of the semiconductor switch. The submodule only needs to be in the first switching state.

[0019] The basis of this approach is the monitoring of the capacitor voltage during the first switching state. This involves the first semiconductor switch being on, the second semiconductor switch being off, or the submodule being pulse-blocked. In conventional component designs, the capacitor voltage remains almost constant or changes very little until the next switching state transition. Discharge of the capacitor occurs only through the presence of equalizing resistors and / or discharge resistors, as well as power supply units for auxiliary components (e.g., for module logic and IGBT drivers), the cutoff current of the switching / freewheeling diodes, and the self-discharge of one or more capacitors.

[0020] Slow discharge caused by a capacitor failure will result in a measurable voltage change in the first switching state, which differs significantly from the voltage curve during normal operation in the same switching state. If the submodule is in the first switching state, the magnitude of the negative value of the voltage change (ΔU) or rate of change (du / dt) across the capacitor is compared to a lower limit. The distinction between voltage change and rate of change is not explicitly made below. Both variables are subject to the following limit. If the voltage change (ΔU) or rate of change (du / dt) is less than the preset lower limit (i.e., greater than the predetermined lower limit due to the negative sign), a fault can be reliably detected, triggering the corresponding fault response in the converter. This can be achieved, for example, by shutting down or disabling / bridging the submodule identified as faulty. The magnitude of the limit is advantageously selected based on the capacitor and converter design used. Importantly, this lower limit (also called the threshold) for the voltage change or rate of change across the capacitor is monitored only in the first switching state of the submodule, because the voltage change of the faulty capacitor will reach a value similar to the voltage change during normal operation in another switching state. This distinguishes the described monitoring mechanism from those that do not consider the switching states of submodules and can only react to large voltage changes that are significantly higher than those occurring during normal operation. The threshold can be set as the voltage difference (ΔU) or the voltage U. C The rate of change (du / dt).

[0021] In another advantageous embodiment of the invention, a capacitor fault is detected by the magnitude of the capacitor voltage change exceeding a preset upper limit. If the submodule is in a first switching state, the magnitude of the voltage change is compared to the upper limit. If the magnitude of the voltage change (ΔU) or the rate of change of voltage (du / dt) is greater than the preset upper limit, a fault is reliably detected and a corresponding fault response of the converter is triggered. In this fault detection method, this can also be achieved, for example, by shutting down or disabling / bridging the submodule identified as faulty. The magnitude of the limit is similarly advantageously selected based on the capacitor and converter design used. Importantly, this upper limit (upper threshold) of the voltage change across the capacitor is monitored only in the first switching state of the submodule (where one of the first semiconductor switches is controlled). In a pulse-blocked state, fault detection cannot be reliably performed without assessing the current flow, as capacitor charging is accompanied by a corresponding, expected capacitor voltage change. Attached Figure Description

[0022] The present invention will now be described and illustrated in more detail with reference to the embodiments shown in the accompanying drawings. Wherein: Figure 1 The voltage U of the capacitor in the fault-free submodule is shown. C The curve of change, Figure 2 The voltage U of the capacitor in the faulty submodule is shown. C The curve of change, Figure 3 An embodiment of the submodule is shown, and Figure 4 An embodiment of a modular multilevel converter is shown. Detailed Implementation

[0023] Figure 1 The voltage U applied to capacitor 12 is shown. C The typical variation curve. During operation, especially when submodule 1 is in a state where the capacitor voltage is applied to terminals 13', 13'' of submodule 1, the voltage will increase or decrease. If submodule 1 is in the first switching state, where zero voltage is applied to terminals 13', 13'' of submodule 1, then the voltage U of capacitor 12... C The voltage remains constant or at least approximately constant. Voltage changes are caused by leakage current from the open semiconductor switch 11, self-discharge of capacitor 12, or by resistor 23, particularly the continuously discharging resistor. Such voltage changes caused by small discharge currents are so small over the typical duration of the switching state that they are sometimes undetectable by the measurement acquisition unit 15. This is because the quantization level of the measurement acquisition unit 15 can be larger than the voltage change caused by the discharge current.

[0024] Figure 2 This shows the voltage U applied to capacitor 12 when submodule 1 is defective (specifically, capacitor 12 of submodule 1 is defective). C An example of a change curve. This defect (also known as a fault condition) refers to, for example, a loss of capacitance in capacitor 12 or a decrease in the internal resistance of capacitor 12. Such a resistance change may be caused, for example, by a decrease in the dielectric resistance. In this case, the discharge current will also cause the voltage U of capacitor 12 to decrease in the first switching state. C A larger change occurs. This change can be measured by the measurement acquisition unit 15 of capacitor 12 because, under fault conditions, the change is greater than one quantization level of the measurement acquisition unit 15. Voltage U C Voltage changes occurring with the quantization level of the measurement acquisition unit 15 can be observed, for example, between the time points shown by the dashed lines in the above figure. It can also be seen that the capacitor fault persists for a relatively long time, and the continuing-operating converter thus feeds energy to the faulty capacitor 12. In this case, for example, due to thermal overload, the affected capacitor 12 may cause a fire inside the converter. The proposed method for monitoring capacitor 12 is suitable for fault monitoring, and particularly suitable for converters with multiple sub-modules 1 and corresponding capacitors 12, such as modular multilevel converters 2.

[0025] Figure 3 An embodiment of submodule 1 is shown. A first of two semiconductor switches 11 is arranged between the two terminals 13', 13'' of submodule 1. The second of the two semiconductor switches 11 is arranged between one of the terminals 13' and the capacitor 12. Different switching states of submodule 1 can be achieved using the two semiconductor switches 11. For example, a first switching state where the voltage between terminals 13', 13'' of submodule 1 is zero can be achieved by turning on the first semiconductor switch 11 and turning off the second semiconductor switch 11. The semiconductor switches 11 are controlled by control unit 14, therefore the switching states of the semiconductor switches and the switching states of submodule 1 are known to the control unit. Measurement acquisition unit 15 acquires the voltage U on capacitor 12. C The measured value is sent to control unit 14. Using this information, control unit 14 can implement the proposed method. Alternatively, control unit 14 can respond to a detected fault by controlling semiconductor switch 11, or alternatively provide a signal for responding to a fault to be transmitted to another protection device. Discharge current present during operation is generated, for example, by resistors 23 in submodule 1, which are used for functions not further described herein.

[0026] Each semiconductor switch 11 includes a switching element 21 and a diode 22, which is arranged in anti-parallel with the switching element 21.

[0027] Figure 4 A modular multilevel converter 2 is shown, comprising multiple sub-modules 1 arranged within it. The modular multilevel converter 2 is powered by a power grid 32, and the grid voltage is regulated by a feed transformer 33. For example, the AC voltage of the power grid 32 is converted to DC voltage using a diode bridge. Sub-modules 1 of the modular multilevel converter 2 generate an AC voltage from this DC voltage to drive a motor 31. To control or regulate the motor 31, the amplitude and frequency of the AC voltage generated by the sub-modules 1 for the motor 31 are adjusted by the modular multilevel converter 2. The control and / or regulation characteristics of the modular multilevel converter 2 can be improved by the inductor 34 shown.

[0028] As an alternative to diodes used to generate DC voltage, submodule 1, especially the series circuit of submodule 1, can also be used.

Claims

1. A method for detecting faults in submodule (1), wherein, The submodule (1) has at least two semiconductor switches (11) arranged in series and a capacitor (12) arranged in parallel with the two semiconductor switches (11). The semiconductor switches (11) enable the generation of zero voltage at the terminals (13', 13'') of the submodule (1) in a first switching state. The voltage (U) of the capacitor (12) is detected during the first switching state. C ), wherein, according to the voltage (U) of the capacitor (12) C The change in voltage (U) of the capacitor (12) is used to detect the fault of the capacitor (12), wherein the fault is detected by the change in voltage (U) of the capacitor (12). C The voltage (U) used for the capacitor (12) was changed. C The measurement value acquisition unit (15) is two or three times the quantization level of the capacitor (12) to detect the fault.

2. The method according to claim 1, wherein, The voltage (U) through the capacitor (12) C If the change is negative and below a preset lower limit, a fault in the capacitor (12) is detected.

3. The method according to any one of claims 1 or 2, wherein, The voltage (U) through the capacitor (12) C If the change in the value of the capacitor (12) exceeds the preset upper limit, a fault in the capacitor (12) can be detected.

4. A submodule (1) for a modular multilevel converter (2), wherein, The submodule (1) has at least two semiconductor switches (11) arranged in series and a capacitor (12) arranged in parallel with the two semiconductor switches (11), wherein, by means of the semiconductor switches (11), zero voltage can be generated at the terminals (13', 13'') of the submodule (1) in a first switching state, wherein the submodule (1) includes a measurement value acquisition unit (15) with a quantization level, wherein the quantization level is designed to acquire zero voltage when the voltage (U) of the capacitor (12) is zero. C When the quantization level is changed by two or three times, a fault in the capacitor (12) is detected.

5. The submodule (1) according to claim 4, wherein, The submodule (1) is connected to the control unit (14), which is configured to perform the method according to any one of claims 1 to 3.

6. A modular multilevel converter (2) having a plurality of sub-modules (1) as described in claim 5.

Images