Method for monitoring the condition of an energy storage device assembly of a modular multilevel converter (MMC)

Abstract

To provide a method for simultaneously monitoring a state of an energy storage device assembly of an MMC.SOLUTION: The MMC further comprises at least one alternating current AC terminal electrically connectable to an AC power source by the pre-charging circuit. During two or more pre-charging processes in which individual energy-storage-device assemblies of the MMC are initially charged to rated voltages by an AC pre-charging current supplied from an AC power source to at least one of the AC terminals, one or more measurements (e.g., v1) of the direct DC voltage across each energy-storage-device assembly are obtained at one or more times (e.g., at time t1) between the start and end of each pre-charging process. The one or more DC voltage measurements obtained during each pre-charging process are compared to determine the health or status of each energy storage device assembly of the MMC.SELECTED DRAWING: Figure 10

Description

[Technical Field]

[0001] The present invention relates to modular multilevel converters (MMCs), and more particularly to a method for simultaneously monitoring the status of the energy storage device assemblies of an MMC using one or more measurements of DC voltage across each energy storage device assembly during a pre-charge process.

[0002] The MMC can include multiple sub-modules connected together in series to define one or more converter arms of the MMC, and each sub-module can include an energy storage device assembly having one or more energy storage devices (e.g., one or more parallel-connected capacitors). [Background technology]

[0003] Modular multilevel converters (MMCs) are popular for a variety of medium to high power applications.

[0004] A typical MMC includes at least one converter arm having multiple series-connected submodules and an inductor. Each submodule can have any suitable topology, such as a half-bridge or full-bridge, as known to those skilled in the art. Each submodule typically has at least two controllable semiconductor switches and an energy storage device (e.g., a capacitor). In particular, if each submodule is a half-bridge submodule, it typically has a leg having two controllable semiconductor switches electrically connected in series. At least one energy storage device is electrically connected in parallel with the series-connected controllable semiconductor switches. In particular, if each submodule is a full-bridge submodule, it typically has a first leg having two controllable semiconductor switches electrically connected in series and a second leg having two controllable semiconductor switches electrically connected in series. The first and second legs are electrically connected in parallel between first and second direct current (DC) submodule rails. An energy storage device (e.g., a capacitor) is electrically connected in parallel with both the first and second legs and between the first and second DC rails.

[0005] An MMC typically has multiple converter arms. For example, an MMC can include one or more converter arms electrically connected in parallel between a pair of DC buses, with each converter leg including an upper converter arm having one or more submodules and a lower converter arm having one or more submodules and defining an alternating current (AC) bus therebetween. An MMC can include multiple converter arms, each electrically connected at one end to a respective AC bus and electrically connected to each other (e.g., in a star or delta configuration) or to another respective AC bus at the other end.

[0006] MMCs typically have a large number of identical submodules. For example, an MMC with a rated power of 100 MVA or greater may have at least 100 full-bridge submodules or at least 200 half-bridge submodules. Often, each submodule may include two or more energy storage devices electrically connected in parallel. Therefore, each submodule may have three to five individual energy storage devices (e.g., three to five parallel-connected capacitors). This means that MMCs typically have a very large number of energy storage devices. While defects in modern DC capacitors are rare, individual capacitors gradually degrade over time, and this degradation typically depends on the operating conditions of the MMC, such as the duty cycle and ambient temperature. Individually testing each energy storage device in an MMC is impractical. The large number of energy storage devices makes costs prohibitive, and accessibility is often limited depending on where the energy storage devices are located in the converter stack. However, the large number of energy storage devices presents a risk that one or more of the individual energy storage devices may begin to degrade sooner than expected, for example, within the first 10 years of operation.

[0007] Therefore, there is a need for a method to reliably determine the state of an energy storage device within an MMC. Summary of the Invention

[0008] The present invention provides a method for monitoring a condition of an energy storage device assembly of a modular multilevel converter (MMC), the MMC comprising a plurality of series-connected sub-modules (e.g., defining converter arms of the MMC), each sub-module comprising a plurality of semiconductor devices, each semiconductor device including at least a controllable semiconductor switch (e.g., IGBT), the energy storage device assembly including one or more individual energy storage devices (e.g., one or more individual capacitors) electrically connected in parallel, the MMC further comprising at least one alternating current (AC) terminal electrically connectable to an AC power source, the method comprising: During a pre-charge process in which individual energy storage device assemblies of the MMC are initially charged to a rated voltage by an AC pre-charge current supplied from an AC power source to at least one AC terminal, obtaining one or more measurements of a direct current (DC) voltage across each energy storage device assembly at one or more times between a start of the pre-charge process and an end of the pre-charge process; determining a state of each energy storage device assembly using one or more DC voltage measurements taken during the pre-charging process; The present invention provides a method comprising:

[0009] In each submodule, each semiconductor device may further include an anti-parallel-connected diode, i.e., a diode connected anti-parallel to its associated controllable semiconductor switch. Typically, the controllable semiconductor switch including the anti-parallel-connected diode includes, for example, an insulated gate bipolar transistor (IGBT). During the pre-charge process, the energy storage device assemblies of each submodule may be charged through the anti-parallel-connected diode, and the controllable semiconductor switch is not switched on and off. Because all energy storage device assemblies of a submodule are charged simultaneously by the AC-side pre-charge process, they are charged in a series of "steps," i.e., periods in which the DC voltage across each energy storage device assembly increases alternate with periods in which the DC voltage across each energy storage device assembly remains substantially constant. Such a pre-charge process may also be described as a "pulse" pre-charge process. However, the overall shape of the DC voltage curve for each energy storage device assembly during pre-charge is substantially exponential. For convenience in the following description, the periods in which the voltage across each energy storage device assembly increases may be referred to as "charging periods," and the periods in which the voltage across each energy storage device assembly remains substantially constant may be referred to as "non-charging periods." Each "step" (or "pulse") includes a charging time and a non-charging time. The transitions between "steps" (or "pulses") and between charging and non-charging periods are regular and well-defined and depend on the frequency of the AC pre-charging current supplied to at least one AC terminal of the MMC during the pre-charging process. For example, if the frequency of the supplied AC pre-charging current is 50 Hz and the sub-module is implemented in a full-bridge topology, the duration of each "step" is 10 ms, and the duration of each charging and non-charging period is 5 ms. If the sub-module is implemented in a half-bridge topology, the duration of each "step" is 20 ms, and the duration of each charging and non-charging period is 10 ms.If the start of the pre-charge process is accurately determined (see below), it is therefore possible to control when the DC voltage across each energy storage device assembly is measured. In particular, it is possible to control the DC voltage measurement unit of the MMC during pre-charge so that one or more DC voltage measurements are obtained for each energy storage device assembly during one or more periods of the non-charge period. It will be appreciated that the non-charge period is an ideal time to obtain DC voltage measurements with minimal variation because the voltage across each energy storage device assembly remains substantially constant. Given a particular non-charge period for a particular energy storage device assembly, a single DC voltage measurement can be obtained, or if two or more DC voltage measurements are obtained during the same non-charge period, an average value of the DC voltage for that particular non-charge period can be determined. This can improve the reliability of the DC voltage measurements obtained during the pre-charge process. The process of obtaining DC voltage measurements for each energy storage device assembly may be repeated during the pre-charge process, and optionally, at least one DC voltage measurement across each energy storage device assembly is obtained during at least a portion of the non-charge period. For example, at least one DC voltage measurement can be obtained for each energy storage device assembly during a selected non-charging period, or during two or more selected non-charging periods. If two or more DC voltage measurements are obtained for each selected non-charging period, an average DC voltage can be determined for each selected non-charging period. The sampling rate for the DC voltage measurements is typically significantly greater than the duration of each non-charging period, typically 10 or 20 ms. As a result, it is usually not difficult to obtain several DC voltage measurements during a particular non-charging period so that an average value can be determined.

[0010] The multiple series-connected sub-modules can define converter arms of the MMC. Each converter arm can further include at least one inductor. If each converter arm includes at least one inductor, it is preferably not bypassed during the pre-charging process. The MMC can include multiple converter arms, which can be electrically connected to each other and / or to any external circuit or load in any suitable arrangement known to those skilled in the art. In particular, the MMC can be utilized in a wide variety of medium- to high-power applications, and the converter arms can be configured accordingly to connect to an external power system, including a transmission system or a power grid, and / or to an electrical load if the MMC is configured for power conversion. A pair of converter arms can be connected in series to define a converter leg of the MMC, and the converter arms define an AC bus therebetween. The converter arms may be connected to each other in a delta or star configuration, for example, when the MMC is configured as a static synchronous compensator (STATCOM). The MMC can have any suitable overall topology with at least one AC terminal. The MMC may have three AC terminals.

[0011] The sub-modules can have any suitable topology, such as a half-bridge or full-bridge, as known to those skilled in the art. In a half-bridge topology, each sub-module includes a leg having two controllable semiconductor switches electrically connected in series, and an energy storage device assembly is electrically connected in parallel with the series-connected controllable semiconductor switches. In a full-bridge topology, each sub-module includes a first leg having two controllable semiconductor switches electrically connected in series and a second leg having two controllable semiconductor switches electrically connected in series. The first and second legs are electrically connected in parallel between first and second DC rails. The energy storage device assembly is electrically connected in parallel with both the first and second legs between the first and second DC rails.

[0012] As noted above, each sub-module energy storage device assembly may comprise two or more individual energy storage devices electrically connected in parallel, e.g., three to five individual parallel-connected capacitors.

[0013] Each sub-module may include a DC voltage measurement unit (e.g., a voltage sensor) adapted to obtain one or more voltage measurements. Each DC voltage measurement unit may be electrically connected to the energy storage device assembly of its respective sub-module. The DC voltage measurements may be transmitted from each DC voltage measurement unit to a controller (e.g., a converter controller) using any suitable protocol. The controller may store or record the DC voltage measurements and perform processing described in more detail below to monitor the status of the energy storage device assembly. The controller may also transmit the DC voltage measurements received from each DC voltage measurement unit to a remote server using any suitable protocol, optionally after temporarily storing or recording the DC voltage measurements but without performing any processing. The DC voltage measurements may also be transmitted directly from each DC voltage measurement unit to the remote server. The remote server may store or record the DC voltage measurements and perform processing described in more detail below to monitor the status of the energy storage device assembly. The processing performed by the controller or remote server may include a calibration or compensation process, as described in more detail below. Each sub-module of an MMC typically includes a DC voltage measurement unit, and this method of condition monitoring does not typically require any modifications to the sub-module design.

[0014] At least one AC terminal of the MMC may be electrically connected to an AC power source by a pre-charge circuit. The AC power source may be, for example, a power grid or other supply network. The power grid or other supply network may be a three-phase power grid or supply network. The pre-charge circuit may include a pre-charge resistor electrically connected in series with each AC terminal of the MMC. For example, if the MMC has three AC terminals, the pre-charge circuit may include three pre-charge resistors, each electrically connected in series with a respective AC terminal of the MMC. Each pre-charge resistor may be bypassed during normal operation of the MMC.

[0015] The pre-charge circuit may optionally further include a switch electrically connected between each pre-charge resistor and the AC power source. The pre-charge circuit may include three switches, each optionally electrically connected between a respective pre-charge resistor and the AC power source. Each switch may be opened or closed by, for example, a controller.

[0016] During the pre-charge process, the MMC and pre-charge circuit may be considered to be a simplified resistor-capacitor circuit (RC circuit).

[0017] The start of the pre-charge process can be determined, for example, by a close signal controlling one or more switches of the pre-charge circuit to close or by measuring the pre-charge current. For example, after one or more switches are controlled to close, the start of the pre-charge process can be determined by detecting the first peak of the pre-charge current supplied from the AC power source to the pre-charge circuit. Using a pre-charge current measurement is often more accurate than using a close signal or some other determination of when one or more switches are closed, and can be particularly important when DC voltage measurements taken during two or more pre-charge processes are compared for status monitoring purposes. Any measurement time can then be determined relative to the start time, which allows DC voltage measurements to be taken exactly simultaneously during different pre-charge processes. For example, the voltage across the energy storage device assembly may be acquired exactly x milliseconds after the start time of the initial pre-charge process. During each subsequent pre-charge process, the voltage across the energy storage assembly may also be acquired exactly x milliseconds after the start of the respective pre-charge process. The end of the pre-charge process can be determined when the voltage across the energy storage device assembly reaches a predetermined value for normal switching operation of the MMC (i.e., the rated voltage). The pre-charge process can typically be completed within a few seconds. After the pre-charge process is complete and the energy storage device assembly is charged to its rated voltage, the MMC can transition to normal operation, where the controllable semiconductor switches in each sub-module are switched on and off as needed.

[0018] At the end of a particular pre-charge process, at least one DC voltage measurement is obtained for the energy storage device assembly of each sub-module. In practice, as described above, two or more DC voltage measurements may be obtained for the energy storage device assembly of each sub-module. Each DC voltage measurement may be associated with a different non-charging period; for example, a first DC voltage measurement may be obtained during a first non-charging period, and a second DC voltage measurement may be obtained during a second non-charging period. Alternatively, two or more DC voltage measurements may be associated with the same non-charging period to obtain an average DC voltage for the non-charging period. It will be readily apparent that the DC voltage measurements for all energy storage device assemblies are obtained simultaneously during the pre-charge process using only existing “hardware” in the form of the pre-charge circuitry and each sub-module’s DC voltage measurement unit. This means that the method can be applied to existing MMCs without requiring significant modifications to installed hardware or control software.

[0019] One or more DC voltage measurements for each energy storage device assembly are preferably stored, for example, in memory. The stored DC voltage measurements are then processed to determine the status of each energy storage device assembly. The DC voltage measurements may be processed by the controller or transmitted to a remote server for processing. This processing does not need to be performed in "real time" and may occur after the pre-charging process has finished and the MMC has begun its normal operation.

[0020] The condition monitoring can be performed using DC voltage measurements obtained from a single pre-charge process. Condition monitoring may also be performed using DC voltage measurements obtained from two or more pre-charge processes, which may be performed at intervals, for example, annually, to provide periodic monitoring of the condition of the energy storage device assemblies of the MMC over its operational life. In this case, the method may further include, during a second pre-charge process in which individual energy storage device assemblies of the MMC are initially charged to a rated voltage by an AC pre-charge current supplied from an AC power source to at least one AC terminal, obtaining one or more measurements of DC voltage across each energy storage device assembly at one or more times between the start of the second pre-charge process and the end of the second pre-charge process, and determining the condition of each energy storage device assembly using the one or more DC voltage measurements obtained during the pre-charge process (i.e., the initial pre-charge process) and one or more DC voltage measurements obtained during the second (or any subsequent) pre-charge process.

[0021] The DC voltage measurements taken from a single pre-charge process can be used in several different ways to determine the state of each energy storage device assembly. For example, the DC voltage measurements taken for each energy storage device assembly at a particular time can be compared. If there are n sub-modules, and therefore n energy storage device assemblies (where n is an integer), then DC voltage measurements v1, v2, ..., v3 taken simultaneously (e.g., at time t1) during the pre-charge process can be compared. n can be compared. (v n is the nth submodule (e.g., SM nIt will be understood that v1 is the voltage across the energy storage device assembly of a first sub-module (e.g., SM1) measured at time t1, v2 is the voltage across the energy storage device assembly of a second sub-module (e.g., SM2) measured at time t1, and so on.) The time at which the DC voltage measurements are taken may be determined with reference to the start time of the pre-charge process. This ensures that the DC voltage measurements for all sub-modules are taken simultaneously. In one example, DC voltage measurements v1, v2, ..., v taken at time t1 may be n may be compared to the expected voltage for time t1. (It will be appreciated that when two or more DC voltage measurements are taken in rapid succession, e.g., during the same non-charging period, for purposes of determining an average voltage, the DC voltage measurements may be taken during a narrow time window, the beginning of which may correspond to a particular time (e.g., time t1) for adequate repeatability of multiple DC voltage measurements in different pre-charge processes.) If an individual DC voltage measurement deviates from the expected voltage by more than a predetermined amount (e.g., more than ±x%), this may indicate that at least one of the energy storage devices of a particular energy storage device assembly is not in the best condition and further testing may be performed. This may be repeated for DC voltage measurements taken at later times during the pre-charge process, i.e., DC voltage measurements v1, v2, ..., v taken at time t2. n may be compared with the expected voltage for time t2, and the DC voltage measurements v1, v2, ..., v taken at time t3 n may be compared to the expected voltage for time t3, and so on. In another example, DC voltage measurements v1, v2, ..., v taken at time t1 nUsing the DC voltage measurements v1, v2, ..., v3 taken at time t2, the average voltage for time t1 can be determined. If the individual DC voltage measurements deviate from the average voltage by more than a predetermined amount (e.g., more than ±x%), this can indicate that at least one of the energy storage devices of a particular energy storage device assembly is not in the best condition and further testing can be performed. This may be repeated at later times during the pre-charge process, i.e., DC voltage measurements v1, v2, ..., v3 taken at time t2 n may be used to determine the average voltage for time t2 and compared to the DC voltage measurements v1, v2, ..., v taken at time t3. nmay be used to determine an average voltage for time t3 and compared to the average value for time t3, and so on. Times t1, t2, t3, etc. may be selected to correspond to non-charging periods during which the voltage across each energy storage device assembly remains substantially constant during pre-charge. Due to manufacturing tolerances, the measured DC voltages of different energy storage device assemblies at a particular time may be expected to vary by up to several percent, even if the individual energy storage devices have the exact same design and nominal capacitance value. Condition monitoring here may be as simple as, for example, classifying any energy storage device assemblies whose DC voltage measurements deviate from an expected or average voltage by less than a predetermined amount as having an acceptable condition, and classifying any remaining energy storage device assemblies as requiring further testing or monitoring. Alternatively, condition monitoring may depend on the amount by which individual DC voltage measurements deviate from an expected or average voltage, or may use two or more predetermined amounts associated with different deviations from the average or expected voltage. For example, if an individual DC voltage measurement deviates from the average or expected voltage by more than a first predetermined amount (e.g., more than ±x1%) and not more than a second predetermined amount (e.g., not more than ±x2%, where x2 is greater than x1), this may indicate that at least one of the energy storage devices of a particular energy storage device assembly is not in the best condition and further testing may be performed. However, if an individual DC voltage measurement deviates from the average or expected voltage by more than a second predetermined amount (e.g., more than ±x2%), this may indicate that more urgent action is needed, including potentially replacing one or more of the energy storage devices of a particular energy storage device assembly. Condition monitoring can be based on any suitable number of predetermined amounts that allow for classification of deviations of individual DC voltage measurements of a particular energy storage device assembly from expected or average voltages and, if necessary, taking appropriate action in response to the indicated degradation.

[0022] Condition monitoring generally preferably uses DC voltage measurements obtained from two or more pre-charge processes. For example, an initial pre-charge process can be performed to establish a "baseline" for the condition of the energy storage device assemblies. Thereafter, pre-charge processes may be performed at intervals, such as every six months, every year, or every two to three years, as part of a regular maintenance program for the MMC, so that the condition of the energy storage device assemblies can be periodically monitored for degradation. For example, if three pre-charge processes are performed, where an initial pre-charge process is performed in year y0, a second pre-charge process is performed one year later in year y1, and a third pre-charge process is performed another year later in year y2, the stored DC voltage measurements for each pre-charge process can be used to determine how the condition of each energy storage device assembly has changed over time. For example, DC voltage measurements v1, v2, ..., v obtained at time t1 in year y1 can be used to determine how the condition of each energy storage device assembly has changed over time. n are the DC voltage measurements v1, v2, ..., v taken at time t1 in year y0. n DC voltage measurements v1, v2, ..., v taken at time t1 in year y2 can be compared. n are the DC voltage measurements v1, v2, ..., v taken at time t1 in year y0. n and optionally DC voltage measurements v1, v2, ..., v taken at time t1 in year y1. nThe DC voltage measurements for the same energy storage device assembly are compared; in other words, the voltage v1 of the energy storage device assembly of a first sub-module (e.g., SM1) measured at time t1 in year y1 is compared to the voltage v1 of the energy storage device assembly of the first sub-module measured at time t1 in year y0, the voltage v2 of the energy storage device assembly of a second sub-module (e.g., SM2) measured at time t1 in year y1 is compared to the voltage v2 of the energy storage device assembly of the second sub-module measured at time t1 in year y0, and so on. Time t1 will be the same for each pre-charge process. In other words, the DC voltages are measured at the same point during each pre-charge process, as determined from the start of each pre-charge process. o The same comparison can be performed for DC voltage measurements taken at time t2, such as year y1, and year y2. This allows for a direct comparison of DC voltage measurements taken for each individual energy storage device assembly simultaneously during each pre-charge process (e.g., yearly). Thus, changes in DC voltage measurements that may indicate degradation of a particular energy storage device assembly of the MMC can be identified.

[0023] The DC voltage measurements obtained for each energy storage device assembly can also be used to determine a characteristic of each energy storage device assembly, which can then be used to determine the state of each energy storage device assembly. For example, the determined characteristic may be the time constant of each energy storage device assembly, e.g., the time it takes for each energy storage device assembly to charge to 63.2% of its rated voltage. The time constant of each energy storage device assembly is known to be equal to the capacitance of the energy storage device assembly multiplied by the total resistance of the pre-charge circuit. As a result, any change in the time constant determined during different pre-charge processes indicates a change in the capacitance of a particular energy storage device assembly. The total resistance of the pre-charge circuit is the same for all energy storage device assemblies being charged simultaneously. In one example, the time constants τ1, τ2, ..., τ2 determined for the pre-charge process are n can be compared with the expected time constant (τ n is the nth submodule (e.g., SM n It will be understood that τ is the time constant of the energy storage device assembly of a first sub-module (e.g., SM1), τ is the time constant of the energy storage device assembly of a second sub-module (e.g., SM2), and so on, until the time constant of the energy storage assembly of SM1 is reached.) If an individual time constant deviates from the expected time constant by more than a predetermined amount (e.g., more than ±x%), this may indicate that at least one of the energy storage devices of a particular energy storage device assembly is not in the best condition and further testing may be performed. In another example, the time constants τ, τ, ..., τ determined for the pre-charging process may be used to calculate the time constants τ, τ, ..., τ ncan be used to determine an average time constant for the pre-charge process. If an individual time constant deviates from the average time constant by more than a predetermined amount (e.g., more than ±x%), this can indicate that at least one of the energy storage devices of a particular energy storage device assembly is not in the best condition and further testing can be performed. Condition monitoring here can be as simple as, for example, classifying any energy storage device assemblies whose time constants deviate from the expected or average time constant by less than a predetermined amount as having an acceptable condition, and classifying any remaining energy storage device assemblies as requiring further testing or monitoring. Alternatively, condition monitoring may depend on the amount by which individual time constants deviate from the expected or average time constant, or may use two or more predetermined amounts associated with different deviations from the average or expected time constant. For example, if an individual time constant deviates from the average or expected time constant by more than a first predetermined amount (e.g., more than ±x1%) and not more than a second predetermined amount (e.g., not more than ±x2%, where x2 is greater than x1), this may indicate that at least one of the energy storage devices of a particular energy storage device assembly is not in the best condition and further testing may be performed. However, if an individual time constant deviates from the average or expected time constant by more than a second predetermined amount (e.g., more than ±x2%), this may indicate that more urgent action is needed, including potentially replacing one or more of the energy storage devices of a particular energy storage device assembly. Condition monitoring can be based on any suitable number of predetermined amounts that allow for classification of deviations of the individual time constants of a particular energy storage device assembly from the expected or average time constant and, if necessary, taking appropriate action in response to the indicated degradation.

[0024] It is generally preferred that the condition monitoring use time constants obtained from two or more pre-charge processes as described above. For example, if three pre-charge processes are performed, where an initial pre-charge process is performed in year y0, a second pre-charge process is performed one year later in year y1, and a third pre-charge process is performed another year later in year y2, the stored DC voltage measurements for each pre-charge process can be used to determine how the time constant of each energy storage device assembly has changed over time. For example, the time constants τ1, τ2, ..., τ determined in year y1 can be used to determine how the time constants of each energy storage device assembly have changed over time. n are the time constants τ1, τ2, ..., τ determined in year y0. n can be compared with the time constants τ1, τ2, ..., τ determined in year y2. n are the time constants τ1, τ2, ..., τ determined in year y0. n and optionally with time constants τ1, τ2, ..., τ determined for year y1. n Time constants for the same energy storage device assembly are compared, in other words, the time constant τ of the energy storage device assembly of a first sub-module (e.g., SM1) determined in year y1 is compared with the time constant τ of the energy storage device assembly of the first sub-module determined in year y0, the time constant τ of the energy storage device assembly of a second sub-module (e.g., SM2) determined in year y1 is compared with the time constant τ of the energy storage device assembly of the second sub-module obtained in year y0, and so on.

[0025] Any change in the DC voltage measured across a particular energy storage device assembly simultaneously during different pre-charge processes (or any change in the determined time constant or any other capacitance-dependent characteristic of a particular energy storage device assembly) indicates a change in the capacitance of the energy storage device assembly. It is important to note that it is not necessary to determine the absolute capacitance of the energy storage device assembly of the MMC, as other electrical parameters, such as the absolute value of the total resistance of the pre-charge circuit, also need to be known. In the above example, if a comparison of the respective DC voltage measurements v1 taken in year y1 and year y0 indicates that the DC voltage has decreased by 1%, this can be considered to indicate that the capacitance of the energy storage device assembly of the first sub-module (i.e., SM1) has decreased by 1% over the course of one year. While this may not require immediate action, if the capacitance continues to decrease, it may exceed the allowable tolerance, and the energy storage device assembly may be scheduled for replacement at an appropriate time in the future. Further testing can also be performed to try and identify whether only specific energy storage devices of the energy storage device assembly are affected. This can avoid the need to replace all energy storage devices of a particular sub-module. This method can indicate whether the capacitance of a particular energy storage device assembly has decreased by more than a threshold amount, which can optionally be determined based on the expected degradation rate of the energy storage devices utilized in the MMC and the time period between each pre-charge process. As described above, the measured DC voltage indicates capacitance. For example, the expected degradation rate of the capacitance of an energy storage device may be 0.5% per year. Thus, a 0.9% decrease in capacitance (indicated by the DC voltage measured across a particular energy storage device assembly) would not exceed the threshold amount if, for example, the time period between each pre-charge process was two years. However, if the pre-charge processes were only one year apart rather than two years apart, the threshold amount would be exceeded.This is because capacitance decreases by more than 0.5% per year. If the capacitance of an energy storage device assembly decreases by more than approximately 5% compared to the initial capacitance value measured during the first commissioning of the MMC, it may have to be replaced. Degradation of modern energy storage devices (e.g., DC capacitors) is typically very slow and can depend on operating conditions (e.g., load cycle conditions, ambient temperature, etc.). Under normal operating conditions, degradation of less than 5% is expected during at least 10 years of operation. However, the normal operating life of an MMC can be expected to be at least 20 years, which means that regular maintenance checks of installed DC capacitors are necessary, especially after the first 10 years of operation. While an online (or "real-time") method for monitoring the condition of the energy storage device assembly is not required, the present method provides a simple, automated, and reliable way to monitor the condition of the energy storage device assembly when performing an AC-side pre-charging process in order to meet customer expectations and improve MMC availability by scheduling preventative maintenance and replacement of degraded energy storage devices. This method typically obtains DC voltage measurements in an automated manner, always using the same pre-charge circuitry and other components, such as the same DC voltage measurement unit, resulting in repeatability with a tolerance of less than 0.2%. The time at which the DC voltage is measured by each submodule's DC voltage measurement unit can be carefully controlled based on the detected start of each pre-charge process. This ensures that DC voltage measurements are taken at exactly the same time during different pre-charge processes, i.e., the times t1, t2, and t3 mentioned in the example above are the same for each pre-charge of the MMC. This also ensures that DC voltage measurements are taken only during non-charging periods, if appropriate.

[0026] It will be appreciated that there may be some variation in electrical parameters when the pre-charge process is performed at different times, e.g., every year. Although the same pre-charge circuit with the same pre-charge resistor is used for each pre-charge of the MMC, there may still be some variation in the absolute value of the total resistance of the pre-charge circuit, for example, as a result of differences in ambient temperature. For example, there may also be some variation in the AC pre-charge current supplied by the AC power source, especially if the AC power source is a power grid. However, such variation may be considered to affect the DC voltage measurements for each energy storage device assembly in the same way and, therefore, can be corrected or compensated for using an appropriate calibration or compensation process. For example, because the total resistance of the pre-charge circuit is the same for all energy storage device assemblies being charged simultaneously, variations in the total resistance when each pre-charge process is performed will have the same effect on the DC voltage measured across all energy storage device assemblies. In other words, the disturbances eliminated by the calibration or compensation process are “common-mode” disturbances. The calibration or compensation process may be performed on the stored DC voltage measurements obtained during a pre-charge process before comparison with stored DC voltage measurements obtained during a previous pre-charge process is performed. Any suitable calibration or compensation process may be used.

[0027] The calibrated or compensated DC voltage measurements obtained after performing the calibration or compensation process may also be stored and then used in any comparisons made with DC voltage measurements obtained in subsequent pre-charge processes.

[0028] The present invention provides an MMC comprising: A plurality of series-connected sub-modules, each sub-module comprising: a plurality of semiconductor devices, each semiconductor device including at least a controllable semiconductor switch; and An energy storage device assembly including one or more individual energy storage devices electrically connected in parallel a plurality of series-connected sub-modules comprising: at least one AC terminal electrically connectable to an AC power source; A condition monitoring device, comprising: obtaining one or more measurements of DC voltage across each energy storage device assembly at one or more times between the start of the pre-charge process and the end of the pre-charge process, wherein during the pre-charge process, the individual energy storage device assemblies of the MMC are initially charged to their rated voltage by an AC pre-charge current supplied from an AC power source to at least one AC terminal; One or more DC voltage measurements taken during the pre-charge process are used to determine the status of each energy storage device assembly and condition monitoring devices adapted for The present invention further provides an MMC comprising:

[0029] Further technical features of the MMC may be as described above.

[0030] The condition monitoring device may include the controller and / or the remote server described above. When each sub-module further includes a DC voltage measurement unit (e.g., a voltage sensor), the condition monitoring device may further include a DC voltage measurement unit. The DC voltage measurement unit is adapted to obtain one or more voltage measurements. Each DC voltage measurement unit may be electrically connected to the energy storage device assembly of the respective sub-module. The DC voltage measurements may be transmitted from each DC voltage measurement unit using any suitable protocol. [Brief explanation of the drawings]

[0031] [Figure 1] 1 shows a first example of a modular multilevel converter (MMC) according to the present invention implemented as a variable speed drive (VSD). [Figure 2] FIG. 2 shows a second example of an MMC according to the present invention implemented as a static synchronous compensator (STATCOM). [Figure 3] FIG. 10 shows a third example of an MMC according to the present invention implemented as a STATCOM. [Figure 4] FIG. 1 shows a first example of a converter arm for an MMC having multiple series-connected sub-modules with a half-bridge topology. [Figure 5] FIG. 10 shows a second example of a converter arm for an MMC having multiple series-connected sub-modules with a full-bridge topology. [Figure 6] FIG. 1 illustrates a pre-charge circuit that may be used to pre-charge an MMC. [Figure 7] FIG. 10 shows DC voltage curves for a capacitor assembly of an MMC during the pre-charging process. [Figure 8] FIG. 8 shows a portion of the DC voltage curve of FIG. 7. [Figure 9] FIG. 10 shows DC voltage curves for multiple capacitor assemblies of an MMC during the pre-charging process. [Figure 10] 10A-10C show DC voltage curves for a capacitor assembly of an MMC during different pre-charging processes. DETAILED DESCRIPTION OF THE INVENTION

[0032] FIG. 1 shows a first example of a modular multilevel converter (MMC) 1A implemented as a variable speed drive (VSD).

[0033] The MMC 1A includes three converter legs 21, 22, and 23 connected in parallel between first and second direct current (DC) buses 4, 6. Each converter leg 21, 22, and 23 is divided into an upper converter arm and a lower converter arm. Each upper converter arm is comprised of n series-connected sub-modules 81, 82, ..., 8 n Each lower converter arm includes n series-connected sub-modules 101, 102, ..., 10 n where n is any suitable integer. Each converter arm includes an inductor.

[0034] The connection between the upper and lower converter arms of the first converter leg 21 defines a first alternating current (AC) bus 12. The connection between the upper and lower converter arms of the second converter leg 22 defines a second AC bus 14. The connection between the upper and lower converter arms of the third converter leg 23 defines a third AC bus 16. The first, second, and third AC buses 12, 14, and 16 each define a respective AC phase (labeled "A," "B," and "C") and may be connected to a three-phase AC load or a three-phase AC supply. The first and second DC buses 4 and 6 may be connected to a DC load or a DC supply and may include a DC link with one or more energy storage devices (e.g., DC link capacitors).

[0035] The MMC1A connects the sub-modules 81, 82, ..., 83 of the first, second, and third converter legs 21, 22, and 23 in a manner well known to those skilled in the art. n and 101, 102, ..., 10 n By operating the converter, a DC input voltage can be converted to an AC output voltage, and vice versa.

[0036] 2 and 3 show second and third examples of MMCs 1B and 1C implemented as static synchronous compensators (STATCOMs).

[0037] Each of MMCs 1B and 1C includes three converter arms 181, 182, and 183. The first converter arm 181 is connected at a first end to a first AC bus 20. The second converter arm 182 is connected at a first end to a second AC bus 22. The third converter arm 183 is connected at a first end to a third AC bus 24. The first, second, and third AC buses 20, 22, and 24 each define a respective AC phase (labeled "A," "B," and "C"). In MMC 1B shown in FIG. 2, the first, second, and third converter arms 181, 182, and 183 are connected to each other at their second ends in a star configuration. In MMC 1C shown in FIG. 3, the first, second, and third converter arms 181, 182, and 183 are connected to each other at their second ends in a delta configuration.

[0038] Each converter arm 181, 182, and 183 is made up of n series-connected submodules 261, 262, ..., 26 n where n is any suitable integer. Each converter arm 181, 182, and 183 includes an inductor.

[0039] Submodules 261, 262, ..., 26 of the first, second, and third converter arms 181, 182, and 183 n By operating MMCs 1B and 1C, MMCs 1B and 1C can act as sources or sinks of reactive AC power, for example, to regulate an AC transmission network.

[0040] 4 and 5 show first and second examples of converter arms 28A and 28B, respectively. Converter arm 28A includes n series-connected sub-modules 301, 302, ..., 303. n Similarly, converter arm 28B includes n series-connected sub-modules 321, 322, ..., 32 nwhere n is any suitable integer. Converter arms 28A and 28B can each be utilized as one of the converter arms of MMC1A shown in FIG. 1 or one of the converter arms of MMC1B or 1C shown in FIGS. 2 and 3.

[0041] Sub-modules 301, 302, ..., 30 in Fig. 4 n is shown as having a half-bridge topology, in which each sub-module 301, 302, ..., 30 n Each sub-module 301, 302, ..., 303 includes a leg having two controllable semiconductor switches S1 and S2 electrically connected in series. Each semiconductor switch is shown as an IGBT with anti-parallel connected diodes D1 and D2, but it will be readily understood that other controllable semiconductor switches may be used. Each sub-module 301, 302, ..., 304 includes a leg having two controllable semiconductor switches S1 and S2 electrically connected in series. Each semiconductor switch is shown as an IGBT with anti-parallel connected diodes D1 and D2, but it will be readily understood that other controllable semiconductor switches may be used. n comprises three energy storage devices (i.e., three capacitors) electrically connected in parallel with series-connected semiconductor switches S1 and S2. The three capacitors form a capacitor assembly CA of the respective sub-module. Although three capacitors are shown, each capacitor assembly CA1, CA2, ..., CA n It will be appreciated that may have any suitable number of parallel connected capacitors.

[0042] Sub-modules 321, 322, ..., 32 in Fig. 5 n is shown as having a full-bridge topology, in which each sub-module 321, 322, ..., 32 nThe sub-module 321 comprises a first leg having two controllable semiconductor switches S1 and S2 connected in series, and a second leg having two controllable semiconductor switches S3 and S4 connected in series. Each semiconductor switch is shown as an IGBT with an anti-parallel connected diode D1, D2, ..., D4, but it will be readily understood that other controllable semiconductor switches can be used. The first and second legs are electrically connected in parallel between the first and second DC rails. Each sub-module 321, 322, ..., 32 n comprises three energy storage devices (i.e., capacitors) electrically connected in parallel with series-connected semiconductor switches S1 and S2. The three capacitors form a capacitor assembly CA of the respective sub-module. Although three capacitors are shown, each capacitor assembly CA1, CA2, ..., CA n It will be appreciated that may have any suitable number of parallel connected capacitors.

[0043] Each sub-module 301, 302, ..., 30 n The semiconductor switches S1 and S2 and the respective sub-modules 321, 322, ..., 32 n The semiconductor switches S1, S2, ..., S4 are connected to the capacitor assemblies CA1, CA2, ..., CA of the converter arm 28A or 28B. n are controlled by respective gate drive commands to be switched between an on state and an off state (ie, switched on and off) to selectively charge, discharge, or bypass.

[0044] Each sub-module 301, 302, ..., 30 n and 321, 322, ..., 32 n are the capacitor assemblies CA1, CA2, ..., CA nEach sub-module 30 includes a voltage sensor 34 electrically connected in parallel with the sub-module 30, i.e., in parallel with each individual capacitor. Each voltage sensor 34 provides a measurement of the voltage across a respective capacitor assembly. n The voltage sensor 34 of the submodule 30 n Capacitor Assembly CA n The measured voltage across v n The voltage sensor 34 of sub-module 301 provides a measurement v1 of the voltage across the capacitor assembly CA1 of sub-module 301, the voltage sensor 34 of sub-module 302 provides a measurement v2 of the voltage across the capacitor assembly CA2 of sub-module 302, and so on. n The voltage sensor 34 of the submodule 32 n Capacitor Assembly CA n The measured voltage across v n The voltage sensor 34 of submodule 321 provides a measurement v1 of the voltage across capacitor assembly CA1 of submodule 321, the voltage sensor 34 of submodule 322 provides a measurement v2 of the voltage across capacitor assembly CA2 of submodule 322, and so on. It will be appreciated that this applies to all submodules within an MMC, i.e., the submodules in all converter arms of the MMC. For example, in MMC 1A shown in FIG. 1 having six converter arms, if each converter arm has n submodules, there will be a total of (6×n) capacitor assemblies, and the voltage sensor 34 of each submodule will provide (6×n) individual DC voltage measurements, i.e., across capacitor assemblies CA1, CA2, CA3, ..., CA (6n-1) , CA 6n are recorded individually, v1, v2, v3, ..., v (6n-1) , v 6nIn MMC1B or 1C shown in Figures 2 and 3, each of which has three converter arms, if each converter arm has n sub-modules, there will be a total of (3 x n) capacitor assemblies, and the voltage sensor of each sub-module will simultaneously acquire (3 x n) individual DC voltage measurements, i.e., the capacitor assemblies CA1, CA2, CA3, ..., CA (3n-1) , CA 3n are recorded individually, v1, v2, v3, ..., v (3n-1) , v 3n For convenience in the following discussion, it can be assumed that the MMC whose condition is being monitored has a total of p sub-modules, where p is an integer (e.g., p=3n or 6n), such that there are a total of p capacitor assemblies and the voltage sensor 34 in each sub-module simultaneously takes p individual DC voltage measurements.

[0045] The individual DC voltage measurements taken by the voltage sensors 34 may be provided using any suitable protocol to the controller 100 or a remote server, which stores or records the DC voltage measurements and performs the processes described in more detail below to monitor the status of the capacitor assemblies of the MMC.

[0046] 6 shows an example of a pre-charge circuit 36 ​​that can be used to pre-charge an MMC. The pre-charge circuit 36 ​​includes three AC terminals 38a, 38b, and 38c that define respective AC phases (labeled "A," "B," and "C") and are electrically connected to corresponding AC buses of the MMC. For example, the first AC terminal 38a is electrically connected to the first AC bus 12 of MMC 1A (or the first AC bus 20 of MMC 1B or 1C), the second AC terminal 38b is electrically connected to the second AC bus 14 of MMC 1A (or the second AC bus 22 of MMC 1B or 1C), and the third AC terminal 38c is electrically connected to the third AC bus 16 of MMC 1A (or the third AC bus 24 of MMC 1B or 1C).

[0047] The first AC terminal 38a is electrically connected to a corresponding AC phase of a three-phase AC power grid 40 by a first pre-charge resistor 42a and a first switch 44a. The second AC terminal 38b is electrically connected to a corresponding AC phase of the power grid 40 by a second pre-charge resistor 42b and a second switch 44b. The third AC terminal 38c is electrically connected to a corresponding AC phase of the power grid 40 by a third pre-charge resistor 42c and a third switch 44c. The first, second, and third switches 44a, 44b, and 44c may be controlled to open and close by a controller (not shown). The power grid 40 may have a grid frequency of, for example, 50 Hz.

[0048] The bypass switches 46a, 46b, and 46c of the pre-charge circuit 36 ​​may be closed to bypass the pre-charge resistors 42a, 42b, and 42c. The bypass switches 46a, 46b, and 46c may be closed during normal operation of the MMC, i.e., after the pre-charge process is completed and when the MMC is receiving power from or supplying power to the power grid 40.

[0049] The health or condition of the capacitor assemblies of an MMC (e.g., MMC 1A, 1B, or 1C shown in Figures 1-3) can be monitored by taking DC voltage measurements during a pre-charge process in which the individual capacitor assemblies are simultaneously charged to their rated voltage by AC pre-charge current supplied from power grid 40 to the MMC's AC bus (or AC terminals) via pre-charge circuit 36.

[0050] FIG. 7 shows an example of how the DC voltage across the capacitor assemblies of an individual sub-module varies during the pre-charging process.

[0051] During the pre-charging process, the parallel-connected capacitors of the capacitor assemblies are charged in a series of "steps," i.e., periods in which the DC voltage across the capacitor assemblies increases (or "charging periods") alternate with periods in which the DC voltage across the capacitor assemblies remains substantially constant (or "non-charging periods"). However, the overall shape of the DC voltage curve for the capacitor assemblies during pre-charging is substantially exponential, as shown in Figure 7. It is readily apparent that a corresponding DC voltage curve can be drawn for each capacitor assembly of the MMC; i.e., if there are p sub-modules, and therefore p capacitor assemblies (p is an integer), there will be p DC voltage curves for the MMC. While all DC voltage curves have the same general exponential shape, the DC voltage curves for each capacitor assembly CA1, CA2, ..., CA p They change slightly because the capacitances have slightly different capacitances and therefore charge slightly differently during the pre-charge process.

[0052] Each "step" (or "pulse") includes a charging period and a non-charging period. The transitions between "steps" (or "pulses") and between the charging and non-charging periods are regular and well-defined and depend on the frequency of the AC pre-charging current supplied from the power grid 40 during the pre-charging process. For example, if the frequency of the supplied AC pre-charging current is 50 Hz and the MMC submodules are implemented in a full-bridge topology, each "step" lasts 10 ms, and each charging and non-charging period lasts 5 ms. If the submodules are implemented in a half-bridge topology, each "step" lasts 20 ms, and each charging and non-charging period lasts 10 ms. The discrete "steps" are more clearly visible in Figure 8, which represents a relatively small portion of the DC voltage curve in Figure 7.

[0053] It will be appreciated that the non-charging period is an ideal time to obtain DC voltage measurements with minimal variation, as the voltage across each capacitor assembly remains substantially constant.

[0054] The pre-charging process (shown as "t start The initiation of the pre-charge process (labeled "t") may be determined, for example, by a close signal controlling the switches 44a, 44b, and 44c of the pre-charge circuit 36 ​​to close or by measuring the pre-charge current. For example, after the switches 44a, 44b, and 44c are controlled to close, the initiation of the pre-charge process may be determined by detecting a first peak of the pre-charge current supplied to the pre-charge circuit 36 ​​from the power grid 40. The pre-charge circuit 36 ​​may include one or more current sensors (not shown) for measuring the pre-charge current. Using a measurement of the pre-charge current is often a more accurate way of determining the initiation of the pre-charge process than using a close signal or some other determination when the switches 44a, 44b, and 44c are closed, and may be particularly important when DC voltage measurements taken during two or more pre-charge processes are compared for status monitoring purposes. The end of the pre-charge process (labeled "t" in FIG. 7) may be determined by detecting a first peak of the pre-charge current supplied to the pre-charge circuit 36 ​​from the power grid 40. end ”) is used to ensure that the voltage across the capacitor assembly is at a predetermined value for normal switching operation of the MMC (i.e., labeled “v rated The pre-charging process can typically be completed within a few seconds.

[0055] FIG. 9 illustrates discrete "steps" of DC voltage curves for multiple capacitor assemblies of the same MMC. While only 12 DC voltage curves are shown in FIG. 9 for clarity, it will be understood that in practice, significantly more DC voltage curves exist for a typical MMC. For example, if an MMC has at least 100 full-bridge or at least 200 half-bridge submodules, there may be at least 100 or 200 individual DC voltage curves (e.g., p = 100 or 200). The start and end of each non-charging period can be precisely determined based on the start time of the pre-charging process, such that one or more DC voltage measurements can be taken during a particular non-charging period, e.g., at a predetermined time from the start time corresponding to a particular non-charging period during which the DC voltage across the capacitor assemblies remains substantially constant, as shown in FIG. 9 . For example, a DC voltage measurement across each capacitor assembly can be taken at time t1 during the pre-charging process, where t1 corresponds to a particular non-charging period. Alternatively, two or more DC voltage measurements across each capacitor assembly may be taken during a narrow time window beginning at time t1 during the pre-charge process, where the time window coincides with a particular non-charging period, so that all measurements are taken with minimal variability and the two or more DC voltage measurements are used to determine an average voltage, thereby improving the reliability of the DC voltage measurements taken during the pre-charge process.

[0056] In FIG. 9, the time t1 at which the DC voltage measurement is taken is shown to be approximately in the middle of a particular non-charging period. p A first DC voltage measurement (e.g., v1) is taken for the capacitor assembly of the first sub-module at time t1, a second DC voltage measurement (e.g., v2) is taken for the capacitor assembly of the second sub-module at time t1, a third DC voltage measurement (e.g., v3) is taken for the capacitor assembly of the third sub-module at time t1, and so on, until a DC voltage measurement (e.g., v1) is taken for the capacitor assembly of the first sub-module at time t1, a second DC voltage measurement (e.g., v2) is taken for the capacitor assembly of the second sub-module at time t1, a third DC voltage measurement (e.g., v3) is taken for the capacitor assembly of the third sub-module at time t1, and so on.

[0057] At the end of a particular pre-charge process, the individual DC voltage measurements (i.e., v1, v2, v3, ..., v p ) is taken for each sub-module's capacitor assembly at time t1. Additional individual DC voltage measurements may also be taken for each sub-module's capacitor assembly at other times during the pre-charge process, e.g., at times t2, t3, etc., which are not shown but may be before or after time t1. Each time t2, t3, etc. may coincide with a different non-charging period.

[0058] The DC voltage measurements are preferably stored, for example, in memory. The stored DC voltage measurements are then processed to determine the status of each capacitor assembly. The DC voltage measurements may be processed by the controller 100 or may be transmitted to a remote server for processing. This processing does not need to be performed in "real time" and may occur after the pre-charge process has finished and the MMC has begun its normal operation.

[0059] Condition monitoring can be performed using DC voltage measurements taken from a single pre-charge process. For example, DC voltage measurements taken for each capacitor assembly at a particular time (e.g., time t1) can be compared. As shown in Figure 9, if there are p sub-modules and therefore p capacitor assemblies (p is an integer), DC voltage measurements v1, v2, v3, v4, ..., v5 taken simultaneously (e.g., at time t1) during the pre-charge process can be compared. p The measurement timing is determined by referring to the start time of the pre-charge process. This allows the DC voltage measurements v1, v2, v3, v4, ..., v for all sub-modules to be compared. p are ensured to be simultaneously acquired by the respective voltage sensors 34. In one example, DC voltage measurements v1, v2, v3, v4, ..., v are acquired at time t1. pmay be compared to the expected voltage for time t1. If an individual DC voltage measurement deviates from the expected voltage by more than a predetermined amount (e.g., more than ±x%), this may indicate that at least one of the capacitors in a particular capacitor assembly is not in the best condition and further testing may be performed. This may also be repeated for any DC voltage measurements taken at other times during the pre-charge process, e.g., at times t2, t3, etc.

[0060] In another example, DC voltage measurements v1, v2, v3, v4, ..., v taken at time t1 p Using the above equation, the average voltage for time t1 can be determined. If an individual DC voltage measurement deviates from the average voltage by more than a predetermined amount (e.g., more than ±x%), this can indicate that at least one of the capacitors in a particular capacitor assembly is not in top condition and further testing can be performed. This may also be repeated for any DC voltage measurements taken at other times during the pre-charge process, e.g., at times t2, t3, etc.

[0061] Due to manufacturing tolerances, the measured DC voltage of different capacitor assemblies can be expected to vary by up to several percent, even when the individual capacitors have the exact same design and nominal capacitance value. Condition monitoring here can be as simple as, for example, classifying any capacitor assemblies whose DC voltage measurements deviate from the expected or average voltage by less than a predetermined amount as having an acceptable condition and classifying any remaining capacitor assemblies as requiring further testing or monitoring. Alternatively, condition monitoring may depend on the amount by which individual DC voltage measurements deviate from the expected or average voltage, or may use two or more predetermined amounts associated with different deviations from the average or expected voltage. For example, if individual DC voltage measurements deviate from the average or expected voltage by more than a first predetermined amount (e.g., more than ±x1%) and less than a second predetermined amount (e.g., less than or equal to ±x2%, where x2 is greater than x1), this may indicate that at least one of the capacitors in a particular capacitor assembly is not in optimal condition and further testing may be performed. However, if an individual DC voltage measurement deviates from the average or expected voltage by more than a second predetermined amount (e.g., more than ±2%), this may indicate the need for more urgent action, including potentially replacing one or more of the capacitors of a particular capacitor assembly. Condition monitoring can be based on any suitable number of predetermined amounts that allow for the deviation of an individual DC voltage measurement of a particular capacitor assembly from the expected or average voltage to be classified and, if necessary, appropriate action taken in response to the indicated degradation.

[0062] Condition monitoring typically uses DC voltage measurements obtained from two or more pre-charge processes. For example, an initial pre-charge process can be performed to establish a "baseline" for the condition of the capacitor assemblies. The initial pre-charge process to establish the "baseline" may be performed during the first commissioning of the MMC. Subsequently, the same pre-charge process may be performed at intervals, such as every six months, every year, or every two to three years, as part of a routine maintenance program for the MMC, so that the condition of the capacitor assemblies can be periodically monitored for degradation. For example, if three identical pre-charge processes are performed—for example, an initial pre-charge process in year y0, a second pre-charge process one year later in year y1, and a third pre-charge process another year later in year y2—the stored DC voltage measurements for each pre-charge process can be used to determine the condition of each capacitor assembly. The same pre-charge circuit 36, i.e., with the same pre-charge resistors 42a, 42b, and 42c, is used for each pre-charge process. FIG. 10 shows discrete "steps" with DC voltage curves for the same capacitor assembly of an MMC taken in years y0, y1, and y2. The DC voltage measurement v1 taken at time t1 in year y1 can be compared to the DC voltage measurement v1 taken at time t1 in year y0 (e.g., when the MMC was first commissioned). The DC voltage measurement v1 taken at time t1 in year y2 can also be compared to the DC voltage measurement v1 taken at time t1 in year y0. DC voltage measurements taken at other times during each pre-charge process (e.g., times t2, t3, etc.) can also be similarly compared.

[0063] Similar comparisons can be performed for DC voltage measurements taken on other capacitor assemblies.

[0064] The time (or times) at which DC voltage measurements are taken is the same for each pre-charge process, and each time is determined based on the start time of the respective pre-charge process. This allows for direct comparison of DC voltage measurements taken for each individual capacitor assembly simultaneously during each pre-charge process (e.g., annually). Therefore, changes in DC voltage measurements that may indicate degradation of a particular capacitor assembly of the MMC can be identified. For example, Figure 10 shows an overall decrease in DC voltage measured across the capacitor assemblies over time, indicating that the capacitor assemblies are degrading at a particular rate.

[0065] Alternatively, DC voltage measurements taken for each capacitor assembly can be used to determine a characteristic of each capacitor assembly, which can then be used to determine the condition of each capacitor assembly. The characteristic can be the time constant of each capacitor assembly, for example, the time it takes for each capacitor assembly to charge to 63.2% of its rated voltage. The time constant of each capacitor assembly is known to be equal to the capacitance of the capacitor assembly multiplied by the total resistance of the pre-charge circuit 36. As a result, any change in the time constant determined during different pre-charge processes indicates a change in the capacitance of a particular capacitor assembly. The total resistance of the pre-charge circuit 36 ​​is the same for all capacitor assemblies being charged simultaneously. For example, for a particular capacitor assembly, the time constant τ1 determined in year y1 can be compared to the time constant τ1 determined in year y0 (e.g., when the MMC was first commissioned). The time constant τ1 determined in year y2 can also be compared to the time constant τ1 determined in year y0.

[0066] Any change in the DC voltage measured across a particular capacitor assembly simultaneously during different pre-charge processes (or any change in the determined time constant of a particular capacitor assembly) indicates a change in the capacitance of the capacitor assembly. It is important to note that it is not necessary to determine the absolute capacitance of the MMC's capacitor assembly, as other electrical parameters, such as the absolute value of the total resistance of the pre-charge circuit 36, also need to be known. In the example above, if a comparison of the respective DC voltage measurements v1 taken in year y1 and year y0 indicates that the DC voltage has decreased by 1%, this can be considered to indicate that the capacitance of the capacitor assembly of a particular sub-module has decreased by 1% over the course of one year. While this may not require immediate action, if the capacitance continues to decrease, it may exceed the allowable tolerance, and the capacitor assembly may be scheduled for replacement at an appropriate time in the future. Further testing can also be performed to try and identify whether only specific capacitors in the capacitor assembly are affected. This can avoid the need to replace all capacitors in a particular sub-module. This method can indicate whether the capacitance of a particular capacitor assembly has decreased by more than a threshold amount, which can optionally be determined based on the expected degradation rate of the capacitors utilized in the MMC and the time period between each pre-charging process. As described above, the measured DC voltage indicates capacitance. For example, the expected degradation rate of a capacitor's capacitance may be 0.5% per year. Thus, a 0.8% decrease in capacitance (as indicated by the DC voltage measured across a particular capacitor assembly) would not exceed the threshold amount if, for example, the time period between each pre-charging process is two years. This is because the capacitance decrease is only 0.4% per year. However, if each pre-charging process is only one year apart, the threshold amount would be exceeded because the capacitance decrease is 0.8% per year.

[0067] It will be appreciated that there may be some variation in electrical parameters when the pre-charge process is performed at different times, e.g., annually. Although the same pre-charge circuit 36 ​​with the same pre-charge resistors 42a, 42b, and 42c is used for each pre-charge of the MMC, there may still be some variation in the absolute value of the total resistance of the pre-charge circuit 36, for example, as a result of differences in ambient temperature. There may also be some variation in the AC pre-charge current supplied by the power grid 40. However, such variation may be considered to affect the DC voltage measurements for each capacitor assembly in the same way and, therefore, can be corrected or compensated for using an appropriate calibration or compensation process. For example, because the total resistance of the pre-charge circuit 36 ​​is the same for all capacitor assemblies being charged simultaneously, variations in the total resistance when each pre-charge process is performed will have the same effect on the DC voltage measured across all capacitor assemblies. In other words, the disturbances eliminated by the calibration or compensation process are “common-mode” disturbances. A calibration or compensation process may be performed on the stored DC voltage measurements taken during a pre-charge process before a comparison is made with stored DC voltage measurements taken during a previous pre-charge process. Any suitable calibration or compensation process may be used.

[0068] The calibrated or compensated DC voltage measurements obtained after performing the calibration or compensation process may also be stored and then used in any comparisons made with DC voltage measurements obtained in subsequent pre-charge processes. [Explanation of symbols]

[0069] 1A modular multilevel converter (MMC), 1B MMC, 1C MMC, 21 first converter leg, 22 second converter leg, 23 third converter leg, 4 first direct current (DC) bus, 6 second DC bus, 81 series-connected submodules, 82 series-connected submodules, 8 nSeries-connected submodules, 101 Series-connected submodules, 102 Series-connected submodules, 10 n Series-connected sub-modules, 12, first alternating current (AC) bus, 14, second AC bus, 16, third AC bus, 181, first converter arm, 182, second converter arm, 183, third converter arm, 20, first AC bus, 22, second AC bus, 24, third AC bus, 261, series-connected sub-modules, 262, series-connected sub-modules, 26 n Series-connected submodules, 28A converter arm, 28B converter arm, 301 series-connected submodules, 302 series-connected submodules, 30 n Series-connected submodules, 321Series-connected submodules, 322Series-connected submodules, 32 n Series-connected submodules, 34 voltage sensor, 36 pre-charge circuit, 38a first AC terminal, 38b second AC terminal, 38c third AC terminal, 40 three-phase AC power grid, AC power source, 42a first pre-charge resistor, 42b second pre-charge resistor, 42c third pre-charge resistor, 44a first switch, 44b second switch, 44c third switch, 46a bypass switch, 46b bypass switch, 46c bypass switch, 100 controller, A AC phase, B AC phase, C AC phase, D1 diode, D2 diode, D3 diode, D4 ​​diode, S1 semiconductor switch, S2 semiconductor switch, S3 semiconductor switch, S4 semiconductor switch, t1 time, t2 time, t3 time, v1 DC voltage measurement value, v2 DC voltage measurement value, v3 DC voltage measurement value, v4 DC voltage measurement value, v n DC voltage measurement, v p DC voltage measurement, v 3n DC voltage measurement, v 6n DC voltage measurement, v (3n-1) DC voltage measurement, v (6n-1)DC voltage measurements, y0 year, y1 year, y2 year, CA1 capacitor assembly, energy storage device assembly, CA2 capacitor assembly, energy storage device assembly, CA3 capacitor assembly, CA n Capacitor Assembly, Energy Storage Device Assembly, CA p Capacitor Assembly, CA 3n Capacitor Assembly, CA 6n Capacitor Assembly, CA (3n-1) Capacitor Assembly, CA (6n-1) Capacitor assembly, τ1 time constant, τ2 time constant, τ n time constant

Claims

1. Energy storage device assembly (CA) of modular multilevel converter MMC (1A, 1B, 1C) １ , C.A. ２ , . . . , CA ｎ ) monitoring the state of the MMC (1A, 1B, 1C), wherein the MMC (1A, 1B, 1C) comprises a plurality of serially connected sub-modules (30 １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ ), and each sub-module (30 １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ ) comprises a plurality of semiconductor devices, each semiconductor device having at least one controllable semiconductor switch (S １ , S ２ , S １ , S ２ ,... ,S ４ ), and an energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) includes one or more individual energy storage devices electrically connected in parallel, the MMC (1A, 1B, 1C) further comprising at least one alternating current AC terminal (38a, 38b, 38c) electrically connectable to an AC power source (40), the method comprising: The individual energy storage device assemblies (CA １ , C.A. ２ , . . . , CA ｎ During a pre-charging process in which each energy storage device assembly (CA) is initially charged to a rated voltage by an AC pre-charging current supplied from the AC power source (40) to the at least one AC terminal (38a, 38b, 38c), each energy storage device assembly (CA) is １ , C.A. ２ , . . . , CA ｎ ) one or more measurements of the direct current DC voltage (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) and The one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) to each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) state and A method comprising:

2. the one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) is a part of each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) during one or more non-charging periods during which the voltage across each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ The method of claim 1 , wherein the .times. ...

3. Two or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) during the non-charging period, each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) and an average value of the DC voltage is determined for the non-charging period.

4. The method of claim 3 , wherein the initiation of the pre-charge process is determined by measuring the AC pre-charge current.

5. The individual energy storage device assemblies (CA １ , C.A. ２ , . . . , CA ｎ During a second pre-charging process in which each energy storage device assembly (CA) is initially charged to a rated voltage by an AC pre-charging current supplied from the AC power source (40) to the at least one AC terminal (38a, 38b, 38c), １ , C.A. ２ , . . . , CA ｎ ) one or more measurements of the DC voltage across １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) and The one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) and the one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) to each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) state and The method of claim 4 further comprising:

6. Specific Energy Storage Device Assembly (CA １ , C.A. ２ , . . . , CA ｎ 6. The method of claim 5, further comprising indicating whether the capacitance of the capacitance sensor has decreased by more than a threshold amount.

7. The method of claim 6 , wherein the threshold amount is determined based on an expected rate of deterioration of the energy storage device and a time period between the first and second pre-charging processes.

8. The DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） 8. The method of claim 5, further comprising performing a calibration or compensation process on the

9. MMC (1A, 1B, 1C), A plurality of serially connected sub-modules (30 １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ ), each sub-module (30 １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ )teeth, a plurality of semiconductor devices, each semiconductor device comprising at least a controllable semiconductor switch (S １ , S ２ , S １ , S ２ ,... ,S ４ ), and a plurality of semiconductor devices including An energy storage device assembly (CA) comprising one or more individual energy storage devices electrically connected in parallel １ , C.A. ２ , . . . , CA ｎ ) a plurality of series-connected sub-modules (30) １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ )and, at least one AC terminal (38a, 38b, 38c) electrically connectable to an AC power source (40); A condition monitoring device, comprising: At one or more times between the start of the pre-charging process and the end of the pre-charging process, each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) one or more measurements of the DC voltage across １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) of the MMC (1A, 1B, 1C) during the pre-charging process. １ , C.A. ２ , . . . , CA ｎ ) is initially charged to a rated voltage by an AC pre-charge current supplied from said AC power source (40) to said at least one AC terminal (38a, 38b, 38c); the one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) to each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) state and condition monitoring devices adapted for MMC (1A, 1B, 1C) comprising:

10. The condition monitoring device At one or more times between the start of the second pre-charging process and the end of the second pre-charging process, １ , C.A. ２ , . . . , CA ｎ ) one or more measurements of the DC voltage across １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) of the MMC (1A, 1B, 1C) during the second pre-charging process. １ , C.A. ２ , . . . , CA ｎ ) is initially charged to a rated voltage by an AC pre-charge current supplied from said AC power source (40) to said at least one AC terminal (38a, 38b, 38c); The one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) and the one or more DC voltage measurements (v １ , v ２ , v ３ , v ４ , v ｎ , v ｐ , v ３ｎ , v ６ｎ , v （３ｎ－１） , v （６ｎ－１） ) to each energy storage device assembly (CA １ , C.A. ２ , . . . , CA ｎ ) state 10. The MMC (1A, 1B, 1C) of claim 9, further adapted to:

11. Each semiconductor device is an anti-parallel connected diode (D １ , D ２ , D １ , D ２ , . . . , D ４ The MMC (1A, 1B, 1C) of claim 10 further comprising:

12. The plurality of series-connected sub-modules (30 １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ ) is a converter arm (18) of the MMC (1A, 1B, 1C). １ , 18 ２ , 18 ３ ), and the converter arm (18 １ , 18 ２ , 18 ３ 12. The MMC (1A, 1B, 1C) of claim 11, further comprising at least one inductor.

13. Each sub-module (30 １ , 30 ２ ,... ,30 ｎ ) has a full-bridge topology, or each sub-module (32 １ , 32 ２ ,... ,32 ｎ 13. The MMC (1A, 1B, 1C) according to claim 12, wherein the MMC (1A, 1B, 1C) has a half-bridge topology.

14. Each sub-module (30 １ , 30 ２ ,... ,30 ｎ , 32 １ , 32 ２ ,... ,32 ｎ 2. The MMC (1A, 1B, 1C) of claim 1, wherein the MMC (1A, 1B, 1C) includes a DC voltage measurement unit (34).

15. 15. The MMC (1A, 1B, 1C) of claim 9, further comprising a pre-charging circuit (36) including at least one pre-charging resistor (42a, 42b, 42c) and at least one switch (44a, 44b, 44c), wherein the at least one AC terminal (38a, 38b, 38c) of the MMC (1A, 1B, 1C) is electrically connected to the pre-charging circuit (36), and the pre-charging circuit (36) is electrically connectable to the AC power source (40).

Images