Auxiliary Submodule Power Supply for a Medium Voltage Modular Multilevel Converter

During the operation of a three line ac link in parallel with a ac-dc-ac link system that employs Back-to-Back (B2B)Medium Voltage Modular Multilevel Converters (MMC), a path for the low frequency zero sequence currents is shown to exist. Apart from causing additional losses, these currents can result in limitations on power capacity. A detailed analysis on the influence of cell capacitance, arm inductance and link conductor length on the circulating zero sequence currents and the means to mitigate them through a simulated control strategy is investigated. The simulations show that the dc link employing B2B-MMC can successfully operate in parallel with ac link without transformer isolation by controlling the zero sequence currents.

Modular Multilevel Converters have been proposed to be a viable topology for medium voltage and high power motor drives due to the stackable structure of the topology. This allows for a transformer-less drive, better output current THD and reduces input and output filter sizes compared to traditional topologies. Replacing the traditional Si IGBT based submodules with SiC MOSFETs has been shown to improve the overall efficiency of the MMC. However, the lower switching losses of SiC MOSFETs have not been fully utilized to realize further system level improvements in the MMC. In this paper, the impact of the SiC MOSFETs on the efficiency, power density and output current THD of the drive is compared against Si IGBTs using analytical and simulation. A 7 kV dc bus, 4.16 kV ac, 1 MVA, 0-1,000 Hz medium voltage variable speed drive based on 1.7 kV devices based is used for the comparison.

We designed an auxiliary power supply for medium voltage MMC-converters with system voltages up to 36 kV rms . The system supplies up to 250 W power and is based on a two-stage design. The first stage is an inductive power transfer system that isolates against the system voltage. It is connected to a low voltage DC-supply on ground potential, and shows DC-DC efficiency between 90–95%. Two distinct insulation concepts for the coupled coil design have been designed and passed partial discharge tests. The second stage supplies the various gate and control units of each MMC power cell. Its insulation requirement is given by the cell voltage. Functional and insulation tests as well as a sensitivity analysis are included. The full auxiliary power supply shows an efficiency of 86%.

Energizing large transformers often cannot be done direct-on-line due to the negative effects of the inrush current. The typical schemes used in the past are energization via a high impedance, or via a tertiary winding connected to an auxiliary AC power source. These schemes require additional equipment resulting in increased foot print and complexity. For offshore and subsea installations, the increase in foot print often greatly exceeds the cost of the additional equipment. This paper presents an alternative solution allowing direct-on-line energization of large transformers, both topsides and subsea. Different solutions for subsea auxiliary AC power supplies not requiring high-voltage circuit-breakers are presented. For offshore and subsea power systems, the circuitbreakers supplying power to the transformers are standard 3.3 kV to 36 kV class devices having a 3-pole operating mechanism. The solution presented allows circuitbreakers to be switched such that the closing of the poles occurs at the point on the voltage waveform where the resulting inrush current is the least. For subsea power distribution systems, the auxiliary power required for the subsea control equipment is provided by a high-voltage DC auxiliary power link from the shore station and the control and communication link is via optical fiber. The solutions presented use standard proven technology and can be integrated within the subsea modules required for supplying power to the loads. This keeps the number of penetrators and subsea connector systems to a minimum. The importance of redundancy and maintenance in obtaining and keeping the required system availability are discussed.

The modular multilevel converter (MMC) has become the most attractive converter technology for medium/high-power applications, specifically for high-voltage dc (HVdc) transmission systems. One of the technical challenges associated with the operation and control of the MMC-based system is to precharge the submodule (SM) capacitors to their nominal voltage during the startup process. In this paper, considering various SM circuits, a generalized precharging strategy is proposed for the MMC-based systems, which can implement soft startup from the dc or ac side. The proposed precharging strategy can be applicable for various SM circuits and MMC configurations. The proposed startup strategy does not require extra measurements and/or auxiliary power supplies. The charging current is controlled by adjusting the changing rate of the number of blocked and bypassed SM capacitors. Based on the proposed startup strategy, the startup processes of MMC/MMC-HVdc systems with various SM circuits are analyzed and a generalized startup procedure for various MMC-HVdc systems is proposed. In addition, the uncontrollable steady-state SM capacitor voltages of various MMC-based systems are analyzed and the associated precharging time is also investigated. Performance of the proposed strategy for various MMC-HVdc systems is evaluated based on time-domain simulation studies in the PSCAD/EMTDC software environment and experimental results are based on a scaled-down prototype.

In the DC power distribution system which needs to supply power for the passive network, because of the modular multilevel converter (MMC) connected to the passive network has no power supply, the MMC control mode can't directly be set to the constant voltage when to start MMC. Accordingly, the particular start-up control strategy must be adopted. This paper presents a novel start-up approach to the DC distribution network. Firstly, the process of the start-up is divided into an uncontrolled charging stage and a controllable charging stage. By placing a current limiting resistor in MMC AC-side, designing the AC-side and DC-side circuit breaker closing sequence of the converter station, and setting the reference voltage by the slope, the start-up of the DC distribution network can be achieved no large impact current. Secondly, by setting one end of the converter station as the master station and designing the control mode of connecting the passive network converter station, the local and the remote MMC sub-module capacitors can be charged at the same time, the start-up of the DC distribution network can be completed without the help of another auxiliary power supply or equipment. Finally, the mathematical model is established by Matlab/Simulink to verify the effectiveness of the proposed method. The studies we have performed showed that the designed start-up strategy is simple, easy to implement, reliable and economical.

The power conversion system for a battery-energy storage system typically employs a conventional voltage-source converter with battery strings directly connected to the dc bus. This system configuration presents several issues, such as limited efficiency of two-level converter systems and the limited reliability associated with the use of long battery strings. This paper examines three viable multilevel converter solutions for integrating battery energy storage that offer the potential for enhanced efficiency and reliability. These solutions are the modular multilevel converter (MMLC) with battery energy storage distributed into its submodules, the cascaded converter, and the MMLC with battery energy storage centralized on its dc link. The three systems are compared in terms of efficiency, reliability, and module redundancy. It is determined that the MMLC with distributed battery energy storage must operate differently from conventional MMLC systems. Its operation is therefore studied in detail and validated through simulation to demonstrate its suitability for distributed energy-storage integration. The analysis shows that the MMLC with distributed battery energy storage requires the largest number of semiconductor devices for a given power level; however, it also provides the most efficient, reliable, and versatile solution of energy-storage integration.

In this paper, the method and a control algorithm for detecting resonant destruction of a resonant auxiliary power supply for railway vehicle are proposed. It is essential to reduce the weight of the electric equipment attached to the lower part of the railway vehicle because it greatly affects the life of the railway vehicle, such as wheel and bearing. In order to reduce the weight and volume of the auxiliary power supply, in the case of windings, high-speed switching should be performed by lowering the input voltage. Therefore, in this paper, the auxiliary power supply control algorithm using an LLC resonant converter to reduce the weight of auxiliary power supply for railway vehicle is proposed, and the method of detecting resonant destruction due to device failure of a resonant converter is proposed. The proposed algorithm steps down the wire voltage using an input buck converter for high-speed switching, and the inverter input voltage is controlled through the LLC resonant converter. In addition, ZVS (Zero Voltage Switching) is operated to minimize the loss of the resonant converter, and the optimal design of the resonant tank is also proposed. In order to detect resonant destruction, a detection method formulated by analyzing the resonant current peak value for the load capacity is proposed. The algorithm proposed in this paper is verified by simulation and experiment.

Resonant converters are being used more and more widely in power supply system, and are potentially adopted as the one part of traction power electronics transformer or auxiliary power supply in locomotives. Series resonant converters (SRCs), which have been researched in many aspects, is relatively focused earlier. In railway transportation locomotive, the output voltage of single-phase rectifiers contains double line ripple component (100Hz or 120Hz), which effects the power quality of the output voltage in auxiliary power converters. This paper focuses on a combination circuit consisting of a Boost converter and a SRC converter in order to reduce the output voltage ripple with a simple control method. The parameters of the DC-DC converter are designed in detail and the coefficients are deduced by small-signal modeling. Detailed analysis about the harmonics distribution and small-signal modeling are shown in this paper, while the input voltage feedforward scheme is adopted to reduce the influence of input voltage ripple and control system of the combination circuit is also presented. Finally, the theoretical analysis is well verified by the hardware-in-loop experiment results.

At present, in the typical design of China’s nuclear power plant, main generation system is connected to the power grid by 500 kV system. 500 kV system as a priority power source, 220kV system as an auxiliary power source. Independent operation of 500kV and 220kV system, improved the reliability of power supply of nuclear power plant. However, the DC 220V power system used to control the 500kV and 220kV system in the switch station of partial nuclear power project is not independently configured, and the design form of one set of DC system is used in the transformer station. In recent years, there are many accidents that AC power enters into the DC power system, resulted in the loss of power source in the transformer station. The loss of external power source in the whole plant is very significant. In this paper, the influence of AC power entering into DC power system on relay protection device is analyzed, the measures to prevent the AC power into DC power system are discussed, the necessity of independent configuration of DC control power system for the 500kV priority power system and 220kV auxiliary power supply system is analyzed.

The series-connected offshore wind energy conversion system (WECS) removes the need for bulky and costly offshore substations. In this system, a current-fed bidirectional isolated DC-DC converter such as the hybrid dual active bridge (HDAB), is essential to supply power from the onshore grid to the wind farm’s local load when there is no wind. However, the reverse-blocking bridge arms on the current-source (CS) side of the HDAB remain off if the auxiliary power supply (APS) is unavailable, preventing the WECS from starting when wind turbines are idle. Moreover, a clamp circuit is needed on the CS side to suppress voltage spikes. This paper presents an improved clamp circuit configuration for HDAB, integrating the CS-side APS with the clamp circuit to provide charging path from the CS-side bus to the APS input. Additionally, a CS-side start-up strategy for HDAB is proposed. The proposed method suppresses voltage spikes and make it possible to power up the APS by the CS-side bus, even without a power source on the voltage side. Simulation and experimental results verify the correctness of the proposed method.

The Austin SuperMOS is an intelligent 7.2kV/60A SiC power switch integrated with a gate driver and an isolated auxiliary power supply. Design considerations for the auxiliary power supply in a 5kV Austin SuperMOS based full bridge power electronic building block (PEBB), in terms of electrical performance and insulation against high voltages, are presented in this paper. In the presented auxiliary power supply, an LLC resonant converter along with 4 individual floated voltage sources including high frequency and high-voltage isolation transformers are developed. The design of individual floating sources is based on circuit and FEM simulations as well as necessary engineering calculation. Partial discharge measurements are also conducted to validate the proposed high voltage transformer design. Electrical measurements performed on the fabricated full bridge PEBB reveal stable and proper operation of the system.

The auxiliary power supply with high voltage isolation and low common-mode capacitance is one of the critical components inside medium-voltage power electronic converters. The isolation transformer design often ends up having a high leakage inductance as a consequence of the low isolation capacitance requirement imposed by the high switching voltage slew rate <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\mathbf{d}\boldsymbol{v}/\mathbf{d}\boldsymbol{t})$</tex> . In the case of series resonant converter this leakage inductance can be compensated by series-resonant capacitor. However, the winding resistance in the transformer still leads to increased voltage drop resulting power limitation. The objective of this paper is to deduce the time domain modelling of series-resonant converter considering isolation transformer winding resistor to analyze the trade-off between transformer winding resistance and voltage-drop.

Low device switching frequency is preferred in high-power converters to achieve higher efficiency, higher device utilization, and feasible cooling system design. However, low device switching frequency operation leads to higher peak voltage in submodule (SM) capacitors of medium-voltage modular multilevel converters (MMC). The main reason is unequal distribution of energy among SMs in each fundamental period due to fewer switching commutations available for balancing. This results in increased converter cost due to higher voltage rating of SM capacitors and semiconductor devices. Therefore, the main objective of this paper is to propose a new SM selection algorithm to reduce the peak voltage of SM capacitors without significantly increasing the device switching frequency. The main idea is to avoid unnecessary switching commutations by ensuring balancing of SM capacitor voltages only at the end of charging period. The proposed algorithm has been verified with extensive simulation studies of a 5L-MMC and it has been observed that peak voltage is kept within 10 % of the nominal value while limiting the device switching frequency in the range of 350 to 400 Hz.

An issue of stand-off energy sources is in an early stage of development. Several concepts have been identified as potential solution of the problem. Those based on the total disconnection between the target assembly and the primary energy source have an obvious advantage in that they allow one to relatively easily protect the permanent part. Indeed, a fast projectile travelling at the velocity of 10 km/s covers the distance of 10 m in 1 ms, the time that is sufficient to mechanically shield the line of site. Auxiliary power supply in the form of an electron beam can be protected by using a magnetic wiggling in the transport channel in the permanent part of the facility. Of some help is also the fact that this auxiliary source operates 10 or so microseconds before the fusion energy release occurs. Another advantage of this approach is related to its compatibility with high rep-rate mode, up to tens pulses per second (because there is no need to insert heavy large-volume parts into the reaction chamber). An obvious disadvantage is that the assembly should contain a more or less complex on-board circuitry. Those concepts that are based on the direct mechanical connection between the external pulse-forming line and the disposable inner part of the transmission line (connected in turn to the Z-pinch diode) have an obvious advantage in eliminating any intermediate steps in delivering the energy to the pinch. They work essentially as the presently used Z-pinch devices and should provide the same high efficiency. A difficulty of this approach is related to the presence of the insulating slot in the walls of the reaction chamber that directly links the disposable inner part of the energy transmission system and permanent external energy source. The slot is vulnerable to mechanical perturbations and neutron irradiation that may propagate into the permanent part of the source and cause a damage to it. Mechanical damage could be reduced by the technique of hydrodynamic channeling or hydrodynamic lensing mentioned in Sets 3 and 4. Those concept that are based on the total disconnection between the target assembly and the primary energy source have an obvious advantage in that they allow one to relatively easily protect the permanent part. Indeed, a fast projectile travelling at the velocity of 10 km/s covers the distance of 10 m in 1 ms, the time that is sufficient to mechanically shield the line of site. Auxiliary power supply in the form of an electron beam can be protected by using a magnetic wiggling in the transport channel in the permanent part of the facility. Of some help is also the fact that this auxiliary source operates 10 or so microseconds before the fusion energy release occurs. Another advantage of this approach is related to its compatibility with high rep-rate mode, up to tens pulses per second (because there is no need to insert heavy large-volume parts into the reaction chamber). An obvious disadvantage is that the assembly should contain a more or less complex on-board circuitry.