Search

Modularity and high reliability from redundancy are the two attractive advantages of modular multilevel converters (MMCs). This paper elaborates a switch open-circuit fault diagnosis and a fault-tolerant operation scheme for MMCs with distributed control. The proposed fault diagnosis and fault-tolerant control method can significantly improve the reliability of the MMC while maintaining the modularity of its software implementation. By distributing fault diagnosis into submodules, its local controller is capable of identifying the switching devices in open-circuit fault without extra hardware circuitry. Based on the real-time measurements of submodule terminal voltage and arm current, single, or multiple faulty switches can be identified within 3.5 ms without triggering faulty alarms. Furthermore, a new fault-tolerant operation is proposed to maintain the output current, internal dynamics, and switching harmonics unchanged after the faulty submodule is bypassed. This is achieved by resetting the period and phase registers in the local controller according to the information of bypassed submodules. The control loops of the MMC are not influenced by the proposed fault diagnosis and fault-tolerant operation, making the operation transition seamless and reliable. Experimental results show that fault identification and system reconfiguration can be completed within 5 ms, and the MMC can seamlessly and smooth ride through the switch open-circuit faults without severe malfunction and catastrophic damages. [ABSTRACT FROM AUTHOR]

Modular multilevel converters (MMCs) in high-voltage dc applications usually adopt a distributed control architecture to manage a large number of submodules (SMs) through a communication network. The communication congestion and network disconnection might lead to communication interruption (CI) and eventually cause the system to malfunction. In this article, a resilient operation strategy is proposed and studied to ride-through the CI fault, in order to prevent frequent fault SM bypassing, replacement, or even system shutdown. The analysis of the MMC distributed control system with the presence of CI indicates that the insertion index of the faulted SM might become constant, which distorts the output current and results in overvoltage of the communication interrupted SM (CI-SM). The CI-SM capacitor voltage prediction can be used to determine the MMC safe operation period after CI occurs. During the safe operation period, the CI-SM power balance is sustained by utilizing prestored phase signals to generate a sinusoidal insertion index according to its capacitor voltage tracking error. Two operation modes are proposed and analyzed to ensure the MMC stable operation under various conditions. The system protection is sensibly used only if the CI duration exceeds a safe operation period, which avoids frequent SM cut-off. Good agreement of the CI-SM capacitor voltage is achieved between the theoretical and simulation results. The effectiveness and robustness of the proposed MMC resilient operation are experimentally confirmed. [ABSTRACT FROM AUTHOR]