Fault Tolerant Control of Five-Level Inverter Based on Redundancy Space Vector Optimization and Topology Reconfigruation

Abstract

Compared with two-level and three-level inverters, five-level inverters have more redundant space vectors to implement flexible control strategy. A fault tolerance method for open circuit fault of power tube in Neutral Point Clamped (NPC) five level inverter is proposed. The open circuit fault of inverter is divided into type Ⅰ and type Ⅱ in this paper. Redundant space vector optimization and topological reconfiguration fault tolerance strategies are used for type I and type II failures, respectively. The basic vector of the synthetic reference vector Uref is reoptimize according to the redundant vector, and get a new control strategy after the decision. This method enables the inverter to achieve fault-tolerant operation without modifying hardware or adding bridge arms. The output performance of the inverter after fault is guaranteed and the cost of fault tolerance is greatly reduced. The feasibility and effectiveness of the proposed fault-tolerant method are verified through experiments.

Redundancy Management Strategy for a Silicon/Silicon-Carbide Hybrid Switch-Based Modular Inverter System

“…In Khanaki et al, 2019, the author integrated a single inverter into an extra low-voltage (ELV) battery module and modularized it and compared the reliability of the 48 V battery module with different redundancy strategies through experiments. In Yu et al, 2020, a five-level inverter single-tube open-circuit fault tolerance method based on redundant space vector optimization and topology reconstruction was proposed, and a new fault circuit space vector pulse width modulation control strategy was obtained, but the maximum output voltage of the inverter will be reduced to 3/4 of the normal inverter voltage. For many types of multilevel circuits, the voltage will be reduced to some extent after the failure occurs (Lee and Lee, 2016;Lee and Lai, 2021;Guan and Liu, 1109).…”

Fault Analysis and Non-Redundant Fault Tolerance in 3-Level Double Conversion UPS Systems Using Finite-Control-Set Model Predictive Control

“…On the other hand, a complete correction of the fault is not always possible. This type of approach can be based solely on software adaptations, relying only on changes to the control and modulation strategies [21][22][23][24][25][26][27][28][29][30], or require some changes to the converter topology . Software-based fault correction presents the lowest possible complexity and cost but typically carries significant operational limitations, resulting in reduced post-fault performance.…”

“…In the case of inner IGBT faults, the effect of the fault can only be attenuated [23][24][25] since a large number of switching states is lost. In DC/AC applications, IGBT faults have a more severe impact on the NPC converter and cannot be fully corrected without topology changes, causing a derating [30].…”

“…This solution requires only one additional bidirectional switch per phase, making it a very cost-effective solution. This solution has been widely used in both 2-level [37,38,[40][41][42][43][44][45][46][63][64][65][66] and NPC [29][30][31][47][48][49][50][51][52][53][54][55] converters. A similar approach has been proposed for 2-level converters in [61,62], connecting the faulty phase to an auxiliary capacitor instead of the DC bus midpoint.…”

“…With these approaches, the controller can only operate normally with modulation indexes below 0.5, which implies a derating of the system. Even though a derating may be acceptable in some applications [30], such as motor drives [37,40,51,54,67] or generation [38,46,47,53,63,64], it is not viable in UPS applications since the system cannot compromise or restrict the operation of the protected critical load.…”