Direct torque control (DTC) has been widely used as an alternative to traditional field-oriented control (FOC) methods for three-phase drives. The conventional DTC scheme has been successfully extended to multiphase drives in recent times, using hysteresis regulators to independently track the desired torque and flux in symmetrical five-phase induction machines (IMs). The fault-tolerant capability of multiphase drives is an interesting intrinsic advantage for safety-critical applications, where recent research has demonstrated the effectiveness of FOC schemes to perform ripple-free postfault operation. In spite of the utility of DTC methods in normal operation of the multiphase machine, no extension to manage the postfault operation of the drive is found in the literature. In this paper, a novel fault-tolerant DTC scheme is presented. The performance of the proposed method is experimentally validated in a five-phase IM drive considering an open-phase fault condition. Provided tests analyze steady and transient states, including the transition from pre- to postfault operation. Obtained results prove the interest of the proposal, which ensures the open-phase fault-tolerant capability of DTC-controlled five-phase IM drives.

Multiphase machines are well known for their fault-tolerant capability. Star-connected multiphase machines have fault tolerance in an open circuit. For an inverter switch short-circuit fault, it is possible to keep a smooth torque of a permanent magnet synchronous machine if the currents of the faulty phases are determined and their values are acceptable. This paper investigates fault-tolerant operations of an open-end five-phase drive, i.e., a multiphase machine fed with a dual-inverter supply. Inverter switch short-circuit fault is considered and handled with a simple solution. Original theoretical developments are presented. Simulation and experimental results validate the proposed strategy.

This paper deals with easy-to-implement control strategies when a seven-phase axial flux permanent magnet machine supplied by a seven-leg voltage source inverter is in fault operation mode. Using a vectorial multimachine description, a seven-phase machine presenting a heightened ability to be controlled with one or two open-circuited phases has been designed. The machine is first presented, and experimental results are provided when one or two phases are open circuited. Based on a vectorial approach, new current references are calculated to avoid high-torque ripples.

This paper deals with the generation of optimal current references for multiphase permanent-magnet (PM) synchronous machines in normal or fault mode (open-circuited phases). Current references are computed in order to keep a constant torque while minimizing instantaneous joule losses. In comparison with commonly used scalar methods, a vectorial approach makes it possible to reduce the number of computations in order to generate optimal current references in real time. In addition to this, since current references are expressed in terms of physical parameters of the machine, this approach can be used to evaluate the influence of machine parameters over the control performances. Finally, experimental results of a surface-mounted PM five-phase synchronous machine are provided in order to demonstrate the proposed strategy.

AC machines supplied in parallel by a common voltage source inverter are sometimes used in industrial drives and railway traction applications. This reduction of power and control electronics leads to a reduction of cost, weight, and dimensions that is very useful for on-board systems. But this common supply imposes common voltages to all machines and the possibilities of independent behavior are reduced. In this paper, a graphical modeling is suggested to model such parallel drives for a railway application. A weighted control is then systematically deduced from this modeling and different possibilities of control are highlighted. Experimental results are provided to compare two of the deduced controls

The performance of a dual energy storage electric vehicle system mainly depends on the quality of its power and energy managements. A real-time management strategy supported by a model predictive control (MPC) using the nonuniform sampling time concept is developed and fully addressed in this paper. First, the overall multiple energy storage powertrain model including its inner control layer is represented with the energetic macroscopic representation and used to introduce the energy strategy level. The model of the system with its inner control layer is translated into the state-space domain in order to develop an MPC approach. The management algorithm based on mixed short- and long-term predictions is compared to rule-based and constant sampling time MPC strategies in order to assess its performance and its ability to be used in a real vehicle. The real-time simulation results indicate that, compared to other strategies, the proposed MPC strategy can balance the power and the energy of the dual energy storage system more effectively, and reduce the stress on batteries. Moreover, battery and supercapacitor key variables are kept within safety limits, increasing the lifetime of the overall system.