Manuscript received 29 April, 2010; revised 6 July, 2010 and 24 September, 2010.The authors are with the Department of Electronic and Electrical Engineering, The University of Sheffield, Sheffield S1 3JD U. K. (email: k.hoang@sheffield.ac.uk; k.d.hoang@ieee.org; z.q.zhu@sheffield.ac.uk; z.q.zhu@ieee.org; m.p.foster@sheffield.ac.uk).Digital Object Identifier I. INTRODUCTIONIRECT torque and indirect torque control techniques are two common methodologies for controlling permanent magnet brushless AC machine [1], [2]. Normally, a sixswitch three-phase (SSTP) inverter [Fig. 1(a)] is used for high performance operation of a three-phase machine and is almost universally considered the industry standard. For economic reasons, however, reducing the cost of the inverter is still under investigation and one obvious way to achieve this aim is to decrease the number of inverter switching devices.Low cost inverter topologies with reduced number of switching devices for an induction machine drive system has been suggested and demonstrated in [3] and [4]. Welchko et al. [3] proposed using a three-switch three-phase inverter with an extra connection from neutral point to DC link mid-point to control torque and speed of an induction machine. Although this reduces the number of active switching devices, it requires modifications of both DC link and a specialized machine stator winding structure. A further disadvantage of this topology is that three phase currents are unidirectional and, hence, this topology is limited to particular applications. In [4] a fourswitch three-phase (FSTP) inverter [Fig. 1(b)] was presented where one of three phase machine terminals was connected to the DC link mid-point and control achieved by manipulating the voltages and currents of the two active phases. The performance of this inverter was comparable to the traditional SSTP inverter. However, machines controlled by a FSTP inverter could only achieve half rated speed due to the fact that the voltage vector value is decreased by a factor of two in comparison with that in a conventional SSTP inverter drive system [5]. It should be noted that for an induction machine 2 drive system, reconfiguration of the windings from star to delta connection can allow full-speed operation [4]. Reduction in high order harmonic currents flowing in the mid-point connected phase current can be seen as a further benefit of the FSTP inverter topology. Other investigations on the FSTP inverter topology include the minimisation of torque ripple by applying space vector modulation schemes [6], the application of pulse width modulation (PWM) control methodologies to improve scalar PWM performance and DC link imbalance [7], and the elimination of current distortion at low-speed operation caused by the limited values of DC link capacitors using a compensation strategy [8]; it was also demonstrated that current waveforms under high-speed and inertial load conditions remain naturally balanced and sinusoidal without compensation [8].Reliability improvement and fault tolerance...

. et al.(1 more author) (2015) Modified switching-table strategy for reduction of current harmonics in direct torque controlled dualthree-phase permanent magnet synchronous machine drives. IET Electric Power Applications, 9 (1).

In this paper, an online control method of interior permanent magnet synchronous machine (IPMSM) drives for traction applications considering machine parameter and inverter nonlinearities is presented. It is shown that the conventional technique using parameter information instantly extracted from premeasured parameter look-up tables (LUTs) only determines the local MTPA operating point associated with this specific parameter information without evaluating the global MTPA achievement. Therefore, global MTPA operation may not be achieved for conventional online control IPMSM drives with extreme nonlinear machine parameters (e.g. short-period overload operations). Thus, a model-based correction method using stator flux adjustment is proposed for an online quasiglobal MTPA achievement. It is also proven that in the fluxweakening region, due to the inverter nonlinearities, a lower than expected maximum achievable torque for a demanded speed and a higher than expected current magnitude for a demanded torque may be obtained. Hence, an inverter nonlinearity compensation (INC) method exploiting the voltage feedback loop is introduced and its advantages over the conventional INC scheme are demonstrated. The proposed online control method is validated via measurements on a 10kW IPMSM. Index Terms-Dead-time compensation, flux-weakening control, interior permanent magnet synchronous machine, inverter nonlinearity, maximum torque per ampere control.

This paper presents a feed-forward torque control (FTC) technique for interior permanent magnet (IPM) brushless AC (BLAC) drives in traction applications. It is shown that by adopting the Newton-Raphson iterative method for solving the proposed high-order nonlinear relationship between the torque demand, flux-linkage and desirable dq-axis currents, FTC with due account of nonlinear machine parameters can be achieved for IPM BLAC drives. It is also proven that the comparison between the reference voltage magnitudes under maximum torque per ampere (MTPA) and field-weakening (FW) operations together with online base speed determination can be utilized for FW operation activation to achieve full exploitation of the available DC-link voltage during the transition between the constant torque and FW operation regions. Since both the dqaxis current references and the base speed for FW operation activation are computed online, the proposed FTC technique provides flexibility for online parameter update or estimation and is able to cope with wide DC-link voltage variation. The proposed FTC strategy is experimentally validated by measurements on a 10kW wide constant power speed range (CPSR) IPM BLAC machine drive. Index Terms-Feed-forward torque control (FTC), field weakening (FW) operation, interior permanent magnet (IPM) brushless AC (BLAC) machine, maximum torque per ampere (MTPA) operation.

Reliability and fault tolerance of drive systems has received renewed research interest over last decade with several fault tolerant drive topologies and control schemes having been proposed and investigated. In this paper, the performance of hysteresis current vector control methodologies when applied to three common fault tolerant drive topologies has been investigated and compared theoretically and experimentally. Their advantages and limitations in terms of current value, current and torque ripples, and operating speed range etc. are considered and verified by both simulated and measured results. It is proven that by simply modifying the hysteresis current control methodology, the maximum torque per ampere (MTPA) can still be achieved in post-fault operation, albeit with potentially increased current value, low frequency current and torque ripples, and compromised speed operating range.