Electric machines are constructed from a variety of different materials which exhibit different mechanical and thermal expansion properties. Losses within the machine cause an increase in temperature, resulting in thermally induced internal stresses in insulation materials. Changes to the operating condition will result in different levels of loss being produced, and give rise to a thermal cycling profile to the machine and cyclic stress in the insulation. This paper studies the impact of these thermal cycles and resultant cyclic stress in the insulation on the lifetime of machines. Through the use of finite element analysis the internal stress levels, and corresponding stress cycle profiles are determined within the machine insulation layer, this data is then combined with experimental data, allowing stress-life (number of cycles) (SN)-curves to be generated for the machine type under test. The correlation between predicted insulation fatigue life with the measurements provides a physical insight of lifetime reduction due to thermal cycling.

Reliability of electric drives is of paramount importance for safety critical applications in transportation, medical and military industries. While numerous fault tolerant machine and inverter topologies have been developed for improving drive reliability, comprehensive reliability evaluation of fault tolerant electric drive systems has not been reported in literature. This paper performs quantitative assessment of mean-time-to-failure of representative fault tolerant permanent magnet AC drive systems by employing Markov chain modelling technique. It is shown that the mean-time-to-failure of a fault-tolerant drive can be significantly increased by employing health monitoring/maintenance check and prompt repair/replacement of faulty components.

The paper describes the analysis, design and control of two new versions of spherical permanent magnet motor, which are capable of two and three degrees-of-freedom of controlled motion, and have a high specific torque capability. Their magnetic field distribution is established using an analytical technique formulated in spherical coordinates. This enables the torque vector and back-EMF to be derived in closed forms. An optimal design procedure, to achieve maximum output torque or maximum acceleration for a given payload, is presented. The control of the motors, whose dynamics are similar to those of robotic manipulators, is facilitated by the establishment of a complete actuation system model and the application of the PD or computed torque control law. The validity of the analysis and design techniques, and the effectiveness of the control strategy, are confirmed by measurements.

A new form of actuator with a spherical permanent magnet rotor and a simple stator winding arrangement is described and analyzed. The actuator incorporates a simple position sensing technique, and, by utilizing a rare-earth permanent magnet and optimizing its design, is capable of developing a high specific torque. The magnetic field distribution, torque vector and back-EMF constants are derived analytically, and a detailed kinematic analysis is presented. These allow a complete and accurate dynamic model of the actuator to be established, and facilitate the development of the closed-loop control law.

A new sliding mode control law based on a linear predictor using a double integrator model for the control of robotic manipulators is presented. The new control law retains the simplicity and robustness of sliding mode controllers while the chattering problem commonly associated with this type of controller is completely eliminated. The control law is directly expressed in a discretized format for digital implementation. Simulations show that the new controller is chatterless, robust with respect to large parameter uncertainties and external disturbances, and simple for digital implementation.>