High Speed PM-Generator Optimized Sizing-Based on Particle Swarm Optimization for Smart Grids
Proceedings of the IAJC/IASM International Conference
High-Speed, Permanent-Magnet (HSPM) types of micro genearators play an important role in power generation involving Smart Grids Applications. This paper illustrates the benefits of HSPM generators compared to the traditional Permanent Magnet (PM) synchronous machines which offer significant reduction in both weight and volume. An optimized analytical design is proposed and compared with the original machine design of a typical 500 kW output power at tip speed of 250 m/s. These two designs take into consideration multiple factors including classical sizing and problem formulation for optimizing efficiency with bounded constraints. A Particle Swarm Optimization (PSO) algorithm was formulated to optimize efficiency as an objective or fitness function and to minimize machine size as a non-linear function with bounded parameter constraints. Particle swarm algorithms use population based on flocks of birds or insects swarming. The parameter variables used for this type of optimaztion consist of rotor length to diameter ratio, rotor radius, and stack length. It was determined that using the PSO algorithm in HSPMSG sizing is able solve constrained and unconstrained optimization problems. Test results including simulations using PSO Tool in Matlab showed significant improvement in machine design and performance. Furthermore, it was observed that the proposed technique has the advantage of limiting losses at higher frequencies with low weight/volume applications thus improving overall efficiency. Other system parameters such as power factor were also shown to improve as well. Finally, several analytical design problems with waveform variations, harmonics distortion, rotor losses, and effects of poles changing were provided to show the merit of the proposed optimization technique.
El-Shahat, Adel, Rami J. Haddad, Youakim Kalaani.
"High Speed PM-Generator Optimized Sizing-Based on Particle Swarm Optimization for Smart Grids."
Proceedings of the IAJC/IASM International Conference: 1-22 Orlando, FL.