Search Results

The optimal design of hybrid-electric vehicle power systems poses a challenge to the system analyst, who is presented with a host of parameters to fine-tune, along with stringent performance criteria and multiple design objectives to meet. Herein, a methodology is presented to transform such a design task into a constrained multi-objective optimization problem, which is solved using a distributed evolutionary algorithm. A power system model representative of a series hybrid-electric vehicle is considered as a paradigm to support the illustration of the proposed methodology, with particular emphasis on the power system's time-domain performance.

A temperature estimation algorithm (thermal observer) that provides accurate estimates of the thermal states of an electric machine in real time is presented. The thermal observer is designed to be a Kalman filter that combines thermal state predictions from a lumped-parameter thermal model of the electric machine with temperature measurements from a single external temperature sensor. An analysis based on the error covariance matrix of the Kalman filter is presented to guide the selection of the best sensor location. The thermal observer performance is demonstrated using a 3.8 kW permanent-magnet machine. Comparison of the thermal observer estimates and the actual temperatures demonstrate that this approach can provide accurate knowledge of the machine's thermal states despite modeling uncertainty and unknown initial machine thermal states.