Modeling the grid interaction of a large, discrete-time controlled active power filter with a multi-level inverter that is part of an electric vehicle charging station

This research project is the first stage of the design of public, electric vehicle charging station with a 3 minute charge time. First, an appropriate topology for the electric circuits of the charging station was selected. Then, control algorithms were designed to perform the charging functions of the station and implement active power filter (APF) functionality. These control algorithms were verified with simulation and high level of performance (bandwidth) was achieved for the APF. Then the control algorithms were redesigned in discrete-time (DT) and with Direct-Quadrature (DQ)-theory to achieve the same bandwidth while substantially lowering switching loss. At this point, the interaction of this charging station/APF with the grid had to be modeled because its large size indicated that it may local voltage stability and quality issues. This part of the project is its primary contribution to the existing body of knowledge. In particular, the DQ-transform and the triangle-hold equivalent were applied in new ways to transform a non-linear, dynamic, closed loop into a linear, time-invariant (LTI) loop. The validity of these new techniques was then verified with simulation. The new techniques and the simulations verifying them both assume that the output current of the charging station makes linear transitions in continuous time (CT) between its discrete-time (DT) values (i.e. it was assumed that the output current has been triangle-held). Therefore, this assumption was verified with additional simulations. Finally, the results of these simulations show that this assumption valid for a large class of next-generation grid interfaces.