We present a model developed to study the increase of self-consumption of photovoltaic (PV) power by smart charging of electric vehicles (EVs) and vehicle-to-grid (V2G) technology. Whereas previous studies mostly use large EV fleets in their models, our focus is on a smaller scale. We apply the model to a microgrid in Lombok, a residential neighbourhood in the city of Utrecht, the Netherlands. The microgrid consists of a 31 kWp PV installation, an office, internet servers, three households, and two EVs. Three control algorithms are presented which manage the charging profile of multiple EVs either in real-time or using linear optimisation with predictions for PV power and electricity demand. We perform one-year simulations using data for PV power, EV use, and electricity demand. Simulations results are evaluated on PV self-consumption and peak demand reduction. In addition, we make qualitative statements on battery degradation resulting from the charging strategies based on several indicators. We also simulate changes in microgrid composition, for example by including more EVs. In the simulations, self-consumption increases from 49% to 62-87% and demand peaks decrease by 27-67%. These results clearly demonstrate the benefits of smart charging EVs with PV power. Furthermore, our results give insight into the effect of different charging strategies and microgrid compositions. (C) 2015 Elsevier Ltd. All rights reserved.