We present a framework and a model for the simulation of collective energy self-consumption strongly coupled to occupant activity in households. The model combines multi-agent simulation of human activity, a dynamic model of building energy including Domestic Hot Water (DHW) and all electrical appliances, and local generation of photovoltaic energy. A cooperative energy exchange policy between households was defined in order to maximise the use of locally generated PV electricity at the neighbourhood scale. In this first implementation, three households A, B and C were simulated, one being the PV energy producer. The DHW model was modified to benefit from PV electricity. Thanks to the energy exchange policy, a total of 83% of the energy generated by household A is used locally in the neighbourhood (compared to 38% when there are no energy exchanges with household B and C). The next step of this study is to add a model of energy storage in several self-consumption configurations.