Cogeneration of heat and electricity constitutes an important pillar in energy and climate policy. The optimal dimensioning and operation of residential micro-combined heat and power (mCHP) systems is an important area of ongoing research, especially because these systems are prohibitively expensive. This paper presents an optimisation model for the capacity and dispatch planning of residential mCHP systems, which is applied with field data from the UK. The method possesses several novelties, including the consideration of technical and economic constraints previously overlooked, such as economies of scale for investments. The application of the model to empirical data demonstrates that economic savings of up to 30% in total annual costs can be achieved through an optimal sizing and operation of a system consisting of a mCHP unit, a peak load boiler and a hot water storage tank. The CO2 emissions in this optimised case are approximately the same as in the case with a mCHP unit only, but substantially (around 22%) below those in a case with a separate generation of heat and electricity from the grid. The new method has been validated in several ways, including a comparison to investments based on constant specific investments, which demonstrate that the latter consistently underestimated the total annual cost by up to 17%. This figure would be higher for larger, or a higher range of electrical CUP capacities. The largest uncertainty relates to the assumed load profiles, from one year, which are taken as representative for the whole time horizon of several years. Further research should focus on more accurately representing discrete operating modi as well as applying the method to other building objects and CHP technologies. (C) 2014 Elsevier Ltd. All rights reserved.