This paper shows the performance curves obtained with an experimental campaign on the following different prime movers: a 100 kW microturbine, a 20 kW internal combustion engine, a 450 kW SOFC-based hybrid system and a 100 kW absorption chiller. While the size related to the microturbine and the engine are actual electrical power values, the hybrid system size is an electrical virtual power (an emulator rig was used for this plant) and the chiller value is a cooling thermal power. These experimental results were obtained with a smart polygeneration facility installed in the Innovative Energy Systems Laboratory by the Thermochemical Power Group of the University of Genoa. This facility was designed to perform tests on smart grids equipped with different generation technologies to develop and improve innovative control and optimization tools. The performance curves were obtained with two different approaches: tests on real prime movers (for the microturbine, the engine and the chiller) or measurements on an emulator rig (for the hybrid system). In this second case, the tests were carried out using an experimental facility based on the coupling of a second microturbine with a modular vessel. A real-time simulation software was used for components not physically present in the experimental plant. These results are a significant improvement in comparison with the available data, because experimental results are presented for different prime movers in different operative conditions (both design and part-load operations). Moreover, since both manufacturers and users are not usually able to control air inlet temperature, special attention was devoted to the ambient temperature impact on the 100 kW microturbine because this property has a strong influence on the performance of this machine. For this reason, empirical correlations on the ambient temperature effect were obtained from the experiments with the objective to perform an easy implementation of the optimization tools. Experimental performance curves (including several off-design conditions) are essential for smart grid management because (if they are implemented in optimization tools) they allow to find real optimal solutions (while tools based on linear or calculated correlations can obtain results affected by significant errors). (C) 2015 Elsevier Ltd. All rights reserved.