The photoelectric power conversion efficiency of polymer solar cells is till now, compared to conventional inorganic solar cells, still relatively low with maximum values ranging from 7% to 8%. This essentially relates to the existence of exciton and charge carrier loss phenomena, reducing the performance of polymer solar cells significantly. In this paper we introduce a new computer simulation technique, which permits to explore the causes of the occurrence of such phenomena at the nanoscale and to design new photovoltaic materials with optimized opto-electronic properties. Our approach consists in coupling a mesoscopic field-theoretic method with a suitable dynamic Monte Carlo algorithm, to model the elementary photovoltaic processes. Using this algorithm, we investigate the influence of structural characteristics and different device conditions on the exciton generation and charge transport efficiencies in case of a novel nanostructured polymer blend. More specifically, we find that the disjunction of continuous percolation paths leads to the creation of dead ends, resulting in charge carrier losses through charge recombination. Moreover, we observe that defects are characterized by a low exciton dissociation efficiency due to a high charge accumulation, counteracting the charge generation process. From these observations, we conclude that both the charge carrier loss and the exciton loss phenomena lead to a dramatic decrease in the internal quantum efficiency. Finally, by analyzing the photovoltaic behavior of the nanostructures under different circuit conditions, we demonstrate that charge injection significantly determines the impact of the defects on the solar cell performance. (C) 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4712622]