Rapidity and accuracy of algorithms evaluating yearly collected energy are an important issue in the context of optimizing concentrated solar power plants (CSP). These last ten years, several research groups have concentrated their efforts on the development of such sophisticated tools: approximations are required to decrease the CPU time, closely checking that the corresponding loss in accuracy remains acceptable. Here we present an alternative approach using the Monte Carlo Methods (MCM). The approximation effort is replaced by an integral formulation work leading to an algorithm providing the exact yearly-integrated solution, with computation requirements similar to that of a single date simulation. The corresponding theoretical framework is fully presented and is then applied to the simulation of PS10.