UTILITY SIZED PHOTOVOLTAIC PLANTS: ANALYSIS OF PERFORMANCE AND INSIGHT ON DRIVERS OF LOW PERFORMANCE

The large-scale grid connected photovoltaic systems have become one of the most used renewable energy sources to support the conventional energy generation and one of the most cost-effective sources of energy, not only as a clean energy but also for the possibility to generate profit and create attractive opportunities of investments. From the operational point of view, the photovoltaic technology has different advantages with respect to other renewable energies because the solar resource is predictable obtaining trustworthy forecasts; the technology implemented is normally fixed reducing the maintenance cost and guaranteeing the long life of operation; scalability and quick construction. Those factors provide to the investors reliability to make a profit and, furthermore, they can predict the production of electricity of each PV system using models and compare them with the real values that are record daily implementing a monitoring system. In photovoltaic technology, reference models to predict the solar irradiance, the electricity generation, the performance ratio of the PV and profit of the owners, one more expansive and the other one thought as a conservative model to be reached to the real production. This reference models are evaluated daily, monthly and yearly with the recorded data by companies like Vector Cuatro that take care of the asset management and technical advisory in green power. In order to have accurate models, each year the models are revised using a theoretical value of degradation rate or using other input parameters; although models are analyzed monthly and yearly, the performance of the dataset models is also evaluated during the lastly years focusing on the most important parameters such as global solar irradiance, energy estimated and the performance ratio expected, implementing statistical analysis to observe how the trends of models are, monthly and yearly, with respect to real parameters to improve the quality of energy prediction and profits for both cases, considering some hypothesis for estimating and re-computing missing information in the databases of client case, bank case and actual values. On the other hand, the models are used as reference point for the evaluation of actual KPI (Key Performance Indicators) of the photovoltaic plant. The models can be used as benchmark to assess the impact of events such as failures of components, technical issues or plant shutdown, in which the performance ratio is affected and reduced, and as a consequence the energy produced decrements. Therefore, in order to study the effect of failures and other possible unforeseen problems, the performance ratio of one PV system is analyzed over multiple years and the most relevant technical issues are detected. The losses associated with different technical causes are quantified. In particular, impact of failures of main components like photovoltaic modules and inverters is assessed. Also the recovery of performance after extraordinary maintenance is accounted. Finally, Measurements were taken making use of Termocamera to observe the effect of the shading in a PV module and study the hotspots in the photovoltaic systems.