Reference strings for statistical monitoring of the energy performance of photovoltaic fields

The paper deals with the issues to monitor the energy performance of Photo Voltaic (PV) fields by means of low cost hardware. In fact, the monitoring systems for low or medium rated power PV plants are often constituted by a limited number of sensors and low processing capacity. These systems allow a supervision of the PV fields when strong reductions of the produced energy happen, but they are ineffective to alert the end user about a gradual energy reduction. These issues are related to the natural ageing of PV modules, the dust or dirt accumulation on the PV modules, and so on. This paper proposes a methodology based on inferential tools, which return information about the correct operation of the PV field. The methodology needs an initial training that allows to define one or more reference strings, which will be used as benchmarks for future comparisons.