Utility scale photovoltaic plant indices and models for on-line monitoring and fault detection purposes

Improving reliability and reducing maintenance and operating costs have become important factors in increasing the competitiveness of photovoltaic (PV) systems. For grid connected PV power plants every kilowatt-hour is crucial, because only kilowatt-hours that are fed into the grid are remunerated. A plant's operator can only adopt prompt measures to eliminate operational faults when these are immediately signaled. In fact, just reading the feed-in meter periodically is not sufficient to promptly recognize faults and to avoid the loss of yields. For this reason, continuous, absolute and comparative measurements are necessary to ensure the highest efficiency and availability of photovoltaic plants. Specifically, comparing the estimated, by means of suitable models of the PV plant, and the measured power allows to provide information on the operating status of the photovoltaic systems and so to perform preventive and/or reparative maintenance. In this context, a physical model has been developed and compared with other models to estimate the DC and AC power. Further, a novel figure of merit, named Non-Uniformity Index is proposed. It permits to individuate those cases the whole system, which is composed of many sections, e.g. PV fields, is working with a sufficient efficiency, but one or more photovoltaic fields are working at lower efficiencies compared to the others. The quality of the measured data has a significant influence on the estimation accuracy of the performance model and thus on the monitoring system, for this reason, models for irradiance and PV module temperature are used to check the measured data for inconsistency and other anomalies. Based on such models, a straightforward methodology for PV plants monitoring and fault detection purposes, which allows to minimize the number of required sensors and that can be applied to PV plants of any dimension, has been developed. The proposed methodology has been applied to a utility scale PV power plant (rated 1 MW), where a Supervisory Control and Data Acquisition system has been installed; operational data are saved on a remote server and available on a dedicated website. The experimental results show the effectiveness of the proposed approach. In order to have the possibility to analyze and compare the performance of the studied PV plant with others, the efficiency of the system has been estimated by both conventional Performance Ratio and Corrected Performance Ratio.