Page 277 - Advances in Renewable Energies and Power Technologies

250 CHAPTER 7 Strategies for Fault Detection and Diagnosis

by the inverter itself offering the evaluation of the PV system in real time, and it is
not necessary to involve complex simulation models to make the supervision and
diagnostic of the system.

7. CONCLUSION
Main strategies for automatic supervision, fault detection, and diagnosis of PV sys-
tems were presented in this chapter.
The simplest supervision procedure is based on the evaluation of the energy pro-
duced by the PV system by comparing the production of monitored and expected
energy over a given period. This evaluation is based on the analysis of the PR and
yields. The efficiency of this supervision technique in the detection of faults present
in the PV system is moderate. However, this procedure is one of the most used
because it is able to detect faults that result in relevant decrease in energy harvesting.
Moreover, the implementation of these kinds of supervision systems is not very
expensive. Nevertheless, the accuracy in the detection of faults depends primarily
on the accuracy of the models and simulation tools used to estimate the expected
values of the PR and yields.
The diagnosis techniques are focused on the identification of the most probable
fault present in the PV system once a fault was detected. Two diagnosis procedures
were presented in this chapter: the first one is based on the power losses analysis, and
the second one is based on current and voltage indicators. Both techniques have
shown good results and high efficiency in the identification of faults in several
real PV systems including different PV module technologies, PV system sizes,
and several configurations of inverters present in the system. However, the diagnosis
based on the current and voltage indicators is more efficient and offers more infor-
mation about the origin and the amount of power losses associated with present
faults in the PV system. The accuracy of this method in the detection of faults is
directly related to the definition of thresholds for current, TNRcfs, and voltage,
TNRvbm. A previous study of the PV system behavior is needed by using statistical
procedures to establish a limit of false fault detections. A good commitment between
the accuracy in the detection of faults and the probability of false fault detection can
be achieved by means of this statistical study and the offsets defined for the NRco
and NRvo values.

REFERENCES
[1] International Energy Agency, IEA, Trends 2016 in Photovoltaic Applications, Report
IEA PVPS T1-30, 2016.
[2] CEC, Joint Research Centre/Ispra. Guidelines for the Assessment of Photovoltaic
Plants, Document B: Analysis and Presentation of Monitoring Data, June 1993.
Issue 4.1.