Ph.D. Thesis title:

Analysis, Diagnosis and Fault Detection in Phtotvoltaic Systems

Author:

Aissa Chouder

Director:

Santiago Silvestre

 

Reading Date:

09/02/2010

 

Abstract:

The sustainability of photovoltaic systems passes through an accompaniment of the photovoltaic system installed in the field. However, to
prevent faults and failures in photovoltaic systems, monitoring facilities have to be installed in order to acquire all the parameters of concern. In addition,
dedicated software will be linked to the monitoring facilities which will automatically diagnose and detect faults or failures in the specific photovoltaic
system and also reliable failure detection routines must be developed.
In the present thesis, we first focused on modelling the different part of a PV grid connected system in which we put enough emphasis on
modelling PV module and PV generator. The objectives assigned to this doctoral thesis work consist of developing an integral system of simulation,
monitorisation, diagnosis, failures detection and remote supervision of photovoltaic systems. These kinds of systems are aimed to designers as well as
technicians to provide them necessary tools for the development of their activities of investigation, design and maintenance.
In order to reach the objectives mentioned before, the detailed analysis of the system and the well knowledge of each part in it is of prime
interest. Hence all components of the photovoltaic system must be modelled in adequately form to approach the real behaviour of the specific
component in order to obtain accurate simulated data in the whole system, and then performing comparison between the real data of the system and
the predicted behaviour. Such approach can be summarized in the following points, developed in this work:

1. A reliable and accurate electrical model of the photovoltaic module has been established. This model is able to determine, at every moment, for a given irradiation and temperature, the power produced by the system and fed into the grid.
2. An accurate efficiency model of the power conditioning unit (MPPT and Inverter), which incorporate losses present during the conversion process as well as most important effects of operation faults, has been carried out.
3. The identification and validation of the system components’ models are necessary before using these models for results analysis and diagnosis. The evaluation and identification of these models (parameters extraction) was based on the measurements data taken from each part of the system and data sheet provided by the component manufacturer or from actual system operation. All the components models have been programmed by using Matlab/Simulink software to provide the system component behaviour as well as the system behaviour when all components of the system are linked together.
4. A monitoring system has been designed in order to validate all the theoretical assumptions and also to supervise and diagnostic the overall system. The measurements include all the DC and AC quantities of the system already in operation as well as irradiation and temperature. The monitoring systems have been installed in the PV grid connected system located at Renewable Energy Centre of Algiers (CDER Algeria).
5. Different models and estimating methods to assess the power losses due to irradiance, reduction of sunlight transmittances on glass of modules, temperature and series resistance, have been revised and simulation procedures allowing these power losses estimation in PV systems have been developed in Pspice and Matlab environments.
6. Finally, a supervision and fault detection procedure for PV grid connected systems, based on the power losses analysis is presented. The procedure analyses the output power losses in the DC side of the PV generator. Processing certain power losses indicators such as Thermal capture losses and Miscellaneous capture losses allow the supervision procedure to generate a faulty signal as indicator of fault detection in the PV system operation. The procedure has been successfully tested experimentally.