Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Afrique du Sud → 2018 → Energy yield modelling and analysis of photovoltaic systems in Namibia

Nelson Mandela University (2018)

Energy yield modelling and analysis of photovoltaic systems in Namibia

Dobreva, Petja

Titre : Energy yield modelling and analysis of photovoltaic systems in Namibia

Auteur : Dobreva, Petja

Université de soutenance : Nelson Mandela University

Grade : Doctor of Philosophy (PhD) 2018

Résumé
Generation of electricity from solar energy by photovoltaic systems has the potential to reduce dependence on non-renewable sources for electrical energy production. It can also reduce the cost of electricity for consumers by moving away from the traditional models of centralised generation combined with extensive transmission. The prices of photovoltaic system components have decreased so dramatically in recent years that nowadays a large portion of the cost of electricity produced by photovoltaic systems is due to the cost of investment capital rather than the cost of the physical parts of the system. The cost of capital is determined by the perceived risk of the investment and is contingent on predictive models whose aim is to project outputs reasonably achievable by the system. The output of a photovoltaic system depends strongly on external factors, like solar radiation and temperature, as well as internal, system specific factors. Models that can accurately predict the output of a photovoltaic system have direct impact on investment cost. Several measures, like root-mean-square-error or coincident of determination, have been used in the past to assess accuracy of models but they appear to be inadequate for the task. I suggest a new method, comprising qualitative and quantitative measures, for model evaluation and introduce two new quantities to assess the predictive capabilities of the models. Models of several photovoltaic systems in Namibia were created, the new evaluation approach was applied to them and the results were compared to the conclusions that would have been drawn if the old measures were used. Outdoor current-voltage measurements are used to assess the power output and general state of the solar generator of a photovoltaic system, but the measured values strongly depend on irradiance and temperature. The prevailing approach has been to transform the current-voltage characteristics to standard-test-conditions with the so-called translation equations and then compare the translated power to the nominal value. The translated value of the power, however, depends on the translation equations used and is not necessarily an accurate reaction of the state of the solar generator. I have proposed a new method for assessment of the state of the solar generator that relies only on the measured quantities and does not require translations to standard-test-conditions. This alternative method has been used in the assessment of the condition of the solar generators in the systems included in this study.

Présentation (SEALS)

Version intégrale

Page publiée le 1er avril 2020