Contribution to the characterization and modeling of photovoltaic generators

• Autor: Angulo Abanto, Jose Ruben
• Formato: tesis doctoral
• Fecha de Publicación: 2022
• Institución: Pontificia Universidad Católica del Perú
• Repositorio: PUCP-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorio.pucp.edu.pe:20.500.14657/188472
• Enlace del recurso: http://hdl.handle.net/20.500.12404/23969
• Nivel de acceso: acceso abierto
• Materia: Sistemas de energía fotovoltaica; Instalaciones fotovoltaicas; Generadores eléctricos--Modelos matemáticos; https://purl.org/pe-repo/ocde/ford#1.03.00

A crucial aspect of evaluating and maintaining a photovoltaic (PV) installation connected to the grid is the availability of models that describe its operation reliably in real operating conditions. The nominal power of the PV generator (P*M) is considered an essential input parameter, and several models have been proposed to estimate P*M for characterizing the PV system. In the case of PV generators in outdoor conditions, the American Society for Testing and Materials, the International Electrotechnical Commission, and others have proposed procedures to determine the P*M of the generator. As part of these procedures, monitoring days with ideal conditions is mandatory, notably days with a clear sky, high irradiance values, and low wind speeds. Such restrictions can limit the number of suitable monitoring days, especially in places where clouds frequently form. This thesis proposes a new approach that allows estimating the P*M with data even from non-ideal, partially cloudy days. Based on non-parametric statistics, this procedure identifies and filters out noise as well as deviations from ideal conditions of irradiance, allowing for an estimation of P*M with similar accuracy as for a clear-sky day. This new procedure enables the characterization of a PV generator on a daily basis without the requirement to meet ideal conditions, thus, considerably enhancing the number of suitable monitoring days. To overcome the limitation in the P*M estimation and considerably extend the number of monitoring days, the new procedure can be applied to ideal and non-ideal conditions, such as partially cloudy days. This procedure determines the most probable nominal power value within one monitoring day using non-parametric statistics. In order to test the new procedure, a 109.44 kW photovoltaic plant in Granada, Spain, was monitored for six months. A referential procedure reported in the literature for large PV plants under ideal climatic conditions is first applied to estimate its nominal power. The results indicate that the nominal power can be estimated reliably in non-ideal conditions, maintaining the same precision as in ideal conditions. Then validating the procedure for a smaller PV generator and under different conditions, two small grid-connected 1.5 kW PV arrays were used. The PV systems in question are located in two different cities in Peru: Chachapoyas (tropical highland) and Lima (coastal desert). The objective of this study in Chachapoyas was to validate the methodology in a tropical climate with a high presence of clouds but at the same time with high irradiance values above 800 W/m2. According to the results obtained, under these conditions, the nominal power of the system can be calculated with reasonable certainty. As a precaution, monitoring for more than one day is recommended to obtain more data (at least 3 hours with high irradiance) to reduce uncertainties. Lima, Peru's second location under study, has a particular climate. Since the capital is located in a desert with high relative humidity values, dust deposition increases and power output decreases due to these conditions. For this purpose, the nominal power was used as a parameter to determine the maintenance schedule. Since keeping the system in optimal performance, considering this in future installations for operation and maintenance costs, is essential. The new procedure developed in this work can be applied to facilitate technical due diligence and quality control processes for PV generators of different sizes and under different operating conditions that are being re-purchased or have been recently installed. The possibility of daily monitoring of the P*M also enables long-term monitoring of a PV generator to ensure the correct operation or identify possible degradation effects