## Abstract

Photovoltaic (PV) power systems are prone to ageing and degradation occurring in the PV cells during their lifespan. Such phenomena cause significant output power degradation which manifests as economic losses. Hence, a reliable condition monitoring procedure to detect and quantify ageing is a necessity. Such procedures do exist, but they are typically laborious and costly. The present study aims at finding a solution for this problem.

A feasible approach for monitoring the condition of a PV system is to measure the current-voltage curves from the terminals of a PV unit and fit a mathematical model describing the operation of a PV module or a larger PV unit to the curves. The widely used single-diode model is a suitable choice. The fitted model parameters provide valuable diagnostic information on the condition of the PV cells.

However, there are some practical challenges. Firstly, the model parameters are affected by the operating conditions, the measurements of which seldom exist at practical PV sites. This makes it necessary to identify the operating conditions jointly with the model parameters. Secondly, using entire measured current-voltage curves in fitting requires the rundown of the PV system for the measurement period, while using partial current-voltage curves reduces the fit quality. Hence, a systematic study of the effect of the limitation of the measurement range on the fitted parameters is needed. Thirdly, the discrepancies in the current-voltage measurement data make such limitation even more involved. Hence a suitable pre-processing procedure for the measurement data is needed.

These issues are addressed in this thesis from an empirical viewpoint. First, a new pre-processing procedure for the measured current-voltage curves is developed. It can be used to improve the quality of such measurement data, making it more suitable for fitting. Thereafter, a novel single-diode model fitting procedure identifying the operating irradiance and temperature jointly with the actual model parameters is developed. It can be utilized fully without external irradiance or temperature measurements. Finally, the effect of limiting the measurement range of the current-voltage curves to the vicinity of the maximum power point is systematically investigated, particularly focusing on ageing detection. It is shown how the measurement range of the current-voltage curves can be limited while maintaining the reliable detection of ageing. As a significant result, the developed single-diode model fitting procedure allows for the usage of suitably formed partial current-voltage curves in the parameter identification.

A feasible approach for monitoring the condition of a PV system is to measure the current-voltage curves from the terminals of a PV unit and fit a mathematical model describing the operation of a PV module or a larger PV unit to the curves. The widely used single-diode model is a suitable choice. The fitted model parameters provide valuable diagnostic information on the condition of the PV cells.

However, there are some practical challenges. Firstly, the model parameters are affected by the operating conditions, the measurements of which seldom exist at practical PV sites. This makes it necessary to identify the operating conditions jointly with the model parameters. Secondly, using entire measured current-voltage curves in fitting requires the rundown of the PV system for the measurement period, while using partial current-voltage curves reduces the fit quality. Hence, a systematic study of the effect of the limitation of the measurement range on the fitted parameters is needed. Thirdly, the discrepancies in the current-voltage measurement data make such limitation even more involved. Hence a suitable pre-processing procedure for the measurement data is needed.

These issues are addressed in this thesis from an empirical viewpoint. First, a new pre-processing procedure for the measured current-voltage curves is developed. It can be used to improve the quality of such measurement data, making it more suitable for fitting. Thereafter, a novel single-diode model fitting procedure identifying the operating irradiance and temperature jointly with the actual model parameters is developed. It can be utilized fully without external irradiance or temperature measurements. Finally, the effect of limiting the measurement range of the current-voltage curves to the vicinity of the maximum power point is systematically investigated, particularly focusing on ageing detection. It is shown how the measurement range of the current-voltage curves can be limited while maintaining the reliable detection of ageing. As a significant result, the developed single-diode model fitting procedure allows for the usage of suitably formed partial current-voltage curves in the parameter identification.

Original language | English |
---|---|

Place of Publication | Tampere |

ISBN (Electronic) | 978-952-03-3156-6 |

Publication status | Published - 2024 |

Publication type | G4 Doctoral dissertation (monograph) |

### Publication series

Name | Tampere University Dissertations - Tampereen yliopiston väitöskirjat |
---|---|

Volume | 904 |

ISSN (Print) | 2489-9860 |

ISSN (Electronic) | 2490-0028 |