Generation of meteorological year for the assessment of photovoltaic systems performance in Togo, West Africa

Abstract

In this paper, the Sandia method, based on the Finkelstein-Schafer (FS) statistical techniques was used to develop a Typical Meteorological Year (TMY) for five cities in Togo (Lomé, Atakpamé, Sokodé, Kara and Mango) with a view to assessing the performance of the 72-cell polycrystalline solar photovoltaic (PV) module in the study areas. Seven daily measurements of meteorological parameters (relative humidity, dew point temperature, minimum and maximum temperatures, mean temperature, wind velocity as well as global solar radiation intensity) obtained from the National Aeronautics and Space Administration (NASA) for 1987–2016 period were used. The TMY predictions were compared with the 30-year Long-Term (LT) average values of the dataset. Two statistical indicators, the Mean Percentage Error (MPE) and the Root Mean Square Error (RMSE), were adopted to evaluate the performance of the PV energy system. Results show close agreements between TMY and LT values for all the parameters considered. The TMY in most cases underestimates the LT mean and fluxes are relatively small. Although the PV energy simulations were found to be underestimated, the generated solar PV energy obtained with TMY compares fairly well with those obtained with LT data. The performance of the TMY data is found to vary over the country but considered ‘very good’ with MPE values ranging between 0.24% and 2.46% as well as RMSE ranging between 8.29 and 16.71 kWh. Slight but insignificant biases (positive in the southern cities of Togo but negative in the north) were found between the estimated PV energy with the TMY when higher weight for global solar radiation was used. This study has implications in building energy simulation for comparative energy efficiency studies.