Browsing by Author "Mahachi, Tafara"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- ItemEnergy yield analysis and evaluation of solar irradiance models for a utility scale solar PV plant in South Africa(Stellenbosch : Stellenbosch University, 2016-12) Mahachi, Tafara; Rix, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: It is critical for electrical utility providers to be certain of the field performance of their systems in order to realize returns on project investments. Due to the actual output of installed solar PV modules being far more involved than in standard test conditions, some real world systems outperform simulations. The output of a solar PV system depends on various factors like local weather, system design characteristics, etc. Various energy forecasting simulation software exist and PVsyst is one of the most reliable when given the right input data. It is vital that software simulated energy yield values closely correlate with the actual field performance. In this study, the cause of the seasonal differences in measured and simulated energy yield for a solar PV power plant in South Africa is investigated. It was observed that a fixed tilt 75 MWP operational system significantly outperformed the software simulations in winter months during its first year of operation. Normally it is expected that simulations perform better than real world systems. In order to achieve our goals, an initial yield assessment was done in PVsyst and simulation results were evaluated against measured data. Differences in actual and estimated weather conditions were avoided by using irradiance and ambient temperature values recorded at the solar PV plant site. Due to the overestimation of losses in the PVsyst model, the simulated results did not correlate with the measured values of energy yield. In the simulations, the real world system was underestimated by a yearly average value of 6.4% and 5% using the Hay and Perez model respectively. Understanding the performance drivers and losses within a solar PV system is crucial for ensuring reliable and high performance solar PV systems. Due to the overestimation of losses in the default PVsyst loss model, system loss values were then derived from the data recorded on the solar PV plant site. Findings from the calculated losses were then used to improve the PVsyst model for the solar PV plant. After using 30 minute average input data and the improved PVsyst loss model, the real operational system was slightly underestimated by the improved simulation model by a yearly average value of 0.78%. Seasonal differences between measured and simulated energy yield were still observed. These differences were larger in winter than in summer and they were also in correlation with the variances in the measured and PVsyst estimated Plane-Of-Array irradiance. Due to the seasonal differences between the improved PVsyst model and the measured energy yield values in winter months, irradiance models that estimate the Plane-Of-Array irradiance from Global Horizontal Irradiance measurements were evaluated for South African climatic conditions. The most common irradiance decomposition models namely Orgill and Hollands, Erbs, Louche, Reindl1, Reindl2, DISC and Dirint were evaluated as well as combinations of the Isotropic, Sandia, Klucher, Hay, Reindl and Perez transposition models with the best performing decomposition models. Based on the root mean square values, the best performing decomposition models were the Dirint, Louche and Disc whereas the Perez, Reindl and Hay transposition models performed the best for South African sites. On a monthly basis, simulations done using measured Plane-Of-Array irradiance were within 2% of the field recorded values. It was then concluded that the choice of irradiance data, transposition and loss models has a significant effect on the short and long term energy predictions.