Evaluation index system for photovoltaic systems statistical characteristics under hazy weather conditions in central China

Abstract

This study investigates different methods of determining the optimum tilt angle (OTA) for photovoltaic (PV) systems to maximise energy yield, and provides an evaluation index system for the static and dynamic performance of PV systems under hazy weather conditions, in view of the nature of solar energy resource and PV energy yield. Due to the fact that the data of total solar radiation on inclined surface in Wuhan is not available and PV power generation is non-linear process, an experimental off-grid PV system with various tilt angles and orientations is established which is located in Hubei Meteorological Administration in Wuhan. The experimental system is then used to analyse the accuracy of five in-plane solar radiation models and the relationship between the model predicted errors and meteorological environmental factors, explore the calculated and measured OTA of PV modules and evaluate the performance of PV system with different tilt angles and orientations based on static and dynamic indexes in real atmosphere condition. Finally, the influence of dust on PV energy production has been reported. After cleaning, the energy yield of PV models can increase by 10-20%.

Evaluation of solar radiation models on vertical surface for building photovoltaic applications in Beijing

“…The coefficient variation ($${V_\sigma }$$), mean anomaly percentage ($${X_d}$$) and stability index ( S ) are the standardised quantification indexes of the fluctuations of meteorological and other variables [32]. The smaller the values of coefficient variation and mean anomaly percentage are, the smaller the variable fluctuations are.…”

A New Data Driven Long-Term Solar Yield Analysis Model of Photovoltaic Power Plants

“…A higher value means lower accuracy of forecasting algorithms whereas lower MAPE value means high accuracy of the forecasting algorithms. The Rvalue is the correlation between the predicted values and the observed values [27]- [29]. It gives an idea about the model generalization.…”

“…An Rvalue close to one means, the forecasting values are highly close to the fitted regression line and it can be used in more generalized cases. The Rvalue can be calculated as in (25) [27]- [29]:…”