Tilted Photovoltaic Energy Outputs in Outdoor Environments

Siwei Lou,Mo Wang,Isaac Y.F Lun,Hainan Chen,Dawei Xia,Wenqiang Chen,Danny H.W Li

doi:10.3390/su11216052

Abstract

The direction and environment of photovoltaics (PVs) may influence their energy output. The practical PV performance under various conditions should be estimated, particularly during initial design stages when PV model types are unknown. Previous studies have focused on a limited number of PV projects, which required the details of many PV models; furthermore, the models can be case sensitive. According to the 18 projects conducted in 7 locations (latitude 29.5–51.25N) around the world, we developed polynomials for the crystalline silicon PV energy output for different accessible input variables. A regression tree effectively evaluated the correlations of the outcomes with the input variables; those of high importance were identified. The coefficient of determination, indicating the percentage of datasets being predictable by the input, was higher than 0.65 for 14 of the 18 projects when the polynomial was developed using the accessible variables such as global horizontal solar radiation. However, individual equations should be derived for horizontal cases, indicating that a universal polynomial for crystalline silicon PVs with a tilt angle in the range 0°–66° can be difficult to develop. The proposed model will contribute to evaluating the performance of PVs with low and medium tilt angles for places of similar climates.

Highlights

There is an increasing concern regarding energy resources, energy use, and its probable effects on the environment
Developing models using data from various PV panels with different tilt angles and climate conditions can result in more universal conclusions
It is expected that a simple and robust model will be developed that estimates the performance of the silicon crystalline PV cells in different routine tilt angles and azimuthal directions that can be applied for engineering use in early design stages

Summary

Introduction

There is an increasing concern regarding energy resources, energy use, and its probable effects on the environment. Developing models using data from various PV panels with different tilt angles and climate conditions can result in more universal conclusions This may involve several large databases and input variables. The boosted model was developed using up to 200 regression trees, resulting in a larger number of coefficients than that for ANN, causing difficulties for it to be used elsewhere In this connection, it is expected that a simple and robust model will be developed that estimates the performance of the silicon crystalline PV cells in different routine tilt angles and azimuthal directions that can be applied for engineering use in early design stages. Importance of the input variables to the PV performance estimation were evaluated to remove those variables of a low significance using the RTree approach This saved the cost of measurement, model development, and curve fitting. The performances of the polynomials in the first and second orders using the identified input variables were evaluated, and their advantages and limitations are discussed

Data Collection of PV and Solar Radiation

Methodologies

Results and Discussion