Predicting diffuse solar irradiance on obstructed building façades under irregular skyline patterns for various ISO/CIE standard skies

Solar irradiance data on building façades are important for BIPV system evaluations and energy-efficient building designs. In urban cities where the building façades are always obscured by surrounding buildings, the determination of sky-diffuse solar irradiance can be complex. Previously, a mathematical model was proposed to estimate the obstructed vertical sky component (OVSC) under various skies by assuming that the obstructions are infinitely long. However, more often, in actual cases, the opposite obstructions are surfaces with different heights in azimuth which can form various irregular skyline patterns. Therefore, this study proposes new methods to determine vertical sky-diffuse solar irradiance on building façades under complex irregular obstructions. Various configurations of irregular obstructions were identified by a proposed quantification method, and the OVSCs were simulated using RADIANCE package. The simulated results were used to correlate with those of infinitely-long urban canyons. Both Artificial Neural Networks (ANN) based model and simple regression model were proposed to find the correlations between various angles defining the irregular obstruction and the equivalent obstruction angles of corresponding infinitely-long cases. The R² values of the ANN model and regression model against simulated results using the testing datasets were above 0.97 and 0.98, respectively. Field measurements were carried out for model validations and the R² value between the proposed methods and the measured readings was 0.8. The findings are significant to PV system designs, daylighting scheme evaluations and low/zero-energy building designs under complex obstructed environments.