Estimation of hourly averaged solar irradiation: evaluation of models

Hourly solar radiation data are required in many building services applications. These are also reported in the Chartered Institution of Building Services Engineers Guides A & J. Data from 16 locations in the UK were used to evaluate the so called Liu and Jordan model(1) for monthly averaged hourly solar irradiation. Individual data sets spanned periods from 12 to 26 years between 1968 and 1994, and overall, provided data from practically the full range of latitude of the UK (50.22 degrees N-58.13 degrees N). For hourly estimation, the model only slightly underestimated both global and diffuse radiation before noon and overestimated, again only slightly, after noon. In addition, a discrepancy was observed between the measured data and the model's predictions at low sunset angles. Following earlier research work, an attempt was made to further improve the Liu and Jordan model. However, it was found that at least for the UK data set, any such attempts were futile. This behaviour was attributed to the highly random nature of UK's solar climate. Practical applications: Most meteorological stations report solar radiation data on a daily averaged basis. However, most building energy simulation software requires hourly radiation. Research studies have confirmed that the well-known Liu and Jordan model, which enables the above conversion, performs well for locations in the US. This paper evaluates the above model for locations in the UK and compares it with previously studied Indian locations. According to the evaluation, the average accuracy of the model to estimate hourly radiation from its daily counterpart is 85%.