Solar radiation resource data is required in designing, modeling and assessing performance of solar energy applications. The preferred method of collecting this resource data is through measurement. However, the problem with this method is that measuring equipment cannot be installed in every location of interest. The equipment is expensive to purchase and maintenance costs are high. An alternative is to estimate the resource data by interpolating existing records.
The study for which this paper was written investigated the performance of five methods in interpolating total solar radiation in Uganda, located in East Africa. The study was meant to propose an interpolation method suitable for the study location. The methods investigated included: nearest point, moving average, moving surface, trend surface and ordinary kriging.
Results showed that the moving average was the most reliable for interpolation of total solar radiation in Uganda and subsequent drawing of solar radiation maps. The results are supported by an error analysis which showed an overall smaller error for the proposed method. The error analysis involved computation of normalized mean bias error and root mean square error. The worst performing method was the nearest point. The principle of computation of an interpolated point value for the nearest point interpolation introduces a significant error which impacts on the accuracy of the interpolated value unlike the moving average where the interpolated point value is a weighted average of neighboring point values.
The paper includes a sample solar radiation map based on the proposed interpolation method.