Hi Jo Ling,

The only way OpticStudio can measure the peak irradiance is by noting the irradiance determined for the detector pixel which received the most power. If the average number of rays in a pixel is small, this could produce large statistical variations. It is also the case that even with good statistics, you are really averaging the  power received over the pixel area to determine each pixels irradiance. So if the power received in a given pixel is non-uniform, then dividing that pixel into a set of smaller pixels will result in one of them having greater irradiance that the original larger pixel.

I think one way to obtain meaningful simulation results is to consider the largest area for which varations in irradiance is insignificant, and choose that as a pixel size. Then use sufficient analysis rays such that pixels receiving significant power have a statistically significant number of ray strikes.

Kind regards,

David