[Radiance-general] clear sky-calculated indoor illuminance is too lower than that of cloudy sky???

[ZZ]

Hi, all,

  I'm trying to use ESP-r coupling Radiance to perform thermal integrating
daylight use simulation.

  And I have performed 7 days simulation with 1 hour timestep, which means I
got 168 illuminance numbers of one indoor test point. I find calculated
illuminances of 6 days is basically close to the measured data. But1 day is
exceptional, there is a big difference between calculated illuminances and
meaured ones. That1 day is a very clear sky because that day has a highest
direct and diffuse radiation of those 7days, but the measured illuminance is
much lower than other six days. Anybody can tell why? and how to modifiy?

Thank you for your time.
ZZ