Slides are here.
The first oil crisis in the 70s was a sort of renaissance for daylighting. The second renaissance came ~20 years ago.
What people mean by daylighting can mean different things (see the 5 different definitions in the sides).
The three main aspects of effective daylighting are daylight availability, visual comfort, and reduction of energy use.
How do we representing daylighting with numerical metrics? There are measures of daylight availability, glare probability, blind status (open/closed), and energy use. In the slides these are combined into a dashboard.
Rule of thumb: with no blinds, daylight penetrate 1.5-2x the window head height, with blinds its a bit lower.
In an urban setting, surrounding obstructions become much more important to daylight availability. Window-to-wall ratio x sky-angle must be larger than 2000 (rule of thumb).
Atrium rule of thumb: to daylight all spaces around an interior atrium, maximum atrium height is about 2.5x its width. Can go pretty far with these three rules of thumb.
From a massing strategy, any big box store can be very well daylight.
To just get uniform daylight in a space you only need 2-3% of the entire roof to be skylights. In California you are only supposed to punch holes in up to 5% of your roof.
Even though we don't talk about them much, skylights are a very effective technology.
(Unsourced) 80% of offics in California are directly under a roof.
Slides are here.
There are 2 types of simulations. Luminance or illuminance. Luminance (candella) might tell you where you will experience high glare. Illuminance (lux) might tell you that tables are well illuminated or not.
Architectural rendering vs daylight simulation. Conveying separate information. But generally both are very similar. In a simulation you have to be more aware of material properties, surrounding objects, and "light leaks".
The course makes an interesting intersection with computer graphics here are is starts to talk about Radiosity, ray tracing, etc.
Radiosity is nice because after one render you can change the camera angle and walk through the scene. This assumes everything has diffuse reflection only. Elements where large gradients would exist need to be subdivided into multiple surfaces.
This class focusses on a raytracer called Radiance. There's a recommended book called Rendering with Radiance.
If you do everything right, your simulation has a mean error of 20%. This is really good because our eyes are logarithmic sensors.
Mirrored curves, e.g. venetian blinds are very troublesome for a raytracer. In this case we have to wait a long time for simulations, or we use a mixture of forward and backward raytracing (photo gathering).
There's a long discussion of all the things people do wrong when trying to set up these daylight simulations. It's very easy to screw up if you don't take care of some common things that can go wrong. Simulations always tend to be better than reality (clean windows, lights off, blinds open, etc). There's a provided list of Simluation Tips.
Use the daylighting rules of thumb to estimate the daylit area of a given building, and rank a series of massing proposals and window layouts.
In a standard, office-type sidelit space equipped with venetian blinds, the depth of the daylit area usually lies between 1 and 2 times of the window head height. For spaces that are not equipped with a dynamic shading system, the ratio range increases to 2.5.
The product of the sky angle theta (in degrees) and the window to wall ratio, WWR, (in percent) of a standard sidelit space should be larger than 2000.
The maximum height for an atrium bordered by daylit spaces is 2.5 times its width.