Optimum daylighting

By - , Build 133

Daylighting design has come a long way in the last 20 years, but the designer needs to understand how to use sky and climate modelling data to deliver optimum lighting in a 21st century office space.

Perez sky-based renders of traditional clear sky (left), cloudy sky (centre) daylight plus the real world partly cloudy sky (right).
Accurate daylight render by Jake Osborne as part of his Master of Building Science from Victoria University.

Computers today can simulate the appearance of buildings under different light conditions in a very believable way. This offers both an opportunity and a risk.

The opportunity is that daylight in buildings can be far more accurately designed than before. On the flip side, the tools are extremely easy to use and can produce convincingly realistic pictures even when the input data is inaccurate. The key areas to get right are sky definition and accurate material and physical dimensions.

Data needs to be accurate

Less than 10 years ago, books describing how to produce realistic images of buildings included sections on faking reality. The focus was on tweaking settings to achieve a realistic appearance, assisting operators to represent what they thought the room should look like, not on how real lights and real building surfaces might function.

Over the past 2 decades, computer models of light have begun to use physics-based renderers and interfaces, and increasingly programs require real-world values for parameters defining real light sources and for building properties.

With the earlier ‘artist’ approach to rendering light, the calculations might look realistic but not represent reality. Although using the new physics-based approach with accurate material properties and dimensions can produce realistic daylight predictions, use inaccurate data, and the results can be just as unreal as before.

Back to top

Real skies as models

The new software can model a far wider range of real daylight conditions experienced in buildings. Daylight design has in the past focused on cloudy skies because they are simpler to model. Before the 1950s, a cloudy sky was assumed to be of uniform brightness. This was replaced by the CIE standard overcast sky model, and techniques and formulae were developed to assist specialist daylight designers to estimate daylight factors (DF) for new building designs.

Perez sky-based renders of traditional clear sky (left), cloudy sky (centre) daylight plus the real world partly cloudy sky (right).
Accurate daylight render by Jake Osborne as part of his Master of Building Science from Victoria University.

The DF is the ratio between the inside and outside illuminance on cloudy days. For example, in a climate where the outdoor illuminance is greater than 6,000 lux, a 5% DF predicts that the light level indoors will be greater than 300 lux. If 6,000 lux is the minimum available outdoors for 80% of the working hours in a year, then a minimum of 300 lux will also be available indoors for 80% of these hours – so electric lights can be replaced for 80% of the working year in offices where 300 lux is the target.

For much of the last century, it was assumed this addressed a worst-case scenario, and as long as the office had sunshades keeping out direct sun, it would produce good daylight design. With physics-based daylight calculation software, designers can examine more than the worst case.

Back to top

From one calculation to 15 skies

At its simplest, the DF approach comprised one calculation or a scale model study for cloudy conditions to represent the whole year. Making scale models is time consuming, so the calculation was made simple enough to be repeated on different designs. More complex schemes tried to incorporate estimates of the CIE standard clear sky on representative days and times throughout the year.

The CIE recently supplemented the Cloudy and Sunny sky mathematical definitions from the 1950s with 13 more skies representative of those in real climates around the world. The challenge for the CIE is educating designers on using these 15 skies. Until this guidance is available, designers are unlikely to incorporate these skies into an analysis of daylight in their designs.

The standard mode of operation of 3D computer render programs today is unfortunately stuck in the last century, selecting only the simplistic CIE standard overcast sky approach for modelling. There is an alternative.

Back to top

Perez All-Weather model

Known as the Perez sky after the researcher who developed it, the Perez All-Weather model automatically selects and generates the most suitable sky type from parameters found in the annual climate datasets used for dynamic thermal modelling, for example, those for New Zealand developed by NIWA for EECA.

While the Perez sky is a huge improvement over the CIE standard clear and cloudy skies, there are two caveats:

  • Developed for weather data from Berkeley, California, it may not be suited to desert or tropical humid climates, but it should suit New Zealand.
  • It produces significant errors at very low sun angles so should not be used for calculation when the sun is below 5 degrees above the horizon, for example, before 9 am and after 6 pm in winter in our southern cities.

Back to top

Green building schemes

In green building rating schemes, daylight autonomy (DA) is used to rate daylighting approaches. DA reports not just that a particular light area is achieved, but what percentage of the floor area achieves this. For example, a point in a room near a window might be lit to a target of 300 lux for 80% of the working year and therefore its DA is 80%, but a point a long way from the window may only have a DA of 20%. The best daylight schemes have a high DA across a large proportion of the floor area and do not have excessive contrasts between parts of the room.

In 2000, new research proposed an improvement on this DF/DA approach. The researchers noted that the DA approach assumed that any DF that led to an illuminance indoors greater than the target was good. The corollary of this is no illuminance is too bright.

Buildings designed using this traditional approach, even with the sun excluded, risk being so overglazed that the shades are pulled all the time to control excess light, defeating the original purpose of the DF study.

Back to top

Minimum and maximum daylight useful

The Useful Daylight Illuminance (UDI) index brings common sense to these calculations, proposing two targets – a minimum and a maximum for the task. In an office, these might be 300 lux minimum and 2,000 or 3,000 lux maximum.

Variations in designs can now be compared on all the hours in a year, rendering the traditional cloudy day DF approach obsolete.

Combined with the UDI, statistics can be calculated about how often anywhere in a building daylight might be sufficient for the task and not too great as to cause glare. With the Perez All-Weather Model of the sky built into a program, it is now possible to calculate the UDI for every hour of the working day all year by linking in the appropriate weather file.

Back to top

Modern daylight design – what’s in the toolbox?

Software currently available can be reliably used to examine new and innovative ideas about daylight design. Used to its potential, it can accurately predict performance at a much more profound level than achievable with traditional methods.

Back to top

Note

This article is based on a keynote address presented at the International Commission for Illuminance (CIE) Congress in Shanghai in September 2012.

Download the PDF

More articles about these topics

Articles are correct at the time of publication but may have since become outdated.

Perez sky-based renders of traditional clear sky (left), cloudy sky (centre) daylight plus the real world partly cloudy sky (right).
Accurate daylight render by Jake Osborne as part of his Master of Building Science from Victoria University.

Advertisement

Advertisement