In my research for The Solar House I came upon this statement by a British engineering professor in 1976:
"If during a 24-hr period the incoming heat is greater than the loss, the window is thermally profitable for that day. Furthermore, if during a year the incoming heat is greater than the loss, the window is thermally desirable. If overall the window gives a net positive gain, its size should be as large as possible. Conversely, if the long-term window net gain is negative, its best size is zero.… On a purely thermal basis there is no optimum intermediate size: The window is either maximum or zero."
The quotation is from Dr. Morris Davies*, an engineer at the University of Liverpool, who was part of a research team that spent about ten years analyzing the performance of the St. George’s School by Emslie Morgan (Wallasey, 1961-62), an immensely interesting building I wrote about in The Solar House.
What an excellent illustration of the difference between scientific optimization and design! Davies' optimized architecture would typically have an all-glass south wall, and all-solid walls on the east, west, and south. In my view, this is why computer simulations can contribute to architectural design, but they can not be generative or determinant of architectural design. Scientific knowledge is one important ingredient in a more complex recipe, and human judgments are essential.
Of course, Davies' position here did not consider that windows have other functions beyond thermal (and he admitted so). They may be desirable for providing daylight, or views, and they may be undesirable for bringing discomfort due to glare. St. George’s School did indeed suffer a glare problem, of which Davies was aware.
But even from a purely technical point of view, Davies was wrong. To focus on annual net gain is too simplistic; it neglects the fact that quite a bit of passive solar heat is collected in afternoons in the Spring and Fall, when it is likely to be unwanted because it causes the building to overheat, depending on the weather and other factors of course. To make a good window "maximum" is to increase the overheating problem. Any number of solar house architects learned that lesson the hard way, long before 1976, as I detail in The Solar House.
Note that overheating does not necessarily imply that the building needs air-conditioning. Opening some windows may suffice, especially on a swing-season afternoon. This was most certainly the case in Wallasey. Still, the solar gains at that time are, for the most part, not beneficial.
Clearly Davies was speaking theoretically, and perhaps he assumed that windows would be shaded at times when the building doesn't need heat. In practice, shading never works in a theoretically-optimum manner. Or perhaps he assumed that thermal mass lagtime and carryover effects would effectively redistribute the heat from day to night. In practice, this is also limited.
Nevertheless, I do enjoy the phrase "thermally desirable," and it wouldn't hurt if most architects adopted a little bit of that mindset when thinking about windows.
*Davies was quoted in Joseph E. Perry, Jr.. “The Wallasey School,” Proceedings of the Conference on Passive Solar Heating and Cooling, 1976.
Previously: Why I care about Building Science