I have fifteen years of temperature data for my high-mass, solar passive, unheated house at Manilla, NSW, Australia. This article has been posted previously here. These graphs show how July temperatures indoors relate to those outdoors. Indoor maxima and minima are not shown, because they are consistently between one and two degrees above and below the indoor mean.
In July, the rooms* in this solar-passive house, heated only by the sun, are much warmer than outdoors. This is shown by the green lines on the graphs, which are drawn to pass through the middle of each cloud of data points. The middle graph shows that, as an average over 15 July months, the rooms have been 8.7 degrees warmer than outdoors. The left graph shows that the rooms have even been 1.4 degrees warmer than the daily maximum outdoor temperatures. The right graph shows that the rooms have been nearly sixteen degrees warmer than the daily minimum overnight temperatures. To stay warm in this way the house must have absorbed many hundreds of kilowatt hours of heat from the sun. I have burned a few kilowatt hours of grid power to maintain my comfort, but this cannot have warmed the house by as much as one tenth of a degree in any month.
Which of the outdoor temperatures affect the room temperature?(dashed black lines)
On each graph there is a dashed black line and an equation (from Excel charts) showing how the room temperature relates to the outdoor temperature (maximum, mean, or minimum) that is plotted along the bottom. In the left graph the temperatures relate quite well. The points seem to line up from bottom left to top right. The equation shows that each extra degree of outdoor maximum temperature gives just over half a degree (0.55) of extra room temperature. Under the equation is an “R-squared” value of 0.47. R-squared indicates how close the relation is. (A perfect relation would give 1.0.) This value (0.47)is quite good for natural relations (but very poor in results of experiments). The middle graph shows that room temperatures relate poorly to outdoor daily mean temperatures. If there is a relation, it causes the room temperature to change by only 0.21 degrees for one degree outdoors. The R-squared value is an abysmal 0.09. The right graph, plotting room temperature against outdoor overnight minima, has a “shotgun” pattern, showing there is no relation at all between these temperatures. The equation, with a zero slope and a zero value for R-squared confirms this.
Why is it so?
In mid-winter, the house functions to receive and trap as much solar heat as possible in day-time, through large north-facing windows. When July is hot and sunny, the rooms get very warm: warmer than when the month is cool and cloudy. The fact that Manilla has a sunny winter climate favours the house. By night, the house is almost immune to outdoor temperature change. While wall and ceiling insulation values are not very high, the windows are double glazed, and they have heavy pelmeted curtains that are closed automatically. There is also a large thermal mass enclosed within the insulating shell, which stabilises the temperature from week to week. The room temperatures, as shown in the right graph, are very much warmer than the outdoor minimum temperatures – in July 2002, 20 degrees warmer. This is less a matter of house design than of climate. While the sunny winter climate makes days warm, it also makes nights cold. As a result, the house, which is able to trap a lot of heat during the day, and stays warm at night, then shows an extreme contrast with the frosty night air. The green lines on the left and right graphs show that July nights in this climate are 14.5 degrees colder than July days. In climates with winters that are less sunny (as Ballarat and Bathurst) the daily temperature range is narrower, and a solar-passive house will perform differently. It may be difficult to raise room mean temperature in July above the outdoor daily maximum there. In those climates, the room temperature cannot possibly be kept so high above the outdoor daily minimum. but it does not matter, because the minimum is not very low.
This post, with July (mid-winter) data for this solar-passive house is one of a pair. January (mid-summer) data are in the post “January “Coolth” in a House without Air-Conditioning”.
*Room temperature is mean daily room temperature, measured at head height on an internal stud wall. The daily temperature range at that point in July is only 3.2 degrees. Outdoor (ambient) temperatures are measured in a Gill-type screen.