Solar Gain?

An Eccentric Anomaly: Ed Davies's Blog

In my previous post I said I thought solar gain would not be a significant influence on the boiler run time. It's worth checking, though.

I don't have an easy way of measuring solar gain directly but the monitor measuring CO₂, temperature, humidity, etc, in my bedroom does have a light sensor. It's primarily intended to set the level of brightness of the device's LED so it's hardly a source of well-calibrate and linear data but it's at least indicative.

Because the room has a south-east facing dormer window and the sensor is on a chest of draws to the north of the window it tends shows a reasonable indication in the morning but drops significantly in its reading in the afternoon. Here's the curve for yesterday (2019-01-22 Tuesday) which was pretty bright most of the day:

By contrast, here's 2019-01-03 Thursday where the sunlight as seen by the sensor never got brighter than my 4W bedside LED, note the very different y-axis with roughly ten times the scale:

Who knows what units those light level measurements are in. Let's call them “fots”. To get an estimate of the solar gain for each day I integrated the readings getting a result in fot·seconds. Because the numbers were large I divided them by a million to get mega fot·seconds (Mfot·s). To avoid most of the effects of the bedroom lights I just did the integration from 09:00Z to 16:00Z.

This gave solar gains from 0.2167 to 6.1868 Mfot·s. My program then divides this range into four equal subranges (in terms of Mfot·s values) and classifies the days by which subrange they're in. For the same 2018 November and December days as my previous post:

The five days in the brightest solar gain range (the brown squares) are in the lower part of the boiler run proportions for their heating requirements but they're not necessarily the lowest for similar requirements. Beyond that, it's not really obvious what effect the solar gain is having.

You might expect more effect earlier in the year when the days were a bit longer and the sun a bit higher. Here's a similar plot for 2018 August, September and October. As the overall range of brightnesses is a bit different the ranges used for classification are also slightly changed:

Again, not much effect.

The two points down towards the lower right and the one on the zero-boiler run axis were days I was away so only used the heating in the morning I was going and the evening I got back.

In conclusion, solar gain isn't a big input to this house. I'd speculate that part of the reason might be that clear skies giving more gain in the day also result in greater cooling overnight which evens things out given its large thermal mass.