Testing cheap insulators on a home-made parabolic trough solar kettle
by Stephen Hewitt | Published
This article reports some empirical measurements of the insulating properties of silicone tubing and plastic bubble wrap on a horizontal copper pipe in open air.
These experiments were motivated by the search for a cheap insulator to help a home-made parabolic trough solar kettle heat water more quickly, but might also have other uses.
The kettle was described in a previous article. The horizontal 22mm copper pipe, 111cm long, holds about 350mL of water. At one end it is closed as shown in Figure 1 and at the other end it has a 3.5cm vertical spout with an open end.
Further articles reported that silicone tube on the copper pipe reduced the efficiency of the kettle on days without much wind. Plastic bubble wrap seemed to do the same, although the results were less clear because there were fewer trials. It remained to be tested whether they could be beneficial on windy days when heat loss from the bare copper would presumably be higher.
These articles are Trying silicone tubing as an insulator for a home-made solar kettle and Trying bubble wrap as an insulator for a home-made solar kettle in Related articles below.
It is hard to draw further conclusions from this because there are two unknowns. One is how much these insulators reduce heating by blocking sunlight and the other is how much they reduce heat loss. Empirical comparison (in England) so far has been hampered by the rareness of consistent sunlight on a windy day.
What is reported here is an attempt to separate the two unknowns and measure the insulating properties alone.
Procedure
Hot water in the hot pipe was allowed to cool in the absence of any heating while temperature was recorded at intervals, with different insulators and under different conditions.
This was repeated with the silicone tube, with the bubble wrap and with no insulation, both outside on a windy day, and indoors.
In these experiments a 1m middle section of the 1.11m copper pipe was insulated, corresponding to the section on the kettle between the supporting clamps.
For the indoor experiments, the water was heated in an electric kettle. First the pipe was filled with boiling water to heat it. After a few minutes the pipe was emptied and a new batch of boiling water introduced, the thermometer probe inserted and recording started. The pipe was not on the reflector of the kettle but was held horizontal in free air by similar brass clamps at each end.
For the outdoor experiments the pipe was on the reflector and pipe and water were heated in the usual way by the sun. When the water was near boiling the kettle was moved to face away from the sun (for example to face north) to start the cooling.
In all cases the temperature was measured as reported in previous articles. A KT-type bare thermocouple was immersed in the water being heated, approximately halfway down the copper pipe. Care was taken not to disturb it during the measurement period.
Results
Results are shown in Figure 2 and Figure 3, with raw measurements in Supplementary data below.
Indoors using hot water from an electric kettle, it proved problematic to get a high starting temperature, especially in the case of the bare pipe. Reasons include the time taken to empty and fill a narrow pipe.
The equivalent power per unit length for a given rate of temperature loss was calculated with the same assumptions and approximations as in the earlier article A home-made solar kettle, its performance and its problems. That means that effects caused by differences at the ends of the pipe were neglected. These differences are that the ends of the pipe in total protrude 11cm beyond the 1m length of insulation and one of them includes a 3.5cm vertical spout. The pipe was treated as a uniform pipe filled with water and the rate of temperature loss was assumed to relate to the heat capacity of a cross section of water and copper in that uniform pipe. The derived figure therefore is approximate. The figures used for the specific heat of water and copper are in that article.
Discussion
Indoors in still air, the bubble wrap reduced heat loss a little and the silicone made almost no difference, at least over the temperature range tested.
Some of the plots show a sharp rise in rate of heat loss over 80°C. This could be suspected to be an artefact of the experiments or the calculations because it occurs here only on the point at the highest temperature plotted. For the indoor measurements, one speculation is that recording started before the thermal capacity of the pipe had been filled by the hot water just introduced. A first batch of hot water had been in the pipe to pre-heat it but the unavoidable delay in emptying and refilling the pipe allowed some cooling.
However this cannot be an explanation for the similar jump for the curves recorded outside on 13 November 2022 because then the heat had come from the pipe and not from the water and they had been in thermal contact during solar heating for many minutes. It could perhaps be an artefact of moving the kettle to halt the heating, although it is not obvious how such an effect would be generated. Perhaps a better way of doing this kind of experiment would be to arrange a curtain or a screen to block the sun without the need to touch the kettle.
The curve of cooling with the silicone tube indoors on the 25 October is notably consistent with the repeat the next day. The reason for the anomalous spike in cooling rate near 70°C on 26 October is unknown.
Further work would be needed to investigate heat loss at higher temperatures above 80°C, including any sharp increase in heat loss.
To start the test indoors at temperatures closer to 100°C a future experiment could either have a different way of heating the pipe and water or an extra, strongly insulating sleeve for the pipe (especially the bare pipe) so that it can be pre-heated and remain hot while it is then refilled with new boiling water and allowed to come into equilibrium before the sleeve insulation is removed to start the test.
From the results on a windy day bubble wrap is a better insulator than the silicone tube and both of them significantly reduce the heat loss from the copper pipe in windy conditions but they don't seem to make much difference in still air, at least in the temperature range compared.
This is consistent with the earlier observations that they both reduce the performance of the kettle on days without wind.
Supplementary data
The raw data for each of the curves shown in Figure 2 is available. It consists of the following files with the following SHA256 hashes at the following locations on this website.
b8d8b4f2fc7dc8287d8254e8290a17a21c22482dafe3d9c223dd99c98e9880cc /download/windy.bare.13nov2022.csv 0a3a0f337092bb09e2c5470c6fb1e35030964eda10dcd160bdf7ff36fc6fc941 /download/windy.bubble.13nov2022.csv cc0736986d7768644f60c00822ac1f2b27552d681790a03dc998df2aee6b8ce3 /download/windy.silicone.13nov2022.csv dd233bf40dced28519d4cb63c58b27796260505495a4b01602a90e16915b4cc4 /download/indoor.bare.25oct2022.csv 13bc285585342046652f2d49fbcc7bfad26985942e324fff3205298997dc85f7 /download/indoor.bubble.17nov2022.csv bbdbc829d0325d2b7793b23e45ecf97e383e6599522f9bf9e196a59f334f74b7 /download/indoor.silicone.25oct2022.csv 64cdd2932305f9c353691468a0ba911398acbfc9c616271142315f0f0c76b733 /download/indoor.silicone.26oct2022.csv
These files are in CSV (comma-separated variable) format, meaning that each data point is represented by one line in the file, with fields on the line separated by commas. The fields for each data point are:
minutes,seconds,temperature
The minutes and seconds are wall-clock time (approximately GMT). When the minutes exceed 59 they roll over to zero. Hours were not recorded but it can be safely assumed that if the minutes on a line are less than the minutes on the previous line then the wall-clock hour has incremented by exactly one.