With the setup as previously described here, we performed a first run of mid-scale experiments.


The first test of the entire installation was done by trying to fill steam into the envelope naked without any insulation. The total condensation rate, according to the figure we obtained from the small-scale tests, should have been about 14 kg of steam per hour. But we didn't get very far with this test, for reasons that are not absolutely clear.... Even with all four of the boiler elements operating (nominal total 12KW, i.e. in theory 18+ kg of steam per hour) neither of the envelopes could be satisfactorily inflated. They swelled up to a certain point, about 2/3 full, but could not be brought beyond that point.

Now these envelopes have areas a little less than three times the area of the second envelope used in the small-scale tests. It will be recalled that it was impossible properly to fill that envelope with 4KW of boiling power. Therefore it is (just) understandable that this envelope cannot be filled using 12KW. The truth is, there must be various small losses in the system which are not obvious... the insulation of the boiler and of the supply tubing is not perfect; there may be some small steam leaks; and, which is more likely, at 100oC the heater elements may not be quite up to their nominal ratings. In any case, without anything seeming drastically amiss, the steaming power was not quite up to filling either envelope (A) or (B) in the naked state. Therefore this first naked test was abortive.

It was fairly clear as we watched that a modest further supply of steam would have done the trick, and we did contemplate adding one or two further boiler elements to augment the power of our boiler, but we abandoned the effort, because (a) we already have a figure for the naked condensation rate, obtained from the small-scale tests; (b) the power supply in our testing facility would not have taken the strain. We may come back to doing this at a later stage however, for the sake of completeness....



We next made an insulating jacket of bubble wrap ("buffer" for our Japanese readers). This was "Aircap" bubble wrap, of a smaller type than before, nominal bubble diameter 10 mm, approximate thickness 0.7 cm, weighing about 30 gm/m2, marketed by Armstrong in London.

We made this jacket according to the same pattern as Envelope (A) (which was the envelope used), sticking the pieces together with tape with the bubbles facing towards the inside. Building the jacket and making it fit was much easier than with the small-scale experiments, and this confirmed our ideas in employing the simpler pattern.

Here is a picture of the completed jacket upon the envelope:

We then started up the boiler and were easily able to fill the envelope with steam, finding that the use of three heater elements (9KW total) provided more than sufficient steam.

We collected the condensate every half hour. The results were;

First 30 minutes, 3185 gm condensate collected;

Second 30 minutes, 3075 gm condensate collected.

Third 30 minutes, 3075 gm condensate collected. Exactly the same!

There are no corrections to be made, in contrast to the small-scale experiments. The envelope area is 8.85 m2, so we get the result:

Condensation rate - 700 gm/m2.hour

In other words, about half of the naked value. It is interesting that this small-sized bubble wrap seems to be marginally more effective for insulation than the large-bubble version tested earlier, which weighs more per square meter. This is not intrinsically impossible or even unlikely, because, although the large-bubble wrap retains an air layer which is much thicker, its bubbles are larger and therefore much more subject to internal circulation, and this effect is by no means linear. We do intend to repeat the test of large-bubble wrap upon this mid-scale rig at some stage, so as to compare oranges with the same general type of oranges....



For the next test, we did not use an insulating jacket. Instead, we prepared several hundred hexagonal pieces of "Mylar" type aluminized plastic film and glued them (using a thin layer of bathroom silicone; what else?) upon the surface of envelope (B) so as more-or-less to cover it.

We then filled in most of the cracks with smaller cut pieces. Here is a picture of the final silvered envelope, hanging deflated upon the steam supply pipe:

We then started up the boiler and were easily able to fill this silvered envelope with steam. We used all four heater elements (12KW) for filling it, and thereafter three (9KW) were able to keep it full and taut:

(The few small non-silvered areas which were left because of our laziness did not constitute a substantial proportion of the total.)

This was the first time that we had been able to appreciate a fairly large envelope, inflated hard and full of steam. It was quite impressive! It really does convince one that steam is a substance when one can thump on the envelope, tight as a drum, and hear it twang.... Actually, most people don't believe, deep down in their guts, that steam is a substance at all.... they confuse it with the visible mist around the spout of a tea-kettle....

We collected the condensate every half hour. The results were;

First 30 minutes, 4300 gm condensate collected;

Second 30 minutes, 4360 gm condensate collected.

Third 30 minutes, 4330 gm condensate collected.

Fourth 30 minutes, 4360 gm condensate collected.

Since now the envelope area is 9.28 m2, so we get the result:

Condensation rate - 935 gm/m2.hour

This is a considerable reduction from the naked value, considering that in principle this metallized plastic film coating should only weigh about 25 gm/m2. Moreover, this result has particular significance for two reasons. First, this film layer is not a loose jacket upon the envelope but is integrated with it and has very little bulk, and accordingly endows the envelope with this substantial insulation performance without making it any harder to handle. Second, the film layer is also beneficial for making the envelope less permeable.

And this suggests a further test, which we have upon our agenda but have not yet performed. I have been told that a few very large hot-air balloons are provided with reflective "Mylar" coatings on the insides of their upper portions, and that this is very effective for reducing heat loss. (This doesn't actually seem to be general knowledge in the ballooning world.) This suggests that we should try a test the same as the above Test Three, but with the reflective coating on the inside. If the benefit for reducing heat loss is comparable, this is important information, because a film coating on the inside of the envelope will be even more effective for helping to keep it gas-tight.



Now, in all the above-described mid-scale tests we had set the superheater control so that the steam coming out of the superheater was at about 105oC to 109oC. This ensured that it was dry without being much superheated, since some heat loss could be expected in the piping enroute to the envelope. However, in one of our sources we had seen the suggestion that superheated steam might be much more reluctant to give up its heat and condense, than steam at saturation temperature, so that, paradoxically, there might be a higher condensation rate if the steam were hotter. Therefore, after the above envelope covered with metallized film was inflated, we turned up the superheater to the maximum, so as to supply steam at about 130oC. And we pierced a small hole in the side of the envelope and inserted a 1 meter long temperature probe, so as to find the steam temperature at various depths.

Basically the results were negative. The temperature inside the balloon was almost exactly 100oC everywhere - very slightly elevated - and the condensation rate was virtually the same as before. We conclude that the hypothesis is without foundation.



We next made an insulating jacket as described here of Primaloft PL1 insulation material as described here, nominal weight 133 g/m2, We made this jacket according to the same pattern as Envelope (B).

Here is a picture of the completed jacket upon Envelope (B), hanging down before inflation:

We then started up the boiler and were easily able to fill the envelope with steam, finding that the use of two heater elements (6KW total) provided more than sufficient steam for inflation, and a single element (3KW) was able to keep the envelope taut once it was full.

We collected the condensate every hour now, because there was much less of it. When we saw how little water was dripping out, even with this rather crude envelope whose performance could certainly be substantially improved, the practicability of a Steam Balloon was strongly brought home to us. The results were;

First hour, 2730 gm condensate collected;

Second hour, 2580 gm condensate collected.

Third hour, 2355 gm condensate collected.

Therefore we obtained the result:

Condensation rate - 275 gm/m2.hour

This is a very effective insulation performance.... We noticed an important fact: the weight of the jacket was of course supported by the envelope, so that the upper portion of the jacket (its "northern hemisphere" above its "tropic of Cancer") was pulled relatively taut against the envelope. And the central portion of the jacket (its "tropical portion" around its "equator") was fairly tight, because the jacket had little spare material around its "equator" so that the envelope was swelled up quite tightly within it. Over these two portions the jacket felt mildly warm to the touch; one could feel the heat coming through it. However, the lower portion of the jacket below its "tropic of Capricorn" was naturally hanging somewhat loose and away from the envelope, and over this portion the jacket did not feel warm at all - very little energy was getting through. This was presumably due to the additional layer of dead air between the jacket and the envelope over this portion. In fact, of course, the above condensation rate of 275 gm/m2.hour is an average value over the whole envelope, and includes this dead air layer effect for the envelope lower portion. This does not however yield an unrealistic final value, because the same effect would be present at full-scale with a man-carrying balloon.


We are aware that further tests are called for.... Watch this space...

A summary of the results of our experiments can be found here, along with extrapolation of their implications.


Back to the mid-scale experiments page...

Back to the experiments page...

Site map

Back to the top page...