In addition, there is the laborious work of transposing the real results into a table that can be read. I decided to do it for the daily figures. Now that the algorithm is working, it could be adapted to a simpler simulation using weekly figures later.
So I thought I would try it myself, using my own computer programming knowledge, and using real data from the last year. I have written 3 books on programming in GDL, so this is something I enjoy doing and it's a challenge, but not impossible. The understanding of the process and the interactions of the elements are still the more difficult task.
I have tried to evolving algorithms that match the real behaviour of each element in the process, and have tried to discover constants that give the results that the actual borehole delivered. This is still a work in progress, but the results are better than nothing - there is progress.
- The GSHP daily consumption figure is used, and if we assume a COP of, say 2.7, that determines the amount taken from the ground. One small error is that my GSHP figure includes the winter overhead of the circulating pump, but that is small in comparison with the main task of evolving an algorithm. There are also complications, in that the COP of the heat pump increases when the deep ground is warmer, so perhaps I should make COP better if the Sunboxes are working.
- The Sunboxes contribute heat through the energy flow meter, so that quantity is known - approx 3,000 kWh per year. This is the easiest part of the whole program.
- The Infinite ground around also tops up the borehole, in fact over a year, it has to contribute about 6,000 kWh, twice as much as the Sunboxes. This is a sliding variable, and not linear. Thus if the ground is very cold, the speed of thermal restoration is very fast, if it is merely cold, then it restores slowly. If the borehole is about 12.8º, there is neither heat gain nor heat loss. If the ground is above 14º due to the work of Sunboxes, some of that heat will be lost to the outside and not recovered. This heatloss will prevent the ground rising to something dramatic like 17º. My algorithm takes all this into account.
- Missing days?: Where I have been away, for example, holidays or conferences, the main spreadsheet calculates the daily average during the time away, so it is possible to fill in the missing days.
As we are in April 2011, I have had to use data from May, June and July of 2010, even though we had a cold May 2010, and the Sunboxes did not have mirrors installed. But for the time, it is good enough to work the animation. When I get time, I will add in for August and September 2010 too.
I am running two simulations, one of the system running without Sunboxes, and one with. A complication is that in the real case, my borehole started from a position of having had an entire summer of solar charging, whereas the simulation assumes the same standing start for both simulations. The previous year, Winter 2009-2010 had no such charging. So I should find a way to compensate for this.
What this is now showing is near to the real result - that without the Sunboxes, the ground needs all of the summer until September to recover its temperature. With the Sunboxes, the winter and spring have more stable temperatures, and during the summer, the ground gets warmed up nicely.
• Video of the simulation, assuming Sunbox Solar charging, assume residual heat from previous summer
These are still not quite right, but I have other work I must turn to so will come back to these.
In the Solar charging one, the real-life temps never got below 10º in winter due to the accumulated heat from previous summer. In the previous winter, with no solar charging, the real-life ground temp went down below 5º in the coldest point.
This simulation uses rules of thumb for expanding and contracting the volume, but I shall try a simulation which really does use Kilowatt Hours in and out.