[jeff5may]

Originally Posted by buffalobillpatrick

I want the pump to burp tanks 2-10 this should get almost all of the air out.

Unless you have a tap in the high point of each series connecting pipe, it will be impractical to purge the air out. It doesn't really matter much unless you want your tanks to thermosiphon through the segments.

I think low flow drip emitters & bedding tanks in sand with EPDM below will increase heat transfer through tank walls.

As top of tank will only be about 3' below grade, I think that 3" of EPS above tanks will reduce winter loss.

Parallel would be hard to balance with tanks back filled.

Naturally, you would want to plumb, fill, and leak check the tanks before back-filling. If you had a header for pipes leading to the tops of the tanks and a header down low (or up top running to dip tubes), so that natural convection could occur, the parallel flows would balance due to the large horizontal cross-section and stratification due to density. With under a gallon per minute flow per tank, density and gravity would overcome your pressure difference.

I think that my 2 ton HP evaporator will need about 6gpm. I could use one of my 10W 3gpm 30VDC El-Sid pumps to continuously move heat between tanks.

If you go with a series connect, you would need the pump to overcome all the air in the tops of the interconnecting tubes. This would work against the "thermal battery" scheme the series connect offers, in effect somewhat swirling the source water a little bit between each tank while the pump is running. The tanks would sequentially charge or discharge heat, with corresponding sequential heat loss or gain with the ground. Looping discharge to intake would deplete your remaining neutralized source water.

The real difference here is not so much in efficiency or total heat transfer, because both configurations will net close to the same heat gain or loss. The big difference is how your supply would act.

With the series connect, any temp change introduced would take a long time to work its way through the tanks. Like a common water heater, you would have a predictable amount of constant-temp source water to draw from, then the source would quickly change behavior for the worse. It would then take a long time to recharge.

With the parallel connect, the source water would react more quickly to any changes. The source temperature would begin to stray sooner than the series connect, but it would exhibit a more gradual drop-off. When the flow was cut, the store would immediately begin to recover.

Here's a research paper that deals with this topic directly:

EDIT: http://ecorenovator.org/forum/attach...1&d=1412291207