Heat travels from the ground to the loop fluid (or back) through three resistances in series, per metre of pipe:
R' = R_conv + R_pipe + R_grout [K·m/W]
R_conv = 1 / (h · π · d_in) (fluid → pipe wall)
R_pipe = ln(d_out/d_in) / (2π k_pipe)
R_grout = ln(r_b/r_pipe) / (2π k_grout)
h from Dittus–Boelter: Nu = 0.023 Re^0.8 Pr^0.4
Along the pipe the fluid temperature relaxes toward the ground temperature exponentially — the same "fin equation" solution used for a pipe losing heat to its surroundings:
T(z) = T_ground + (T_in − T_ground)·exp(−z / L_c)
L_c = ṁ·c_p·R' (characteristic decay length)
A slower flow (small ṁ) or a poor grout (small k_grout, big R_grout) shrinks L_c, so the fluid equilibrates faster — it leaves closer to the ground temperature, but moves less total heat because ṁ is small. A faster flow moves more heat per trip but each parcel spends less time in contact, so it approaches the ground temperature less closely. This trade-off is exactly why ground-loop design optimizes flow rate rather than just maximizing it.
- Mode — heating pulls heat out of the ground (fluid enters cold); cooling rejects heat into the ground (fluid enters warm), as in a real reversible heat pump.
- Flow rate — sets both the convective resistance R_conv and the residence time, so it moves the exponential decay length L_c.
- Grout conductivity — a poured grout with more solids (e.g. thermally-enhanced grout) lowers R_grout and raises heat transfer for the same pipe.
Simplification: this model treats each U-tube leg as a single pipe centered in the borehole and ignores "thermal short-circuiting" (heat leaking directly between the down-leg and up-leg through the grout) — real design software adds a correction for that, but the resistance stack and exponential approach shown here are the same physics used in first-pass borehole sizing (Kavanaugh & Rafferty method).