Skip to main content
Version: Latest

Scheduled water and heat sources in a closed box

One point supplies water and heat while another extracts them, over time windows that end inside time steps.

Largest stored-volume error: and largest stored-heat error: of the final net supply.

At a glance
Dimension3D
Buildmake ndims=3 from the poromechanics source (not yet merged; DynEarthSol #134)
Inputssources.zip
Reproduce withDynEarthSol 3774df0

Problem setup​

Setup

Closed domain. Unit 3D box; all water and conductive-heat boundaries sealed. Normal motion fixed on every face. Initial pressure 0 Pa, temperature 300 K.

Two signed sources. Point 1: (0.37, 0.43, −0.53) m, fluid +8×10⁻⁸ m³/s and heat +4000 W, active 0.25–2.75 s. Point 2: (0.61, 0.57, −0.41) m, fluid −3×10⁻⁸ m³/s and heat −1500 W, active 0.8–2.1 s.

Isolated balance. κ = λ = 0; mechanical pressure feedback and thermal expansion disabled. This case compares integrated volume and heat; it does not claim a spatial point-source solution.

Input and how to run​

Every file this example needs is in sources.zip. Run from that folder:

# Runs the closed 3D box for 4 steps of 1 s with the two scheduled water and heat sources. DES writes fresh.save.* frames at every step.
dynearthsol3d input.cfg

Settings that define this example​

SettingValueWhat it does here
control.has_hydraulic_diffusiontrueturns on the pore-pressure equation that receives the water sources
control.has_poroelastic_pressure_feedbackfalseoff: deformation does not change the pressure, so stored water is storage times pressure change
control.has_fluid_thermal_advectiontrueturns on groundwater heat transport with the combined rock and water heat capacity, which receives the heat sources
control.fixed_dt1steps of 1 s, so the source windows start and end inside steps
sim.max_steps44 steps of 1 s, to 4 s
mat.hydraulic_perm[0]zero permeability: water stays where it is supplied, so only the totals are tested
mat.therm_cond[0]zero conductivity: heat stays where it is supplied
mat.alpha[0]no thermal expansion of the solid (mat.fluid_alpha is also 0), so heat does not change the pressure
mat.fluid_heat_capacity[4000]4000 J/(kg K); with the rock values it gives C = 2.95e6 J/(m³ K)
injection.points_x[0.37,0.61]x of the two source points; points_y and points_z complete their positions (heat_injection uses the same points)
injection.rate[8e-08, -3e-08]water rates +8e-8 and −3e-8 m³/s: the first point supplies, the second extracts
injection.start_time_in_yr[7.922191133769641e-09, 2.5351011628062857e-08] yr = [0.25, 0.8] sstart times 0.25 s and 0.8 s, in years (also used by heat_injection)
injection.end_time_in_yr[8.714410247146606e-08, 6.6546405523665e-08] yr = [2.75, 2.1] send times 2.75 s and 2.1 s, in years
heat_injection.rate[4000, -1500]heat rates +4000 W and −1500 W at the same points and times

input.cfg holds all 68 settings that differ from DES’s defaults; dynearthsol3d -h lists every option with its default.

Output​

See Running the examples for output files, the Python reader and ParaView conversion.

FieldWhat to look at
pore pressurechange since frame 0, times the nodal storage, gives the stored water
temperaturechange since frame 0, times the nodal heat capacity, gives the stored heat
coordinatewith connectivity, gives the element volumes shared out to the nodes
connectivityelement-to-node map for the nodal volumes

Result: DES and the analytical solution​

Result

(a) Net water supplied to the closed box, in µL: the DES storage change from the saved pore pressure (circles) against the time integral of the two signed source schedules (line). (b) The same for heat, in kJ, from the saved temperature and the combined heat capacity.

This checks that every source acts exactly once over each integration interval. It does not test the spatial point-source solution.

To reproduce the whole study behind this figure, from the DES source folder:

python3 tests/functional/hydromechanics/check_point_sources.py dynearthsol3d OUT3D --dimension 3
python3 tests/functional/hydromechanics/check_point_sources.py dynearthsol2d OUT2D --dimension 2

The study’s full numbers: record.json.

Analytical reference​

and are signed fluid-volume and heat rates, positive for supply; and are start and end times in seconds. The plotted run has J/(m³ K), in 3D. In 2D the in-plane mean-stress split uses and rates are per unit out-of-plane thickness. At 4 s the exact net supply is 1.61×10⁻⁷ m³ and 8050 J.

Derivation​

A point rate is shared among the nodes of the element that contains it by the element shape functions. They sum to one, so the nodal loads add up to the requested signed rate. A source window that starts or ends inside a time step is integrated over its exact overlap with the step, so every source acts exactly once.

Summing the lumped nodal update cancels internal transport contributions. With sealed boundaries, fixed capacities and the mechanical source disabled, only the prescribed source integrals remain. Apply the same argument to thermal mass and heat rate. A total-amount schedule uses rate equal to amount divided by (bi−ai).

Direct time integral of the supplied schedules plus the assembled closed-domain storage/heat balances; independently evaluated from simplex volumes and saved fields in check_point_sources.py.

Validity and limits​

This is an exact integral balance for an isolated component test with fixed capacity and mechanical pressure feedback disabled. It does not qualify the spatial singular solution, nonlinear storage, the enthalpy of injected fluid (see the injection-temperature page) or conservative remesh transfer. Fluid volume and heat sources are independent inputs. Separate forced-remesh runs check source relocation and continuation.