HomeFeaturesExamplesTutorialsHelpTheoryVerificationPricingContact
BoreTherm / TutorialsTutorial 14 / 19
Workflow and theory · Advanced physics

Borehole thermal energy storage

How BoreTherm simulates a seasonal store: the operation series, the reversed flow, the ground coupling per ring, and how the Storage page counts the heat that goes in and comes back. Example project: 37 boreholes of 50 m, 3 m apart in a hexagon, charged May–September at 70 °C, discharged November–March at 30 °C, 3 years from April.

01 / The one idea

The one idea: store heat in summer, take it back in winter

A borehole thermal energy store (BTES) is a compact field. In summer hot fluid enters at the centre and heats the ground. In winter the flow is reversed: cooler fluid enters at the edge and leaves through the hot centre. The store keeps a radial temperature gradient: hot in the middle, cooler towards the edge, where the heat losses happen.

Plan of the store (the example)hot centre · rings 0–3 · cooler edgegreen: one of the six strings, charging directionCharging (summer)hot fluid entersat the centreDischarging (winter)cool fluid enters at theedge: flow reversedThe storeground between the boreholesEinEoutEin − Eout: heat left in the ground,mostly around and below the storeRecovery η = Eout / Ein
Left: the 4 rings of the hexagonal field and one of its six strings, which runs ring by ring from the centre outwards. Right: what the Storage page counts. Not all heat comes back; the rest stays in the ground around the store.
Heat into the ground in a stepQ = ṁ cp (Tin − Tout)
Charged and dischargedEin = Σ max(Q, 0) h
Eout = Σ max(−Q, 0) h
Recovery of a yearη = Eout / Ein

Q > 0 is heat into the ground (ṁ mass flow, cp specific heat). The sums run over the control steps with flow of one operating year (step length h). How a store differs from a normal heating and cooling borefield (Tf: mean fluid temperature, the mean of inlet and outlet; Rb*: resistance between fluid and borehole wall; g-function: the field’s temperature response to a constant heat rate; SPF: seasonal performance factor):

Heating / cooling borefieldThermal energy store
Loadsmonthly or hourly ground loads (or building loads through the SPF)a time series of inlet temperature, flow and direction; Q is a result
Borehole modelsteady (Rb*), optional short-time capacitydynamic: transit time, depth levels, grout capacity; steady pipe model in between
Groundone wall temperature for the field (g-function)one wall temperature per ring or per borehole: hot centre, cool edge
Flowone direction, pump always onreversed for discharging; pump off in rest months
Design questionTf within the limits; Calculation › Size borehole length sizes the lengthhow much heat comes back, and at what temperature
Main resultsTf,min, Tf,max, required lengthcharged, discharged, recovery per year; T in / T out; ring temperatures
02 / The workflow

The workflow

The design starts from the layout and the operation. The operation is a time series: BoreTherm prescribes the inlet temperature (or the load), the flow and the flow direction step by step, and calculates how much heat the ground takes or gives.

Field layouthexagonal or concentric rings; BTES pattern, Centre → edgeIs there a measured orplanned operation series?noyesLoads › Generate BTES operation…each month: charge, rest or discharge;inlet temperature (or load), flow,daily hours; one year, repeatedLoads › Time seriesimport a CSV: time, inlet temperature(or load), flow or mass flow, direction+1 charging, −1 reversed, 0 keepDesign limits & simulation › Dynamic modelInlet: temperature given · Flow and direction: from the time seriesGround coupling: per ring or per boreholeDynamic periods: whole period, custom periods or steady model onlyCalculation › Calculate, then Results › Storage / Dynamic modelper year: charged, discharged, recovery; per ring: Rings tab
Without a measured series, Generate BTES operation builds one year (365 days, from the start month, repeated every year). It can also switch the borehole model to Dynamic with the inlet, flow and direction taken from the series.
Field layoutrings around thecentre; patternLoadsTime series orGenerate BTES…Design limits &simulationDynamic model boxCalculatesteady model +dynamic periodsStorageper operating yearDynamic modelseries, profile, ringsResults
The pages involved. Storage and Dynamic model are result pages; they show results only when the borehole model is Dynamic.

Settings that matter

  • Connection pattern (Field layout): Radial strings, Sectors or Ring-wise, all with the direction Centre → edge. The network as drawn is the normal direction, used for charging.
  • Ground coupling: Uniform gives the whole field one wall temperature and averages the hot centre and the cool edge. Per ring resolves them at little extra cost. Per borehole is for unsymmetric networks; it is slow for large fields (above 300 boreholes the program asks first).
  • Flow direction: From the time series (the direction column), or Automatic: normal when the inlet is at least 1 K warmer than the fluid of the inlet stage, reversed when it is at least 1 K colder than the outlet stage; at least 1 h between two reversals.
  • Dynamic periods: the dynamic borehole model is slow. The example runs the steady per-borehole model hourly for the three years and the dynamic model only from day 940 to 954, plus one day of warm-up.
03 / How the calculation works

How the calculation works

BoreTherm steps through the whole period at the control step (1 h in the example). Every step does the following.

1 Read the stepinlet T, flow, direction(the step means of the series)2 Direction changed?reverse every string: cellsre-indexed, no fluid movedin a dynamicperiod?3a Dynamic boreholeplug flow, depth levels,grout capacity3b Steady pipe modelper borehole or group;no transit time4 Group wall temperaturesall groups solved together:history + responses gij5 Fluid resultQ into the ground, T out;loads pushed to the ground6 Storage summary of the stepyear ⌊t / 8760 h⌋ · Q > 0 charged, Q < 0 discharged · hours · T in / outper ring: wall heat; wall temperature at the end of charging / dischargingyesnoGreen: the next step. At the end of the run: recovery per year and from year 1 on (Results › Storage).
One control step. Inside a dynamic period the borehole interior is resolved (fluid cells that move with the flow, grout nodes on 25 depth levels of 2 m); outside it the steady pipe model of each borehole is used. Both push their heat into the same ground history, so the ground sees one continuous record.
  1. Read the step. Inlet temperature, flow and direction are the means of the series over the step. A mass flow is converted with the density at the inlet temperature; a flow per borehole is converted to the pump flow of the network.
  2. Reverse if needed. At the first step with flow in the other direction every string is run backwards: the fluid cells are re-indexed, so the fluid that entered last leaves first. With the pump off the direction is kept.
  3. Borehole model. Dynamic inside the dynamic periods, steady pipe model per borehole outside.
  4. Ground. Each group (ring or borehole) has its own wall temperature, uniform along the depth. It follows from the group's history and the responses gij: the temperature rise of group i caused by a unit heat rate on group j. All groups are solved together.
  5. Fluid result. Q, Tout and the mean fluid temperature of the step.
  6. Storage summary. The step is added to its operating year by the sign of Q.
04 / The theory

The theory: why the first year is poor and the centre matters

A new store must first warm its own ground and the ground around it. That heat does not come back in the first winter, so the recovery of year 1 is low. Each year the surroundings are warmer, the losses smaller, and the recovery rises.

Monthly mean fluid temperature, three years061218243036020406080month of operation (from April)Tf [°C]charging, 70 °C indischarging, 30 °C inrest months: white
Monthly mean fluid temperature (the last hour of each 730 h month) from the stored result of the example. Orange: charging months, blue: discharging months. The store gets a little warmer at the end of each discharge.
Charged and discharged heat per year010020030040050060058092year 1426138year 2401152year 3MWhEin chargedEout dischargedRecovery η = Eout / Ein12301020304050operating year%15.9 %32.5 %37.9 %27.2 %solid: the year · dashed: from year 1 on
Left: heat charged and discharged per year. Right: the recovery of each year and the cumulative recovery from year 1 on.
The point that is easy to miss. Recovery is not a property of the store alone. It depends on the operation: the discharge temperature sets how far the store can be cooled, so a lower discharge temperature gives a higher recovery. In the example the store is discharged at 30 °C; every kilowatt-hour that cannot be taken out above that temperature stays in the ground. Year 1 gives 15.9 %, year 3 37.9 %.

Why the coupling per ring. The centre ring loses heat only to its warm neighbours; the outer ring loses it to the cold ground outside. With the coupling per ring BoreTherm sees this. With the uniform coupling it averages it away.

Ring wall temperatures, year 3012320406080ring (0 = centre)°Cend of chargingend of dischargingRecovery per ring, year 302040608059 %0 (1)56 %1 (6)46 %2 (12)27 %3 (18)ring (boreholes)%
Year 3 of the example (coupling per borehole, shown per ring). Left: mean wall temperature of each ring at the end of the last charging step and the last discharging step. Right: discharged / charged wall heat of each ring. The outer ring returns the least.

Flow reversal. In discharging the cool fluid enters at the outer ring, picks up heat ring by ring and leaves through the hot centre. The outlet is then as hot as the store can deliver. At a switch nothing is mixed: the fluid already in the pipes leaves first.

05 / How each result is derived

How each result is derived

The Storage page shows one row per operating year. The values are taken from every control step, not from the output series, so an output interval that mixes charging and discharging does not net them. The last column is year 3 of the example.

ResultHow BoreTherm derives itYear 3
Charged [MWh]Σ max(Q, 0) · h over the steps with flow; Q = heat the fluid gives to the ground in the step401
Discharged [MWh]Σ max(−Q, 0) · h over the steps with flow152
Recovery [%]discharged / charged of the year37.9
Cumulative recovery [%]Σ discharged / Σ charged from year 1 to this year27.2
Charging / discharging [h]steps with flow and Q > 0 / Q < 0 (by the sign of Q, not by the schedule)3672 / 3624
Pump off [h]steps without flow (rest months, hours outside the daily window)1464
Reversalssteps with flow whose direction differs from the last step with flow; the run starts normal2
T in min / max [°C]step means over the steps with flow30.0 / 70.0
T out min / max [°C]step means over the steps with flow30.8 / 64.2
Ring: charged / discharged (Dynamic model › Rings)wall heat of the ring’s boreholes, Σ Hb qw · h, split by sign; differs from the field values by the heat stored in fluid and groutring 3: 210 / 57
Ring: wall temperature at the end of charging / discharginglength-weighted mean wall temperature of the ring at the end of the year’s last charging / discharging stepcentre: 67.0 / 32.0 °C

An operating year is 8760 h from the start of the simulation, so it begins in the start month (here: April), not in January. A year with no charged heat has no recovery.

06 / The example in numbers

The example in numbers

Hexagonal field of 37 boreholes, 50 m, 3 m apart, six strings ring by ring from the centre. Operation from the series in the project: charging in 3672 h of the year at 70 °C, discharging in 3624 h at 30 °C, pump off 1464 h; 0.25 kg/s per borehole. Ground λ = 2.0 W/(m·K), undisturbed ground temperature Tm = 10.0 °C.

YearCharged [MWh]Discharged [MWh]Recovery [%]Cumulative [%]Hours charging / discharging / offReversalsT out range [°C]
15809215.915.93889 / 3407 / 1464129.9 – 62.3
242613832.522.93672 / 3624 / 1464230.6 – 63.9
340115237.927.23672 / 3624 / 1464230.8 – 64.2
  • Over 3 years: 1407 MWh charged, 382 MWh discharged.
  • Year 1 counts 217 h more charging than the schedule has. At the end of the first winter the store has cooled to about the discharge temperature (T out min 29.9 °C): the 30 °C inlet is then warmer than the ground, and Q > 0. The hours are counted by the sign of Q.
  • Recovery per ring in year 3: 59 %, 56 %, 46 %, 27 % from the centre to the edge.
  • Mean fluid temperature 10.0 – 67.1 °C: within the limits 2 / 80 °C. The minimum is the undisturbed ground before the first charge.
07 / Limits to keep in mind

Limits to keep in mind

  • Charged and discharged count all heat the fluid exchanges with the ground. BoreTherm applies no useful-heat threshold or return-temperature cut-off; heat discharged at a temperature the plant cannot use still counts. Judge the outlet temperatures (T out) together with the recovery.
  • The wall temperature of a group is uniform along the depth. Vertical stratification in the store is not resolved.
  • Outside the dynamic periods the steady pipe model has no transit time and no grout capacity. For short cycles (days) use the dynamic model over the whole period, which is slower.
  • Judge a store on its later years. Year 1 always includes the heating-up of the surroundings.
  • Size borehole length sizes the length to the fluid temperature limits only; there is no sizing for recovery. Compare variants (length, spacing, temperatures) with Calculate runs.
  • Tf,max also stands for the pipe material limit: in a storage operation BoreTherm warns when the inlet or outlet exceeds it, and when the fluid reaches its upper property limit (water 100 °C).
  • The 3 % gate on the discharged energy against the benchmark of an independent finite-element code is not met by all stored benchmarks; the Verification and Known issues pages of the help give the details.

Numbers calculated with BoreTherm. Theory: BoreTherm theory manual.