Walkthrough: Design a District Energy System (Heating + Cooling)

A district energy system (DES) couples many buildings to a shared thermal infrastructure so that heating, cooling, and domestic hot water are produced at higher efficiency than each building could achieve alone. The economics depend on three structural facts: (1) waste heat in cities is enormous and persistently underused — server rooms, sewers, supermarkets, ice rinks, industrial flue gas; (2) the ground is a thermal battery that can be charged in summer and discharged in winter; (3) the marginal cost of electricity, not natural gas, now sets the price of heat in most temperate developed economies. This walkthrough takes a 1-million-m² (10.76-million-ft²) mixed-use research-housing-commercial campus in ASHRAE Climate Zone 5A (Chicago- or Boston-like) from program brief to commissioning and traces the engineering choices, real OEM equipment, energy modeling, controls, regulatory hurdles, and economics that decide the project.

1. Project Brief and Loads

• Site area: 1,000,000 m² gross floor area (10.76 million ft²) across 40 buildings.
• Mix: 45% research/lab (high cooling load, simultaneous heating reheat), 35% residential (high DHW, low cooling), 20% commercial/retail/conference (peaky cooling).
• Climate zone: ASHRAE 169 Zone 5A — cooling design 33 °C dry bulb / 23 °C wet bulb (91 °F / 73 °F), heating design −20 °C (−4 °F), heating degree days 6,500 °F·d, cooling degree days 1,000 °F·d.
• Peak space heating: 25 MW thermal (85.3 MMBtu/h).
• Peak space cooling: 30 MW thermal (8,530 ton, 102 MMBtu/h).
• Annual heating energy: 45 GWh-th (154,000 MMBtu).
• Annual cooling energy: 35 GWh-th (120,000 MMBtu).
• Annual DHW: 8 GWh-th (27,300 MMBtu).
• Simultaneity factor: research buildings drive a 12 MW year-round heat-recovery base load — this is the most valuable load in the entire campus.
• Connected peak electrical load (HVAC only): ~6 MW with COP averaging.

2. Generation Architecture — Choosing the Generation

District energy generations are a clean way to frame the design space.

• 1st generation (1880s–1930s): saturated steam at 200 °C / 14 bar, cast-iron mains, condensate return. New York Steam Company still operates this in Manhattan.
• 2nd generation (1930s–1970s): pressurized hot water 100–200 °C, steel pipes, fuel oil/coal CHP.
• 3rd generation (1970s–2010s): hot water 80–120 °C, pre-insulated bonded steel pipes, gas CHP, peaking boilers. Most existing European DH systems.
• 4th generation (4GDH): 50–70 °C supply, 25–40 °C return, designed around low-temperature radiators and condensing return temperatures that enable flue-gas heat recovery, gas CHP, large heat pumps, biomass. Copenhagen, Vienna, Helsinki post-2010 upgrades.
• 5th generation (5GDH/C, anergy/ambient loop): 8–25 °C bidirectional loop, distributed building-side water-to-water heat pumps boost or reject to the loop. ETH Hönggerberg (Zurich, commissioned 2013 with progressive expansion through 2025), EPFL Lausanne, ETH Hönggerberg’s “Anergy Network.” Suurstoffi (Risch-Rotkreuz, Switzerland) is a fully built-out 5GDH residential-commercial neighborhood.

Decision: 5GDH ambient loop. Reasons:

1. Simultaneous heating and cooling on a research/lab/data-center campus means the loop is naturally balanced — one building’s condenser water is another building’s evaporator water. A 4GDH hot-water system cannot recover this directly without a second cold network.
2. Loop temperatures of 8–25 °C remove the need for high-grade pre-insulated steel mains; PE-Xa or PE-RT pipe rated PN 10 to PN 16 is permissible because the temperature is below polymer derating limits and heat losses to a 12 °C ground are negligible (<2% over the network).
3. The borefield doubles as inter-seasonal storage. Surplus cooling in summer charges the ground; the ground gives that heat back in winter through the building heat pumps.
4. Building owners install their own water-to-water heat pumps (the “substation”) and pay only for thermal energy transferred from/to the loop. This decouples thermodynamic risk from infrastructure risk.
5. Future-proof: as buildings electrify and densify, the loop accepts any new heat source (process waste, data center, sewage HX) without re-temperaturing the network.

Trade-off accepted: peak electrical load is higher than a 4GDH system because every kWh of heat at the user point passes through a heat pump. This is paid for in lower distribution losses, no peaking gas boilers (almost), and the ability to monetize cooling that 4GDH cannot.

3. Heat Sources and the Energy Balance

The annual energy balance has to be solved before any pipe is sized. The campus needs +45 GWh of heating + 8 GWh of DHW (+53 GWh-th) and −35 GWh of cooling rejection. Net annual injection to the source pool is 18 GWh.

3.1 Geothermal borefield (BTES — Borehole Thermal Energy Storage)

• Vertical closed-loop borehole heat exchangers (BHE).
• Borehole depth: 180 m (590 ft). Drilling deeper than 200 m crosses into hot-rock regulatory regimes in most US states (e.g., Massachusetts DEP, Illinois IDNR).
• Spacing: 7 m grid (23 ft). Tighter spacing increases storage density but reduces extraction rate per borehole during winter peak.
• U-tube: double-U high-density PE-Xa SDR 11, 32 mm OD (1¼ in.) per leg, grouted with thermally enhanced bentonite (k ≥ 1.6 W/m·K — e.g., GeoPro TG, CETCO Thermal Grout 85).
• Pre-design extraction rate: 35 W/m drilled length (winter sustained), 50 W/m peak. Total drilled length needed for 25 MW peak heating (after 7 MW reduction by HPs lifting from loop): 25,000 kW / 50 W/m = 500,000 m = 2,778 boreholes at 180 m each.
• Footprint: 2,778 × 49 m² (7×7 spacing) = 136,000 m² = 13.6 ha (33.6 acres). Located beneath surface parking, athletic fields, and unpaved campus open space.
• Thermal Response Test (TRT) per ASHRAE Handbook–HVAC Applications Chapter 35: drilled pilot bore, 6 kW heater, 48 h test, infer ground thermal conductivity (target 2.0–2.5 W/m·K for glacial till common in Zone 5A) and undisturbed ground temperature (target 11–13 °C).
• Sizing software: EED 4.0 (Earth Energy Designer, BLOCON), GLD (Ground Loop Design, Gaia Geothermal), Comsol Multiphysics with subsurface module for transient long-term simulations across 25 years.
• Manifold: HDPE flow centers with electrofusion fittings, reverse-return for hydraulic balance, isolation valves per branch of 25 boreholes.

3.2 Sewage heat recovery

• Campus sewage outfall: 60 L/s (950 gpm) average, 90 L/s (1,425 gpm) peak, 15–22 °C year-round.
• Inline wastewater HX or sidestream:
  • Inline: Uhrig Therm-Liner ribbed steel HX laid in 1.2 m (DN 1200) sewer pipe, 80 m run, ~250 kW thermal recovery.
  • Sidestream: SHARC Energy systems (Vancouver, BC) — wastewater is pumped through a self-cleaning rotating drum filter and then through a brazed-plate HX (Alfa Laval AlfaNova fully welded stainless to avoid biofouling).
  • Huber ThermWin or KASAG wastewater HX as alternative.
• Annual recoverable energy: 4 GWh-th.

3.3 Data center waste heat

• Campus research compute: 4 MW IT load (anticipated to double by 2030).
• PUE 1.3 → 5.2 MW total electrical → 5.2 MW thermal rejection at chillers.
• Cooling fluid: 25–35 °C return from CRAH coils when configured for warm-water cooling (ASHRAE TC 9.9 Class W4 / W5 liquid-cooled servers).
• Direct integration into 5GDH loop on the “warm” side: condenser water from the data center chillers enters the loop’s warm leg with no additional heat pump.
• Annual contribution: 25 GWh-th (highest-value source — runs 24/7 with no diurnal swing).

3.4 Solar thermal + Pit Thermal Energy Storage (PTES)

• 12,000 m² (129,000 ft²) flat-plate collector array (Arcon-Sunmark HT-A 35/10 panels, common in Scandinavian DH).
• Annual yield in Zone 5A: ~450 kWh/m²·yr → 5.4 GWh-th.
• Pit thermal energy storage: 50,000 m³ insulated lined pit (referencing Vojens, Denmark — 200,000 m³, the world’s largest at construction in 2015 — and Marstal Fjernvarme — 75,000 m³). Lined with HDPE geomembrane and floating insulating cover. Charged at 80 °C in summer, drawn down to 25 °C by spring.
• For 5GDH, PTES is decoupled to a higher-temperature pre-heat loop that lifts the warm leg from 18 °C to 25 °C during heating season, raising COP of building HPs by 0.7–1.2 points.

3.5 Industrial process heat (campus laboratory loops)

• Autoclaves, vacuum pumps, semiconductor lab cleanrooms: 1.5 MW year-round at 30–60 °C.
• Captured via plate HX into the warm leg.

3.6 Trim — air-source heat pumps and gas peaking

• Air-source booster heat pumps: Mitsubishi e-Series (60 °C output, R-32), 4 × 1 MW.
• Gas-fired condensing peaking boilers: Viessmann Vitocrossal 200 CT3, 4 × 2 MW. Sized for N+1 redundancy at design winter day with borefield offline. Run < 200 h/year (regulated as cold-start emergency by EPA NESHAP Subpart JJJJJJ).

4. Distribution Network — Pipes, Pumps, Substations

4.1 Pipe selection

• For a 5GDH ambient loop (8–25 °C):
  • Polyethylene (PE-Xa, PE-RT, or HDPE PE 100) pre-insulated twin-pipe: BRUGG CalPex, LOGSTOR FlexPipe, ISOPLUS FlexClassic. Diameters available DN 25 through DN 200 in flexible coils up to DN 110, beyond that rigid lengths.
  • Pressure rating PN 10 (10 bar / 145 psi) sufficient; PN 16 for high points and pumping stations.
  • Insulation: PUR foam, lambda 0.022–0.024 W/m·K, with a HDPE jacket and (optionally) a leak-detection wire pair.
  • Compared to traditional 4GDH bonded steel: 60–70% installed cost reduction in trenching because trenchless plowing or narrow trench (0.6 m wide) is permissible, no welding/X-ray, no expansion compensators (PE-Xa accommodates thermal strain through cold flow).
• For the high-temperature pre-heat loop from PTES:
  • LOGSTOR pre-insulated bonded steel, P235GH carbon steel, alarm-wire monitored, DN 200, expansion loops and U-bends every 80 m.

4.2 Hydraulic sizing

• Loop topology: looped ring main with branch laterals — N+1 redundant feed for every building. Two 24 in. (DN 600) trunk lines circling the campus.
• Loop velocity: 1.5 m/s (5 ft/s) design, peaking 2.2 m/s; pressure drop limit 150 Pa/m (1.5 mbar/m).
• Operating pressure: 4 bar static, 2 bar dynamic differential at the worst-case substation.
• Glycol: 25% propylene glycol in any segment exposed to outdoor air or buried < 1.2 m (freeze protection). Inside the borefield manifolds, 25% PG. In the main loop where buried > 1.5 m, plain water.

4.3 Pumps

• Distributed pumping per building substation (each building has its own circulator) plus a small central booster:
  • Grundfos TPE 100-360/4 inline single-stage end-suction, IE5 PM motor, integrated VFD, EuP-compliant. 200 kW frames at trunk booster.
  • Wilo Stratos GIGA-N for branch substations, 7.5–22 kW.
• Variable primary flow (no secondary distribution pumping); pressure-independent control valves (PICVs) at every substation — Belimo EnergyValve or Danfoss AB-QM.
• Hydraulic balancing via remote-controlled PICVs, balanced at commissioning per ASHRAE Guideline 22.
• Pump energy budget: 3.5% of delivered thermal — i.e., 2.8 GWh-electric/year. This is the single largest operating cost line item after building-side HP electricity.

4.4 Energy transfer station (ETS) / building substation

Each building gets a self-contained skid:

• Two brazed-plate HX (Alfa Laval AlfaNova 76 or SWEP B649) — one for the building’s water-to-water HP evaporator/condenser, one bypass for direct heat exchange (free cooling in spring/fall).
• Belimo PICV control valves on both sides.
• Kamstrup MULTICAL 803 ultrasonic heat/cooling meter with redundant communication (LoRaWAN + M-Bus + Modbus TCP), accuracy Class 2 per EN 1434 — replacing older Diehl SHARKY 775 or Itron CF Ultramax that some legacy substations carry.
• Variable-speed pumps Grundfos MAGNA3 D 80-100 F (twin-head, ECM motor).
• Strainers, air separator (Spirovent SpiroTrap), magnetic dirt separator, expansion tank (10 bar nominal, 50 L nominal per 100 kW connected load).
• BACnet/IP controller (Distech Controls EC-NetAX, Schneider Electric SmartX, or Siemens Desigo PXC).

4.5 Building-side heat pumps (water-to-water)

• Modular reciprocating + screw chillers/heat pumps:
  • Trane Sintesis Helical Rotary RTAF (200–500 kW): R-1234ze (low-GWP, A2L). SCOP 4.5–5.5.
  • Carrier 30XW-V variable-speed screw, R-1234ze.
  • York YK two-stage centrifugal, R-1233zd.
  • Daikin McQuay WMC magnetic-bearing centrifugal — oil-free, 200–1,500 kW.
• For high-DHW residential dorms with low-temp radiators (60 °C supply max), Mitsubishi Ecodan QAHV CO2 transcritical (R-744) booster reaches 90 °C output from 25 °C loop. 24 kWth/unit, paralleled.
• For ammonia industrial-scale boost in central plant: Sabroe HeatPAC HPX (R-717 NH3), 800 kWth, 80 °C output, dual-stage screw.

5. Storage Strategy

Three timescales:

5.1 Diurnal — water tanks and ice

• 4 × 1,500 m³ stratified hot-water buffer tanks at 65 °C / 25 °C (1,500 m³ × 40 K × 1.16 = 70 MWh capacity each, 280 MWh total). Smaller than a single seasonal store but enough to time-shift 18 hours of heating against electricity price.
• 6 × 500 ton-hour ice storage tanks (CALMAC IceBank model 1500 modules) for cooling peak shaving. Stored ice melts during 14:00–20:00 peak electricity hours.

5.2 Inter-seasonal — borefield

• The borefield’s secondary function is the seasonal store. Drake Landing Solar Community (Okotoks, AB, commissioned 2007, operating to date with 144 boreholes × 35 m) demonstrated 97% solar fraction for space heating of 52 single-family homes — the canonical case study.
• Drake Landing’s lessons: ground temperature climbs over 5 years to a steady-state cycle (winter low 35 °C, summer high 80 °C in their high-temperature design). For our 5GDH ambient design, the swing is much narrower (winter min 5 °C, summer max 25 °C); the ground recovers fully each year.

5.3 Inter-seasonal high-temp — PTES

• See Section 3.4. PTES is the most cost-effective per kWh stored at scales > 30,000 m³ (Marstal, Vojens, Aalborg, Brædstrup). At smaller campus scales, tank thermal energy storage (TTES) — steel tanks, often above-ground — is more competitive.

5.4 Optional — molten salt or PCM

• Not selected. Molten salt (NaNO3/KNO3 60/40 solar salt — Sandia) is justified above 250 °C; the campus has no use case. Phase-change materials (microencapsulated paraffin from Microtek Labs, BASF Micronal) are too expensive per kWh at building scale.

6. Energy Modeling — From Brief to Sizing

• Building-by-building hourly load model: EnergyPlus 23.2 with Eppy/Python orchestration. Each of the 40 buildings has a generated IDF file from a typology template (lab, dorm, office, retail, mixed). Climate file: TMY3 for ORD (Chicago O’Hare) or BOS (Boston Logan).
• District network model: TRNSYS 18 with TESS Loops library. Types:
  • Type 557 (vertical U-tube ground heat exchanger, DST model — Hellström).
  • Type 668 (water-to-water heat pump).
  • Type 4 (stratified storage tank).
  • Type 39 (pump with VFD).
  • Type 60 (flat-plate solar collector).
• Co-simulation: EnergyPlus exports 8760-hour load profile per building → TRNSYS solves loop hydraulics and source/storage dispatch at 5-minute resolution → results fed back into EnergyPlus for HP sizing iteration.
• Validation: IES-VE for compliance reporting; IDA-ICE if AHU coil simulation needs more fidelity than EnergyPlus offers.
• Code compliance: ASHRAE 90.1-2022 Appendix G, IECC 2021 Section C405, LEED v4.1 EAp2, BREEAM 2023 New Construction Ene 01. Aim Energy Star Portfolio Manager score ≥ 95 per building.
• Spec sheets archived: every chiller/HP model is selected through manufacturer’s selection software (Trane TRACE 3D Plus, Carrier ECAT, York Optiview) and saved as PDFs with the operating conditions called out.

7. Controls and Optimization

7.1 BMS hierarchy

• Level 0: Field devices — sensors, actuators, VFDs. Protocols: Modbus RTU, BACnet MS/TP.
• Level 1: Unitary controllers — Distech ECY-S1000 or Schneider AS-P at substation skids.
• Level 2: Building automation — Siemens Desigo CC, Honeywell Niagara N4 (Tridium), Schneider EcoStruxure Building Operation.
• Level 3: District controller — supervisory layer. Custom Python on Linux server with InfluxDB time-series + Grafana dashboards + Volttron (PNNL) for legacy interop.
• Level 4: Enterprise — energy accounting, billing, ESG reporting. Tools: Tableau, custom Streamlit dashboards.

7.2 Model predictive control (MPC)

• 24-hour rolling horizon, 15-minute time step. Objective: minimize electricity cost + carbon (weighted) subject to building setpoints, storage state-of-charge limits, equipment minimum turndown, and grid demand-response commitments.
• Predicts: weather (Solcast 24-hour forecast for solar irradiance + temperature), occupancy from building access controls and Wi-Fi association count, day-ahead electricity price (PJM or ISO-NE LMP via API).
• Solver: Pyomo + Gurobi for MILP; Casadi + IPOPT for nonlinear sub-problems (HP efficiency curves).
• Reference deployments:
  • Stanford Energy System Innovations (SESI, 2015) — heat-recovery chillers (Carrier 19DV magnetic-bearing centrifugals, originally; expanded to 6 units totaling ~33 MWth). Saves 65% in heating energy and 50% in water vs prior cogen plant. Predictive optimization layer custom-built.
  • ETH Hönggerberg Anergy Network (Zurich, 2013–) — 12 buildings on 8–22 °C loop, BTES under campus, MPC over electricity price.
  • Cornell Lake Source Cooling (LSC, commissioned 2000) — 76 m intake from Cayuga Lake at 4 °C, 20,000 ton cooling, almost zero compressor energy.
  • HafenCity Hamburg — 4GDH evolving to 5GDH, integrated waste-incinerator heat + industrial.
  • Princeton co-gen + chilled water — example of disciplined district control for a mid-sized campus.
  • MIT central utility plant — district steam and chilled water, modernizing to electrify.

7.3 Demand response and virtual power plant participation

• PJM Capacity Performance (PJM tariff) — campus enrolls 5 MW shed-capable load through a curtailment service provider (Enel X, Voltus, CPower). Pays ~ $30k/MW-year.
• NYISO Demand-Side Ancillary Services Program — frequency regulation via BESS + HP turn-down.
• FERC Order 2222 — distributed energy resource aggregation into wholesale markets.

8. Power Supply

• Peak electrical: 25 MW (HVAC + lighting + plug + IT).
• Utility service: 34.5 kV primary, two redundant feeds, automatic transfer switch.
• On-site PV: 5 MWac rooftop + parking canopy. Module: First Solar Series 7 thin-film CdTe (530 W modules) — chosen over crystalline-Si for low temperature coefficient (better in Zone 5A summer). Inverters: SMA Sunny Tripower Core2 60 kW string inverters. Annual yield 6,500 MWh.
• BESS: 20 MWh / 5 MW Tesla Megapack 2 XL, four-hour duration. Used for (a) demand-charge avoidance, (b) frequency regulation, (c) emergency backup.
• Backup: two 2 MW Caterpillar XQP2000 diesel gensets, paralleled, EPA Tier 4 Final, in N+1.
• Renewable PPA: 25 MW notional via virtual PPA from a Texas wind farm — RECs delivered to retire Scope 2 emissions.

9. Refrigerants and Safety

Refrigerant choice is governed by GWP regulation, flammability, and toxicity.

• EU F-Gas Regulation 2024/573 (new May 2024) tightens GWP limits and phases down HFC quotas to 5% of 2015 baseline by 2030.
• US EPA AIM Act and SNAP rules; California CARB further restricts to GWP ≤ 750 for chillers > 200 RT after 2025.
• Selected:
  • R-1234ze(E) — GWP 7, A2L mild flammability. Trane RTAF, Carrier 30XW-V.
  • R-32 — GWP 675, A2L. Mitsubishi e-Series air-source.
  • R-744 (CO2) — GWP 1, A1 non-flammable. Mitsubishi Ecodan QAHV transcritical for DHW; Mayekawa NewTon C transcritical for industrial cycles.
  • R-717 (NH3) — GWP 0, B2L toxic + mildly flammable. Sabroe HeatPAC, restricted to mechanical-room locations meeting ASHRAE Standard 15-2022 occupied classification, with ammonia leak detection (Calibration Technologies CT-100) and ventilation interlocks.
• Inventories logged to F-gas registry; refrigerant tracking software (Trakref or refnet) maintains the EPA Section 608 record.

10. Permitting and Regulatory

• Groundwater extraction permits: state-level in the US. Illinois IDNR, Massachusetts MassDEP Groundwater Discharge Permit (closed-loop borefields generally need only a “groundwater heat pump well registration”). EU: Water Framework Directive 2000/60/EC compliance plus member-state hydrogeology survey.
• Drilling: state well-drilling contractor license (NGWA), Underwriters Laboratories or equivalent listing for surface manifolds.
• Air quality: NSR / Title V for backup gensets (limited operating hours keep them under major-source threshold). Boilers under NESHAP DDDDD.
• Electrical: NEC 2023, NFPA 70E, IEEE 1547 for grid interconnection. Utility approval for 5 MW PV + 5 MW BESS — typically a 6-12 month interconnection study with the utility.
• Fire and life safety: NFPA 855 (energy storage), NFPA 70, ASHRAE 15 (refrigerants), ASHRAE 34 (refrigerant designations).
• District energy network operator status: In Denmark, the municipality typically owns the network (Heating Supply Act). In the Netherlands, the Warmtewet (Heat Act 2014, revised 2023) regulates tariffs. In the US, the operator is usually a private utility or campus entity (e.g., Stanford Utilities, Princeton University Facilities, Cornell Energy and Sustainability), with rates set by long-term contracts not state public-utility regulation.

11. Construction and Commissioning

• Trenchless installation where possible (HDD horizontal directional drilling under existing roadways). Vermeer D40x55 S3 Navigator or Ditch Witch JT60 directional drill rigs.
• Open-cut on green-field sections; 0.6 m wide trench, 1.2 m bury depth, 100 mm sand bed.
• Pipe joining: PE-Xa electrofusion for couplings; butt-welding for HDPE PE 100 mains (McElroy TracStar 412).
• Pressure testing: 1.5 × design pressure, 24-hour hold, < 0.1% leak (EN 13941 standard).
• Borehole drilling: Atlas Copco TH60 or Sandvik DI65 rotary-percussion top-hammer rigs. Drilling cost (2025 USD): $25–45/m drilled depending on geology, casing requirements, and disposal of drill cuttings.
• BTES commissioning: thermal response test per BHE on every 50th borehole, sustained extraction test for 7 days on the full field once piped.
• ETS commissioning: ASHRAE Guideline 0 + Guideline 1.1; functional performance testing per equipment.
• Energy balance verification: 12-month performance period before final acceptance; SCOP for the whole network ≥ 4.0.

12. Economics

• CAPEX breakdown for 1 M m² campus (2025 USD, approximate):
  • Borefield (2,778 boreholes × 180 m × $35/m):$17.5 M.
  • PE-Xa twin-pipe distribution (12 km buried, $400/minstalled):$4.8 M.
  • Substations + building HPs (40 buildings × $250kaverage):$10 M.
  • Central plant building, peaking boilers, NH3 boost HP: $8 M.
  • Sewage HX + data-center HX: $1.5 M.
  • Solar thermal + PTES (50,000 m³): $5 M.
  • BMS + MPC + commissioning + engineering: $6 M.
  • Electrical infrastructure incl. PV + BESS: $14M(BESSalone$8 M at $400/kWh).
  • Contingency 15%: $10 M.
  • Total: ~$77 M, or ~ $77/m² connected building floor — at the low end of the €2,000–5,000/m² European 5GDH range because the campus model bundles building HPs with their host real-estate budgets.
• OPEX:
  • Electricity: ~25 GWh/yr × $0.09/kWhblended(afterPVself-consumption+PPA)=$2.25 M/yr.
  • Gas (peaking): negligible, < $50k/yr.
  • O&M (1.5% of CAPEX): $1.15 M/yr.
  • Borefield monitoring + drilling reserve (0.5%): $385k/yr.
• LCOE (thermal): $65–90/MWh-thover25-yearhorizon,6$40–80/MWh in Zone 5A depending on commodity strip) but with full Scope 1 decarbonization.
• Payback vs gas: 14–22 years at current US gas prices, faster if a carbon tax is introduced.
• Comparable European deployments report €2,500–4,000/m² CAPEX, often subsidized through national heat funds (Denmark, Germany BEW programme, NL warmtefonds).

13. Operating Phase — Years 1 to 25

13.1 First-year ramp

• Substations brought online by building completion phasing. Loop charging takes 4–6 weeks to reach steady-state temperatures.
• Initial COP often disappoints (3.5–4.0) due to overshoot in HP staging, conservative MPC tuning. Year-2 retuning is normal.
• Common Year-1 issues:
  • Air entrainment in flexible PE pipe (resolve with auto-air separators and post-commissioning bleed).
  • PICV authority mismatch (rebalance after 90-day operation).
  • HP refrigerant leaks at flanges (re-torque per manufacturer at 1,000 h).

13.2 Long-term ground thermal balance

• Annual borehole heat-flux audit. If extraction exceeds injection by > 5% per year for three consecutive years, the field will drift colder — predictable from TRT-calibrated simulation but easily caught with sensor strings (e.g., DTS — distributed temperature sensing — fiber optic in every 20th borehole, Silixa Ultima DTS).
• Reinject excess summer heat from data center directly into borefield to maintain energy balance.

13.3 Equipment refresh cycle

• HP compressors: rebuild at 60,000–80,000 h, replace at 120,000 h.
• Pumps: VFD replacement at 12-year average.
• Heat exchangers: gasket replacement at 8 years (brazed-plate units are essentially indefinite).
• Borehole U-tubes: design life 50+ years, no field replacement is practical; redundancy in field sizing accommodates 5% borehole loss.

14. Case Study References

• Drake Landing Solar Community (Okotoks, AB, 2007–): 52 homes, 144 BHE × 35 m, 800 m² solar collector, 97% solar fraction for heating. Operated by Sterling Homes consortium. Documented in Sibbitt et al., Energy Procedia 30 (2012).
• Stanford Energy System Innovations (SESI) (Stanford CA, commissioned 2015): replaced 60-year-old gas cogen plant with electric heat-recovery chillers. ~33 GWh-th annual recovered. $485Mproject.Saves$425 M cumulative through 2050 vs prior plant.
• Cornell Lake Source Cooling (Ithaca NY, 2000): 76 m intake from Cayuga Lake, 20,000 RT cooling capacity. ~85% electrical energy reduction vs vapor-compression chillers.
• ETH Hönggerberg Anergy Network (Zurich, 2013–): 12 buildings, 8–22 °C ambient loop, 700-borehole BTES under campus, 14,000 t CO2/yr avoided.
• Suurstoffi (Risch-Rotkreuz CH, 2010–): purpose-built 5GDH residential-commercial neighborhood, fully built-out reference.
• HafenCity Hamburg (2008–): waterfront redevelopment using 4GDH with industrial waste integration; planned 5GDH expansion.
• Vauban Freiburg (Germany, 1996–): biomass + solar 4GDH in eco-district.
• Vojens Fjernvarme (Denmark, 2015): 200,000 m³ PTES, world’s largest at construction.
• Princeton University and MIT central utilities: long-running campus-scale operators currently electrifying away from steam.
• New York Steam Company (Manhattan, 1882–): canonical 1st-generation system, still in service for legacy buildings.

15. Common Failure Modes and Lessons

• Under-sized borefield — most expensive mistake; retrofit drilling at $50–80/m doubles the cost vs greenfield. Mitigation: TRT before sizing, conservative thermal conductivity assumption (lower end of literature range), 20% safety margin on drilled length.
• Over-temperature in summer — happens when cooling load is over-predicted and injection swamps extraction. Mitigation: enable dry-cooler bypass to reject excess heat to atmosphere when ground > 25 °C.
• Under-temperature in winter — opposite failure. Mitigation: auxiliary air-source HP available; PTES drawdown.
• Flow imbalance — manifold reverse-return + PICV needs commissioning at design and shoulder seasons.
• Refrigerant leakage tracking — F-gas inventory drift exceeds annual self-report limit if undetected; remote leak detection (Cool-IR sensors, Bacharach H10 PRO) on every machine room.
• MPC drift — when occupancy patterns change (e.g., COVID, lab decommissioning), MPC’s learned models go stale; quarterly retraining recommended.
• Sewage HX fouling — wastewater HX requires periodic high-pressure jetting (every 6–12 months). Sidestream filter (SHARC drum filter) reduces frequency.
• Pump shaft seal failure — most common single failure mode in the network; carry rotating-element spares for top three pump models.

16. Closing Notes

The hardest engineering decision in a district energy project is not the equipment selection. It is the energy balance — committing to a borefield size and storage strategy that will perform 25 years from now with loads that nobody can predict precisely. The conservative move is to over-drill the borefield because everything else in the system can be retrofit or replaced, but the boreholes cannot. The second hardest decision is the governance and contracting model — who owns the network, how building owners are billed, who carries the risk of underperformance. Both decisions are made in the first six months of design and shape every dollar spent for the next quarter-century.

Adjacent

• design-residential-solar-battery-system — building-scale electrification analogue, where heat pumps and PV interact at the meter.
• design-utility-scale-solar-pv-plant — grid-scale renewables that supply the PPA underwriting district electrification.
• design-turbomachinery-cooling-loop — large-scale closed thermal loops and pump/HX hydraulics.
• design-hospital-mri-installation — building MEP and substation interface practice on a research campus.
• design-ev-traction-inverter — power-electronics primer relevant to large variable-speed drives and BESS interconnection.
• design-container-ship-propulsion-system — large-scale waste heat recovery and thermal-cycle analogues.