Guide

District heating and multi-energy system software: how to choose

There is no single best tool for district heating or multi-energy systems, because "modelling the system" covers several different decisions. Some tools specialise in one of them; others cover several at different levels of detail. This guide sorts the main open-source and commercial tools by the decision they support, notes what each does well and where it stops, and shows how teams combine them.

Last reviewed: 17 November 2026

Christopher McLaren, PhD
Written by Christopher McLaren, PhD · Head of AI & Commercial Enablement
Questions? Ask Christopher →
Jump to the comparison →

About this guide

We make one of these tools, Sympheny. We selected tools for their relevance to district heating and cooling or integrated multi-energy planning, the availability of public documentation, and distinct capabilities. It is a representative selection, not a ranking, and descriptions are based on each tool's own public documentation, not on independent testing. Features and licences change: if something here is out of date, tell us and we'll correct it.

Send a correction
Start with the decision

Five decisions, five kinds of tool

Decision The question Typical method
1. Map and screen Where could a heat network make sense, which buildings or sites should connect, and where might pipes run? GIS analysis, demand and resource mapping, route optimisation
2. Optimise supply and investment Which technologies, what capacities, which connections, built when, for the best cost and CO₂ trade-off? Mathematical optimisation (LP or MILP) over full or representative time periods
3. Design and verify the network Will the pipes deliver the required flow, pressure and temperature everywhere? Steady-state or quasi-dynamic thermo-hydraulic calculation
4. Simulate dynamics and controls How does the coupled system behave over time, with its control strategy? Physics-based dynamic simulation
5. Optimise operation and markets How should existing assets run, given forecast demand, prices and constraints? Dispatch optimisation and forecasting, linked to SCADA and markets

These are not stages in a fixed order. A municipality exploring heat-network areas starts with mapping; a utility with low pressure at remote substations starts with hydraulics; a plant losing money in electricity markets starts with dispatch. For a new, complex, multi-energy site with many supply options, investment optimisation often comes before detailed network design. Many projects go back and forth between these steps.

Two points that cut across all five

Scale. The same kind of tool can work at building, site, district, city or national scale. Check that a tool's intended scale matches your project.

Commercial and legal structuring. Who pays, who owns which asset, and how heat is metered and priced between parties. Most tools support this only at cost and cash-flow level; contracts and settlement are usually worked out outside them.

Decision 1

Map and screen

Tools for finding where a heat network could work and which buildings or sites belong in it.

THERMOS

Centre for Sustainable Energy (UK); originally an EU Horizon 2020 project

Free, open-source web tool that optimises which buildings to connect and where to route a heat or cold network, using OpenStreetMap or your own GIS data. Fast early screening for municipalities.

Website →

Citiwatts (formerly Hotmaps)

Successor to the EU Hotmaps toolbox, run by HES-SO Valais

Online toolbox for mapping heating and cooling demand, resources and district heating potential across Europe.

Website →

City Energy Analyst (CEA)

ETH Zurich; open source (MIT)

Urban building energy modelling: building demand, solar potential, emissions, and district heating and cooling supply options. Python, with a new graphical interface in version 4.

Website →

nPro (see decision 2) also covers district mapping, network routing and pipe sizing.

Decision 2

Optimise supply and investment

These tools search across technology combinations, sizes and connections and return designs that are optimal, or within a stated tolerance of optimal, under the model's assumptions. They suit questions where the design is still open.

Open-source frameworks

Free, transparent and extensible. Your team builds, validates and maintains the model, and picks a solver (see the note on solvers below).

PyPSA

Open source (MIT); community led from TU Berlin

Investment and operational optimisation with sector coupling (electricity, heat, hydrogen, transport), supporting LP and selected MILP formulations such as unit commitment. A very large community; widely used for regional and national studies (PyPSA-Eur). Heat is modelled as an energy carrier, without pipe hydraulics.

Website →

Calliope

Open source (Apache 2.0)

Planning optimisation from districts to continents, with model data kept in YAML and CSV files separate from code, which keeps scenario studies clean. LP and MILP.

Website →

oemof-solph and DHNx

Open source (MIT); oemof developer group

A flexible LP and MILP model generator built from generic buses and components. The DHNx extension optimises district heating network topology and pipe sizes.

Website →

ETHOS.FINE

Open source; Forschungszentrum Jülich

Multi-region, multi-commodity investment optimisation with multiple investment periods.

Website →

SpineOpt

Open source; Spine project (Julia)

Flexible MILP framework for multi-energy systems with customisable temporal and stochastic structure.

Website →

urbs

Open source (GPL-3.0); TU Munich

Multi-commodity, multi-site linear optimisation model.

Website →

Free tools from public laboratories

DER-CAM

Berkeley Lab (USA)

MILP investment and dispatch of distributed energy resources for multi-energy microgrids, including heat, cooling and multi-node sites.

Website →

REopt

Developed at the US National Renewable Energy Laboratory

Free web tool, API and open-source Julia engine for techno-economic optimisation of distributed energy resources at buildings, campuses and microgrids, including resilience. Not a heat-network design tool.

Website →

Commercial platforms

Sympheny

Urban Sympheny AG (Switzerland); Empa spin-off

Browser-based MILP optimisation of multi-energy systems across connected hubs: heat at several temperature levels, cooling, electricity, gas and hydrogen in one model. Candidate network links are drawn on a map and the optimiser decides which to build and how large. Design and hourly operation are solved together, with typical days for large models, multi-stage investment pathways and cost-versus-CO₂ Pareto fronts. Users set the optimality gap (1% by default) and the solver runs until it reaches it; Gurobi is included. No coding needed; REST API and Python SDK for automation. Works well with AI assistants: a connector for Claude is available and an official MCP server is coming soon.

Limits, from its own documentation: best suited to early planning phases; sites above about 15 to 20 buildings are modelled by grouping buildings into hubs; heat networks are energy flows with linear losses, without pressure or mass flow; no detailed electrical-network optimisation; no control or operational optimisation.

Sympheny and AI assistants → Website →

energyPRO

EMD International (Denmark)

MILP plant design, asset sizing, dispatch, finance and simultaneous electricity-market participation, with a module for several sites. The long-standing reference for district heating plant economics in Denmark and beyond.

Website →

nPro

nPro Energy (Germany)

Web app for early-phase planning of districts and heat networks, including 4th- and 5th-generation (anergy) networks: hourly demand, network routing, pipe sizing, pressure losses and energy-system sizing, with German data and funding (BEW) reporting.

Website →

TOP-Energy

GFaI (Germany)

Flowsheet-based simulation and optimisation of industrial and municipal energy systems, including steam and compressed air.

Website →

XENDEE

Xendee (USA)

Web-based MILP design and operation of microgrids, including CHP and thermal storage, across multiple nodes.

Website →
Decision 3

Design and verify the network

Tools that check whether a specific pipe network works: pressures, flows, temperatures, pump sizes and bottlenecks. Utilities often run them on the same network for decades.

pandapipes

Fraunhofer IEE and University of Kassel; open source (BSD)

Python solver for gas, water and heat networks, coupled to electricity grids through pandapower.

Website →

STANET

Ingenieurbüro Fischer-Uhrig (Germany)

Steady-state and dynamic calculation of gas, water and district heating networks, widely used by German utilities.

Website →

NetSim

Vitec (Nordics)

Static and dynamic simulation of district heating and cooling networks, used by many Nordic network owners and consultants.

Website →

OpenUtilities sisHYD

Bentley Systems

Hydraulic and thermal modelling of hot-water, steam, cooling and gas networks, with automated pipe sizing.

Website →

EcoStruxure District Energy

Schneider Electric; formerly Termis

Network simulation and operational optimisation for district energy utilities.

Website →

Fluidit Heat

Fluidit (Finland)

City-scale network modelling with time-series simulation and operational analysis.

Website →

Leanheat Network

Danfoss (Denmark)

Thermo-hydraulic planning and online operational optimisation of heating and cooling networks.

Website →
Decision 4

Simulate dynamics and controls

Tools that show how a chosen design behaves over time, down to its controls.

IDA Districts

EQUA (Sweden)

Dynamic thermo-hydraulic simulation of heating, cooling and anergy networks together with buildings, plants, storage and controls, including meshed, multi-source and bidirectional networks. A QGIS plugin imports network layouts.

Website →

Modelica and the Modelica Buildings Library

Library: Berkeley Lab, open source; tools such as Dymola (Dassault Systèmes) or OpenModelica

Equation-based, configurable physics down to control loops, including district heating and 5th-generation networks. Very flexible, with a corresponding modelling effort.

Website →

Hysopt

Hysopt (Belgium)

Web-based hydronic design and dynamic simulation, from building systems to large heat networks.

Website →

TRNSYS

Established transient simulation environment

Transient simulation of thermal energy systems, widely used for solar thermal, storage and heat-pump systems.

Website →
Decision 5

Optimise operation and markets

Tools for running existing assets well.

energyPRO

EMD International (Denmark)

Plant dispatch and simultaneous participation in day-ahead, reserve and balancing markets (see decision 2).

Website →

Gradyent

Gradyent (Netherlands)

Real-time digital twin that optimises supply temperatures and multi-source dispatch for heat networks, linked to SCADA.

Website →

Optit

Optit (Italy)

Decision support and optimisation for district heating dispatch, network operation and investment roadmaps, alongside logistics and industry.

Website →

Leanheat Network, EcoStruxure District Energy and Fluidit (decision 3) also offer operational modules.

Related tools for specific applications

HOMER Pro (UL Solutions) for electricity-focused microgrid sizing, with CHP and hydrogen modules; Polysun (Vela Solaris) for simulating solar, heat-pump and storage systems; and, for national and regional energy systems, EnergyPLAN (Aalborg University), Balmorel and PyPSA-Eur.

Comparison

Side by side

"Native" means built in; "Custom" means possible with custom modelling; "Partial" means limited support; a dash means it is not a primary function. Based on public documentation, November 2026.

Scroll sideways to see all columns.

ToolLicenceMain decisionTypical scaleMulti-energy investment optimisationNetwork routingDetailed hydraulicsDynamic simulationMulti-year investmentInterfaceMain limitation
THERMOSOpen source, free webMap and screenDistrict, city–Native–––WebHeat and cold networks only; early stage
City Energy AnalystOpen sourceMap and screen; supply optionsDistrict, cityPartialPartial–––Python, GUISet-up effort
PyPSAOpen sourceSupply and investmentRegion, nation; also localNativeCustom––NativePythonNeeds programming; no pipe hydraulics
CalliopeOpen sourceSupply and investmentDistrict to continentNativeCustom––NativePython, YAMLNeeds programming; no pipe hydraulics
oemof-solph + DHNxOpen sourceSupply and investment; network layoutSite to regionNativeNative (DHNx)––NativePythonNeeds programming; modules to combine
ETHOS.FINE, SpineOptOpen sourceSupply and investmentSite to nationNativeCustom––NativePython, JuliaNeeds programming
DER-CAM, REoptFreeDistributed energy sizing and dispatchBuilding to microgridNative–––PartialDesktop, web, APIMicrogrid focus; no heat-network design
SymphenyCommercialSupply and investmentSite to city (hubs)NativeNative (links between hubs)––NativeWeb, API, Python, Claude connectorNo hydraulics; above 15 to 20 buildings, group into hubs
energyPROCommercialPlant design, sizing, dispatch, marketsPlant, several sitesNative (plant level)–––PartialWindows desktopPlant-centred; no network design
nProCommercialEarly district and network planningDistrictNativeNativePartial–Not verifiedWebEarly phase; German focus
TOP-EnergyCommercialSupply optimisation, industrySiteNative–––Not verifiedDesktopSite focus
pandapipesOpen sourceNetwork designNetwork––NativePartial–PythonNeeds programming
STANET, NetSim, sisHYD, FluiditCommercialNetwork designNetwork–PartialNativePartial to native–Desktop, cloudNo supply optimisation
IDA DistrictsCommercialDynamics and controlsNetwork and buildings––NativeNative–Desktop, QGISSimulates a given design
Modelica / DymolaLibrary open; Dymola commercialDynamics and controlsComponent to district––NativeNative–ModelicaModelling effort
GradyentCommercialOperationExisting network––Native (digital twin)Native–Software as a serviceNeeds an operating network
In practice

How teams combine them

A common sequence for a new multi-energy district: screen where a network makes sense (THERMOS, Citiwatts, City Energy Analyst, nPro), optimise the supply and investment concept (Sympheny, PyPSA, Calliope, oemof, energyPRO, nPro), verify the chosen network (pandapipes, STANET, NetSim, sisHYD), simulate dynamics and controls where the design is tight (IDA Districts, Modelica), and optimise operation once it is built (energyPRO, Gradyent).

A published example: in "Platform-based design for energy systems" (Applied Energy, 2023), researchers from Berkeley Lab and Empa sized a district system with Sympheny's MILP optimisation, then simulated it in Modelica, and published the model.

Read the paper →
A note on solvers

Free solvers work. On hard problems, the gap is still large.

Every optimisation tool needs a solver. Free solvers (HiGHS, SCIP, CBC, GLPK) are capable and improving fast, so open-source frameworks do not require a commercial licence. On difficult mixed-integer problems, though, there is still a large gap.

In the 2025 Open Energy Benchmark, run by Open Energy Transition on 80 mixed-integer problems from energy modelling frameworks (problems under one million variables, one-hour time limit), Gurobi solved 84% of the problems, HiGHS and SCIP 66% each, CBC 56% and GLPK 29%. On average, HiGHS took about four times as long as Gurobi.

Solve time is also cost, because cloud computing is billed by the hour. That is why Sympheny uses Gurobi: in our own tests, free solvers needed far more computing time on our models, which cost more in cloud resources than the licence saved. Your models may behave differently, and the benchmark's scripts let you test your own.

MILP problems solved within one hour
  • Gurobi 84%
  • HiGHS 66%
  • SCIP 66%
  • CBC 56%
  • GLPK 29%
Source: Open Energy Benchmark, aggregated results, MILP problems, short configuration (accessed 17 November 2026).
Choosing

Questions that decide it

Which decision are you making now?

Screening, investment, network capacity, controls or operation. Pick a tool from the matching group above.

One plant, or many connected sites and carriers?

For a single energy centre, a plant tool may be simpler. Interconnected districts, campuses and portfolios need multi-node modelling.

Who will build and maintain the model?

Open-source frameworks give full control and transparency at no licence cost, and need modelling and programming skills. Commercial tools trade licence fees for packaged workflows, data and support.

How will you defend the result?

For optimisers, check the solver status and the optimality gap actually reached, and whether runs are reproducible. For simulators, check validation and calibration.

Where is the project?

Some tools ship strong national data (nPro for Germany, energyPRO for Danish markets); others are country-neutral.

For a longer checklist on optimisers, see How to choose an energy system optimiser →
Honest limits

When Sympheny is not the right tool

  • You need pressure, flow and temperature in individual pipes: use a hydraulic or dynamic simulator.
  • You are optimising the daily operation of an existing plant on energy markets: energyPRO, Gradyent or Optit.
  • You are modelling a national energy system: PyPSA-Eur, EnergyPLAN or Balmorel.
  • You want a German heat-network study with BEW reporting and no wider system questions: nPro is quicker.
  • You want building-by-building optimisation of hundreds of buildings without grouping them into hubs.
  • Your team wants full control of the model code at no licence cost: an open-source framework fits better.

Multi-hub investment optimisers such as Sympheny, PyPSA, Calliope, oemof or ETHOS.FINE are most useful when the design is still open, several energy carriers and sites interact, and the investment decision has to be justified to others.

FAQ

Frequently asked questions

What is the best software for district heating modelling?

It depends on the decision. For screening where a network makes sense: THERMOS, Citiwatts, City Energy Analyst or nPro. For deciding what to build across heat, electricity and storage: an optimiser such as Sympheny, PyPSA, Calliope, oemof, energyPRO or nPro. For pipe hydraulics: pandapipes, STANET, NetSim, sisHYD or Fluidit. For dynamics and controls: IDA Districts or Modelica. For operation: energyPRO, Gradyent or Optit.

Which software can size district heating pipes?

nPro, oemof DHNx and THERMOS size or route pipes at planning stage. pandapipes, STANET, NetSim, sisHYD, Fluidit and IDA Districts verify them hydraulically.

Which tools support multi-year investment optimisation?

Among others: Sympheny (multi-stage runs), PyPSA, Calliope, oemof-solph, ETHOS.FINE and SpineOpt.

Which open-source tools can model electricity and heat together?

PyPSA, Calliope, oemof-solph, ETHOS.FINE, SpineOpt and urbs optimise them jointly. pandapipes with pandapower couples them in network simulation.

Do I need a commercial solver?

No. Free solvers work, especially for linear problems. On large mixed-integer problems, commercial solvers such as Gurobi still solve more problems and solve them faster in public benchmarks, which matters for project time and cloud costs.

What is the difference between optimisation and simulation?

A simulation tells you how a design you specify will perform. An optimisation searches many designs and finds the best one for your objective under the model's assumptions. They answer different questions and are often used one after the other, sometimes repeatedly.

Does Sympheny do hydraulic network simulation?

No. Sympheny models thermal networks as links between hubs with capacity, losses and cost. That suits concept-stage comparisons but does not replace pipe-level engineering, which is done in a dedicated tool.

Can I use Sympheny with AI assistants?

Yes. Sympheny works well with AI assistants: a connector for Claude is available, and an official MCP server is coming soon. The optimisation itself remains a deterministic MILP.

Planning a multi-energy system?

See whether Sympheny fits your project.

If your design is still open and several energy carriers and sites interact, we'll show you what an optimised concept for a project like yours looks like.

Book a demo