DDM Feldkirchen

Demonstration of digitalization measures in heating networks using the example of Feldkirchen district heating network

Short Description



Starting point / motivation

"DDM Feldkirchen" deals intensively with various digitization measures in the district heating sector. Due to the increasingly complex topic, digitization is becoming increasingly important, not only for the operators, but for all stakeholder groups. The reasons for the increasing complexity are, on the one hand, decentralization and, on the other hand, the ongoing reduction in all network parameters such as temperature and pressure in the network.

The two ongoing research projects in the "Industrial Research" category, namely "lowTEMP4districtheat" and "Brainy Heat Grids" deal with the challenges that these developments entail and are preliminary projects to the current demonstration project.

Contents and goals

Modern district heating systems are designed to connect heterogeneous energy sources and technologies such as renewable energies, waste heat, heat storage, power grids, heating networks and heat pumps.

The design process is both complicated and critical to the performance of the district heating system. Therefore, a number of intelligent approaches (digital twin, model predictive controls, AI-based forecasts) should not only be used to optimize the design of the grids, but also during their operation for

  1. optimal control of operational processes,
  2. accurate error detection and
  3. a control of transfer stations,

created and tested in the district heating system of Feldkirchen. There is a need for intelligent monitoring and control of district heating systems controlled by traditional PID controllers, with most adjustments being made manually due to grid operators' experience.

To support the transformation to automated heating systems, model-predictive control strategies should be used, including AI-based forecast models

  1. be able to compensate for changes in the network and data failures to a large extent,
  2. are robust to stochastic uncertainties, and
  3. are flexible and scalable.


Due to the dynamic nature of district heating networks, due to the constant expansion of these, and due to frequent data failures or the lack of connection options for various measuring points to the central control system, a complete data situation for the exact depiction of the thermo-hydraulic situation in the network, with the help of a purely physical model (white box model) is only possible to a very limited extent. For this reason, a digital twin equipped with soft sensors is used in "lowTEMP4districtheat". This gray box model allows for multiple degrees of freedom to exist.

With this model and moderate computing power, the overall thermal-hydraulic situation can be depicted in the network at any time as realistically as possible in real time using standard load profiles and AI methods. For this purpose, the necessary interfaces to the project partner Hoval Gesellschaft m.b.H. TopTronic® Supervisor software already installed in the demonstration system. With these interfaces, it must be possible to monitor, record data and optimize energy production systems and the entire district heating network in real time.

This real-time simulation is to be used in "DDM Feldkirchen" on the one hand as a basis for an extended error analysis by comparing simulated with measured data. On the other hand, the control of the network pumps and the heat-generating systems should be optimized through the simulation. Finally, an optimization algorithm should adjust the storage management of the primary heat storage to the current or forecast heat demand and to the operation of a heat pump supplied by a local PV system.

Another focus of "DDM Feldkirchen" will be on the implementation of the higher-level control of various district heating transfer stations, developed in "Brainy Heat Grids", for minimizing peak loads and reducing return temperatures. Here, individual transfer stations activated by the secondary side serve as a tool for increasing flexibility, primarily by shifting the loading processes of secondary hot water storage tanks in time.

These loading processes always result in high return temperatures on the primary side. Due to the time shift, a generally lower return temperature can be achieved, especially in the heating plant, with the associated advantages for the heat-generating systems and for the network itself.

Expected results

In "DDM Feldkirchen" some results of the research projects "lowTEMP4districtheat" and "Brainy Heat Grids" are applied. From "lowTEMP4districtheat" the transient network simulation with soft sensors is to be applied. For example, the real-time data from TopTronic® Supervisor should be accessed via an MQTT broker and a real-time simulation of the district heating system should be carried out in Python.

The results of the various forecast methods based on AI and the control algorithms for controlling individual transfer stations from "Brainy Heat Grids" are to be used with the aim of minimizing peak loads and return temperature. Based on this preliminary work, various predictive control algorithms are to be examined with a focus on the heating plant.

For example, the current location in the grid with the lowest pressure can be determined from the transient grid simulation. In this way, this dynamic grid wide lowest pressure can be used as the control variable for the grid pumps and not the specified static location in the grid with the lowest pressure.

Furthermore, the operation of the central heat-generating systems and the centrally located heat storage tanks should be coordinated with the control of the heat transfer stations, which has already been optimized using the results from "Brainy Heat Grids".

In addition to the field tests of the technologies from the "lowTEMP4districtheat" and "Brainy Heat Grids" projects, various questions about further expansion stages of the district heating system in Feldkirchen (e.g. new heat pump) still have to be clarified in "DDM Feldkirchen". Suitable business and operator models must be defined for this.

Project Partners

Project management

4ward Energy Research GmbH

Project or cooperation partners

  • BC Regionalwärme Gruppe GmbH
  • Hoval Gesellschaft m.b.H.
  • GEF Ingenieur AG
  • Energieagentur Obersteiermark GmbH
  • Ingenieurbüro Jaindl & Garz GmbH
  • Prozess Optimal CAP GmbH

Contact Address

4ward Energy Research GmbH
Dr. Markus Rabensteiner
Reininghausstraße 13A
A-8020 Graz
Tel.: +43 (664) 88251830
E-mail: markus.rabensteiner@4wardenergy.at
Web: www.4wardenergy.at