The EnEff:Heat project, carried out from 1 January 2020 to 31 December 2023, was funded under the German federal code 03EN3012A and brought together the Technical University Darmstadt, Entega AG and Siemens AG. Its goal was to accelerate the decarbonisation of the heat sector by creating a modular toolbox—MeFlexHeat—for the design, optimisation and operation of next‑generation flexible heat networks. The collaboration combined academic expertise in system modelling and optimisation with industrial experience in heat generation, storage and market participation. Entega supplied real‑world data from the city of Darmstadt’s existing district‑heat network, while Siemens contributed advanced simulation tools and market‑integration capabilities. The project’s four work packages focused on state estimation, local market mechanisms, mathematical optimisation of heat‑balance equations, and joint simulation and techno‑economic assessment.

In the state‑estimation phase, the team built a comprehensive measurement network across the Darmstadt district‑heat system and developed algorithms that infer the internal state of the network from sparse sensor data. These algorithms use a combination of Kalman filtering and deep‑neural‑network techniques to reconstruct temperature, flow and pressure profiles with an average root‑mean‑square error below 2 % of the measured range. The resulting state estimates provide the necessary transparency for subsequent optimisation and market‑based dispatch.

The market‑mechanism work package produced a suite of dispatch models that require only limited information about individual participants’ internal heat‑distribution systems. Three variants were implemented: a linear‑programming model for fast, real‑time bidding; a mixed‑integer linear‑programming model that captures discrete on/off decisions of heat‑turbines and boilers; and a nonlinear‑programming model that incorporates detailed thermodynamic constraints. All models were calibrated against historical operation data and demonstrated that a market‑based dispatch can reduce overall network losses by up to 8 % compared with conventional, centrally controlled operation, while maintaining consumer comfort levels.

Mathematical optimisation of the heat‑balance equations was tackled in the third work package. The team formulated a global optimisation problem that couples heat‑production, storage, and consumption decisions with the physical constraints of the network. Convex relaxations and robust optimisation techniques were introduced to keep the problem tractable for real‑time application. The resulting optimisation framework was embedded in the Modelica environment, enabling rapid prototyping and integration with existing simulation tools.

The final work package combined all components into a unified simulation platform. The platform was used to validate the current network configuration and to explore a 2030 scenario in which renewable heat sources, such as industrial waste heat and heat pumps, are significantly expanded. Techno‑economic analysis of the 2030 scenario indicated a potential cost reduction of 12 % for the utility and a 15 % increase in renewable heat penetration, while keeping peak demand within existing infrastructure limits. The results were presented to the utility’s decision‑makers and served as a basis for a planned pilot deployment.

Throughout the project, the partners published 12 peer‑reviewed papers, presented at five international conferences, and organised two workshops for industry stakeholders. Several graduate students completed theses that contributed to the state‑estimation and optimisation modules. The EnEff:Heat project therefore delivered a scientifically robust, industry‑ready toolbox that demonstrates how flexible heat networks can be optimised and market‑oriented, paving the way for a more sustainable and resilient energy system.