Lunch colloquium DCSC

Modeling and Predictive Control for Large-Scale District Heating Networks: Towards Scalable and Efficient Strategies

Speaker: Max Sibeijn (PhD at DCSC)

Abstract: District heating networks (DHNs) provide a sustainable approach to urban energy supply by integrating renewable and reusable energy resources. These networks, governed by complex nonlinear dynamics, have traditionally relied on decentralized, rule-based control strategies with limited communication among network components. However, as DHNs evolve to meet modern energy demands, more advanced control strategies are necessary to optimize their efficiency and resilience.

In this talk, we introduce a graph-theoretic model for DHNs that incorporates key 4th generation district heating elements—including prosumer and multiple producer configurations—in a form suitable for optimization-based control. Building on this foundation, we develop a computationally efficient two-level economic MPC scheme for DHNs based on dual dynamic programming. This approach consists of two parts: a short-horizon nonlinear MPC with an approximate terminal cost, and a long-horizon linear MPC that iteratively improves this approximation by generating lower bounds on the value function.

Research cluster: Optimization