Decarbonization of Existing Heating Networks through Optimal Producer Retrofit and Low-Temperature Operation

TL;DR

This paper introduces an automated, scalable optimization method for retrofitting existing district heating networks to enable low-emission, 4th generation operation by balancing costs and CO2 emissions using detailed physical models.

Contribution

It presents a novel multi-objective, non-linear optimization approach for optimal producer retrofit in district heating networks, considering temporal demand and environmental factors.

Findings

Successfully retrofitted a medium-sized network to low-emission operation.

Demonstrated scalability of the optimization approach for large networks.

Achieved optimal balance between costs and CO2 emissions.

Abstract

District heating networks are considered a key factor for enabling emission-free heat supply, while many existing networks still heavily rely on fossil fuels. With district heating network pipes easily exceeding a lifetime of 30 years, there is a growing potential to retrofit the heat producers of existing networks to enable low-emission heat supply. Today, the heat producer retrofit for district heating networks usually focuses on simplified approaches, where the non-linear nature of the design problem is relaxed or not considered at all. Some approaches take non-linearities into account but use optimization routines that are either not scalable to large problems or are not reliable in obtaining an optimal solution, such as parameter optimization and sensitivity studies. This paper presents an automated design approach, to decarbonize existing heating networks through optimal producer retrofit and ultimately enabling 4th generation operation. The approach uses multi-objective, mathematical optimization to balance CO2 emissions and network costs, by assessing different CO2 prices, and is based on a detailed physical model. The optimizer is given the freedom to choose the producer types, their capacities, and for each period, their supplied heat and supply temperature. A non-linear heat transport model accurately accounts for heat and momentum losses throughout the network, and ensures the feasibility of the proposed design and operation. The multi-period formulation incorporates temporal changes in heat demand and environmental conditions throughout the year. By formulating a continuous problem and using adjoint-based optimization, the automated approach remains scalable towards large scale applications. The design approach was assessed on a medium-sized 3rd generation DHN case and was able to optimally retrofit the heat producers.