Optimal Integration of Thermal Energy Storage and Conversion in Fourth Generation Thermal Networks
Published in KU Leuven, 2019
Recommended citation: B. van der Heijde, "Optimal Integration of Thermal Energy Storage and Conversion in Fourth Generation Thermal Networks", PhD Thesis, KU Leuven, Belgium, 2019. https://lirias.kuleuven.be/handle/123456789/638748
This dissertation investigates the integration of thermal energy storage and conversion systems in fourth generation thermal networks, with a focus on district heating networks. Energy storage, more specifically in the form of heat or cold, will play a crucial role in future energy systems where large shares of renewable and residual energy sources are included. Due to the variability and seasonality of particular renewable sources, energy storage is needed to bridge the mismatch between moments where energy is needed by end-users and when it is available. Thermal networks are a second technology that may be crucial in the energy system of the future. Whereas energy storage systems solve mismatches in time, thermal networks are able to solve misalignments in space, between locations with large heat or cold surpluses on the one hand and large demands on the other. Moreover, collective systems may exploit the economy of scales, and the lower supply temperatures of the newest generation of district heating allow for more efficient heat production systems, both renewable (such as solar thermal collectors or low-temperature geothermal heating) and electricity based (such as heat pumps). This thesis investigates how thermal energy storage and conversion systems can be integrated in an optimal way in future thermal networks. Therefore, an optimisation algorithm is set up with the goal to find the optimal sizes of several network components. The first part of the thesis focuses on the thermal and hydraulic behaviour of thermal network pipes. Firstly, steady-state and dynamic simulation models are derived and validated to understand which effects are most important. Based on these detailed models, a model for optimisation is derived. In order to keep the solution of the mathematical model feasible within reasonable time, a number of assumptions need to be made. In the second part, additional models for thermal energy storage and conversion systems are described. The final part of the thesis integrates all previously derived models into a large optimisation problem. Given the complex interactions between different system components, the operational aspect of the network is essential. Whereas a number of previous studies employed simulation-based control evaluations, this thesis introduces modesto, a Python toolbox for the compilation of optimal control problems for district energy systems and district heating systems in particular. This toolbox is used as the lower layer of a two-layer optimisation structure. The upper layer comprises an evolutionary algorithm that explores the design space with respect to multiple objective functions. The objective function values result from the lower layer’s evaluation of individual network designs. The lower, operational layer’s calculation time is reduced by using an optimal representative days selection algorithm, compatible with seasonal thermal energy storage systems. As such, an integrated optimal control and design of district energy systems can be performed. The resulting design optimisation algorithm is shown to be very flexible in terms of which design variables are included. The design optimisation algorithm is applied to a fictitious case study, based on real geographical data and building geometries from the city of Genk. Although the results show that the algorithm is capable of proposing a set of optimal designs with respect to multiple objective functions, it is hard to derive general design rules or rules of thumb that can be easily applied in real designs. On the other hand, it is shown that seasonal thermal energy storage will be a crucial component of future renewable-based energy systems, since every proposed design includes substantial storage volumes that act on a seasonal time scale.