GPU accelerated power flow calculation of integrated electricity and heat system with component-oriented modeling of district heating network
Zhang Chen,Jun Liu,Xinglei Liu
DOI: https://doi.org/10.1016/j.apenergy.2021.117832
IF: 11.2
2022-01-01
Applied Energy
Abstract:Due to its advantage of improving energy efficiency and promoting sustainable development, integrated electricity and heat system (IEHS) has been widely studied in recent decades. However, the traditional network-oriented district heating network (DHN) model in IEHS could only deal with DHNs of supply-return-parallel topologies, and the employ of constant thermodynamic properties could incur inaccurate power flow results. With the increasing requirements on operation flexibility and system resilience of IEHS, it has become a necessity to develop a superior power flow model of DHN. This study presents a novel component-oriented modeling method in which the models of the three basic components in DHN, the pipelines, pressure sources and junctions, are investigated in detail. Formulas of the fundamental physical processes including pressure, temperature loss and enthalpy transfer are derived based on the variable thermodynamic state of the fluid rather than predetermined constants in the traditional simplified models. Variables associated with these basic components are discussed in detail and their respective constraints are expounded. To overcome the huge amount of computation in the IEHS analyzing process, GPU is introduced as a coprocessor and a parallel algorithm is designed accordingly. The versatility of the proposed model, including providing accurate, more detailed power flow results and analyzing DHN of general topologies, is presented in a small-scale DHN case. And the practicality of the proposed model is demonstrated in the ensuing practical-scale IEHS case. Meanwhile, the proposed GPU-based parallel algorithm has attained more than 3 times of performance boost compared to single CPU computing.
energy & fuels,engineering, chemical