Full Program »
Application of Artificial Intelligence In Geothermal Energy Utilization From Underground Heat Transfer To Integrated System Operation
Geothermal energy is becoming increasingly attractive due to its wide distribution, low operating costs and carbon emissions, as well as uninterrupted and sustainable energy supply merits over other renewable energy resources. As the most important approach to utilise geothermal energy, ground source heat pump technology (GSHP) has been widely applied in space heating and cooling. GSHP features with dynamic thermal response and complicated energy exchange between the coupled subsystems, including large scale borehole heat exchangers, energy plant and building terminals. Under ground heat transfer of large scale borehole heat exchangers as well as system optimization of GSHP which involvs multiple objectives are generally extremely time-consuming problems. Although considerable research has been carried out to analyze the transient process of geothermal energy utilization especially the thermal performance of the borehole heat exchangers (BHEs), few studies have investigated their impact on dynamic operation performance of GSHP due to the inefficiency of traditional numerical approach. This paper addresses the simulation and optimization of the integrated GSHP model by way of artificial intelligence. Two key issues are concerned in this paper. Firstly, with the assistance of multiscale convolution neutral network and long short term memory neutral network, an efficient and high-fidelity simulation method is developed for large scale heat transfer problem in complex porous media underground over different operation periods. Secondly, a model-based Artificial Neural Network (ANN) is presented to characterize the transient behavior of integrated GSHP system, which is combined with a multi-objective Genetic Algorithm (NSGA-II) for optimization. The methodology has been validated in the current study for the simulation of heat transfer underground of GSHP in a typical residential building. It is found that the proposed model based on artificial intelligence algorithms can appropriately predict the dynamic heat transport and migration between BHEs array with various configurations. Results also show significant reduction of time consumption for simulation. On the other hand, optimization of the GSHP system in terms of energy consumption as well as thermal comfort. Thanks to the artificial intelligence approach, dozens of potential operation strategies are revealed, with a wide range of trade-offs between thermal comfort and energy consumption. Finally, economic performance and operating savings of the system are also analyzed.