期刊
RENEWABLE & SUSTAINABLE ENERGY REVIEWS
卷 153, 期 -, 页码 -出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.rser.2021.111685
关键词
Geothermal energy; Ground source heat pump; Control; Predictive model; Machine learning
资金
- European Union's Horizon 2020 research and innovation program [792355]
- H2020 Societal Challenges Programme [792355] Funding Source: H2020 Societal Challenges Programme
Geothermal energy has great potential for CO2 reduction, but is underutilized due to high initial costs. Artificial Intelligence and Machine Learning play important roles in optimizing Ground Source Heat Pump control, achieving long-term performance and reduced payback time.
Geothermal energy has the potential to contribute significantly to the CO2 reduction targets as a renewable source for building heating and cooling but is yet under exploited, mostly due to its high initial investment cost. A lot of research is being carried out to optimise Ground Source Heat Pump (GSHP) systems' design, but a good control strategy is also fundamental to achieve long-term performance and reduced payback time. GSHP control optimisation is a non-linear dynamic optimisation problem that is influenced by multiple parameters. It can thus not be fully optimised with traditional methods. Artificial Intelligence, and in particular Machine Learning, is suited for this type of optimisation as it can learn implicit relations between parameters and can address non-linearity. This paper reviews the challenges of GSHP control and the strategies for control optimisation found in the literature, from basic rule-based system to artificial neural network-based strategies. Two principal uses of Artificial Intelligence for ground source heat pump control are identified: building a predictive model of the system that reflects its real performances and optimising the control decision in real time. However, the examples found in the literature are limited and the need to further explore the benefits of Machine Learning is identified. The latest developments in the field are reviewed to explore their potential to further improve GSHP control. The challenges of the full implementation of such algorithms are also discussed.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据