Machine Learning Techniques for Supporting Renewable Energy Generation and Integration: A Survey

Kasun S. Perera, Zeyar Aung, Wei Lee Woon

Research output: Contribution to journalArticlepeer-review

62 Scopus citations

Abstract

The extraction of energy from renewable sources is rapidly growing. The current pace of technological development makes it commercially viable to harness energy from sun, wind, geothermal and many other renewable sources. Because of the negative effects on the environment and the economy, conventional energy sources like natural gas, crude oil and coal are coming under political and economic pressure. Thus, they require a better mix of energy sources with a higher percentage of renewable energy sources. Harnessing energy from renewable sources range from small scale (e.g., a single household) to large scale (e.g., power plants producing several MWs to a few GWs providing energy to an entire city). An inherent characteristic common to all renewable power plants is that power generation is dependent on environmental parameters and thus cannot be fully controlled or planned for in advance. In a power grid, it is necessary to predict the amount of power that will be generated in the future, including those from the renewable sources, as fluctuations in capacity and/or quality can have negative impacts on the physical health of the entire grid as well as the quality of life of its users. As renewable power plants continue to expand, it will also be necessary to determine their optimal sizes, locations and configurations. In addition, management of the smart grid, in which the renewable energy plants are integrated, is also a challenging problem. In this paper we provide a survey on different machine learning techniques used to address the above issues related to renewable energy generation and integration.

Keywords

  • Machine learning
  • Renewable energy
  • Smart grids

Fingerprint

Dive into the research topics of 'Machine Learning Techniques for Supporting Renewable Energy Generation and Integration: A Survey'. Together they form a unique fingerprint.

Cite this