Forecasting Hydrogen Production from Wind Energy in a Suburban Environment Using Machine Learning

There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

Abstract

The environment is seriously threatened by the rising energy demand and the use of conventional energy sources. Renewable energy sources including hydro, solar, and wind have been the focus of extensive research due to the proliferation of energy demands and technological advancement. Wind energy is mostly harvested in coastal areas, and little work has been done on energy extraction from winds in a suburban environment. The fickle behavior of wind makes it a less attractive renewable energy source. However, an energy storage method may be added to store harvested wind energy. The purpose of this study is to evaluate the feasibility of extracting wind energy in terms of hydrogen energy in a suburban environment incorporating artificial intelligence techniques. To this end, a site was selected latitude 33.64° N, longitude 72.98° N, and elevation 500 m above mean sea level in proximity to hills. One year of wind data consisting of wind speed, wind direction, and wind gust was collected at 10 min intervals. Subsequently, long short-term memory (LSTM), support vector regression (SVR), and linear regression models were trained on the empirically collected data to estimate daily hydrogen production. The results reveal that the overall prediction performance of LSTM was best compared to that of SVR and linear regression models. Furthermore, we found that an average of 6.76 kg/day of hydrogen can be produced by a 1.5 MW wind turbine with the help of an artificial intelligence method (LSTM) that is well suited for time-series data to classify, process, and predict.

Related collections

Most cited references 58

Record: found
Abstract: found
Article: not found

A Review of Recurrent Neural Networks: LSTM Cells and Network Architectures

Yong Yu, Xiaosheng Si, Changhua Hu … (2019)

Recurrent neural networks (RNNs) have been widely adopted in research areas concerned with sequential data, such as text, audio, and video. However, RNNs consisting of sigma cells or tanh cells are unable to learn the relevant information of input data when the input gap is large. By introducing gate functions into the cell structure, the long short-term memory (LSTM) could handle the problem of long-term dependencies well. Since its introduction, almost all the exciting results based on RNNs have been achieved by the LSTM. The LSTM has become the focus of deep learning. We review the LSTM cell and its variants to explore the learning capacity of the LSTM cell. Furthermore, the LSTM networks are divided into two broad categories: LSTM-dominated networks and integrated LSTM networks. In addition, their various applications are discussed. Finally, future research directions are presented for LSTM networks.