Using IoT to assist monitoring of the methane gas extraction at Lake Kivu
Methane gas is a powerful greenhouse gas with global warming potential. The current techniques being used to monitor the leaks are expensive and likely onerous and demands for trained operators. There are available solutions tried by the space agencies such as National Aeronautics and Space Administration (NASA) and European Space Agency (ESA) using satellites to better understand the distribution of greenhouse gases on regional and global scales. Those are ENVISAT, GOSAT, OCO-2, and the recently launched TROPOMI instrument on the Sentinel 5P satellite, but all these, regardless of the advanced technology associated cannot pinpoint the source of emissions. In this study, the performance of low-cost Internet of Things (IoT) sensors and isolation forest anomaly detection machine learning technique was implemented. Isolation Forest is one of the outstanding outlier detectors in the real-time DataStream for faulty detection, and money laundering in banking industry. It was tested in this system to improve the accuracy in detecting the methane gas leak. According to the experimental results, the anomaly detection based on isolation forest achieved an excellent performance in terms of accuracy of outlier detection while minimizing the false positives. Decarbonization is an essential component in the climate system, and this plays a key role in reducing methane emissions. Finally, the study presents future research directions to carry out research on the machine learning with Internet of Things (IoT).
Capstone Project submitted to the Department of Engineering, Ashesi University in partial fulfillment of the requirements for the award of Bachelor of Science degree in Computer Engineering, May 2021