Master of Science (MS)
Hua H. Li
Artificial Neural Network, Chemical Oxygen Demand, Cloud Computing, Embedded System, Spectroscopy
Computer engineering; Water resources management
One of the many parameters indicating water quality is chemical oxygen demand (COD), which is an indirect measurement of the amount of organic compound material in water. There have been many studies, in both academia and the industry, to analyze the COD content of water using spectral analysis. The proposal of this thesis was to study, analyze, and identify methods to determine the presence of COD using UV spectroscopy data and an artificial neural network (ANN) in a cloud-connected embedded system. The system was implemented using an ARM11 board and a portable spectrometer. Light in the UV range was used to analyze the water sample. As an analysis strategy, the spectral data were converted into a real number value in the range of 0 to 1. Twenty equidistance samples were taken out of the converted data to be fed into the ANN, and the ANN was trained with known samples to identify any presence of COD. Experiments used laboratory-calibrated water samples with known COD and some real-life water samples. All the experiments showed that the implemented system could successfully indicate the presence or absence of COD in the given water sample. The system also successfully demonstrated the application of a cloud-connected embedded system to an area in environmental engineering. This indicated that the system was a bridge between computer and environmental engineering.
Patra, Kaushik, "Embedded Cloud System for Ann-Cod Analysis Using UV Spectroscopy" (2013). Master's Theses. 4401.