Estimation of oxygen demand levels using UV-Vis spectroscopy and artificial neural networks as an effective tool for real-time, wastewater treatment control

Abstract

In order to effectively manage and treat wastewater, it is important to constantly monitor the oxygen demand levels, so that remediation process can be implemented immediately if problems are found. To enable constant monitoring of oxygen demand levels, a simple and effective method based on the mathematical treatment of spectral absorbance patterns, using artificial neural networks (ANNs) is demonstrated fro rapidly estimating biochemical oxygen demand (BOD) and chemical oxygen demand (COD) values of wastewater samples. The method involves recording spectrum absorbance patterns from 190 to 350 nm and processing the patterns obtained using an ANN to indirectly estimate BOD and COD values. The results indicated that in most cases the proposed technique (UV-ANN) worked well, as UV-ANN derived BOD and COD values were very close to those estimated using the standard BOD and COD methods. The UV-ANN derived values also followed the trends of the standard methods closely, which make them ideal for real-time, on-line process control of wastewater effluents, as problems such as exceeding critical limits could be identified easily.