Automatic detection of exogenous respiration end-point using artificial neural network

When aerobic bacteria receive a biodegradable material such as wastewater, then respiration changes from endogenous to exogenous. The reverse occurs when biodegradation is complete. When using respirometry a respirogram is recorded showing those changes in respiration, and for an expert it is not difficult to point the moments at which they occur. The area corresponding to the exogenous respiration phase is a measure of the easily biodegradable fraction of material, also called the short-term BOD or BODST. That value, in combination with a value for COD, can be used to determine the treatability of wastewater. Respirometry can also be applied on-line, e.g. for on-line monitoring of wastewater. However, automatic detection of the end-point of exogenous respiration is difficult. The first step towards on-line monitoring of wastewater treatability is to make automatic detection of this end-point possible. In this study the use of a neural network for detection of this end-point was investigated. Results are promising; after training the neural network is able to detect the correct end-point in the majority of the studied cases