A soft sensor development with dsPIC implementation for water quality monitoring

Abstract EN

In the field of water quality monitoring, there is a huge number of heterogeneous sensors which may be time-consuming in the measurement and processing stages.

For several reasons, such as the high cost of some sensors, their number, the cleaning operation, and the calibration routines, make their proper operation hard to ensure.

However, soft sensors approach can provide an effective and economic way to solve this problem.

This work is part of a clean and wastewater monitoring system design, where dissolved oxygen (DO) represents a very important parameter of water quality.

The measurement of DO is done conventionally by specific probes but with very strict conditions of measurement hence the emergence of prediction methods to make a soft measurement.

In this paper, we propose a soft sensor allowing the DO estimation in the surface water where Artificial Neural Network (ANN) architecture is used as a prediction model.

The developed model is then implemented in a dsPIC microcontroller for a possible use in a water quality monitoring system