Paper Title
SMART WATER QUALITY PARAMETERS MONITORING FOR ESTIMATION OF BIOCHEMICAL OXYGEN DEMAND (BOD): DEVELOPMENT OF BOD SOFT SENSOR
Abstract
BOD serves as a crucial indicator of organic pollution in water bodies, reflecting the impact of human
activities and potential harm to aquatic ecosystems due to oxygen depletion caused by the decomposition of organic matter.
The fluctuation of BOD, the need for time-consuming testing procedures, resource limitations, and inconsistent regulatory
criteria hinder comprehensive efforts to measure water quality. Innovative technologies like sensor-based systems and
advanced analytical techniques are being developed to improve BOD monitoring, aiming for more precise and timely
measurements to support informed decision-making for protecting and managing aquatic ecosystems. This project aims to
create an intelligent water quality monitoring system that would promote sustainable aquaculture through effective water
quality management by constructing a BOD soft sensor and forecasting biochemical oxygen demand (BOD). With an Rsquared
of 0.80 and a Root Mean Squared Error (RMSE) of 0.37, the Multi-modal Ensemble (KNN, RF, and XGB) had the
lowest Mean Squared Error (MSE) of 0.14. Temperature, dissolved oxygen, pH, and conductivity do not significantly differ
between the four groups, according to the analysis; conductivity has the lowest p-value of the comparisons between the four
groups, at 0.185. Successful implementation of the BOD Soft Sensor holds significant implications for efficient water quality
management, enabling proactive interventions to mitigate pollution and safeguard aquatic ecosystems to revolutionize
environmental monitoring practices, facilitating more informed decision-making and enhancing sustainability efforts.
Keywords - Biological Oxygen Demand, Soft Sensors, KNN, Random Forest, XGBoost, Smart water quality.