Appling machine learning for estimating total suspended solids in BFT aquaculture system

Abstract

Biofloc Technology (BFT) systems are used to improve water quality and the production of aquatic organisms, and they influence dissolved oxygen, alkalinity, and pH, directly affecting the efficiency and success of this production system. Measuring total suspended solids (TSS) in water demands substantial investments and involves a time-consuming process to obtain results. This delay in obtaining results poses a significant challenge to the operations of these farms. In this study, we applied Artificial Neural Networks (ANN) and Support Vector Machine (SVM) methods based on artificial intelligence and water quality parameters (easy to measure, low cost, and quick response) to identify the most accurate method for measuring TSS. The best TSS estimate was achieved with SVM using nitrite and turbidity as predictive variables, which tended to overestimate the real value by 19 %, presenting a potential for application in estimating TSS in the BFT aquaculture system.