On-line evaluating the SS removals for chemical coagulation using digital image analysis and artificial neural networks

Abstract

Chemical coagulation is one of the most important processes for industrial wastewater treatment plants to remove the suspended solids (SS), which depend significantly on particle characteristics. A digital image analysis system was set up in this study for the on-line measurements of particle characteristics, including particle size distribution, equivalent diameter, total area, total volume, and the fractal dimension in the both coagulation and flocculation periods in chemical coagulation. Two real industrial wastewaters, textile wastewater and landfill leachate, were used for conducting the coagulation and flocculation processes with different polyaluminum chloride dosages in a batch reactor. The artificial neural network (ANN) models were used to construct the correlations between the monitoring data acquired and the SS removal efficiencies. The experimental results indicated that the ANN models were able to precisely predict the SS removal efficiencies and effluent SS concentration after the chemical coagulation, with the correlation coefficient (R 2) of 0.96–0.97 for real landfill leachate and R 2 of 0.93–0.97 for real textile wastewater, which provided significant benefits for the control of chemical coagulation.