Decision Model for Optimization of Coagulation/Flocculation Process for Wastewater Treatment

The dye wastewater is an important environmental problem. Coagulation-flocculation process is an available technology in treatment of this kind of wastewater. The aim of the present study was to develop a model for coagulation-flocculation process using ferric chloride as a flocculant and the model could provide an alternative to the experimental jar test for determining the operational variables for treatment of the row wastewater. Also, some parameters affecting the degradation degree such as coagulant dosage, wastewater initial pH, agitation speed and time, were examined. The results showed that the linear correlation coefficient (R) between the simulation values and the expected values was 0.94 after network training and 0.92 after testing. The simulation effect of RBF neural network is related to the maximum number of neurons of network and the spread of radial basis functions. When the maximum number of neurons of network is 28 and the spread of radial basis functions is 2.1, the ratio of test error to training error is l.36 and the simulation result is the best. The experiment was carried out based on better simulation results and the degradation degree could reach 93% when coagulant dosage, wastewater initial pH, agitation speed and rapid agitation time are 0.3 ml/l, 7.6, 300 rpm and 1 min respectively. In addition, the importance of the parameters on degradation degree was investigated by the model. The degradation degree was enhanced significantly as listed herein decreasing order of effectiveness: coagulant dosage >; agitation speed >; agitation time >; wastewater initial pH.