Decision model for coagulant dosage using genetic programming and multivariate statistical analysis for coagulation/flocculation at water treatment process

In this research, genetic programming and multivariate statistical analysis techniques have been applied for decision support on the coagulant dosage and the mixing ratio as two kinds of coagulants have been injected at the same time in the coagulating sedimentation process of water treatment. The coagulant dosage has typically been determined through the Jar-test, which requires a long experiment time in a field-water treatment plant. It is difficult to efficiently determine the coagulant dosage since water quality changes with time. As there are no human experts who have sufficient knowledge and experience in the field, coagulants may be injected with an improper mixing ratio, which causes poor performance in the coagulating sedimentation process. In this study, a model for the approximation of coagulant dosage has been developed using genetic programming (GP). The performance of this model was evaluated through validation. A guideline on the optimal mixing ratio between PACS (Poly Aluminum Chloride Silicate) and PAC (Poly Aluminum Chloride) has been provided through statistical analysis.