Affiliations: UM Power Energy Dedicated Advanced Centre (UMPEDAC), Level 4, Wisma R&D, Malaysia | Department of Chemical Engineering, Faculty of Engineering, University of Malaya, Kuala Lumpur, Malaysia | Department of Electrical Engineering, Faculty of Engineering, University of Malaya, Kuala Lumpur, Malaysia
Note: [] Corresponding author. M.A. Hussain, UM Power Energy Dedicated Advanced Centre (UMPEDAC), Level 4, Wisma R&D, Malaysia. Tel.: +60 379675206; Fax: +60 379675319; E-mail: mohd_azlan@um.edu.my
Abstract: This paper proposes a hybrid model to identify the on-line pH characteristic of a neutralization plant. The hybrid model is the combination between neuro-fuzzy identification technique and first principle model. The neuro-fuzzy identification technique used training dataset to map the neutralization response curve in full ranges. The first principle model is based on material balances and chemical equilibrium equation. The objective of the proposed model is to extend the robustness effect for the on-line titration characteristic without having to re-design the model if the plant undergoes different conditions. In the experiment, the proposed model's dynamic response was compared with the on-line pH data. It showed the best fit for hybrid model with dynamic weight adjustment in nominal condition (RSME = 0.1013) and in altered condition (RMSE = 0.5616) proved it capability in capturing the additional variations to a pH neutralization plant.
