Modeling of submerged membrane bioreactor filtration process using NARX-ANFIS model

This paper presents modeling techniques for submerged membrane bioreactor (SMBR) filtration process using. The Nonlinear Auto Regressive with Exogenous Input (NARX) structure was used with adaptive neuro-fuzzy interface system (ANFIS) and feed forward neural network (FFNN) are employed to model the filtration system. The transmembrane pressure and the permeate flux were model during the relaxation and permeate cycle. In this work diluted palm oil mill effluent (POME) will be used as an influent of the treatment process. The performance of the models was measured using the R², mean square error (MSE) and mean absolute deviation (MAD). The result showed that the ANFIS with NARX structure perform slightly better compare with ANN with NARX structure.