Identification of pH process using Hammerstein-Wiener model

Abstract

Potential of Hydrogen (pH) neutralization process plays a significant role in process industries. In pH neutralization process, identification can be done in many ways. Modeling of pH neutralization process is necessary for proper control. Neural networks used for modeling the process had a larger training error and learning rate selection was complex. Wiener model used for pH neutralization process identification had less model fit and larger training error. This paper presents the use of Hammerstein-Wiener model for identification of pH neutralization process. The proposed model has a dynamic linear part between two static nonlinear parts. Piecewise linear, sigmoidal, saturation, dead zone and one dimensional polynomial nonlinearities are used as nonlinear part and transfer function is used as linear part. Model is found for different nonlinearities and their model fits are compared. A model fit of 65.46% is obtained using Hammerstein-Wiener model. This paper also compares the results of Wiener model and Hammerstein-Wiener model.