Abstract
In this work, the pH neutralization process is described by a neural network Wiener (NNW) model. A nonlinear Model Predictive Control (NMPC) is established for the considered process. The main difficulty that can be encountered in NMPC is solving the optimization problem at each sampling time to determine an optimal solution in finite time. The aim of this paper is the use of global optimization method to solve the NMPC minimization problem. Therefore, we propose in this work, to use the Self Organizing Migrating Algorithm (SOMA) to solve the presented optimization problem. This algorithm proves its efficiency to determine the optimal control sequence with a lower computation time. Then the NMPC is compared to adaptive PID controller, where we propose to use the SOMA algorithm to formulate the PID in order to determine the optimal parameters of the PID. The performances of the two controllers based on the SOMA algorithm are tested on the pH neutralization process.
Highlights
The pH neutralization process is characterized by a nonlinear behavior
We propose in this work to use the Self Organizing Migrating Algorithm (SOMA) algorithm to solve the presented optimization problem
The pH neutralization process will be described by a neural network Wiener (NNW) model, where the linear block is described by an autoregressive model and the nonlinear block is given by a multilayer feed-forward neural network with one hidden layer
Summary
The pH neutralization process is characterized by a nonlinear behavior. The high nonlinear characteristic of this process makes the control of the pH a hard task. To avoid solving the nonlinear optimization problem and reduce the implementation complexity of the NMPC, [20], proposed to describe the nonlinear process by a set of uncertain linear models instead of one nonlinear model. We can note that all these works avoid to solve the nonconvex optimization problem regarding the difficulty of implementation and the high computation burden necessary at each sampling time. The sampling period of the process under consideration must be respected at each iteration when solving the NMPC problem This constraint is difficult to be satisfied using global optimization method due to their slow convergence. We propose to use the Self Organizing Migrating algorithm (SOMA) in this work to solve the optimization problem.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
