Design and Control of Steam Flow in Cement Production Process using Neural Network Based Controllers

TL;DR

This paper presents a neural network-based control system for steam flow regulation in cement production, demonstrating improved tracking performance through simulation comparisons of different controllers under various conditions.

Contribution

It introduces a neural network predictive controller for steam flow control in cement industry, validated through simulation with noise disturbance and comparison with other controllers.

Findings

Neural network predictive controller effectively tracks desired steam flow.

Controller performance remains robust under sensor noise.

Proposed system outperforms traditional controllers in simulations.

Abstract

In this paper a NARMA L2, model reference and neural network predictive controller is utilized in order to control the output flow rate of the steam in furnace by controlling the steam flow valve. The steam flow control system is basically a feedback control system which is mostly used in cement production industries. The design of the system with the proposed controllers is done with Matlab/Simulink toolbox. The system is designed for the actual steam flow output to track the desired steam that is given to the system as input for two desired steam input signals (step and sine wave). In order to analyze the performance of the system, comparison of the proposed controllers is done by simulating the system for the two reference signals for the system with and without sensor noise disturbance. Finally the comparison results prove the effectiveness of the presented process control system with model reference controller.