Discrete-Time Neural Network Modelling Of Industrial Air Compression System

Khong, Fan Hao (2021) Discrete-Time Neural Network Modelling Of Industrial Air Compression System. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (24 Pages) Discrete-Time Neural Network Modelling Of Industrial Air Compression System.pdf - Submitted Version Download (331kB)

Abstract

System identification is an approach of constructing the mathematical model of a dynamical system using the instrumentation signal of input and output of the system. This approach could be integrated into industrial processes such as air compression system by using the NARX model as a base model with the functions of the neural network. Basically, in any industrial process, the non-linearity behaviour of the system makes the predictive control to be complicated because the existence of random variable is unavoidable. Furthermore, the random behavior must not be ignored as it may indicate any unknown event occurring during the process. This project aims to perform system identification using neural network techniques for industrial air compression system. Besides, the validation of model’s predictive performance is also included in this report. The proposed methodology of this project involves the data acquisition stage until the end of simulation and validation stages. The outcome of the simulation would undergo a series of analysis to determine the most suitable NARX-NN model architecture configuration. Finally, the predicted data is compared to the industrial data to verify its accuracy and difference, which shows that this model had successfully ruled out the suspicious random event data

Actions (login required)

View Item