SIMULATING FAULT IN DISTILLATION COLUMN USING ASPEN HYSYS

Overview

For the upcoming Machine Learning course, we would like to give a case study to develop a machine learning model which can be used to predict if there's any fault that occurs in the distillation column and identify the type of fault.

As we might have known, to develop a machine learning model, we need data for it to train and test it, so those data will be generated using Aspen HYSYS (the original reference simulated this in Aspen Plus)

Distillation Column Environment (Credit to the author and owner)

The simulated fault are as shown in the table below:

Next, zero-mean normal distributed noise is added to the data to simulate noise from sensor reading

The readings goes haywire once the fault is introduced and starts to stabilise after a while

This is how the simulation were made:

Reference

Multiple Fault Diagnosis in Distillation Column Using Multikernel Support Vector Machine Syed A. Taqvi, Lemma Dendena Tufa, Haslinda Zabiri, Abdulhalim Shah Maulud, and Fahim Uddin Industrial & Engineering Chemistry Research 2018 57 (43), 14689-14706 DOI: 10.1021/acs.iecr.8b03360