System Identification and Control for Scrubber Unit at MLNG

Nowadays, there are a lot of research that focus on plant improvement;
especially in Malaysia Liquefied Natural Gas (MLNG). Without these improvement, the
outstanding cost for the plant operation could not be reduced; these improvement is
needed to reduce the cost operation so that more revenue will be gained. Thus,
optimization are needed to develop profit in any methods available; for this case is to
reduce cost of plant operation. Before the project can proceed, research on System
Identification and the process are needed. In this paper, the first approach for
optimization is to model the plant step test data. Step test data was done in MLNG.
Then, obtained data was identified by using System Identification Toolbox in
MATLAB. The data was modeled by using the System Identification methods available
such as Auto Regression eXogenous (ARX), Box Jenkins (BJ), and also State Space
Analysis. The model will be estimate and validate, hence the best model method will be
determined by Best Fit Analysis. Then, the model will be control by using poleplacement
method and projected output will be compared with the plant data. The
outcome of this project proven that the plant can be more optimized with deeper
research and analysis of the plant. The more optimized the plant will be, the more profit
gain the plant will be.