Online Top Composition Prediction of a Distillation Column Using Feedforward Neural Network

The development of neural network prediction had widely discovered by the industrial world nowadays. The efficiency of predicting and manipulating the outcome of the input based on the given functions makes neural network favourable. Neural Network which inspired by the brain neural system itself also capable of dealing with the non-linear functions which make it suitable for controlling complex operation units such as distillation column. By the mean of study, this paper will be looking into feed forward neural network prediction. The main focus is to validate the efficiency of feed forward neural network in producing the expected composition as the output. The research will be conducted step by step by firstly gathering the relevant data from the distillation column. It will then be proceed with data analysis and validation of each component. This study will utilized the MATLAB software to produce the expected results. The study is expected to come with persuasive results on the effectiveness of feed forward neural network prediction.