Fouling resistance prediction using artificial neural network nonlinear auto-regressive with exogenous input model based on operating conditions and fluid properties correlations
- Department of Engineering Physics, Institute Technology of Sepuluh Nopember Surabaya, Surabaya, Indonesia 60111 (Indonesia)
Fouling in a heat exchanger in Crude Preheat Train (CPT) refinery is an unsolved problem that reduces the plant efficiency, increases fuel consumption and CO{sub 2} emission. The fouling resistance behavior is very complex. It is difficult to develop a model using first principle equation to predict the fouling resistance due to different operating conditions and different crude blends. In this paper, Artificial Neural Networks (ANN) MultiLayer Perceptron (MLP) with input structure using Nonlinear Auto-Regressive with eXogenous (NARX) is utilized to build the fouling resistance model in shell and tube heat exchanger (STHX). The input data of the model are flow rates and temperatures of the streams of the heat exchanger, physical properties of product and crude blend data. This model serves as a predicting tool to optimize operating conditions and preventive maintenance of STHX. The results show that the model can capture the complexity of fouling characteristics in heat exchanger due to thermodynamic conditions and variations in crude oil properties (blends). It was found that the Root Mean Square Error (RMSE) are suitable to capture the nonlinearity and complexity of the STHX fouling resistance during phases of training and validation.
- OSTI ID:
- 22608926
- Journal Information:
- AIP Conference Proceedings, Journal Name: AIP Conference Proceedings Journal Issue: 1 Vol. 1737; ISSN APCPCS; ISSN 0094-243X
- Country of Publication:
- United States
- Language:
- English
Similar Records
Effects of mitigating fouling on the energy efficiency of crude-oil distillation
Artificial neural networks estimate evapotranspiration for Miscanthus × giganteus as effectively as empirical model but with fewer inputs