Abstract
Since natural gas liquefaction operation is an energy-intensive technology. Research has been published for the optimization of the process by utilizing the first principal complex nonlinear SMR model. The process is computationally expensive and requires several hours or even days for optimization. To address this issue, this study proposes a metamodel for natural gas liquefaction featuring a single mixed refrigerant cryogenic cycle. The SMR model was modeled and simulated by utilizing Aspen Hysys software. However, the metamodel was developed by utilizing radial basis function methodology and optimization work using Matlab. The results summary comparison for specific compression duty for metamodel was 0.3863 kW/kg-NG. The first principle-based published study of this duty was 0.3625 (kW/kg-NG) which is approximately 6.5% lower than the current study. However, a huge reduction in computational time was obtained. The metamodel building and optimization time-lapse was 292.96 s while the same first principle-based model lapse 201.24 h using 300 iterations. In comparison, it can be inferred that metamodel could capture almost all-important characteristics of the firs-principle based model. Hence, the proposed study could be considered as an alternative to the first principle model for optimization purposes. This research may prove to be more vital especially in any abrupt changing or uncertain conditions or for real-time plant optimization.
•Optimization of the first principle-based model is computationally expensive.•Metamodel minimizes energy with huge timesaving in optimization.•The study may be applied to other computational expensive engineering problems.•Several optimal operating solutions can be obtained in a few minutes.•The optimal solutions may be useful for the operation of plants in uncertain conditions.