Sequential coordinate random search for optimal operation of LNG (liquefied natural gas) plant

This study exploits the sequential coordinate randomization search method for optimizing LNG (liquefied natural gas) process plants. The coordinate search is based on the idea of minimizing the multivariable function considering one variable at a time. The random element is incorporated in the coordinated search for an exhaustive exploration of decision variable space. A simple implementation with few operating parameters makes the proposed approach suitable for the optimization of highly non-linear LNG process plants. The efficacy of the proposed methodology was tested on the well-known SMR (single mixed refrigerant) and propane pre-cooled mixed refrigerant (C3MR) process of NG liquefaction. The main decision variables in SMR and C3MR process plants were identified and optimized in terms of the compression energy. The results were compared with the heuristic results, which revealed the superiority, simplicity and suitability of the proposed sequential coordinate random search algorithm for LNG process plants. •LNG process plants are optimized using SCRS (sequential coordinate random search).•Simple implementation of coordinate descent and inclusion of random element are main attributed of SCRS.•LNG plant optimization with SCRS methodology bring consistent results.•SCRS methodology can be exploited to any process model developed in simulator.