Evaluating and optimizing the cold energy efficiency of power generation and wastewater treatment in LNG-fired power plant based on data-driven approach.

Energy shortage, greenhouse effect and environmental pollution are the three major crises faced by mankind. How to solve these three problems effectively has become the focus of all countries. Therefore, this paper proposes an innovative LNG cold energy utilization method that can simultaneously alleviate these three problems. This method includes an improved structure of the organic Rankine cycle, CO 2 capture cycle and supercritical CO 2 extraction of wastewater. In addition, automated machine learning based on data-driven approach is effectively used for modeling the wastewater extraction. Then, after verifying the accuracy of the proposed model, some parameters have great influence on the system performance. For instance, the temperature of supercritical CO 2 and working fluid water affect the local exergy efficiency and their regression coefficients are −0.29%/°C and 0.13%/°C, respectively. These parameters in wastewater extraction process are studied. Finally, under the optimization of genetic algorithm, the maximum wastewater treatment can reach 62.7 t/h increasing by 14.0%, and the LNG cold energy generation can reach 110.1 kWh/tLNG with an increase of 28.9%, while the exergy efficiency of the system is 50.4% increasing by 24.1%. Thus, organic Rankine cycle power generation, carbon capture and wastewater treatment play the roles of energy saving, emission reduction and environmental protection. At the same time, the results are compared with those of other scholars to verify the superiority of the proposed scheme. [Display omitted] • LNG cold energy reuse covering electricity, CO 2 capture and extraction is proposed. • Automated machine learning approach is used to build wastewater extraction model. • Reasonable arrangement of regenerative heat exchanger enhances system performance. • The maximum wastewater treatment reaches 62.7 ton/h utilizing genetic algorithm. [ABSTRACT FROM AUTHOR]