基于粒子群算法优化LNG(C3-MR)流程的研究

Particle swarm optimization of the parameters for C3-MR process of LNG

关键词:  LNG  C₃-MR  粒子群算法  优化  比功耗 

Key words:  LNG  C₃-MR  particle swarm optimization (PSO)  optimization  specific work

摘要:

丙烷预冷混合制冷液化流程(C₃-MR)在液化天然气生产中应用最广。该流程的优化属于非线性问题,优化结果受到过程变量和算法的影响。基于HYSYS软件模拟,对C₃-MR流程用MATLAB嵌入粒子群算法(PSO)优化制冷剂组分、流量以及流程压力以降低过程能耗。研究结果表明,对C₃-MR流程使用PSO算法优化迭代20次便收敛,优化后理论能耗低于公开文献报导的序列二次规划(SQP)和BOX方法的结果。 

Abstract:

Propane pre-cooled mixed refrigerant (C₃-MR) process is the most widely used process for LNG production. Nonlinear optimization of the process depends on process variables and algorithm. The refrigerant components, flow rate and process pressure of the C₃-MR process are optimized by HYSYS with particle swarm optimization (PSO) embedded in MATLAB. The results show that optimized parameters in PSO converge after 20 iterations for the C₃-MR process. The resultant theoretic energy consumption is lower than that obtained by the sequential quadratic programming method (SQP) and BOX method in reported literatures.