基于动态等级PSO-ELMAN的乙烯裂解深度模型及其优化控制

提出一种自适应动态等级粒子群算法(ADHPSO)。该算法保持粒子多样性,能摆脱局部极值,有良好的全局收敛性。将ADHPSO训练ELMAN神经网络,建立乙烯裂解炉裂解深度的在线预测模型。研究一种集成ADHPSO-ELMAN过程建模的裂解深度智能优化控制方法,得到裂解过程的最优操作条件。仿真计算表明,该方法显著提高了乙烯及丙烯的收率,具有良好的稳定性和适应性,对实际生产具有极大的应用潜力。

Optimal control of cracking depth based on ADHPSO-ELMAN for ethylene cracking furnace

An adaptive dynamic hierarchical version of the particle swarm optimization (ADHPSO) metaheuristic is proposed in this paper, which gets rid of immersing to the local optimization and has good global searching performance. Depending on the quality of their so-far best-found solution, the diversity of the particles had been maintained. Meanwhile, the ELMAN neural network is trained by ADHPSO , and established the forecasting model of ethylene cracking depth online, study a intelligent optimize control method of integrating ADHPSO with ELMAN process model, got the optimal operating conditions for cracking process. The simulation results show that the method can greatly improve the ethylene and propylene yield with good stability and adaptability to the actual production, and has great potential application.