基于改进PSO算法和集成神经网络的裂解炉在线优化  

Online Optimization of Cracking Furnace Based on Advanced PSO Algorithm and Neural Network Ensembles

作　　者：庄敏慧[1] 张照娟[1] 王振雷[1] 钱锋[1] 

机构地区：[1]华东理工大学化学工程联合国家重点实验室,上海200237

出　　处：《华东理工大学学报（自然科学版）》2009年第5期756-761,共6页Journal of East China University of Science and Technology

基　　金：国家杰出青年科学基金(60625302);高等学校学科创新引智计划资助(B08021);国家973项目(2009CB320603);国家863计划课题(2006AA04Z168;2007AA041402);长江学者和创新团队发展计划资助(IRT0721);国家科技支撑计划(2007BAF22B05);上海市科技攻关项目(08DZ1123100);上海市重点学科建设项目资助(B504);上海市科技启明星计划(07QA14015)

摘　　要：针对传统粒子群算法(PSO)寻优时易陷入局部最优、后期全局搜索能力下降等不足,提出了基于载波的粒子群算法(CWPSO)。通过粒子基于载波的搜索和载波扩展精确寻优,较好地克服了上述缺点,且寻优时间明显减少。同时,针对工业裂解炉在线优化要求,采用了权值动态集成的集成神经网络(NNE)对双烯收率进行建模预测,并结合CWPSO算法进行了在线滚动优化。仿真结果表明,该方法对裂解炉的优化效果明显,双烯平均收率有了明显提高。The traditional Particle Swarm Optimization （PSO） algorithm is easily trapped in the local optimum and converges slowly. Due to the shortcomings above, a novel PSO algorithm based on the carrier-wave （CWPSO） is presented in the paper, which searches through the carrier-wave and takes a precise search by means of carrier-wave extending. As a result, it overcomes the above shortcomings better, and has a shorter searching time as well. In addition, towards the online optimal requirements of the industrial cracking furnace, a neural network ensembled with dynamic weights is applied in the predictive modeling of C2 H4 and C3 H6 yield rates, then the online rolling optimization is carried out. The simulating result shows that the optimal method has sound effects for the cracking furnace, and there is a palpable improvement of C2 H4 and C3 H6 yield rates.

关 键 词：CWPSO 集成神经网络 在线优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]