摘要
针对锅炉燃烧系统智能算法建模,为克服标准粒子群(PSO)优化算法对最小二乘支持向量机(LSSVM)模型参数进行优化时容易陷于局部最优解的缺点,提出一种改进型的蚂蚁−粒子群算法(MAPSO)对LSSVM模型参数进行优化。根据模式搜索的“探测”思想,通过与蚁群算法移动规则的结合改进粒子群算法,增加粒子群算法的小步长局部搜索过程,让每一步迭代所得的最优粒子在其邻域内进行小步长的局部精细搜索,以便找到更好的全局最优解。MAPSO算法的先验性预判断组合共有8种,通过对比先验性粒子各个方向的适应度,保证粒子群在最优粒子邻域内向正确的方向进行小步长局部搜索。研究结果表明:本文提出的改进型MAPSO算法有效保留了标准PSO优化算法的全局寻优性能,并增强了算法的局部寻优性能。MAPSO算法在建立锅炉燃烧系统模型过程中能有效避免陷入局部极小,寻找到全局最优解。所建立的MAPSO-LSSVM模型与标准PSO-LSSVM模型相比,具有更快的收敛速度,更高的预测精度,更强的拟合能力和泛化能力。
Aiming at the intelligent algorithm modeling of boiler combustion system,in order to overcome the shortcoming that the standard particle swarm optimization(PSO)algorithm is easy to fall into the local optimal solution when optimizing the parameters of the least squares support vector machine(LSSVM),a modified antparticle swarm optimization(MAPSO)was proposed to optimize the parameters of the LSSVM model.According to the"detection"idea of pattern search,the particle swarm algorithm was improved by combining with the moving rules of ant colony algorithm.By adding a small-step local search process of the particle swarm optimization algorithm,the optimal particles obtained in each step of iteration can perform a small-step local fine search in its neighborhood,and then find a better global optimal solution.There are eight combinations of a priori pre-judgment of the MAPSO algorithm.By comparing the fitness values of the prior particles in all directions,the particle swarm is guaranteed to perform a small-step local search in the correct direction in the optimal particle neighborhood.The results show that the improved MAPSO algorithm proposed in this paper effectively retains the global optimization performance of the standard PSO algorithm and enhances the local optimization performance of the algorithm.In the process of establishing the boiler combustion system model,the MAPSO algorithm can effectively avoid falling into a local minimum and find the global optimal solution.Compared with the standard PSO-LSSVM model,the established MAPSO-LSSVM model has faster convergence speed,higher prediction accuracy,stronger fitting ability and generalization ability.
作者
蓝茂蔚
李杨
赵国钦
周元祥
江政纬
甘云华
LAN Maowei;LI Yang;ZHAO Guoqin;ZHOU Yuanxiang;JIANG Zhengwei;GAN Yunhua(School of Electric Power Engineering,South China University of Technology,Guangzhou 510640,China;Xi'an Thermal Power Research Institute Co.Ltd.,Xi'an 710054,China;Jinghai Power Generation Co.Ltd.,Guangdong Energy Group,Jieyang 515223,China)
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2022年第4期1506-1515,共10页
Journal of Central South University:Science and Technology
基金
广东省基础与应用基础研究基金资助项目(2020B1515020040)
国家自然科学基金资助项目(51776077)
天津大学内燃机燃烧学国家重点实验室开放基金资助项目(K2021-01)。
关键词
电站锅炉
锅炉效率
NO_(x)排放
改进型粒子群算法
最小二乘支持向量机
power plant boiler
boiler efficiency
NO_(x) emission
modified particle swarm optimization
least square support vector machine
