基于PSO-SVR算法的燃煤锅炉结渣特性预测

针对燃煤锅炉结渣特性的有限样本、非线性和高维数问题,提出了一种基于粒子群优化（PSO）和支持向量回归（SVR）的预测模型。对于支持向量回归机在建模中存在的参数选取问题,采用改进的粒子群算法（PSO）对模型参数进行优化,该方法结合了PSO的快速全局优化能力和SVR的结构风险最小化理论,精确地逼近非线性映射关系的能力。仿真结果表明：相比遗传算法（GA）SVR预测模型和模拟退火（SA）SVR预测模型,PSO-SVR模型预测燃煤锅炉结渣特性具有较高的准确率。

Predicting Coal-fired Boiler Slagging Characteristics Based on PSO-SVR Algorithm

Aiming at the finite sample,nonlinearity and high dimension of the coal-fired boiler slagging characteristics,a prediction model based on improved particle swarm optimization（PSO） and support vector regression（SVR） was proposed.As for SVR＇s parameter selection in modeling,the improved PSO algorithm was used to optimize model parameter,this method which having minimum structure risk theory combined with SVR＇s accurate non-linear simulation and the PSO＇s fast global optimization can accurately approximate nonlinear mapping relationship.Simulation results show that the proposed PSO-SVR model outperforms both GA-SVR and SA-SVR models while accurately predicting coal-fired boiler＇s slagging characteristics.