摘要
为解决煤矿单传感器瓦斯浓度预测精度不足的问题,将自适应人工免疫系统(AIS)与自适应粒子群(PSO)相结合,建立多参数并行双自适应AIS-PSO算法的瓦斯浓度软测量模型。通过分析煤矿井下环境参数对瓦斯浓度监测的影响,将矿井下温度及风速等环境参数作为软测量模型输入,上隅角瓦斯浓度作为模型输出。利用并行双自适应AIS-PSO算法对最小二乘支持向量机(LS-SVM)的核参数σ和正则化参数γ进行寻优,并与PSO-LSSVM、LS-SVM结果进行对比。结果表明:PSO-LSSVM平均相对误差为5.5083%,LS-SVM平均相对误差为8.6883%,并行双自适应AIS-PSO软测量模型的平均相对误差为2.0165%,最小相对误差为1.194%,与另两种方法相比具有较高的预测精度和泛化能力。
To deal with the problem in the low forecasting accuracy of coal mine single sensor gas cc,a parallel double adaptive AIA-PSO gas concentration soft-sensing model was built on the basis of combining of adaptive Parti-cle Swarm Optimization(PSO)and adaptive Artificial Immune System(AIS).Through analyzing the effect of envi-ronment parameters of mine on gas concentration monitor,the wind speed and environment temperature and etc.were extracted as the inputs to the model which subsequently gave the top corner gas concentration.the parallel double a-daptive AIA-PSO algorithm was used to optimize the kernel parameterσand the regularization parameterγof LS-SVM.The results show that relative errors in prediction with the model are not greater than 2.0165%,and that the model has higher prediction accuracy and stronger generalization ability than both the PSO-SVM and LS-SVM in pre-diction accuracy.
作者
单亚锋
高振彪
SHAN Ya-feng;GAO Zhen-biao(Faculty of Electrical&Control Engineering,Liaoning Technical University,Huludao Liaoning 125105,China)
出处
《计算机仿真》
北大核心
2020年第1期338-342,393,共6页
Computer Simulation
基金
国家自然科学基金项目(71371091,71771111)。
关键词
粒子群
人工免疫系统
瓦斯浓度
最小二乘支持向量机
软测量
Particle swarm optimization(PSO)
Artificial immune system(AIS)
Gas concentration
Least square support vector machine(LS-SVM)
Soft-sensing
