基于智能算法优化SVM的横向通风过程中温度场预测方法探究

将改进的智能预测技术应用于储粮横向通风过程中的粮堆温度预测,为粮食通风智能预测与决策提供了一种新思路。选取河北清苑国家粮食储备库冬季横向通风的实时监测数据,在分析主要影响因素的基础上,应用三种智能优化算法——网格寻优算法、GA遗传算法寻优、PSO粒子群算法,结合回归支持向量机理论,对粮堆的通风过程进行建模。结果表明,优化过的回归预测模型能较好地拟合粮食温度与其他变量之间的非线性关系,尤其是当样本数量较为有限时,该方法具有更高的拟合精度,更适合对储粮通风这一强非线性过程的预测研究,对于人工干预操作具有一定的现实指导意义。

RESEARCH ON HORIZENTAL VENTENLATION TEMPERATURE FORECAST BASED ON IN TELLIGENT OPTIMISTIC SVM

Applied improved intelligent prediction technology to transverse-ventilation temperature forecast of grain storage,provides a new approach to grain ventilation prediction and decision.Based on the analysis of main factors,this paper selects the experimental data of qingyuan national grain depot,combines optimization algorithm-grid optimization algorithm,GA algorithm and PSO algorithm with the SVM regression theory,modeling the grain ventilating process.Results show that optimized prediction model better fitting the nonlinear relationship between grain temperature and other variables.Especially when the sample size is limited,this method has a high fitting accuracy,and more suitable on prediction of nonlinear process,which possess great realistic meaning.