摘要
以PC(聚碳酸酯,简称PC工程塑料)板温室为研究对象,通过相关系数显著性检验,确定了影响温室内温度的主要因子。应用系统辨识技术,建立了温室温度的ARX模型(扩展的自回归模型)。采用残差分析法确定模型的阶,使用递推最小二乘法估计模型参数,并根据另一组试验观测资料对模型进行检验。结果表明,温室温度模拟值与实测值的变化趋势一致,模型能有效地模拟温室内温度,其模拟值与实测值的决定系数R2=0.84,均方根误差(RMSE)=0.96℃,相对误差(RE)=2.51%。
A PC （polycarbonate） greenhouse was used as research object, and the main factors affecting temperature in the greenhouse was determined through the significance test of correlation coefficient. By application of system identifica- tion technology, a greenhouse temperature ARX model （extension of autoregressive model） was established. The residual analysis method was used in model order identification, the recursive least squares method was applied in model parameter estimation, and the model was verified by other observation data. The results showed that the ARX model could simulate the temperature inside the greenhouse reliably, the coefficient of determination（R2） between the simulated and the measured being 0. 84,the root mean squares error （RMES） being 0. 96 ~C, and the relative error （RE） being 2.51%.
出处
《江苏农业学报》
CSCD
北大核心
2013年第1期46-50,共5页
Jiangsu Journal of Agricultural Sciences
基金
安徽省科技攻关项目(1201a0301008)
关键词
温室
温度
系统辨识
ARX模型
greenhouse
temperature
system identification
ARX model
