基于BP神经网络的大棚作物腾发量预测模型.pdfVIP

2O 中国农村水利水电·2008年第2期 文章编号：1007—2284(2008)02—0020—02 基于BP神经网络的大棚作物腾发量预测模型 温耀华 ，罗金耀 ，李小平 ，孙 俊 ，高金花 ，程国银。，赵秀江 (1．武汉大学水资源与水电工程科学国家重点实验室，武汉430072；2．长春工程学院，长春130012； 3．湖北省水利厅，武汉430070；4．湖北省水利厅节水灌溉试验示范基地，湖北鄂州436013) 摘 要：在分析影响大棚作物腾发量的气象因子的基础上，以气象因子为输入向量，以大棚作物腾发量为输出向量， 构建了气象资料不足情况下计算大棚作物腾发量的BP神经网络模型。利用鄂州节水灌溉实验基地逐日气象资料对所 建模型进行反复训练和预测，并把预测结果与用传统的Penman Monteith公式计算而得的同期作物ET值相比较。BP 神经网络模型的预测值与公式计算的ET值的相关系数为0．986。研究结果表明：构建的模型计算精度较高，方法简便 可行，能满足实际生产需要。 关键词：BP神经网络；腾发量；大棚 中图分类号：$152．7 文献标识码：A Crop Evapotranspiration of Greenhouse Based on BP Neural—Network Models Yao-hua ，LUO Jin-yao ，LI Xiao-ping ，SUI Jun ，GAO Jin-hua2，CHENG Guo-yin。，Z姒 o Xil卜ji~ng (1．State Key Laboratory of Water Resources and Hydropower Engineering Science，Wuhari University，Wuhan 430072。China； 2．Changchun Institute of Technology，Changchun 130012，China； 3．The Water Resources Department of Hubei Province，Wuhan 430070，China； 4．The WSI Experiment and Demonstration Base of Water Resources Department of Hubei Province，Ezhou 436013，China) Abstract：Based on the analysis of meteorological factors affecting the evapotranspiration(ET)of plastic greenhouse crops，BP neural —network models for evapotranspiration of plastic greenhouse crops were established in this paper，which can be used for limited me— teorological data．The input vectors of the model are various combinations of climatic factors，while the output was the daily value of E By using the daily climatic data during March 23 to March 29 of 2007 in Ezhou Water-Saving Irrigation Experiment Station，the models were calibrated and trained．The calculated values of the models were compared with the ET value computed with the conven— tional Penman Monteith equation．The correlation coefficient of two methods was 0．986．It clearly