基于模糊神经网络的参考作物蒸腾量预测研究

Research on Reference Crop Evapotranspiration Prediction Method Based on Fuzzy Neural Network

为准确估算灌区作物需水量,建立了基于模糊神经网络的参考作物腾发量时间序列预测模型。采用宝鸡地区1954—2004年逐月气象资料,利用主成分分析法提取影响参考作物蒸发蒸腾量的主要影响因子,获得4个综合变量作为输入向量,用彭曼-蒙蒂斯公式计算的参考作物蒸发蒸腾量作为目标向量。运用matlab进行编程,应用模糊神经网络模型预测参考作物腾发量。结果表明:12组测试集样本的平均相对误差绝对值为5%,最大相对误差为11.4%,最小相对误差为0.4%;模糊神经网络模型与用彭曼-蒙蒂斯公式计算值有很高的一致性。

In order to estimate crop water requirement of irrigation area accurately, a forecasting model had been built on reference crop evapotrans- piration time series based on fuzzy neural nelwork. Meanwhile, monthly meteorological data came from Baoji region during the period of 1954 to 2004 and then used principal component analysis method to extract the main impact factor of ETo. Finally, four integrated variable acquired as the input vector and the estimation of ETo using PM formula set as the target vector. Fuzzy Neural Network Model of MATLAB was used to predict refer- ence crop evapotranspiration. The results show that the absolute value of the average relative error of the 12 test samples is 5%, with the maximum relative error of 11.4% and the smallest relative error of 0.4%. It is evidenced that the prediction of fuzzy neural network model has very high uni- formity with that of PM formula.