CAJ | 학술논문

农田灌溉量预测对水资源合理规划和优化配置具有重要意义。为此，以农田水分平衡公式为基础，导出灌溉量I与前各时期的降雨量pt-1、pt-2、pt-3灌溉面积S之间存在非线性关系，由此建立了农田灌溉量的BP神经网络预测模型。检验结果表明，灌溉量BP 神经网络预测法精度约为94．5％。因而，以灌溉面积、前各时期降雨量为输入变量，结合BP神经网络技术来预测农田灌溉量，具有很好的应用前景。
농전관개량예측대수자원합리규화화우화배치구유중요의의。위차，이농전수분평형공식위기출，도출관개량I여전각시기적강우량pt-1、pt-2、pt-3관개면적S지간존재비선성관계，유차건립료농전관개량적BP신경망락예측모형。검험결과표명，관개량BP 신경망락예측법정도약위94．5％。인이，이관개면적、전각시기강우량위수입변량，결합BP신경망락기술래예측농전관개량，구유흔호적응용전경。
Prediction of farmland irrigation is of great significance to rational planning and optimal allocation of water re -sources .The article based on the farmland water balance equation ,deduced that nonlinear relationship between irrigation amount I and pt-1 、pt -2 、pt-3 that rainfall irrigation amounts of each period、irrigation area S;which established a BP net-work prediction model irrigation amount nerve .Test results show irrigation amount BP neural network prediction accuracy is about 94 .5%.Thus we can see that in the irrigated area , rainfall before each period as input variables, combined with BP neural network techniques to predict the amount of irrigation, has good prospects.