Estimation of Advance Time in Furrow Irrigation Using Artificial NeuralNetwork and Principle Component Analysis (PCA) technique

In this study, Neural Network Technique is used to predict advance time using inflow rate, n coefficient, slope, length of furrow, infiltration curve number, initial soil moisture and bulk density by ANN and PCA Technique. Field measurements on furrows of different length and slopes in Mashhad, Dezful, Orumia, Birjand and Karaj having various soil Textures were used in this study. In the Training phase 144 advance time measured data were initially used and then 96 other field measurements were used for cross validation (48) and evaluation (48) phase. The enter parameters determined by using the sensitivity analysis Network and Principle Component Analysis (PCA) technique The obtained results showed Neural Network Technique is well capable of estimating advance Time with high accuracy. The best results (R2= 0.995) obtained from models that used of Principle Component Analysis as enter parameters. The models that used of initial soils moisture content (R2= 0.848) have higher accuracy in comparison to models that used of infiltration curve number (R2= 0.417) and Coefficient of Manning formula (R2= 0.492).