شبيه سازي هوشمند تغييرت الگوي توزيع آب از آبپاش منفرد در اثر باد

Intelligent Simulation of Wind Distorted Water Distribution Pattern of Single Sprinkler

A new approach based on Artificial Neural Networks (ANNs) is presented to simulate the distortion of single sprinkler water distribution pattern influenced by wind under center pivot irrigation systems. Field experiments were conducted at the Research Center of Agricultural Faculty, University of Tabriz-Iran. The distribution pattern of Nelson R3000 Rotator® sprinkler was tested at various wind velocities and directions. Experimental dataset were divided into three subsets named as training, validation, and testing dataset. Affecting parameters were defined and various networks with different structures were trained and validated. A Multi Layer Perceptron (MLP) neural network with backpropagation training algorithm having two hidden layers, 20 neurons in the first hidden layer, and 6 neurons in the second hidden layer and tangent-sigmoid transfer function was selected as the optimal network and produced 0.929 and RMSE=6.69 for test dataset. Using the structure of the optimal network, a model named ISSP (Intelligent Simulator of Sprinkler Pattern) was developed as a MATLAB® m-file. Entering wind velocity and direction as the inputs, this model could simulate the distribution pattern of R3000 Rotator® sprinkler and have ability to display graphical and Excel® file of results. Simulated distribution patterns of this model for three different wind speeds were compared with observed distribution patterns. Obtained results showed that there is no meaningful difference between R values for these three conditions and these values are in a very appropriate range (^>0.935).