Estimation of reference evapotranspiration in greenhouse by Artificial Neural Network

Nowadays Artificial Neural Networks (ANNs) are being applied in several problems of water engineering where there is no clear relationship between effective parameters on the estimation of phenomenon. This research was used to measure aerodynamic data inside and outside greenhouse for estimating reference evapotranspiration in greenhouse by using ANNs. ANN was used with perceptron multilayer structure and Back Propagation with one hidden layer for estimating evapotranspiration by using meteorological parameters. Results showed, with regard to Root Mean Square Error (RMSE), ANNs wasable to estimate reference evapotranspiration with low error. Inside greenhouse, ANN showed a best estimation maximum temperature (Tmax), minimum temperature (Tmin), extraterrestrial radiation (Ra), actual vapor pressure (ea) and sunshine (n) in entrance layer and found as the best model for estimating inside greenhouse reference evapotranspiration with RMSE equal to 1.1 mm day^-1. Outside greenhouse, ANN was found as best model which can use maximum temperature (Tmax), minimum temperature (Tmin), and sunshine (n) in entrance layer estimating inside greenhouse reference evapotranspiration with RMSE equal to 1.01 mm day^-1.