Estimation of the Evapotranspiration using ANFIS algorithm for Agricultural Production in Greenhouse

Abstract

In this work it is showed fuzzy model, trained with ANFIS algorithm, to estimate water evapotranspiration (ET) of a tomato plant, as a witness in the production in a greenhouse. There are two variables used, the VPD that represents the microclimate of the greenhouse and the mass of the plant, that was measured with a digital balance (lysimeter). Additionally, a second lysimeter was used to prove the result of the structure of the fuzzy model, by means of mass balance principle, it was proved the effectiveness of the model that it is proposed. The experimental results here presented are from two days, the pot with the tomato plant was saturated with water and no irrigation cycles were performed. The approximate model error was 20.3239 (g) of a total 453.3239 (g), that represents approximately 4.48 %. The experimental results indicate that the modeling proposal shown here, reduces the number of variables that it is actually suggest by the organization of the united nations for the agriculture and food (FAO), ET fuzzy model proposed can success in the economy for its implementation in protected agriculture.