Abstract:
Irrigation has been the backbone of civilized society since time immemorial. The. urge to meet one's survival needs civilization to learn and embrace different agricultural practices, hence irrigational methods and techniques. With population increasing at an exponential rate and land areas being curved short to provide lodging for the enormous population, several new and innovative practices have become inevitable for prolonged nourishment for the human race. Water which happens to be the most central resource for survival is becoming scarce these days and we will be in jeopardy if adequate measures are not taken right away,
The focal point of this project seeks to solve the above mentioned challenges at hand and frying to figure out some better techniques in this technologically prolific era however, the design crop is tea. Although many and innovative techniques have been employed towards Automatic irrigation, they mostly involve simple moisture monitoring control logic for their operations which come with wastage of water resources if not monitored properly.
An Evapotranspiration controller based on adaptive neural fizzy inference structure was adopted
in this project because of it being a powerful tool in prediction of complex phenomena. The ANFIS model was first tested with regression to check. its validity against Penman Evapotranspiration model and proved to be outstanding with MSE of 0.037 and Coefficient of determination of 0.960 while regression gave MSE of 0.96 and Coefficient of determination of 0.886.
The proposed algorithm for irrigation scheduling prohibits water stress because it ensures that moisture depletion does not reach 100% that represents permanent wilting point and irrigation will always occur when depletion reaches 500/0 from which it will deliver water up to 60% of the available water from where it begins applying real time crop water requirement and stops after delivering 95% of field capacity. All this was to ensure that there is negligible or no loss through depletion beyond the root zone.
The model was prototyped and simulated in Simulink Matlab 2013 environment.
