TURNELL, D. J.; http://lattes.cnpq.br/9509471389644717; TURNELL, David John.
Resumo:
Around 20% of the world's cultivated area depend upon irrigation. One type of irrigation, popular
in the semi-arid regions, including those in the Northeast of Brazil, is furrow irrigation. In spite of its
popularity, this type of irrigation is characterized by low water use efficiency when controlled
manually. The automation of furrow irrigation typically is limited to the controlling of valves and
gates, at predetermined times, to direct water between irrigated sets. To increase the efficiency of
furrow irrigation and make it viable in semi-arid regions, this thesis research investigates alternatives
in hardware and software for the automated management of water use. A development approach is
proposed for distributed automated control based upon a set of control strategies based upon the
detection, in real time, of the presence of water at multiple points in an irrigated area.
Three control strategies were proposed that aim to optimize the use of water in furrow irrigation.
These stratagies are based upon the use of simulation to support the control algorithm. On of these
strategies, 'Advance Analysis', consists of the optimization of the water inlet based upon a set of water
infiltration coefficients. These coefficients are identified in real-time, during the first phase of the
irrigation, using water advance data. The technique utilized in the identification of the coefficients e
adopted from Agricultural Engineering.
The optimization strategy 'Runoff Control' aims at minimizing the runoff of water at the ends of
the furrows. The identification of the infiltration coefficients arises from the advance/recession
characteristics of the water.
For all of the proposed strategies, the control algorithm implements the duty cycle variation of the
water inlet supported by hardware for remote sensing and distributed control and by software for
irrigation management support. The work analyzes the influence upon the control of errors in sensing
of field conditions and input parameters for the irrigation model adopted.
From the hardware point of view, based upon the need for low-cost, low-power telemetry method,
for the transmission of data over short distances, three sensor networks were developed to detect the
presence of water in the furrow and to allow the water inlet control of the irrigation system. Two of the
networks are based upon resistive sensing elements and one network is based upon capacitive sensing
elements. Also developed were a field station and the means of communication between these
elements and the control structure (sensor networks and central computer). The development of
hardware took into consideration the powering available in the field, the noise conditions imposed
upon the communication, the robustness necessary to cope with the hostile field conditions and the
need to offer cost-effectiveness to motivate the adoption of irrigation automation.
From the software point of view, a system called Farm Web was developed based upon the object
action model adopted for distributed systems, in order to offer support to the development of
integrated agricultural automation applications. Farm Web is a system that offers: applications for the
real-time control of furrow irrigation, with features for the simulation and simultaneous visualization,
in graphical form, of the progress of the irrigation process; mechanisms for communication between
applications; multiprocessing support, and a set of application support services that aid their
development. This system is offered as 1) a platform to study control alternatives based upon
simulation of the agricultural processes to be controlled, 2) a base for the development of control
systems that can be used in real applications, and 3) a research environment for use with agricultural
process simulators. Investigated in this research were aspects of hardware and software integration
which allow distributed real-time control with the objective of creating an agricultural automation
system with broad capabilities.
This thesis research resulted in the development of both hardware and software, that due to its
characteristics, can be applied to the development of distributed applications in other contexts of
agricultural automation. The proposed solutions are compatible with the concept of integrated
precision agriculture, in which diverse processes backed by a large number of field stations, embody
an integrated agricultural automation environment.