Systems Engineering for Irrigation Systems: Successes and Challenges

Abstract

In Australia gravity fed irrigation systems are critical infrastructure essential to agricultural production and export. By supplementing these large scale civil engineering systems with an appropriate information infrastructure, sensors, actuators and a communication network it is feasible to use systems engineering ideas to improve the exploitation of the irrigation system. This paper reports how classical ideas from system identification and control can be used to automate irrigation systems to deliver a near on-demand water supply with vastly improved overall distribution efficiency.

A modelling methodology for natural dam-river network systems

“…Such a dam-river system is very complex and at present no simple mathematical model is being applied to the system operation. A model established in line with the proposed approach is very similar to the grey-box model in Weyer (2001Weyer ( , 2003aWeyer ( , 2003b, Ooi and Weyer (2001), Ooi et al (2003Ooi et al ( , 2005, Li et al (2005), Mareels et al (2005), and Maxwell and Warnick (2006), but is much simpler because it includes water level measurements at some points and control variables, while in the references the flow rates are included in the models. As is known, the flow rates are difficult to measure, especially for natural channels (the velocity distribution in a cross-section of the channel is very complex).…”

“…The parameters in data-driven models are identified from real time data. Such models include black-box models in Elfawal-Mansour, Georges, and Ohnishi (2000), greybox models in Weyer (2001Weyer ( , 2003aWeyer ( , 2003b, Ooi and Weyer (2001), Ooi, Weyer, and Campi (2003), Ooi, Krutzen, and Weyer (2005), Li, Cantoni, and Weyer (2005), Mareels et al (2005) and Maxwell and Warnick (2006), high order transfer function models in Sawadogo, Faye, Malaterre, and Mora-Camino (1998) and Sawadogo, Faye, Benhammou, and Akouz (2000) and neural network models in Gallinari (1996, 1997). Neural network techniques are also employed in Paravan, Stokelj, and Golob (2004) to forecast theshort-term water inflow in a hydroelectrical power plant.…”

Decomposition of a fixed-profile load scheduling method for large-scale irrigation channels

“…The idea is to represent the interconnection between two pools as a function of the already scheduled offtakes in the downstream pools. For example, from (7), the interconnection v i−1 is expressed as a function of y i , which is a function of d i (see (6)), 8 then from (6), y i−1 can be written as a function of d i by direct substitution of variables. In this way, y 1 , · · · , y i are written as a function of d i .…”

“…there is no pumping). The flow of water through the network is regulated by automated gates positioned along the channels [3], [8], [10]. The stretch of a channel between two gates is commonly called a pool.…”

A distributed accelerated gradient algorithm for distributed model predictive control of a hydro power valley

“…Large-scale system control has been an active research area that has resulted in a variety of control techniques, which can be classified in three main categories: decentralized control, distributed control, and centralized control. The application of these approaches can be found in a rich literature on control of water canals for irrigation and hydro systems [16,14]. We are interested in applying model predictive control (MPC), a control method that has been successfully used in industry [25], thanks to its capability of handling hard constraints and the simple way of incorporating an economical objective by means of an optimization problem.…”