Modeling and control in open-channel irrigation systems: A review

Highlights

• In OCIS, the direct use of the SVE for control systems design is often impractical.

• Diverse control-oriented models for OCIS have been identified.

• Several approaches that solve the control problem in OCIS have been reported.

Abstract

Water is the most important element of food production, and the easiest and most cost-efficient way to transport it is through open-channel irrigation systems (OCIS). These types of systems have a high agricultural and ecological impact. However, in most countries, OCIS lack automation and efficiency at mitigating the economic and environmental costs that the waste of water is causing. In order to identify modeling strategies and the best control practices, this paper presents an overview of the main factors of control-oriented models and control strategies for OCIS. In modeling, two fields are considered: (i) models that come from simplifications of the Saint-Venant Equations (SVE); and (ii) approximate models. For each category, a brief description of the control-oriented modeling strategies is given. In the control field, five relevant aspects are considered: (i) centralized, decentralized, and distributed architectures; (ii) control objectives; (iii) regulation structures and control-action variables; (iv) feedback and feedforward configurations; and (v) control strategies. For each aspect, the most important features are explained. Finally, with the aim of establishing the acceptability of the reported modeling and control techniques, as well as challenges that remain open, a discussion and a case study are presented.

Introduction

Through irrigation, it is possible to compensate the amount of water that crops need in dry seasons and to extend the productive land away from natural water sources. The easiest and most economical way to transport water in agriculture is through open-channels. Currently, water is taken from rivers and transported by using an intricate network of channels to each user. These networks are called open-channel irrigation systems (OCIS). Nearly 70% of the water consumed in the world is used for irrigation (OECD, 2018), and most of the water is transported through open-channels. Moreover, the world population grows continuously. In 1980, the world population was around 4.4 billion. Now, there are about 7.4 billion people and in 2060 the population will likely increase to 10.2 billion (United Nations Department of Economic and Social Affairs Population Division, 2017). Consequently, in 40 years food production must increase by 40%.

On the other hand, the irrigation process has a high environmental impact since the water taken from a river reduces its flow, affecting life in the river and the surrounding ecosystem. Therefore, as it is highlighted by Lamnabhi-Lagarrigue et al. (2017), it is necessary to develop new approaches to increase food production by increasing the efficiency of the OCIS, where “efficiency is seen as the ratio of the volume of water delivered to the users and the volume of water extracted from the source” (Mareels et al., 2005).

The OCIS are complex systems. In most countries, their operation is in charge of user associations, which maintain the system in operational conditions, manage the economic resources, and calculate, assign, and supply the appropriate amount of water to the users. The water assignment process can be performed in multiple modes such as: (i) rotational mode, where the central administration develops the supply polices and allocates the amount of water and time duration of the flow delivered to each user; (ii) on-request mode, where the user must request in advance the amount of hydraulic resource that will be used; and (iii) on-demand mode, where the user is free to take water from the system when it is needed. According to the assignment process and the hydraulic characteristics of the system, the central administration must calculate the water levels and flows throughout the systems, which are regulated by gates and weirs, and their positions are calculated with the aim of assigning a specific amount of water to each user. Most of the OCIS operate in rotational and in on-request modes in absence of automatic control systems. Therefore, each regulation structure is manually adjusted by operators, who must carry out this task throughout many kilometers of channels and hundreds of regulation structures. In the normal operation of the OCIS it is common to find disturbances such as flow variation at the source, channel obstructions, leaks, overflows, and demand changes. These types of disturbances lead to water spillages that affect the OCIS efficiency (Litrico & Fromion, 2006a).

In order to promote the implementation of automatic control in OCIS, in the last three decades, multiple works that review the advances in modeling and control of OCIS have been reported. For instance, Malaterre (1995) presents an exhaustive characterization of regulation methods for OCIS, showing the need to unify definitions and concepts in a field where there is a convergence of civil, hydraulic, and control engineers. Schuurmans (1997) shows basic principles for understanding the control problem in OCIS, explaining the finite-difference model and proposing the integrator delay model to adjust the real dynamic behavior of OCIS in a simple way. Moreover, in the control area, Schuurmans (1997) presents the implementation of traditional controllers such as the linear quadratic regulator (LQR) and the linear quadratic Gaussian regulator (LQG). Malaterre et al. (1998) review and classify the implemented controllers according to the variables (controlled, measured, control-action), the logic of control (type and direction), and design technique. Furthermore, Malaterre and Baume (1998) explore several modeling techniques and control strategies. Cantoni et al. (2007) and Mareels et al. (2005) discuss some aspects such as infrastructure automation, control objectives, and system identification. Weyer (2008) shows alternatives in centralized and decentralized control. Moreover, Malaterre (2008) reviews the main concepts and strategies in the control of OCIS. Over the last decade, the task committee on recent advances in canal automation provides a practical guide on OCIS automation (Wahlin & Zimbelman, 2014). This guide covers topics about supervisory control and data acquisition, as well as fundamentals in the design and implementation of control strategies. Finally, Ding et al. (2018) provide a review focused on applications of model predictive control in agriculture, where it can be highlighted that the control of OCIS is the area that shows more development of this kind of control strategy.

According to the presented information, the OCIS control problem is an issue of interest, which has been continuously studied and summarized in several works. On the other hand, in control systems, usually, the selection of an accurate control-oriented model of the system is an important stage that must be addressed before selecting, designing, and implementing a control strategy. However, it has been identified that in the reported reviews, the control-oriented modeling topic has not been broadly addressed. Moreover, control of OCIS is a relevant and challenging field, where there is a continuous generation of contributions. Therefore, there is a need to: (i) review recent modeling and control techniques that have been reported; (ii) establish the acceptability of existent techniques; and (iii) report challenges that remain open for future research.

In this way, the motivation of this paper is to provide a detailed review¹ of modeling and control of OCIS towards providing useful information for researchers interested in contributing to the OCIS modeling and control area. This review focuses on the two main aspects in control of OCIS: control-oriented models and control design, which are firstly presented, subsequently discussed, and finally illustrated through a case study.

The remainder of the paper is organized as follows. Section 2 starts by presenting the proposed notation and a description of the OCIS. In Section 3, a classification of the control-oriented models for OCIS is given. In Section 4, the multiple approaches that in control of OCIS can be established are presented and classified. In Section 5, it is given a discussion around the reported modeling and control approaches for OCIS. In Section 6, it is presented a case study that explains the development of the most common control-oriented modeling strategy, and the most common control strategies for OCIS. Finally, in Section 7, some conclusions are drawn.