IOP Conference Series: Materials
Science and Engineering
PAPER • OPEN ACCESS You may also like
- Infrastructure performance of irrigation
Characterization of Water Dynamics and Modelling canal to irrigation efficiency of irrigation
area of Candi Limo in Mojokerto District
of an Open Channel Irrigation System S Kisnanto, R R R Hadiani and C Ikhsan
- Constriction canal assisted artificial lateral
line system for enhanced hydrodynamic
To cite this article: P Narendra Ilaya Pallavan et al 2020 IOP Conf. Ser.: Mater. Sci. Eng. 955 012104 pressure sensing
Zhiqiang Ma, Yonggang Jiang, Peng Wu
et al.
- A model of the lateral line of fish for vortex
sensing
View the article online for updates and enhancements. Zheng Ren and Kamran Mohseni
This content was downloaded from IP address 139.135.192.224 on 12/10/2024 at 12:06
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
Characterization of Water Dynamics and Modelling of an
Open Channel Irrigation System
P Narendra Ilaya Pallavan1, V S Abhirame2 and V Suriya Prabha3
1
Assistant Professor, Department of Electronics and Instrumentation Engineering,
Bannari Amman Institute of Technology, Sathyamangalam, India.
2&3
Department of Electronics and Instrumentation Engineering, Bannari Amman
Institute of Technology, Sathyamangalam, India.
Corresponding author’s email address: [email protected]
Abstract. Management of open channel canals is very essential to avoid overflows and
overcome water demand. Automatic control of delivery canals has been gaining significance in
recent years with the purpose of improving the efficiency of water management. This paper
presents the development of single reach irrigation canal with the objective of controlling the
water level. Upstream gate opening of an irrigation canal reach is operated based upon a
constant-level reservoir height. The canal reach between two gates is divided into N nodes, and
the finite-difference forms of the continuity and the momentum equations (saint venant
equations) are written for each node. The Taylor series expansion is applied to linearize the
equations around the equilibrium conditions. The state space model of the canal system is
developed and the same will further be used to design an efficient controller for maintaining
the level of water in the canal.
1. Introduction
Incompetent usage of water networks is the major reason for the water scarcity faced by several
developing countries. In South Asian countries, the infrastructure is on a constant decay whilst water
resources are going dry besides the vast and complex water network system present. Reliability of
water supply can be achieved through improved operation of partial and full automation in the main
and branch canals [1].
Open channel flow is governed by steady and unsteady flow equations. Unsteady flow saint-venant
equations are nonlinear hyperbolic partial differential equations which can be solved only by
numerical methods. Published literature on unsteady flow modelling is numerous [3,4].
Mohan Reddy [10] linearized the saint-venant equation and local optimal control technique is
applied for a single reach canal. Malaterre and Baume [11] addresses about all type of model that
include: saint venant linearization, nonlinear features, strong unknown perturbations, and interactions
among subsystems, neural network based models, fuzzy logic based models. Litrico and Fromion [6]
developed frequency response approach for saint-venant equations. It is linearized around any
stationary regime. Weyer [9] was performed the Model identification of the Haughton channel reaches
in Australia. Ooi [7] was applied System identification in saint-venant equations. P.J.Van overloop [8]
was used system identification method in open channel system. Klaudia Horvath, Eduard Galvis,
Manuel Gomez, Jose Rodellar [5] linearized the saint venant equations and transformed in Laplace
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
domain. In this paper, linearization is explained elaborately by using Taylor series expansion [12,13,
14].
The objective of this study is to model the unsteady flow for a single reach system with rectangular
cross section in an irrigation canal. The controller is designed to achieve the target depth at the
downstream end of the canal so as to satisfy the downstream water demand.
2. Mathematical Modelling
An open channel is a conduit in which a liquid flows with a free surface which is actually an interface
between the moving fluid and an overlying fluid medium. The schematic diagram of an open channel
with a single reach is shown in Figure 1.
Figure 1: Definition sketch for a canal reach
The unsteady flow Saint-Venant equations are given by
𝜕𝑦 1 𝜕𝑄
+ + =0 (1)
𝜕𝑥 𝑇 𝜕𝑥
𝜕𝑄 𝜕 𝑄2 𝜕𝑦
+ ( ) + gA ( − 𝑆0 + 𝑆𝑓 ) = 0 (2)
𝜕𝑥 𝜕𝑥 𝐴 𝜕𝑥
In which y = depth of flow in the channel, m; Q = flow rate through the channel, m3/s; T = flow
surface width, m; g = gravitational acceleration, m/s2; S0 = channel bottom slope; x = longitudinal
direction of channel, m; A = flow cross section, m2; t = time, s; and Sf = friction slope, which is
expressed as
𝑄|𝑄|𝑛2
𝑆𝑓 = 2 4/3 (3)
𝐴 𝑅
Where n = Manning's roughness coefficient; and R = hydraulic radius, m. It is given by
𝐴
𝑅= (4)
P
Where P = wetted perimeter in m. Equation 1 and 2 are quasi linear, partial differential equations of
the hyperbolic type and defy a closed-form solution.
In this study the hyperbolic equations are discretized using finite difference schemes and are
linearized using Taylor’s series approximations by neglecting higher order derivatives. The
equilibrium condition is assumed to be the steady flow values and is governed by steady state
gradually varied flow equation.
2
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
𝜕𝑦 𝑆0 − 𝑆f
= 𝛼𝑄2 B
(5)
𝜕𝑥 1−
𝑔𝐴3
In which x = distance along the channel; S 0=longitudinal slope of the channel bottom; S f = slope of
the energy line; B = top water surface width; g = acceleration due to gravity. A single reach canal is
divided into ‘N’ nodes in each reach shown in Figure 2. It is divided into N-1 sub reaches of length Δx
(distance between the nodes). The continuity and momentum equations are written for each of these N
nodes.
Figure 2: Depth-time grid for finite difference scheme
2.1. Upstream Node:
Forward difference scheme is applied for the upstream node, Continuity equation:
𝑑𝑦 −1 𝑞
𝑑𝑡
=
Δx 𝑇1,m
(𝑄2 − 𝑄𝑔 ) − Δx1,m
𝑇
(7)
1,m
Where,
𝑄𝑔 = 𝐶𝑑 𝑈𝑚 b √2g √𝑦N,m−1 − 𝑦1,m (8)
Where, Cd = gate coefficient; b=gate width, m; Um = opening at gate, m; g = gravity acceleration,
m/s2; yN,m-1 = depth of previous pool; Δx = length of segment.
2.2. Interior Node:
Central difference scheme for the interior nodes, Continuity equation:
𝑑𝑦 −1 𝑞j,m
= (𝑄j+1,m − 𝑄𝑗−1,𝑚 ) − 2Δx 𝑇 (9)
𝑑𝑡 2Δx 𝑇j,m j,m
3
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
Momentum equation:
𝑑𝑄 −1 𝑄 2 j,m
= (𝑄j+1,m − 𝑄𝑗−1,𝑚 ) + 2Δx 𝐴2 + (𝐴j+1,m − 𝐴𝑗−1,𝑚 ) −
𝑑𝑡 Δx 𝐴j,m j,m
gAj,m gAj,m 𝑄|𝑄|
2Δx
(𝑦j+1,m − 𝑦𝑗−1,𝑚 ) − gAj,m S0 + 𝐾 2 j,m
(10)
2.3. Downstream Node:
Backward difference scheme for the downstream nodes Continuity equation:
𝑑𝑦 −1 𝑞
𝑑𝑡
=
Δx 𝑇N,m
(𝑄g − 𝑄𝑁,𝑚 ) − Δx N,m
𝑇
(11)
N,m
Where,
𝑄𝑔 = 𝐶𝑑 𝑈𝑚+1 b √2g √𝑦N,m − 𝑦1,m+1 (12)
In which Cd= discharge coefficient of the downstream gate in reach; b = width of downstream gate
in reach, m; um+1 =height of downstream gate opening in reach, m. For the last pool, y1,m+1 is replaced
with constant reservoir height of downstream end. Ordinary differential equations is still nonlinear in
both y, Q, u.
The state variables chosen are the depth of flow and the flow rate. The gates are the controls and
the gate discharge equations in conjunction with continuity equations serve as boundary conditions in
upstream node and downstream nodes. These equations are,
𝑄𝑁,𝑚−1 = 𝑄1,𝑚 +𝑄𝑜𝑢𝑡,𝑚−1 (13)
𝑄𝑁,𝑚−1 = 𝐶𝑑 b𝑈𝑚 √2g √Δh (14)
The depths at each node and the gate openings at the equilibrium conditions is obtained from the
steady flow equations. In the linearization, we have linearize in the equilibrium states. The equilibrium
e e e
variables are 𝑦
𝑗,𝑚 ,𝑄 𝑗,𝑚 ,𝑢 𝑗,𝑚 . Actual variables are 𝑦𝑗,𝑚 , 𝑄𝑗,𝑚 , 𝑢𝑚 . Difference
between the equilibrium variables and actual variables are deviation
variables 𝛿𝑦𝑗,𝑚, 𝛿Q 𝑗,𝑚 , 𝛿𝑢𝑚 .
𝑦𝑗,𝑚 = 𝛿𝑦𝑗,𝑚 + 𝑦 e 𝑗,𝑚 (15)
𝑄𝑗,𝑚 = 𝛿Q 𝑗,𝑚 +𝑄e𝑗,𝑚 (16)
𝑢𝑚 = 𝛿𝑢𝑚 + 𝑢e𝑗,𝑚 (17)
The linearization results in linear matrix state space equation.
𝛿𝑥 (𝑡) = 𝐴𝛿𝑥 (𝑡) + B𝛿𝑢(𝑡) (18)
4
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
The objective is to maintain constant discharge at the gate. If there is a disturbance in the system,
system should be bring back to the original equilibrium condition. In order to achieve this, a feedback
controller is designed. For that problem, suitable controller is used to maintain the depth and constant
discharge.
3. Results and Discussions
The cross sectional details and the parameters of the hypothetical canal system used in the model is
given below Table 1.
Table 1: Cross sectional parameters of the hypothetical canal
Length of canal reach 6250 m
Number of nodes 5
Number of sub reaches used 4
Distance between the node 1250 m
Discharge required at end of canal 47 m3/s
Upstream reservoir height 3.24 m
Downstream reservoir height 1.25 m
Target depth at downstream end 1.99 m
Gate width 18.25 m
Gate discharge coefficient 0.9
With the simulated model, the depth of water along the length of the canal system can be estimated
for the input flow conditions. It should be noted that the depth of water along the length of the canal
will not be uniform. Rather the depth of water gradually increase along the length due to the slope of
the canal. The open loop response of the depth of the canal system for different nodes is given in
Figure 3.
Figure 3: Variation of depth of water in different nodes of the canal system
The above results may be helpful to efficiently control the level of water in the canal system by
manipulating the control gates placed at the upstream side of the canal.
5
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
4. Conclusion
In order to assure the availability of water for the final users and in order to prevent the canal
overflows during the flood, water discharge, water depth level and stored water volume of the canal
systems are the important control variables. Controlling the water level contributes additional
advantages like, preventing canal overflows and increase in the stability of the system. Hence, the
control of water level in the extraction zone is considered as the control objective.
The modelling of the irrigation canal is carried out by dividing the canal into reaches, and then
characterizing the water dynamics at each reach separately. The boundary conditions between reaches
also included. Steady flow condition is modelled by applying newton method. Open channel dynamics
for unsteady flow are described using non- linear partial derivatives equations (saint-venant
equations). The modelling of the channel using saint venant equations is developed and linearized for
unsteady flow condition. The depth is obtained for steady flow and unsteady flow conditions of the
canal. Controller will be designed to achieve water level control by manipulating the position of the
control gates.
References
[1] Aberle J and Smart G M 2003 Journal of Hydraulic Research. 41 pp 259–269
[2] Anderson W and Meneveau C 2011 Journal of Fluid Mechanics. 679 pp 288–314.
[3] Barros J, Schultz M and Flack K 2015 Bulletin of the American Physical Society. 60.
[4] Henderson F M 1966 Open channel flow MacMillan Publishing Co., Inc., New York.
[5] Klaudia Horvath, Eduard Galvis, Manuel Gomez Valentin and Jose Rodellar 2015 Control
engineering practice. 41 pp 13-25.
[6] Litrico and Fromion V 2004 Journal of Hydraulic Engineering. 130 pp 806–815.
[7] Ooi S K, Krutzen M and Weyer E 2005 Control Engineering Practice. 13 pp 461–471.
[8] VanOverloop P J, Schuurmans J, Brouwer R and Burt C M 2005 Journal of Irrigation and
Drainage Engineering. 131 pp190–196.
[9] Weyer 2001 Control Engineering Practice. 9 pp 1289–1299.
[10] Mohan Reddy J 1990 Journal of Irrigation and Drainage Engineering. 116 pp 616-631.
[11] Bathurst J C 2002 Journal of Hydrology. 269 pp 11–26.
[12] Bendat J and Piersol A 2010 Random data: Analysis and measurement procedures, Hoboken,
NJ: Wiley.
[13] Cameron S M, Nikora V I and Stewart M T 2017 Journal of Fluid Mechanics. 814 pp 416–429.
[14] Chan L, MacDonald M, Chung D, Hutchins N and Ooi A 2015 Journal of Fluid Mechanics. 771
pp 743–777.
Appendix
The following symbols are used in this paper:
A = flow cross section, m2;
A = continuous-time feedback matrix;
B = continuous-time control matrix;
bi = width of upstream gate in reach;
bi+1 = width of downstream gate in reach;
C = continuous-time disturbance matrix;
Cd = discharge coefficient of upstream gate in reach;
Cdi+1 = discharge coefficient of downstream gate in reach i+1;
D = hydraulic depth, m;
g = acceleration gravity, m/s2;
6
FIC-SISTEEM-2020 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 955 (2020) 012104 doi:10.1088/1757-899X/955/1/012104
N = number of nodes in each reach of canal;
N = manning coefficient;
P = wetted perimeter;
Q = flow rate through channel, m3/s;
Qg = discharge rate through gate, m3/s;
q = turnout demand per unit length of channel, m2/s;
R=hydraulic radius;
Sf = friction slope;
S0 = channel bottom slope;
T = flow surface width;
ui = height of upstream gate opening in reach i, m;
ui+1=height of downstream gate opening in reach i+1,m;
x = longituidinal direction of channel, m;
Δx = length of subreaches;
Y =depth of flow in channel;
δQ =deviation in flow rate at node,m3/s