Sie sind auf Seite 1von 11

IRRIGATION AND DRAINAGE

Irrig. and Drain. 58: 6171 (2009)


Published online 20 June 2008 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/ird.392

DESIGN OF MULTIQUALITY IRRIGATION WATER SUPPLY SYSTEMS


USING THE QC FEASIBILITY DOMAIN CONCEPT: II. QCFD MODEL
AND APPLICATIONSy
G. SINAI1*,z AND BEN-ZION DALINS2x
1

Faculty of Civil and Environmental Engineering, Technion, Haifa, Israel


T.L.M. Cons. Engr. Ltd., 3 Hayozma St., Tirat HaCarmel 39120, Israel

ABSTRACT
The world shortage in water sources for irrigation indicates a need to exploit poor water quality sources. Traditional
design methods of irrigation systems consider primarily the hydraulic aspects. Inclusion of water quality aspects
requires new tailored design methods. The method suggested herein is part of a continuous effort taken in that
direction. Earlier works paved the way for the presently suggested design model which computes the feasibility
domain of water discharge and quality of supply networks. Steady-state conditions, conservative water quality
parameters, instantaneous and complete mixing of water from diverse water quality sources were assumed. The
QC (chemical transport sub-problem) was dealt with and the QC feasibility domains of sources for the inner nodes
and outlets were computed for every feasible flow pattern in a given network. Feasibility of supply was evaluated by
comparing demand (water discharge and quality) values with the supply values computed by the QCFD. Real water
supply in a rural community in the Arava desert, Israel, was designed using the QCFD model. Feasibility of various
irrigation schedules was tested and a guideline for operation derived. The method, however, needs further
improvement. Copyright # 2008 John Wiley & Sons, Ltd.
key words: irrigation; irrigation systems; irrigation water supply systems; poor quality water; water quality; feasibility domain
Received 3 August 2005; Revised 15 February 2007; Accepted 18 December 2007

RESUME
La rarete de leau dans le monde rend necessaire lexploitation des eaux de mauvaise qualite. Les methodes de
conception traditionnelles des syste`mes dirrigation conside`rent principalement les aspects hydrauliques.
Linclusion des aspects de qualite de leau exige de nouvelles methodes de conception. La methode suggeree
ici sinscrit dans un effort continu dans cette direction. Les premiers travaux ont prepare le terrain pour le mode`le ici
suggere qui calcule le domaine de faisabilite du debit et de la qualite dans les reseaux. On suppose des conditions de
stabilite, de conservation des parame`tres de qualite, de melange instantane et complet des eaux de qualite
differentes. Le QC (sous-proble`me de transport des sels) a ete traite et les domaines de faisabilite des sources, des
nuds interieurs et des sorties ont ete calcules pour chaque serie de debits dans un reseau donne. La faisabilite de la
fourniture deau a ete evaluee en comparant des valeurs de demande (debit et qualite) aux valeurs de fourniture
calculees par le QCFD. La fourniture reelle pour une communaute rurale dans le desert dArava, Israel, a ete concue
en utilisant ce mode`le de QCFD. La faisabilite de divers programmes de pilotage de lirrigation a ete examinee et
* Correspondence to: G. Sinai, Faculty of Civil and Environmental Engineering, Technion-Israel Institute of Technology, Haifa 32000, Israel.
E-mail: aruti@tx.technion.ac.il
y
Conception des syste`mes dalimentation en eau dirrigation de multi-qualite en utilisant le concept de domaine de faisabilite QC: II. Mode`le et
applications de QCFD.
z
Associate Professor
x
M.Sc. in Agricultural Engineering

Copyright # 2008 John Wiley & Sons, Ltd.

62

G. SINAI AND B.-Z. DALINS

une directive dexploitation a ete elaboree. La methode necessite cependant des ameliorations. Copyright # 2008
John Wiley & Sons, Ltd.
mots cles: irrigation; syste`mes dirrigation; syste`mes dalimentation en eau dirrigation; leau de mauvaise qualite; qualite de leau; domaine
de faisabilite

INTRODUCTION
Increasing demand for irrigation water is a worldwide problem. Use of poor quality water sources is among the
resources which have not yet been fully exploited. Although poor quality (especially saline) water is available in
semiarid climate zones, they are not successfully used for several reasons, among which is the lack of knowledge
required for efficient design and operation of irrigation systems which consider water quality aspects in addition to
traditional hydraulic design.
Attempts have been made by experts to develop methods for planning, design, analysis and operation of water
supply systems with water quality aspects. Several earlier works were conducted on this issue. Liang and Nangi
(1983), and Males et al. (1985) discussed mixing in water supply systems. Sinai et al. (1985) suggested optional
operation of irrigation systems considering water quality issues. Liang and Nangi (1983) suggested control of water
quality by mixing water from different sources. The term multiquality network was suggested by Shah and Sinai
(1985) for water supply systems where water of different qualities is supplied. Mixing and dilution of various water
quality flows inside the supply network has been dealt with in detail by many researchers: Shah and Sinai (1985);
Percia et al. (1997), Ostfeld (2005); Cohen et al. (2000, 2004).
Most of the works assume instantaneous and complete mixing in network junctions where two or more inflows of
different water quality meet. The outflow concentration is therefore a weighted average of the inflow concentration.
A nodal equation can be delivered by imposing water balance and solute flow balance on a dilution/mixing junction
inside the network. Substitution of the dilution equality (Shah and Sinai, 1985), which reads equal concentration in
every outflow of a dilution junction, yields the following, so-called, dilution equation for steady flow conditions:
P
Cip

in in
j Qij Cijp

ex
qi Cip

out
j Qij

8j 2 di

(1)

8i2N

where
Cip
out
Qin
ij ; Qij
in
Cijp
qi
ex
Cip
di
N

concentration of substance p inside junction j and its outflows


inflows and outflows of junction i respectively
concentration of substance p in the inflows of junction i
external flows to junction, qi > 0 source, qi < 0 - consumer
ex
concentration of external flows, if qi is a source, Cip
Csource ,
ex
if qi flows away from i Ci Cip
set of pipes connected to junction i
set of junctions in the network.

QC FEASIBILITY DOMAIN (QCFD)


Problems of solute transport in networks considers water discharge (Q), hydraulic and elevation/pressure heads (H)
and solute concentration (C) as the state variables. The so- called QCH problem can be decomposed into two
coupled subproblems; the hydraulic (QH) and the chemical transport (QC) one. Following Cohen et al. (2000) we
deal here within the QC subproblem assuming conservative water quality substances and steady flow conditions.
The concept of QC feasibility domain (QCFD) was originally suggested by Dalins (1986) for analysis of
multiquality networks. Water discharge (Q) and solute dischargeJ Q  C (where C solute concentration) are
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

QCFD MODEL AND APPLICATIONS

63

drawn in a QJ diagram. Every vector point to a point inside the QCFD represents a supply condition of given Q
and C (note C dJ/dQ) Properties and methods to compute QCFD of any junction in a network were given in an
earlier paper in this series (Sinai et al., 2008). Two methods were developed: (1) vector additions for the case of
fixed concentration C 6 f(Q); and (2) a numerical method for the case where C is a function of Q, i.e. C f(Q).
Application of these methods for dilution junction requires the use of the dilution equation (Equation 1). The
result is a QCFD of the dilution junction.

THE QCFD MODEL


A computer model was developed to calculate the QCFD of every inner node and of the outlets of a given network.
(A) The inputs were as follows: layout of the network, with all the inlets (sources) and outlets (consumers); pipe
diameters; value of concentration (salinity) in the sources; demand data (water discharge and range salinity
level).
(B) Computation of all the FFPs of the network and maximal discharges: the following computations were
carried out off-line: transformation of the network layout to topological matrices (adjacency and incidence
matrices); setting the sources/consumer data as I/O vectors; computation of all feasible flow patterns
(FFPs); computation of maximal discharge value in every pipe based on pipe diameter, pipe material and
design hydraulic gradient.
(C) Running the QCFD model using the QCFD algorithm: for every FFP the following computation was made:

(D)
(E)
(F)
(G)
(H)

Jij Qij Ci 8i 2 S
where Jij salt flow rate in pipe ij, Qij maximal water discharge in pipe ij, Ci concentration (salinity)
in the upstream node of pipe ij and S set of all the sources.
Identify all the mixing nodes in the network for a given FFP. The methods are: (1) visual analysis of the
directed graph of that FFP; and (2) using the node rank method on the adjacency/incidence matrices.
Identifying dilution junction by examining the flow paths connecting the mixing junction to upstream nodes
of different water quality (e.g. salinity).
Setting the dilution equation (Equation 1) to every dilution junction. Equation (1) can now be solved
explicitly for Ci since flow directions are known for a given FFP.
Computing the QCFD of every dilution junction successively, starting with the closest dilution junctions to
the sources.
Continued computation of QCFD up to the outlet exits.

QCFD algorithm
Several methods can be used to compute the QCFD (step G). Two of them were presented in Sinai et al. (2008),
namely: the vector addition method for fixed concentration, and the numerical method for concentrations which
change with the discharge level C f (Q).
The output of this QCFD model were QCFD values for every node in the network and, which is of special
interest, that of the consumption outlets. The above procedure is continuous to every one of the FFPs found earlier.

EVALUATION OF SUPPLY DEMAND


Given the QCFD of the consumer outlets, assessment of supply feasibility was made by inserting the demand for
discharge and concentration range into the QJ diagram of the QCFD of the examined outlet. A criterion could be
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

64

G. SINAI AND B.-Z. DALINS

imposed for evaluation of partial coverage of the demand by the QCFD of that outlet. A decision would then be
made whether the demand could be met in that water supply system under the FFP examined. Subsequently,
the next FFP was examined using the same QCFD algorithm. Finally, all the FFPs were tested and the most
appropriate ones selected. The criteria for selecting the best FFPs are not trivial, since the hydraulic control
requirements depend among others on the actual FFP selected. This issue is, however, beyond the scope of the
present paper.

DEMONSTRATION OF QC DESIGN OF A REAL MULTIQUALITY IRRIGATION WATER


SUPPLY SYSTEM
The Hazeva region in the Central Arava Valley in southern Israel was chosen for demonstration of the QC design.
The climate of this region is hot and dry, similar to many semiarid zones worldwide. A local groundwater aquifer is
the only source of water. It is a phreatic shallow aquifer with diverse values of salinity. Shallow wells (boreholes)
were drilled and low-lift pumps were used to avoid upconing of the groundwater level in that shallow aquifer.
Therefore six water sources were available, all of which were boreholes, equipped with pumps and discharging
their water directly into the water supply system. Three types of consumer existed: five settlements, where one of
them is a school campus and the other four are villages which demand drinking water, two agricultural fields for
field crops, three for vegetables, and three for dates. Table I summarizes data for the six sources including their
discharge and salinity levels. Location of sources, villages, and irrigated fields is shown in Figure 1.
The hot climate of the Arava valley is an advantage for winter crops, in particular winter vegetables, which are
exported to Europe and attract high prices. However, winter crops are quite sensitive to water salinity level. Local
experience shows a range of 250450 mg 1 Cl as suitable for winter vegetables. Unfortunately, the necessary
volume of such high-quality water was not available locally, so poor-water quality in the range of
450900 mg 1 Cl, which was available, was utilized. Field crops, especially cotton and dates, are more tolerant
to water salinity. Therefore the farmers grow these tolerant crops in addition to vegetables. The vegetable crops
were grown in the Hazeva area only, due to the availability of a good water source nearby (Hazeva III, Figure 1). A
general allocation of crop types in the individual fields was planned by a local planner. This plan was based on water
quality availability as well as soil conditions and other agronomic factors. This plan can be seen in Figure 1.
Table II summarizes data of the consumers of every settlement, including demand for drinking water, for
irrigation of field crops, vegetable and date fields. The required salinity ranges for every consumer are also
presented in Table II. The demand for water was computed with the assumption that night to day storage tanks were
erected at key points of the water supply system for regulation of the stochastic demand for drinking water. It was
therefore assumed that the demand discharge represented daily or even weekly time averaged values.
The sources and consumers are illustrated in Figure 1 and the water supply system in Figure 2 on a real scale.

ANALYSIS OF QCFD AT THE SOURCESCONSUMERS LEVEL


The first step in the Q-C design of the water supply system was a sourcesconsumers analysis before imposing the
network constraints. The proposed method of QCFD was used. Figure 3 shows the cumulative supply QCFD of all
Table I. Water sources in the Hazeva region
Source name in Figure 1

Node no. in Figure 2

Discharge (m3 h1)

Salinity (mg 1 Cl)

A
B
C
F
E
D

210
250
280
200
150
200

700
680
1040
500
450
260

Ein Ofarim IV
Ein Ofarim V
Ein Ofarim VI
Hazeva X
Idan III
Hazeva

Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

65

QCFD MODEL AND APPLICATIONS

Figure 1. Water sources, agricultural fields and drinking water consumers in the Hazeva region Israel (on a real scale)

the six sources. Several demand ranges were marked by vertical thick lines in the QJ diagram (Figure 3): (i) all
the drinking water (Q 145 m3 h1, C 150300 mg 1 Cl); (ii) all the vegetable fields (Q 270 m3 h1,
C 250450 mg 1 Cl); (iii) all the field crops (Q 300 m3 h1, C 500900 mg 1 Cl); and (iv) all the date
fields (Q 400 m3 h1, C 600800 mg 1 Cl). The following symbols are used in Figure 3: K drinking water,
V vegetable field, FC field crops, D date fields.
The analysis shown in Figure 3 revealed the following: (i) all the demands for drinking water, vegetables, field
crops, and dates can be met since all these demand ranges were inside the cumulative QCFD of the combined sources.
All these demands can therefore be met if the supply for every consumer group (i.e. drinking, vegetables. . .) was
sequential. However, simultaneous supply was preferable since occasionally the demand for drinking water occurs at
the same time as irrigation. In addition, individual crop types require different irrigation schedules, so several fields
may be irrigated at the same time. Simultaneous supply should therefore be examined as well.
Several combinations of simultaneous supply were tested. They are shown in Table III.
Combinations (V+ K), (V+ FC), (V+ D), (FC+ D), (V+ FC+ D), and (V+ FD+ D+ K) are shown in Figure 3. The
combination V+ K did not meet the supply QCFD in the lower part of salinity range, i.e. if water was supplied
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

66

G. SINAI AND B.-Z. DALINS

Table II. Type of consumers, their node number, water discharge demands, and tolerance range of water salinity
Consumer in Figure 1
Idan
Drinking
Field crops
Field crops
Dates
Ovot
Drinking
Dates
Hazeva
Drinking
Vegetables
Vegetables
Vegetables
Dates
Ein Hazeva
Drinking
School
Drinking

Demand (m3 h1)

Water quality salinity range (mg 1 Cl)

d
FC9
FC10
D7

35
150
150
150

150300
500900
500900
600800

a
D6

15
150

150300
600800

e
V11
V12
V13
D8

40
50
100
120
100

150300
250450
250450
250450
600800

35

150300

20

150300

Node no. in Figure 2

simultaneously to all the vegetable fields and drinking water to the five settlements, there would be insufficient good
quality water to meet all drinking water demands of those five settlements. Similarly, the combination (V+ D) was
also not feasible since the lower range of salinity demand (i.e. of the vegetables) could not be met. The
combinations (V+ FC+ D) and (V+ FC+ D+ K) presented similar problems implying unfeasible combinations. The
combinations (V+ FC) and (FC+ D) were generally feasible with a slight disagreement between supply and demand
QCFD at the extremes of the salinity tolerance range.
The analysis shown in Figure 3 was a limited demonstration, which covered just a part of all possible
combinations of simultaneous supply. We did not show here a subpartition of the different crops to individual fields,
which possibly may reveal more feasible combinations of simultaneous irrigation.

THE MULTI-QUALITY WATER SUPPLY SYSTEM


Based on the previous design step at the sourcesconsumers level, the second step the network layout design
was taken. A design of that multiquality network layout was shown in Figure 2. It employs the concept of a
combined-separate dilution network. The concept of separation is essential for protection of the drinking water
consumers from the saline water sources. Therefore a single subnetwork connects the drinking water source D with
the five settlements. Another, saline, subnetwork connects the saline sources to the agricultural fields. These two
subnetworks were connected at node 21 where dilution takes place. The diluted (medium quality) water
downstream to node 21 is used for irrigating the three vegetable fields V11, V12 and V13 (see Figure 2).
Four other dilution junctions were located: one near field D6 which controls the supply to D6, and another
downstream of D6 which directly controls the quality of irrigation water supply to D7 and indirectly that to D8, FC0
and FC10. At nodes 26 and 30 two other dilution junctions exist which directly control the supply to FC9, FC10 and
D8. A valve preventing backflow was located on the pipeline connecting the two subnetworks upstream of node 21.
It prevents intrusion of saline water into the drinking water subnetwork, an essential device indeed! Our experience
shows that some states (e.g. California) require an air gap device instead of a valve preventing backflow. They act
under the US Uniform Planning Code which requires an air gap separation to protect drinking water from
intrusion of contaminant water by vacuum which may be formed in pipes due to pump activation or draining out of a
pipe section full of water.
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

67

QCFD MODEL AND APPLICATIONS

Figure 2. The water supply network of the Hazeva region Israel (names of sources and consumers are listed in Figure 1 and Tables I and II)

ANALYSIS OF QCFD AT THE NETWORK LEVEL


Analysis at the network level is the following logical design step after the sourcesconsumers analysis. The
designed layout of the supply network shown in Figure 2 was examined. Values of maximal discharge in every
pipe were computed based on the diameter of the pipes and assuming an average hydraulic gradient. Maximum
source discharge (in m3 h1), source water salinity (in mg 1 Cl) and maximum pipe discharge in every pipe
section were the input data for the computer program which calculated the supply QCFD in every node and
consumer outlet of the network. Results are shown in Figure 4 (left column), where QCFD of three nodes 21, 26
and 30 in Figure 2 are drawn. These nodes were chosen since they are the inlets of fields V11, V12, V13, FC9,
FC10 and D8. The demands of the relevant fields were also plotted on the same QJ diagram for evaluation. Some
interesting conclusions could bedrawn: (i) node 21 supplied irrigation water to the three vegetable fields V11, V12
and V13 in Figure 2. The demand of field V11 could be met, but that of fields V12 and V13 could not. These two
fields should be subdivided into two irrigation blocks of 50 m3 h1 each in field V12, and of 60 m3 h1 each in field
V13. Five irrigation blocks were therefore determined for the three vegetable fields, three of 50 m3 h1 and two of
60 m3 h1 discharge. These five irrigation blocks should be irrigated sequentially, one after the other. This was the
only way to meet the demands of these fields by the supply QCFD of node 21. A similar phenomenon was observed
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

68

G. SINAI AND B.-Z. DALINS

Figure 3. The cumulative QCFD of all the sources in the Hazeva region with several demand ranges under sequential and simultaneous water
supply

Table III. Demand values of consumers according to crop type and drinking
Code

Description

K
V
FC
D

All
All
All
All

drinking water
vegetable fields
field crops
date fields

Q(m3 h1)

DC (mg 1 Cl)

145
270
300
400

150300
250450
500900
600800

Combinations of consumers for simultaneous supply


KV
KFC
KD
VFC
VD
FCD
KVFC
KVD
VFCD
DVFCD

Drinkingvegetables
Drinkingfield crops
Drinkingdates
Vegetablesfield crops
Vegetablesdates
Field cropsdates
Drinkingvegetablesfield crops
Drinkingvegetables/dates
Vegetablesfield cropsdates
Drinkingvegetablesfield cropsdrinking

Copyright # 2008 John Wiley & Sons, Ltd.

145 270 415


145 300 445
145 400 545
270 300 570
270 400 670
300 900 700
145 270 300 715
270 400 300 970
270 400 300 970
145 270 400 300 1115

150450
150800
150800
250900
250800
500900
150900
250900
250900
150900

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

QCFD MODEL AND APPLICATIONS

69

Figure 4. QCFDs of nodes 26, 30 and 21 in FFP1 shown in Figure 2 (left column) and the QCFDs of node 26 and 30 in FFP2 of Figure 5
(right column)

at node 26 which supplied the field crops FC9 and FC10 (Figure 2). Just one of these two fields could be irrigated
within the QCFD of node 26. Therefore, sequential irrigation of FC9 and FC10 was essential. The demand of field
D8 can be met by the supply QCFD of node 30. The concluding irrigation schedule is therefore: (1) FC9, D8 and
V11, (2) FC10, V12 (only 50 m3 h1), (3) V12 (the second plot 50 m3 h1), (4) V13 (first 60 m3 h1), (5) V13
(second 60 m3 h1), repeat the cycle 15. This is only a schematic example; specific growth factors, crop types, and
weather conditions obviously affect the actual irrigation schedule.

EFFECT OF FLOW DIRECTIONS ON THE QCFD


Change in flow direction affected the QCFD of the nodes in two aspects: (i) locations of dilution junctions were
changed; and (ii) the shape of the QCFD of individual nodes were also changed.
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

70

G. SINAI AND B.-Z. DALINS

This phenomenon was demonstrated by comparison of two feasible flow patterns FFP1 and FFP2. FFP1 is shown
(not completely) by arrowheads of several pipes in Figure 2. Similarly FFP2 is shown on the network layout in
Figure 5. Flow directions were changed in pipes (27, 29) and (25, 30) resulting in change in the location of dilution
junctions. Node 30 was a dilution junction in Figure 2 (FFP1), and an ordinary junction in Figure 5 (FFP2),
representing constant salinity of a value lower than the demand of D8, hence a waste of good quality water (Figure 4
node 30, FFP2). Nodes 29 and 30 were dilution junctions in FFP1 (Figure 2) but not in FFP2 (Figure 5). Nodes 25
and 27 were ordinary (dilutionless) junctions in FFP1 (Figure 2), but dilution junctions in FFP2 (Figure 5).
These changes were also noted in the relevant QCFDs, and were shown in Figure 4, right column. The shape of
the QCFD of node 26 was changed in terms of Q and range of salinities DC. The QCFD of node 30 which was of a
dilution junction in FFP1, was changed to a single inclined line in FFP2 since node 30 is an ordinary one in FFP2.
The QCFD of node 21 remains unchanged in FFP2 and was therefore not drawn. One can realize that fact just by
observing the change in flow directions between FFP1 and FFP2. The drinking water subnetwork which was being
fed by source D was connected to node 21 by a pipe with a backflow-preventing valve. Therefore, pressure changes
that may occur due to changes in flow direction cannot be propagated upstream from node 21 to the drinking water
subnetwork. Similarly, the other pipe connecting source C with node 21was subjected to one-way flow from source
C to node 21. Therefore, the QCFD of node 21 remains unchanged in all the feasible flow patterns. Now we can

Figure 5. The same network as in Figure 2 but with different FFP2. Note changes in flow directions in pipes (27, 29) and in the locations of the
dilution junction at nodes 25, 29 and 30
Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

QCFD MODEL AND APPLICATIONS

71

understand why the supply from node 21 had to be divided into several irrigation events in a sequence to
accommodate the demand of fields V11, V12, V13.
A more complete design of similar multiquality networks requires, therefore, a consideration of feasible flow
pattern effects. This aspect, however, was beyond the scope of the present paper.

CONCLUSIONS
Design of multiquality water supply systems, for drinking and irrigation consumers with diverse water quality
demands, needs special care. A method for analysis of the solute transport problem of water supply from diverse
water quality sources was successfully decomposed into two subproblems: the hydraulic one (QH) and the
chemical transport one (QC). We were able to apply the QC feasibility domain (QCFD) concept to a real water
supply system for drinking and irrigation. A computer model was developed to calculate the QCFD of any node/
junction in the network, and especially of the demand outlets. Given demand data (water discharge and water
quality concentration), feasibility of supply can be assessed by comparing the demand to outlet QCFD on the same
QJ diagram. This test was first conducted at the sourcesconsumers level, ignoring network constraints. In the
second step, network layout was assumed and all feasible flow patterns (FFPs) were found. The QCFD model from
above was run for every FFP, computing the QCFD of the outlets so that feasibility of supply could be evaluated for
that FFP. Finally the performance of the design network was compared for selected operational FFPs. This subject
is, however, beyond the scope of the present paper.
All these design steps were demonstrated in a detailed example of a rural community in the desert climate in
Israel. The agro-climatic conditions of the example farm were, however, similar to many semiarid lands worldwide.
However, the design method should have been improved to include cost aspects as well as loss of yield due to
irrigation with poor quality water.

ACKNOWLEDGEMENTS
This research was partially supported in its early stage by Grant no 1-no2-81, Bard, USIsrael Binational
Agricultural Research and Development. Thanks are expressed to Mrs Olga Kaplan for editing and to Mr Arieh
Aines for graphics.

REFERENCES
Cohen D, Shamir U, Sinai G. 2000. Optimal operation of multi-quality water supply systems: I Introduction and the Q-C Model. Engineering
Optimization 32: 549558.
Cohen D, Shamir U, Sinai G. 2004. Water quality aspects of optimal operation of rural water distribution systems for supply of irrigation and
drinking water. Irrigation and Drainage 53: 339361.
Dalins B-Z. 1986. Synthesis of the network graph in multiquality water systems. MSc thesis, Faculty of Agricultural Engineering, The
Technion-IIT, Haifa, Israel; 140 pp. (in Hebrew).
Liang T, Nangi S. 1983. Managing water quality by mixing water from different sources. Journal of Water Resources and Planning Management,
ASCE 109(1): 4857.
Males RM, Clark RM, Weahrman RJ, Gates WE. 1985. Algorithm for mixing problems in water systems. Journal of Hydrologic Engineering
ASCE 111(2): 206219.
Ostfeld A. 2005. Optimal design and operation of multiquality networks under unsteady conditions. Journal of Water Resources and Planning
Management, ASCE 131(2): 19.
Percia C, Oron G, Mehrez A. 1997. Optimal operation of regional systems with diverse water quality sources. Journal of Water Resources and
Planning Management, ASCE 123(2): 105115.
Shah M, Sinai G. 1985. Salinity control in multiquality irrigation networks, Kibbutz Hammadia case study. Agricultural Water Management
Journal 10: 2331.
Sinai G, Koch E, Farbman M. 1985. Dilution of brackish waters in irrigation networks an analytical approach. Irrigation Science 6: 179190.
Sinai G, Dalins B-Z, Cohen D, Shamir U. 2008. Design of irrigation water supply systems using the Q-C feasibility domain concept: I
Introduction and theory. Irrigation and Drainage (in press). DOI: 10.1002/ird.391

Copyright # 2008 John Wiley & Sons, Ltd.

Irrig. and Drain. 58: 6171 (2009)


DOI: 10.1002/ird

Das könnte Ihnen auch gefallen