Acta Sci. Pol.

Formatio Circumiectus 18 (1) 2019, 27–37

ENVIRONMENTAL PROCESSES

www.formatiocircumiectus.actapol.net/pl/ ISSN 1644-0765 DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

O R I G I N A L PA P E R Accepted: 19.11.2018

DISTRIBUTION OF FLOWS IN A CHANNEL NETWORK UNDER STEADY

FLOW CONDITIONS
Dariusz Gąsiorowski, Wojciech Artichowicz
Faculty of Civil and Environmental Engineering, Gdańsk University of Technology, Gabriela Narutowicza 11/12, 80-233 Gdańsk

ABSTRACT

Aim of the study

The article presents an algorithm for calculating the distribution of flow in a junction of open channel network
under steady flow conditions.

Material and methods

The presented algorithm is based on the continuity equation and a simplified energy equation. To describe
the relationship between the depth of water and the flow rate, the Manning’s equation was used to express
discharge in the main channel, whereas in case of other channels, with hydraulic structures, appropriate
equations describing the flow over a weir and through the culverts were used. Substitution of the abovemen-
tioned relationships into the continuity equation leads to a non-linear algebraic equation with respect to the
water level. The resulting equation can be solved with iterative numerical methods. Calculation example
using the proposed algorithm was carried out for the hydraulic system located on the Strzyża stream in
Gdańsk.

Results and conclusions

The presented approach can be an alternative to analytical-graphic method and it does not require formulation
of the boundary problem for the system of ordinary differential equations describing the gradually varied flow
in the open channel network.

Keywords: division of flows, open channel network, steady gradually varied flow, nonlinear equation, flow
through weir and culvert

INTRODUCTION other two channels – J and K, respectively – with the

known flow rate of QI.
When a steady gradually varied flow in a network of The above problem is a typical example of the
open channels is considered, often it is necessary to steady gradually varied flow in a channel network. In
estimate the distribution of flow in its branches. Usu- the classical approach, the problem of this kind can be
ally the flow rate Q in one of the channels is known solved graphically and analytically based on the flow
(or the water level h corresponding to this flow rate), curves of individual channels. A detailed description
whereas the percentage shares of flows in the remain- of this approach is presented by Chow (1959) and
ing channels are unknown. Assuming the connection French (1985). In order to obtain the solution, one can
of the three channels presented in Figure 1a, it is nec- also formulate the boundary problem for the system of
essary to determine the flow rates QJ and QK in the ordinary differential equations (Szymkiewicz, 2000).


e-mail: [email protected]

© Copyright by Wydawnictwo Uniwersytetu Rolniczego w Krakowie, Kraków 2019

 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

tions, one at the beginning, and one at the end of the

channel L (see: Fig. 2):

h(x = 0) = h0 and h(x = L) = hL (4a,b)

where h0 and hL are the water levels in cross sections x

= 0 and x = L, respectively.

Fig. 1. Three connected channels: general scheme (a), discreti- The system of differential equations (1–2) written
sation scheme (b) for a single channel can be solved, for example, by
applying the implicit trapezoidal rule (Szymkiewicz,
Such approach is valid if the flow in the channels is 2010; Artichowicz and Gąsiorowski, 2018). As a re-
not known. Equations of this type are derived from de sult of the discretization of equations (1–2) and addi-
Saint Venant system of equations, with the assumption tionally noticing that Qi = Qi+1 = QI = constant, we
of steady flow conditions. In the case of a single chan- obtain the following algebraic equation:
nel, these equations take the following form:
αi + 1 ⋅ QI2 αi ⋅ QI2 ∆x
hi + 1 +
2 g ⋅ Ai2+ 1
= hi +
2 g ⋅ Ai2
−
2
( )
Si + Si + 1 (5)
d  α ⋅ Q2 
h + =−S (1)
dx  2 g ⋅ A2 
where:
dQ i – cross-section index in the channel, i = 1, ....,
=0 (2) N-1,
dx
N – number of cross-sections,
where: I – channel index,
Q – flow rate, Δx – distance between the sections.
x – spatial variable, Similar equations can be derived for the remain-
h (x) – water level above the assumed datum, h = ing channels J and K. As a result, a system of (N-1
z + H (see: Fig. 2), + M-1 + P-1) algebraic equations with (N + 1 + M
H (x) – water depth, + 1 + P + 1) unknowns is obtained. There are (N +
z (x) – bottom elevation above the assumed da- M + P) unknown values of water levels in particu-
tum, lar cross-sections of the channels, and one unknown
Downloaded from mostwiedzy.pl

A (x) – wetted cross-sectional area, value of flow rate per channel (Szymkiewicz, 2010).
S – friction slope, To solve the abovementioned system of equations it
α – de Saint-Venant coefficient. is necessary to impose additional conditions at the
channel connections, resulting from the conservation
The friction slope can be estimated using the trans- of mass principle:
formed Manning equation:
QI = QJ + QK (6)
n2 ⋅ Q 2
S = 2 4/ 3 (3) and from the energy conservation principle:
A ⋅R
α ⋅ QI2 α ⋅ QJ2 α ⋅ QK2
where R is the hydraulic radius, and n is Manning’s hi + = h j + = hk + (7)
roughness coefficient. 2 g ⋅ Ai2 2 g ⋅ A2j 2 g ⋅ Ak2

The system of differential equations (1–2) de- where:

scribes the changes of the water level h(x) along the I, J, K – channel designations (see: Fig. 1a),
channel at a constant flow rate Q. That system of equa- i, j, k – designation of nodes at the channels’ con-
tions should be supplemented by two boundary condi- nection, i = N, j = 1, k = 1 (see: Fig. 1b).

28 www.formatiocircumiectus.actapol.net/pl/
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

Fig. 2. Calculation scheme of longitudinal profile for the system of ordinary differential equations describing steady gradu-
ally varied flow

Algebraic equations (5) written for individual chan- Besides the application of an appropriate method to
nels constitute a global system of nonlinear equations. solve such a system of equations, the convergence of
This system, supplemented with the known boundary the iterative process itself may become an issue. An-
conditions (4a, b) and with the compliance conditions other problem complicating the solution is the variable
(6,7), can be solved using the Picard method (Szym- number of roots of the resulting algebraic equations
kiewicz, 2010). As a result, we obtain the flow rates (5) (Artichowicz and Szymkiewicz, 2014). Therefore,
in individual channels QI, QJ, QK and the water level in order to avoid the aforementioned problem, alterna-
profile – h (x) – along the channels. tive methods can be proposed which result from appro-
In the above-mentioned formulation of flow distri- priate simplifications. If the scope of interest is only
bution issues, the boundary conditions for obtaining the information regarding the percentage distribution
the correct solution must be imposed. In other words, of flows, this issue can be significantly simplified by
the water level values at the end of channel hi = 1, hj = N, omitting information related to the water level profile
hk = N are required (see: Fig. 1b). When these values are along the channels. Then, the solution of the flow dis-
unknown, this requirement may constitute a significant tribution problem can be based only on the conditions
limitation. Most often, the end section is an ordinary (6–7). In this approach, the mass balance equation (6)
Downloaded from mostwiedzy.pl

channel cross-section, and then the boundary condition is adopted in an unchanged form, while in the energy
can be determined on the basis of the relationship be- equation (7) for small differences in velocity, only the
tween the flow rate and the water elevation in the form of compliance of water levels can be assumed:
the flow curve Q = Q (h) or in the form of the Manning
hi = N = hj = 1 = hk = 1 = h (8)
equation. On the other hand, if the water in the channels
flows through hydraulic structures, then the known for- Additionally, using appropriate relationships be-
mulas describing the relationship between the flow rate tween the water level elevation h, and the flow rate
and the hydraulic head of a given structure can be used. in individual channels Q = Q (h) (for instance, the
Manning equation, or the flow formula through weir
FLOW DISTRIBUTION CALCULATED ON THE BASIS or culvert), the mass balance equation for the open
OF SOLVING THE NONLINEAR MASS BALANCE channel network node can be ultimately rewritten in
EQUATION the following form:
QI (h) – QJ (h) – QK (h) = 0 (9)
Due to the approximation of the system of ordinary
differential equations, a system of non-linear algebra- In the presence of hydraulic structures, it is as-
ic equations arises (Eq. (5)) which has to be solved. sumed that these structures are located at a sufficient-

www.formatiocircumiectus.actapol.net/pl/ 29
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

ly close distance from the network junction, so that Fig. 3a). The flow in the diversion channel depends on
the water level in the junction simultaneously corre- its current water level. Flow can take place through the
sponds to the level determining the hydraulic head of circular culverts with flow rate Qk only or additionally
the structure. The mass balance equation obtained in through the trapezoidal weir with flow rate Qt. Taking
this way is a nonlinear equation with one unknown, into account these hydraulic schemes, the non-linear
that is, the water level h. This equation can be solved mass balance equation with respect to the depth H writ-
by any method used for solving algebraic nonlinear ten for the analysed junction takes the following form:
equations, for instance Newton’s, Picard’s, false po-
sition, or simple fixed point iteration method. Due to Qg – Qp (H) – (Qk (H) + Qt (H)) = 0
this approach, the flow distribution problem is reduced or
to solving only one algebraic nonlinear equation in- (10a, b)
stead of a system of equations. Qg – Qp (H) – Qo (H) = 0

APPLICATION EXAMPLE In the considered task, it was assumed that in the

main channel a subcritical flow occurs, i.e. the depth
Flow distribution calculations were performed for the H is greater than the critical depth (H > Hkr). In addi-
river network junction near the Srebrniki reservoir lo- tion, because the capacity of the diversion channel, as
cated on the Strzyża stream in Gdańsk (see: Fig. 3). well as the channel transporting water to the reservoir
The junction consists of a section of the main channel is smaller than in the main channel at normal depth
with a trapezoidal cross section (see: Fig. 4a), con- Hn, the water in the cross-section of the computation-
nected to two other channels. One of the two channels al node will increase to the depth of H (H> Hn). As
transports water to the reservoir through a rectangular a result, on the analysed section in the main channel,
culvert (see: Fig. 4b), while the other one is the diver- a backwater profile is created. This situation is sche-
sion channel of this reservoir, including two circular matically illustrated in Figure 3b.
culverts over which the trapezoidal weir is located (see: In the main inflow channel and the diversion chan-
Fig. 4c). Therefore, the flow in the main channel Qg is nel (see: Fig. 4a, c), the relation between the flow rate
distributed to flow through the culvert towards the res- and water depth has been determined using the Man-
ervoir Qp and the flow in the diversion channel Qo (see: ning equation for the trapezoidal cross-section:
Downloaded from mostwiedzy.pl

Fig. 3. “Srebrniki” hydraulic node: plan view (a), longitudinal section (b).

30 www.formatiocircumiectus.actapol.net/pl/
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

and for supercritical flow (s > skr):

( )
2 5/ 3
1 b⋅H + m⋅H
Qg = s1/ 2 (11)
( ) 2 g (H − H kr)
2/ 3
n Q = µ N ⋅ Akr (14)
b + 2 H 1 + m2

where:
where: s – culvert bed slope,
b – width of the channel at the bottom level, skr – critical slope,
s – channel bed slope, µN – discharge coefficient (non-submerged in-
m – bank slope parameter. let),
ΔH – difference of water levels at the inlet and
In the case of the hydraulic structures, the appro-
at the outlet of the culvert, ∆H = H – Hd,
priate formula was adopted for a given culvert or weir
Hkr – critical depth in the culvert,
depending on the current water level in the channel.
Ad, Akr – cross-sectional area at the depth of Hd,
The water level in turn determines the corresponding
Hkr, respectively
hydraulic condition, in which the structures operate.
HL – height of hydraulic losses along the
The culvert may operate with a free-surface flow or
length.
a pressurized flow. The first type of flow occurs when
the following conditions for non-submerged inlet and Additionally, the flow with a free surface, sub-
outlet are fulfilled (Bodhaine, 1976; French, 1985): merged inlet and with a non-submerged outlet (see:
H/D < 1.5, Hd < D (12a, b) Fig. 5c) should be taken into consideration. For this
case, the following formula can be used:
where:
H – water depth above the inlet bottom, Q = µ z ⋅ A0 ⋅ 2 g ⋅ H (15)
D – height or diameter of the culvert,
Hd – water depth above the bottom of the outlet.
where:
In addition, for the flow with a free surface, we need μz – discharge coefficient (submerged inlet),
to distinguish between the conditions of subcritical A0 – cross-sectional area at the culvert inlet cor-
flow and those of supercritical flow (see: Fig. 5a, b). responding to the area completely filled with
In this case, the formulas for calculating the discharge water.
take the following forms, respectively – for subcritical
Downloaded from mostwiedzy.pl

In order to determine the flow rate in the culvert

flow (s < skr) (French, 1985; Szpakowski, 2015):
operating in pressurized flow conditions (see: Fig. 5d),
the formula (13) can be used, with the only difference
Q = µ N ⋅ Ad ⋅ 2 g (∆H − H L) (13) being that the cross-sectional area Ad will correspond

Fig. 4. Cross sections of: the main channel (a), rectangular culvert (b), trapezoidal weir and circular culverts in the diversion
flow channel (c)

www.formatiocircumiectus.actapol.net/pl/ 31
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

to the area of the culvert cross-section completely where:

filled with water, i.e. Ad = A0. Hp – hydraulic head of the weir, Hp =H – Hg,
The values of the discharge coefficients μ depend Hg – height of the weir,
mainly on the geometry of the culvert, the type of inlet B – width of the weir at the threshold level,
construction and the depth of water H before the inlet µPT – discharge coefficient,
(French, 1985; Kubrak and Kubrak, 2004). α – angle of inclination of the weir edge to the
The trapezoidal weir, located above the circu- vertical.
lar culverts, can operate as a non-submerged or sub-
merged weir structure depending on the hydraulic con- The weir is considered to be submerged when the
ditions (see: Fig. 6). In non-submerged conditions, the following conditions are fulfilled (see: Fig. 6a):
discharge of the weir can be determined on the basis of Ho > Hg and DH/Hg < 0.7 (17a, b)
the following formula (Sawicki, 2009):
where:
2 Ho – water depth after the weir,
Q=
15
µ PT ( )
2 g 5B ⋅ H 3p / 2 + 4tg α ⋅ H 5p / 2 (16) ΔH – difference of water levels before and after
the weir, ∆H = H – Ho.
Downloaded from mostwiedzy.pl

Fig. 5. Hydraulic schemes of the culvert operation during the flow: with free surface, non-submerged inlet (subcritical flow)
(a), with free surface, non-submerged inlet (supercritical flow) (b), with free surface, submerged inlet (c), pressurized flow (d)

Fig. 6. Hydraulic schemes of the weir (a), cross-section of the trapezoidal weir (b)

32 www.formatiocircumiectus.actapol.net/pl/
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

Then the discharge of the weir can be estimated reservoir through the rectangular culvert Qp(Hk). Us-
according to the following expression: ing the mass balance equation (10b), the flow rate in
the diversion channel Qo = Qg – Qp is also computed.
Qz = s ∙ Q (18)
The character of the flow in the diversion channel near
where Q is the discharge determined for a non-sub- the circular culverts and the trapezoidal weir depends
merged weir, while σ is a submergence factor of the on the hydraulic conditions before and after the given
weir, the value of which can be computed according structure. In the case of a culvert, this may be flow with
to the following formula (Kubrak and Kubrak, 2004): a free surface or the pressurized flow, whereas the weir
may work either as non-submerged or submerged.

σ = 1.05 1 + 0.2
(
Ho − H g ) ∆H 1/ 3 (19)
Therefore, in order to identify the proper oper-
ation of the hydraulic system, it is also necessary to
 Hg   H p  determine depth Hok in the diversion channel after the
 
structures using the Manning equation. Based on
the knowledge of the depth approximations H k and Hok
COMPUTATIONAL ALGORITHM in the given iteration k, it is possible to determine – by
adopting the appropriate calculation scheme – the val-
The presented formulas make it possible to estimate the ues of flow through the circular culverts Qk (H k, Hok)
discharges of individual structures in the mass balance and through the trapezoidal weir Qt(H k, Hok). The iter-
equation (10a). The non-linear equation obtained in this ative process of approximating the depth H is carried
way, was solved with respect to water depth H using out until the mass balance equation (10a) is fulfilled
the false position method. While solving the equation, with the given required accuracy, in other words, the
first the initial approximation of the normal depth H0 calculations are completed when the following con-
in the main channel has to be determined for the given vergence criterion is fulfilled:
flow rate Qg (see: Fig. 7). This depth is determined on
the basis of the Manning equation, which is equivalent H (k + 1) − H (k ) ≤ ε (20)
to solving the non-linear equation (11). The determined
value of the depth Hk = H0 (where k denotes the iter- where:
ation index) is at the same time the hydraulic head of ε – required accuracy, the value of ε = 0.0001 m
the culverts that allows estimating the discharge to the was adopted in the study.
Downloaded from mostwiedzy.pl

Fig. 7. Computational algorithm for flow distribution in the analysed hydraulic node

www.formatiocircumiectus.actapol.net/pl/ 33
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

EXAMPLE OF COMPUTATIONS tions. The assumed values of geometrical and hydrau-

lic parameters of individual structures are presented
In the analysed hydraulic system, the channels have in Table 1.
a trapezoidal shape in the cross-section, with a bank The flow distribution calculations were performed
slope of m = 1.5 and a bottom slope of i = 0.5%. In for the assumed flow rate in the main channel Qg.
turn, the widths at the bottom of the main channel The highest value of the flow rate corresponded to
and the diversion channel are assumed to be equal to the maximum flow with the probability of exceed-
bg = 3.5 m and bo = 1.0 m, respectively (see: Fig. 4a, ance p = 10%, that is, Qg = Q10% = 6 m3 ∙ s–1. The cal-
c). In these channels, a constant Manning coefficient culations were carried out for the existing system as
value n = 0.02 m–1/3 s was assumed. Due to the signif- well as for a hypothetical system without the circular
icant bottom slope in the rectangular culvert (s = 3%), culverts and trapezoidal weir in the diversion chan-
it was assumed that this culvert would operate only in nel. The results of the flow distribution calculations
the supercritical flow. In such case the possibility of are presented in the form of the characteristics show-
submerging the inlet does not occur and the discharge ing the relation between the flow rate and the depth
of the culvert Qp can be determined by means of the Q = f (H) (see: Fig. 8a, 9a) as well as the percentage
formula (14). In the case of circular culverts and trap- share of individual flows with regard to the flow Qg
ezoidal weir, it was assumed that they could operate in the main channel (see: Fig. 8b, 9b). This relative
in the submerged and non-submerged flow condi- share was estimated according to the following rela-
tions. The discharge Qk in this case is determined by tionship:
the equations (13–15), whereas the weir discharge Qt
is determined by the equations (16) and (18). In or- Qi
∆Q = 100% (21)
der to simplify the calculations, the hydraulic losses Qg
(HL = 0) in circular culverts were omitted. Also fixed
values of discharge coefficients were assumed for cul- where Qi denotes flow through the appropriate channel
vert and weir operating in the given hydraulic condi- or structure (Qo, Qp, Qk or Qt).

Table 1. Assumed values of the parameters of structures operating in the analysed hydraulic conditions

Geometrical Discharge
Device Hydraulic scheme Criterion
Downloaded from mostwiedzy.pl

parameters coefficients

Non-submerged inlet and

H/D < 1.5, Hd < D, s < skr 0.85
outlet, subcritical flow

Flow with a free Non-submerged inlet and

H/D < 1.5, Hd < D, s > skr 0.90
Circular surface outlet, supercritical flow
D = 0.4 m
culvert Submerged inlet,
H/D > 1.5, Hd < D 0.50
non-submerged outlet

Flow under Submerged inlet and

H/D > 1.0, Hd > D 0.60
pressure outlet

Rectangular Flow with a free Non-submerged inlet and D = 1.3 m,

H/D < 1.5, Hd < D, s > skr 0.90
culvert surface outlet, supercritical flow B = 2.0 m

Non-submerged – B = 3.0 m 0.60

Trapezoidal
– tgα = 1.5
weir
Submerged Ho > Hg, ∆H/Hg < 0.7 Hg = 0.7 m Formula (19)

34 www.formatiocircumiectus.actapol.net/pl/
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

When analysing the graph in Figure 8b, it can the position of the water level above the threshold of
be observed that in the existing system for various the trapezoidal weir (see: Fig. 3b and Fig. 6a), which
depths H, there are different distributions of the flow. in turn causes the inclusion of that trapezoidal weir in
This relationship has a nonlinear character and re- the operation with discharge Qt. Therefore, the flow
sults from the fact that in individual channels the in the diversion channel takes place both through the
structures can operate in different hydraulic con- circular culverts and through the weir, thus Qo = Qk +
ditions. For the depth range of 0.1–0.5 m, a nearly Qt. In addition, for the appropriate value of the depth
constant trend in flow distribution is obtained (see: Ho after the culverts, inlets of the culverts may be-
Fig. 8b), where the flow to the reservoir through the come submerged, which forces the pressurized flow.
culvert Qp is about 70% of the flow in the main chan- A further increase in the depth of Ho also causes sub-
nel Qg, whereas the flow to the diversion channel Qo merging of the weir.
is 30% of that flow. In this depth range, the flow in The situation is different in the flow distribution
the diversion channel takes place only through cir- when the lack of culverts and weir in the diversion
cular culverts, i.e. Qo = Qk. For depths above 0.6 m, channel is assumed (see: Fig. 9). In this case, the in-
the condition for non-submerged inlet (Eq. (12a)) is crease in value of depth causes a gradual increase in
not fulfilled, which causes a change in the hydraulic the percentage share (from 40% to 65%) of the flow
conditions of the culverts’ operation, and thus affects Qo in the diversion channel, while at the same time
the value of the flow rate. Above the depth of 0.7 m, the flow Qp to the reservoir is reduced from 60% to al-
there is a gradual increase in the flow Qo in the diver- most 35%. Significantly lower values of inflow to the
sion channel, which is accompanied with the percent reservoir, in comparison to the existing system shown
decreasing discharge Qp to the reservoir. This results in Figure 8, are obviously due to an absence of back-
from the fact that the depths H > 0.7 m correspond to water effect in the diversion channel.
Downloaded from mostwiedzy.pl

Fig. 8. Results of flow distribution computations for the existing system: characteristics of Q = f(H) for individual channels and
structures (a), percentage share of the flow Qg (b)

www.formatiocircumiectus.actapol.net/pl/ 35
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

Fig. 9. Results of flow distribution computations without the hydraulic structures in diversion channel: characteristics of Q = f(H)
for individual channels (a), percentage share of the flow Qg (b)

CONCLUSIONS Sample computations using the proposed algo-

rithm were carried out for the hydraulic system of the
The article presents a computational algorithm for de- network of channels located on the Strzyża stream
termining the flow distribution in the junction of open in Gdańsk. Flow distribution calculations were per-
channels network under steady gradually varied flow formed for different values of the determined flow rate
conditions. The algorithm is based on the adequate in the main channel – for the existing system, as well
conditions at the junction of the channels, resulting as for the hypothetical system without the hydraulic
from the continuity equation and a simplified energy structures in the diversion channel.
equation, in which only the conformity of water levels
has been taken into consideration. Additionally, us- REFERENCES
Downloaded from mostwiedzy.pl

ing the relationships between the water level and the

flow rate in individual channels including hydraulic Artichowicz, W., Gąsiorowski, D. (2018). Numerical Anal-
structures, the simplified mass balance equation was ysis of Steady Gradually Varied Flow in Open Channel
derived. For this reason, the obtained mass balance Networks with Hydraulic Structures. In: M. Kalinows-
ka, M. Mrokowska, P. Rowiński, Free Surface Flows
equation is a non-linear one with regard to the wa-
and Transport Processes, Springer International Publish-
ter depth. In the present work, the false point method
ing, 127–142.
was used in order to solve this equation. Consequent-
Artichowicz, W., Szymkiewicz, R. (2014). Computational
ly, due to this approach the flow distribution problem issues of solving the 1D steady gradually varied flow
was reduced to solving only one algebraic non-linear equation. Journal of Hydrology and Hydromechanics,
equation instead of a system of equations. Therefore, 62(3), 226–233.
the presented flow distribution algorithm can offer Bodhaine, G. L. (1976), Measurement of Peak Discharge
an alternative to the analytical-graphic method or a At Culverts by Indirect Methods. Techniques of Water
method based on the solution of the boundary prob- Resources Investigation of the United States Geological
lem for the system of ordinary differential equations Survey. U.S. Geological Survey, Washington.
describing the steady gradually varied flow in the Chow, V. T. (1959). Open channel hydraulics. McGraw-Hill
open channel network. Book Company.

36 www.formatiocircumiectus.actapol.net/pl/
 Gąsiorowski, D., Artichowicz, W. (2019). Distribution of flows in a channel network under steady flow conditions. Acta Sci. Pol., For-
matio Circumiectus, 18(1), 27–37. DOI: http://dx.doi.org/10.15576/ASP.FC/2019.18.1.27

French, R. H. (1985). Open –channel hydraulics. New York: wlotu. Acta Sci. Pol. Formatio Circumiectus, 14 (2)
MacGraw-Hill. 2015, 195–207.
Kubrak, E., Kubrak, J. (2004). Hydraulika techniczna – Szymkiewicz, R. (2000). Matematyczne modelowanie prze-
przykłady obliczeń. Warszawa: Wydawnictwo SGGW. pływów w rzekach i kanałach. Warszawa: PWN.
Sawicki, J.M. (2009). Mechanika przepływów. Gdańsk: Szymkiewicz., R., (2010). Numerical modeling in open
Wydawnictwo Politechniki Gdańskiej. channel hydraulics. Water Science and Technology Li-
Szpakowski, W. (2015). Przepustowość hydrauliczna brary, New York: Springer.
przepustu kołowego w warunkach niezatopionego

ROZDZIAŁ PRZEPŁYWÓW W SIECI KANAŁÓW W WARUNKACH RUCHU USTALONEGO

ABSTRAKT

Cel pracy
W artykule przedstawiono uproszczony algorytm obliczeniowy służący do oszacowania rozdziału przepły-
wów w sieci kanałów w warunkach przepływu ustalonego.

Materiał i metody
Algorytm obliczeniowy opiera się na warunkach zgodności w połączeniu kanałów wynikających z równania
zachowania masy oraz uproszczonego równania energii, w którym przyjęto tylko zgodność poziomów wody.
Wykorzystując dodatkowo związki pomiędzy rzędną zwierciadła wody, a natężeniem przepływu w poszcze-
gólnych kanałach oraz urządzeniach występujących w tych kanałach wyprowadzono odpowiednie równanie.
Otrzymane w ten sposób równanie bilansu masy jest równaniem nieliniowym względem tylko głębokości
wody. W pracy do rozwiązania tego równania wykorzystano metodę siecznych. Przykładowe obliczenia
z wykorzystaniem zaproponowanego algorytmu przeprowadzono dla układu hydraulicznego znajdującego
się na potoku Strzyża w Gdańsku.

Wyniki i wnioski
Zaproponowany algorytm rozdziału przepływów może stanowić alternatywę do metody analityczno-graficz-
nej lub metody opartej na rozwiązaniu zagadnienia brzegowego dla układu równań różniczkowych zwyczaj-
nych opisujących przepływ ustalony niejednostajny w sieci kanałów.
Downloaded from mostwiedzy.pl

Słowa kluczowe: rozdział przepływów, sieć kanałów otwartych, ruch ustalony niejednostajny, przepływ
przez przelew oraz przepust, równanie nieliniowe

www.formatiocircumiectus.actapol.net/pl/ 37