See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/272489063

The Design and Analysis of Labyrinth Weirs

Conference Paper · January 2011

CITATIONS READS
2 839

2 authors:

Brian Crookston B. P. Tullis

Utah State University Utah State University
82 PUBLICATIONS   288 CITATIONS    99 PUBLICATIONS   582 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Size Scale Effects of Nonlinear Weir Hydraulics View project

Dam Safety View project

All content following this page was uploaded by B. P. Tullis on 18 April 2019.

The user has requested enhancement of the downloaded file.

 Hydraulic Design and Analysis of Labyrinth
Weirs. I: Discharge Relationships
B. M. Crookston, A.M.ASCE 1; and B. P. Tullis, M.ASCE 2

Abstract: A method is presented for the hydraulic design and analysis of labyrinth weirs based upon the experimental results of physical
Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.

modeling. Discharge coefficient data for labyrinth weirs with quarter-round and half-round crest shapes are presented for sidewall angles
ranging from 6 to 35°. Cycle efficiency is also introduced as a design aid, which compares the hydraulic performance of different cycle
geometries. Geometric parameters that affect flow performance are discussed. The predictive accuracy of the design method is evaluated
through comparisons to previously published labyrinth weir head-discharge data. The companion paper examines nappe behaviors that affect
flow performance and presents hydraulic design considerations specific to nappe characteristics. DOI: 10.1061/(ASCE)IR.1943-4774
.0000558. © 2013 American Society of Civil Engineers.
CE Database subject headings: Weirs; Water discharge; Coefficients; Hydraulics; Design; Irrigation.
Author keywords: Labyrinth weir design; Discharge coefficients; Weir crest shape; Cycle efficiency; Local submergence.

Introduction direction) of equivalent length. However, as the head increases,

A labyrinth weir is a linear weir that is folded in plan-view to in-
crease the crest length for a given channel or spillway width (Fig. 1).
There are an infinite number of possible labyrinth weir configura-
tions and design variations; however, labyrinth cycles are typically
placed in a linear fashion (i.e., upstream apexes align at a common
channel cross section; Fig. 1), have a sidewall angle (α) less than
30°, and are oriented towards the approaching flow.
A labyrinth weir is able to pass large discharges at relatively low
heads compared to traditional linear weir structures of equal width.
As a result of their hydraulic performance and geometric versatility,
labyrinth weirs have been placed in streams, canals, rivers, ponds,
and reservoirs as headwater control structures, energy dissipaters,
flow aerators, and spillways. Labyrinth weirs are well suited for
spillway rehabilitation where aging infrastructure, dam safety con-
cerns, freeboard limitations, and revised and larger probable maxi-
mum flows have required increased spillway capacity. Recently
constructed examples are: Lake Brazos spillway in Texas (Vasquez
et al. 2007) and Lake Townsend spillway in Greensboro, North
Carolina (Tullis and Crookston 2008).
Flow Characteristics
The geometry of a labyrinth weir produces complex three-
dimensional flow patterns. At very low heads, it behaves similar
to a linear weir (α ¼ 90°, oriented normally to the approach flow
1 design methods; a selection is presented and discussed. Hay and
Postdoctoral Researcher, Utah Water Research Laboratory, Dept. of
Civil and Environmental Engineering, Utah State Univ., 8200 Old Main
Hill, Logan, UT 84322-8200. E-mail: bcrookston@gmail.com
2
Associate Professor, Utah Water Research Laboratory, Dept. of Civil
and Environmental Engineering, Utah State Univ., 8200 Old Main Hill,
Logan, UT 84322-8200 (corresponding author). E-mail: blake.tullis@
usu.edu
Note. This manuscript was submitted on July 6, 2011; approved on
October 25, 2012; published online on October 29, 2012. Discussion per-
iod open until October 1, 2013; separate discussions must be submitted for
individual papers. This paper is part of the Journal of Irrigation and
Drainage Engineering, Vol. 139, No. 5, May 1, 2013. © ASCE, ISSN
0733-9437/2013/5-363-370/$25.00.
labyrinth weir discharge efficiency, as quantified by the discharge
coefficient value, begins to decline as nappe collision and local
submergence regions develop (Crookston and Tullis 2012c).
Previous Studies
Labyrinth weir head-discharge relationships have been described
by various empirical equations. These relationships vary based
on different definitions of the discharge coefficient, the character-
istic weir length, and the upstream driving head (e.g., the inclusion
of the velocity head component V 2 =2g, described in the following).
In the present study, a standard form of the weir equation, Eq. (1),
was selected with the centerline length of the crest (Lc ) as the char-
acteristic weir length:
2 pffiffiffiffiffi
Q ¼ Cdðα°Þ Lc 2gH3=2 ð1Þ
3 T
where Q = labyrinth weir discharge; Cdðα°Þ = dimensionless dis-
charge coefficient; Lc ¼ Nð2lc þ A þ DÞ where N = number of
cycles, lc = centerline length of the sidewall, A = inside apex length,
and D = outside apex length; g = acceleration constant of gravity;
and HT = total upstream head (unsubmerged) measured relative
to the crest elevation [H T ¼ V 2 =2g þ h (V is the average cross-
sectional velocity at the gauging location, and h is the piezometric
head upstream of the weir)].
Several earlier labyrinth weir studies resulted in published
Taylor (1970) presented parameter guidelines, based upon research
by Taylor (1968), for sharp-crested triangular and trapezoidal
labyrinth weirs. Discharge rating curves for h=P < 0.6 were pre-
sented in terms of a labyrinth-to-linear weir discharge ratio (based
on a common channel width, W, and h), requiring discharge infor-
mation for a linear weir (α ¼ 90°) of equivalent weir height (P),
wall thickness (tw ), and crest shape. The Bureau of Reclamation
(USBR) conducted model studies to aid in the design of Ute Dam
(Houston 1982). Discrepancies between their experimental results
and the recommendations by Hay and Taylor (1970) were attributed
to different definitions of upstream head [h, Hay and Taylor (1970);

JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013 / 363

J. Irrig. Drain Eng. 2013.139:363-370.


Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.
Fig. 1. Labyrinth weir schematic including geometric parameters
H T , USBR] and limited geometric data. From the physical model
studies of Ute Dam and Hyrum Dam, Hinchliff and Houston (1984)
developed new design guidelines. Despite scope limitations, they
provided valuable insights regarding labyrinth weir orientation and
placement in reservoir and channel applications.
Based upon model studies of Avon and Woronora Dam, Darvas
(1971) simplified labyrinth weir design by introducing an empirical
discharge equation and a discharge coefficient to accompany dis-
charge rating curves for labyrinth weirs. However, Magalhães and
Lorena (1989) juxtaposed this method with their own experimental
results for labyrinth weirs with a truncated ogee crest shape and
reported their curves to be systematically lower than the data
presented by Darvas (1971).
Lux and Hinchliff (1985) and Lux (1984, 1989) developed a
new empirical equation that includes the cycle width ratio, w=P,
and an apex shape constant to determine the discharge of a single
labyrinth cycle. Although this dimensionless equation applies to
trapezoidal and triangular weirs, the inclusion of w=P complicates
the weir equation and was limited to w=P ≥ 2.0. Similar parameter
limits have been set by other design methods that do not explicitly
include w=P in the head-discharge equation.
Tullis et al. (1995) developed a design method based upon
Eq. (1) and research conducted at the Utah Water Research Labo-
ratory (UWRL) by Waldron (1994), Tullis (1993), and Amanian
(1987). Instead of using W or Lc, Tullis et al. (1995) introduced
an effective weir length, Le , as the characteristic weir length in
Eq. (1) to define the head-discharge relationship for trapezoidal,
quarter-round labyrinth weirs, Le < Lc , and was intended to ac-
count for apex influences on discharge efficiency. They also
presented a spreadsheet-based labyrinth weir design program for
quarter-round crest applications. Willmore (2004), however, noted
the following discrepancies: the α ¼ 8° data fall above the α ¼ 9°
presented by Waldron (1994), and a minor mathematical error
was found in the weir length geometric calculations. The α ¼ 6°
data are also significantly lower and trend differently than the
adjacent curves. The supporting quarter-round crest shape data
364 / JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013
J. Irrig. Drain Eng. 2013.139:363-370.
for the Tullis et al. (1995) method were limited to 6° ≤ α ≤ 18°
and α ¼ 90°; the α ¼ 25° and 35° Cdðα°Þ curves were linearly in-
terpolated. Falvey (2003) reviewed many past research studies
and added new interpretations; this document has contributed to an
increase in labyrinth weir use in practice.
Melo et al. (2002) expanded the work of Magalhães and Lorena
(1989) by adding an adjustment parameter for labyrinth weirs lo-
cated in a channel with converging sidewalls. Tullis et al. (2007)
developed a dimensionless submerged head-discharge relationship
(tailwater submergence) for labyrinth weirs that was verified by
Lopes et al. (2009).
Physical models have proven to be highly useful in designing
and analyzing specific labyrinth weir geometries, and hydraulic de-
sign methods are useful tools to estimate the discharge performance
of labyrinth weir geometric configurations. The purpose of this
study is to provide new insight into the performance and operation
of labyrinth weirs and to improve the currently available design and
evaluation tools. This is to be accomplished by utilizing experimen-
tal results from physical modeling to provide a design optimization
program, an analysis program, and additional hydraulic informa-
tion. Paper II presents additional labyrinth weir hydraulic informa-
tion and design considerations based on nappe behavior (Crookston
and Tullis 2013). The design program developed during this study
is similar to that of Tullis et al. (1995), but accommodates a user-
specified footprint size (channel width, W, and apron length, B).
This method contains new, extensive, and accurate Cdðα°Þ data sets
for quarter-round and half-round crest shapes, utilizes Lc of the
crest as the characteristic length, and includes additional design
considerations such as cycle efficiency (ε 0 ), aeration device place-
ment, and tailwater submergence effects. The labyrinth weir design
method presented in these two companion papers is complimented
by the additional research and design information on: labyrinth
weirs in reservoir applications (Crookston and Tullis 2012b); linear
and arced labyrinth weir cycle configurations (Crookston and Tullis
2012a); and labyrinth weir nappe interference and local submer-
gence (Crookston and Tullis 2012c).
Experimental Method
Physical modeling of labyrinth weirs (Crookston 2010) was con-
ducted at the UWRL. Labyrinth weirs were fabricated from high
density polyethylene (HDPE) sheeting and tested in a rectangular
flume (1.2 × 14.6 × 1.0 m). The influence of acrylic sidewalls of
the flume on the sectional model hydraulic performance was con-
sidered to be negligible, based upon the findings of Johnson (1996).
Details of the tests are summarized in Table 1.
Calibrated orifice meters in the flume supply piping, differential
pressure transducers, and a data logger were used to determine flow
rates. The flume was equipped with a headbox and baffle (to create
uniform and tranquil approach flow conditions), a stilling well (hy-
draulically connected to the flume 3P upstream from the labyrinth
weir), and a rolling instrument carriage. The point gauge instru-
mentation was carefully referenced to the crest elevation of the
weirs. The labyrinth weirs were installed on an elevated horizontal
apron with a ramped upstream floor transition. Experimental data
were collected under steady-state conditions. The Q measurements
were recorded for 5 to 7 min with the data logger to determine an
average flow rate and h was determined with the stilling well
equipped with a point gage accurate to
0.15 mm.
In an effort to accurately characterize the labyrinth weir behav-
ior, ∼85 head-discharge data points were collected for each tested
weir geometry (∼2,600 total). A system of checks was established
 Table 1. Physical Model Test Program The influence of labyrinth weir orientation in the channel [distal
α a
P Lc-cycle Lc-cycle =w w=P N Crest apexes connecting to channel sidewalls as upstream apexes (re-
Model (°) (mm) (mm) (−) (−) (−) (−) ferred to as normal orientation) or as downstream apexes (inverse
orientation)] was evaluated by testing the α ¼ 6° labyrinth weirs
1b 6 304.8 4,654.6 7.607 2.008 2 HR
2–3 6 304.8 4,654.6 7.607 2.008 2 QR, HR
with both orientations. No measurable variations were observed
4–5 8 304.8 3,544.9 5.793 2.008 2 QR, HR between Cdðα°Þ data sets (data not presented). Consequently, the
6–7 10 304.8 2,879.1 4.705 2.008 2 QR, HR data in Figs. 2 and 3 are assumed to be applicable, independent
8–9 12 304.8 2,435.1 3.980 2.008 2 QR, HR of weir orientation. Crookston and Tullis (2012b) present addi-
10–11 15 304.8 1,991.4 3.254 2.008 2 QR, HR tional design information and discussion regarding labyrinth weirs
12 15 152.4 1,991.4 3.254 4.015 2 QR in reservoir applications and abutment effects on discharge.
13 15 152.4 995.7 3.254 2.008 4 QR For convenience, the labyrinth weir Cdðα°Þ data in Figs. 2 and 3
14 15 304.8 995.7 3.254 1.019 4 QR were curve-fit per Eq. (2), and the corresponding coefficients are
15–16 20 304.8 1,548.1 2.530 2.008 2 QR, HR presented in Tables 2 and 3. Eq. (3) was used for α ¼ 90° data
Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.

17–18 35 304.8 983.5 1.607 2.008 2 QR, HR

19–20 90 304.8 1,223.8 1.000 4.015 — QR, HR
a
Linear configuration and normal orientation for all models unless noted.
b
Inverse orientation.
in which at least 10% of the data were repeated to ensure accuracy
and to determine measurement repeatability.
Experimental Results
Discharge Rating Curves
Eq. (1) was used to quantify the labyrinth weir head-discharge re-
lationship with Lc representing the characteristic weir length. The
term Cdðα°Þ can be influenced by weir geometry (e.g., P, tw , A, w, α,
and crest shape), weir abutments, flow conditions (H T , approach-
ing flow angle, local submergence, and nappe interference), and
nappe aeration conditions (clinging, aerated, partially aerated,
and drowned). The two-cycle sectional labyrinth weir models
evaluated in this study did not account for the influence of abut-
ments on discharge. Data for Cdðα°Þ are presented in terms of
H T =P for nonvented trapezoidal labyrinth weirs for 6° ≤ α ≤ 35°
in Fig. 2 (quarter-round crest shape) and Fig. 3 (half-round crest
shape). The data for α ¼ 90° (linear) weirs are also included for
comparison.
and the corresponding coefficients are presented in the aforemen-
tioned tables. The curves have been validated for 0.05 ≤ HT =
P < ∼0.8–0.9; however, due to the well-behaved nature of the
data and Eq. (2), the Cdðα°Þ curves have been extrapolated to
HT =P ¼ 1.0. Eqs. (2) and (3) were selected over polynomial rela-
tionships because of their improved data representation (R2 ≥ 0.99)
and extrapolation performance (they remain well-behaved up to
HT =P ≤ 2.0). Crookston et al. (2012) evaluated Eq. (2) for HT =P <
2 via physical and numerical modeling and found that Eq. 2 may be
used as a good first order approximation. When the Tullis et al.
(1995) polynomial Cdðα°Þ relationships are extrapolated, they incor-
rectly compute Cdðα°Þ (even producing negative values) beyond the
upper limit of their experimental data (experimental data limited to
HT =P < 0.9). For labyrinth weirs:
  b H T  c
HT P
Cdðα°Þ ¼ a þd ð2Þ
P
For linear weirs:
1
Cdð90°Þ ¼ þd ð3Þ
a þ b HPT þ HTc=P
A comparison between the half-round and quarter-round exper-
imental data is presented in Fig. 4 as the ratio of the half-round over

Fig. 2. Values of Cd versus H T =P for quarter-round trapezoidal labyrinth weirs

JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013 / 365

J. Irrig. Drain Eng. 2013.139:363-370.

Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Values of Cd versus H T =P for half-round trapezoidal labyrinth weirs

Table 2. Curve-Fit Coefficients for Quarter-Round Labyrinth and Linear

Weirs, Validated for 0.05 ≤ H T =P < 0.9
α a b c d
6° 0.02623 −2.681 0.3669 0.1572
8° 0.03612 −2.576 0.4104 0.1936
10° 0.06151 −2.113 0.4210 0.2030
12° 0.09303 −1.711 0.4278 0.2047
15° 0.10890 −1.723 0.5042 0.2257
20° 0.11130 −1.889 0.5982 0.2719
35° 0.03571 −3.760 0.7996 0.4759
90° −2.3800 6.476 1.3710 0.5300

Table 3. Curve-Fit Coefficients for Half-Round Labyrinth and Linear

Weirs, Validated for 0.05 ≤ H T =P < 0.9 Fig. 4. Comparison of half-round and quarter-round crest shape on
hydraulic performance of labyrinth weirs
α a b c d
6° 0.009447 −4.039 0.3955 0.1870
8° 0.017090 −3.497 0.4048 0.2286
10° 0.029900 −2.978 0.4107 0.2520 HT =P. Improving weir approach flow conditions and using a more
12° 0.030390 −3.102 0.4393 0.2912 efficient crest shape can obtain further gains in efficiency. Brazos
15° 0.031600 −3.270 0.4849 0.3349 Dam (Waco, TX), for example, features an ogee-type crest [modi-
20° 0.033610 −3.500 0.5536 0.3923 fied half-round crest with an upstream radius of 1=3tw and a down-
35° 0.018550 −4.904 0.6697 0.5062 stream radius of 2=3tw (Willmore 2004)].
90° −8.60900 22.650 1.8120 0.6375
Nappe Behavior and Artificial Aeration
Labyrinth weir flow can produce several nappe behavior phenom-
the quarter-round Cdðα°Þ values (Cd-HR =Cd-QR ) versus HT =P. ena that should not be overlooked: nappe aeration conditions,
A crest that is rounded on the downstream face helps the flow stay nappe instability (also termed flow surging), and nappe vibrations.
attached (clinging flow) to the downstream weir wall at smaller These behaviors, conditions, and remedial actions are discussed
H T =P values, thus increasing flow efficiency and discharge capac- in detail in the companion paper (Crookston and Tullis 2013).
ity. As the discharge and the corresponding momentum of the flow The artificial aeration, vent pipes, and nappe breakers are also
passing over the weir increase, the nappe becomes aerated and the discussed.
streamlines will eventually detach from the weir crest, creating a
similar nappe profile to the quarter-round crest. Once nappe detach-
Labyrinth Design and Analyses
ment occurs, the gains in the half-round crest flow efficiency are
lost relative to the quarter-round crest. All of the Cd-HR =Cd-QR The recommended procedure for designing a labyrinth weir is pre-
curves are anticipated to eventually converge to 1.0 with increasing sented in Table 4, which includes a design example for illustration

366 / JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013

J. Irrig. Drain Eng. 2013.139:363-370.

 Table 4. Recommended Design Procedure for Labyrinth Weirs
Parameter Symbol Value Units Notes
Hydraulic conditions—input data
Design flow Qdesign 1,500.00 (m3 =s) g ¼ 9.81 m=s2
Design flow water surface elevation H 1,680.00 (m)
Approach channel elevation H apron 1,674.00 (m)
Crest elevation H crest 1,678.00 (m)
Unsubmerged total upstream head HT 2.00 (m) Piezometric head + velocity head - losses
Downstream total head Hd 0.50 (m)
Labyrinth weir geometry—input data
Angle of side legs α 12 (°) α ∼ 6°–35°
Number of cycles N 9 — Whole or half cycles
Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.

Cycle width w 11.10 (m)

Crest height P 4.00 (m) P ∼ 1.0H T
Thickness of weir wall at the crest tw 0.50 (m) tw ∼ P=8
Inside apex width A 0.50 (m) A ∼ tw
Crest shape Crest shape Quarter — Quarter- or half-round
Aeration device (nappe breakers, vents) — Breakers — Breakers, vents, or none
Calculated data
Headwater ratio H T =P 0.50 —
Labyrinth weir discharge crest coefficient Cdðα°Þ 0.429 — Cdðα°Þ ¼ fðH T =P; α; Crest ShapeÞ
Total centerline length of weir Lc 418.28 (m) Lc ¼ 3=2Qdesign =½ðCdða° Þ H 3=2
T Þð2gÞ
1=2

Centerline length of sidewall lc 2.33 (m) lc ¼ ðB − tw Þ= cosðαÞ

Outside apex width D 1.30 (m) D ¼ A þ 2tw tanð45 − α=2Þ
Width of labyrinth (normal to flow) W 99.87 (m) W ¼ Nw
Length of apron (parallel to flow) B 22.35 (m) B ¼ ½Lc =ð2NÞ − ðA þ DÞ=2 cos ðαÞ þ tw
Magnification ratio M 4.19 — M ¼ Lc =ðwNÞ
Cycle width ratio w=P 2.77 — Normally 2 ≤ w=P ≤ 4
Relative thickness ratio P=tw 8.0 — ∼8
Apex ratio A=w 0.05 — <0.08
Cycle efficiency ε0 1.80 — ε 0 ¼ Cdðα°Þ M
Efficacy ε 2.23 — ε ¼ Cdðα°Þ M=Cdð90°Þ
Number of nappe breakers or vents — 9 — Breaker on ds apex, 1 vent per sidewall
Linear weir discharge coefficient Cdð90°Þ 0.808 — Cdð90°Þ ¼ fðH T =P; α; Crest ShapeÞ
Length of linear weir for same flow Lcð90°Þ 222.33 (m) Lcð90°Þ ¼ 3=2Qdesign =½ðCdð90°Þ H 3=2
T Þð2gÞ
1=2 

Submergence (Tullis et al. 2007)

Downstream/upstream ratio of unsubmerged head H d =H T 0.25 (m)
Submerged head discharge ratio H  =H T 1.013 — Piecewise function Tullis et al. (2007)
Submerged upstream total head H 2.025 (m)
Submergence level S 0.247 — S ¼ H d =H 
Submerged weir discharge coefficient Cd−sub 0.421 — Cdðα°Þ ðH T =H  Þ3=2

Note: Design is limited to the extent of experimental data; designs that exceed these limits may warrant a physical model study. Table 4 can be modified so that
the weir footprint (B and W) and N are independent variables and α and w are dependent variables.

(input data, computed results). Table 4 can be adjusted for SI or
Imperial units and includes a “Notes” column that presents appli-
cable design guidance for each design variable (formula and geo-
metric limitations.). The top section of the design table includes the
user-defined hydraulic conditions or requirements for the labyrinth
weir. The design flow rate (Qdesign ) typically represents a specific
return-period flood discharge from a hydrologic analysis, HT is
based on upstream freeboard or flood plain constraints, and H d
is determined by a backwater curve based on Qdesign . User-defined
labyrinth weir geometric parameters (e.g., α, P, tw , crest shape, N)
are entered in the second section of the design table. Also, a place is
provided to specify a nappe aeration device if desired. The third
section contains the labyrinth weir design geometry, design ratios,
and hydraulic performance output data (Cdðα°Þ , ε 0 ), which are
calculated using the input data from the previous sections. Cells
are provided to annotate anticipated nappe aeration behavior,
determined from information presented in Paper II. If a specific
labyrinth weir footprint size is required, Table 4 can be adjusted
by switching B to an independent variable and N to a dependent
variable using the equations provided in Table 4. For comparison,
Cdð90°Þ and the required linear weir length (with the same crest
shape as the labyrinth weir) are reported, which match the labyrinth
weir head-discharge condition. The last section of the design
method includes the submerged head-discharge relationships de-
veloped by Tullis et al. (2007).
Per Figs. 2 and 3, Cdðα°Þ decreases with decreasing α. For a
given weir footprint (W and B held constant); however, the
labyrinth weir crest length increases with decreasing α. The net
effect of the α-dependent opposing influences of Cdðα°Þ and Lc
on discharge efficiency should be considered when optimizing a
labyrinth weir design. To account for the influences of Cdðα°Þ
and Lc on discharge capacity and to provide a guide in the selection
of α, cycle efficiency [ε 0 ¼ Cdðα°Þ Lc-cycle =w] is introduced,
representative of the discharge per cycle (at a specific H T value)
for a given labyrinth weir geometry. Values of ε 0 for the experimen-
tal labyrinth weir data are presented in Figs. 5 (quarter-round) and 6
(half-round) as a function of H T =P. The data indicate that the in-
crease in weir length more than compensates for the decrease in

JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013 / 367

J. Irrig. Drain Eng. 2013.139:363-370.

Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.
Fig. 5. Cycle efficiency versus H T =P for quarter-round labyrinth weirs
Cdðα°Þ with smaller α values. Figs. 5 and 6 also show that the maxi-
mum ε 0 values occur at relatively low HT =P (as delineated by the
dashed line), ε 0 increases as α decreases, and the benefit of in-
creased ε 0 of smaller α angles decrease with increasing HT =P.
The value of ε 0 only evaluates the discharge efficiency per cycle.
In choosing a labyrinth weir design, a complete feasibility study
(construction cost, environmental impact, and permitting issues)
is recommended.
Beyond the ability to design a labyrinth weir for a particular
head-discharge condition using Table 4, the ability to determine
a complete head-discharge relationship for specific labyrinth weir
geometries is also important. Such a procedure, which also adapts
easily to a spreadsheet program format, is outlined in Fig. 7. The
known labyrinth weir geometries are entered. Missing geometric
parameters and labyrinth weir ratios are calculated and a head-
discharge rating curve is produced. The effects of tailwater sub-
mergence may be determined by solving for Q or submerged
upstream total head, H , by iteration (Tullis et al. 2007). The
dimensionless submerged head relationship, based on Tullis et al.
(2007), is presented in Fig. 8, which includes H and tailwater
submergence levels (S).
The design method and support data are limited to the geom-
etries (Table 1) and hydraulic conditions tested in this study
(e.g., 0.05 ≤ H T =P ≤ 0.9). These results may be conservatively
applied (with sound engineering judgment) to other labyrinth
weir geometries and flow conditions (design verification with a
hydraulic model study is recommended). Linear interpolation is
recommended to determine Cdðα°Þ for α values other than those
presented. As previously discussed, the form of the predictive
Fig. 6. Cycle efficiency versus HT =P for half-round labyrinth weirs
368 / JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013
J. Irrig. Drain Eng. 2013.139:363-370.
Fig. 7. Recommended procedure for analyses of labyrinth weirs
Cdðα°Þ relationships [Eq. (2)] was chosen because of improved
extrapolation behavior for values of 1.0 ≤ HT =P ≤ 2.0, relative
to the polynomial functions used by Tullis et al. (1995). Although
insufficient data are currently available to validate the predictive
accuracy of all of the α-specific Cdðα°Þ relationships over the range
of 1.0 ≤ H T =P ≤ 2.0, the α ¼ 15° labyrinth weir model (Model 13
in Table 1) was tested up to HT =P ¼ 2.0 and good agreement
was noted between the experimental and predictive Cdðα°Þ data.
Using the labyrinth weir Cdðα°Þ relationships for 1.0 ≤ H T =P ≤
2.0 conditions is recommended as a first-order approximation of
head-discharge performance. Physical modeling is recommended
to validate high-head labyrinth weir hydraulic performance.
Crookston and Tullis (2012a, b, c, 2013) and Crookston et al.
(2012) provide additional design information intended to compli-
ment the design method presented herein.
Data Verification
An uncertainty analysis was performed, as outlined by Kline and
McClintock (1953) for single-point experimental data. The percent
Fig. 8. Dimensionless submerged head relationship for labyrinth weirs
(data from Tullis et al. 2007)
 uncertainty for each Cdðα°Þ was computed by determining individ-
ual parameter uncertainties (e.g., measurement accuracy of Q, Lc ,
h, and V) and taking partial derivatives of Eq. (1), following the
methodology prescribed by Kline and McClintock. The average
single sample uncertainties for each data set were 1–1.5%. Uncer-
tainties were largest (∼4% for HT =P ≤ 0.075) for very small
values of Q and h (instrument accuracy) and smallest (≤0.9% for
H T =P ≥ 0.3) for large values of Q and h.
The Cdðα°Þ data from the current study were also compared with
the results of previous studies. The data in Fig. 9 show relatively
good agreement between the nonvented, half-round labyrinth weir
Cdðα°Þ data and the data from Willmore (2004), with the largest dis-
crepancy appearing for large α labyrinth weirs at HT =P ≤ 0.2.
Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.
Fig. 10 compares the quarter-round, nonvented labyrinth weir
Cdðα°Þ data of this study with Cdðα°Þ data from Tullis et al.
(1995) (presented in terms of Lc rather than Le ). There appears
to be relatively good agreement at large values of HT =P; however,
large differences are present for HT =P ≤ 0.4. This may be attrib-
uted to the smaller sized labyrinth weir models used by the Tullis
et al. (1995) method (potential size-scale effects and different val-
ues of P), a potentially higher level of data uncertainty (e.g., their
flow measurement technique), and the fact that B was held constant
(not N), which in some cases resulted in partial cycles. The α ¼ 25°
and α ¼ 35° Cdðα°Þ curves linearly interpolated by Tullis et al.
(1995) (based on α ¼ 18° and α ¼ 90° experimental data) also
Fig. 9. Comparison between Cd values obtained by Willmore (2004)
and Crookston (2010) for nonvented, half-round labyrinth weirs
Fig. 10. Comparison between proposed Cd design curves by Tullis
et al. (1995), Willmore (2004), and Crookston (2010) for nonvented,
quarter-round labyrinth weirs (based upon Lc )
JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013 / 369
J. Irrig. Drain Eng. 2013.139:363-370.
do not agree with the data from the current study. The α ¼ 8° data
from Tullis et al. (1995) is incorrect; it has been replaced in Fig. 10
with the data from Willmore (2004) α ¼ 8° quarter round, which
shows good agreement with the data from the present study.
Summary and Conclusions
A labyrinth weir design and analysis procedure is presented
(Table 4) based upon the results of physical modeling in a labora-
tory flume. The discharge is calculated based on the traditional
weir equation [Eq. (1)], utilizing H T as the representative upstream
head and Lc as the characteristic weir length. Tailwater submer-
gence for labyrinth weirs is included, as presented by Tullis et al.
(2007). The proposed design and analysis method was validated by
juxtaposing the experimental Cdðα°Þ data of this study with other
physical model studies presented in Figs. 9 and 10. Paper II is
devoted to nappe behavior and discusses aeration conditions, insta-
bility or flow surging, artificial aeration and structure placement,
and nappe vibration.
Figs. 2 and 3 present a dimensionless discharge coefficient,
Cdðα°Þ , as a function of HT =P for quarter-round and half-round
labyrinth weirs (6° ≤ α ≤ 35°) and for linear weirs. The test results
indicate that the increase in efficiency provided by a half-round
crest shape (relative to a quarter-round crest) is more significant
for HT =P ≤ 0.4 (Fig. 4).
Cycle efficiency, ε 0 , represents the labyrinth weir cycle-specific
discharge (for a given HT condition) and is a tool for examining
the discharge capacity of different labyrinth weir geometries
(Figs. 5 and 6). An ε 0 analysis found that for a fixed cycle width
(w), the cycle efficiency increased with decreasing values of α be-
cause the increase in cycle crest length, Lc-cycle (with decreasing α),
more than compensates for decreasing Cdðα°Þ values.
Although the methods and tools presented in this study are
recommended for use in the design and analyze of labyrinth
spillways, a physical model study is recommended to verify hy-
draulic performance. A model study would include site-specific
conditions that may be outside the scope of this study and may
provide valuable insights into the performance and operation of
the labyrinth weir.
Additional components of this study not presented here include
arced labyrinth weirs (Crookston and Tullis 2012a), reservoir-
specific labyrinth weir applications (Crookston and Tullis 2012b),
local submergence and nappe interference (Crookston and Tullis
2012c), scale effects, 1 ≤ H T =P ≤ 2 (Crookston et al. 2012),
and the influence of the cycle width ratio and labyrinth apexes
on discharge (2=3 ≤ w=P ≤ 4).
Acknowledgments
This study was funded by the State of Utah and the Utah Water
Research Laboratory at Utah State University.
Notation
The following symbols are used in this paper:
A = inside apex width;
A=w = apex ratio;
a = curve-fit coefficient for quarter-round and half-round
labyrinth and linear weirs (Tables 2 and 3);
B = length of labyrinth weir (apron) in flow direction;
b = curve-fit coefficient for quarter-round and half-round
labyrinth and linear weirs (Tables 2 and 3);
 Cd = discharge coefficient, data from current study;
Cdðα°Þ = discharge coefficient for labyrinth weir of sidewall
angle α;
Cdð90°Þ = discharge coefficient for linear weir;
Cd-sub = discharge coefficient for a labyrinth weir with tailwater
submergence;
c = curve-fit coefficient for quarter-round and half-round
labyrinth and linear weirs (Tables 2 and 3);
D = outside apex width;
d = curve-fit coefficient for quarter-round and half-round
labyrinth and linear weirs (Tables 2 and 3);
g = acceleration constant of gravity;
H = design flow water surface elevation;
Downloaded from ascelibrary.org by UTAH STATE UNIVERSITY on 04/11/13. Copyright ASCE. For personal use only; all rights reserved.
Hapron = approach channel elevation;
Hcrest = elevation of labyrinth weir crest;
Hd = total downstream head measured relative to the weir
crest;
H d =HT = downstream/upstream ratio of total unsubmerged head;
HT = total upstream head (unsubmerged) measured relative to
the weir crest;
HT =P = headwater ratio;
H = submerged total upstream head measured relative to the
weir crest;
H =HT = submerged head discharge ratio;
h = depth of flow over the weir crest;
Lc = total centerline length of labyrinth weir;
Lc-cycle = centerline length for a single labyrinth weir cycle;
Le = total effective length of labyrinth weir;
lc = centerline length of weir sidewall;
M = magnification ratio, Lc-cycle =w;
N = number of labyrinth weir cycles;
P = weir height;
P=tw = relative thickness ratio;
Q = discharge over weir;
Qdesign = design discharge over labyrinth weir;
Rcrest = radius of rounded crest (e.g., Rcrest ¼ tw =2);
S = submergence level;
tw = thickness of weir wall;
V = average cross-sectional flow velocity upstream of weir;
W = width of channel;
w = width of a single labyrinth weir cycle;
w=P = cycle width ratio;
α = sidewall angle;
ε = efficacy; and
ε 0 = cycle efficiency.
References
Amanian, N. (1987). “Performance and design of labyrinth spillways.”
M.S. thesis, Utah State Univ., Logan, UT.
Crookston, B. M. (2010) “Labyrinth weirs.” Ph.D. dissertation, Utah State
Univ., Logan, UT.
Crookston, B. M., Paxson, G. S., and Savage, B. M. (2012). “It can be
done! Labyrinth weir design guidance for high headwater and low cycle
width ratios.” Proc., Dam Safety 2012, ASDSO, Denver, CO.
Crookston, B. M., and Tullis, B. P. (2012a). “Arced labyrinth weirs.”
J. Hydraul. Eng., 138(6), 555–562.
370 / JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / MAY 2013
View publication stats J. Irrig. Drain Eng. 2013.139:363-370.
Crookston, B. M., and Tullis, B. P. (2012b). “Discharge efficiency of
reservoir-application-specific labyrinth weirs.” J. Irrig. Drain. Eng.,
138(6), 773–776.
Crookston, B. M., and Tullis, B. P. (2012c). “Labyrinth weirs: Nappe
interference and local submergence.” J. Irrig. Drain. Eng., 138(8),
757–765.
Crookston, B. M., and Tullis, B. P. (2013). “Hydraulic design and analysis
of labyrinth weirs. II: Nappe aeration, instability, and vibration.”
J. Irrig. Drain. Eng., 139(5), 371–377.
Darvas, L. (1971). “Discussion of ‘Performance and design of labyrinth
weirs’ by Hay and Taylor.” J. Hydraul. Eng., 97(80), 1246–1251.
Falvey, H. (2003). Hydraulic design of labyrinth weirs, ASCE, Reston,
VA.
Hay, N., and Taylor, G. (1970). “Performance and design of labyrinth
weirs.” J. Hydraul. Eng., 96(11), 2337–2357.
Hinchliff, D., and Houston, K. (1984). “Hydraulic design and application of
labyrinth spillways.” Proc., 4th Annual USCOLD Lecture, Dam Safety
and Rehabilitation, Bureau of Reclamation, U.S. Dept. of the Interior,
Washington, DC.
Houston, K. (1982). “Hydraulic model study of Ute Dam labyrinth
spillway.” Rep. No. GR-82-7, U.S. Bureau of Reclamation, Denver.
Johnson, M. (1996). “Discharge coefficient scale effects analysis for weirs.”
Ph.D. dissertation, Utah State Univ., Logan, UT.
Kline, S., and McClintock, F. (1953). “Describing uncertainties in single-
sample experiments.” Am. Soc. Mech. Eng., 75(1), 3–8.
Lopes, R., Matos, J., and Melo, J. (2009). “Discharge capacity for free-flow
and submerged labyrinth weirs.” Proc., 33rd IAHR Congress: Water
Engineering for a Sustainable Environment, IAHR, Madrid, Spain,
1054–1061.
Lux, F., III (1984). “Discharge characteristics of labyrinth weirs.” Proc.,
Conf. on Water for Resources Development, ASCE, Reston, VA.
Lux, F., III (1989). “Design and application of labyrinth weirs.” Design
of hydraulic structures 89, M. Alberson and R. Kia, eds., Balkema,
Rotterdam, Netherlands, 205–215.
Lux, F., III, and Hinchliff, D. (1985). “Design and construction of labyrinth
spillways.” 15th Congress ICOLD, Vol. IV, Q59-R15, ICOLD, Paris,
249–274.
Magalhães, A., and Lorena, M. (1989). “Hydraulic design of labyrinth
weirs.” Rep. No. 736, National Laboratory of Civil Engineering, Lisbon,
Portugal.
Melo, J., Ramos, C., and Magalhães, A. (2002). “Descarregadores com
soleira em labirinto de um ciclo em canais convergentes. Determinação
da capacidad de vazão.” Proc., 6th Congresso da Água, CD-ROM,
Associação Portuguesa dos Recursos Hídricos, Lisboa, Portugal,
(in Portuguese).
Taylor, G. (1968). “The performance of labyrinth weirs.” Ph.D. thesis,
Univ. of Nottingham, Nottingham, England.
Tullis, B., and Crookston, B. M. (2008). “Lake Townsend Dam spillway
hydraulic model study report.” Utah Water Research Laboratory,
Logan, UT.
Tullis, B., Young, J., and Chandler, M. (2007). “Head-discharge relation-
ships for submerged labyrinth weirs.” J. Hydrol. Eng., 133(3), 248–254.
Tullis, P. (1993). “Standley Lake service spillway model study.” Hydraulic
Rep. No. 341, Utah Water Research Laboratory, Logan, UT.
Tullis, P., Amanian, N., and Waldron, D. (1995). “Design of labyrinth weir
spillways.” J. Hydrol. Eng., 121(3), 247–255.
Vasquez, V., Boyd, M., Wolfhope, J., and Garret, R. (2007). “A labyrinth
rises in the heart of Texas.” Proc., 28th Annual USSD Conf., USSD,
Denver, CO, 813–826.
Waldron, D. (1994). “Design of labyrinth spillways.” M.S. thesis, Utah
State Univ., Logan, UT.
Willmore, C. (2004). “Hydraulic characteristics of labyrinth weirs.”
M.S. Rep., Utah State Univ., Logan, UT.