Experimental study of Scour around a complex pier with elliptical pile-

cap
Priyanka Gautam
Ph.D. Research Fellow, Department of Civil Engineering, Indian Institute of Technology Bombay, India
T. I. Eldho
Professor, Department of Civil Engineering, Indian Institute of Technology Bombay, India
M. R. Behera
Assistant Professor, Department of Civil Engineering, Indian Institute of Technology Bombay, India

ABSTRACT: A complex bridge pier is an assembly of a column and foundation such as pile foundation with
a pile-cap, partially or completely exposed to the flow. Considering the interaction of various factors around
the pier, the pile-cap elevation in relation to the sand bed is expected to differ with time. Therefore, it is nec-
essary to assess scouring over various potential pile-cap elevations. In this paper, laboratory experiments were
carried out to analyze scour and mean velocity profiles around complex pier under the effect of low Reynolds
numbers. Three pile-cap elevation cases were studied here, viz, the sand bed depth taken above the pile-cap
level; the sand bed depth taken below the pile-cap top; and the sand bed depth taken at the level of pile-cap
bottom. The results showed different scour patterns in the three cases considered. It was also noted that the
least scour-depth occurred in the second case and maximum in the first one.

1 INTRODUCTION the flow patterns around a complex bridge pier

structure (e.g., Beheshti and Ashtiani, 2010; 2016),
A complex bridge pier is an assembly of bridge mostly with rectangle shaped pier and pile-cap or
foundation consisting of a pier column built on a rectangle with round ends.
pile-cap or footing, supported by a number of piles Beheshti and Ashtiani (2010) investigated the
(Coleman, 2005). Because of its physical and finan- three-dimensional turbulent flow field around a rec-
cial considerations, they are more preferred to be tangular complex bridge pier (with rectangular pile-
constructed in rivers these days. Over its lifetime, cap) placed on a fixed rough bed with the aid of ex-
the different components of a complex pier gets re- periments. In their setup, they exposed all the com-
vealed to the flow because of varying amount of ponents of the pier (column, pile-cap and piles) to
sand aggradation or degradation around the pier as- the incoming flow. They found out the governing
sembly. Hence, it becomes necessary for the engi- features of the flow around the pier which are re-
neers and hydraulicians to know what change in vor- sponsible for the scouring process around a complex
tex pattern is caused as a result of the varying pile- pier. The main reasons being the constricted flow
cap depth relative to the sand bed depth, and hence below the pile-cap and towards the piles, a strong
the scouring around it. downward flow along the sides of the pile-cap in the
The complications in the local scouring process at upstream region of the pier and a vortex flow behind
complex bridge pier locations are because of the in- the pile-cap.
teraction of flow structures (i.e., down flow and Few researchers have conducted experiments to
horseshoe vortex) with the three components of the study scour around complex bridge piers, with vary-
complex pier, generating new vortex structures, as ing shape of the pier structure, flow conditions as
witnessed by Beheshti and Ataie-Ashtiani (2010). well as at different pile-cap elevations. Melville et
This augments to the difficulty in predicting local al. (2006) gave a methodology to predict local scour
scour depth at these complex pier structures. depth at a complex pier to identify the relative scour-
Research on scouring at complex bridge piers is ing capacity of the various components of complex
comparatively newer than on single uniform piers, piers and the evolution of scouring phenomenon oc-
with lesser number of studies present in the literature curring for variable pile-cap elevations. Ferraro et al.
and thus demanding additional research on the flow (2013) performed experimental tests on scour
changes occurring around complex bridge piers around complex bridge piers of different arrange-
(HEC-18, 2001). Few studies have been done for the ments for studying the influence of pile-cap thick-
calculation of local scour-depth (e.g. Melville et al., ness on the temporal development of maximum
2006; Akib et al., 2014) and few for investigating scour-depth. Akib et al. (2014) presented an experi-
mental study focusing on the relationship between modifiable tailgate which was positioned in the
scour depth in complex pier groups and combined downstream end of the flume. The water was lifted
piles bridge and various factors including the varia- from the main reservoir to a constant head reservoir
tion of inflow velocity, distance, and time. Moreno by pump. The flow rate was monitored using a
et al (2015) carried out experimental campaign to Flow-meter attached to the feeder pipe from pump.
study three complex bridge pier models (rectangular
shape of pier with round ends) under steady state,
clear-water flow conditions. They assessed the effect
of relative column and the pile-cap widths on equi-
librium scour depth in three different situations, viz,
when the pile-cap was above the sand bed (Situation
1), partly submerged in the sand bed (Situation 2),
and completely buried in the sand bed (Situation 3).
Various conclusions have been made through the
experimental findings of the aforementioned investi-
gators. Ferraro et al. (2013) found out that greater
the thickness of the pile-cap, the deeper was the cor- Figure 1. Schematics of the flume used for experiments.
responding scour-hole. The thickness of the pile-cap
also affected the temporal variation of the maximum A schematic diagram of the experimental setup is
scour-depth, depending upon the height of the pile- shown in Figure 1. A sand box (consisting of test
cap with respect to the sand bed. Moreno et al. section) of length 4.38 m was provided at a distance
(2015) found out that the temporal variation of max- of about 2.0 m from the inlet. The location of test
imum scour-depth is affected by various factors, viz, section was selected in such a way that the flow be-
pile-cap elevation pertaining to sand, relative col- comes uniform and fully developed as it reaches the
umn width (main column width to pile-cap width), test section. Non-uniform Quartz sand of median
pile-cap thickness and the geometrical shape of the diameter d50 = 0.81 mm. The geometric standard de-
various bridge pier components. viation or gradation coefficient, σg (= (d84/d16)0.5) =
From the results obtained by Melville at al. 1.431 (found after performing Sieve Analysis of the
(2006), they observed that the pile-cap elevation sand), was filled in the sand box. The relative densi-
with respect to the sand bed may differ with respect ty of sediment, s was measured as 2.66 for the sam-
to time, therefore, local scouring needs to be evalu- ple.
ated over an array of pile-cap elevations. So in this The complex pier model was placed centrally and
paper, a complex bridge pier structure with cylindri- vertically in the sand box, at a distance of 4.5 m
cal main column and elliptical pile-cap is studied from the water inlet of flume. A sand recess box of
experimentally, with 3 different pile-cap elevations. length 0.9 m was provided at a distance of ~2.0 m
The effect of using an elliptical shape of pile-cap is from the pier location, to trap the eroded and trans-
studied, assuming it may help in avoiding excessive ported sand and prevent settlement of sand in the
scour around them because of its streamlined shape. water reservoir.
The experiments were conducted under clear water
conditions, with low Reynolds numbers. The time
evolution of scour, equilibrium scour depth, scour
patterns and mean velocity profiles around the com-
plex pier were analyzed and are presented here.

2 EXPERIMENTAL SETUP AND PROCEDURE

The experiments were carried out in a horizontal, re-

circulating flume of length 8.4 m, width 0.3 m and
depth of 0.55 m in Hydraulics laboratory, Indian In-
stitute of Technology (IIT) Bombay, India (Fig. 1). Figure 2. Geometric characteristics of complex pier: (a) side
The sidewalls of the flume were made of transparent view; (b) front view; (c) top view; (d) bottom view.
Plexiglas to facilitate visual observation of the flow.
At the inlet section, a stilling ramp was provided for The complex pier model was designed with a cyl-
uniform entry of water into the flume consisting of a inder resting on an elliptical shaped pile-cap rein-
metallic grid and an ascending ramp with sand glued forced by an array of 2 x 2 piles (Fig. 2).
to its surface for damping the flow disturbances The diameter of the main column (Dc) was taken
leading to uniform flow distribution. Required nor- as 3 cm. Chiew and Melville (1987) recommended
mal flow depth over the sand bed was achieved by a B/b ≥ 10 (B = flume width) for single uniform piers
of width b so that the blockage effect of the pier be- were tested. These pile-cap positions were selected
comes insignificant in pier scour experiments. For to identify relative scouring potential of the various
this study, the ratio B/b = 10, ensuring negligible components of the complex pier.
blockage effect. Rest of the definition of the parame- The cases are shown in Figure 3, as follows:
ters of complex pier and their dimensions are pre-
sented in Table 1. 1. Case (a) showing pile-cap buried in the sand
bed (sand bed thickness over the pile-cap top
Table 1. Complex pier definition and dimensions being 5 mm);
Parameter Definition Dimension 2. Case (b) showing pile-cap exposed to the
(cm)
flow (pile-cap thickness exposed to the flow
Lc Height of main column 25 is 10 mm); and
Dpc Height of pile-cap 5 3. Case (c) showing pile-cap resting on the bed.
Lp Height of pile 20
Dp Diameter of pile 1
Lpc Length of pile-cap 12 Each experiment was run for an average time du-
Wpc Width of pile-cap 4 ration of 9 hours. This duration was adopted after
observing the time taken to achieve equilibrium
The complex pier model was built using Acrylic scour depth around the piers. The magnitude of the
rod and sheets and then painted for better observa- maximum scour depth remained same, while the ex-
tion under water. The complex pier was placed in tent of the scour area increased in the third case,
the flume such that the longitudinal axis of the com- even after 9 hours. The 3D velocities were recorded
plex pier was in line to the approaching flow. For using Nortek Acoustic Doppler Velocimeter (ADV).
ensuring clear water conditions. Shen et al. (1969) The Sampling rate was set at 50 Hz. At the end of
suggested that the flume width should be at least each experiment run, the water flow in the flume
eight times the diameter or size of the pier, which was slowly reduced, so as not to cause any change in
agrees well in the adopted experimental setup and the scour pattern obtained. After all the water got
pier model. For attaining clear-water conditions in drained out of the flume, scour-depth readings were
the flume, the experiments were conducted for a taken using a Vernier Point-Gauge.
number of discharge, viz, 0.009 m3/s, 0.012 m3/s,
0.015 m3/s and 0.018 m3/s. In this paper, results ob-
tained from only the highest discharge are presented, 3 RESULTS
having a flow intensity U/Uc ≈ 0.7 (< 1 for clear wa-
3.1 Temporal variation of scour
ter scour), where average flow velocity, U = 0.31
m/s and critical flow velocity, Uc = 0.44 m/s. Uc was
obtained from Shield’s diagram (Miller et al., 1977). The temporal variation of scour was measured at the
After compaction and levelling of the sand, the point where the scour initiated (mostly the position
pump was started initially with a very low discharge of maximum scour-depth obtained). All the cases
so that the flume gets filled up with water without took almost 3-4 hours to reach the stage of equilibri-
any uncontrolled scour happening near or far from um.
the pier model. Once the flume was filled with the
required water depth, the discharge was increased to
the required value. Local scour around the pier initi-
ated immediately, which was carefully observed and
noted.

Figure 4. Comparison of temporal evolution of maximum scour

depths for the 3 cases.

Figure 4 shows the comparison between the de-

Figure 3. Photographs showing: (a). pile-cap buried in bed; (b). velopments of maximum scour-depth as function of
pile-cap partially buried in bed; (c). pile-cap resting on the bed. time for the 3 pile-cap elevations considered in the
study. As evident from the figure, the temporal vari-
For the purpose of evaluating scour and velocity ation of maximum scour-depth is extremely diverse
profiles around complex piers, 3 pile-cap positions for the 3 different cases. Here the maximum scour-
ing took place for case (a) (where pile-cap was bur- tively (arrow indicates flow direction). For case (b),
ied in the sand), whereas for case (b), absolutely no almost no scour was observed for the adopted flow
scouring was observed. conditions (Fig. 6).
It can also be observed that the maximum scour For case (a), the scour initiated at the downstream
took place in the initial 50 minutes only, and after half of the pier (near the tail of pier). Within a dura-
that, the scour started to stabilize, reaching equilibri- tion of 5 minutes, the scouring reached the nose of
um stage in almost 3-4 hours. In case (a), the scour the pier and subsequently the pile-cap also got ex-
depth increased rapidly in the first 30 minutes (until posed to the flow. After 15 minutes, the scouring
the pile-cap top got exposed to the flow ) and then it halted at the upstream region of the pier, owing to
increased at a much slower pace, whereas, in case the fact that the down flow cannot erode the exposed
(b), the scour depth increased gradually for almost 3 rigid pile-cap, but because of the flow separation and
hours and then it stopped. velocity increase at the pier sides, the scouring am-
plified at the sides of the pile-cap, thus giving rise to
3.2 Scour patterns and Equilibrium scour-depth maximum scouring at the sides of the pile-cap.
Scour depth reached equilibrium stage after 4 hours
From the observations made after stopping the flow, of run. Based on this observation, conclusion could
it was clear that the pile-cap elevation (with respect be made that when the pile-cap was submerged in
to sand bed) holds an important part in the formation the sand, after sometime, it gets exposed because of
of scour-hole around the pier. The maximum equi- erosion above it, and then the scour takes place ac-
librium scour-depths for each case and their location cording to the shape and overhang of the pile-cap af-
are presented in Table 2. terwards.

Table 2. Maximum scour depth obtained for the 3 cases and

their location of occurrence
Case Maximum scour depth Location of maximum
(cm) scour depth
a 1.67 Side of pile-cap
b 0 -
c 1.05 Tail of pile-cap

Figure 6. Photograph showing no scour in case b.

For case (b), no scour was observed at all around

the pier (Fig. 6). The partially buried pile-cap acted
as a scour countermeasure against bridge pier scour
Figure 5a. Measured scour contours at equilibrium conditions
here, which resulted in completely no scour around
for case a. the pile-cap. The elliptical shape of the pile-cap aid-
ed in the streamlining of the flow responsible for
less turbulence and hence no scour.
For case (c), the scour only took place mostly in
the downstream region of the pier. The scour in case
(c) was still more than that in case (b) and less than
that in case (a). So, it can be established that case (b)
is the most suitable pile-cap location in order to
avoid scour around a complex pier with elliptical
shaped pile-cap.

3.3 Mean velocity profiles

Figure 5b. Measured scour contours at equilibrium conditions

To facilitate the velocity measurements in stream-
for case c. wise and vertical directions, the ADV was mounted
on a movable trolley kept on the flume rails. The in-
Figures 5a, b shows the measured scour contours strument was connected to the PC for real-time visu-
at equilibrium conditions for case (a) and (c), respec-
alization of the velocities, using ExploreV software
(Nortek Product).
The raw velocity data collected during the exper-
iments using ADV were post-processed to remove
the spikes (noise) using phase space threshold de-
spiking method given by Goring and Nikora (2002).
The data was sorted by removing the errors with low
Signal to Noise ratio (SNR) (<15 dB) and low corre-
lation (<70%) samples. This way, about 3 to 5 % of
the raw data was excluded. The excluded data was
replaced by cubic polynomial interpolation tech- Figure 8. Stream-wise mean velocity profiles at the plane of
nique. The despiked data was then further processed symmetry upstream and downstream of pier for cases a, b and
to find out 3 dimensional mean velocity values from c, respectively.
the instantaneous data obtained from ADV.
The velocities were measured at six different lo-
cations on the centerline of the flume; three in the
upstream region of the pier and three in the down-
stream region of the pier, as shown in Figure 7.

Figure 9. Lateral mean velocity profiles at the plane of sym-

metry upstream and downstream of pier for cases a, b and c, re-
Figure 7. Locations of velocity measurements in the test sec- spectively.
tion.

The locations of measurements are as follows:

A 20 cm upstream of pier center

B 9 cm upstream of pier center
C 5 cm upstream of pier center
D 5 cm downstream of pier center
E 9 cm downstream of pier center
F 20 cm downstream of pier center

The vertical mean velocity profiles for all the 3 Figure 10. Vertical mean velocity profiles at the plane of sym-
metry upstream and downstream of pier for cases a, b and c, re-
cases at six different selected locations demonstrate spectively.
the characteristic flow features upstream and down-
stream of the pier. Beheshti and Ashtiani (2010)
gave velocity profiles around a complex pier struc- In case a, the stream-wise velocities (fig. 8(a)) on
ture with rectangular pier and pile-cap, but no study the upstream of the pier followed logarithmic law,
has been undertaken for other shapes of pier and but it was observed that the velocities decreased as
pile-caps for comparative study. the flow approached the pier. On the downstream
Figures 8, 9 and 10 shows the stream-wise (x- side of the pier, a backward flow was observed. At
direction), crosswise (y-direction) and vertical (z- 20 cm downstream of the pier, the velocity profiles
direction) velocity profiles for the 3 cases, measured seemed to regain back its original logarithmic na-
at afore-mentioned locations, respectively. ture. The lateral velocity profiles (Fig. 9 (a)) show
The effect of the exposed pile-cap is evident from very low flow velocity values far upstream of the
the obtained velocity profiles. The effect of the pier. Near the nose of the pier, the velocities reached
thickness of the pile-cap (i.e. 5 cm) is also clearly near zero velocities. But, at the downstream side of
visible in the velocity profiles. The varying velocity the pier, the lateral velocities show an oscillatory
zone has increased from 4-5 cm from case b to 6-7 behavior, contributed by the existence of the Horse
cm in case c. shoe Vortex severity on the downstream of pier.
Again, at 20 cm away from the pier tail, the veloci-
ties regained their original values. The vertical ve-
locity profiles (Fig. 10(a)) show that the velocity
values are more or less zero from a distance of 4 cm shore gravity based structure. Although, it will take
from the bed till near water surface. In the 4-5 cm greater time period and stronger currents to happen.
region near the bed, the velocity values show sudden
and random variation, and surprisingly this variation
isn’t limited to near field of the pier, at far field of 5 CONCLUSIONS
the pier also, this variation is observed.
In case b, the stream-wise velocity profiles (Fig. In this study, the effects of using an elliptical shaped
8(b)) followed logarithmic law away from the pier. pile-cap in reducing scour depth at various pile-cap
The velocities increased as they approached the pier. elevations with respect to the sand bed was investi-
On the downstream side, on the other hand, the ve- gated. The mean velocity data were examined to de-
locities are higher near to the bed as compared to the termine the relative importance of flow on scour
velocities above the pile-cap region but interestingly, depth as well as the effect of various components of
the vertical velocity profile (Fig. 10(b)) is almost the pier. Based on the experimental results, the fol-
constant with near zero velocity. It is assumed that lowing conclusions are drawn from the present
since the vertical velocity profile is almost non- study, as compared to existing literature:
changing, the scour initiating downward flow was • Use of elliptical shape of pile-cap can act as
non-existent because of the presence of the pile-cap. an effective way of reducing scour around a
Also the exposed sides of the pile-cap lead to pier.
smooth transition of the streamlines around it lead- • For pile-cap completely buried in the sand,
ing to reduction in the severity of the horse shoe vor- maximum scour depth was witnessed for the
tex mechanism, and hence no scouring in this case. given discharge.
In case c, the vertical velocities (Fig. 10(c)) show • Best elevation for the elliptical pile-cap to be
variations even on the far downstream of the pier, placed is partially buried in the sand. Null or
explaining the downstream scour of the pier. The very slight scour was observed for this eleva-
change in lateral velocities (Fig. 9(c)) is almost neg- tion. So, to dampen the vortex mechanism
ligible on the upstream side of the pier, though on around the pier, it seemed necessary for the
the downstream side, the velocities are found to be pile-cap to be placed partially or completely
fluctuating. But, this fluctuation was not strong over the sand bed.
enough to cause any erosion around the pier. No • When the pile-cap was resting on the sand
scour was observed on the front of the pile-cap in bed (completely exposed to the flow), there
case c, little scour on the sides of the pile-cap, and was slight scour around the pier, but interest-
comparatively greater scour on the downstream of ingly the scour occurred at mostly far down-
the pile-cap tail. stream of the pier.

4 DISCUSSIONS REFERENCES

The purpose of this experimental study was to find Akib, S. and Jahangirzadeh, A. (2014). Local scour around
out the scour pattern around the pile-cap when it is complex pier groups and combined piles at semi-integral
at different elevations with respect to the sand bed. bridge, J. Hydrol. Hydromech., 62(2): 108-116.
Ataie-Ashtiani, B., Baratian-Ghorghi, Z., and Beheshti, A. A.
Previous studies on complex piers are mostly on rec- (2010). Experimental investigation of clear-water local
tangular pile caps ignoring other shapes of the pile- scour of compound piers. J. Hydraulic Engineering, 6(343):
cap. The drive to use an elliptical pile-cap in this 343–351.
study was also to test its effectiveness against scour. Beheshti, A. A., and Ataie-Ashtiani, B. (2010). Experimental
Case b here resembles foundation caisson protruding study of three-dimensional flow field around a complex
above the original bed level for which Melville and bridge pier. J. Engineering Mechanics, 73: 143–154.
Beheshti, A. A., and Ataie-Ashtiani, B. (2016). Scour hole in-
Raudkivi (1996) stated that the scour depth gets in- fluence on turbulent flow field around complex bridge
creased if the caisson top level relative to the bed is piers. Flow Turbulence Combust., DOI 10.1007/s10494-
kept at higher level than the sand bed, because of the 016-9707-8.
fact that greater flow interferes with the pier as com- Chiew, Y. M., and Melville, B. W. (1987). Local scour around
pared to a single column. Our observations showed a bridge piers. J. Hydraulic Research, 25(1): 15–26.
Coleman, S. E., (2005). Clearwater scour at Complex Piers. J.
contradiction to the above stated results. But case b Hydraulic Engineering, 4: 330-334.
can also be compared to the effectiveness of collar Ferraro, D., Tafarojnoruz, A., Gaudio, R. and Cardoso, A. H.
studied by Sonneville et al. (2010). They found out (2013). Effects of pile cap thickness on the maximum scour
that a collar attached to a pier, placed on the bed depth at a complex pier. J. Hydraulic Engineering, 704:
level acts as an effective scour protection technique. 482–491.
The pile cap exposed to the flow (case b) resembles Goring D.G., Nikora V., (2002). Despiking acoustic Doppler
Velocimeter data. J Hydraulic Engineering, 128(1):117–
a similar configuration. Case c resembles an off- 126.
Hydraulic Eng. Circular (HEC-18) manual, (2001). Evaluating
Scour at Bridges. Federal Highway Administration
(FHWA), USA.
Melville, B.W., Raudkivi, A. J. (1996). Effect of foundation
geometry on bridge pier scour. J. Hydraulic Engineering,
4(203): 203–209.
Melville, B. W., Coleman, S.E. and Priestley, S., (2006). Local
scour at complex piers. World Environmental and Water
Resource Congress, ASCE.
Miller, M. C., McCave, I. N. and Komar, P. D., (1977).
Threshold of sediment motion under unidirectional cur-
rents. Sedimentology, 24: 507-527.
Moreno, M., Maia, R. and Couto, L., (2015). Effects of relative
column width and pile-cap elevation on local scour depth
around complex piers. J. Hydraulic Engineering, ASCE.
Shen H.W., Schneider, V.R. and Karaki, S.S. (1969), Local
scour around bridge piers. Proc ASCE, J Hydraulic Div.; 95
(HY6): 1919-1940.
Sonneville B.de, Rudolph, D., Raaijmakers, T.C., (2010),
Scour reduction by collars around offshore monopiles. Fifth
International Conference on Scour and Erosion, San Fran-
cisco, California, United States 2010.