Vous êtes sur la page 1sur 23

Sardar Vallabhbhai National Institute of Technology, Surat

Under the subject of

Environmental Geotechnology

Failure analysis of a geomembrane

lined reservoir embankment
Presented by
Aakash Kamthane P17SM003

Guided by
Prof. (Dr.) Satyajit Patel
About the paper

Riya Bhowmik, J.T. Shahu, Manoj Datta.
Department of Civil Engineering, IIT Delhi, Hauz Khas, New Delhi
110 016, India
Geotextiles and Geomembranes volume 46, issue 2, April
2018 page no. 52–65


• The paper aims to have a case study and failure analysis of

geomembrane of an embankment enclosing a raw water reservoir at
a coal based thermal power plant.
• To identify the type, degree and causes of the failure.
• To provide appropriate remedial measures to the existing structure
and to determine the precautions and set up guild lines for future thus
avoiding the such a failure from occurring again.


• A coal-based thermal power plant was set up in
the state of Punjab in India and the first unit
was commissioned in November 2013.
• The twin reservoirs, RWR-I and RWR-II, are
separated by a partition embankment.
• RWR–I had storage capacity of 2,077,800 m3,
whereas RWR–II was partly constructed.

Plan view of raw water reservoir

2) Site investigation
Based on borehole data, the soil profile at the site was as follows:
• Top 2–4 m thick layer consisted of loose to medium dense (bulk unit weight, γ= 18 kN/m3),
greyish brown, silty sand (SM) with an average SPT-N value of 10.
• This stratum was followed by medium dense to very dense (bulk unit weight, γ=20 kN/m3),
brownish grey, silty sand (SM) up to the maximum depth of boring of 50 m.
• The average N value for the lower strata varied from 10 at a depth of 2 m to 70 at a depth of
20 m.
• Water level was encountered at an average depth of 7 m


3) Embankments
• The partition embankment is 8 m high on both upstream side and downstream side.
• The peripheral embankments are 8 m high on the upstream side with height varying from
5 m to 8 m on the downstream.
• Both the embankments have an upstream slope of 2H:1V and a downstream slope of
• Both upstream and downstream slopes are provided with 1 m wide berm at 5 m below
the crest of the embankment.
• The embankments were constructed with the soil excavated from the site, which was
compacted to the optimum moisture content of 12%, and at 98% of Proctor density.


3) Downstream Slope Protection :

• The downstream slope of embankments were protected by two layers of 100 mm thick
graded and compacted filter material overlain by a single layer of 300 mm thick stone-

4) Internal Drainage :
• The central drain was trapezoidal in shape (0.3 m deep and 0.3 m wide at bottom), filled
with gravel–sized stones and had a longitudinal slope of 1:400.
• The central drain was surrounded by two 0.1 m wide layers of graded filter material.
• The sump of the central drain was located in the outlet chamber



Typical details of section of partition embankment


Typical details of section of peripheral embankment.


5) Liner :
• The liner on the upstream slope of
embankment consisted of 1 mm thick smooth
HDPE geomembrane overlain with
consecutive layers of 12 mm thick cement
mortar and 50 mm thick precast cement tiles
(Fig. a).
• The liner in reservoir bed consisted of 1 mm
thick smooth HDPE geomembrane overlain
with 300 mm thick soil cover (Fig. b).


• In October 2014, leakage was observed at two

locations on the downstream side of the partition
embankment at approximately 1 m above the
ground level.
• The leakages (Y and Z) were located at a
distance of approximately 100 m - 150 m from the
inlet duct corner of the partition embankment (X).
• The rate of leakage from these points was approx
20–40 L per second. `

Locations of initial leakage and schematic view of first trial to identify source of leakage. 11

Two probable sources of leakage were as follows:

(1) leakage from upstream side of the embankment
(2) leakage from the bottom of the reservoir.
• On the upstream side of the embankment and in the vicinity of locations Y and Z
(Fig. below), the tiles and the joints between the tiles were observed to be
satisfactory it was clear that the leakage was from the bottom.
• And it was confirmed with when the geomembrane at location A was found
punctured. This exposed the drain pipe beneath causing a cavity approx. of 1 m
dia to washing of fines (fig below).

• Two major breaches were observed in the partition embankment



• The reservoir was not completely empty after the breach and some water was still
present in the reservoir.
• During the site visit, it was observed that drops of water were still jumping/trickling
out of the floor at one spot in one of the breach locations (Fig. b).

• Tile slippage was also observed at few locations on the upstream side of the
embankment. Few tiles had lifted up on the berm (Fig.(a), (b)).
• No flow was reported at the sump of the central drain during the operation of the


The findings of the above site observations and review of design are
• The liner at the base of the reservoir was found torn in the vicinity of the pipe
drain adjacent to the location of the leakage and a large cavity was found
beneath the location of tear in the liner.
• The leakage stopped when the pipe drains located in the vicinity of the
leakage points were grouted. The reservoir could then be operated for nearly
seven months before the major failure occurred at a different portion of the
• Direct flow path was available from reservoir base to beneath the partition

• No flow at the outlet sump of central drain was observed at the time of
• No other source of infiltration of large quantity of water into the embankment
at very high pressure existed that could threaten the stability of

Seepage and stability analyses of
The stability analyses of the embankments were undertaken for the following
four possible cases of seepage conditions

• Both liner at base of the reservoir • Liner on the side slope of

and on slope of embankment are embankment was considered intact
intact and functioning. and functioning, while leakage was
• This is the best case scenario considered at base of the reservoir.
• Represents the condition for • This is the most probable reason
which the embankments were for the major breach in the partition
originally designed. embankment. 18

• Both liners were considered leaking, • All liners were considered leaking
but no flow pathway (representing and additionally, a flow pathway
pipe drains) was modelled from the (representing pipe drains) was
base of the reservoir to the central modelled from the leakage point to
drain beneath the embankment. the defunct central drain beneath
• This case is also relevant in case of the partition embankment.
the partition embankment for the • This case pertains to the worst case
situation wherein all pipe drains are scenario for the partition
grouted and blocked. embankment.


Remedial measures involved strengthening of the embankment

1. to make it stable for the representative seepage condition of leaked liner.
2. to control seepage and piping in embankment.


For reducing leakage and To strengthen the

precautions against piping embankment

Schematic diagram of the remediated section of the partition embankment


• It is important that a secondary liner in form of clay liner or geosynthetic clay liner
should be provided. As n case underdrains are provided beneath the liner
because a secondary liner would prevent development of high flow seepage paths
through a tear in geomembrane liner to the slotted pipes.
• if provision of single layer of geomembrane as the only barrier layer is
unavoidable, then the seepage and slope stability analysis may be carried out
with the assumption that the geomembrane barrier layer may leak.


• Adequate internal drainage should be provided even in a small embankment, if the

embankment is made up of semi-pervious fill material. Absence of internal
drainage, like vertical/inclined filter, horizontal filter and rock toe, may result in
high seepage flow in the embankment and cause erosion of fill material.

• Field and construction quality control for geomembrane should be adopted to

minimise the frequency of holes, tears and punctures.

• Use of tiles over geomembrane should be used as a protection measure for

geomembrane with care. The primary reason being bonding between tiles and
smooth geomembrane is generally not satisfactory. Moreover, laying of tiles over
geomembrane may induce damage to the geomembrane, and thus jeopardise the
safety of the embankment.