INVESTIGATIONS RELATING TO SEEPAGE THROUGH A NATURAL SIDE EMBANKMENT FOR A RESERVOIR PROJECT

 

Heigerth Günther

 

Technical University Graz, Dep. of Hydraulic Structures

Stremayrgasse 10, A-8010 Graz / Austria

Tel.: 0043-316-873-8360, Fax: 0043-316-873-8357, E-mail: heigerth@kwb.tu-graz.ac.at

 

 

Abstract

Seepage through reservoir slopes to a neighbouring valley may involve serious problems. For a hydro project in southern Turkey (Client DSI/Ankara, Consultant chaired by Verbund-Plan/Vienna), it was necessary to investigate an unfavourably bedded flysh "side embankment" showing slide scars, which forms one of the reservoir slopes. A large number of in-situ and laboratory investigations were carried out to obtain information on the hydro-geological and geotechnical details. Using a mathematical model, underground flow patterns and pressures were analysed using the potential method, and the sliding stability of the downstream slope was analysed assuming a variety of different characteristic values and engineering measures. This led to a proposal providing for a grout curtain and a drainage curtain as well as a monitoring system, which would ensure the required safety and allowed adjustment during operation.

 

Keywords: Seepage, Hydro-geology, Slope stability, Potential method, Reservoir

embankment

 

1. Introduction

Among the main elements to be considered in the design of a dam are hydraulic questions such as seepage through the abutments and the appropriate engineering measures to be taken. In addition, depending on the topography and morphology of the reservoir area, seepage through the reservoir slopes to adjacent valleys may pose a major problem. Special attention must in this case be given to slope stability at the point of emergence of seepage flow. The following is a report on a project where seepage through a "natural" embankment adjacent to the abutment of a fill dam had to be studied with a view to developing measures for slope stabilisation and monitoring.

 

The Turkish overall development plan for central and eastern Anatolia in the early eighties included investigating the Seyhan river basin for potential utilisation. The master plan provided for 7 multi-purpose stations serving for irrigation, flood control, and power generation. The client was Turkish General Directorate of State Hydraulik Works DSI (Devlet su isleri); the design studies were carried out by a joint venture consisting of sponsor Verbund-Plan/Vienna, Romconsult/Bucharest, and Temelsu/Ankara. Meanwhile the lowest plant has been constructed.

 

 

Fig. 1 Plan view of Yedigöze project (showing planned measures)

 

One of the next plants to be realised is another storage development ("Yedigöze"): A more than 100m high rockfill dam with an earth core will create a reservoir with a total storage of 660 hm³ (see fig. 1, Plan view). The spillway, designed for a discharge of 8760 m³/s, will be situated on the right-hand bank. The left bank will accommodate the powerhouse with two tunnels with a discharge of 194m³/s each as well as the diversion tunnels, later to be converted to bottom outlets. The dam site has been found to afford optimal overall conditions in both topographical-morphological and geological-geotechnical terms. Situated on the outer side of a meander-like river loop that follows the lines of the main geological structures, the project has involved the need to study in some detail the mountain ridge enclosed by the loop as part of the reservoir boundaries. So, following the completion of the Feasibility Study, a work group under the direction of Prof. W.Schober from Innsbruck university (Austria), in which the author participated, prepared a study dealing with this problem.

 

2. Hydrogeological and Geological Situation, In-situ Investigations

The river valley has been carved with medium steep slopes some 150 m deep into a mid-Miocene flysch zone south of the Taurus mountains. The flysch is characterised by alternating layers of calcareous sandstone and clay-stone. The studied ridge is more than 2 km long and 750 to 1000 m wide at the base. It mainly consists of thick-bedded sandstone with a bed thickness of several metres, and intercalations of clay-stone. The bedding shows a fairly general parallel dip to the south; the strike is approximately parallel to the slope. The downstream slope exhibits major slide scars. The river valley itself is filled with alluvial material.

 

Following first geological mapping carried out by DSI, core drillings were sunk, mainly along a characteristic cross section, an exploratory gallery was driven, and water pressure tests were carried out. This gave the following picture: The slope shows evidence of extensive sliding at the surface. The slide scars decrease with depth, so as to pass into the intact bedding-plane slope of 8 to 12 %. Three levels were established where continuous clay-stone layers consitute the active or potential sliding planes. The investigations showed different permeabilities to exist in these sections, permeability parallel to the bedding planes being naturally higher than that vertical to the bedding planes. Further details completed the overall picture.

 

These findings suggested that reservoir filling might result in seepage through the slopes, mainly parallel to the bedding planes. Due to the presence of the less permeable clay layers, such seepage would risk to mobilise uplift forces likely to destabilise major slope portions and cause slides at the points of emergence of the seepage water. In order to obtain more details, rock mechanics in-situ and laboratory tests were performed (by DSI and Middle-East Technical University at Istanbul), in particular to establish residual shear strengths and residual shear angles (27° - 23°).

 

The relevant department at Innsbruck university, carried out complementary laboratory tests, which included permeability tests. The sandstone was not found to show any major anisotropy of permeability; the samples turned out to be surprisingly impermeable (K=10^-9 - 10^-10 m/s). Evaluation of the tests showed that the slide phenomena of the slope could not be explained by underground joints and fissures alone. The presence of fissure water pressure was suggested as a further main cause. In addition, earthquake effects had to be considered.

 

3. Possible Preventive Measures and Analytical Studies

In order to ensure the stability of the downstream slope after reservoir filling and, as far as possible, to improve the conditions as compared with the present situation, it is necessary to prevent the build-up of fissure water pressures in the critical area. For this purpose, we had to investigate several potential engineering measures and study and compare their effects in terms of slope stability by analytical means. Thanks to the parallelism of the geological situation it was possible to adopt a two-dimensional investigation programme to be carried out along what was selected as the most unfavourable slope cross section. For this purpose, the three above-mentioned series of beds were defined by different characteristic numbers.Therefore, it appeared useful to adjust a FEM grid to this geological structuring for both the flow and the slide studies. Different ratios, varying from 1 : 1 to 100 : 1, were assumed for the permeabilities parallel and vertical to the dip of strata.

 

For calculating steady seepage flow with a free groundwater surface, we used a program FEFLOW, assuming Darcy's law and complete saturation in the groundwater zone, which means that it was possible to apply the potential method. The results were plotted as flow nets (with potential lines and velocity vectors) as well as fields of pressure lines; figs. 2a and 2b are examples showing the result of one loading case analysis.

 

 

Fig. 2a Flow pattern -Orthotrope case (K-10:1) in 4 layers, with grout and drainage curtains

 

 

 

Fig. 2b Pressure lines - diagrammatic example showing layer structure and potential slip surfaces

 

For the slide studies, we defined for each case two slide masses of different depths with their failure surfaces naturally in a bedding plane (fig. 2b). The exterior loadings were taken as the fissure water pressures obtained from the flow analysis as well as- in additional loading cases - a quasi-static earthquake acceleration. According to the usual practice, we determined the minimum shear angle and defined the resistance to sliding as the quotient of residual shear strength to minimum shear strength. Assuming prototype conditions, however, the sliding resistance is definitely higher.

 

As preventive engineering measures we first suggested a grout curtain with a realistic permeability equal to the lowest natural permeability as well as a parallel drainage curtain in two different lengths. The following cases were finally studied:

 

- no engineering measures (reference case)

- grout curtain (impervious), drainage curtain, k =1 : 1 to 100 : 1

- grout curtain (partly pervious, k 50 times as large), k = 1 : 1 to 100 : 1

- grout curtain (partly ineffective), k = 100 : 1

 

The results were as follows:

In the first case, seepage flow, as expected, affects the areas vulnerable to sliding; so this case is unacceptable.

All the other cases show that the fissure water level downstream of the grout and drainage curtains is low enough to remain practically clear of the critical areas. As expected, it is necessary to provide a drainage curtain as a safety measure in case the grout curtain is only partly effective for technical reasons. The flow that would enter the drainage curtain has been calculated to range between about 14 and 37 l/sec x km, which would correspond to 88 - 93 % of the total seepage flow.

 

The calculated resistance to sliding is slightly above 1 for the case without preventive engineering measures, but is less than 1 on the assumption of added earthquake effects. For the suggested measures and a partly permeable grout curtain, we obtained 2.03, and 1.42 assuming earthquake effects, as most unfavourable values. On the extremely pessimistic assumptions, these values can safely be considered as sufficient.

 

4. Conclusions and Planned Preventive Measures

The engineering measures to be considered will necessarily be a combination of grout curtain and long drainage curtain. The system developed is shown in fig. 3:

 

 

Fig. 3 Typical profile - Grout and drainage system including monitoring system

 

·      Sink a grout curtain along the crest, determining the respective depths, by individual sections, from preceding investigations. Grouting pressures will be selected as a function of depth, no subsequent near-surface pressure grouting under surcharge being needed. This systematic grouting should naturally be supplemented by grouting as required for geological reasons.

·      Downstream, two tunnels will be provided from which a not more than two-level drainage curtain will be sunk. The two tunnels should be provided with a longitudinal slope and one access and drainage gallery each. The curtain, in particular with respect to borehole spacing, can be selected to suit the local conditions. Grout curtain and drainage curtain connect with the corresponding structures in the abutment of the fill dam and should extend sufficiently far into the area outside the region of the loop.

·      This scheme might prove redundant in the light of further details being obtained or in the course of the construction operations and should be adjusted, and thus possibly be reduced, as the work proceeds. Supplemental measures should be taken on the downstream slope surface

·      A measuring and monitoring system is planned to form an integral part of the scheme. This consists mainly of piezometers arranged at several cross sections and installed from the two tunnels. Furthermore, the scheme provides for drainage channels equipped with instruments measuring the entering seepage flows. Naturally, total flow has to be measured at the portal. At least those values should be recorded continuously and should possibly be included in an alarm system. Furthermore, the two access galleries offer the possibility of installing simple measuring equipment to detect potential sliding. This monitoring system can be supplemented as required by geodetic and other aboveground and underground measurements.

 

Measuring and monitoring will have to be particularly intensive during the phase of first filling. This will at the same time allow comparison with the analytical results and calculation of the actual seepage conditions.

 

A main idea in our concept has been the possibility of supplementing even in the long term the preventive measures as experience is gathered during operation. This mainly refers to the provision of additional drainage holes, which could probably be sunk without interference with reservoir operation.

 

By way of summary, it can be stated that a solution has been found to a fairly unusual problem, affording the required safety even if allowance is made for the variance of the natural parameters. Monitoring and the possibility of subsequent supplementary measures are to ensure that this safety is a lasting one.

 

Reference:

Republic of Turkey, General Directorate of State Hydraulic Works:

Yedigöze, dam and hydroelectric power plant project;

Special Study: Left bank seepage and landslide problems 1984 (not published).