SAN ROQUE MULTIPURPOSE PROJECT PERFORMANCE

21st Century Dam Design — Advances and Adaptations

31st Annual USSD Conference San Diego, California, April 11-15, 2011

Hosted by Black & Veatch Corporation GEI Consultants, Inc. Kleinfelder, Inc. MWH Americas, Inc. Parsons Water and Infrastructure Inc. URS Corporation

On the Cover Artist's rendition of San Vicente Dam after completion of the dam raise project to increase local storage and provide a more flexible conveyance system for use during emergencies such as earthquakes that could curtail the region’s imported water supplies. The supplies. The existing 220-foot-high dam, owned by the City of San Diego, will be raised by 117 feet to increase reservoir storage capacity by 152,000 acre-feet. The project will be the tallest dam raise in the United States and tallest roller compacted concrete dam raise in the world.

U.S. Society on Dams Vision

To be the nation's leading organization organization of professionals professionals dedicated dedicated to advanci advancing ng the role of dams for the benefit of society. Mission — USSD USSD is dedicated to:

• Advancing Advancing the knowledge of dam engineering, engineering, construction, construction, planning, planning, operation, operation, performance, rehabilitation, decommissioning, dec ommissioning, maintenance, security and a nd safety; • Fostering Fostering dam technology for socially, socially, environmentally environmentally and financially financially sustainable sustainable water resources systems; • Providing Providing public awareness awareness of the role of dams in the management management of the nation's nation's water resources; • Enhanci Enhancing ng practices practices to meet current and future future challenges challenges on dams; and • Representing Representing the United United States as an active member of the International International Commission Commission on Large Dams (ICOLD).

The information contained in this publication regarding commercial projects or firms may not be used for advertising or promotional purposes and may not be construed as an endorsement of any product or from by the United States Society on Dams. USSD accepts no responsibility for the statements made or the opinions expressed in this publication. Copyright © 2011 U.S. Society Society on Dams Printed in the United States of America Library of Congress Control Number: 2011924673 ISBN 978-1-884575-52-5 U.S. Society on Dams 1616 Seventeenth Street, #483 Denver, CO 80202 Telephone: 303-628-5430 Fax: 303-628-5431 E-mail: stephens@ussdams. [email protected] org Internet: www.ussdams. www.ussdams.org org

On the Cover Artist's rendition of San Vicente Dam after completion of the dam raise project to increase local storage and provide a more flexible conveyance system for use during emergencies such as earthquakes that could curtail the region’s imported water supplies. The supplies. The existing 220-foot-high dam, owned by the City of San Diego, will be raised by 117 feet to increase reservoir storage capacity by 152,000 acre-feet. The project will be the tallest dam raise in the United States and tallest roller compacted concrete dam raise in the world.

U.S. Society on Dams Vision

To be the nation's leading organization organization of professionals professionals dedicated dedicated to advanci advancing ng the role of dams for the benefit of society. Mission — USSD USSD is dedicated to:

• Advancing Advancing the knowledge of dam engineering, engineering, construction, construction, planning, planning, operation, operation, performance, rehabilitation, decommissioning, dec ommissioning, maintenance, security and a nd safety; • Fostering Fostering dam technology for socially, socially, environmentally environmentally and financially financially sustainable sustainable water resources systems; • Providing Providing public awareness awareness of the role of dams in the management management of the nation's nation's water resources; • Enhanci Enhancing ng practices practices to meet current and future future challenges challenges on dams; and • Representing Representing the United United States as an active member of the International International Commission Commission on Large Dams (ICOLD).

The information contained in this publication regarding commercial projects or firms may not be used for advertising or promotional purposes and may not be construed as an endorsement of any product or from by the United States Society on Dams. USSD accepts no responsibility for the statements made or the opinions expressed in this publication. Copyright © 2011 U.S. Society Society on Dams Printed in the United States of America Library of Congress Control Number: 2011924673 ISBN 978-1-884575-52-5 U.S. Society on Dams 1616 Seventeenth Street, #483 Denver, CO 80202 Telephone: 303-628-5430 Fax: 303-628-5431 E-mail: stephens@ussdams. [email protected] org Internet: www.ussdams. www.ussdams.org org

SAN ROQUE MULTIPURPOSE PROJECT PERFORMANCE MONITORING ASSESSMENT

Michael Pavone, P.E.1 Joseph Ehasz, P.E.2 Stephen Benson, PE.3 Bonnie Witek, L.G., L.E.G.4 ABSTRACT

The San Roque Multipurpose Project (SRMP) is a major hydroelectric and flood-control project in Asia. The 200-meter-high, central clay core, rock-fill dam is the 12th highest dam of its kind in the world. It is located on the Agno River in the Philippines and impounds a reservoir with a surface area of about 12.8 square kilometers that provides flood attenuation benefits downstream of the dam. The SRMP has an installed rated capacity of 411 megawatts. URS was awarded two contracts totaling $705 million for the engineer-procure-construct (EPC) work by San Roque Power Corporation. Performance monitoring of the project began prior to first filling of the reservoir reservoir and continues on a regular schedule. Key design requirements included stringent leakage criteria and reliability in this seismically active region. The performance monitoring program for San Roque Dam includes instrumentation monitoring and routine visual inspections. The program is intended to provide verification of design parameters, analyze adverse effects, verify performance, and identify any potential safety concerns. Instrumentation includes: 1) piezometers, 2) movement survey monuments, 3) seepage flow measurement stations, 4) rainfall gauge, and 5) turbidity meters. This paper presents the evaluation of monitoring program data with regards to pore water pressures, seepage, and deformation considering the designer’s (URS’) prediction of performance. Performance prediction is generally tied to the appropriate factors of safety which, along with other design parameters such as calculated deformations, seepage flows and piezometric pressures, determine the desired threshold limits for the design conditions. INTRODUCTION

Located in the mountains 200 kilometers north of the capital of Manila in the Philippines, the SRMP is a major hydroelectric and flood-control project in Asia. Its major feature is a 1

Manager of Engineering, URS Energy and Construction, Bellevue, Washington, [email protected] Vice President, URS Energy and Construction, Bellevue, Washington, [email protected] 3 Manager of Geotechnical Engineering, URS Energy and Construction, Bellevue, Washington, [email protected] 4 Senior Engineering Geologist, URS Energy and Construction, Bellevue, Washington, [email protected] 2

Performance Monitoring Assessment

1479

200 meter high central core, rock-fill dam which is the 12th highest dam of its kind in the world. Often called a national flagship project for the Philippines, in addition to providing flood control and irrigation, the SRMP supplies clean hydroelectric power for commercial and industrial use to a region that desperately needs it. In 1998, URS was awarded contracts totaling $705 million for the engineer-procure-construct (EPC) work by San Roque Power Corporation, an international consortium led by United States-based Sithe Energies, and presently under the joint venture of Marubeni and Kansai Electric. San Roque Power worked under a build-operate-transfer contract with the state-run National Power Corporation. The dam, spillway and powerhouse and appurtenant structures were constructed entirely by URS owned equipment and staff composed of 80 expatriate employees and as many as 4,000 local Pilipino workers. The Project was designed by URS Office staffs in New York and Bellevue, Washington, as well as local site support staff at the Project Site. Major project features include: •

•

•

•

•

• •

An earth and rockfill dam, 200 meters high. The dam consists of 41 million cubic meters of combined gravel fill and rockfill shell zones, filter, drain and transition zones and an impermeable core. A concrete spillway with six 15-meter-wide by 18.6-meter-high radial gates, and a 485-meter-long, 100-meter-wide concrete chute ending in a flip bucket. The spillway is designed to pass 12,800 cms. A power tunnel, 8.5 meters in diameter and 1,300 meters long which includes a pair of wheel gates in a shaft near the intake to shut off flow, a surge shaft for control of hydraulic gradients and a steel-lined high-pressure segment. A low-level outlet tunnel, 5.5 meters in diameter, 1,300 meters long with an intake and flow control by a set of slide gates near the dam centerline. Three diversion tunnels, two which are 10 meters wide by 15 meters high and one which is 6 meters wide by 6 meters high. Below ground powerhouse with three 137 MW turbine generators. Electrical substation and nine kilometer long 230 kW transmission line.

In addition to the dam’s production of non-polluting hydroelectric power, its five-squaremile reservoir serves as a settling basin that entraps sediment above the dam, thus improving water quality below the dam. These project features are shown on the accompanying photograph.

1480


PROJECT DESCRIPTION General Geology

The San Roque Multipurpose Project is located on the southern flank of the Central Cordillera, the mountain highlands that make up the northwestern part of the island of Luzon. The project takes advantage of the major change in topography at the southern edge of the Cordillera where the Agno River flows from a steep, narrow-walled canyon in the mountains out into the relatively flat Central Valley basin of Pangasinan Province. The basement rocks of the southern part the Central Cordillera include pre-Tertiary phyllite, schist, plutonic rocks, pillow basalt, chert, and a variety of clastic rocks. The specific lithologies and deformational history of this sequence varies widely throughout the region. Overlying the basement complex are probable Eocene to lower Miocene volcanic rocks composed of andesite flows, basalt flows, and breccias and other pyroclastic rocks intercalated with chert, argillite, sandstone, and conglomerate, known as the Pugo Formation. Both of these rock sequences have been intruded by Tertiary plutons, generally middle Oligocene to late Miocene. Deformation associated with the intrusives has locally deformed the Pugo Formation. A thick, early Miocene to Pleistocene, sedimentary section with some volcanic components is draped over the southern and western flanks of the Central Cordillera and is present as the Klondyke Formation in limited areas at the project. A number of early Quaternary intrusive rocks are associated with the extensive mineralization that has been exploited in the Baguio district in the upper reaches of the reservoir.


1481

Most of the major project structures (dam, spillway, and most of the tunnels) are founded on or excavated in metamorphosed volcanic, sedimentary, and volcaniclastic rocks, probably equivalent to the Pugo Formation. Various diorite intrusions of the Agno Batholith are also present, especially in the powerhouse area, on the left bank of the river at its last major bend between the dam and the powerhouse, and in the spillway chute area. The Klondyke Formation forms a sedimentary sequence of conglomerates, sandstones, and claystones that thickens down dip south and east of the dam. Bedrock at the dam is primarily volcanic and volcaniclastic rock, which has been subject to low grade metamorphism and intruded by diorite. The dominant foundation rock type is metamorphosed volcanic breccia, with minor intercalations of fine-grained metavolcanic/metasedimentary rocks. Numerous joints are infilled with calcite, quartz, and other minerals. Pleistocene to Recent alluvial channel deposits of the Agno River are found in the steep narrow-walled canyon where the dam is located as well as in the broad valley downstream. Alluvial terrace deposits of the Agno River and alluvial fan deposits from tributary drainages are also present along the sides of the riverbed in some areas. Landslide deposits are also widespread upstream in reservoir and along the front of the Central Cordillera, most notably east of the powerhouse. Abutments

The abutments of the dam are primarily underlain by volcanic breccia with minor amounts of metavolcanic/metasedimentary rocks. The diorite intrusion crops out at the surface in only a few places, but is present in the subsurface of the left abutment and has probably contributed to the alteration of the volcanic breccia by contact metamorphism or the source of hydrothermal fluids. Both abutments had a significant thickness of overburden, consisting of soil and completely weathered rock. The overburden thickness ranged from more than 20 meters on the upper left abutment to zero where slightly weathered rock crops out along the left bank of the Agno River. In general, the overburden was thicker on the left abutment than the right abutment, where soil and completely weathered rock was generally 5 to 15 meters thick. Below the overburden, the volcanic breccia of the dam abutments generally becomes highly weathered. The rock has significant cohesion. Joints and other structures are measurable in the rock. They are generally tight but can be open with or without filling. Although weathering intensity generally decreases with depth, the weathering profile is highly variable. Weathering appears to be controlled by permeable fractures and zones of relatively competent rock are underlain by moderately to highly weathered material. The crystalline bedrock at the dam site exhibits varying degrees of fracturing as a result of the stresses associated with tectonic history. In general, the meta-volcanic and metasedimentary rock and the volcanic breccia are more highly fractured than the diorite unit. In most locations, more than 3 joint sets can be observed. Joint spacing for each joint set varies from about 2 to 20 cm. The most highly fractured areas are often at the margins of

1482


the contact with diorite intrusions. A number of shear zones are present on the abutments as observed both in borings and in the excavation. These zones are characterized by crushed rock and clay gouge that may vary in thickness from less than a centimeter to several tens of meters. Almost every core boring in the dam footprint encountered sheared zones. Most of the shear zones observed in the dam excavation are less than two meters wide. Some shear zones are associated with diorite intrusions. Slickensides were observed in many of the shear zones. The presence of calcite veins and other mineral infillings in many of the shear zones indicates that these are not geologically recent features. The combination of the above poor rock conditions made foundation treatment very difficult. The design-build nature of the contract and project enabled the close coordination of design and construction to meet the foundation treatment objectives. River Channel

The riverbed area is generally covered by various alluvial deposits. Most of the alluvium at the dam site consists of Agno River channel deposits of sand, gravel, cobbles, and boulders. Older channel alluvium can be distinguished from younger more recent deposits, generally on the basis of density. The deposits vary in thickness from 4 to 15 meters, but average close to 15 meters in the main river channel under the dam. Bedrock beneath the alluvium is similar to that found on the abutments, primarily volcanic breccia with lesser amounts of meta-volcanic/meta-sedimentary rock and diorite. Unlike the abutments, the weathering profile was much shallower, with residual soil and completely weathered rock being absent, and a rather quick transition into moderately to slightly weathered rock. Like the abutments, there are numerous joints and shear zones in the bedrock mass, however several features are especially distinctive. A narrow deeply incised old river channel, eroded into the bedrock, crosses the main dam axis. Two major shear zones, one approximately 20 meters wide and the other about 30 meters also cross the dam axis. These shear zones are characterized by alternating zones of clay gouge, crushed rock, and highly fractured rock. Where intercepted, these shear zones are associated with significant water inflows. Foundation Excavation Criteria

All areas of the core trench are founded on rock. The depth of excavation to the final surface were based on the assessment of rock quality and consisted of two requirements: (1) the rock must be treatable by grouting and be hard enough to allow setting of packers for grouting and (2) where geologic features (joints, fissures, or shear zones), are encountered, the features must be tight and characterized by a general absence of internal erosion of erodible material (piping). A method specification based on equipment performance was developed to achieve the above design objectives. Based on the results of field trials, the limit of core zone excavation is specified as rock of hardness and structure such that the material can no


1483

longer be effectively removed by a three-tooth ripper of a Caterpillar D-9 tractor dozer equipped with a ripper of standard manufacturer’s design, working on a production basis, and operating in low gear. The design of core zone foundation required shaping of the rock to provide a generally uniform foundation surface that improves long-term performance of the dam by reducing the potential for differential settlement, cracking, and seepage paths. The design of the shell zone foundations required removal of any unsuitable materials to a level where the foundation materials have strength and elastic characteristics equal to or greater than the overlying shell zone materials. The acceptance criterion for shell zone foundation on rock was based on equipment performance and verification of rock quality conditions. The limit of shell zone excavation in rock was specified as excavation to a rock surface such that materials can no longer be effectively removed by the blade of a Caterpillar D-7 tractor dozer equipped with a blade of standard manufacturer’s design, working on a production basis, and operating in low gear. Alluvium at the dam site consisted mostly of gravel and cobbles, with lesser amount of sand and boulders. Older alluvium at depth was distinguished from younger alluvium near the surface on the basis of density. Older alluvium is very dense and can be excavated in near-vertical cuts. Correlation of shear wave velocity with the potential for soil liquefaction and the results of large-scale density and gradation test were used as a basis for determining the thickness and extent of alluvium to be removed and replaced with compacted fill for the foundations of the shell zones of the dam. Grouting Program and Foundation Preparation

The dam foundation grouting program included a combination of single and double line grout curtain, consolidation grouting, and stitch grouting. Grouting was accomplished from both the surface and the grouting galleries (Figures 8 and 9) in the foundation rock. Curtain grouting consisted of a row of grout holes near the main dam axis, designed to decrease seepage through the foundation rock and abutments. The grout curtain is a plane inclined 70 degrees upstream to improve its intersection with vertical joints and shear zones. In general, the grout curtain extends 80 meters below the core trench foundation. On the upper abutments, the depth is gradually reduced to 40 meters as the reservoir head is less in these areas. Because of water inflows into the gallery associated with the shear zones, a double curtain was installed between stations 7+95 and 8+95 from the lower gallery. Consolidation grouting consists of a grid of relatively shallow grout holes in the main dam core trench foundation to decrease the permeability of surficial fractured zones adjacent to the grout curtain and reduce the potential for embankment materials to be carried into the foundation rock mass (piping). Consolidation grouting in the grouting galleries was done to repair rock that may have been opened by excavation and to reduce seepage flow into the galleries on the upstream (grout curtain) side.

1484


Stitch grouting was used for treatment of features such as a shear or fracture zone in a pattern of holes designed to intersect that feature at various depths at angles favorable to grout penetration. Stitch grouting was done from the surface and extends the full width of the core trench if the feature is a significant seepage pathway. Stitch grouting was performed for the full core trench width in the 20 and 30 meter wide shear zones in the riverbed portion of the core trench. Embankment Section — As Constructed

The general arrangement and dimensions of the zones within the embankment dam are shown on the typical section on Figure 1. The maximum height of the dam from the core trench to the crest is 200 meters. Both the upstream and downstream slopes are 2H:1V. The crest width is 12 meters and the width of the core at the top is 6 meters. The central impervious core has an upstream and downstream slopes of 0.2H:1.0V.

Figure 1. Idealized Section on Alluvial Foundation

• • •

• • • •

Zone 1 Zone 1A Zone 1B Zone 2 Zone 3 Zone 4 Zone 5

•

Zone 6 Zone 7A

•

Zone 7B

•

Upper Core Material ( PI ≥15) Base of Core (PI > 20) Core - Material obtained from overburden excavation (% fines> 20) Transition Filter Drain Clean Shell (Processed from alluvial borrow or rock quarry areas – 0-30% passing No. 4 sieve) Shell (0-50% passing No. 4 sieve) Random Rockfill (obtained from rock excavation – 0-60% passing No. 4 sieve) Select Rockfill (obtained from rock excavation – 0-30% passing No. 4 sieve)


1485

• • • •

Zone 8 Zone 5/7B Zone 6/7B Zone 6/7A

Riprap Shell (co-mingled with alluvium and rockfill) Shell (co-mingled with alluvium and rockfill) Shell (co-mingled with alluvium and rockfill)

The zone arrangement depicted on Figure 1, evolved as construction progressed in order to make maximum utilization of the required excavation material from the dam foundation and spillway. A plan view of the embankment and project is shown on Figure 2.

Figure 2. Plan view of the SRMP INSTRUMENTATION

The instrumentation monitoring and periodic visual inspections taken collectively provided an overview of dam performance during construction, first reservoir filling, and long-term operation. The main objectives of instrumentation monitoring and visual inspections are: verify project performance, identify any potential safety concern, analytical assessment and prediction of future performance. Within the a nalytical assessment, instrumentation data can verify design assumptions and construction techniques as well as analyze adverse events that may occur during the construction of the dam. Prediction of future performance is associated with deviations of data trends and identification of unusual data. To perform these verifications and analyses, various instruments were placed at specific locations of the structure during construction. These instruments measure: 1486


• • • • •

pore-water pressures within the dam and its foundation seepage deformation precipitation turbidity

Most of the embankment piezometers and settlement cells are located at five different elevations along four lines perpendicular to the dam axis. The four lines are at dam station 5+00, 7+00, 9+25, and 11+00. Station 9+25 is near the maximum section of the dam. Foundation piezometers are located in the grouting galleries underneath the dam in holes oriented upstream and downstream of the grout curtain. Additional abutment piezometers are located in the gallery access adits (Figure 9). Flow measurement stations are located in the galleries, adits, and at the downstream collector pipe near the toe of the dam. Movement survey monuments are located on the surface of the dam. A plan view and cross-sections of the instrument locations are shown on Figures 3 through 7.


1487

Figure 3. Plan View of Instrumentation

1488


Figure 4. Instrumentation Section – Station 5+00




1489

Figure 7. Instrumentation Section – Station 11+00 Pore Water Pressure

Piezometers are used to monitor pore water pressure and free water surfaces within the dam foundation and abutments and to monitor uplift pressure beneath the spillway ogee section. Pore water pressure measurements are necessary to determine the phreatic surface in the embankment and thus evaluate slope stability, evaluate design assumptions, and seepage or other changes which could indicate changing conditions within the core or fill. During early reservoir filling, piezometer readings were used to verify phreatic assumptions used in the analyses, evaluate pore pressure dissipation across a given section of the dam and to assess the effectiveness of the impervious core and foundation grouting. Four types of piezometers were used for the project in order to increase reliability of results. These include: •

Vibrating Wire Piezometers

The vibrating wire (VW) piezometer is an accurate piezometer that is relatively easy to install and monitor. The VW piezometer converts water pressure into a frequency signal via a diaphragm, a tensioned steel wire, and an electro-magnetic coil. The instrument is designed so that a change in pressure on the diaphragm causes a change in tension of the wire. When excited by the electro-magnetic coil, the wire vibrates at its natural frequency. The vibration of the wire in the proximity of the magnetic coil generates a frequency signal that is transmitted to the readout device. The readout device processes the signal, applies calibration factors, and displays a read ing in the required engineering unit. A total of 27 VW piezometers were installed at various levels in the dam. Twentyfive additional VW piezometers were installed in the up stream holes drilled from the galleries. •

Pneumatic Piezometers

Pneumatic piezometers were used as an alternative measuring method to the VW piezometers and, as a back-up in the event that a VW piezometer(s) were damaged. The

1490


advantage of the pneumatic piezometer is that it uses a simple diaphragm with no electrical parts and therefore is not susceptible to lightening damage. In order to enhance long-term performance, corrosion-resistant plastic construction, polyethylene tubing, and in-line filters were used at all connectors. A total of 7 pneumatic piezometers were included in the dam. •

Open Standpipe Piezometers

Open standpipe piezometers were installed along the downstream toe of the dam to measure pore water pressure in the abutments. Four open standpipe piezometers were installed at locations shown on Figure 3. The open standpipe piezometers included a drill hole approximately 15 meters deep, 75 mm diameter and having a PVC pipe extending approximately 0.5 meters above grade. The PVC pipe extends into and was grouted into the hole. •

Pressure Gage Piezometers

Pressure gages piezometers were installed at three locations in the dam and spillway drainage galleries. These pressure gage piezometers were primarily oriented downstream of the gallery. A section view showing location of the gallery piezometers is shown on Figure 8.

Figure 8. Gallery Piezometers Seepage

The amount of seepage migrating through, under, and around the embankment is necessary information as it relates to the internal stability of the structure and the verification that the dam drainage zones are functioning properly. Monitoring seepage is essential during construction, first filling, and in establishing long-term trends. Seepage is measured by means of weirs, volume vaults or wet wells (calibrated containers). A system of flow measuring weirs was established to collect and measure seepage and leakage that passes the dam. Leakage through the dam’s grout curtain and rock foundation is collected by the foundation drainage system. The foundation drainage Performance Monitoring Assessment

1491

system consists of drain holes drilled from within the grouting galleries under the dam. The galleries run primarily along the centerline of the dam. Water collected in the galleries and access adits flow by gravity to flow measuring stations. Several flow measurement stations are provided within the galleries to isolate and measure the quantity and locations of leakage at specific points along the gallery. The locations of the flow measurement stations are shown on Figure 9. Seepage through the dam is collected along the inclined drain zone adjacent to the core and collected by the blanket drain beneath the downstream shell. This flow is conveyed to a buried collection pipe system just beyond the toe of the dam where it is measured by a weir. The project’s drainage collection system was sized to accommodate a total of 320 liters / second (l/s), which includes an allowance for rainfall infiltration, natural groundwater, and backflow from tailwater. One hundred-fifty l/s was assumed to flow into the galleries and adits and 170 l/s into the embankment drain zones.

Figure 9. Flow Measurement Stations in Galleries Deformation

Surface deformation includes horizontal movement, settlement and heave of the embankment. To monitor these changes, Movement survey monuments (MSM) are installed on the dam crest, downstream slope, upstream slope, and on the spillway ogee section. The monuments are surveyed using conventional survey equipment to measure overall vertical and horizontal movements of the dam. A total of 67 MSM were installed. Ten are located on the dam crest, 35 are located on the downstream slope, 3 on the spillway structure piers, and 19 are installed on the upstream slope. In addition, a settlement monument was installed immediately in front of each monitoring station. Locations of movement survey monuments are show in Figure 10.

1492


Figure 10. Location of Survey Monuments Vibrating wire (VW) settlement cells were installed in the core and downstream shell zones zone to monitor settlement during construction and first filing. A total of 14 vibrating wire settlement cells were installed in the downstream half of the impervious core and downstream shell fill at the locations shown on Figures 3 through 7.


1493

Rainfall Measurement Devices

Rain gages were installed in order to measure rainfall in the immediate vicinity of the project structures in order to correlate rainfall infiltration with measured drainage gallery flows and the flows measured at the blanket drain weir beyond the toe of the dam. Baseline data and correlations between precipitation and drain response were developed prior to the initiation of reservoir filling operations in order to estimate flows resulting from precipitation and infiltration versus those resulting from seepage and leakage past the dam. Turbidity Measurement Devices

The purpose of measuring turbidity of the seepage water collected in the dam’s drainage system is to reveal if there is any piping or loss of material through the drain system taking place. A turbidity meter measures turbidity in nephelometric turbidity units (NTU) that can be generally correlated to the dissolved and suspended sediment concentration in a water sample. Nephelometric refers to the way the instrument, a nephelometer, or turbidity meter, measures how much light is scattered by suspended particles in the water. Sampling is conducted at the various seepage and leakage measurement stations located in the galleries and adits, and at the collection and measurement structure for the horizontal drain of the main dam. Similarly, the turbidity meter can be used in the sampling and measuring of turbidity at any location along the river. Low NTU values indicate high water clarity, while high NTU units indicate low water clarity. Leakage flow is sampled and measured for turbidity at the various weirs located in the adits and galleries and at the blanket drain collection location PREDICTED PERFORMANCE Pore Water Pressure

Pore water pressures for design of the dam were computed from the location of the theoretical piezometric surface, which varied for the various loading conditions considered. The piezometric surface for the normal operation steady-state seepage condition was conservatively assumed to be a horizontal line at elevation 280 m in the upstream rockfill shell and the core to a point 3 m beyond the core in the Zone 3 filter. From that point, the phreatic surface was assumed to be parallel to the downstream face of the core extending to the Zone 4 blanket drain. Beyond the limits of the shell, the piezometric line was assumed at elevation 108 m at the maximum downstream section. The downstream phreatic line for the maximum upstream section was assumed at the base of the embankment on the excavation surface. The results of steady-state seepage finite element analyses were used to verify the locations of the phreatic surfaces described above. In general, the pore pressures indicated by the finite element analyses are lower than the values assumed for the design phreatic surface. Therefore, the stability analyses results were based on conservative pore pressures. Piezometric head levels

1494


predicted by design computations for steady state conditions for piezometers located in the maximum section of the dam at Station 9+25 are summarized in Table 1. Monitoring data from these piezometers are discussed below in Evaluation and Performance. Table 1. Maximum Section (Station 9+25) Piezometric Design Levels Piezometer No.

Location

VWP - 11 VWP - 12 VWP - 13 VWP - 14 VWP - 15 VWP - 16 VWP - 17 VWP - 18 VWP - 19 VWP - 20 VWP - 21 VWP - 22

Core Core Core Core Core Shell Shell Core Core Shell Shell Shell

Distance from Axis (m) 5 8 20 10 30 80 175 10 40 125 225 330

Tip Elevation (m)

Design Head (m)

240.0 184.6 184.5 131.3 131.4 130.4 131.1 102.5 103.4 106.7 106.5 105.8

280 280 224 274 174 108 108 274 124 108 108 108

Seepage

Seepage analyses were performed using the two-dimensional finite element computer program SEEP/W (Geo-Slope International Ltd., 1998). Numerous analyses were performed as the design and construction was advanced. Two-dimensional finite element analysis models were developed for several selected transverse sections along the length of the dam. Total seepage through the dam and foundation was obtained by using the end-area method to extrapolate two-dimensional results at each cross section and summing them along the axis of the dam. For the initial analysis, four sections were selected to represent the dam, the arrangement of the drainage gallery, variations in foundation excavation geometry and foundation stratigraphy. Permeability values for various zones within the bedrock foundation were initially estimated based on field permeability testing of borings, performed as pa rt of the foundation core drilling exploration program at selected locations within the main dam footprint area. With the completion of the construction of the dam, but prior to reservoir filling, the designers availed themselves to as-built excavation topog raphy and extensive water pressure test data obtained from Primary (P) and Verification (V) grout ho les to better model pre- and post-grouted rock permeabilities, respectively. Therefore, as the design and construction advanced, the number of representative sections was increased, ultimately up to fifteen, to accommodate the actual foundation excavation geometry, expanded subsurface information obtained from foundation mapping, water pressure test results from foundation grout holes, and changes to the embankment cross section and drainage gallery arrangement. The seepage prediction analysis results described herein are for the final post-construction, pre-filling “as-built” conditions.


1495

The permeability values, calculated from the field exploration permeability tests and P and V grout hole pressure testing were plotted, by dam station, versus depth in order to evaluate the grouted and ungrouted rock permeability values representative of each analysis cross section. The permeability values, calculated from the field permeability tests performed within each section, were evaluated versus depth and representative values for analysis were selected for ungrouted rock. The anisotrophy ratio was assumed to be 1.0 for all rock material. The results are summarized in the following Table 2. Table 2. Representative Permeability Values for Ungrouted Foundation Materials Station 2+50 to 6+00

6+00 to 8+12.5

8+12.5 to 9+00

9+00 to 10+37.5 10+37.5 to 13+80

13+80 to 15+00

Depths (m) 0 - 20 20 – 60 60 – 110 >110 0 – 20 20 – 35 >35 0 - 30 30 – 50 >50 0 – 30 >30 0 – 30 30 – 75 >75 0 – 30 30 – 75 >75

Permeability (cm/sec) 1x10 -3 3x10-4 1x10-4 2x10 -5 3x10 -4 2x10-4 8x10-5 6x10 -4 1x10-4 5x10-5 2x10 -4 5x10-5 3x10 -4 2x10-4 4x10-5 3x10 -4 2x10-4 4x10-5

Again, it should be reiterated that these are the values used in the final “as-built” analysis. Leading up to this analysis were numerous other analyses that also examined lower and upper bound ungrouted rock permeability ranges versus depth. The original permeability (hydraulic conductivity) goals for the consolidation grout zone and curtain grout zone were 20 Lugeons and 10 Lugeons, respectively and were based on early grouting test section results. These values were used in the earlier seepage analyses. The “as-built” analysis used grouted rock permeability coefficients that were determined from V grout hole pressure testing results. Final “as-built” grouted rock permeabilities of 2.64x10-4 cm/sec. (20 Lugeons) and 6.6x10-5 cm/sec. (5 Lugeons) were used in the consolidation grout zone and curtain grout zone, respectively. The two impervious core zone materials, Zone 1 and 1B, were modeled with permeability coefficients of 2.0x10-8 cm/sec. and 5.0x10-7 cm/sec., respectively. These values were based on laboratory determinations in accordance with ASTM D-5084. The anisotropy ratio (the ratio of the hydraulic conductivity in the horizontal direction to the vertical direction) was assumed as 2.0 for Zone 1 and 5.0 for Zone 1B and were based on specified compaction methods.

1496


The total seepage was estimated based on fifteen selected analysis sections. The end area method was used to compute the total seepage between sections. The first and last sections with assumed seepage equal to zero were STA 2+50 on the left abutment and STA 15+00 on the right abutment. Calculated seepage rates are summarized in Table 3 by station interval and by seepage collection feature (i.e. drainage gallery segment and blanket drain). The predicted combined gallery flow was calculated to be 240 l/sec. Similarly, the predicted downstream blanket drain flow was calculated to be 67 l/sec. for a combined predicted seepage rate of 307 l/sec. Table 3. “As Built” Seepage Analysis Results Station Interval

2+50 to 4+00 4+00 to 5+00 5+00 to 6+00 6+00 to 6+70 6+70 to 7+00 7+00 to 8+00 8+00 to 8+12.5 8+12.5 to 9+00 9+00 10+00 10+00 to 10+37.5 10+37.5 to 11+50 11+50 to 13+00 13+00 to 13+80 13+80 to 14+50 14+50 to 15+00 Total Seepage by Collection Feature (l/sec.) Total Gallery Seepage (l/sec.) Total Downstream Blanket Seepage (l/sec.) Combined Total Seepage (l/sec.)

Upper Left Gallery (l/sec.) 16.5 27.0 53.5 97.0

Lower Left Gallery (l/sec.)

Upper Right Gallery (l/sec.)

Lower Right Gallery (l/sec.)

Downstream Blanket Drain (l/sec.)

17.2 17.2

42.8 15.5 8.3 6.1 72.7

9.7 43.7 53.4

2.9 0.7 0.0 8.1 6.3 19.2 2.2 9.0 15.5 0.5 2.5 0.1 67

240 -

-

-

-

67

307

Deformation

Long term deformation of the dam was evaluated for final design by two methods: 1) review of data related to the long term deformation of large dams with features similar to San Roque Dam, and 2) determine the rough magnitude of long term consolidation settlement of the core zone using Terzaghi’s one-dimensional consolidation theory.


1497

The results of the first approach (case history study) were used to calibrate numerical static stress and deformation analyses of the dam. Because of the many limiting assumptions related to pore water pressure generation during construction, onedimensional consolidation modeling, and the variability associated with large-scale placement and field compaction of cohesive soils, it is generally accepted that numerical models tend to overestimate the magnitude of post-construction deformation of dam cores. A more pragmatic and reliable approach to evaluating the effects of core deformation (and the associated redistribution of stresses) was adopted that considered case histories of post-construction settlement of dams coupled with stress analysis using a well-validated generalized soil constitutive model. Well-documented case histories of the long-term settlement behavior of rockfill dams with finer-grained cores (Dascal 1987, Clements 1984, Kollgaard and Chadwick 1988) among others) served as the basis of the case history study. Approximately 90 case histories were initially considered, and of these, fifteen were judged to be similar enough to San Roque Dam to be included in the deformation evaluation. The selected case histories were generally similar to San Roque Dam in core, shell, and foundation materials; dam height; and the construction technique. The soil compaction effort for San Roque was judged to be higher than those of the older dams. Therefore, settlement estimates based on case histories were somewhat conservative to use for this project. The case history data indicated that dams with characteristics similar to the San Roque dam typically experience long term (about 40 years) maximum vertical strains (settlement per unit height) of about 0.4 percent. Based on this evaluation of performance of existing dams similar to San Roque, long-term post-construction vertical strains of the dam core were predicted to be on the order of 0.5 percent at the maximum station 9+00 (1m). These strains were expected to decrease to close to 0 percent near the abutments of the dam. Although it was believed that long-term vertical strains will be limited to about 0.5 percent, to assess the sensitivity of the results and possible variation in total settlement, the stress analysis were performed considering 1.0 percent vertical strain (2m). The results of the second approach (consolidation settlement analysis) are known to overestimate long term consolidation settlement of central core embankment dams; however the analysis was performed to obtain an upper bound settlement value for comparison to the results computed using more detailed long term deformation analysis techniques. A one-dimensional long-term consolidation settlement analysis was performed on the longitudinal profile of the dam. The analysis was performed using a spreadsheet to determine the order of magnitude of long-term consolidation settlement using Terzaghi’s one-dimensional consolidation theory. Compression index values (CC) were determined from laboratory testing and used as input for the spreadsheet. Consolidation settlement of the core was calculated at 14 stations. At each of the stations, the core was divided into sublayers with heights corresponding to the height of core elements along the central axis of finite element models. The results of the analysis using the range of Cc values

1498


predicted that total long term settlement of the core would vary from 1.4 to 3.1 meters, at stations 8+50 to 10+00 (closest to the maximum section). An average value of 2 meters was chosen to account for the long term consolidation settlement of the dam. Camber was incorporated into the embankment design geometry to account for potential earthquake-induced deformations of the crest of the dam and long-term static deformation caused by consolidation of core material. In order to compute camber, the estimated long-term deformation was added to the maximum anticipated dynamic deformation. Details regarding the earthquake-induced deformation computation analyses are outside the scope of this paper. The results of the analyses indicate a maximum dynamic deformation of 0.5 m and approximately 2.0 m of long-term static deformation, for a cumulative total of 2.5 m. Therefore, in addition to 15 m of freeboard above the normal operating reservoir elevation 280, the crest of the dam was constructed with a maximum 2.5 m of camber in the center of the dam. EVALUATION OF PERFORMANCE Performance During First Filling

Impoundment of the reservoir began on August 8, 2002 with closure of diversion tunnels 2 and 3. Diversion Tunnel No. 1 was permanently closed earlier in the year. As of November 15, 2002, the reservoir had filled to elevation 268.8, about 11.2 meters below the normal maximum operating pool of elevation 280.0. The reservoir did not meet the design maximum pool until the 2003 wet season. Evaluation of monitoring program data at the end of 2002 indicated that the dam and its foundations performed as anticipated during the first impoundment of the reservoir, and there were no concerns with respect to the performance and safety of the dam. The following observations were made based on the inspections and data obtained at that time: •

•

•

•

•

A number of the piezometers in the core were then responding to reservoir filling. A steady state condition had not yet been established but the trend was in that direction. The downstream shell piezometers indicated that drainage occurred with essentially no increase in piezometric head. The piezometers in the foundation indicated head loss generally in accordance with the seepage analyses. Total seepage from the galleries, as of November 15, 2002, is approximately 55 l/s; seepage at the downstream toe of the dam is about 127 l/s. These flow rates were reasonable and acceptable considering the geology of the site. There was some indication, based on the response of the core piezometers, that the recent increase in the seepage flow rate at the toe of the dam was the result of the normal development of saturation and seepage through the core. The maximum vertical settlement measured to date (November 2002) at the crest of the dam is 309 mm. This settlement was considerably less than the 2.5 meters of


1499

•

camber and is well within the anticipated post-construction settlement after six months of the dam being completed. No indication of longitudinal cracking at the upstream edge of the crest was observed as a result of saturation of the upstream shell.

In 2003, the reservoir reached full operating level several times, and was above normal operating level of elevation 280, for short periods during extreme typhoon events. Evaluation of monitoring program data at the end of 2003 again indicated that the dam and its foundations performed as anticipated during the first impoundment of the reservoir, and there were no concerns with respect to the performance and safety of the dam. The following observations were made based on the inspections and data obtained at that time: •

•

•

•

•

•

The piezometers in the core have responded to reservoir filling. A steady state condition has not yet been established in some lower piezometers, but the trend is in that direction. The downstream shell piezometers indicated that drainage occurs with no increase in piezometric head. The piezometers in the foundation indicated head loss generally in accordance with the seepage analyses. Total seepage from the galleries, as of October 14, 2003, was approximately 80 l/s; seepage at the downstream toe of the dam was about 88 l/s. These flow rates were reasonable and acceptable considering the geology of the site. There was some indication, based on the response of the total seepage, that the slight decrease in the seepage flow rate from the galleries was a result of calcification or natural sealing of the rock fractures. In addition, the recent reduction of the toe drain seepage recorded, was a result of the drier weather conditions being experienced over the past few weeks. The maximum vertical settlement measured to date at the crest of the dam was 443 mm. This settlement was considerably less than the 2.5 meters of camber and was well within the anticipated post-construction settlement after 17 months of the dam being completed. No indication of longitudinal cracking at the upstream edge of the crest was observed as a result of saturation of the upstream shell.

Following first filling in November 2003, it was concluded that there were redundancies in the instruments designed into the dam instrumentation system, in particular, the piezometers and the settlement devices. The purpose of these instruments was for use during construction and first filling of the reservoir. Based on the piezometer redundancy and the fact that the vibrating wire piezometers were performing well, the non-working pneumatic piezometers were decommissioned based on recommendations from the Board of Consultants. Since the pneumatic piezometers were the redundant instruments, it was concluded that if or when the remainder do not perform then they should also be decommissioned. These instruments are no longer read. In addition, since the shell piezometers were indicating little or no water pressure, all of the reading frequency was reduced read to bi-monthly.

1500


Evaluation of monitoring data also indicated that the internal settlement cells were not performing properly. It also appeared that the dam settlements, as measured by the external settlement monuments, were consistent and valid, and were approaching stable values. Since the internal settlement cells were intended to measure settlement of the dam during construction and first filling, it was concluded that the usefulness of the internal settlement cells was in question, and the internal settlement cells were decommissioned upon recommendation of the Board of Consultants. Subsequently, the monitoring of deformation for long term conditions has relied on the external settlement monuments. Long Term Performance

Pore Water Pressure: Vibrating wire piezometers are installed at various elevations in the downstream portion of the embankment core and in the downstream shell (Figures 11 through 13). Piezometers in the core quickly responded to the filling of the reservoir and mimic the rising and falling levels of the reservoir. Data from the piezometers in the core show head levels that are consistently less than the phreatic surface assumed for the design analyses. For the vibrating wire piezometers in the shell, the piezometers indicated constant stable readings near the design assumption, indicating the internal drain is functioning as intended by the deign with no increase in piezometric head in the downstream shell zone of the dam. Table 4 lists the highest reading for the calendar year for vibrating wire piezometers in the core and shell zones where data is continually available since the beginning of reservoir filling operations. The core piezometers show an increase with reservoir filling and then stabilization over time. Water levels in the shell are typically dry or respond to precipitation and rises in the tailwater level. The 2010 water levels in these piezometers are shown on Figure 11. Superimposed on this figure is the estimated phreatic surface calculated during design using the Casagrande method. This level can be compared to the 2010 measured levels.


1501

Figure 11. 2010 Water Levels in Core and Shell, Station 9+25

1502


Table 4. Highest Piezometer Reading per Calendar Year Core Piezometers VWP-11 VWP-13 VWP-14 VWP-18 Shell Piezometers VWP-16 VWP-21 VWP-22

Elevation (m) 240 184.5 131.3 102.5

2002

2003

2004

2005

2006

2007

2008

2009

2010

254 208 183 112

260 218 212 163

259 216 217 172

258 216 216 172

260 213 211 178

259 210 207 179

259 210 206 179

263 210 207 183

256 208 201 181

130.35 106.53 105.76

130 109 108

130 108 108

130 107 113

130 106 108

130 105 107

130 105 107

130 105 107

131 105 109

131 105 107

Figure 12. Water Levels in Core Piezometers


1503

Figure 13. Water Levels in Shell Piezometers Monitoring of the foundation gallery piezometers indicate head loss generally in accordance with the design seepage analyses. The gallery piezometers respond to changes in reservoir elevation. Their pattern tends to mirror that of the change in reservoir levels. There is no direct link to changes in rainfall patterns, other than in that the rainfall affects the reservoir elevation. Table 5 shows for a typical group of gallery piezometers at or near the maximum section, the piezometric head after the first filling and at current high reservoir levels and compares it to the analysis piezometric head. This table illustrates that most of the piezometric heads are at or below the analysis piezometric head.

1504


Table 5. Lower Gallery Piezometers Number

Location

GP-12-50-D GP-12-20-D GP-12-60-U GP-12-50-U GP-12-20-U GP-13-70-D GP-13-50-D GP-13-20-D GP-13-60-U GP-13-50-U GP-13-20-U

8+85 8+85 8+85 8+85 8+85 9+00 9+00 9+00 9+00 9+00 9+00

Piezometric Head at Tip (m) October 29, 2002 157.46 152.30 120.54 145.76 185.02 126.62 141.10 178.69 165.13 177.16 198.77

Piezometric Head at Tip (m) October 29, 2010 138.16 137.13 102.11 130.19 173.71 114.90 127.31 169.04 168.31 135.33 189.14

Analysis Piezometric Head at Tip (m) 130 160 130 150 190 90 130 170 160 170 190

Seepage: The long term, total seepage reflects the reservoir level with the total amount of seepage decreasing over time and appears to be approaching steady-state conditions. Total seepage from first filling to February 2010 is shown on Figure 14. Total seepage consists of flow measured from the galleries and the toe drain. The monitoring data shows that the total flow has gradually decreased over time with a maximum flow of 220 l/s in 2002 to a low value of 100 l/s in 2008. The total flow in 2009 spiked to 190 l/s; however, this increase is due to infiltration from extremely high precipitation that influenced the toe drain readings. The gallery seepage has also had a net decrease over time. There is some indication that the decrease in the seepage flow rate from the galleries is a result of calcification or natural sealing of the rock fractures. In the upper left grouting gallery, a limited remedial grouting program was conducted in areas of very high seepage which resulted in a reduction of the inflow. Table 6 shows the decrease in seepage over time. Seepage quantities reflect highest measured amounts per calendar year at high reservoir levels. The high value for the toe drain in 2004 reflects high precipitation.

Location Toe Drain Galleries Total

Table 6. Measured Seepage Flows (l/s) at Highest Reservoir Level in Year Shown Predicted Value 2004 2007 67 112 45 240 71 55 307 183 100


2010 36 41 77

1505

Figure 14. Total Seepage Deformation: As predicted by design analyses and evaluations, the maximum amount of settlement has occurred at the maximum section of the dam, and appears at this time to be reaching equilibrium, as shown on the crest settlement profile on Figures 15 and 16. Figure 15 shows the longitudinal settlement profile across the crest of the dam. As anticipated, the greatest amount of settlement has occurred in the highest central sections of the dam and decreases towards the abutments. Settlement at the maximum section of the dam is 756 mm (29.8 inches) and reduces to 230 mm (9.1 inches) at the left abutment and 265 mm (10.4 inches) on the right abutment. Figure 16 shows a transverse settlement profile of the maximum section. As expected, at the deepest sections, monuments SP5 and SP49, show the greatest amount of settlement, 756 mm (29.8 inches) and 802 mm (31.6 inches), respectively. The monitoring data indicates that the long term settlement is much less than the two meters of settlement predicted by design analyses and evaluations. There has been no indication of longitudinal cracking at the upstream edge of the crest as a result of the saturation of the upstream shell.

1506


Figure 15. Crest Settlement Profile


1507

Figure 16. Settlement Profile – Maximum Section CONCLUSIONS

The San Roque rockfill-central core dam has performed very well for the last eight years; essentially since its first filling in 2002. The embankment dam was designed using sophisticated methods and checked using empirical methods and performance characteristics from similar dams. The foundation conditions were the most difficult to ascertain and as a result considerable on-site modifications were made to accommodate the poor geologic conditions. These modifications were necessary to ensure the safeguards against piping as well as meeting the seepage and leakage objectives. The fact that the design-build nature of the contract and project facilitated the rapid recognition of unforeseen and unknown features during construction and enabled efficient and effective treatments made the project a success. The result, as indicated above, is that this large rockfill dam meets and exceeds its performance criteria and reinforces the current engineering methods of design. ACKNOWLEDGEMENTS

The authors would like to acknowledge the assistance of Mr. William Connell – Senior Vice President for Plant Operations and Site Administrator, Mr. Jeric Codinera – SRPC 1508


Instrumentation Engineer and Mr. Raymund Mariano – SRPC Engineering Manager, for their help in providing the monitoring data and reviewing this paper. REFERENCES

San Roque Power Corporation, Weekly Reports of Instrumentation Results, from 2002 to 2010. San Roque Consulting Panel, San Roque Multipurpose Project, SRPC Consulting Panel Meeting No. 12, July 17, 2002. San Roque Consulting Panel, San Roque Multipurpose Project, SRPC Consulting Panel Meeting No. 13, November 18, 2002. San Roque Consulting Panel, San Roque Multipurpose Project, SRPC Consulting Panel Meeting No. 14, October 26, 2003. United Engineers International, Inc., Addendum 1, Design Statement – Level 2, Main Dam Embankment, for the San Roque Multipurpose Project for the San Roque Power Corporation, July 2000. United Engineers International, Inc., Instrumentation Monitoring and Inspection Manual, for the San Roque Multipurpose Project for the San Roque Power Corporation, August 2002.


1509

SAN ROQUE MULTIPURPOSE PROJECT PERFORMANCE

Recommend Documents