UPDATED ASSESSMENT OF CONCRETE GROWTH EFFECTS

ON HIWASSEE DAM

Dan D. Curtis1 Bradley J. Davis

2

Shahzad Rahman3 Rex C. Powell

4

ABSTRACT

Tennessee Valley Authority’s Hiwassee Dam is experiencing concrete growth caused by

Alkali-Aggregate Reaction (AAR). Extensive investigations and remedial actions were

undertaken in the early 1990s. The concrete growth investigations were recently updated

and this paper presents the results of the updated investigations.

Hiwassee Dam is a 307-ft high concrete gravity dam located near Murphy, North

Carolina. The project was completed in 1940 and consists of a non-overflow gravity

dam, a spillway with seven bays, an intake and a two-unit powerhouse.

A review of survey and instrumentation data is presented. The survey and

instrumentation data is processed to provide estimates of concrete growth rates in the

dam. The survey, instrumentation data and measurements from stress cells and concrete

overcoring are used to calibrate a global finite element model of the dam. The finite

element program ANSYS was modified using the user-programmable features to include

the nonlinear stress-dependent nature of concrete growth and enhanced creep behavior of

AAR-affected concrete. The finite element model was calibrated to field measurements

and then used to predict the future conditions in the dam and spillway. The finite element

model was used to assess the behavior of slots, which were cut in the dam in the early

1990s. The finite element model results were also used to assess the effects of concrete

growth on the sliding stability of the dam.

A local finite element model was used to investigate the effects of concrete growth of the

spillway piers. The spillway has seven radial gates and the gate anchorage steel is

embedded in the spillway pier concrete. The analysis showed that the anchorage steel

had become highly stressed due to the concrete growth and additional anchorage was

recommended and installed. A thorough review of instrumentation data confirmed the

conclusions from the finite element modeling.

1 Senior Civil Engineer, Acres International, 4342 Queen Street, P.O. Box 1001, Niagara Falls, Ontario,

Canada, L2E 6W1, Tel: (905) 374-5200, Fax: (905) 374-1157, dcurtis@acres.com 2 Civil Engineer, Tennessee Valley Authority, 1101 Market Street, LP1F, Chattanooga, Tennessee 37402-

2801, Tel: (423) 751-4844, Fax: (423) 751-3548, bjdavis@tva.gov 3 Civil Engineer, Acres International, 4342 Queen Street, P.O. Box 1001, Niagara Falls, Ontario, Canada,

L2E 6W1, Tel: (905) 374-5200, Fax: (905) 374-1157, srahman@acres.com 4 Senior Civil Engineer – Acres International, 100 Sylvan Parkway, Amherst, New York 14228-1146, Tel:

716-689-3737 • Fax: 716-689-3749, rpowell@acres.com

INTRODUCTION

Project Description

Hiwassee Dam is located on the Hiwassee River in Cherokee County, North Carolina.

Construction of Hiwassee Dam started in July 1936 and was completed in December

1940. The dam is a 307-ft high concrete gravity structure. The total length of the dam is

1285 feet and is composed of a 260-ft long central spillway section, and two non-

overflow sections on the left and right side of the spillway, measuring 597 and 428 ft,

respectively. The main spillway consists of seven radial gated bays. The powerplant has

two units, one of which is a pump-turbine. There is a roadway over the top of the dam

with bridge spans at the spillway bays. The dam is owned and operated by the Tennessee

Valley Authority.

The concrete structures at Hiwassee Dam are affected by Alkali-Aggregate Reaction

(AAR), which causes the concrete to expand over time. The AAR-induced deformations

can have adverse effects on the functioning of mechanical equipment such as spillway

gates and turbines. At Hiwassee, the AAR-induced stresses have caused cracking at

several locations, most noticeably along the horizontal construction joints at both ends of

the dam. The non-overflow blocks adjacent to the spillway have deflected into the

spillway opening, causing the spillway gates to bind. There was also concern that these

deflections may eventually overload the spillway bridge. As a result, TVA has in place an

extensive program consisting of monitoring by instrumentation, finite element based

investigations, and a slot cutting program aimed at extending the service life of the

project.

The objective of this paper is to present an updated assessment of the effects of concrete

growth due to AAR at Hiwassee Dam.

Previous AAR Assessments and Remedial Actions

In 1991 a three-dimensional, linear-elastic finite element model of the dam and its

foundation was developed by Stone & Webster Engineering Corporation using ANSYS, a

general-purpose finite element analysis program developed by ANSYS Inc. In these

analyses AAR-induced concrete growth was simulated by an equivalent thermal

expansion of concrete. The model was used as an aid for devising remedial measures to

control stresses and deflections resulting from concrete growth. From the analysis, it was

concluded that remediation would include the following: post-tensioning of the dam,

cutting of expansion joints in the bridge, and cutting of four transverse slots; two adjacent

to the spillway and two near the abutments. This would help control the stresses and

allow additional expansion of the concrete in the spillway area without causing binding

of the spillway gates. Figure 1 shows a downstream elevation of the dam and the location

of the four slots in the dam.

The remedial repairs were completed using a phased approach. The non-overflow

sections of the dam were post-tensioned to an average depth of 75 feet, using two rows of

multi-strand tendons. This was included in the repair program, primarily to improve

stability under seismic loading and during slot cutting. The spillway section was not

post-tensioned.

Figure 1. Downstream Elevation Showing Four Slots in Upper Portion of Dam

In 2002, Acres International was contracted to provide an updated assessment of the

condition of the dam. The updated assessment included: a review of instrumentation

data, global finite element analysis (FEA) of concrete growth effects, stability

assessment, local FEA of spillway piers and recommendations for remediation. The

following sections summarize the results of these updated assessments undertaken by

Acres.

REVIEW OF SURVEY AND INSTRUMENTATION DATA

Survey and instrumentation data includes the following:

• top of dam settlement and alignment surveys

• Electronic Distance Measurement (EDM) surveys

• spillway pier stationing surveys measuring tilt of end piers

• upstream-downstream tilt of spillway Block 14 plumb line

• growth meters and extensometers in Block 16 of the left non-overflow section

• joint meters at slots and adjacent block joints

• stress measurements from overcoring and stress cells

• temperature measurements from various instruments in the dam.

The top of dam survey data was processed in two ways. The vertical rise measurements

from the special survey of 1991 were used to estimate the total rise of the dam since the

end of construction (1940). The top of dam surveys after 1990 were used to estimate

rates of vertical rise using the 1990 survey as a baseline. The survey measurements

provided very useful information on the average rate of concrete expansion over the

height of the dam. The various surveys were quite consistent with respect to the rates of

concrete growth in the vertical direction.

The EDM surveys measure the longitudinal length changes of the right non-overflow

section, the spillway and the left non-overflow section. The measurements are taken at

the top of the dam. The measurements from 1975 onward provide useful information on

the change in the width of the spillway as concrete growth causes the spillway end piers

to tilt inwards.

The Block 14 plumb line provides upstream-downstream tilt measurements of spillway

Block 14. The plumb line was operational prior to reservoir impoundment in 1940,

therefore it provides useful data on the dam response to impounding and also the tilt in

the upstream direction caused by continued concrete growth.

Growth meters and extensometers were installed in Block 16 adjacent to the spillway.

The meters were installed at the top of the dam and in an access gallery near the base of

the dam. These instruments were installed in 1991; hence this new data was not

available in the previous evaluations. The measurements, together with the survey data,

were used to estimate the distribution of vertical concrete growth as shown in Figure 2.

Figure 2. Extensometers and Growth Meters in Block 16

Joint meters are installed across the two spillway and the two abutment slots. The joint

meters monitor the closure of the slots over time.

Concrete overcoring stress measurements provide data on the stress condition in the dam

at key points in time. In several cases, stress cells were installed in the overcored holes to

allow monitoring of the increase/decrease of concrete stress over time. The overcoring

stress measurements showed that relatively large longitudinal compressive stresses occur

at the crest of the spillway near the spillway end piers. The large longitudinal

compressive stress in this area is attributed to stress flow from the upper portion of the

non-overflow section down into the crest of the spillway with the spillway acting as a

wide notch in the dam. Several stress cells were located adjacent to and below the four

slots. The stress cells adjacent to the slots showed an increase in stress during slot cutting

followed by a complete unloading as the diamond wire saw proceeded deeper below the

stress cell. The stress cells under the slots showed an increase in stress due to

longitudinal flow of stress under the slot.

GLOBAL FINITE ELEMENT ANALYSIS

TVA supplied Acres with the original finite element (FE) model as used in the 1991

studies. The finite element model is shown in Figure 3. The finite element model is

relatively coarse but it was considered sufficiently detailed to allow an assessment of the

condition of the dam. The finite element program ANSYS was used to perform the time-

dependent nonlinear concrete growth analyses. ANSYS has a feature of allowing user-

specified material subroutines whereby the user can program their own material laws into

ANSYS. This feature was used to implement the time-dependent and stress-dependent

concrete growth law from Acres GROW3D program into ANSYS. The concrete growth

law is described in Curtis (1995). The solution procedure is shown in Figure 4.

Figure 3. Global Finite Element Model Mesh

Figure 4. Concrete Growth Law Solution Procedure in ANSYS

DAM AND FOUNDATION

6881 Elements 9301 Nodes

Establish Initial Conditions

(Gravity, Hydrostatic Loads etc)

Access Element Stresses

Compute Strain / Stress Increments in

Global Coordinates Based on Growth Law and Impose Them at Each Gauss Pt.

Carry out Equilibrium Iterations till Convergence

Save Results for Load Step

Next Load Step

The finite element model of the dam was calibrated using concrete growth rates

determined from data review. Temperature measurements at instruments within the dam

were used to assess the effect of temperature on the concrete growth rates. Similar to

previous work, it was found that the higher temperature on the downstream portion of the

dam leads to higher growth rates in this area.

From Figure 4, the initial conditions are established prior to running the nonlinear

concrete growth analysis. The model is stepped through time using one-year time

increments to determine the effects of continued concrete growth. The finite element

model was calibrated with vertical displacement survey data as shown in Figures 5 and 6.

Figure 7 shows the calibration to the Block 14 plumb line data. Figure 8 shows the

comparison of the measured and computed spillway closure from the EDM survey.

Figure 9 shows a comparison of measured and computed longitudinal stresses from

Block 16 adjacent to the spillway. A good match to the measured stresses was achieved.

The slot closure measurements at Block 16/17, shown in Figure 10, also compare well

with the closures predicted by the model.

The model results taken prior to slot cutting computed quite large horizontal shear

stresses acting longitudinally at the spillway end pier/spillway crest junction. The fact

that the model matches the measured longitudinal stresses and slot closures at this

location suggests that other stress components are probably reasonably well estimated by

the model. There was no evidence of Block 16 sliding longitudinally into the spillway

opening, therefore the construction joints at this location must possess quite high

cohesive strengths in excess of 400 psi!

Conclusions From Global FEA

The model has provided useful insight into the behavior of the dam and this information

was compared to other projects to assess rehabilitation requirements and life extension

planning. The calibrated model was used to investigate the various proposed remedial

measures, such as deepening of the slots, additional spillway slots, and the effect of

allowing the spillway slots to close.

Investigation of the mechanisms of crack formation revealed that the cracking in the

curved abutment blocks is due to the interaction of the longitudinal compression and the

curved geometry of the dam at these locations. The abutment slots prevent the

development of compressive force in the curved abutment blocks, which should prevent

future crack formation in these blocks. Therefore the abutment slots should be

maintained open.

The stability of the dam sections most affected by concrete growth and slot cutting was

examined. Forces acting on sections adjacent to the slots were determined from the FE

model. In some cases, it was found that blocks adjacent to slots may, in the future, move

slightly towards the slot to relieve longitudinal forces caused by continued concrete

growth. Stability in the upstream-downstream direction of the blocks in the vicinity of

Figure 5. Measured versus Computed Total Vertical Rise

Figure 6. Measured versus Computed Vertical Rise Since 1990

.

DISTANCE FROM SURVEY MARKER 3-1 (FT)

Settlment Surveys After 1990

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

-200

-100 0

100

200

300

400

500

600

700

800

900

1000

1100

DISTANCE FROM SURVEY MARKER 3-1 (feet)

SE

TT

LE

ME

NT

(-)

---

Vert

ical

Mo

vem

en

t (i

nch

) --

- E

XP

AN

SIO

N (

+)

Special Survey

of 1991

Calibration

Result

Vertical Movement (1940 to 1991)

Right Abutment Spillway Left Abutment

-0.60

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

1.10

1.20

-200

-100 0

100

200

300

400

500

600

700

800

900

1000

1100

SE

TT

LE

ME

NT

(-)

---

Mo

ve

men

t (in

ch

) --

- E

XP

AN

SIO

N (

+)

NOV 1990 - INITIAL SURVEY

Dec 2002

Nov 1995

Calibration

Results

Right Abutment Left AbutmentSpillway

1240

1260

1280

1300

1320

1340

1360

1380

1400

1420

1440

1460

1480

1500

1520

1540

-0.6

0

-0.4

0

-0.2

0

0.0

0

0.2

0

0.4

0

0.6

0

0.8

0

1.0

0

1.2

0

1.4

0

EL

EV

AT

ION

(f

eet)

EL 1316

EL 1277

EL 1423

EL 1503

1940

Refe

ren

ce

May 27, 1944

(observed)

Foundation, EL 1245 +/,

assumed to be zero.

1992 (Mean)

Downstream Upstream (inches)

Calibratio

n

Calib

ratio

n

Figure 7. Measured versus Computed Spillway Tilt due to Impounding (1944)

and Concrete Growth (1992)

Figure 8. Measured versus Computed Longitudinal Length Change

of Spillway from EDM

1540

Spillway Block 14 Plumbline Tilt

-1.10

-1.00

-0.90

-0.80

-0.70

-0.60

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

SP

ILL

WA

Y C

LO

SU

RE

(In

ch

es

)

EDM Survey Data

Calibration

Spwy Slots cut in 1993.

2/

3

S

LO

T

H D - 1H D - 2H D - 4 H D - 3

7/8

SL

OT

16

/1

7

S

LO

T

24

/2

5

S

LO

T

BL

K

0

BL

K

8

BL

K

16

BL

K

2

7

2/

3

S

LO

T

H D - 1H D - 2H D - 4 H D - 3 H D - 1H D - 2H D - 4 H D - 3

7/8

SL

OT

16

/1

7

S

LO

T

24

/2

5

S

LO

T

BL

K

0

BL

K

8

BL

K

16

BL

K

2

7

Right Side Left SideSpillway

YEAR

Abutment Slots cut in 1994.

0.34 " (Immediate Recovery)

0.37 " (Additional

Recovery by 2003)

Longitudinal Stress Hole 3, Phase VI (1992)

Figure 9. Measured versus Computed Longitudinal Stresses in Block 16

Adjacent to Spillway

Closure of Spillway Slot 16/17

Figure 10. Measured versus Computed Slot Closure at Block 16/17 Slot

0

5

10

15

20

25

30

35

40

45

50

-500 0 500 1000 1500 2000

Tension (-) --- STRESS (psi) --- Compression (+)

Dep

th o

f Ov

erc

ore

Te

stin

g (fe

et)

Longitudinal Stress (RAW)

Calibration (AAR + Temp)

Longitudinal Stress (AVG)

Spillway Crest, EL

1503.5

Roadway, EL 1537.5

End of beam

touching

concrete

BAY 7

EN

D P

IER

2

-1.6

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Year

CL

OS

E <

Mo

vem

en

t (i

nch

) >

OP

EN

JM 16/17 US-L

JM 16/17 DS-L

JM 16/17 DS-L (Gallery)

Calibration

Slo

t C

ut

Re

cu

t

Re

cu

t

Re

cu

t

the slots is adequate even if longitudinal movement into the slot has occurred. No

additional anchors to tie the upper portion of the dam to the dam concrete beneath the

slots are considered necessary at this stage.

A number of remedial measures such as deepening of existing slots and allowing the slots

to close were examined using the FE model with concrete growth. The need for

additional spillway slots was also reviewed. Results indicated that there is little

justification for deepening the slots or providing additional slots. Both instrumentation

data and model predictions suggest that the time interval between recutting of the

spillway slots can be increased considerably without impairing gate operation.

The strategy for slot cutting depends on the performance of the spillway gates. At

present, the gates operate satisfactorily without jamming. The survey data and the model

prediction indicate slight continued opening of the spillway width. However, given that

there is tilting of the end piers and the gate clearances are unknown at this time, the

spillway pier movements should be monitored and, if necessary, future measures such as

minor deepening of spillway slots could be undertaken if gate jamming becomes a

potential problem. Such measures would only be pursued after monitoring movements

over the next few years of operation.

LOCAL SPILLWAY PIER ANALYSIS

Regular inspections of the dam have identified cracking in the spillway piers near the

trunnion pins. The cracks are typically vertical and run above and below the concrete

around the trunnion pins. The spillway piers were not included in the global FE model.

In order to investigate the condition of the spillway pier, a local finite element model of

one spillway pier, including the trunnion anchorage steel, was developed. The model of a

typical pier is shown in Figure 11 and the model of trunnion pin and anchorage steel is

shown in Figure 12. The steel reinforcement in the pier was modeled with shell elements

using a smeared equivalent area of steel.

From the analysis, the anchorage steel restrains the concrete and the restrained concrete

expansion leads to higher horizontal concrete compressive stresses in the vicinity of the

anchorage steel. The concrete above and below the trunnion pin anchorage steel expands

horizontally at a greater rate than the concrete in the immediate vicinity of the trunnion

anchorage steel. The result of this differential horizontal expansion is the vertical crack

observed in the concrete downstream of the trunnion pin.

The leveling surveys show the total rise of the dam at the spillway section to be less than

the adjacent non-overflow section. The local finite element analysis also predicts lower

expansion rates in the spillway pier concrete due to the restraining compressive stress

imposed on the concrete from reinforcing steel. As the concrete expands, the reinforcing

steel resists the expansion, and this small restraint induces concrete compression. The

magnitude of this additional compressive stress is 100 to 200 psi.

Figure 11. Local Spillway Pier Model Mesh

Figure 12. Trunnion Pin and Anchorage Steel Plate Mesh (Concrete not shown)

A variety of modelling assumptions were made regarding anchorage connectivity to

concrete and concrete growth rates in the secondary concrete. Regardless of the

assumption used, the model predicts very high stresses in the anchorage steel. The

computed stresses are approaching the yield stress of the steel in both the bracket plate

and double plate anchorage steel. From the analysis, it was concluded that the anchorage

steel is probably yielded at the bolted connection upstream of the trunnion pin.

Subsequently, the trunnion pin extensometer data was re-examined. The measurements

are plotted as a function of time in Figure 13. The data was re-plotted in Figure 14 such

that trunnion pin movements versus reservoir elevation are shown. From the analysis and

measurements, the gate anchorage system is behaving inelastically at high reservoir

levels.

Figure 13. Measurements of Trunnion Pin Movement versus Time

Figure 14. Measurements of Trunnion Pin Movement versus Head Water Elevation

It is anticipated that inelastic behavior is occurring at the bolted connection of the

trunnion pin bracket plate to the double plate anchorage steel. As local yielding develops

at this location, the surrounding concrete takes on additional load which limits the

downstream movement of the trunnion pin. In time, as the steel stresses become higher,

the pin deflections for a given load will increase and the ability of the pier to sustain

additional load will be limited by the shear strength of concrete downstream of the

trunnion pin. The concrete downstream of the trunnion pin was not designed to carry this

Hiwassee Project - BAY 5 - Left & Right Trunnion Pins

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

2001 2002 2003 2004 2005 2006

US

(-)

---

Pin

Mo

ve

me

nt

(in

ch) -

-- D

S (

+)

1440

1450

1460

1470

1480

1490

1500

1510

1520

1530

1540

He

ad

wa

ter

Ele

vatio

n (

fee

t)

LEFT PIN (Corrected) RIGHT PIN (Corrected) HW

The meter is attached to the pin

and anchor ed in the concrete

ups tr eam.

Trunnion Pin Movement versus Head Water Elevation for Period of May 1

to May 31 , 2003

1512

1514

1516

1518

1520

1522

1524

1526

1528

0.005 0.010 0.015 0.020 0.025

Pin Movement (inch)

He

ad

Wa

ter

Ele

va

tion

(ft)

Left Pin Right Pin

additional load. Also, it is not accepted practice to rely on the downstream pier concrete

alone to carry the gate load. Therefore remedial action was recommended to maintain the

safety and reliability of the dam.

The geometry of the end piers differs from that of the intermediate piers and provides

significantly more concrete downstream of the trunnion pin. It was determined that the

concrete and reinforcement downstream of the trunnion pin for the end piers has surplus

capacity to resist the gate loads. Even though the finite element analysis estimated that

the reinforcement in the end piers has already yielded due to concrete growth, the

reinforcement still provides confinement and strength. The end piers therefore do not

require remedial repairs. Nevertheless, the end piers will be monitored along with the

intermediate piers to track their behavior in the future.

SPILLWAY PIER REHABILITATION

A remediation scheme was needed to allow repair of the intermediate spillway piers

during service. It was also imperative that the reservoir be maintained to a predetermined

elevation in order to provide a margin of safety against inelastic movement of the

trunnion pins. Therefore, the selected scheme needed to be reasonably simple and

require as short an installation time as possible. The physical dimensions and

accessibility of the existing piers also limited available options.

Six schemes were considered in the evaluation of repair options. The selected scheme

was comprised of a pair of post-tensioned anchors and a cast-in-place concrete bearing

pad at each of the six intermediate piers. The anchors were aligned parallel with the face

of the piers and inclined so that the upstream ends could be grouted into the mass

concrete of the spillway as shown in Figure 15.

Figure 15. Spillway Pier Anchor Repair

Design of Anchors

The existing anchorage was designed to restrain the full gate load without relying on the

strength of the concrete downstream of the trunnions. Seismic, flood, and normal

maximum headwater loading conditions were considered for design. Normal maximum

headwater was the governing load case, with a horizontal load of 295 kips at each end of

each gate. Based on an anchor inclination of 25 degrees from horizontal, the design post-

tension load in the new anchors was required to be 325 kips.

As the concrete in the piers will continue to grow after installation of the anchors, the

anchors will stretch and increase the post-tension force over time. The finite element

analysis estimated the annual strain rate due to AAR to be on the order of 14 microstrain

per year, corresponding to an increase in anchor force of about 2.1 kips per year. The

design accommodates this phenomenon by essentially over-sizing the anchors and

initially locking them off at a stress well below the normal allowable. Nevertheless,

anchors will require periodic relaxation over the life of the structure.

While the intermediate piers are normally subjected to nearly equal loading from the

gates on each side, a potential situation can occur where one gate may be fully open or

one gate may be dewatered for maintenance. In such a case, eccentric loading will be

applied to the pier. The anchor loads were estimated to increase by approximately

48 kips under such loading due to prying action at the trunnion pin. This allowance for

prying was accounted for in the design recommendations for future relaxation of the

anchors.

According to the guidelines established by the Post Tensioning Institute (PTI), the design

load in an anchor should not exceed 60% of its guaranteed ultimate tensile strength

(GUTS). The anchors used in the repairs each have a GUTS of 778 kips. In order to

verify timing for future relaxation of the anchors, provisions were made to install load

cells to monitor the anchor loads. The anchors will be monitored for increased loads due

to continued concrete growth and when they approach a load of 419 kips, they will be

relaxed to the initial lock-off load.

The vertical contraction joints within the spillway section of Hiwassee Dam are located

coplanar with one face of most piers. Nearly half the anchors are therefore located in

close proximity to those existing contraction joints. However, due to concrete growth,

the contraction joints are tightly closed and in compression, estimated to be on the order

of 600 psi at the spillway crest. Free-edge effects at the contraction joints were therefore

not considered in the design. In fact, future longitudinal stress increase in the spillway is

expected to improve available bond zone capacity over time.

Installation Access to the work area was accomplished from either a man basket or drill platform

suspended by a crane from the bridge deck over the spillway. The drill operation is

shown in Figure 16.

Figure 16. Anchor Hole Drilling

The existing semi-circular downstream face of the piers did not provide a flat surface

necessary for applying anchors loads. Also, as a result of the AAR, the outer 4 inches of

the pier were heavily cracked and deteriorated. The existing concrete was removed to an

approximate depth of 16 inches to allow construction of a reinforced concrete bearing

pad, or saddle, and to provide a substrate of sound concrete. The new saddle was

constructed to transfer the anchor load into the pier over a sufficient area and to provide a

bearing surface perpendicular to the axis of the anchors. Lateral ties were also installed

through the sides of the pier to resist bursting forces near the anchor heads.

After construction of the concrete saddles, two 6-in diameter holes for the anchors were

drilled into each pier. The holes were approximately 38 feet long, with the bond zone

extending approximately 16 feet into the spillway concrete mass. The bond zone at the

bottom of the hole was water pressure tested and verified to be suitably watertight. The

upper portion, along the free-stressing length, was water tested using only static head in

the hole. Two of the piers exhibited significant leakage from cracks through the side of

the pier near the nose. In order to provide an adequate substrate for the anchor loads and

ensure that the encapsulation grout would not be lost, the cracked concrete was removed

and replaced with new concrete. The new concrete was tied to the existing pier with

grouted steel reinforcement dowels.

The anchors were 2 1/2-inch diameter by 40-ft long ASTM A722 (150 ksi) epoxy-coated

bars with Class I corrosion protection. They were fabricated in the factory with pre-

grouted corrugated poly tube along the bond zone. The anchors are re-stressable, with

the free-stressing length comprised of a grease-filled smooth sleeve factory-sealed to the

top of the corrugated tubing. The entire assembly was encapsulated in non-shrink cement

grout by partially filling the hole and inserting the anchor such that grout was displaced

out the top of the hole. Centralizers were used to ensure uniform encapsulation. The

bearing plate trumpet was sealed to the top of the smooth sleeve with an o-ring. The

remaining annulus outside the trumpet was injected with grout after the bearing plates

were installed. Following anchor installation and tensioning, corrosion protection was

completed with a pre-fabricated grease-filled PVC cap installed over the exposed end of

the anchor bar and sealed to the top of the bearing plate. All anchors were either proof or

performance tested in accordance with PTI guidelines and then locked off at a load

nominally 15% higher than the minimum design load of 325 kips. During the testing of

the first two anchors, the piers and adjacent spillway blocks were monitored with

instrumentation for any adverse movements. No adverse affects were noted and

monitoring was discontinued. The anchored piers are shown in Figure 17.

CONCLUSIONS

This paper has shown the importance of both instrumentation and analysis in the

assessment of concrete dams affected by AAR. Both the instrumentation data and the FE

analysis identified a structural problem with the spillway gate trunnion pin anchorage.

The method of analysing the instrumentation data also proved key in recognizing the

inelastic behavior of the trunnion pin anchorage. In the global analysis of the dam, the

instrumentation data was extremely useful in model calibration and understanding the

present and future behavior of the dam

Figure 17. Completed Spillway Pier Repair

ACKNOWLEDGEMENTS

The authors wish to thank the Tennessee Valley Authority for permission to publish this

paper. Numerous TVA staff were involved in this project including M. Cones, R. James

and J. Peyton. The bulk of the instrumentation data presented in this paper was collected

and prepared by D. Tanner of TVA. Also, F. Feng of Acres participated in the anchor

design and analysis.

REFERENCES

Acres International for Tennessee Valley Authority. “Hiwassee Dam Spillway Pier

Repair – Design Basis”. Document No. 14520.29.09. May 20, 2004

Acres International for Tennessee Valley Authority. “Hiwassee Dam – Spillway Pier

Repair Design Calculations”. Document No. 14520.37.03.04. August 23, 2004.

Acres International for Tennessee Valley Authority. “Stress Analysis of Hiwassee

Spillway Pier”. Document No. P14520.29.07. June 2004.

Curtis, D.D. “Modeling of AAR Affected Structures Using the GROW3D FEA

Program”, Proceedings for the Second International Conference on Alkali-Aggregate

Reaction in Hydroelectric Plants and Dams. USCOLD, Chattanooga, Tennessee. 1995.

pp 457-478.

Grenoble, B.A., J.K. Meisenheimer, D. D. Wagner, and V. A. Newell. "Finite Element

Analysis of Three TVA Dams with Alkali-Aggregate Reaction”, Proceedings for the

Second International Conference on Alkali-Aggregate Reaction in Hydroelectric Plants

and Dams. USCOLD, Chattanooga, Tennessee. 1995. pp 67-82.

Newell, V.A. and C. D. Wagner. “ Modifications to Hiwassee Dam and Planned

Modifications to Fontana and Chickamauga Dam by Tennessee valley Authority to

Manage Alkali-Aggregate Reaction”, Proceedings for the Second International

Conference on Alkali-Aggregate Reaction in Hydroelectric Plants and Dams. USCOLD,

Chattanooga, Tennessee. 1995. pp 83-100.

Tanner, D. T. “The Use of Monitoring and Finite Element Analysis in Evaluating

Remedial Measures at TVA’s Hiwassee Dam”. August 1992.