PERFORMANCE OF RETROFITTED DETERIORATED ARCH DAM[1]

By Avinash C. Singhal[2], Fellow, ASCE, and Larry K. Nuss[3]

ABSTRACT: The Stewart Mountain Dam is a 212 ft (64.6 m) high multi-curvature thin arch dam. The structure was completed in 1930. It has experienced alkali-silica reactions within the concrete and exhibited no bond across horizontal construction lift surfaces. The dam could be subjected to upgraded maximum credible earthquake (MCE) or probable maximum flood (PMF) loadings. Alkali-silica reactions and expansions have caused visible surface cracking. This structure was analyzed for gravity-, reservoir-, temperature-, and earthquake-induced loads. Results indicated an unsafe structure for earthquake conditions. Several measures for prevention of further deterioration and strengthening were considered. Post-tensioned cables were selected to provide seismic strengthening. Cable design uses the stiffness, and acceleration response spectra methods. Vertical post-tensioned cables were installed during 1990-92 construction phase. This paper summarizes various field and laboratory investigations, structural analyses, and design parameters needed for post-tensioning a deteriorated arch dam. Post-tensioned cables are found to be a viable solution for the dynamic stability of a thin-arch dam. Methodology presented in this paper is applicable to other deteriorated dams.

INTRODUCTION

The Stewart Mountain Dam located 41 miles (66 km) east of Phoenix, Ariz., on the Salt River, was completed in March 1930. The structure contains an arch dam, two thrust blocks for simulating abutments for the arch dam, three gravity dams, and two spillways. The arch dam measures 212 ft (64.6 m) high at the maximum section, 8 ft (2.44 m) thick across the crest, 34 ft (10.36 m) thick across the base, and 583 ft (177.7 m) in length along the crest. Four keyed vertical contraction joints with copper water-stops separate the arch into distinct concrete sections called cantilevers. The concrete structure has experienced alkali-silica reactions and has exhibited no bond across horizontal construction lift surfaces. In addition, the dam could be subjected to upgraded maximum credible earthquake (MCE) or probable maximum flood (PMF) loadings.

Numerous investigations, field measurements, laboratory test, inspection, and on-site tests have been performed over the years to assess material properties, deformation, and deterioration. Concrete cores were extracted in 1943, 1946, 1947, 1948, 1968, 1977, 1979, 1982, and 1985. The many engineering questions that arose during the investigations and inspections included the following: (1) What caused the poor lift surface bond and what was its extent?; (2) What is the serviceability life expectancy of the existing or deteriorating concrete?; (3) At what rates are the alkali-silica reactions deteriorating the concrete? Have the reactions stopped? Could changing reservoir levels or other conditions accelerate the reaction?; (4) is the concrete more brittle due to micro-fracturing from the reactions?; (5) Why does the upper arch appear more susceptible to the alkali-silica reactions than other areas of the dam?; and (6) Why do deflection measurements of the crest indicate a slowing or stopping of the rate of permanent drift toward the up-stream direction?

LITERATURE REVIIEW

Many dams built before 1945 and located in the southwestern United States, such as the Coolidge, Stewart Mountain, and Parker dams in Arizona and the Riant and Matilija dams in California, have shown signs of alkali-silica reactivity in the concrete. The Matilija Dam showed permanent displacement upstream at the crest ("Railroad" 1984), with concrete cores indicating alkali-silica reactions and deterioration in the uper 25 ft (7.6 m). Modifications made to the Matilija Dam included notching and enlarging the spillway. The Railroad Canyon Dam in southern California has similar horizontal lift surface bond problems (Matilija 1972). The dam, completed in 1928, consists of an arch dam portion with supporting thrust blocks. The dam was stabilized by placing additional concrete on the thrust blocks and installing six 200 kips (890 kN) post-tensioned cables in each abutment.

MODIFICATION CONSIDERATION

Seismic analysis of the dam showed that the arch portion of the dam is potentially unstable during a MCE seismic event. Justification of the decision to modify the structure and the chosen method of modification is based upon the following investigation findings.

  1. Inertia forces at the crest of the arch are probably quite large judging from the resulting peak accelerations of 2.32 g at the crest.
  2. A linear finite element analysis of calculated tensions indicates that the arch dam pulls apart horizontally with a duration of upu to 0.1 seconds, long enough for concrete blocks to slide.
  3. Horizontal construction lift surfaces are laitance filled and exhibit little or no cohesion.
  4. Vertical contraction joints are keyed but provide little resistance against sliding of the massive concrete blocks.
  5. Uniaxial compression tests on 6-in. (15.2 cm) cores extracted from the dam interior indicate very strong concrete of about 5,400 lb/sq in. (37.21 MPa). Alkali-aggregate reaction has not deteriorated the dam to the point requiring its total replacement.

Seismic analysis revealed that the dam will not perform dynamically as a monolithic unit, because of the unbonded horizontal lift surfaces.

POSTTENSIONING

Post-tensioned tendons increase the normal force on the unbonded horizontal arch lift line surfaces and consequently the frictional component of sliding. Cables also produce three-dimensional stresses throughout the arch section depending on orientation and eccentricities. Post-tensioning induces two equal and opposite loads at the ends of the free length. Load at the top or head transfers through the bearing plate into the concrete. This load can be considered a concentrated, or point, force. Load at the bottom develops through bond along the embedment length of the cable.

CABLE CAPACITY AND CONSIDERATIONS

The aforementioned inertia forces were computed at 15 locations along the crest and substituted into Eq. 3. A design cable load is 700 kips (3,114 kN) pre 10 ft (3.05 m0 spacing along the crest. The cables were positioned within the arch were as close as possible to the centerline of the vertical radial section. Finite element studies showed a beneficial stress distribution within the arch dam created by the cable load during normal operating conditions. Special design considerations and requirements were developed for drilling methods, drilling accuracy and tolerances, tensioning sequence, placement within the arch, corrosion protection, grouting, monitoring, and pre-stressing.

CONCLUSIONS

The Stewart Mountain Dam has been deteriorated by alkali-silica reactions and exhibits no bond across horizontal lift surfaces. In addition, it is now required to be subjected to an upgraded maximum credible earthquake. Trends from historic deflection measurements, concrete coring programs, and laboratory tests indicate that the deterioration from alkali-silica reactions is contained. A system of post-tensioning for arch stabilization was chosen. Ease of design and cable-load control were among the factors in this selection. Post-tensioned cables are a viable solution for the dynamic stability of a thin arch dam.

APPENDIX REFERENCES

Mather, B.M. (1967). "Factors which influence the deterioration of concrete dams and measures for preventions of deterioration." Trans. 9th Int. Congress on Large Dams, Int. Congress on Large Dams, Paris, France.

"Matilija Dam - Stress investigations." (1972). Report for the Department of Public Works, County of Ventura, International Engineering Co., Inc., Ventura, Calif.

Raphael, J. M. (1984). "Tensile strength of concrete," ACI J., 81(2), 158-165. "Railroad canyon dam safety evaluation." (1984). Final Report for TEMESEAL Water Company, Woodward Clyde Consultants, San Francisco, Calif.

Seed. H.B., and Idriss, I.M. (1982). Ground motion and soil liquefaction during earthquakes, Earthquake Engineering Research Institute, Berkeley, Calif.

"Seismo-tectonic investigation for Stewart Mountain Dam -- Salt River Project. Arizona" (1986). Seismo-tectonic Report No. 86-2, Bureau of Reclamation, Denver, Colo.

"Static and dynamic structural analysis of the arch, thrust locks, and gravity sections at Stewart Mountain Dam." (1987). Tech. Memorandum SM-220-01-87, Bureau of Reclamation, Denver, Colo.

"Structural designs of the post-tensioned cables for the dynamic stability of Stewart Mountain Dam, Phoenix, Arizona." (1990). Tech Memorandum SMC-3110-01, Bureau of Reclamation, Denver, Colo.

Von Thun, J. L., Roehm, L., Scott, G., and Wilson, J. (1988). "Earthquake ground motions for design and analysis of dams." Earthquake engineering and soil dynamics II: Recent advances in ground motion evaluation, geotechnical special publication no. 20, ASCE, New York, N.Y.

Westergaard, H. M. (1931). "Water pressure on dams during earthquakes." Trans., ASCE, Paper No 1835, ASCE, New York, N.Y., 418-433.

Figure 1. Upstream Acceleration Response Spectra

Figure 2. Finite Element Model

Figure 3. Cantilerver No. 6 Roadway Station 3 + 61.26: (a) Plan of Road Centerline and Radii; (b) Section

Figure 4. Profile of Arch Looking Downstream Showing Location of Posttensioned Cables

1

Avi/Retrofit2.doc

[1] Submitted for the session on "Performance of Retrofitted Structures," 12th ASCE Engineering Mechanics Conference, LaJolla, Calif, May 17-20, 1998.

[2] Prof. Of Civil Engrg., ECE-5306, Arizona State Univ., Tempe, AZ 85287-5306 Tel: 602-965-6901 Fax 602-727-6192, E-mail:

[3] Struct. Engrg., Div. Of Civil Engrg., Concrete Dams Branch, Bureau of Reclamation, Denver, CO 80225