Proper design and application of shotcrete could have prevented failures in the 1995 repair efforts on Barre, Vt.'s Dix Reservoir dam. Photos: CTLGroup
Proper design and application of shotcrete could have prevented failures in the 1995 repair efforts on Barre, Vt.'s Dix Reservoir dam. Photos: CTLGroup
Test panels are the best way to ensure that the shotcrete mix design and application will be acceptable.
Test panels are the best way to ensure that the shotcrete mix design and application will be acceptable.

The concrete dam that created Dix Reservoir in Barre, Vt., was constructed in 1950. Over decades of service, erosion and deterioration became increasingly apparent on the dam's spillway and training walls. Water began seeping through fine cracks and other damaged areas. In 1978, when the U.S. Army Corps of Engineers first conducted a condition survey of the training walls and spillway, they discovered a loss of concrete section and exposed reinforcing steel.

Years later, in conjunction with the building of a new water filtration plant in 1995, the city set out to repair and resurface the spillway and training walls. As city engineer and director of public works, I was tasked with hiring an engineering consultant to design and specify the repairs, and to oversee and inspect the construction. Unfortunately, we hired a consultant who had limited experience and expertise with the specific materials planned for this project. As a result, we paid for a generally ineffective repair project and endured a lengthy process to reach a financial settlement. Along the way, however, we learned valuable lessons about hiring a consultant, ensuring quality construction, and repairing concrete.

The original repair design called for removing the outer 4 inches of concrete from the spillway and training walls, then cleaning out and pressure-grouting visible cracks. The contract called for the application of a bonding agent to the sealed surface of existing concrete, and the installation of horizontal and vertical steel reinforcing bars spaced at about 16 inches. The spillway and training walls were then to be refaced with a shotcrete overlay approximately 4 inches thick. The shotcrete specifications called for a prebagged concrete mix, which would be applied in layers to reach the desired thickness. A natural gun finish was specified for the shotcrete overlay.

Within days of shotcrete application, new seepage was noticeable on the surface. The contractor suggested and tried a series of remedies, but without success. At one point, for example, a number of weep holes were drilled through the shotcrete layer in an effort to help direct water out.

By August 1995, when the shotcrete project was substantially complete, various defects were visible in the overlay. These included surface cracks in a pattern that followed the location of reinforcement, discoloration at the cracks, and water leakage through the cracks. The overlay's condition continued to deteriorate over the next four years, as the city, design engineer, and contractor disputed the causes of and liability for the problems.

Finally, in July 1999, the city retained Skokie, Ill.-based Construction Technology Laboratories Inc. (now CTLGroup), to conduct a new investigation. The firm is known nationally for its technical knowledge of concrete and its breadth of services, encompassing structural engineering, field inspection, and materials testing on-site and in the laboratory.

Document Review

The investigation began with a review of pertinent documents, including the contract specifications and correspondence between the city, design engineer, and contractors involved with the overlay. This review alerted the investigators to several problems:

  • The prebagged dry shotcrete mix should not have been used on a project of this magnitude. Because the maximum recommended layer thickness of this product is 2 inches, it would require three layers to produce the 41/2-inch dam overlay. However, multiple layers generally lead to high-porosity interfaces, so application of the total thickness in one layer is advantageous. A single layer is achievable with shotcrete.
  • A bonding agent, generally not used in shotcrete repairs and “not recommended” according to American Concrete Institute (ACI) 506R Guide to Shotcrete, was specified for and reportedly applied on this project.
  • Rather than a natural gun finish, the specifications should have required a trowel finish for the spillway face. This would have resulted in less abrasion and better resistance to the local soft water.
  • The section of the contract specifications on polyurethane injection to seal the concrete substrate against the passage of water should have more strongly emphasized the importance of complete sealing. Incomplete sealing of the leakage caused much of the subsequent trouble with the hardened shotcrete.
  • Field Inspection

    On July 22, 1999, I accompanied CTLGroup investigator Steve Gebler as he examined the training walls and spillway face. We observed several forms of distress, namely delaminations, spalls, soft surfaces, and cracks. Gebler sounded the walls with a 2-pound hammer to detect delaminations near the surface. He also used a magnetic device (pachometer) to locate and measure the concrete cover over the steel reinforcement.

    Both the east and west training walls exhibited cracks, some in a gridlike pattern that generally coincided with the embedded steel reinforcement. Delaminations about 1/2 inch thick were apparent in the upper section of the west training wall. Most cracks in both training walls had calcium carbonate exuding from them, indicating that moisture traversed the cracks.

    The spillway face showed considerable spalling and delamination; in some areas, there were depressions as deep as 2 inches where the shotcrete was missing. In some areas, the spillway face was soft and crumbly; in others, the shotcrete was soft beneath the delaminated surface.

    Lab Testing

    During the field inspection, shards of delaminated shotcrete and core samples were removed from both training walls and the spillway, then sent to CTLGroup's laboratory for examination. Test reports stated most of the core samples containing steel reinforcement had voids or porous areas behind the steel. The voids weakened the concrete section in these areas, resulting in a crack between the steel and the shotcrete surface. Movement of the soft water seeping through cracks and through porous areas of the shotcrete leached calcium from the hardened shotcrete, leaving calcite deposits at the surface adjacent to the cracks.

    The report also explained that porous areas between shotcrete layers would permit movement of water that would accelerate the leaching action. It stated further that the porous areas would be especially vulnerable to freeze-thaw damage. The conclusion was that the shotcrete overlay would continue to deteriorate over the course of a shortened service life.

    CTLGroup's investigation offered authoritative evidence that the overlay project's problems stemmed from a combination of factors: flawed design and material specifications, poor workmanship, and inadequate inspection and quality control during construction. The firm's work helped us reach a financial settlement with the responsible parties and reminded us how important it is to verify the competence of individuals or companies that seek to work for the city of Barre.

    It also taught us more about how to recognize and obtain quality concrete repairs. When performed by experts, field inspection, nondestructive testing, petrographic examination, and other lab testing of core samples provide information that can help determine what is acceptable, what is wrong, and who is responsible for problems.

    Reg Abare is the city engineer and director of public works in Barre, Vt.