The River Inside the Dam

When the United States Bureau of Reclamation sat down to build Hoover Dam in the early 1930s, its engineers did a calculation that changed how mass concrete is placed to this day. The dam would hold more than three million cubic yards of concrete. Cement gives off heat as it hardens, and in a mass that size the heat has almost nowhere to go. If the dam were cast as one continuous block, they found, the concrete would not have finished cooling for lifetimes.

It would have taken 125 years for the concrete to cool to ambient temperatures.

US Bureau of Reclamation, "The Story of Hoover Dam: Concrete."

And as it cooled it would have contracted, unevenly, and torn itself apart. The problem was not the recipe. It was the heat, and the fix was a way of building rather than a better mix.

So the dam was not poured as one block. It was built as a grid of separate vertical columns, each raised in five-foot lifts, and no column was allowed more than one lift every seventy-two hours. Shallow layers, spaced out in time, let each lift shed some of its heat to the air before the next arrived on top of it. That alone was not enough for a structure this thick, so the builders put a cooling system inside the concrete itself: coils of one-inch steel pipe cast into every form, more than 582 miles of it in all, chilled by a plant that could make a thousand tons of ice a day.

When the concrete was first poured, river water was circulated through these pipes. Once the concrete had received a first initial cooling, chilled water from a refrigeration plant on the lower cofferdam was circulated through the coils to finish the cooling.

US Bureau of Reclamation, "The Story of Hoover Dam: Concrete."

River water to take off the worst of the hydration heat, then chilled water to bring each block down the rest of the way. Here is the part that is easy to miss. The cooling was not only there to stop cracks. It was there to hit a number. Each column had to be brought down to a target temperature so that the contraction joints between the columns would open to a designed width, be pressure-grouted, and knit the separate blocks into one monolithic dam. Cooling was a measured process with a target and a thermometer, not a hope that the heat would sort itself out.

That is the whole discipline, and it is worth being precise about what you are fighting, because mass concrete cracks in two different ways. Early on, while the core is hot and the surface is cool, the two pull against each other, and if that difference gets too large the surface cracks. Later, as the entire mass cools from its peak back toward the temperature of the valley around it, it wants to contract, and the foundation and the older concrete beneath it will not let it, so it cracks through. You beat the first by holding the gap between core and surface small. You beat the second by holding the peak down and letting the mass cool slowly and evenly.

India codified all of this. IS 14591 is the Bureau of Indian Standards guideline for temperature control of mass concrete in dams, and it points to the same levers Hoover used: limit the placing temperature, limit the lift height, cool the interior, insulate the surface. In an Indian summer the first lever is the hard one. When the ambient sits above forty degrees, you cannot place hot concrete and hope, so the mix is pre-cooled, with chilled water and with flaked ice standing in for part of the mixing water, before it ever reaches the block. Hot-weather placement has its own code, IS 7861 Part 1, for exactly this reason. Sardar Sarovar, among the largest concrete gravity dams anywhere, was built with chilled concrete on the Narmada for the same physics that drove the pipes into Hoover.

The lesson: Thermal control is a construction decision made before the first pour, not a mix-design afterthought. You cap the lift height, you space the lifts in time, you cool the interior on a schedule to a target you can measure, and you grout the joints only once the block has actually reached that target. The dam does not care what you intended its temperature to be. It cracks on the temperature it actually reaches, so measure it, and design the sequence that keeps it in bounds.

Read more: Thermal Control of Mass Concrete: ACI 207, IS 7861 and IS 14591 →