On February 13, 2017, all of California watched in disbelief as water overtopped Oroville Dam through its “emergency spillway.” But, the real emergency started a few days earlier when a small hole in the dam’s main spillway allowed the torrent from a controlled water release to get under the spillway’s concrete and peel it from its foundation. That is not supposed to happen, even on a 50-year-old dam.
Like so many other professional civil engineers, I can speculate on what may have caused this major failure.
Based on my education and professional experience, I speculate Oroville Dam’s main spillway just may have failed due to two contributing factors:
- Microcracking of the spillway concrete due to the harsh environmental extremes (freeze/thaw and wetting/drying) it has experienced over the last 50 years, and
- Cavitation occurring on the surface of the concrete as the high-velocity flows raced down the vertically-curved spillway
This speculation is not based on any real familiarity with the spillway construction or with maintenance and testing practices by the California Department of Water Resources, only by studying photos and reports in the news media. So, I could be completely wrong.
Microcracking is the tiny cracking of dried cement paste and aggregate (rocks, gravels, and sands) that make up concrete. Many times, the cracking can only be seen under a microscope, so they may not be detected by a visual inspection of the spillway concrete, especially if the cracks are only in the cement, and do not extend into the aggregate. The photo below depicts microcracking in the cement paste only.
According to researchers and experts, microcracking can be caused by a variety of fractures, including repeated freeze/thaw cycles, excessive heat during cure, or repeated wetting/drying cycles. This 2004 Purdue University research report describes the connections in detail.
Developing this concept as it applies to Oroville Dam, the surface of the spillway concrete can be subject to diurnal freezing and thawing cycles during winter. This could create microcracks in the surface of the spillway concrete, weakening it at least slightly over the years.
Next, the spillway faces the southwest, subjecting it to the direct sun in hottest daytime temperatures of the summer. Over the last several years, the drought likely exacerbated the desiccation of the concrete, again potentially contributing to microcracking.
Finally, DWR investigators will also want to go back to original concrete pour data to find out if careful attention was paid to internal concrete temperatures as the concrete cured. If it got too hot, that could have contributed to the failure 50 years later.
But, concrete microcracking can’t be the whole story. Some sort of forces had to be at play to physically pop pieces of concrete out of the spillway. That is where cavitation comes in.
Recent news reports about the Oroville spillway failure included theories about the cause proposed by one of my former graduate school professors, Dr. J. Paul Tullis, of Utah State University and the Utah Water Research Laboratory. Dr. Tullis blames cavitation.
Cavitation occurs when water flows at high velocity (like the maximum spillway flow of 150,000 cfs) along a smooth surface causing a vacuum at the interface between the water and the surface. It is especially prevalent on an “inside” curve where gravity or centripetal force are tugging the water away from the surface, like when water cascades over the top of a vertical curve. The vacuum pulls dissolved gases out of the water and forms tiny air bubbles. When the water slows down again or becomes turbulent, those gas bubbles collapse or implode, usually with a lot of energy. When you are dealing with boatloads of water in big structures, like dams similar to Oroville, the released energy is multiplied enough to blow holes in concrete.
Over the last 40 years, Dr. Tullis’ computer models have simulated this path of concrete failure on several dam structures throughout the world. Here is a link to a listing of some of his writings.
So, my speculation is that if you apply cavitation dynamics to aged concrete filled with microcracks, maxing out the flow of water down the Oroville Dam spillway for the first time after years of on and off drought may just have been enough to pop Prius-sized pieces of concrete into the air. Once the initial hole was opened up, the energy of the turbulent water just peeled away most of the remaining downstream concrete.
It will be interesting to hear what the investigators find out. Could it be microcracks and cavitation? It may be a few years to get that answer.