they had been experiencing major leakage in the reservoir and last year, they lined the reservoir to prevent it. I wouldn't be surprised to see that a defect in the new lining or in the reservoir itself caused the breach.
Instrument failure, an expression roughly equivalent to down-sized maintenance, got the blame:
"Gary Rainwater, AmerenUE chairman and chief executive, said it appeared that the plant's automated instruments had pumped too much water into the reservoir and caused it to rupture. A backup set of instruments should have recognized the problem but didn't, and the utility was trying to figure out why, AmerenUE said."
-- Associated Press, appearing in Los Angeles Times, Dec. 15, 2005
You're absolutely right.
However, it's a mountaintop reservoir that gets filled by pumping, so maybe inadequate overflow spillway.
My guess is that there had been some weakening of the walls, so it was being kept at lower capacity. However, failure of liquid level sensors and/or human error in programming the reduced level into the controls; especially, the backup controls.
As it turns out (According to the news reports I read this morning) the dam was not properly constructed. Supposedly, the construction documents show that the earth dam was built with a granite core with the rock that was excavated from the top of the mountain. Field reports from engineers on-site after the failure noted that the dam cross section was only earth with small rocks and boulders mixed in.
What happened to the 1 million cubic yards of granite that was excavated that should have been used to construct an impermeable core? Nobody seems to know.
I'm not sure how you construct an impermeable dam with granite - it should have had a clay core. But the granite probably would have been very good material for the shell of the embankment. Was the granite intact, or decomposed?
"Instrument failure, an expression roughly equivalent to down-sized maintenance, got the blame:"
"Gary Rainwater, AmerenUE chairman and chief executive, said it appeared that the plant's automated instruments had pumped too much water into the reservoir and caused it to rupture. A backup set of instruments should have recognized the problem but didn't, and the utility was trying to figure out why, AmerenUE said."
"However, failure of liquid level sensors and/or human error in programming the reduced level into the controls; especially, the backup controls."
"As it turns out (According to the news reports I read this morning) the dam was not properly constructed."
Hmmm. Maybe is wasn't an I&C (Instrument & Controls) fault.
Charles Morris, a professor at University of Missouri Rolla who teaches water resource engineering, hydrology and hydraulics, said if it was an instrument failure and water was spilling over the top in the way described by the company, he would expect the inspectors to see evidence of erosion all around the top of the site, not just in one particular area where the failure occurred.
He has inspected many dams that have failed and said he thinks a more likely scenario based on the history of the dam is that “piping” has been carrying away material over a long period of time, weakening the reservoir's holding walls.
He explained that all dams made of earth materials will lose some water no matter what is done. The goal is to make sure there is not too much leakage. If there is, then material begins to be carried away, bit by bit, weakening the structure. This is referred to as “piping.”
He pointed out that the dam had been losing two foot of water over it's 55-acre surface for quite some time. “That's not an insignificant amount of water,” he said.
When piping occurs, the structure can settle in, which means water may start flowing over its top - but that is because a weakened area in the dam has settled in.
An annual survey can help point out whether that is occurring. It was not known at press time whether that was part of the company's routine inspections.
Morris also suggested that a record of water levels along the affected rivers could be used to backtrack and simulate the rate of failure. That will provide additional clues to what happened.
Morris was among those who saw the structure when it was first built. The top of the mountain was blasted out and the walls built up, he said.
Among those checking the failed reservoir were members of the state dam inspection staff. Chief engineer James Alexander said he was "shocked to learn that the (failed section) was not granite, but mainly a mixture of soil and rock fill." He said most of the rest of the upper reservoir's wall was made of granite.
If there is active settlement on the crest of the dam - which I assume there was - then the water could overtop the dam in one location and cause a failure. A low spot normally forms directly over the point where piping is occuring. Sometimes, if you have a lot of piping and internal erosion, then you will even see sinkholes form in the embankment. Normally, frequent inspection of the crest elevation of dams is required (by the regulator of the dam) to alert the dam owner of the situation. The crest would then be required to be brought back up to proper levels. FERC was the regulator of the dam and should have been on top of this. An emergency action plan should have been activated to prevent the failure and to evacuate the area.
The key issue in the failure of Taum Sauk Dam is whether or not the dam overtopped during filling. As someone else noted, there should be evidence of any overtopping. Any emergency spillway capacity should have exceeded the maximum filling rate. This may be called an "instrumentation malfunction," but in reality, the condition of the remaining embankment may prove this event to be a result of INSPECTION failure.
The more that you read about this failure, the more questions arise.
At Taum Sauk the upper reservoir is a 32-acre pool constructed atop a mountain, at about 1,500 feet above sea level, and the lower reservoir, impounded by a 60-foot dam, covers about 370 acres and has a usable storage capacity of 4,350 acre-feet of water. The two are connected by a pressure tunnel and conduit, with a pumping and generating station on an open channel running to the lower pool.
Seems that there was no overflow spillway. Others have reported that there was a maximum fill level that was 6 feet below the top. If the reservoir was filled in 16 hours and is 32 acres across, that would calculate to about 1 ft of water rise in 7 minutes. If the staff was not alert, the reservoir would have overflowed in just 42 minutes. 1.5 billion gallons capacity in 16 hours equates to 1.476 million gallons per minute fill rate. It seems like an accident was waiting to happen.
If a big portion of the retaining wall was mostly soil and smaller rock, it was likely doomed once too much water was pumped into the reservoir”, said Charles Morris, an environmental engineering professor at the University of Missouri-Rolla. Like earthen levees that crumbled during the Great Flood of 1993, soil-based retention walls will erode when overtopped, he said.
If you look quick, you can still see the google cache of the geo--synthetics website.
Interesting quotation on the website:
"Since it's construction in the 1960's, cracks had formed in the 1.3 to 1 side slopes, resulting in two feet of water loss from the upper reservoir every day."
AmerenUE Taum Sauk Reservoir Lining Project
14 pictures and story about project to stop leaks in upper reservoir.
At Taum Sauk the upper reservoir is a 55-acre pool.
If the reservoir was filled in 16 hours and is 55 acres across, that would calculate to about 1 ft of water rise in 12 minutes. If the staff was not alert, the reservoir would have overflowed in 72 minutes.
"The more that you read about this failure, the more questions arise." -- bimr.
It seems like a total fiasco, from initial construction errors, inadequate repairs, lax state & federal regulators, human error & equipment failures [lack of maintenance & periodic testing]. Kind of like the Conemaugh Lake dam failure/Johnston flood of 1889. Not much progress in 100+ years!
The original design had the maximum water level 6 feet below the top. Was this level ever reduced when leakage persisted after repairs???
AmerenUE experienced an overtopping event on September 25th, 2005 that should have raised alarms bells. Then, just 2 months later, a catastrophic overtopping failure occurs on December 14th. Somebody definitely will have some explaining to do.
Thanks Boo1 for the update. It would seem to be common sense to put the high water sensor/emergency shutoff lower than the reservoir crest....(maybe they were using the wrong datum)
