AmerenUE was told reservoir might fail

Monday, January 16, 2006

ST. LOUIS -- Less than three months before the Taum Sauk reservoir collapsed, a plant superintendent sent an e-mail to several AmerenUE supervisors warning about a possible collapse, according to a newspaper report.

The e-mail was prompted by a surge of water that overflowed the northwest wall of the reservoir on Sept. 25, the St. Louis Post-Dispatch reported in Sunday's editions.

Company e-mails also show that the plant had problems for months accurately gauging water levels in the reservoir.

Richard Cooper, superintendent of Ameren's Taum Sauk Hydroelectric Plant, said in his Sept. 27 e-mail that the overflow had washed away rock at the base of the reservoir wall, leaving trenches a foot deep in some areas.

"Overflowing the upper reservoir is obviously an absolute 'NO-NO,'" Cooper's e-mail said, noting that it would "cause eventual failure ... Those kind of headlines we don't need."

Ameren never reported the September overflow to federal regulators, who now are investigating, the Post-Dispatch said.

On Dec. 14, the reservoir ruptured in the same spot as the September overflow, releasing more than a billion gallons of water that caused extensive damage Johnson's Shut-Ins State Park and swept park superintendent Jerry Toops' home away. Toops' three children were hospitalized but have since been released.

The internal e-mails and documents were given to the Post-Dispatch by Don Giljum, business manager for Operating Engineers Local 148, the union that represents Taum Sauk workers.

"Our operators were letting (Ameren managers) know that something was wrong," Giljum said.

Federal regulations require companies to report conditions affecting the safety of a project or its operation "as soon as practicable" after discovered.

But the Federal Energy Regulatory Commission, which regulates Taum Sauk and is investigating its failure, wasn't aware of the Sept. 25 overflow until last week, according to a letter issued to Ameren on Friday. The agency gave Ameren a week to respond.

Ameren spokeswoman Susan Gallagher said the company didn't consider the incident reportable. She said the company took action after the overflow, but refused to give specificcs, citing the FERC investigation.

Dam expert Charles Morris, a civil engineer at the University of Missouri at Rolla, said any previous overtopping at Taum Sauk should have been a strong warning of danger.

"That structure was not designed to be overtopped. Period. If it did so, (Ameren) should have gone berserk," Morris said. "If they knew about it, they were very negligent."

The 55-acre Taum Sauk reservoir, near Lesterville, Mo., is the upper part of a pumped-storage plant designed to meet peak daytime electricity needs. At night, Ameren pumps water from a 370-acre lower reservoir to the upper reservoir.

The upper reservoir leaked for decades, forcing Ameren to install an impermeable liner in 2004. But after the liner was installed, the utility had problems with instruments that measure water depths and are supposed to trigger alarms if water levels rise too high, according to company e-mails.

One of the instruments measured elevation by the pressure of the water column over it. In an Oct. 7 e-mail Cooper said that instrument was malfunctioning and he had scheduled divers to inspect it.

"This bend in the pipes gives us a false reading and causes the reservoir level to look lower than it actually is," Cooper wrote. "Until these pipes can be re-attached we are lowering the pumpback shutdown setpoint to 1594 down from 1596. We want to give ourselves enough cushion so that we won't pump over the reservoir walls."

At an elevation of 1,594 feet above sea level, employees thought the reservoir was within 3 feet of the crest, which is 1,597 feet.

But according to other Ameren data obtained by the Post-Dispatch, the instruments may have registered a 3-foot cushion that didn't exist.

On the day of the collapse, Ameren's chief executive, Gary Rainwater, said at a news conference that an instrumentation failure led to overpumping, then overtopping, of the reservoir.

The data show that Ameren had been pumping to 1,594 feet six times in the two weeks before the failure.

"If (the levels) were consistently in error, then they overtopped previously -- all those other nights that they pumped it up to that level," Morris said.

Respond to this story

Posting a comment requires free registration: