Search This Blog

Saturday, February 18, 2017

Hydroelectric near-miss: Oroville Dam ***Update #2***

Ha!  This has been serendipitous...  I've been doing a series of posts about dam failures and near misses.  The posts are here, here, and here (with another post on the St. Francis dam in draft) - when lo and behold, we have an ominous near failure in real-time on the evening news right now.

As I've said in the past, I'm not a hydro engineer.  However my line of work lends itself to appreciating thoughtful and solid engineering.  It's therefore easy to note when the engineering is less than stellar.  With regard to hydro, one thing I've been surprised to learn is how dependent large dams are on using the reservoir as a buffer from large inflows.

I get that this works most of the time.  Let's say you are a dam operator (an organization that operates a dam), and there is a huge snow pack up in the mountains.  You monitor this snow and its water content constantly at several hundred points in the watershed.  When it looks like things will begin warming up, you begin drawing down the reservoir to make room for the estimated snowmelt runoff.

This works - until it doesn't.

I'm surely not the only one who has noticed that weather doesn't always cooperate with us.  Given that fact, I find it shocking how little excess water many large dams are able to safely bypass.  Glen Canyon dam nearly failed in spring 1983 when warm weather and heavy rain produced a sudden runoff, with the reservoir already full.  This year, it's another dam's turn.

This past week, we were able to witness hydro history, as authorities evacuated 188,000 people from the California cities of Oroville, Marysville, and Yuba City.

The reason for the evacuation?  Fear of imminent failure of the Oroville reservoir.

Below is an image of the Oroville Dam and the cities immediately downstream.  Oroville Dam is just right of the center near the top of the image.  Flow of the river after it exits the dam is north to south. The state capital of Sacramento is further to the south, and is apparently in no danger.   Click the image to enlarge it.

Below is a closer view of the dam.  The main spillway is the white concrete chute.  It has gates that allow bypass flow to be adjusted.  The emergency spillway is a fixed weir that when reservoir level rises high enough, water flows over, rather than over the dam itself.

So what happened?  In a word, weather.

Rain, coupled with the inability of the dam to bypass large amounts of runoff, nearly led to loss of the reservoir.  The dam itself was not in danger of failing.  The bypass works (called spillways) were damaged and inadequate for the amount of runoff.

On Feb 7, 2017 a crater appeared in the main spillway while the dam operators were releasing 50,000 cubic feet of water per second.  Continuing high inflows to the reservoir forced the operators to continue using the main spillway.  In just three days (Feb 10), the crater had grown to 300 ft wide, 500 ft long, and 45 ft deep.

Below are images of the initial damage to the main spillway.  The spillway was briefly taken out of service for an inspection, but had to be quickly re-opened due to rising reservoir levels.  There was no time for repairs.  The soil on the right side of the spillway, downstream of the hole has eroded and undercut the spillway.

Dam operators were hoping to lower the reservoir level in advance of additional runoff, but erosion of the hillside around the damaged main spillway was endangering transmission line towers, and so the spillway flow was reduced from 65,000 to 55,000 cubic feet per second.  This caused the reservoir level to rise to the point where the emergency spillway would come into service.

On Feb 11, the emergency spillway saw its first-ever use since the dam was completed in 1968.  The emergency spillway was "designed" to handle 250,000 cubic feet per second of excess water flow. After one day of 12,600 cubic feet per second (just 5% of "design" capacity), the hillside downstream of the emergency spillway showed so much erosion that evacuation orders were issued on Feb 12 for low-lying areas of Butte, Sutter, and Yuba counties.

Below: Feb 11 - Oroville dam emergency spillway first use.

Below:  Oroville dam before water flowed over the emergency spillway, which is in the foreground.

Below:  Erosion following just 24 hours at 5% of the rated flow of the emergency spillway.  The emergency spillway is the concrete weir at the bottom left.  Note the finger of erosion in the bottom left corner has almost reached the base of the weir.  Further erosion would undercut the weir, causing a large, uncontrolled release of water.

On Feb 12, the main spillway flow was increased to 100,000 cubic feet per second, even though it was badly damaged by this point.  This was done to prevent imminent failure of the emergency spillway.  The reason for the hasty evacuation orders was the probable failure of the emergency spillway concrete weir, due to erosion undercutting it.

The lower half of the main spillway at this point can probably be considered a write-off.  One side is missing most of the supporting hillside, and the damage is ongoing as I write this.

Below:  This is what 100,000 cubic feet per second of water looks like going down a spillway that has already suffered massive damage:

Things take an interesting turn though.  The dam was re-licensed in 2005, and three environmental groups filed a motion with the Federal Energy Regulatory Commission (FERC), to cover the emergency spillway with concrete.  These groups warned that erosion might cause the emergency spillway (and the entire reservoir) to fail.  The motion was rejected by FERC after several water agencies disagreed with the motion.

Here's some awesome reporting on the failure to upgrade the spillway by the San Jose Mercury News.  Oddly enough, none of the water agencies who refused to strengthen this reservoir are inside the flood path.  Go figure...

Meanwhile, we continue to wait for the relentless rain to end, so that the damage to both spillways can be assessed, and eventually repaired.  It will be a big repair project - bigger even than covering the emergency spillway with concrete.  For now the emergency seems to have receded, and people downstream of the dam have been allowed to return to their homes.

All of this begs a question:  Where did the the emergency spillway 250,000 cubic feet per second "design capacity" come from?  Was it a number some engineer grabbed out of the air?  Clearly this is an under-designed flow path...  Therefore we have to conclude the "design capacity" of the emergency spillway is based on either poor or non-existent engineering - And that is really, really shameful.  Oroville dam is run by and serviced by a government agency, so I don't anticipate anyone being held accountable for this fictitious value of 250,000 cfs of water.

FERC, to their credit, has belatedly ordered an investigation into this event.

The Federal Energy Regulatory Commission also ordered the Department of Water Resources, which manages the dam, to perform a “forensic analysis aimed at determining the cause” of the failures in both the main spillway that fractured on Tuesday.

To ensure objectivity, “the forensic analysis must be performed by a fully independent third party with no previous involvement in assessing the spillway structure at this project,” the letter reads.

Finally some wisdom...


It appears that the weather has finally let up, allowing the dam operators to finally shut the main spillway down for a look-see.  As suspected, there is a great deal of damage.  A great deal of the right hillside is gone, as well as the middle section of the main spillway.  Tons of debris is blocking the river.  I'm not sure what effect, if any, this debris will have on the power plant output.  That water has to flow over/around the debris, and I don't know if that causes issues for the turbines or the surrounding environment (if there is any remaining).

Pictures follow:

Note the scale of the damage.  That's a helicopter flying into the new canyon.

Below is video taken from the above helicopter

Lastly, in the event you were curious about how the Oroville reservoir failure might have unfolded, watch the video below on the Auburn Cofferdam.

A further update, March 26.  The requested investigation report for FERC has arrived (pdf).  It's a long read, and I'm not familiar with a lot of the terminology used by the investigators.

A brief summary of their findings:

  • The Main spillway needs a short-term emergency repair, to get through next year.  Time is short to decide what to do, plan repairs, award a contract, and make repairs before next rainy season starts.
  • Use of the Emergency spillway must absolutely be avoided.
  • The apparently undamaged sections of the main spillway are probably damaged
  • There are voids under the main spillway concrete, which have been backfilled with clay (should be concrete).
  • The main spillway thickness is 12", which these engineers believe is too thin.  It's also thinner in places than 12"
  •  There needs to be a long-term solution, which will probably involve replacing the main spillway, and it's important to keep the long-term and short-term repairs segregated. 
***Second Update***

It seems that I was correct that nobody will be held accountable for this mess.  The State of California and Federal Government have claimed "security" is the reason that the investigative report won't be released.  More likely they are trying to keep bumbling incompetence from public view.  

The optics on this behavior aren't good at all.  It looks like a cover-up.  They might also like to prevent hard questions from being asked about maintenance and operation at many other dams.
If you want some light reading to confirm my opinion that these guys are engaging in a cover-up, someone else does too.  Robert Bea of the Center for Catastrophic Risk Analysis at UC Berkeley performed an RCFA (Root Cause Failure Analysis) on the spillway failure.

He gathered the available dam construction information before it was all taken offline by the State of California and the Federal Government.  His assessment is that the original construction was inadequate, and that the spillway was doomed to fail.  Here is his 78 page Root Cause Failure Analysis  

No comments: