Failure of the Kingston Ash Facility: What caused it and why it - PowerPoint PPT Presentation

Failure of the Kingston Ash Facility: What caused it and why it matters TVA Forensic Team’s “Creep of the Slimes/Static Liquefaction” Theory (Ash Dredge Cell 2 Failure) vs. “It’s the Water, Stupid” Theory (Clay Dike C Failure) by Barry Thacker, P.E. Geo/Environmental Associates, Inc. PEM Seminar ‐ 21 August 2009 – Lexington, KY

KINGSTON ASH FACILITY FAILURE The failure occurred around 1:00 a.m. EST, on Monday, December (Background Information from Forensic Report ) 22, 2008, when the north and central portions of Cell 2 at the Kingston Fossil Plant ash disposal site suddenly failed. An estimated 5.4 million cubic yards estimated 5.4 million cubic yards of material, consisting primarily of hydraulic ‐ filled ash and intermediate stage containment dik dikes, were released in a l d i progressive sequence of flow slides over a period of approximately one hour. pp y Ultimately, the flow slide would extend northward approximately 3,200 feet beyond the limits of , y the original ash pond over the Swan Pond Creek flood plain, a back water slough of the Emory River and into the former Emory River and into the former Emory River Channel of Watts Bar Reservoir.

TVA’S FORENSIC TEAM: First ‐ time in Missed c Missed clue #1?... ue #1?... Ho How could w could N history phenomenon, denoted as ash tr ash travel el near nearly 2/3 of 2/3 of a mile a mile “Creep of the Slimes/Static Creep of the Slimes/Static to the to the nor northw thw est fr st from f om failur ilure e Liquefaction”, is modeled using of of a a dr dredge cell outslope w ith edge cell outslope w ith undrained shear strength for a buried a nor a northeaster theastern orienta orientation? ion? slimes layer and the overlying sluiced ash, and predicts failure of the Hint: Dike C had a dredge cells. northwestern ‐ facing limb Dike C Dredge Cells

OPINION BY AECOM AS TO CAUSE OF THE KINGSTON ASH FACILITY FAILURE (From Forensic Report): The north end of Dredge Cell 2 was on the verge of failure due to the high stresses and creep in the loose wet layer of weak slimes. The deformation of the slimes in turn caused p y the overlying collapsible wet ash to liquefy. Failure of the Kingston dredge cells was sudden and complex in nature due its geographic setting and being built within the Watts Bar Reservoir after the lake was formed. It took a forensic type study to determine the propensity of the ash to liquefy at low strain levels when the material cannot drain and thus becomes of the ash to liquefy at low strain levels when the material cannot drain and thus becomes undrained, and to locate the slide plane in the unusual, creep susceptible, low undrained shear strength slime layer that underlies Cell 2. In AECOM’s opinion, subsurface conditions at the dredge cells were unusual and rarely found. The consequence of failure in the slimes led to the collapse of the dredge cell and loss of the saturated contents of the ash landfill due to the breach of perimeter Dike C. (NOTE: AECOM’s primary failure (NOTE: AECOM s primary failure surface was analyzed at the location of a former knob, whereas the thickness of ash is greater beneath Dike C’s northwestern limb.) th t li b ) Dike C

Initial Burst Location of Dike C (Based on “It’s the Water, Stupid” A Analysis) l i ) Initial Static Liquefaction Location of Dredge Cells (Based on TVA’s Forensic Dredge Cells (Based on TVA s Forensic AECOM Team ) RELIC SURVEY SURVEY DATA

CORROBORATING RESULTS OF ENGINEERING ANALYSIS USING FORENSIC EVIDENCE: My buddy, Ray Bob, offered some of his observations about the forensic study. He is proud to describe himself as, “a snuff ‐ dippin’, coal ‐ minin’, 6 ‐ foot ‐ 6, 250 ‐ pound, American Veteran”; and his friends call him “Tiny”. Tiny brought to my attention the Expected Failure Mode slide from the forensic study shown below that he says “Must be one of them cartoons they make in the forensic study shown below that he says, Must be one of them cartoons they make in Hollywood.” Tiny says to look real close at how real close at how the ash from the “ dredge sales ” supposedly pp y “lickerfide” and ran over Dike C.

Tiny gave me two Dike C figures/photos from the AECOM forensic report, which are shown at right. According to Tiny, “Do ya seed whar I drawt ‘ SEE ‐ DIK (THEN)’ and ‘SEE ‐ DIK (NOW)’? “ “If that ‐ there Hollywood cartoon from them forensic fellers was true, then the grass on top of the See ‐ Dik slope they say got flooded and pushed they say got flooded and pushed seven ‐ hundurd feet to the northeast by the tidal wave of lickerfide ash from failurt of the dredge sales would at least be tarnished and flat, and not as fresh as a Daisy Duke.” “At least in the Road Runner and Wile E. Coyote cartoons, they showt the coyote with matted fur, and seaweed stuck in his ear, after he gets hammered by the tidal wave ” hammered by the tidal wave.”

Tiny gave me two photos of the relic cattails from the AECOM report as shown below, and informed me that you can’t push a rope, you have to pull a rope to get it to move. In his words, “Do ya seed whar I drawt ‘CATTAILS (THEN)’ and ‘CATTAILS (NOW)’? Thur is no way on God’s Green Earth them dredge sales could push a clump of cattails two ‐ thirds of a mile up the holler to the northwest without destroyin’ them. Do them forensic fellers expect us to believe that the dredge sales had the force to push See ‐ Dik seven ‐ hundurd feet to the northeast, yet not hurt a hair on the fur of a cattail’s haid? not hurt a hair on the fur of a cattail s haid?... Yea, rite! The only way them cattails could Yea rite! The only way them cattails could have moved that furr a piece up the holler to the northwest is by ridin’ on the back of the sluiced ash after the northwestern limb of See ‐ Dik burst.” “Don’t them forensic fellers knows that a cattail is as fragile as… ummmmm, a cattail”? Don t them forensic fellers knows that a cattail is as fragile as… ummmmm, a cattail ?

Relics from Dredge Cell 2 Outfall Pipe Found Northeast of Original Location – illustrating that the dredge cells failed last after the bursting of Dike C undermined after the bursting of Dike C undermined the northeastern outslope of Dredge Cell 2 N Initial burst location of Dike C Cell 2 Outfall Pipe (Original Location) Location) Dredge Cell 2

OPINION BY BARRY THACKER, P.E.: Dike C burst at its northwestern limb due to artesian pore water pressures p p in the underlying sluiced ash; then it was like Dominoes, the game, not the pizza.

Piezometer locations and water level data provided by TVA/AECOM: For reference purposes, the ground surface at MW ‐ 15 is at elevation 771 feet; thus, water levels in MW ‐ 15 above elevation 771 feet represent artesian conditions in the sluiced ash. Trend in water levels in piezometers after last measurement on 11/19/08 validates Thacker seepage model. i f l 11/19/08 lid Th k d l Note: The alluvium beneath Dike C where MW ‐ 4B was screened was not included in the Thacker model.

LET’S NOW EXAMINE THE DETAILS: Missed Clue #2? Mitigation report prepared after a 2003 incident at the northwestern toe of the dredge cells concludes that a “blowout” was caused by “ excessive seepage and piping ”, yet y p g p p g , y internal drains were proposed to increase the rate of seepage – Is that logical? 2003 “Blowout” at toe of Northwestern Outslope of Dredge Cells Hint: Hinge No 1 Hint: Hinge No. 1 Hint: Bottom Ash Lenses

In reviewing the input data used in the AECOM seepage modeling, I see that AECOM used a ratio of horizontal hydraulic conductivity (kh) to vertical hydraulic conductivity (kv) = 1. Maybe, the engineers at AECOM do not realize the impact the black, more pervious bottom ash lenses, shown in the photo at left from the AECOM report, have on the kh/kv ratio; or how a high kh/kv ratio can in some circumstances cause significant uplift pore water can in some circumstances cause significant uplift pore water pressure in sluiced ash, hydraulic fill. Hinge (i.e. thin, potentially weak zone subject to failure by high pore water seepage pressures) I learned that lesson decades ago from a gray ‐ haired engineer with the Note: kh/kv = 100 U.S. Mine Safety and Health Administration (MSHA) who referred Administration (MSHA) who referred me to the results of the forensic study of the 1972 Buffalo Creek Slurry Impoundment failure, shown at right. One of the early practitioners in design of ash disposal facilities was Professor Arthur Casagrande of Harvard University. A 1971 design report he co ‐ authored states: “ When fly ash is deposited from a slurry in a pond, considerable segregation develops according to grain size p f y p g g p g g and specific gravity. The resulting stratification and loose structure produces relatively high horizontal permeability.“ I have found that kh/kv can vary from 1 to 100 at the same site, so I believe use of kh/kv = 100 for hydraulic fill in seepage analysis is prudent.