Thread Status:
Not open for further replies.
  1. MortarBoarder

    MortarBoarder Member

    In the main thread, the joints were described as step joints. Perhaps the leakage through those joints are expected source of the water. 100,000 cfs is a lot of water.

    • Useful Useful x 1
  2. Balkan

    Balkan New Member

    This is a very 'common sense' observation.
    However, divergence concentrates internal stress of a structure, i.e. makes it fragile. Not to mention that thickness differences of concern question compliance to the design.
  3. CRM114

    CRM114 Member

    From deep in, Drawing A-389-4:


    This doesn't mean this is the way it ended up, but this is what they were thinking at one point. It sure does match the photo evidence, and the text descriptions, right down to the polyethylene sheet. We know the VCP pipes got resized to six inches at the recommendation of the Oroville Dam consulting board. I'll stand corrected, there is no gravel over the top of the pipe and there is no trench. If it really was built like this the perforations may only be 1/2 to 3/4 effective. This is an odd design, and could explain these cracks:

    Oroville Dam Drains in The Spillway Walls - How Do They Work?

    Also, the pounding of the water above puts a direct point load on the pipe from the concrete as does the rock below. Gravel all around would distribute the load. VCP pipe is subject to cracking.

    This is how it's done these days:


    The foam is for freeze protection and probably would not be necessary in Oroville. Sand is standard as a filter around the gravel to prevent fines from being removed from the foundation. In hard rock it's probably not necessary, but as we have observed, it is questionable whether the rock here is hard.

    Attached Files:

    Last edited: Feb 21, 2017
    • Informative Informative x 3
    • Like Like x 1
  4. wrorke

    wrorke New Member

    i apologize for lack of picture but I am on road w only tablet.

    In a prior post i referred to the concrete under the sidewall as a "footing". That is inaccurate. A footing has a horizontal surface on the bottom to support loads without those loads generating sliding forces. This especially important on slopes.

    The concrete under the sidewalls (and laterally doweled into the outer slabs) appears to slope with the hillside without any flat steps on the bottom side. Instead of being a footing it might be more accurate to characterize it as a "grade beam".

    It appears resistance to downhill creep was provided by the bolts along the outside and the rebar angles beneath the adjacent slabs.

    In the detail in the post above, note the upper slab could press against the downhill slab (consistent w concrete being strong in compression) but the upper slab did not restrain the downhill slab from sliding (thus avoiding possible tension in the concrete).

    Could a phenomenon such as thermal ratcheting gradually shift the lower slabs downhill and open gaps? The less sloped upper spillway would be less prone to sliding than the portion below the knee. Sliding of the lower section could create gaps and leaks in the knee region.
  5. wrorke

    wrorke New Member

    My prior post erroneously assumed facts not in evidence. Was the concrete poured directly on the bare rock cut or was there an intervening fill to bring up the low spots in the uneven rock cut? Fill would be bad for possible downhill shifting. Uneven bare rock could provide some "tooth" which would contribute to resistance against slippage.

    DWR must have tracked spillway movement via periodic surveys.
  6. sushi

    sushi Member

    In this image it is evident that the "grade beam" exhibited significant variation in width. The image also shows what appear to be bolt holes which would normally be found well set back from the outer face of the beam. In the image the bolts, or other form of reinforcement is completely missing and the bolt holes are fully exposed with only the half diameter being visible on the side face of the grade beam.

    I have been puzzling over this arrangement for some time. Your comment opens up the possibility that this section of the spillway "migrated" down-slope. One possible effect of that migration was to cause the outer face of the grade beam to shear off due to tension.

    The problem with this interpretation is that it is insufficient to serve as the initial cause of failure. It is more likely to be the result of secondary failure once the original trigger fault took place.

    Also of interest is the fact the reinforcement rods beneath the sidewall extend straight down. I interpret this to have been a consequence of the fact that whatever they were anchored to fell away beneath them rather than making them subject to a lateral displacement.

    A possible sequence of events is as follows:
    1) The initiating failure occurs. Water in volume enters beneath the slab
    2) The flow begins to erode the weathered gabbro visible in this area [ one caveat - we do not know what the subsurface was composed of. The adjacent hillside appears to have the characteristics of glacial till rather than any form of weathered igneous rock.]
    3) As the anchors are undermined, the 52' section of sidewall and the 9' 4" margin slab slips down hill and creates increased opportunity for flow beneath the slab. This movement may have been quite minor, a few millimeters or centimeters, but the displacement is sufficient to permit the entry of high speed water in volume. The Oroville area is renowned for its use of hydraulic mining.
    4) The water cuts an erosion channel to the SE working beneath the sidewall.
    5) The outer depth of the grade beam shears off in tension along the line of bolt holes. This is a relatively common failure location. Whatever was connected to those vertical anchors either erodes away, or drops away, or some combination.

    • Like Like x 1
  7. sushi

    sushi Member

    How do you know this?

    If you review the photographic evidence in the thread you will find little basis for your assertions.
  8. MortarBoarder

    MortarBoarder Member

    #11 anchor bars in grout were installed in holes drilled into bedrock. This has been covered in the original thread, and remains of anchors found in the debris, on 10 foot separations.
  9. MortarBoarder

    MortarBoarder Member

    Not proved. Erodible bedrock or shifting fill are possibilities. There will be reports later with the results of expert analysis of the cause of the failure. We're trying to study the drains here, the main thread has wider discussion.
    • Like Like x 1
  10. Pozzolith

    Pozzolith Member

    I looked through those dwgs the other day and went right past that joint detail. That dowel bar is close to the bottom. If the gravel base were eroded, load transfer at the joint could spall the concrete off below the bar.
    • Like Like x 1
  11. Pozzolith

    Pozzolith Member

    • Agree Agree x 1
    • Informative Informative x 1
  12. CRM114

    CRM114 Member

    It's good to compare these details with the photos compiled by wrorke:

    Oroville Dam Spillway Failure

    It looks like they may have added dowels to the longitudinal expansion joints between this study and final construction, but a lot of those photos seem to match these details.
  13. CRM114

    CRM114 Member

    A zoom of a pic by dierdre on the historical pics thread:

    Shows relatively low flow, but drains appear mostly dry or just trickling.


    But it's useful to compare to recent pics where much more flow is evident:

    Previously, I did a comparison between 1986 and 2017:

    It's possible that between 1969 and 1986, flow went up.
  14. sushi

    sushi Member

    Great find. I was looking at that and trying to decipher what it represented.

    What I do not yet understand:

    1) The vertical dimension form top of slab to the C/L of the dowel is indicated as 1'-0"
    There is no dimension given for the depth of slab from C/L of the dowel to the bottom of the slab. By eyeball this dimension appears to be no more than 2" for a full slab depth of 12" The figure of 15" for depth of the cast slab has been thrown around. This 15" depth now seems incorrect.

    2) Regardless of the given dimension, the photographic evidence strongly suggests that this would have been a minimal distance and that slab pour would conform with the bedrock and act to compensate for variances in the bedrock surface.

    3) Para 2 suggests that in those areas of relatively soft and friable material (I will call this type of subsurface "garbbro") the contractor would be able to rip the surface and bring it to within a relatively close tolerance of the desired depth. This depth would be planar across the spillway in the area of the gabbro. In those areas where the subsurface was composed of a much harder, more competent rock (I will call this type of rock "chert") it would not be so easy to achieve a smooth and uniform surface. The slab in this area appears to vary considerably in depth.

    4) It therefore seems to be a valid conclusion that the spillway is composed of two distinct slab types: 4.1) the relatively uniform gabbro slab which is situated over areas of subsurface which are known to be subject to weathering such that the uniform and smooth slab base is bonded to less competent rock. This raises a question as to the strength of that bond; and 4.2) the areas of chert in which the rock will have been resistant to ripping, in which it would have been much more difficult to obtain a uniform and relatively smooth surface. The slab in these areas would exhibit the variable depth noted in the photographic evidence, and would likely achieve a bond with a much stronger grade of rock. In addition to this bond, the slab in these areas would be mechanically locked to the subsurface due to the slab pour conforming to the high tooth of the rock surface.

    5) This perspective suggests a number of variable stress risers present in the slab. It is noted that the spillway failure occurred in an area of the slope in which the material on each side of the point of failure is the reddish material which appears to be a loose and unconsolidated aggregate, easily eroded. This perspective also lends weight to the suggestion raised by wrorke with respect to possible slip.

    6) From the C/L of the VCP to the joint between slabs is given the dimension 2'-0." I do not understand the reason for this dimension. Does it suggest a relationship between slab joints and the first diagonal in the french drain?

    7) There is a dimension of 6" shown from the slab joint to some variation in the slab surface downstream of the joint. There is a second unreadable dimension given for something indicated just beneath the slab surface. What might this represent?
  15. Balkan

    Balkan New Member

    This is a very reasonable assumption, but let's put this model into the light of above-mentioned drainage issues. We shall agree that the spillway drainage is composed of two distinct water seepage-collection conditions. One for 4.1, where gabbro lays, and another for 4.2 where the slabs are locked to rock subsurface. Thus, quantities of drained water shall significantly differ between the two slab types. IMHO.
  16. Pozzolith

    Pozzolith Member

    Best I can see after blowing up the PDF, it looks to be 1/2". It looks like the edge of the slab downstream of the expansion joint is feathered 1/2" below that of the upstream edge for a 6" distance. Perhaps to be sure not to present an upraised edge to the flowing water?​
  17. CRM114

    CRM114 Member

    I think the spillway was running at 55,000 CFS when the hole developed, but yes, it's still a lot of water. From all the photo evidence, it's apparent to me flow in the drains goes up with more flow in the spillway. There is some evidence that flow in the drains has gone up over time, imo. Stagnation pressure is the flipside of cavitation:

    Various excerpts from:

    It's really only been since 2007 or so, as far as I can tell that stagnation pressure has been quantified.

    The evidence suggests Oroville had joint sealant, but no waterstops per se.

    Water is flowing down the spillway at speed V, then hits a small bump. The stagnation pressure can be as high as V^2 / 2 g, where g is gravity, or the water velocity head for those familiar with the Bernoulli equation. Velocity head is converted to static head that acts as a hydraulic jack on the bottom of the slab (if not completely relieved by drains). If the displacement is the reverse in the above pic, pressures tend to go negative, and water can actually be sucked from under the slab - if bad enough, and there is no source of air to relieve the negative pressure, cavitation can occur. But the high flow in the drains may suggest cavitation is not the issue.

    In stagnation, hydraulic jack occurs, meaning the high pressure develops under the slab, and it can exacerbate over time, if erosion is occuring:

    So it can be longitudinal (down spillway cracks or joints), but lateral or transverse crack are the worst. Erosion can create a void that creates even more jacking pressure. Spalling can cause it:

    If spalling occurs on the upspillway slab at a lateral joint, it can create stagnation pressures. Also, it doesn't take much of a crack to generate stagnation:

    With regard to drains:

    Stagnation can pump a LOT of water beneath the slab:


    Unit discharge is CFS per foot of crack. My rough estimate is the velocity was around 60 40 ft/s at time of failure, possibly a little more. (90' 178 feet wide spillway, flowing 8 feet deep at 55,000 CFS). The Bureau cautions the numbers in the chart above may be high, but my estimate of what is coming out each of those sidewall drains at ~50,000 cfs is on the order of 1-2 cfs. The chart says 10-20 13-26 feet of the right crack could produce that. What isn't relieved by drains, tends to build up as jacking pressure:


    The media is speculating on cavitation, but given the drain flows, cracks, joint details, and various pics, stagnation should be considered as well. There certainly are clues:

    Rock anchors resist jacking pressure uplift, but if they aren't in something solid....

    The apparent lack of reinforcement on longitudinal joints didn't help resist jacking pressures if they did develop:
    Last edited: Feb 23, 2017
    • Informative Informative x 1
  18. MortarBoarder

    MortarBoarder Member

    Usually, the load across the joint is carried by the overlapping lip, which is concrete under compression. No problem with that. To apply downward pressure on the dowel bar, the lower slab's support would have to be significantly eroded, because the leverage of the lower part of the slab would continue to try to lift the top of the slab until the fulcrum was far enough down the length of the slab.
  19. CRM114

    CRM114 Member

    I missed this when posted, but note that the pipes are likely at different grades (slopes), giving them different carrying capacity. From all the evidence, the collector pipe is at the slope of the spillway.
  20. deirdre

    deirdre Moderator Staff Member

    Moderator Note - deirdre

    Just a reminder, this thread is specifically about "how the drains work". It is not about 'what caused the spillway to fail'.
    As Mick mentioned a few days ago

    • Like Like x 1
  21. Pozzolith

    Pozzolith Member

    Some more important details from dwg A-389-4 that answer some questions. (page 151)

    sec f-f.
    Here, we learn that the longitudinal contraction joints are 20'-0" wide and the wall footing extends 9'-4" into the chute. From that, the overall inside width of the spill way would be 178'-8".

    detail b.

    This detail shows the slab thickness to be 1'-3" and the reinforcing bar to be #5 (5/8" dia) and spaced at 12" each way. Also, at the anchor bar locations, there is 3'+ wire mesh mat buried in the slab. Presumably, that would resist pullout of the anchor bar.

    edge ex jt.
    Here, we learn that the wall base slab (wall footing) varies in thickness as suspected by the photographs. It also proves that the wall footings were placed prior to the slabs because of the way the joint is offset.

    sec e-e.

    Here, we learn the wall height to be 16'-0".

    sec g-g.

    This view shows the collector pipe (6" shown, but suspected to be 12") to also be vitreous clay pipe (VCP). To me, this is important, as that type construction must have stable bedding. Otherwise, the joints will open up and allow leakage that could exacerbate a washout.

    Finally, the notes on the dwg specify 3,000 psi concrete, and specify the reinforcement to be covered by 3" of concrete unless otherwise noted.
    • Like Like x 2
  22. CRM114

    CRM114 Member

    upload_2017-2-24_21-8-58. upload_2017-2-24_21-9-47.

    Cracks are herringbone pattern, and spaced about 20 feet apart. (slab panels are 50 long in flow direction per Google Earth). I'm not saying it's underdrains, but....

    Attached Files:

    • Like Like x 3
  23. chriswlan

    chriswlan New Member


    I need to post about the drains UNDER the chute pavement: this thread title is about the drains in the walls... Where should I go?

  24. deirdre

    deirdre Moderator Staff Member

    the Main spillway drain system includes under the cute. Please read the thread through before posting. We have enough repeat posts happening.

    PS thank you though for asking.
    • Like Like x 1
  25. MortarBoarder

    MortarBoarder Member

    Has anyone found a drawing of how the drains under the spillway are connected to their discharge point on the wall? We haven't seen pipes running along the outside of the wall, which implies that they're somehow embedded inside the wall and running for dozens of feet without weakening the wall. The drains also are somehow crossing the expansion joint which they're shown as being upstream of, but that might be happening under the fill which is next to the wall.
  26. sweepleader

    sweepleader Member

    Yes, see sushi's post above #126, note blue circle. They are not embedded in the wall.

  27. aczlan

    aczlan Member

    I am fairly certain that the drains are buried in the gravel outside the wall, so they are not visible from the outside and the pipes are protected.

    Aaron Z
  28. sweepleader

    sweepleader Member

    I should have elaborated perhaps. In the picture noted, "see sushi's post above #126", the blue circle highlights the elbow fitting which was used to connect the under drain system to the wall discharge point. There is a second such fitting shown in the right photo in deirdre's post just above, #147, at the upper edge of the photo, partially cut off. I would like to have been able to mark up both photos to show better what I was talking about but I do not have the needed capability. If this is not clear, perhaps I am missing the point, in which case this post and #146 should be removed. Thanks.
    • Like Like x 1
  29. deirdre

    deirdre Moderator Staff Member

    that's what i was thinking, would make the most sense. i had read early on an article that mentioned the drains were to take care of water fowing along the wall so the wall doesnt erode. but when i went back to look for it i couldnt find it again. ; (
  30. deirdre

    deirdre Moderator Staff Member

    duh. i havent had my coffee yet. thats EXACTLY what i was looking for but didnt notice it. heres the full photo. Thanks. :)
    • Like Like x 1
  31. EricL

    EricL Member

    In addition to the info provided in the first few posts just above this one, I think Mortarboarder is asking about the route that the pipe actually takes on its way to those wall outlets. I was talking with an engineer just last night, and we concluded that the most likely situation is that a series of adjacent under-floor drains connect to a common pipe located just outside the bottom of the wall, with that pipe matching the slope of the wall. Then, at intervals, that pipe connects with a vertical riser, the same structures that are seen in this photo, indicated by the yellow arrows:
    Due to the downhill location of each riser from the series of drains that feed it, the head pressure is sufficient to bring the water to the top of the riser, and from there, the water takes the short path to the next wall outlet located downhill from there (note that every riser is just a short distance uphill from an outlet). What is less certain, is whether each riser is part of an independent system consisting only of a short series of drains located farther up the hill, or if the common pipe is continuous, connected to all under-floor drains and all risers in series. The first of these two options would require that there are overlapping sets of common pipes outside the wall, and though I once thought something like that was involved, I now lean toward the idea that there is single pipe outside the wall, with all drains and all risers connected to that pipe in series.

    I believe that the pipe leading from near the top of the vertical riser to the wall outlet is what can be seen in the photo, gushing water from its open end (provided by Pozzolith in Post #30):
    It might appear at first glance that the elevation of the broken end of the pipe is not correct to be feeding the wall outlet just downstream of there, but considering that the slope of the spillway is about 20 degrees (I seem to remember that someone posted in the spillway thread that the exact figure is 24 degrees), then when using the top of the wall as a reference to identify an imaginary line on the wall which is horizontal, the elevation of the gushing end of the pipe looks about right to have been connected there before being exposed and broken off (the photo's perspective won't allow a person to use a protractor to find a horizontal reference line, so I just "eyeballed" it, with the aid of the most prominent "drippage stain" running down from the top of the wall, which must be a vertical line).

    Using this idea of the design, the common pipe that's connected to the bottom of the riser would be below the depth of erosion shown in that second photo, and thus out of sight in this view.

    I had previously thought (and posted) that pipes outside the wall ran at only a slight downhill slope, thus "getting higher and higher relative to the wall", but such a system would have resulted in a very complex backfilling scenario. The situation utilizing risers which I'm now hypothesizing would work exactly the same, but would be simpler to build and would present no backfilling difficulties.

    Maybe someone who knows how to find and look through the drawings that have been mentioned elsewhere can find a detail that might verify or otherwise clarify this idea.
    Last edited: Feb 27, 2017
  32. Pozzolith

    Pozzolith Member

    This is the dwg that I most want to see.

    On your first point, the wall outlets are located in the top half of the wall, and would necessarily have to connect to the exterior collection drain considerably upstream from the outlet. If the collection pipe runs alongside the wall footing at its base, then there must be a wye connection in that pipe that then drains into the wall outlet at some distance downstream. I believe these cutouts in the footing base (see picture below) is where the underslab VCP drains run beneath the footing and connect to the collection pipe.

    I don't think the collection drain are embedded inside and running lenghwise in the walls. The only wall penetrations, imo, are where the outlets are located. An internal pipe would be an expensive design. Also, none of the fractured wall pictures show any evidence of internal piping.

    To you last point, the underslab drains wouldn't necessarily have to cross the expansion joints. If the joints are spaced at 60' or so, there is plenty of distance to lay those drains in a herringbone pattern, with a slight elevation decline to assure drainage.
  33. Pozzolith

    Pozzolith Member

    022717 KCRA screencap.

    A screengrab from the KCRA live video.

    On the left, drainage from the collection drain. In the center, what appears to be a piece of VCP drain. In some of the live shots, the piece in the center was clearly visible. Also, on the right, the collection drain was sometimes clearly visible.
  34. StickyMockwell

    StickyMockwell New Member

    What struck me watching that video is how long the flow at the bottom of the spillway continued after the intake gates were fully closed, i.e. flow was being sustained entirely by what was coming out of the drains. Obviously the flow was much lower than when the gates were open but was still a decent amount of water.
  35. deirdre

    deirdre Moderator Staff Member

    how long?. i scrolled through it twice and dont see where (timestamp) the flow from the gates actually stopped.

    and re: @Pozzolith i see a drain but it isnt int he location your pic identifies
    pipe.JPG pipe2.JPG

    Moderator Note - deirdre

    Please be aware that this is not a private chat between you all posting on this thread. Metabunk is a public information site, the goal is that ALL readers can gather information quickly and accurately without having to verify every post made.

    Specifics are required. Timestamps, photos etc.
    If long explanations are necessary try to start posts with either bullet points or a brief summary of the point of your post.

    Thank you.

  36. StickyMockwell

    StickyMockwell New Member

    At about 1:13 in the video the camera zooms in to what I think is the bit you've circled and it then breaks off and is carried away downstream (blue circle in image below). The red circle shows what looks like another smaller diameter pipe or bar sticking out - as Deirdre has just circled

    Attached Files:

    • Disagree Disagree x 1
  37. StickyMockwell

    StickyMockwell New Member

    Sorry, only the last gate is open at around 1hr06 (29:49 remaining) shown by the white plume on the right side of the immediate downstream face of the intake gates (there are closer shots of gates just before this):

    By around 1hr09 (26:31 remaining) that white plume has disappeared, from which I have interpreted that gate was shut too:

    Attached Files:

    • Like Like x 1
  38. MortarBoarder

    MortarBoarder Member

    It was pointed out in the main thread that the drain pipe should run only slightly downhill to its end, and not along the bottom of the wall with 90 degree bend to go up to its exhaust port. That might look tidy in drawings, but sediment would clog the pipe at the bend.

    Also, there should not be a continuous drain pipe down the 3,000 feet of spillway because it would easily develop huge velocities and pressures.
    • Agree Agree x 1
  39. Mick West

    Mick West Administrator Staff Member

    • Useful Useful x 3
    • Like Like x 1
  40. Pozzolith

    Pozzolith Member

    • Like Like x 2
Thread Status:
Not open for further replies.