Predicting Spillway Failure

Lab studies undertaken by the Bureau of Reclamation, U.S. Department of the Interior, provide data needed to assist in predicting spillway failure modes related to stagnation pressure. Reclamation is using these results to develop risk analysis methodologies and analytical tools for a full range of spillway failure modes.

By K. Warren Frizell

Scientists at the Hydraulic Laboratory of the U.S. Department of the Interior’s Bureau of Reclamation completed a multi-year study intended to provide design information concerning uplift and structural collapse of spillway slabs with open offset joints or cracks. This data was needed for many aging facilities where cracks and joints open up. The use of risk-based processes for scheduling repairs is only as good as the data that is fed into the process. Designers had been using data from the 1970s and scaling it up, not knowing if the scaling procedures were entirely correct. In addition, there was no data available on the quantity of flow that could be expected for various specific crack and joint geometries.

The lab model developed for this study provided data in the range of most prototype structures, both in velocities and the resulting stagnation pressures (hydraulic jacking) developed for typical joint offset heights and gaps. In addition, the studies provided the first measurements of unit discharges inducted by the stagnation effects into the gaps that could flow beneath a concrete slab.

Understanding spillway failures from stagnation pressure

Spillway failures related to stagnation pressure are not a widely significant problem for Reclamation structures. What Reclamation was most concerned about was the lack of appropriate data to use in the risk analysis process that now drives repair/rehabilitation decisions.

Most Reclamation structures have been designed with joint details that protect the slab foundation, but this cannot be said for all spillways. There are concrete slab structures on earth foundations that do not have appropriate (waterstop) protection. When water penetrates beneath the slab, it can lead to erosion and potentially structural collapse. As structures age, there also are uncertainties regarding the condition of waterstops and their ability to inhibit water from reaching the foundation through an open-offset joint or crack. If it is a crack, there is no protection and so they perhaps are more critical.

These types of spillway failures typically are the result of high-velocity water flowing over and into open, offset joints or cracks along the invert of the spillway channel. The failure can result from two common mechanisms — hydraulic uplift and structural collapse — or a combination of these processes. Stagnation pressure, sometimes referred to as total pressure, is described by the following:

Equation 1
Pstagnation = ½ ρu2 + P
— ρ is fluid density;
— u is velocity; and
— P is static pressure.

Offsets that encroach into the flow from the channel boundary are prime sources to increase local pressures due to deceleration (or stagnation) of the fluid stream. If a water path to the foundation exists through the joint or crack, an elevated pressure or at least a fraction of the magnitude of the stagnation pressure can be transferred to the foundation. If large enough, this pressure can cause a net uplifting force on the slab. In addition, if flowing water can attack the spillway foundation, materials can be carried away in the drains or through other open cracks and joints. Eventually, large voids can form beneath the chute slab that raise the possibility of structural collapse.


In reinforced concrete-lined chutes, the stability of the slab depends on the overall concrete design, including: joint and waterstop details; reinforcement; anchorage; and a functioning, filtered underdrain system. Usually, drainage under the slab is provided to prevent the build up of uplift pressure and subsequent instability due to seepage and natural foundation groundwater conditions. Typically, damage resulting from hydrodynamic uplift on slabs (jacking) begins at the joints, where offsets or spalling has occurred. Spillway flows over these offsets can introduce water into the foundation, which can lead to structural damage as a result of either uplift or erosion of the foundation material.

If this problem persists, there can be complete failure and removal of chute slabs. Structural collapse due to undermining of the chute slabs has been a difficult problem to evaluate due to lack of applicable data or analyses. This problem is generally more of a concern for structures where the chute and underdrain systems may be in poor condition due to aging or improper design. This problem is especially critical for chutes that are founded on soil because joint/crack flow can lead to erosion and undermining of the chute foundation and structural collapse of a chute slab.

With the advent of the risk assessment process as an approach to assist dam owners and dam safety decision-makers on repairs or replacement of hydraulic structures, Reclamation designers identified a critical need for additional data that would reduce uncertainty in predicting possible uplift or structural collapse failure modes. These data were collected as part of a multi-year Dam Safety Research program at Reclamation’s hydraulic laboratory and are presented in Reclamation Report DSO-07-07.1 It extends the available data for predicting uplift and provides estimates of crack and joint flows that previously were unavailable.

Summary of lab studies

The basis of most of Reclamation’s analyses for stagnation pressure failure date to laboratory studies by Johnson in the 1970s in which he tested a model depicting a two-dimensional open channel flow on a steep canal wasteway with a range of offsets and gap dimensions.2 A wasteway is a steep chute that may come off a canal and is designed to prevent overtopping of the canal. Johnson did not model flow through the joint but did measure uplift pressures resulting from a variety of offset dimensions (both vertical and horizontal) for flow velocities up to 15 feet per second. One of the major uncertainties in applying this lab data to prototype spillway analyses was the method that should be used to extrapolate the data up to more typical spillway flow velocities (50 to 90 feet per second).

To produce high velocities in the laboratory, a reduced sectional model was needed (see Figure 1). With this water tunnel-type representation, offsets and velocities of prototype scale could be studied. Offsets (both horizontally and vertically) could be varied up to 0.75 inch, and velocities up to 60 feet per second were possible.

FIGURE 1 Lab Model Test Section
A reduced sectional model of a spillway was built in a lab by Reclamation to allow testing of high flow velocities. With this water tunnel -type representation, offsets and velocities of prototype scale could be studied. Offsets (both horizontally and vertically) could be varied up to 0.75 inch, and velocities up to 60 feet per second were possible.

Flow entering the test section was measured, and a mean velocity was calculated from known dimensions. The pressure differential across the “slab” downstream from the joint/crack was measured using an electronic pressure transducer. The cavity beneath the slab could be sealed or opened to allow flow through the gap. This flow initially was measured volumetrically and then correlated with the pressure differential measurements using physically-based theory. Using the sealed cavity, the maximum possible uplift pressure for a given joint/crack geometry could be measured. Then, when the cavity was opened, an estimate could be made of the amount of unit discharge through the gap.

A complete presentation of the laboratory data is available.1 Figures 2 and 3 present a sample of the type of data available. This data is for a 0.125-inch horizontal gap with sharp-edged joint details. It shows the mean uplift pressure due to stagnation for a sealed lower cavity and the unit discharge for similar conditions with the lower cavity vented to the drain system. These figures show that substantial differential pressures across a slab can be produced due to this stagnation effect, even with what would seem to be very small gaps and offsets. The flow that is “pumped” into the foundation area by these conditions also can be quite large, and actually smaller gaps or cracks can have the higher unit discharges transferred into the openings.

Application of the research

The use of risk assessment techniques is growing as a method to facilitate dam safety decisions for prioritizing rehabilitation or replacement of structures or critical features. Thus, improved methods, new data, and reduced uncertainty become increasingly important to the overall process. These present studies have increased the data quantity and quality for conditions more relative to typical spillway flows. The data presented in DSO-07-07 can be used to evaluate existing or new spillways or other high-velocity conveyance channels. The application of the results can provide information for two of the basic issues regarding structure performance: slab stability and underdrain performance.

Typically, the first step would be to predict water surface profiles for the range of flows that are important. These profiles will provide depths and velocities at various positions on the chute for various discharges. With this information, slab stability can be estimated using weight of the water and slab as typical downward loads and then by either assuming some gap and offset dimensions or using information from a recent inspection to estimate the possible uplift due to stagnation effects. These loads have to be resisted by bending and shear strength of the slab design, i.e., the design reinforced concrete properties.

FIGURE 2 Mean Uplift Pressure
Mean uplift pressure due to stagnation for a sealed lower cavity with a 0.125-inch gap shows measured differential pressures across the slab for a variety of offset heights into the flow.

Depending on the worst-case scenario, which would include no waterstops at the joints or at least damaged and degraded waterstops, a net force can be estimated that will be used to predict stability of the slabs along the length of the chute (or at a particular location identified by inspection) for the variety of discharge conditions.

Depending on the assumptions made in this process, the evaluation could have some probabilities estimated as well that could depend on waterstop condition or other features that are not readily known. Estimates of inflow to the subsurface drains or areas beneath the slab can be estimated from the information contained within DSO-07-07.

The most interesting finding of the research is that unit inflows can be substantial and the larger the gap in the flow direction does not necessarily mean increased flow volume. If unfiltered drains are present or the structure is on a soil foundation, the inflow through open joints or cracks can be quite damaging, causing foundation erosion. Materials could be carried out through the drainage system or possibly through other open joints or cracks further downstream. Depending on the slab design, most specifically the level of reinforcement, structural collapse may be possible if large voids form beneath the slab.

FIGURE 3 Unit Discharge
Unit discharge for a 0.125-inch joint/crack, sharp -edged geometry, with the lower cavity vented to the drain system shows flow into the simulated foundation area without waterstops. The flow rate is dependent on the actual head loss of the joint or crack gap, plus the drainage system.

Prediction of damage or failure due to a structural collapse is a bit more complex and not easily addressed through known parameters. Thus, if the conditions exist for foundation erosion to occur, it may be best to be conservative about the end result.

In summary, if application of the data and conditions present would indicate a possible instability for a spillway slab due to stagnation-induced uplift, the decision-maker must evaluate the possibility of the flow conditions occurring that might cause the problem. The decision-maker then must factor that probability into the risk assessment process to make the decision that might drive replacement or modification of the structure.

This data has been used at Hyrum Dam, on the Little Bear River in Utah, to verify prediction of a partial failure that occurred. The data will be used in a subsequent risk analysis for this dam to determine long-term repair strategies. The data also is used to evaluate joints and cracks during comprehensive facilities reviews, especially in cases where problems are predicted as a result of stagnation pressure effects.

Undermining of this spillway chute slab occurred due to flow entering at joints and cracks and transporting foundation material through unfiltered drains.

Future work

Work is ongoing in several areas. Reclamation recently published updated best practices that include evaluation of all types of failure modes for spillways (including stagnation pressure failures).3

In addition, a joint team is working on a spillway failure modes toolbox to be used for risk analyses. The team consists of representatives from the U.S. Army Corps of Engineers, Reclamation, and URS Corp., as well as consultants from the U.S. and Australia. The team is developing risk analysis methodologies and analytical tools for a full range of hydrologic failure modes related to spillways, including ball milling, overtopping of spillway walls, cavitation, hydraulic jacking, foundation erosion from stagnation pressure, and plunge pool scour. All of the failure modes include a common element of foundation exposure and erosion and headcutting within the foundation.

The team is working toward developing an initial draft of the toolbox document by April 2009.

  1. Frizell, K. Warren, “Uplift and Crack Flow Resulting from High Velocity Discharges Over Open Offset Joints – Laboratory Studies,” Bureau of Reclamation, Dam Safety Technology Development Program, Report DSO-07-07, Denver, Colo., 2007.
  2. Johnson, Perry L., “Research into Uplift on Steep Chute Lateral Linings,” Memorandum to the Open and Closed Conduit Systems Committee, Bureau of Reclamation, Denver, Colo., 1976.
  3. Dam Safety Risk Analysis Best Practices Training Manual, Version 1.3, Bureau of Reclamation, Technical Service Center, Denver, Colo., 2010.

Warren Frizell, research hydraulic engineer with the U.S. Department of the Interior’s Bureau of Reclamation, was the principal investigator for the laboratory studies on predicting spillway failure modes.


This article has been evaluated and edited in accordance with reviews conducted by two or more professionals who have relevant expertise. These peer reviewers judge manuscripts for technical accuracy, usefulness, and overall importance within the hydroelectric industry.

More Hydro Review Current Issue Articles
More Hydro Review Archives Issue Articles


Previous articleA-Power signs US279 million contract for design, construction of hydro plants
Next articleScottish & Southern Energy to restart Glendoe hydropower plant in 2012

No posts to display