Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
162 4.1 Applicability of Results to Highway Practice The Interstate Highway (I-90) Bridge over Schoharie Creek in New York failed during a flood in 1987 and led FHWA to issue a mandate for all highway bridges over water to be evaluated for scour-critical conditions. As state departments of transportation complied with the mandate, a number of bridges with shallow foundations were identified on rock-bed channels. Available pro- cedures for evaluating scour of sand-bed channels produced scour-depth estimates that seemed to be unrealistic or unbelievable for rock materials. The need for improved methods for evaluating scour at bridge foundations on rock became widely recognized because of these experiences. The results of the research described in this report are applicable to design of proposed bridges founded on rock-bed channels, as well as to evaluation of existing bridges on rock-bed chan- nels and design of foundation rehabilitation or scour countermeasures. An immediate benefit of these research results is likely to be the removal from the scour-critical list of bridges with shallow foundations on rock. The guidelines and methods that resulted from this research provide tools for bridge owners and their technical staff members to use for evaluating modes of rock scour that may be relevant to par- ticular bridges. Procedures are provided for estimating time-rate of scour and design scour depths for progressive and cumulative scour of degradable rock materials. Guidance also is provided for threshold-controlled scour processes of cavitation and quarrying and plucking of durable rock blocks. The recommended procedures are relatively simple and quantitative and use equipment and methods familiar to transportation agency personnel and other bridge owners. 4.2 Conclusions and Recommendations Scour at bridge foundations traditionally is evaluated by hydraulic engineers with input from engineering geologists and geotechnical engineers. NCHRP Project 24-29 focuses on recognition of rock and rock-like materials that may be susceptible to scour processes and characterization of bridge foundation conditions in terms that accurately reflect the scour susceptibility and can be used by hydraulic engineers to calculate design scour depths. In essence, the research strives for geotechnical site characterization expressed in scour-relevant terms for use by hydraulic engineers. The following four modes of rock scour were defined in this research: 1. Dissolution of soluble rocks, 2. Cavitation, 3. Plucking of durable jointed rock blocks, and 4. Gradual wear of degradable rock material. C h a p t e r 4 Conclusions, Recommendations, and Suggested Research
Conclusions, recommendations, and Suggested research 163 The time between flood events can contribute to reduction in scour resistance through weath- ering or slaking of rock materials, or enhanced circulation of water in joints held open by gravel fragments wedged into the joint planes during turbulent flow causing blocks of durable rock to vibrate or jostle. One of the most important conclusions of this research is that the scour resistance of degrad- able rock materials is not solely a function of the rock propertiesâit is a rock-water interaction phenomenon. The hardest of rock materials will wear away in response to sustained powerful discharges, whereas the softest of rock materials may resist erosion for a long period of time in response to tranquil water flow. Waterjet cutters are used to strip concrete away from reinforcing steel for bridge-deck rehabilitation, and the waterjets will cut the steel if the force is applied to the steel for sufficient periods of time. For example, the claystone from the Montezuma Creek Bridge on State Route 262 in southeastern Utah was the most scour-susceptible of the rocks tested with the Modified Slake Durability Test procedure. The hydraulic loading in the Chipola River at the twin span I-10 bridges in northwestern Florida was so low (500 ft-lb/s/ft2) over the 32-year period between 1976 and 2008 that the highly scour-susceptible claystone would be expected to scour less than 1 foot over that period and might be defined as non-susceptible to scour for Chipola River flow conditions. Soluble rock material that dissolves in engineering time is not used for foundation sup- port of bridges. Therefore, the scour-related issue regarding dissolution of soluble rocks is the complex-scour-response of the heterogeneous earth materials that may fill solution cavities. Typical cavity-filling earth material consists of blocks of relatively hard rock (limestone, dolostone, or marble) in a soil matrix that commonly is clay. The clay will wear away progres- sively, whereas the blocks of rock will be plucked as the threshold conditions are attained. In some cases, loose blocks of rock may accumulate in the filled cavity until they form a self- armoring layer. Cavitation has produced spectacular scour holes in rock in spillway tunnels. However, natural open channels where bridges are likely to be located typically do not have the water depth and velocity conditions that are needed to support cavitation. Thus, a simple check of expected maxi- mum flow depth and velocity may be used to determine if cavitation is likely or even possible. In natural channels where bridges are likely to be located, hydraulic conditions for cavitation generally do not appear likely to occur. Plucking of durable rock blocks is governed by the size and shape of the rock blocks, the nature of the discontinuity surfaces, the hydraulic loading at peak discharge, and turbulence intensity fluctuations created by flow around bridge piers and across irregularities at block edges across the channel. The Comprehensive Scour Model applied to open channels provides some guidance on the velocity at the onset of block plucking; this model appears to be promising, but model calibra- tion could not be done as part of this research project. Calibration is needed and should be done with actual natural channel sites with blocky rock-bed channels; alternatively, flume tests could be conducted to validate and refine the model for application at bridges on natural channels. The Headcut Erodibility Index and the Erodibility Index Methods were evaluated as part of this research. These two threshold-controlled index methods are similar and both were devel- oped for unlined spillway channels with significantly more peak hydraulic energy than exists in normal natural channels where bridges are likely to be located. The results of these two index methods applied to channels at bridges evaluated in this research show that the peak hydraulic energy generally is insufficient for erosion of the local rock masses. Bridge owners might choose to define rock masses with low RQD values (e.g., less than 75 percent) as being susceptible to scour regardless of the stream flow conditions. A discontinu- ous rock mass comprised of cubes of durable rock 4 inches on each side could have an RQD
164 Scour at Bridge Foundations on rock value of 100 percent if the core is oriented parallel or perpendicular to the discontinuity planes. The same rock mass would have an RQD value of zero percent if the core is oriented 45 degrees to two of the discontinuity planes. Consequently, a high RQD value by itself is not suitable to demonstrate that a rock mass is likely to be resistant to quarrying and plucking without under- standing the peak discharge characteristics of flowing turbulent water in the channel. Similarly, a low RQD value by itself would not necessarily demonstrate that a rock mass is susceptible to quarrying and plucking under all stream flow conditions. Progressive scour of degradable rock material was documented at three of the bridge sites discussed in this report, and a fourth bridge was documented to have experienced no scour over a period of several decades. The progressive nature of scour in susceptible rock materials sug- gested that cumulative hydraulic loading needed to be considered; stream power is a hydraulic parameter that captures flow velocity, flow depth, and slope, and logically can be accumulated over time. Stream power is calculated from commonly available daily flow series, so the accumu- lated hydraulic parameter is cumulative daily stream power. This cumulative parameter could be converted to unit energy (i.e., kW-hr/m2), but the calculations (shear stress à velocity) can be expressed conveniently in terms of unit energy dissipation (i.e., ft-lb/s/ft2), which appears to have meaningful units. A probability-weighted approach was used to convert conventional flood-frequency events into event-based scour depths by using stream power and channel response based on observed scour from repeated cross sections or approximated from specialized geotechnical laboratory test results. Durations of flows associated with flood frequencies must be included in the analy- ses. The annualized scour depths associated with the spectrum of flood-frequency events can be combined to produce the time-rate of scour, which is one of the objectives of this research. The service life of the bridge in years times the average annual scour depth produces the design scour depth, which is another research objective. The results of this research can be applied with the greatest confidence to the scour mode of progressive wear of degradable rock material at sites of existing bridges with repeated cross sec- tions. For such bridges, past scour depths are documented and can be compared to cumulative daily stream power to produce an empirical scour number (i.e., scour depth per unit of stream power). Without repeated cross sections to document past scour, the analysis must rely on the results of the Modified Slake Durability Test. These test results appear to provide a promising opportunity to quantify rock-bed channel response based on an index property of the rock material expressed in stream power units. Only one of the bridge sites studied gives an opportunity for calibration of the geotechnical test results with repeated cross section data (State Route 273 at the Sacramento River in Redding, California). A second bridge site (Interstate 10 at the Chipola River in Jackson County, Florida) provides a limiting condition of no measurable scour with calculated cumulative daily stream power that can be reconciled with the geotechnical test data to explain why no measurable scour occurred. Analysis of the I-10 bridge site at the Chipola River in Florida provides a useful exam- ple to emphasize the importance of the rock-water interaction phenomenon. The cumulative stream power delivered by the Chipola River at the I-10 bridges was 500 ft-lb/s/ft2 over the 32-year period between 1976 and 2008 (Section 3.8.7.3). The most scour-susceptible rock mate- rial measured during this research was the claystone in Montezuma Creek at the State Route 262 Bridge, with a geotechnical scour number of 0.001 ft/unit of stream power (linear regression slope in Table 3.4). This scour number indicates that the Montezuma Creek claystone would be expected to scour 0.5 ft (= 0.001 ft/unit of stream power à 500 ft-lb/s/ft2) over the 32-year period during which the Chipola River limestone did not scour a measureable amount. Additional bridge sites are needed for calibration and validation of the procedure, but it appears to be useful and consistent with observed scour response.
Conclusions, recommendations, and Suggested research 165 Repeated cross sections are the best way to document scour. The cross sections used in this research posed some challenges for interpretation. The cross section data was not well docu- mented and locations along the bridge where measurements were made were inconsistent. The value of repeated cross sections would be improved if a larger number of measurements were taken at consistent locations. It would be helpful for scour at piers to be differentiated from scour between piers. Five bridges were visited as part of this research. The geologic conditions at each bridge site are listed in Table 3.6. It can be seen in Table 3.6 that one bridge was founded on Quaternary-age ice- contact stratified glacial drift (till) that has rock-like qualities (I-90 over Schoharie Creek in New York). Two bridges were founded on Tertiary-age marine sedimentary rock formations (I-10 over Chipola River in Florida was founded on dolomitic limestone, whereas State Route 22 over Mill Creek in Oregon was founded on blocky siltstone). Two bridges were founded on Cretaceous- or Jurassic-age (Mesozoic) marine or fluvial sedimentary rock formations (State Route 273 over Sacramento River in California was founded on thinly bedded siltstone, whereas State Route 262 over Montezuma Creek in Utah was founded on sandstone underlain by claystone). The geology at all five bridges consisted of sedimentary formations. Igneous and metamor- phic rock types probably will respond to the modified slake durability test in ways consistent with the response of the sedimentary formations described in this report. Coarse-grained granitic rock that is weathered or altered would be expected to wear more rapidly than simi- lar fine-grained granitic rock that is fresh or unaltered. Metamorphic rock can range from sedimentary-like low-grade slate to medium-grade schist and high-grade gneiss, depending on the mineralogy. Some metamorphic rocks can be quite durable (e.g., quartzite) whereas other metamorphic rocks can be quite degradable (e.g., mica schist). The degree of fracturing and weathering or alteration, of course, may control the response of the rock to flowing water. For these reasons, the geology of each bridge site must be evaluated specifically for its likely response to hydraulic forces. 4.3 Suggested Research A number of research activities appear to be potentially useful in calibrating or refining methods for evaluating rock scour. These suggested research activities include field investiga- tions, laboratory testing, and engineering analyses. Promising field investigations could be undertaken to locate and characterize sites where durable rock blocks have been plucked by flood flows, as well as evaluation of additional well-documented sites where degradable rock materials have undergone progressive scour. The durable block-plucking process on natural channels apparently has no engineering calibration at this time. Geomorphology literature regarding mountain-scale processes has documented block plucking and cavitation in some steep-gradient, high-energy natural channels in Alaska and Pakistan, but these channels are not locations where bridge foundations would be placed. The limiting block sizes and shapes susceptible to quarrying and plucking need to be refined for the range of typical flow velocities and depths at bridge sites. Further assessment of a limiting value of RQD may be productive for durable rock quarrying and plucking as a simple approach to defining susceptible rock masses. For example, durable rock masses with RQD values less than 75 percent might be defined as being susceptible to scour; however, without appropriate procedures for determining the scour depth, the HEC-18 meth- odology for sand-bed channels likely would be applied, possibly with predictably conservative results. Perhaps the minimum dimension for calculating RQD (100 mm) could be used with the minimum RQD value (e.g., 75 percent) to prescribe an equivalent D50 grain size (i.e., 75 mm) for use with the HEC-18 methodology.
166 Scour at Bridge Foundations on rock Progressive scour of degradable rock materials appears to be the most important rock-scour process on natural channels where bridges might be located. The results from the Sacramento River in Redding, California, are encouraging and indicate that the probability-weighted approach to defining the average annual scour is valuable. Additional similar sites are needed for calibrating empirical and geotechnical scour numbers, as well as validating and refining the approach. Flume tests could provide some valuable information regarding plucking durable rock blocks and progressive wear of degradable rock materials. The Comprehensive Scour Model used sim- ple friction coefficients for inclined joints defining rock blocks, but considered vertical joints to be essentially frictionless because no normal stresses were applied across the joint surfaces. Flume tests could provide useful information regarding this aspect of block plucking at scaled flow velocities and depths that are expected on natural channels where bridges might be located. The Modified Slake Durability Test results are consistent among the rock types that were tested in that the least durable claystone samples wore away most rapidly and the most durable sandstone samples wore very little during the entire test, with intermediate durability samples responding within the range according to the sample durability. Testing is needed to document the scour rate of blocks of different rock types when subjected to water flow in a test flume. These flume-test results should be correlated with the Modified Slake Durability Test results so that the less expensive Modified Slake Durability Test results can be used with higher confidence. Physical properties of weak rock materials and strong soil deposits can be similar to overlap- ping. The results from the Modified Slake Durability Test seem to be useful for quantifying equivalent average scour rates in terms of equivalent cumulative daily stream power. Samples of clay soil that have sufficient toughness at field moisture content to be cut into cubes or lumps, each weighing approximately 50 grams and placed in a slake durability test drum, should be suitable for testing provided that oven drying is not used on the samples at any time. This prom- ising area of research might lead to an understanding of the index property associated with the Modified Slake Durability Test. The importance of air slaking deserves additional consideration. Hydraulic engineers are accustomed to performing flood frequency analyses, but scour in channels with slaking rock materials may be dominated by frequency and severity of droughts. Procedures for assessing drought frequency should be developed so that a process model can be formulated for incorpo- rating the rate of slaked-material production into channel-scour assessments. It would be desirable for a database of rock scour evaluation results to be established that may lead to generalizations about scour resistance of rocks based on rock type, basic physical char- acteristics, and hydraulic loading. Rock mass characterization approaches, such as rock mass rating (RMR) or the Norwegian rock mass quality (Q) systems (Bieniawski, 1989), may capture sufficient detail about the rock mass to provide a reliable basis for a scour-response index. A sub- stantial number of rock scour assessments for a variety of igneous, metamorphic, and sedimen- tary rock types would be needed to permit a reliable rock-scour-response index to be developed.
