Bridge Scour and Stream Instability Countermeasures: Experience, Selection, and Design Guidance-Third Edition

Chapter 5

RIPRAP DESIGN, FILTERS, FAILURE MODES, AND ALTERNATIVES

5.1 OVERVIEW

Riprap consists of a layer or facing of rock, dumped or hand-placed on channel and structure boundaries to limit the effects of erosion. It is the most common type of countermeasure due to its general availability, ease of installation and relatively low cost. Any successful riprap design must account for several possible modes of failure. These include riprap particle erosion, substrate material erosion and mass failure. Riprap particle erosion is minimized by sizing the riprap to withstand hydraulic and turbulence forces, but is also affected by riprap slope, impact and abrasion, ice, waves and vandalism. Substrate particle erosion occurs when the base material erodes and migrates through the riprap voids causing the riprap to settle. Substrate particle erosion is limited by placing a granular or geotextile filter between the riprap and the base material. Mass failure occurs when large sections of the riprap and/or base material slide or slump due to gravity forces. Mass failure can be caused by excess pore water pressures, bank steepness and loss of basal support through scour or channel migration. Also, a filter fabric that is too fine can clog and cause the buildup of pore water pressures in the underlying soil.

Riprap that is large enough to resist all the hydraulic forces can fail if channel migration or scour undermines the toe support. When the riprap toe is undermined it can shift and remain functional to some degree. Often an extra volume of riprap is included at the toe for this purpose, or the riprap toe is trenched to the depth of potential degradation and contraction scour.

Graded riprap is more stable than uniform riprap because the range of sizes helps the riprap layer to interlock. Care must be taken during construction to ensure that the graded rocks are uniformly distributed. If large rocks roll to the base of the bank and the smaller rocks accumulate at the top, the benefits of using graded riprap will be lost. Also, a relatively uniform riprap surface will be more stable than an extremely uneven riprap surface.

Riprap design requires hydraulic, scour, and stream instability analyses as well as geotechnical investigations of channel and bank stability. Pier riprap can fail if contraction scour or channel bed degradation causes the stones to launch and roll away from the pier, or on rivers with mobile bed forms, by bedform undermining. Abutment riprap can fail if channel migration undermines the toe support of the rock. Channel bank riprap can fail if excess pore pressures or toe scour produce a mechanically unstable bank. These failures could occur even if the riprap size was appropriate for the particular application.

In summary, design of a riprap erosion control system requires knowledge of: river bed, bank, and foundation material; flow conditions including velocity, depth and orientation; riprap characteristics of size, density, durability, and availability; location, orientation and dimensions of piers, abutments, guide banks, and spurs; and the type of interface material between riprap and underlying foundation which may be geotextile fabric or a filter of sand and/or gravel. Adequate "toe down" and termination details are essential to the performance of the riprap system. Thus, riprap should be considered an integrated system where successful performance of a riprap installation depends on the response of each component of the system to hydraulic and environmental stresses throughout its service life.

5.2 RIPRAP DESIGN

5.2.1 Introduction

Most of the guidelines and recommendations of this chapter are derived from NCHRP Report 568, "Riprap Design Criteria, Recommended Specifications, and Quality Control," the final report for NCHRP Project 24-23 (Lagasse et al. 2006). The basic objectives of this study were to develop design guidelines, material specifications and test methods, construction specifications, and construction, inspection and quality control guidelines for riprap for a range of applications, including: revetment on streams and riverbanks, bridge piers and abutments, and bridge scour countermeasures such as guide banks and spurs. NCHRP Project 24-23 was a synthesis study and did not involve any original laboratory experimental work, but some analytical work (specifically 1- and 2-dimensional computer modeling) was necessary to address issues related to input hydraulic variables for design.

Additional guidance for riprap at bridge piers is based on results of the NCHRP Project 24-07(2) provided in NCHRP Report 593, "Countermeasures to Protect Bridge Piers from Scour" (Lagasse et al. 2007). This study involved extensive laboratory testing at Colorado State University of a range of bridge pier scour countermeasures, including: riprap, partially grouted riprap, articulating concrete blocks, gabion mattresses, and grout-filled mattresses. NCHRP Report 593 includes detailed, stand-alone guidelines for the design of these five pier scour countermeasure alternatives.

Sizing the stone is only the first step in the comprehensive design, production, installation, inspection, and maintenance process required for a successful riprap armoring system. Filter requirements, material and testing specifications, construction and installation guidelines, and inspection and quality control procedures are also necessary. This section recommends riprap design approaches for a range of riprap applications. Riprap design (sizing) is covered in more detail in application-specific design guidelines in Volume 2.

Subsequent sections provide an overview of filter design requirements and recommendations for specification, testing, and quality control for revetment riprap installations. In general, these recommendations are applicable to riprap for other applications such as at bridge piers and abutments, and for countermeasures such as spurs and guide banks.

Generalized construction/installation guidance is also summarized in this chapter. Failure modes for revetment and bridge pier riprap are described to underscore the integrated nature of riprap armoring systems and as a basis for developing inspection and maintenance guidance. Finally, an overview of concrete armor units (artificial riprap) that could be used in lieu of rock for selected applications is provided.

5.2.2 Riprap Revetment

Based on a screening of the many riprap revetment design equations found in the literature, seven equations were evaluated with sensitivity analyses using both field and laboratory data during NCHRP Project 24-23 (Lagasse et al. 2006). One, the U.S. Army Corps of Engineers EM 1601, was recommended for streambank revetment design (USACE 1991). Factors considered were the ability of the basic equation to discriminate between stable and failed riprap using field and laboratory data, bank and bend correction factors, and the reasonableness of safety/stability factors. Detailed design guidance using the EM 1601 equation is provided in Volume 2, Design Guideline 4. A standard riprap gradation specification which considers design, production, and installation requirements is recommended in Design Guideline 4, together with a standardized riprap size classification system. Installation guidance for toe down and transitions is also provided for the revetment application. General riprap specifications, testing, and quality control guidance can be found in Design Guideline 4.

5.2.3 Riprap for Bridge Piers

According to Hoffmans and Verheij (1997) riprap can be sized using either the Isbash or Shields stability criteria if turbulence intensity is incorporated into the velocity component. The effect of turbulence is to increase instantaneous velocities well above the levels for unobstructed flow. This concept is particularly applicable to the pier riprap equations.

The standard Isbash (1936) formula for sizing riprap on a channel bed is:

(5.1)

where:

To incorporate the effects of turbulence intensity, Hoffmans and Verheij (1997) recommend that the value of K be adjusted above a value of 1.0. In the specific case of circular piers, they recommend using the local velocity upstream of the pier and values of K up to 2.0. This amount of adjustment is equivalent to increasing shear stress by a factor of four.

This approach is similar to the equations presented in Design Guideline 8 and in the riprap sizing formula presented by Parola (1993). The only difference is the recommended values of K in the design guideline are 1.5 for circular piers and 1.7 for square piers. The recommended values of K by Parola ranged from 1.44 to 1.90 depending on pier and footing geometry and approach flow angle of attack.

After a preliminary screening during NCHRP Project 24-23, the HEC-23 (Second Edition) equation, which was derived from work by Parola and Jones, was compared to several other equations using three laboratory data sets. Based on this sensitivity analysis, it was concluded that the HEC-23 and Parola equations provide the best balance between the objective of rarely (if ever) undersizing bridge pier riprap and the desire to not be overly conservative. As these equations are very similar, the HEC-23 (Second Edition) equation was recommended for design practice.

The laboratory results and design recommendations from a concurrent study of countermeasures to protect bridge piers from scour (NCHRP 24-07(2)) were evaluated regarding filter requirements, riprap extent, and other construction/ installation guidelines for pier riprap (Lagasse et al. 2007). Specifically, guidelines for the use of sand-filled geotextile containers as a means of placing a filter for pier riprap derived from European practice were investigated. Construction and installation guidelines and constructability issues were also addressed. The findings and recommendations from NCHRP Projects 24-23 and 24-07(2) are combined in Volume 2, Design Guideline 11 to provide comprehensive design guidance for bridge pier riprap.

5.2.4 Riprap for Bridge Abutments

For NCHRP 24-23, only the abutment riprap sizing approach as developed by FHWA (Pagán-Ortiz 1991, Atayee 1993) and presented in HEC-23 (Second Edition) was considered to be a candidate for further investigation. The approach consists of two equations, one for Froude numbers less than 0.8 and the other for higher Froude numbers. There are no field data available to test these equations and the only available laboratory data set was used to develop the equations. The FHWA equations rely on an estimated velocity, known as the characteristic average velocity, at the abutment toe. Rather than evaluating these equations using the same laboratory data set used to develop them, the method for estimating the velocity at the abutment was investigated in detail. Two-dimensional modeling was performed to evaluate the flow field around an abutment and to verify or improve the Set Back Ratio (SBR) method for estimating velocity for the design equations. Results of the modeling indicated that if the estimated velocity does not exceed the maximum velocity in the channel, the SBR method is well suited for determining velocity at an abutment as a basis for riprap design.

The findings and recommendations from NCHRP Project 24-23 (Lagasse et al. 2006) and NCHRP Project 24-18 (Barkdoll et al. 2007) are presented in Volume 2, Design Guideline 14 for the sizing, filter, and layout of abutment riprap installations. Material and testing specifications, construction and installation guidelines, and inspection and quality control for revetment riprap are suitable for abutment riprap (see Section 5.5 and Design Guideline 4).

5.2.5 Riprap Protection for Countermeasures

In general, design guidelines and specifications for riprap to protect countermeasures are similar to those for bankline revetment or abutments. Consequently, recommendations for revetment riprap can be adapted to the countermeasure application. Guidance for sizing and placing riprap at zones of high stress on countermeasures (e.g., the nose of a guide bank or spur) was developed during NCHRP Project 24-23 (Lagasse et al. 2006). The feasibility of using an abutment-related characteristic average velocity for countermeasure riprap sizing was also evaluated, and a recommendation on an adjustment to the characteristic average velocity approach for guide bank riprap design was developed. Guidance from the U.S. Army Corps of Engineers (EM 1601) can be used for sizing riprap for spurs (USACE 1991). The findings and recommendations from NCHRP Project 24-23 are the basis for design guidance for sizing riprap for spurs in Design Guideline 2 and for guide banks in Design Guideline 15.

NCHRP 24-23 also investigated methods for sizing riprap under overtopping conditions on roadway embankments and the embankment portion of countermeasures. The recommended methodology, based on laboratory testing at Colorado State University, is presented in Design Guideline 5.

5.2.6 Riprap for Special Applications

Environments subject to wave attack frequently require some type of protection to ensure the stability of highway and/or bridge infrastructure. Design Guideline 17 provides information on wave characteristics and procedures for designing rock riprap as protection against wave attack.

Bottomless (or three-sided) culverts are structures that have natural channel materials as the bottom. These structures may be rectangular in shape or may have a more rounded top. They are typically founded on spread footings and can be highly susceptible to scour. Recent laboratory studies by FHWA (Kerenyi 2003, 2007) show that scour is greatest at the upstream corners of the culvert entrance. Based on these studies and other guidance (MDSHA 2005), Design Guideline 18 presents riprap sizing, filter, and layout details to protect against scour at bottomless culverts.

5.2.7 Termination Details

Undermining of the edges of armoring countermeasures like riprap is one of the primary mechanisms of failure (see Section 5.4). The edges of the armoring material (head, toe, and flanks) should be designed so that undermining will not occur. For channel bed armoring, this is accomplished by keying the edges into the subgrade to a depth which extends below the combined expected contraction scour and long-term degradation depth. For side slope protection, this is achieved by trenching the toe of the revetment below the channel bed to a depth which extends below the combined expected contraction scour and long-term degradation depth. When excavation to the contraction scour and degradation depth is impractical, a launching apron can be used to provide enough volume of rock to launch into the channel while maintaining sufficient protection of the exposed portion of the bank. Additional guidelines on edge treatment for riprap countermeasures can be found in Design Guidelines 4, 11, and 14.

5.2.8 Riprap Size, Shape, and Gradation

Riprap design methods typically yield a required size of stone that will result in stable performance under the design loadings. Because stone is produced and delivered in a range of sizes and shapes, the required size of stone is often stated in terms of a minimum allowable representative size. For example, the designer may specify a minimum d₅₀ or d₃₀ for the rock comprising the riprap, thus indicating the size for which 50% or 30% (by weight) of the particles are smaller. Stone sizes can also be specified in terms of weight (e.g., W₅₀ or W₃₀) using an accepted relationship between size and volume, and the known (or assumed) density of the particle.

Shape: The shape of a stone can be generally described by designating three axes of measurement: Major, intermediate, and minor, also known as the "A, B, and C" axes, as shown in Figure 5.1.

Figure 5.1. Riprap shape described by three axes.

Riprap stones should not be thin and platy, nor should they be long and needle-like. Therefore, specifying a maximum allowable value for the ratio A/C, also known as the shape factor, provides a suitable measure of particle shape, since the B axis is intermediate between the two extremes of length A and thickness C. A maximum allowable value of 3.0 is recommended:

(5.2)

For riprap applications, stones tending toward subangular to angular are preferred, due to the higher degree of interlocking, hence greater stability, compared to rounded particles of the same weight.

Density: A measure of density of natural rock is the specific gravity S_g, which is the ratio of the density of a single (solid) rock particle γ _s to the density of water γ_w:

(5.3)

Typically, a minimum allowable specific gravity of 2.5 is required for riprap applications. Where quarry sources uniformly produce rock with a specific gravity significantly greater than 2.5 (such as dolomite, S_g = 2.7 to 2.8), the equivalent stone size can be substantially reduced and still achieve the same particle weight gradation.

Size and weight: Based on field studies, the recommended relationship between size and weight is given by:

(5.4)

where:

Table 5.1 provides recommended gradations for ten standard classes of riprap based on the median particle diameter d₅₀ as determined by the dimension of the intermediate ("B") axis. These gradations were developed under NCHRP Project 24-23, "Riprap Design Criteria, Recommended Specifications, and Quality Control" (Lagasse et al. 2006). The proposed gradation criteria are based on a nominal or "target" d₅₀ and a uniformity ratio d₈₅/d₁₅ that results in riprap that is well graded. The target uniformity ratio d₈₅/d₁₅ is 2.0 and the allowable range is from 1.5 to 2.5.

To specify riprap using the standard classes shown in Table 5.1, the "next larger size" approach should be adopted. For example, if a riprap sizing calculation results in a required d₅₀ of 16.8 inches, Class V riprap should be specified because it has a nominal d₅₀ of 18 inches.

Based on Equation 5.4, which assumes the volume of the stone is 85% of a cube, Table 5.2 provides the equivalent particle weights for the same ten classes, using a specific gravity of 2.65 for the particle density.

5.3 FILTER REQUIREMENTS

5.3.1 Overview

The importance of the filter component of a countermeasure for stream instability or bridge scour installation should not be underestimated. Filters are essential to the successful long-term performance of countermeasures, especially armoring countermeasures. There are two basic types of filters: granular filters and geotextile filters. Some situations call for a composite filter consisting of both a granular layer and a geotextile. The specific characteristics of the base soil determine the design considerations for the filter layer. In general, where dune-type bedforms may be present during flood events, it is strongly recommended that only a geotextile filter be considered for use with countermeasures.

The filter must retain the coarser particles of the subgrade while remaining permeable enough to allow infiltration and exfiltration to occur freely. It is not necessary to retain all the particle sizes in the subgrade; in fact, it is beneficial to allow the smaller particles to pass through the filter, leaving a coarser substrate behind. The filter prevents excessive migration of the base soil particles through the voids in the armor layer, permits relief of hydrostatic pressure beneath the armor, and distributes the weight of the armor to provide more uniform settlement.

Guidance for the design of both granular and geotextile filters is provided in National Cooperative Highway Research Program (NCHRP) Report 568, "Riprap Design Criteria, Recommended Specifications, and Quality Control" (Lagasse et al. 2006), and is found in Volume 2 as Design Guide 16. When using a granular filter, the layer should have a minimum thickness of 4 times the d₅₀ of the filter stone or 6 inches, whichever is greater. When placement must occur under water, the layer thickness should be increased by 50%. In flowing water, the placement of both granular and geotextile filters becomes challenging. Under these conditions, special materials and placement techniques have been developed to ensure that a quality filter installation is achieved, as discussed in the next section.

5.3.2 Placing Geotextiles Under Water

Placing geotextiles under water is problematic for a number of reasons. Most geotextiles that are used as filters beneath riprap are made of polyethylene or polypropylene. These materials have specific gravities ranging from 0.90 to 0.96, meaning that they will float unless weighted down or otherwise anchored to the subgrade prior to placement of the armor layer (Koerner 1998). In addition, unless the work area is isolated from river currents by a cofferdam, flow velocities greater than about 1.0 ft/s (0.3 m/s) create large forces on the geotextile. These forces cause the geotextile to act like a sail, often resulting in wavelike undulations of the fabric (a condition that contractors refer to as "galloping") that are extremely difficult to control. In mild currents, geotextiles (precut to length) have been placed using a roller assembly, with sandbags to hold the fabric temporarily.

To overcome these problems, engineers in Germany have developed a product known as SandMat^TM. This blanket-like product consists of two nonwoven needle-punched geotextiles (or a woven and a nonwoven) with sand in between. The layers are stitch-bonded or sewn together to form a heavy, filtering geocomposite. The composite blanket exhibits an overall specific gravity ranging from approximately 1.5 to 2.0, so it sinks readily.

According to Heibaum (2002), this composite geotextile has sufficient stability to be handled even when loaded by currents up to approximately 3.3 ft/s (1 m/s). At the geotextile - base soil interface, a nonwoven fabric should be used because of the higher angle of friction compared to woven geotextiles. Figure 5.2 shows a close-up photo of the SandMat^TM material. Figure 5.3 shows the SandMat^TM blanket being rolled out using conventional geotextile placement equipment.

In deep water or in currents greater than 3.3 ft/s (1 m/s), German practice calls for the use of sand-filled geotextile containers. For specific project conditions, geotextile containers can be chosen that combine the resistance against hydraulic loads with the filtration capacity demanded by the application. Geotextile containers have proven to give sufficient stability against erosive forces in many applications, including wave-attack environments. The size of the geotextile container must be chosen such that the expected hydraulic load will not transport the container during placement (Heibaum 2002). Once placed, the geotextile containers are overlaid with the final armoring material (e.g., riprap or partially grouted riprap) as shown in Figure 5.4 .

Figure 5.2. Close-up photo of SandMat^TM geocomposite blanket.
(photo from NCHRP Project 24-07(2), courtesy Colcrete - Von Essen Inc.)

Figure 5.3. SandMat^TM geocomposite blanket being unrolled.
(photo from NCHRP Project 24-07(2), courtesy Colcrete - Von Essen Inc.)

Figure 5.4. Sand-filled geotextile containers.

Figure 5.5 shows a geotextile container being filled with sand. Figure 5.6 shows the sand-filled geotextile container being handled with an articulated-arm clam grapple. The filled geotextile container in the photograph is a nominal 1-metric-tonne (1,000 kg or 2,200 lb) unit. The preferred geotextile for these applications is always a non-woven needle punched fabric, with a minimum mass per unit area of 500 grams per square meter. Smaller geotextile containers can be fabricated and handled by one or two people for smaller-sized applications.

As a practical minimum, a 200-lb (91 kg) geotextile container covering a surface area of about 6 to 8 square ft (0.56 to 0.74 m²) can be fashioned from nonwoven needle punched geotextile having a minimum mass per unit area of 200 grams per square meter, filled at the job site and field-stitched with a hand-held machine. Figure 5.7 illustrates the smaller geotextile containers being installed at a prototype-scale test installation for NCHRP Project 24-07(2) (Lagasse et al. 2007) in a pier scour countermeasure application (see also Design Guidelines 11 and 12, Volume 2).

5.4 RIPRAP FAILURE MODES

As discussed in Section 5.1, riprap can be considered an integrated system for which successful performance of a riprap installation depends on the response of each component of the system to hydraulic and environmental stresses throughout its service life. This section provides an overview of failure modes for revetment and bridge pier riprap to underscore the integrated nature of riprap armoring systems and support development of inspection guidance.

Figure 5.5. Filling 1.0 metric tonne geotextile container with sand.
(photo from NCHRP Project 24-07(2), courtesy Colcrete - Von Essen Inc.)

Figure 5.6. Handling a 1.0 metric tonne sand-filled geotextile container.
(photo from NCHRP Project 24-07(2), courtesy Colcrete - Von Essen Inc.)

Figure 5.7. Two hundred lb (91 kg) sand-filled geotextile containers,
NCHRP Project 24-07(2).

5.4.1 Riprap Revetment Failure Modes

In a preliminary evaluation of various riprap design techniques, Blodgett and McConaughy (1986) concluded that a major shortcoming of all present design techniques is the assumption that failures of riprap revetment are due only to particle erosion. Procedures for the design of riprap protection need to consider all the various causes of failures which include: (1) particle erosion; (2) translational slide; (3) modified slump; and (4) slump.

Particle erosion is the most commonly considered erosion mechanism (Figure 5.8). Particle erosion occurs when individual particles are dislodged by the hydraulic forces generated by the flowing water. Particle erosion can be initiated by abrasion, impingement of flowing water, eddy action/reverse flow, local flow acceleration, freeze/thaw action, ice, or toe erosion. Probable causes of particle erosion include: (1) stone size not large enough; (2) individual stones removed by impact or abrasion; (3) side slope of the bank so steep that the angle of repose of the riprap material is easily exceeded; and (4) gradation of riprap too uniform. Figure 5.9 provides a photograph of a riprap failure due to particle displacement.

Figure 5.8. Riprap failure by particle erosion (Blodgett and McConaughy 1986).

Figure 5.9. Damaged riprap on left bank of Pinole Creek at Pinole, CA, following flood of January 4, 1982. Note deposition of displaced riprap from upstream in channel bed (photographed March 1982) (Blodgett & McConaughy 1986).

A translational slide is a failure of riprap caused by the downslope movement of a mass of stones, with the fault line on a horizontal plane (Figure 5.10). The initial phases of a translational slide are indicated by cracks in the upper part of the riprap bank that extend parallel to the channel. This type of riprap failure is usually initiated when the channel bed scours and undermines the toe of the riprap blanket. This could be caused by particle erosion of the toe material, or some other mechanism which causes displacement of toe material. Any other mechanism which would cause the shear resistance along the interface between the riprap blanket and base material to be reduced to less than the gravitational force could also cause a translational slide. Probable causes of translational slides are as follows: (1) bank side slope too steep; (2) presence of excess hydrostatic (pore) pressure; and (3) loss of foundation support at the toe of the riprap blanket caused by erosion of the lower part of the riprap blanket. Figure 5.11 provides a photograph of a riprap failure due to a translational sliding-type failure.

Modified slump failure of riprap (Figure 5.12) is the mass movement of material along an internal slip surface within the riprap blanket. The base soil underlying the riprap does not fail. This type of failure is similar in many respects to the translational slide, but the geometry of the damaged riprap is similar in shape to initial stages of failure caused by particle erosion. Probable causes of modified slump are: (1) bank side slope is so steep that the riprap is resting very near the angle of repose, and any imbalance or movement of individual stones creates a situation of instability for other stones in the blanket; and (2) material critical to the support of upslope riprap is dislodged by settlement of the submerged riprap, impact, abrasion, particle erosion, or some other cause. Figure 5.13 provides a photograph of a riprap failure due to a modified slump-type failure.

Figure 5.10. Riprap failure by translational slide (Blodgett and McConaughy 1986).

Figure 5.11. Riprap on Cosumnes River at Site 2 near Sloughhouse, CA, looking downstream, showing translational slide failure (photographed May 31, 1983) (Blodgett & McConaughy 1986).

Figure 5.12. Riprap failure by modified slump (Blodgett and McConaughy 1986).

Figure 5.13. Riprap on Consumnes River at Site 3 near Sloughhouse, CA, looking downstream, showing modified slump failure (photographed May 31, 1983) (Blodgett & McConaughy 1986).

Slump failure is a rotational-gravitational movement of material along a surface of rupture that has a concave upward curve (Figure 5.14). The cause of slump failures is related to shear failure of the underlying base soil that supports the riprap. The primary feature of a slump failure is the localized displacement of base material along a slip surface, which is usually caused by excess pore pressure that reduces friction along a fault line in the base material. Probable causes of slump failures are: (1) non-homogeneous base material with layers of impermeable material that act as a fault line when subject to excess pore pressure; (2) side slopes too steep and gravitational forces exceeding the inertia forces of the riprap and base material along a friction plane; and (3) too much overburden at the top of the slope (may be caused in part by the riprap). Figure 5.15 provides a photograph of a riprap failure due to a slump-type failure.

Figure 5.14. Riprap failure due to slump (Blodgett and McConaughy 1986).

Figure 5.15. Riprap on left bank Cosumnes River at Site 1 near Sloughhouse, CA, showing slump failure (photographed May 31, 1983) (Blodgett & McConaughy 1986).

Summary: Blodgett and McConaughy (1986) conclude that certain hydraulic factors are associated with each of the four types of riprap failure (particle erosion, translational slide, modified slump, and true slump). While the specific mechanism causing failure of the riprap is difficult to determine, and a number of factors, acting either individually or combined, may be involved, they identify the reasons for riprap failures as:

1. Particle size was too small because:
   • Shear stress was underestimated
   • Velocity was underestimated
   • Inadequate allowance was made for channel curvature
   • Design channel capacity was too low
   • Design discharge was too low
   • Inadequate assessment was made of abrasive forces
   • Inadequate allowance was made for effect of obstructions
2. Channel changes caused:
   • Impinging flow
   • Flow to be directed at ends of protected reach
   • Decreased channel capacity or increased depth
   • Scour
3. Riprap material had improper gradation
4. Material was placed improperly
5. Side slopes were too steep
6. No filter blanket was installed or blanket was inadequate or damaged
7. Excess pore pressure caused failure of base material or toe of riprap
8. Differential settlement occurred during submergence or periods of excessive precipitation

5.4.2 Pier Riprap Failure Modes

A study of the causes of riprap failure at model bridge piers (Chiew 1995) under clear-water conditions with gradually increasing approach flow velocities defined three modes of pier riprap failure:

1. Riprap shear failure - whereby the riprap stones cannot withstand the downflow and horseshoe vortex associated with the pier scour mechanism.
2. Winnowing failure - whereby the underlying finer bed material is removed through voids or interstices in the riprap layer.
3. Edge failure - whereby instability at the edge of the coarse riprap layer and the bed sediment initiates a scour hole beginning at the perimeter and working inward that ultimately destabilizes the entire layer.

Since live-bed conditions are more likely to occur during flood flows, additional experiments were conducted to evaluate the stability of pier riprap under live-bed conditions with migrating bed forms (Lim and Chiew 1996). These experiments and subsequent research Melville et al. (1997), Lauchlan (1999), and Lauchlan and Melville (2001) indicates that bed-form undermining is the controlling failure mechanism at bridge piers on rivers with mobile bed forms, especially sand bed rivers. The most important factors affecting the stability of the riprap layer under live-bed conditions were the turbulent flow field around the pier and the fluctuations of the bed level caused by bed forms (e.g., dunes) as they migrate past the pier. The three failure modes defined for clear-water conditions also exist under live-bed conditions and they may act independently or jointly with migrating bed forms to destabilize the riprap layer.

Once sediment transport starts and bed forms associated with the lower flow regime (i.e., ripples and dunes) begin to form, the movement of sediments at the edge of the riprap layer remove the support of the edge stones. When the trough of a bed feature migrated past the riprap layer, stones would slide into the trough, causing the riprap layer to thin. Depending on the thickness of the remaining riprap layer following stone sliding and layer thinning, winnowing may occur as a result of exposure of the underlying fine sediments to the flow. Winnowing can cause the entire remaining riprap layer to subside into the bed. With thicker riprap layers winnowing is not a factor and there is no subsidence.

Under steady flow conditions, the inherent flexibility of a riprap layer can provide a self-healing process (Chiew 1995). As scour occurs and sediment is removed from around the riprap layer through the three modes of erosion described above, the riprap layer, if it has sufficient thickness, can adjust itself to the mobile channel bed and remain relatively intact while providing continued scour protection for the pier. However, when flow velocity is steadily increased, riprap shear, winnowing, and edge erosion combine to cause either a total disintegration or embedment failure of the riprap layer in the absence of an underlying filter (either geotextile or granular).

Total disintegration, which is characterized by a complete breakup of the riprap layer whereby the stones are washed away by the flow, occurs when the self-healing ability of the riprap layer is exceeded by the erosive power created by higher flow velocity. Total disintegration occurs when the riprap stone size to sediment size ratio is small. Embedment failure occurs when: (1) the riprap stones are large compared to the bed sediment and local erosion around the individual stones causes them to embed into the channel bed (i.e., differential mobility); and (2) the riprap stones lose their stability as bed forms pass and drop into the troughs of the migrating bed forms (i.e., bed feature destabilization).

5.4.3 Pier Riprap Failure Modes - Schoharie Creek Case Study

The failure of the I-90 bridge over Schoharie Creek near Albany, New York on April 5, 1987, which cost 10 lives, was investigated by the National Transportation Safety Board (NTSB) (Richardson and Davis 2001). The peak flow was 64,900 cfs (1,838 m³/s) with a 70- to 100-year return period. The foundations of the four bridge piers were large spread footings 82 ft (25 m) long, 18 ft (5.5 m) wide, and 5 ft (1.5 m) deep without piles. The footings were set 5 ft (1.5 m) into the stream bed in very dense ice contact stratified glacial drift, which was considered nonerodible by the designers (Figure 5.16). However, flume studies of samples of the stratified drift showed that some material would be eroded at a velocity of 4 ft/s (1.5 m/s), and at a velocity of 8 ft/s (2.4 m/s) the erosion rates were high.

Figure 5.16. South elevation - Schoharie Creek Bridge showing key structural features and a schematic geological section (Richardson et al. 1987).

A 1 to 50 scale, 3-dimensional, model study indicated a prototype flow velocity of 10.8 ft/s (3.3 m/s) at the pier that failed. Also, the 1 to 50 scale and a 1 to 15 scale, 2-dimensional model study gave 15 ft (4.6 m) of maximum scour depth. The scour depth of the prototype pier (pier 3) at failure was 14 ft (4.3 m) (Figure 5.17).

Figure 5.17. Pier scour holes at Schoharie Creek Bridge in 1987. Pier 2 in the foreground with pier 3 in the background.

Design plans called for the footings to be protected with riprap. Over time (1953 to 1987), much of the riprap was removed by high flows. NTSB gave as the probable cause "… the failure of the New York State Thruway authority to maintain adequate riprap around the bridge piers, which led to severe erosion in the soil beneath the spread footings. Contributing to the severity of the accident was the lack of structural redundancy in the bridge" (NTSB 1988).

The NYSTA inspected the bridge annually or biennially with the last inspection on April 1, 1986. A 1979 inspection by a consultant hired by NYSDOT indicated that most of the riprap around the piers was missing (Figures 5.18 and 5.19); however, the 1986 inspection failed to detect any problems with the condition of the riprap at the piers. Based on the NTSB findings, the conclusions from this failure are that inspectors and their supervisors must recognize that riprap does not necessarily make a bridge safe from scour, and inspectors must be trained to recognize when riprap is missing and the significance of this condition.

Summary : Examples of the most common modes of riprap failure at piers provide guidance for post-flood and post-construction performance evaluation. Inspectors need to be aware of, and understand, the causes of riprap inadequacies that they see in the field. While the specific mechanism causing failure of the riprap is difficult to determine, and a number of factors, acting either individually or combined, may be involved, the reasons for riprap failures at bridge piers can be summarized as follows:

Figure 5.18. Photograph of riprap at pier 2, October 1956.

Figure 5.19. Photograph of riprap at pier 2, August 1977 (flow is from right to left).

1. Particle size was too small because:
   • Shear stress was underestimated
   • Velocity was underestimated
   • Inadequate allowance was made for channel curvature
   • Design channel capacity was too low
   • Design discharge was too low
   • Inadequate assessment was made of abrasive forces
   • Inadequate allowance was made for effect of obstructions (such as debris)
2. Channel changes caused:
   • Increased angle of attack (skew)
   • Decreased channel capacity or increased depth
   • Scour
3. Riprap material had improper gradation
4. Material was placed improperly
5. No filter blanket was installed or blanket was inadequate or damaged

5.5 RIPRAP INSPECTION GUIDANCE

5.5.1 General

Inspection of riprap placement typically consists of visual inspection of the installation procedures and the finished surface. Inspection must ensure that a dense, rough surface of well-keyed graded rock of the specified quality and sizes is obtained, that the layers are placed such that voids are minimized, and that the layers are the specified thickness.

If the riprap installation is part of channel stability works in the vicinity of a bridge, it is typically inspected during the biennial bridge inspection program. However, more frequent inspection might be required by the Plan of Action for a particular bridge or group of bridges. In some cases, inspection may be required after every flood that exceeds a specified magnitude. Underwater inspection of a riprap system should only be performed by divers specifically trained and certified for such work.

The following general guidance for inspecting riprap is presented in the National Highway Institute (NHI) training course 135047, "Stream Stability and Scour at Highway Bridges for Bridge Inspectors:"

1. Riprap should be angular and interlocking (Old bowling balls would not make good riprap). Flat sections of broken concrete paving do not make good riprap.
2. Riprap should have a granular or synthetic geotextile filter between the riprap and the subgrade material.
3. Riprap should be well graded (a wide range of rock sizes). The maximum rock size should be no greater than about twice the median (d₅₀) size.
4. For bridge piers, riprap should generally extend up to the bed elevation so that the top of the riprap is visible to the inspector during and after floods.
5. When inspecting riprap, the following are strong indicators of problems:
   • Has riprap been displaced downstream?
   • Has angular riprap blanket slumped down slope?
   • Has angular riprap material been replaced over time by smoother river run material?
   • Has riprap material physically deteriorated, disintegrated , or been abraded over time?
   • Are there holes in the riprap blanket where the filter has been exposed or breached?

5.5.2 Guidance for Recording Riprap Condition

To guide the inspection of a riprap installation, a recording system is presented in Appendix D . This guidance establishes numerical ratings from 0 (worst) to 9 (best). Recommended action items based on the numerical rating are also provided (Lagasse et al. 2006).

5.5.3 Performance Evaluation

The evaluation of any revetment system's performance should be based on its design parameters as compared to actual field experience, longevity, and inspection / maintenance history. To properly assess the performance of revetment riprap, the history of hydraulic loading on the installation, in terms of flood magnitudes and frequencies, must also be considered and compared to the design loading. Guidance for the performance evaluation of riprap armoring systems is provided in NCHRP Report 593 (Lagasse et al. 2007).

Changes in channel morphology may have occurred over time subsequent to the installation of the riprap. Present-day channel cross-section geometry and planform should be compared to those at the time of installation. Both lateral and vertical instability of the channel can significantly alter hydraulic conditions at the site. Approach flows may exhibit an increasingly severe angle of attack (impinging flow) over time, increasing the hydraulic loading on the riprap.

Deficiencies noted during the inspection should be corrected as soon as possible. As with any armor system, progressive failure from successive flows must be avoided by providing timely maintenance intervention.

5.6 GROUTED AND PARTIALLY GROUTED RIPRAP

Grouted riprap is rock slope paving with voids filled with concrete grout forming a monolithic armor. Because fully grouted riprap is a rigid structure, it will not conform to bank settlement or toe undermining as loose riprap does. Therefore, fully grouted riprap is susceptible to mass failure, especially if pore water is not allowed to drain properly. Although the revetment is rigid, it is not particularly strong and even a small loss of toe or bank support can result in the failure of large portions of the structure.

The primary advantage of fully grouted riprap is that the grout anchors the rock and eliminates particle erosion of the revetment. Therefore, smaller rock can be used for the revetment, and the total thickness of the revetment can be reduced as compared with traditional riprap revetment. Another advantage is that a relatively smooth surface can be achieved and, therefore, the hydraulic efficiency of the waterway is improved. Filters are not required for fully grouted riprap but drainage of pore water must be provided. A significant disadvantage of fully grouted riprap is that a complete layer of grout converts a flexible revetment to a rigid cover, subject to the potential problems of any rigid slope paving, including undercutting at the toe, out flanking, and the possibility of sudden catastrophic failure .

An alternative to fully grouted riprap is partially grouted riprap. In general, the objective is to increase the stability of the riprap without sacrificing its flexibility. Partial grouting of riprap may be well suited for areas where rock of sufficient size is not available to construct a loose riprap revetment.

The River and Channel Revetments design manual published by H.R. Wallingford in the United Kingdom (Escarameia 1998) provides design guidance for grouting "hand pitched stone" with both bituminous and cement grout. For grouting riprap in the United Kingdom, bitumen is the material most commonly used. Although various degrees of grouting are possible, effective solutions are usually produced when the bituminous mortar envelopes the loose stone and leaves relatively large voids between rock particles. The degrees of bituminous grouting available are:

• Surface grouting (which does not penetrate the whole thickness of the revetment and corresponds to about one-third of the voids filled)
• Various forms of pattern grouting (where only some of the surface area of the revetment is filled, between 50 to 80% of voids)
• Full grouting (an impermeable type of revetment)

Partial grouting of riprap with a cement slurry is presented as one of several standard design approaches for permeable revetments in a discussion of considerations regarding the experience and design of German inland waterways (Heibaum 2000). Partially grouted riprap consists of appropriately sized rocks that are grouted together with grout filling only 1/3 to 1/2 of the total void space (Figure 5.20). In contrast to fully grouted riprap, partial grouting increases the overall stability of the riprap installation unit without sacrificing flexibility or permeability. It also allows for the use of smaller rock compared to standard riprap, resulting in decreased layer thickness. Design, specification, and construction guidance for partially grouted riprap is provided in Design Guideline 12, Volume 2.

The holes in the grout allow for drainage of pore water so a filter is required. The grout forms conglomerates of riprap so the stability against particle erosion is greatly improved and, as with fully grouted riprap, a smaller thickness of stone can be used (Figure 5.21). Although not as flexible as riprap, partially grouted riprap will conform somewhat to bank settlement and toe exposure.

An important consideration for partially grouted riprap is that construction methods must be closely monitored to insure that the appropriate voids and surface openings are provided. Contractors in Germany have developed techniques and equipment to achieve the desired grout coverage and the right penetration. Various European countries have developed special grout mixes and construction methods for underwater installation of partially grouted riprap (see Design Guideline 12).

Figure 5.20. Close-up view of partially grouted riprap.

Figure 5.21. "Conglomerate" of partially grouted riprap, Federal Waterway Engineering and Research Institute, Karlsruhe, Germany (Heibaum 2000).

5.7 CONCRETE ARMOR UNITS

Concrete armor units, also known as artificial riprap, consist of individual pre-cast concrete units with complex shapes that are placed individually or in interconnected groups. These units were originally developed for shore protection to resist wave action during extreme storms. All are designed to give a maximum amount of interlocking using a minimum amount of material. These devices are used where natural riprap is unavailable or is more costly to obtain than fabrication of the artificial riprap units. Parker et al. (1998) provide a review of studies conducted on the use of concrete armor units as pier scour countermeasures.

Various designs for size and shape of concrete armor units are available and include such commercial names as Tetrapods, Tetrahedrons, Toskanes, Dolos, Tribars, Accropodes, Core-Loc^TM, and A-Jacks^® (Figure 5.22). Because concrete armor units are similar to riprap, they can be susceptible to the same failure mechanisms as riprap. The use of a filter layer or geotextile in conjunction with these types of devices is often required, especially in coastal applications, and a geotextile or filter may be critical to the stability of these devices when used as pier scour protection (see Design Guideline 19 for design procedures for Toskanes and A-Jacks^®).

Figure 5.22. Concrete armor units.

The primary advantage of armor units is that they usually have greater stability compared to riprap particles of equivalent weight. This is due to the interlocking characteristics of their complex shapes. The increased stability allows their placement on steeper slopes or the use of lighter weight units for equivalent flow conditions as compared to riprap. This is significant when riprap of a required size is not available.

The design of armor units in open channels is based on the selection of appropriate sizes and placement patterns to be stable in flowing water. The armor units should be able to withstand the flow velocities without being displaced. Hydraulic testing is used to measure the hydraulic conditions at which the armor units begin to move or "fail," and dimensional analysis allows extrapolation of the results to other hydraulic conditions. Although a standard approach to the stability analysis has not been established, design criteria have been developed for various armor units using the following dimensionless parameters:

• Isbash stability number (Parola 1993; Ruff and Fotherby 1995; Bertoldi et al. 1996)
• Shields parameter (Bertoldi et al. 1996)
• Froude number (Brown and Clyde 1989)

The Isbash stability number and Shields parameter are indicative of the interlocking characteristics of the armor units. Froude number scaling is based on similitude of stabilizing and destabilizing forces. Quantification of these parameters requires hydraulic testing and, typically, regression analysis of the data. Prior research and hydraulic testing have provided guidance on the selection of the Isbash stability number and Shields parameter for riprap and river sediment particles, but stability values are not available for all concrete armor units. Therefore, manufacturers of concrete armor units have a responsibility to test their products and to develop design criteria based on the results of these tests. Since armor units vary in shape and performance from one proprietary system to the next, each system will have unique performance properties.

Installation guidelines for concrete armor units in streambank revetment and channel armor applications should consider subgrade preparation, edge treatment (toe down and flank) details, armor layer thickness, and filter requirements. Subgrade preparation and edge treatment for armor units is similar to that required for riprap. Considerations for armor layer thickness and filter requirements are product specific and should be provided by the armor unit manufacturer.

Concrete armor units have shown potential for mitigating the effects of local scour in the laboratory; however, only limited data are available on their performance in the field. Research efforts are currently being conducted to test the performance of concrete armor units as pier scour countermeasures in the field.

Design methods which incorporate velocity (a variable which can be directly measured) are commonly used to select local scour countermeasures. Normally an approach velocity is used in the design equation (generally a modified Isbash equation) with a correction factor for flow acceleration around the pier or abutment (see Section 5.2.3).

Although tetrahedrons are currently used for bank protection (Fotherby 1995), they have garnered very little interest with regard to pier scour protection in the United States. This may be primarily related to their lack of appendages and interlock (i.e., their simple compact shape is similar to riprap and spheres). Dolos also have not been seriously considered for use as pier scour protection because they have no inherent interlocking property to resist movement under steady state turbulent flow (Brebner 1978). Extensive testing and research has been conducted on the Core-Loc^® system, which was developed by the U.S. Army Engineer Waterways Experiment Station, but the testing was limited exclusively to coastal applications. Accropode and tribar systems are used almost exclusively in coastal applications as well.

In contrast, tetrapods have been extensively studied and evaluated for use as pier scour protection (Fotherby 1992, 1993; Bertoldi et al. 1996; Jones et al. 1995; Bertoldi and Kilgore 1993). Fotherby (1992, 1993) and Stein et al. (1998) suggest that tetrapods offer little advantage compared to riprap in terms of stability. Layering and density had no appreciable effect on the stability of the tetrapods, although the stability increased with the size of the tetrapod pad. Work by Bertoldi et al. (1996) and Stein et al. (1998) indicates that riprap and tetrapods behaved comparably when both stability number and spherical stability number were compared, and also suggest that fixing the perimeter and varying the number of tetrapod layers may have an effect on stability.

A specific design procedure for Toskanes has been developed for application at bridge piers and abutments where the Toskanes are installed as individual, interlocking units. The design procedures for Toskanes are based on extensive research conducted at Colorado State University (Ruff and Fotherby 1995; Fotherby 1995; Burns et al. 1996; Fotherby and Ruff 1998, 1999). Based on hydraulic model studies conducted at Colorado State University for the Pennsylvania Department of Transportation, Burns et al. (1996) presented procedures for the design of Toskane pads, provided criteria for sizing Toskanes, and suggested techniques for installation of Toskanes (Figure 5.23). No other concrete armor unit has been as extensively tested and evaluated for use as a pier scour countermeasure (see Design Guideline 19).

Another approach to using concrete armor units for pier scour protection involves the installation of banded modules of the A-Jacks^® armor unit (Ayres Associates 1999; Thornton et al. 1999). Laboratory testing results and installation guidelines developed at Colorado State University by Ayres Associates (1999) for the A-Jacks^® system illustrate the "modular" design approach in contrast with the "discrete particle" approach for Toskanes (Figure 5.24).

The discrete particle design approach illustrated by the Toskane design guidelines concentrates on the size, shape, and weight of individual armor units, whether randomly placed or in stacked or interlocked configurations. In contrast, the basic construction element of A-Jacks^® for pier scour applications is a "module" comprised of a minimum of 14 individual A-Jacks^® banded together in a densely-interlocked cluster, described as a 5x4x5 module. The banded module thus forms the individual design element as illustrated in Figure 5.24 (see Design Guideline 19).

As with any countermeasure, the armor units must withstand several potential failure modes. Providing resistance against the hydraulic stresses may not be sufficient for structure success. If the armor units are used to counter pier scour, they must also remain stable for channel degradation, contraction scour, and the passage of bed forms (dunes). If armor units are used as bank revetment, then the stability of the bank must be analyzed for potential toe scour, pore water pressures and saturated soil strengths.

It should also be noted that concrete armor units, depending on their size, may be very susceptible to vandalism. In addition, there may be maintenance and degradation issues associated with any cables used to tie groups of concrete armor units together.

Figure 5.23. Laboratory study of Toskanes for pier scour protection.

Figure 5.24. Installation of A-Jacks® modular units installed by Kentucky DOT for pier scour protection.