PhD thesis Design of Pattern placed revetments - D.J. Peters
The Dutch coast is protected by beaches, dunes and dikes. Erosion of the outside slopes of dikes by waves is a potential threat during storm conditions when high water levels and severe waves occur. Revetment systems prevent erosion of the dikes and they need to be stable under wave attack.
A sufficiently large size and weight of revetment elements is the main contribution to their stability. Traditional smooth revetment systems often consist of relatively small single blocks or column-shaped natural stones or concrete elements that are placed in a regular pattern. The pattern creates a regular distribution of joints and voids which limit the build-up of water pressure in the system. The pattern also contributes to the mechanical resistance against wave pressures. When applied on a slope gravity induces in-plane forces in the revetment and provides coherence through so-called frictional interlocking. In this thesis this mechanism is studied in detail with model simulations and field measurements. The contribution of frictional interlocking makes the observed high resistance of pattern-placed revetment structure against concentrated loads comprehensible. This phenomenon can be used in the development of new revetment systems and in optimization of the design of revetment slopes and dikes.
Dikes and dunes prevent low-lying land against flooding. High tides and surges create conditions under which dikes and dunes are actually in service. Waves from the sea tend to erode beaches and dunes. Sandy material is taken away and disposed elsewhere on the coast, where it contributes to accretion of dunes again. For man-made dikes this survival mechanism is not an option. They have too little volume and are designed to maintain their shape. Although made of sand, they are covered with impermeable clay which is, when overgrown with grass, to some extent resistant to erosion caused by waves. This resistance however falls short in the tidal zones, where grass has no chance to survive and at high tides the waves can become too high for grass-and-clay dikes. Dikes are therefore covered with armour layers in the appearance of stones, rock or concrete elements.
In order to understand the development of the system a historic perspective is useful. The present-day typical revetment system is a result of developments through the last 200 years. As in many fields of engineering, revetment systems could be developed because of an increase in knowledge as well as improved economical means.Amongst the variety of available systems the pitched Basalt-type column revetment on an open granular filter layer became dominant since 1870. They have become an icon of the Dutch hydraulic engineering structures, and are even valued as monumental elements of the coastal scenery in Zeeland and elsewhere.
During the 20th century a number of experimental concrete protection systems were developed and tested in practice. After the big flood of 1953, in the beginning of the dike upgrade under the Delta Act, concrete blocks on clay were popular but rejected later. More recent concrete revetments are again of the column type, with an open granular filter layer, more or less inspired by the Basalt slopes. Studying the history of revetments it can be noticed that some technical improvements did not meet the expectations. Although empirical knowledge was dominant, important principles and valuable experiences were not always valued and incorporated in more recent developments.
Non-technical influences play a role as well. Maintenance, periodical upgrade and replacement of the revetment structures have always been costly activities. An interesting coincidence with the economic and regional societal background of the protected polders can be observed. Remarkable recent arguments for the persistent preference for gentle slopes and smooth pitched revetments are their accessibility and esthetical appearance.
Designing a revetment slope as a concrete structure within a stand-alone activity is not potentially successful. The system is influenced by the wave conditions, the foreland geometry, the slope angle and slope length and consequently interacts with the complete dike design. Many authors provide supporting data obtained in field observations. The phenomena of wave loads on slopes are explored through literature, data on flume tests and field observations. In this thesis recent findings with respect to these subjects are collected from international literature. The findings are described in detail, discussed and where possible improved.
The principle of placing the stones or elements in a regular pattern is vital to the functioning of the armour.Firstly, pattern-placing is creating a pretty much closed slab, the so-called top-layer, which controls the water flows during the cyclic wave loading. When the holes and joints of the slab are adequately designed and the permeability properties are balanced with the gravel and rubble layers underneath they determine and control the build-up of water pressure under the top-layer, and therewith play an essential role in the stability of the system under wave loads.
It seems likely to believe that the pattern-placing as such also makes the structure stronger. The elements cannot be lifted out by just exerting them with forces that balance their weight. This aspect of the pattern-placing has not been researched nor extensively verified and has been the focus of the research conducted for and described in this thesis.A definition of how pattern-placed revetment and the associated element-to-element interaction works relies on gravity forces.
The elements encounter frictional forces which cause them to respond like interlocking elements, and not as single elements. This is called frictional interlocking and the forces that compress the elements and joints are called in-plane forces, or normal forces.
As such, available knowledge of structural mechanics can be applied to the unknown behaviour of the revetments by considering the revetment as a coherent structural system. Various authors have suggested solutions in this direction. This study provides a comprehensive introduction of state-of-art structural modelling dedicated to modelling revetments as elements with joints, beams, slabs, dynamic systems, etcetera. This way of modelling of the revetment structure creates a possibility of a better evaluation of test data collected by different authors.
Theoretical evaluation of the stability of pattern-placed revetments has begun with testing. Pull tests have been designed and executed by various authors. This started in the 1970s and 80s and ultimately resulted in a very large test campaign in the 1990s with over 10,000 single element field tests.
In this study the evaluation of these tests was revisited and interpreted based on a newly developed structural slab model. From the experimental results, evidence was derived for the latent presence of normal force of a certain level. The experiments however fall short in really testing the wave-resisting capabilities of the revetments.
An important original element in this thesis is the design of new tests. Tests that simulate wave action in a measurable and comprehensible way were not executed before. The existing single stone pull tests do not represent wave loads on the revetment. The wave tests in wave flumes entangle the unknowns in the hydraulic and the mechanical performance of the system. Therefore the results of the real wave tests could not be correctly nor reliably interpreted andcould not be translated to design rules up to now. The newly designed tests fill the gap between the existing test methods. New pull tests focus on the mechanical behaviour and were performed using a load pattern that represents real wave loads.
A first new experimental verification has been obtained with trivial full scale load tests in the laboratory.From those tests evidence was found for the principle of the revetment acting as a beam or slab forming a coherent structure consisting of loose elements. Moreover the exact behaviour could be measured. Model predictions could be verified and further interpretation and refinement became possible.
Partly similar new field tests were executed on real structures. Those experiments were useful to investigate real properties, parameter values as well as for correlating them with initial conditions of in-plane stress states in the slopes. The well-known finding that pull-out strengths are location dependent was confirmed. They are stronger at lower positions on the slope because a larger portion of the slope, with a larger weight, is providing interlock forces. Weak spots were however found close to the toe structure at the bottom edge of the revetment slab. Additionally, a remarkable and undesirable influence of the ambient temperature was found for some types of revetments. It was also found that old revetment slopes are more resistant than young slopes. This was already common knowledge through past experiences. Maiden revetment slopes are much more vulnerable to waves. Cases of revetments that showed damage and even large failures in their first year are well-known.
In order to find out what series of waves do to this in-plane stress state on the slope, extra data were extracted form large-scale flume tests using additional instrumentation. It was found that during the exposure to waves the revetment slope tends to build-up normal force and hence becomes stronger over time.
The undervalued effect of the structure at the toe of the dike slope as an important stiff boundary and important condition for effective development of interlocking action is included in the models and verified by means of field tests.A dominant observation in many tested structures was the substantial uncertainty in the experimental results. The principle of frictional interlocking has – in the dry condition – a huge inherent uncertainty caused by the undetermined equilibrium of the element of the slope. In theory, the elements do not need the toe and their downstairs neighbours to withstand gravity. Concrete elements positioned on an e.g. 1:4 rubble slope can find equilibrium based on friction forces between the top-layer and the bedding. The elements can remain in position without any in-plane force. In practise the in-plane force was not zero, but has low values now and then. Wave action needs to shake up the elements, re-arrange their positions and with that their in-plane stress state. That mechanism develops easier at steeper slopes. Steeper slopes are therefore considered more reliable for their dependency on frictional interlocking.
Full scale revetment slope testing in wave flumes has been performed by various researchers. Those experiments are used to verify model predictions of wave heights, breaker heights on the slope, differential water heads over the top-layer and ultimately in the achieved stability numbers. Those experiment results were used as a calibration of the total model approach.
A design philosophy for pattern-placed revetments that is based on present day structural design methods is now available, rather than a design method solely based on empirical methods and the results of arbitrarily interpreted tests. This results in a calibrated design model that helps to understand the real behaviour and resistance of revetments under wave attack. The model more clearly distinguishes between safe and unsafe structures. Based on the model better structures can be designed.
The findings are in summary:
- A strong and stiff toe structure capable of supporting the component of the weight of the full revetment slope length is a key factor for the development of the potential of resistance of the pattern-placed revetment.
- The tallness of the individual elements (or a high height over width ratio) is more important than their individual weight.
- It is beneficial if a revetment type generates in-plane forces in both directions. Plunging breakers form the highest load on the revetment. They create high pressures with a short peak and a limited footprint in both the transverse and the long direction of the dike. Frictional interlocking in both directions is hence contributing to the resistance of the system.
- Recent innovations of revetment types have concentrated on wave run-up reduction. A next step might be possible in unbundling the mechanical and permeability properties of the joints.
- Step slopes experience a larger contribution from frictional interlocking forces and are hence stronger than gentle slopes. Combined with the effect that a gentle slope is longer, gentle slopes require significantly more revetment material. Steep slopes however have a higher risk of instability of the impermeable cover layers and have an impact on the required dike height in order to prevent overtopping. In order to avoid both drawbacks, dikes with steep top and bottom slopes and wide berms might be an economic solution.
This study forms a part of a wider research effort which has led to an integrated approach of flood defence safety evaluation that was implemented and used during the recurring national dike ring evaluations of 2007 and 2012 in the Netherlands.