Erosion d’une vase de l’estuaire de la Loire sous l’action du courant

Résumé

Le phénomène d’érosion des vases sous l’action d’un courant est abordé expérimentalement. Cette étude comporte six séries d’essais d’érosion réalisées en laboratoire sur une vase de Loire pour six concentrations différentes. Pour chaque concentration, trois essais successifs ont été effectués afin de pouvoir analyser les résultats statistiquement. La rigidité initiale et la concentration des vases sont mesurées en suivant un protocole d’essai. Des lois phénoménologiques sont proposées pour relier le taux d’érosion mesuré à la rigidité initiale du sédiment et à la contrainte hydrodynamique. Un piège à sédiments intégré dans le dispositif d’essai, permet d’obtenir un échantillon de matériau érodé. La taille des copeaux de vase érodés et leur densité sont mesurées grâce à une technique expérimentale originale. Une expression mathématique est ainsi avancée pour relier la taille des copeaux à la densité et à la rigidité initiale du dépôt dont ils proviennent.

Abstract

This research work deals with an experimental study on the erosion phenomenon of a mud under the action of a water current. It is observed from research works carried out by Partheniades (J Hydraul Div ASCE 91(HY1):105–139, 1965), Migniot (La Houille Blanche 1&2:11–29, 95–111, 1989), Ockenden and Delo (GeoMar Lett 11:138–142, 1991), Aberle et al. (Mar Geol 207:83–93, 2004), among others, that mud erosion process by an hydrodynamic action depends mainly on sediment properties. Based on a literature study, this critical stress is assumed to be proportional either to the effective cohesion (Eq. 2) or to the yield stress (Eq. 4) of the sediment bed. Six erosion test series have been performed at six different concentrations of a mud from the Loire estuary. Some properties of the tested sediments are: a solid particles density ρ _s = 2,550 kg m⁻³, a liquidity and plasticity limit at 140 and 70% of the water content, respectively, a mean size of the dispersed mineral fraction determined by laser techniques of 10 μm, and a volatile matters content of 11.86% by total dry weight burned at 550°C. As the rheological behaviour is difficult to describe, the sediment strength is characterized by only a parameter, namely, the yield stress τ _y . It is measured with a coaxial cylinder Brookfield LVT viscosimeter following a defined procedure (Hosseini in Liaison entre la rigidité initiale et la cohésion non drainée dans les vases molles—Relation avec la dynamique sédimentaire. Thèse, Université de Nantes, 167 p, 1999). For each studied concentration, three successive erosion tests are carried out, and for every erosion test, 15 successive measurements of τ _y are made. The mean values and the standard deviations of τ _y are shown in Table 1 as a function of the bed sediment concentration C. A confined flume has been conceived and built to characterize the erosion rates. With this device, a current-induced shear stress is generated above an homogeneous deposited sediment (Fig. 1). The bed shear stress τ _o is calculated from the measured mean velocity V by Eq. 5. The friction coefficient c _f involved in Eq. 5 has been evaluated from measurements of the hydraulic pressure loss. Finally, the validity of Eq. 5 has been confirmed by five calibration tests on the incipient of the movement of sands for which results are compared with Shields diagram in Fig. 2. The observed erosion mechanisms affecting cohesive sediments depend mainly on the value of the yield stress. For a fluid mud (τ _y less than 3 N m⁻²), the bed shear stress produces at first a wavy motion on the bed surface with a progressive undulation. When the shear stress increases, resuspension produces a diluted sediment cloud which is entrained and dispersed by the flow. For a plastic mud (τ _y greater than 3 N m⁻²), erosion occurs by a wrenching of aggregates which are transported near the bottom. Initially, the eroded aggregates measure several millimetres in size; but, once transported, aggregates break into very rigid and compact aggregates of maximum size close to 1 mm in all the cases. During erosion tests, erosion volumetric rates E _v have been evaluated under steady-state bed shear stress τ _o from the observed variation in time of suspended sediment concentration. A generalized erosion is assumed for E _v ≥ E _vo = 3 × 10⁻⁷ m s⁻¹ (that is the equivalent of 1 mm sediment layer eroded per hour). The generalized erosion occurs above a critical bed shear stress τ _oe which is linked to τ _y and mud density through Eq. 7. A phenomenological law defined by Eq. 8 is proposed to calculate the erosion rate as a function of yield stress and hydrodynamic shear stress at the bottom. For τ _o < τ _oe , a small erosion is observed, which is described by a formulation given by Cerco et al. (Water quality model of Florida Bay. U.S. Army Engineer Research and Development Center, ERDC/EL TR−00-10, Vicksburg, USA, 260 p, 2000). Above τ _oe , a small increase in τ _o produces an important erosion rate rise which is described by a formulation suggested by Mehta and Partheniades (Resuspension of deposited cohesive sediment beds. In: Proceedings of the 18th coastal engineering conference, Cape Town, South Africa, 2:1569–1588, 1982). Figure 3 shows as a function of τ _o the observed values of E _v (discrete symbols) as well as the proposed model (in solid lines) for the six concentrations. For the theoretically maximum value of E _v , the hydrodynamic shear stress is very high with regard to the sediment cohesion; and then, the water flux produces an entrainment of underlying fluid mud layers. The asymptotic law obtained from the model of Kranenburg and Winterwerp (1997) plotted in dashed lines in Fig. 3, is expressed by E _v = 0.3 u _* . A sediment trap inserted in the experimental system allows a sample of eroded mud aggregates to be obtained. It is observed that the maximum value of the diameter D _M of the eroded aggregates depends on the density and yield stress of the initially deposited mud according to Eq. 10. In the same way, the density of the aggregates issued from plastic mud erosion is measured following an original experimental method and procedure (Table 3). The erosion of plastic muds with a concentration from 310 to 420 kg m⁻³ produces aggregates with a concentration close to 400 kg m⁻³ and yield stress a little greater than 100 N m⁻².