A brief review is presented of the importance of dilatancy prior to rock failure and especially of its fundamental role in earthquake prediction. Constitutive equations for the visco-elastic-dilatant behaviour of rocks are presented. The dilatancy parameters are determined from various triaxial tests. The process of dilatancy prior to earthquakes is discussed on the basis of this hypothetical model.


Cette communication examine le rôle important des dilatations survenant avant la rupture de roches et plus particulièrement leur rôle fondamental dans la prediction des tremblements de terre. On presente des equations pour l'etude du comportement visco-elastique des roches et on determine les paramètres de dilatation pour divers essais triaxiaux. C'est sur la base de ce modàle hypothetique qu'on analyse le processus de dilatation avant les tremblements de terre.


Die Arbeit gibt einen kurzen Überblick ueber die Bedeutung der Dilatanz vor dem Bruch des Gesteins und besonders ihre grundlegende Rolle in der Voraussage der Erdbeben. Die Ausgangsgleichungen fuer das elasto-visko-dilatante Verhalten der Gesteine werden aufgestellt und die Dilatations-Parameter fuer verschiedene Triaxial-Versuche bestimmt. Die Dilatations- Vorgange vor den Erdbeben werden aufgrund dieses hypothetischen Modells behandelt.

The concept of dilatancy is well known in rheology and soil mechanics. Its origin is in the observation that wet sands dilates under the action of shearing stresses (Reynolds1885). Dilatancy in granular media is associated with the overall decrease of packing density due to relative movements of groups of grains; it is a geometrical necessity in the deformation process. Dilatancy of rocks was first observed by Bridgman (1949); Handin et al (l963) measured dilatancy of rock samples of Berea sandstone at low confining pressures, while decrease in porosity (volume hardening)at high confining pressures. Dilatancy under deviatoric stresses is observed as a time dependent volume increase; a dilatancy model is shown in Fig.1; later many experimental results and their analysis have been published from laboratory testing by Paterson (l963), Edmond and Paterson (1972), Brace, Paulding and Scholz (1966), Bienawski (1967), Crouch (1970), Zobackand Byerlee (1975), Perkins, Green and Friedman (1970), Rummel (1974), Mogi (1977),Tan (1964) and others which are not listed here. The results of these investigators will be summarised and analysed in the next paragraphs. In Rock-engineering the phenomenon of dilatancy in rocks is often observed; although it is a fundamental factor in the stability of underground structures it has not yet got the attention it deserves. In another paper to this Congress (Tan 1983) it is emphasized that dilatant volume increase in rock masses plays a crucial role in the stability of potentially swelling rocks. Rock dilatancy has gained worldwide attention since it has been observed that the earthcrust dilates prior to earthquakes; direct evidence of dilatancy have been observed in the San Andreas Fault near Parkfield (Cherry and Savage1972); a very clear evidence have been reported from the bulging of the earthcrust prior to the large Haycheng Earthquake (1975). This has been one of the crucial scientific materials for the prediction of this earthquake. The Haycheng earthquake was the first earthquake, which was predicted correctly in site, magnitude and time (Fig.7A,B,C). Indirect evidence for crustal dilatancy has been reported from the lowering of deep water wells situated in a radius of 100 km around the epicentre of the large Tangshan Earthquake (Tan, He1982); a further indirect indication is the steady decrease in earth resistivity of the upper layer of the crust (Fig.2a-b).· The changes in vp/vs ratio (velocities of normal wave to shear waves) which is considered to be directly related to dilatancy, have been frequently measured in our country (Feng et al 1976; Duan et al 1976). Such changes in this vp/vs ratios have been earlier reported by Semenov(1969), Aggarwal (l973). Some U.S. scientists belief that fluid inflow during dilatancy is a crucial factor leading to earthquakes as it is accompanied by the generation of water pressure (i.e. decrease of the effective normal stress, thus strength) and the lubrication of fissures. A physico rheological model for Earthquake fore runners has been recently suggested (Tan 1982); it is based on the fundamental assumption that the earthcrust is a rheological dilatant body traversed by a network of planes of easier glide, the seismic belts. Time dependent dilatancy is an important problem, which has not yet been explored extensively. It is generally believed that this is due to "creep" but it is not clearly specified what is understood under "creep". As it will be discussed next, creep is due to the continuous compatible straining of grains which is increasing with the time with decreasing rate. This is the case as far as the deviatoric stresses remain below an upper yield limit f3(f3 for shear and f = 3f3 for compression). As soon as this upper limit is exceeded, then non compatible anelastic deformations will occur leading to void and crack formation and opening of inborn cracks.

This content is only available via PDF.
You can access this article if you purchase or spend a download.