About deterioration in the abutment rocks of some large dams in Northern Italy

Experimental data concerning to the velocities of longitudinal elastic waves obtained in abutment rocks of some large dams in North Italy are considered; also the time variation of values is studied. In the different cases, crack density has been then estimated, on the basis of O'Connell and Budiansky' paper, regarding the seismic wave velocities in dry cracked bodies. Study carried out a decrease in the values of velocity and a consequent increase of crack density caused, at first, by the breaking of explosive, and then by the microseismicity due to the dynamic of structure in abutment rocks. Phenomenology is different in the various cases; however, the effect is generally strong.


INTRODUCTION
Starting 1949, various large dams in Northern Italy along with sustaining rocks underwent geodynamic controls. Some of these results have already been published. Herewith some data are given which for the most part are inedited and adjourned up to 1973 (after that year it has been impossible to continue these specific geophysical studies) concerning the velocity of longitudinal elastic waves within the abutment rocks of certain dams; further on their elaboration results are discussed.

DAMS UNDER OBSERVATION
In correspondence to the dam crowns -on rock systems building up their downward side -tape recordings were made of waves produced by blastings on the valley bottom. These experiments were carried out over the years, during the same seansons and along the same portion of rock. Every time precise sights made it possible to reproduce exactly sources and stations. The distances between sources and stations were determined by means of precision topographic levels; the uncertainty does not exceed 1%. A scheme of the stations and the sources for a dam is shown in fig. 1, and explored abutment rock is pointed out too. During the first years, an Askania vibrometer with optical recordings and incorporated time signal was used. Later on, an Hellige vibrometer with three indipendent components was used and recordings were made on heat-sensitive paper. In all cases the origin time was obtained by an electromagnetic impulse via cable. Recordings were made at an average speed of 10 cm/s (tape speed); thus it was possible to evaluate the velocity of the longitudinal waves with a uncertainty of about 2%.
The dams considered in this report, are those whose data had been gathered methodically enough to permit the elaboration of this information. Table I refers to the main characteristics of the dams under study; the relative longitudinal sections are in diagrams in figure 2, with the indication of the lithology of the abutment rocks.

EXPERIMENTAL RESULTS AND ELABORATION OF THE DATA
The values of the velocity on the abutment rocks which were obtained in the course of the following years are different: for the most part it was possible to verify a specific trend toward a reduction of the same. This can be explained as a consequence of the force of the explosive used in the construction phase and later on, as a consequence of a microseimicity produced in the rock, whereby the dam is anchored to, by the constant movements that the construction undergoes when it is subject to variations of the reservoir water load, and by variations of the outside temperature due to daily insolation and change of seasons. Lastly, the velocity decrease which has been observed is to be attributed to the increased crack density in the rock.
The study on the connections between these factors has been carried out by many authors. During the last few years, R.J. O'Connell and B. Budiansky have statistically obtained the following equations (values with index « o » refer to rocks which are not cracked) in the hypothesis of dry circular cracks: 3, being Poissons's report, e the density of the cracks, K the bulk coefficient and v p the velocity of the longitudinal waves. From previous formulae it is possible to obtain E as a function of (if a = 0.25

Vajont dam
As it is known, on October 9, 1963, about one year after the dam entered into operation a big landslide broke away from Mount Toe, on the left side of the dam causing tragic consequences.
Much has been written on the causes of this event; in this report we are going to limit ourselves to illustrate the pertinent results to the velocity's values and to the consequent crack density within the abutment rocks.
Geophysical studies in the narrow of the Vajont began in 1953, that is to say, three years before construction began. As  fig. 6 shows, following the initial period whereby values remain practically constant, the velocity in both abutment rocks drops abruptly in time with a consequent 35-40% increase of the crack density. This effect is more evident in the left and it is no doubt the result of the disruptive force of the explosive, of the structure's construction and of the preliminary operations of hollowing and stemming. The excessive speed of the phenomenon seems, at any rate to point out a certain tendency (referred as " latent microseismicity" by P. Caloi) of the rock to deteriorate whenever subject to perturbing causes, both natural or artificial.

Pieve di C ado re dam
First tests were made in 1949 after the construction was completed. As far as the left hand abutment rock is concerned, fig. 5a shows a gradual and modest decrease of the velocity values which has brought an increase of about 18% of the crack density, over the period of time during which these observations were made.
The condition of the right hand abutment rock is different. The trend of the obtained points ( fig. 5b) has permitted to point out three distinct phases: the first goes to about 1960, the second which ends around 1963-1964, the third being underway during the last study. During the first period, the about 18% increase of the crack density indicates the structure's tendency to reach a new equilibrium, first as a result of alterations caused by the ground removal and by dam's construction, and later, by micromovements of the structure.
The second phase, characterized by a sudden increase of the crack density which in two year's time has reached 45%, is probable cause by the increase dynamic of the dam following increased water supplies consequent to completion of the Piave-Boite-Mae-Vajont's system ( fig. 7) to which the dam belongs.
One should not be surprised for the lack of a similar difference of phases on the left, considering the considerable asymmetry of the structure ( fig. 2c) whereby its left abutment rock bears the burden of the efforts connected to the dynamic of the dam. Later, a similar phenomenon occurs again despite reinforcement's work although it is resulted considerably slower in the time, must be noted however that following the Vajont's incident, system's production was reduced to that of the years preceding 1960.

Ambiesta dam
The various geophysical studies were carried out after the dam entered into operation. Figure 3 shows in this case too, that the two abutment rocks took a different course; while velocity's values in the left were practically constant in time, the velocity's values in the right shwed a considerable decrease with a consequent 40% increase of the crack density. This can be possible attributed to the different consistency of the dolomite's limestone and not to asymmetry in the structure. Because of the scarse rigidity of the left rocks, the right abutment rock, as a hinge, bears the highest weight as to the dynamic of the dam. This asymmetry is also likely responsible for a transverse fracture of a lateral section which we noticed during the last geophysical testing.

Maina di Sauris dam
As in the previous case, there are no available experimental data obtained during the construction. In effect, the first tests were carried out the year the dam entered into operation. At any rate, the velocity's values which have been obtained during a 20 year's period show that both abutment rocks have a similar trend ( fig. 4): i.e. a slight decrease and a crack density's increase of about 15%. This phenomenon, quite unlike the preceeding ones, is justified by the symmetrical construction of the dam ( fig. 2b) and by the same high rigidity of the massive limestone of the two abutment rocks, whereby obtained velocity's values, since from the beginning, have been the highest among all those obtained in several parts of the Alps for rock formations of the same lithology.

CONCLUSIONS
In view of the above, cases pertinent to the velocity's variations within the abutment rocks of the dams are in effect rather different. In some cases there is a slow and gradual decay of the values during the time, in others there is a sudden change; at times the situation is similar in both abutment rocks, while in others the decrease is differentiated due to intrinsic or estrinsic causes to the structure.
Reinforcement operations do somewhat enable the rock to regain something of its elasticity without preserving them from further variations.
The theoretical handling of the crack density to which one attributes the decreasing of velocity's values has enables us (for dry circular cracks) to estimate the crack density in various cases. The maximum values obtained indicate an increase of 40-45% compared to initial value.
In conclusion, studies indicate that deterioration in the abutment rocks of some large dams is generally a strong phenomenon; therefore, a constant control must be taken place.
At any rate, it is possible (in some large dams) that this effect reduces the average life span of the plant remarkably. The average life span is generally calculated only on the basis of the silting up of the reservoir.