Elaboration d'une methodologie de detection et de caracterisation des courts-circuits hydrauliques causes par des defauts de scellement de puits

An experimental and theoretical approach has been developed for detecting and quantifying hydraulic short-circuits and cross-contamination between two aquifers separated by an aquitard and intersected by a borehole. Such leaks can be caused, for example, by improperly sealed pumping or monitoring wells and can have serious economic and health consequences if the underlying aquifer is being pumped for water supply. The research was undertaken within the global perspective of sustainable development and as an application of water resources protection.; The research project consisted in developing and testing a technical approach and associated instrumentation for detection of short-circuits, and a mathematical approach to characterize and quantify the fluid and mass leakage rates. The methodology was successfully tested by numerical modelling, small-scale laboratory experiments, then finally at a large-scale field research site.; The method consists in executing a series of successive, variable-rate pumping tests in the lower aquifer and measuring the leakage rates associated with each of the pumping steps. The leakage rate is estimated by a non reactive tracer test. At each pumping step, the tracer is injected under constant concentration from a piezometer installed in the upper aquifer. If a seal defect exists, the tracer will leak past the seal defect and will be recovered from the lower aquifer. The theoretical equations presented here describe the relationship between the leakage rate, the initial concentration of the injected tracer, the pumping rate and the recovered concentration. Hydraulic and geometrical properties of the seal defect can then be determined.; This newly developed methodology has several advantages. It is relatively simple and inexpensive to apply in the field, for wells to be newly dug as well as for existing wells. The piezometer, which also functions as the injection port, can be installed within the same borehole as the well (in the case of a new well), or drilled separately (for an existing well). The pumping well itself is used to monitor the tracer breakthrough.; One limitation has been identified in regards to the applicability of this method. It concerns the contrast of hydraulic conductivity between the aquitard and the defective seal. It has been numerically shown that in order for the method to function, the hydraulic conductivity of the aquitard must be inferior to the hydraulic conductivity of the defective seal by at least two orders of magnitude. The theory upon which the method is based supposes that the aquitard leakage rate is negligible compared to that of the defective seal. In the opposite case, the method will not provide an accurate prediction of the leakage rate through the defective seal.; Although some assumptions must be made, the method works for a wide range of aquifer/aquitard properties and leakage rates. However it must remembered that the properties of a seal can change with time, which can affect the degree of cross-contamination. Thus a suspect well should be re-tested fairly often. (Abstract shortened by UMI.)...