Un système d'aide à la gestion des ressources en eaux souterraines. Application au site de Bordeaux.
Résumé
Integration of groundwater management into a service that produces and distributes water, helps in determining a strategy for optimum abstraction that is in harmony with the constraints and objectives related to the resource. The constraints in this case are mainly defined by technical and conservation parameters, the first being related to the movements of the water levels in a pumped borehole, e.g. the fact that screens should not fall dry. Conservation parameters Concern points such as the conservation of the confined character of an aquifer, or the protection against salt-water intrusion. The main objective is to minimize the lowering of the water table from one year to the next, thus ensuring the medium- and long-term sustainability of the groundwater resource. This type of management first of all requires a sufficiently detailed modeling of flow within the aquifer. This will, in turn, enable the definition of each pumping weIl and the determination of ifs associated constraints. Methodology First, a hydrodynamic finite-difference model is created and calibrated on the aquifer of interest. This model will help in predicting the influence of an exploitation scheme on the evolution of potentiometric levels. The drawback is that, especial/y in transient state, this model requires a large amount of computer time. This means that such models are mainly used for the one-time calculation of coefficients of the influence exerted by ail existing pumping wells on all sensitive potentiometric levels. The coefficient of influence of a well on an observation point, is the response, at this point and in transient state, of a unit increase in the pumping flow rate. In steady state, this coefficient is the differential of potentiometric variation divided by the variation in flow rate. Assuming that the aquifer can be represented as an invariant linear system and applying the superposition theorem, it is possible to calculate the influence of any pumping scenario through the simple convolution of the coefficients of influence. In reality, the system is not perfectly linear, as the coefficients of influence depend upon transmissivity values that, in an unconfined aquifer, are partial/y dependent upon the potentiometric levels, which in turn are related to the abstraction, i.e. the influences. However, if the variations in abstraction are small, the linear assumption generally is entirely acceptable. BRGM researchers have developed the CAPUCINE software (Thiery, 1993a), which is based on an optimization under constraint and works with coefficients of influence that are derived from a hydrodynamic model. This software enables the optimization of a parameter function, e.g. the sum of abstracted flow rates or exploitation cost, under various constraints. The last are expressed in terms of heads or flow rates to be respected, either locally or in a geographic region. The original aspect of this optimization software is that is directly integrates the diameters of boreholes and head losses coefficients in boreholes, which are indispensable data. As the influence of borehole diameter and head loss is non-linear, the software calculates a series of successive linear optimizations in an iterative manner, to arrive at an optimum scenario.
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...