- A field and computer model investigation was conducted of the ground-water flow system at the radioactive waste management area of the Savannah River Plant in South Carolina. Numerical ground-water flow simulations using the trial-and-error and automatic parameter estimation approaches produced estimations of the distribution of transmissive characteristics of the water-bearing formations in the study area. This distribution was also estimated by geostatistical analysis of its spatial structure using the results of aquifer tests. The distributions estimated by the three approaches were compared and a qualitative calibration methodology was developed that enhances the integration of field-measured data into the numerical modeling process., Numerical modeling by the trial-and-error and automatic history matching approaches produced estimates for the distribution of transmissive characteristics that were comparable between the methods. Variability in the conductivity distribution in the water-bearing units underlying the study area is high; kriging estimates of the conductivity distribution in each unit demonstrated this variability., The higher conductivities numerically estimated for portions of the water-bearing units are attributed to the presence of zones of higher hydraulic conductivity which control ground-water flow. Locating these zones by drilling wells and performing aquifer tests would require considerable numbers of wells and large financial resources. The numerical modeling approach identifies these areas because of its integration of all aspects of the flow system., and The comparison of the three approaches to estimating the distribution of transmissive characteristics suggests a feedback relationship which reduces uncertainties in modeling by incorporating both the numerical estimation method and the data base of measurements. Using the numerical simulation, zones of characteristic conductivity can be defined. Comparing these results to the existing data base evaluates the adequacy of the data base. A data base demonstrated to be inadequate can be improved, if desired, with more field work. The conductivity distribution from an adequate data base can be incorporated into the numerical simulation to constrain estimates for one or more zones. Then, either more zones may be estimated, or the number of zones to be estimated may be reduced. In either case, the progress of the model calibration is enhanced.
- Dissertation Note:
- Ph.D. The Pennsylvania State University 1987.
- Source: Dissertation Abstracts International, Volume: 48-10, Section: B, page: 2903.
View MARC record | catkey: 13613234