Application, Comparison, and Optimisation of Geophysical Techniques in Detection of Groundwater

           

 

In this project, the problem of detection of groundwater contamination was investigated by using a combination of different electrical resistivity configurations. New theory has been developed and tested on a field area in which groundwater in the subsurface is contaminated. Theoretical computations of the electrical response for real geological models are often lengthy. Thus, for these computaions, a fast machine with large amounts of memory is needed. The results of the theoretical computations are used for interpretation of the electrical resistivity data obtained from the field, resulting in detection of subsurface groundwater contamination zones.


   

Principal Investigator

Paul G. Wilkes

Department of Exploration Geophysics

Curtin University of Technology

Co-Investigators

Abolghasem Kamkar-Rouhani

Department of Exploration Geophysics

Curtin University of Technology

Projects

g99 - PC

     
         
           

What are the results to date and the future of the work?

In 1998, we implemented in our computer codes two computational methods using integral equations which enhanced the accuracy of the results. These two methods use the biconjugate gradient method and new digital linear filters for computation of Hankel transfoms. The codes were run on PC to obtain the electrical responses of different 3-D subsurface models indicating different kinds of subsurface contamination zones. Relatively good results were obtained for some kinds of models. However, we had difficulty in obtaining satisfactory results for models indicating very shallow contamination. The difficulty was mainly due to dividing the 3-D body or bodies into a great number of elementary cells to get more accurate results. As a result, the excution of the codes with large two-dimensional arrays needed a considerable amount of time, large stack frame sizes, and thus, we always used the bigmem queue to submit the jobs for running in PC. Although theoretically the accuracy of results increased with increasing the number of cells, the accumulation and also probably propagation of (round-off) errors was a reason for not getting good results for such models.

What computational techniques are used?

The method of integral equations was used in our Fortran codes to obtain good accuracy in calculation of electric potentials, which are converted to apparent resistivity by a simple transformation. The anomalous subsurface body indicating the contamination zone, was divided

           
- Appendix B

 
           

       

into a number of cells. In this project, between 4000 and 10000 cells were used to study different models of contamination. For computation of electric charge density on the surface of each cell, which was required to obtain the electric potential, the method of integral equations, moment method with pulse basis function and point collocation was used. In this method, distribution of the electric charge density on the surface of each cell was assumed to be constant. The computations were carried out for 3-D models, and large and small arrays were used for defining different computational elements in the integral equations method.

Publications

A. Kamkar-Rouhani, Development and Application of Processing Techniques for Signal Enhancement Using Multisystem Resistivity Measurements, PhD thesis, Department of Exploration Geophysics, Curtin University of Technology, 1998.

A. Kamkar-Rouhani, U. C. Das, Resistivity Responses of Mineral Deposits Under a Transitional Overburden, 60th Annual International Meeting, European Association of Geosciences and Engineers, Extended Abstracts, 1998, P148.

       
Appendix B -