Two dimensional direct current resistivity forward modelling finite element C++ source code

Finite Element source code is written in C++ for two dimensional direct current resistivity surface geophysics problem. Rayleigh-Ritz energy minimisation approach is followed for formulation of the problem. Three aspects of the problem are studied, viz. (i) Effect of matrix partitioning in reducing the storage and time for inversion of the sparse matrix. (ii) Effect of number of integration variables and their distribution for accurate computation of potentials (iii) effect of element size in implementation of Neumann's boundary condition in the vicinity of the current electrode. It is observed that matrix partitioning significantly reduces the computation time and space required for storage. Surface element in z direction should preferably be smaller for accurate implementation of Neumann's boundary condition. Number of integration variable should be chosen according to (a) the accuracy demanded by the problem (b) the distances till which match between the analytical and computed curve is required (within the error limits) and (c) the computational facility available. C++ language has been used because the program utilises an object oriented concept in storage and inversion of the sparse matrix. Finite Element formulation and important source code modules are presented in Appendix A and B. The model was tested showing the comparison of the variation of analytical and numerical potential values with distance from the source for different number of integration variables. Two electrode apparent resistivity profiles over an outcropping vertical dyke and an outcropping vertical contact are presented to show the nature of response and the level of accuracy achieved by working with 6264 three noded triangular elements on an AcerPower 6500 PC (Intel 233 MHz chip).