DEVELOPMENT OF CLUSTERING-BASED 2D EARTH RESISTIVITY IMAGING SYSTEM

Abstract

The advent of microcomputers and micro electronics in equipment manufacture have made Geophysical data acquisition, analysis and interpretation more advanced, especially in the area of data interpretation. The development of sophisticated hardware has made data acquisition to be faster and more efficient than ever before. This calls for rapid development in data interpretation. However, the challenge in data analysis and interpretation is beyond manual techniques which had hitherto been acceptable. Data acquisition in 2D becomes popular, economical and efficient, hence, the development of 2D data interpretation techniques. Many researchers have developed complex models of subsurface over the years with attendant computational cost and complexities resulting in the problem of non uniqueness in Geophysical data interpretation. Computer software applications were developed to handle Geophysical data interpretation in 1D, 2D and recently 3D and more. In this research, problems of computational cost and model non uniqueness have been addressed. A novel 2D interpretation technique based on clustering has been formulated and application software based on the technique produced; Clustering-Based 2D Earth Resistivity Imaging System (CB2DERIS). The algorithm has been developed to image the earth’s electrical resistivity distribution. Clustering technique has been used to group 1D data according to a stratified earth model. Problem of non uniqueness of models have been addressed by using a methodology that imitates seismic-well-tie process in Seismic data interpretation. Observations are repeated at certain depths during both 1D and 2D data acquisitions, which are used to establish a relationship between the two sets of data using Regression analysis. Hence, 2D interpretation is obtained from 1D interpretation. Data have been collected in specific locations whose geology is well known and the developed software application has been used to interpret the acquired Geophysical data. Results obtained have been compared with that obtained by the commercially available software application; DIPROfWIN. Interpretations were same to an acceptable degree. They both image same 2D geologic structure based on electrical resistivity distribution. CB2DERIS spent 3.56s to process same data that DIPROfWIN used 4.20s in FE mode and 1.30s in FD mode to process. However, CB2DERIS does not require many iterations and therefore, faster. The two applications clearly identify same number of layers representing the true geology of the locations. The CB2DERIS definitely produces clearer image of the subsurface. The new scheme developed also provides a way to mitigate the problem of non uniqueness in resistivity modeling.