Uniqueness and resolution in resistivity interpretation /
| Main Author: | |
|---|---|
| Other Authors: | , , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1991.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | The problems with model uniqueness and resolution in one-dimensional resistivity interpretation are studied. Seven theoretical examples, representing both Schlumberger and Wenner soundings, are evaluated. The examples include three- and four-layer models and can be considered to be unique, nonunique, or exhibit poor resolution. Gaussian noise is added to the data to simulate actual field data. A three-step approach is taken. First, the noniterative direct method of interpretation is compared with the iterative indirect inversion method to evaluate which is better for interpreting resistivity data. The method of Pekeris (1940) is used for the direct interpretation and a nonlinear least-squares method for the indirect interpretation. It is found that for noisy data, the indirect inversion method is more reliable in finding a solution and gives better results. The second step involves inverting the data of each example using four nonlinear least-squares inversion schemes. Three schemes hold the layer thickness fixed, with only the resistivity varying, and the other allows both the layer resistivities and thicknesses to vary. The fixed-layer thickness schemes are uniform thickness, geometrical progression of thickness, and logarithmic progression of thickness. The variable parameter scheme is best at simultaneously minimizing the data and model rms errors for models that can be uniquely determined through the inversion process. A statistical parameter, the F-ratio, is used to determine the optimum number of layers in the model. The last step is to estimate the range of variation in the layer parameters in order to determine if a solution can be considered unique. The use of error ellipses, standard deviation, and forward modeling are investigated for this purpose. The correlation between parameters of different layers is not considered; this investigation is only concerned with the correlation between parameters of a given layer. Forward modeling gives the best results and is easy to incorporate at the end of an inversion routine. Although the problems with nonuniqueness and poor resolution will always be present in resistivity interpretation, there are techniques available which can aid the interpreter in determining the optimum layer model and warn he/she when other information is needed to interpret the data properly. |
|---|---|
| Item Description: | Typescript (photocopy). Vita. "Major subject: Geophysics." |
| Physical Description: | xiii, 221 leaves : illustrations ; 29 cm |
| Bibliography: | Includes bibliographical references. |