FLOAT-ENCODED GENETIC ALGORITHM USED FOR THE INVERSION PROCESSING OF WELL-LOGGING DATA

Szabó, Norbert Péter and Dobróka, Mihály (2013) FLOAT-ENCODED GENETIC ALGORITHM USED FOR THE INVERSION PROCESSING OF WELL-LOGGING DATA. In: Global optimization - Theory, Developments and Applications. Nova Science Publishers Inc., New York, pp. 79-104.

Preview

Text edited_Chapter.ID_12387_6x9_rev.pdf Download (1MB) | Preview

Abstract

In this chapter a Float-Encoded Genetic Algorithm is presented for solving the well-logging inverse problem. The aim of the global inversion of well-logging data is to provide a robust and reliable estimate of petrophysical properties of geological structures such as porosity, water saturation, shale volume and mineral content. There are two possible ways to solve the interpretation problem. The first is a conventional inversion scheme, which estimates the unknowns to different depths separately. In the forward modeling phase of the local inversion procedure the theoretical well-logging data are calculated by using locally defined probe response functions, which are then fitted to real data in order to estimate model parameters only to one depth. This procedure leads to a marginally over-determined inverse problem, which results in relatively poor parameter estimates. A further disadvantage of the above technique is that some crucial quantities such as the thickness of layered geological formations cannot be extracted by inversion, because it does not appear explicitly in local response equations. A new inversion methodology introduced by the authors gives much more freedom in choosing the inversion parameters. The so-called interval inversion method inverts all data measured from a greater depth interval in a joint inversion process. By a series expansion-based discretization of the petrophysical model a highly over-determined inverse problem can be formulated, which enables to estimate the petrophysical parameters including new unknowns such as zone parameters and layer thicknesses more accurately compared to local inversion methods. The authors give further references for several applications of the global inversion method. In this chapter, a synthetic and two field examples are presented to demonstrate the application of the Genetic Algorithm-based inversion method. It is shown that the combination of the new inversion strategy and global optimization tools forms a highly effective and adaptive algorithm for earth scientists who are interested in a more reliable calculation of the reserves of hydrocarbons and other mineral resources.

Actions (login required)

Edit Item