Study on Identification of Oil/Gas and Water Zones in Geological Logging Base on Support-Vector Machine

Abstract

Support Vector Machines (SVM) represents a new and very promising approach to pattern recognition based on small dataset. The approach is systematic and properly motivated by Statistical Learning Theory (SLT). Training involves separating the classes with a surface that maximizes the margin between them. An interesting property of this approach is that it is an approximate implementation of Structural Risk Minimization (SRM) induction principle, therefore, SVM is more generalized performance and accurate as compared to artificial neural network which embodies the Embodies Risk Minimization (ERM) principle.

We studied the theory and method of Support Vector Machines based the Statistical Learning Theory and proposed a pattern recognition method based Support Vector Machine to determine oil, gas and water zones in geological logging. The outline of the method is as follows: First, we researched the basic parameters of the gasometry logging and geochemistry logging, and induced some significant parameters, such as hydrocarbon moistness index, pyrolysis hydrocarbon equilibrium index, gasometry hydrocarbon equilibrium index, etc., which can help to distinguish the feature of reservoir; then we used the Support Vector Machine to study the relationship of those parameters, and set up the recognition mode and develop its program to determine of oil, gas and water zones. Application and analysis of the experimental results in Xinjiang oilfield proved that SVM could achieve greater accuracy than the BP neural network do, which also proved that application of SVM to identification of oil/gas and water zones in geological logging, even to the other theme in petroleum engineering, is reliable, adaptable, precise and easy to operate.