A Verıfıcatıon And Qualıty Control Method For Permeabılıty Dıstrıbutıon Data In Oıl And Gas Fıelds

Abstract

The purpose of this research is to test the applicability of the Benford’s First Digit law (FDL) as a data quality control method for permeability distribution data in oil and gas fields. The FDL involves the distribution of occurrence of first digits (from 1 to 9) in measurements emanating from natural processes. Distribution of permeability is an example of such a natural process. The FDL has been used successfully as a tool in the field of financial accounting for the detection of fraud and misrepresentative data thus giving financial auditing professionals a method of probing non-conformant FDL data sets. In this thesis, permeability distribution data from two major fields in the Norwegian continental shelf have been investigated in the light of the FDL to confirm if the data is truly from a natural geological process. In achieving this objective, samples of permeability distribution data from both fields were examined. Their first digit distributions and goodness-of –fit to the ideal Benford’s FDL distribution was evaluated using the chi-square statistic. In doing this, two hypotheses were considered. These are the Null Hypothesis and the Alternative Hypothesis. The Null Hypothesis was stated as; Ho: Permeability distribution in oilfields is a non- random natural geological occurrence which conforms to the Benford’s FDL. The Alternative Hypothesis was stated as; H1: Permeability distribution in oilfield is a non- random natural geological process which does not conform to Benford’s First-Digit Law. The Null hypothesis that their respective permeability distributions follow the Benford’s FDL distribution was clearly established and accepted based on the results of the statistical goodness-of-fit test. Although it may not be immediately concluded that non-compliant datasets are not representative of the field under investigation, a non-compliant data set is an invitation for the Petroleum Professional to ask important questions like why certain permeability range of values seem to distort this trend. Trend distortions could be as a result of geological misinterpretations, non- standard permeability measurement techniques, data transmission, storage and decoding errors or outright fictitious data entries. All of these possibilities have to be investigated for final data validation.