WAVELET ANALYSIS OF WELL LOG DATA

Abstract

Well log analysis is conventionally carried out by visual inspection to identify different lithology and formation interfaces. However, in this study a combination of continuous wavelet transform and wavelet entropy is used to identify a zone containing gas hydrates. The application of continuous wavelet transform on input geophysical well logs result in wavelet coefficients. The wavelet entropy is computed using these wavelet coefficients. This whole process requires optimum choice of wavelet and scale range to carry out continuous wavelet transform and suitable entropy criteria to obtain a clear signal for identification of different lithology. A set of synthetic well log were created to analyze the validity of above developed procedure and to identify which wavelet, scale range and entropy criteria are suitable for lithology identification through case studies. The results of the synthetic analysis suggest that Haar wavelet at lower scales and Norm entropy criteria are most suitable among a range of choice for wavelet, scale range and entropy criteria for lithology identification. A set of field well logs (GR, Density, Resistivity, Porosity and Velocity) from Cascadia basin were utilized to test the performance of the developed procedure. Hole U1327A from IODP Expedition 311 has been inferred to show a 18m gas hydrate zone between 120-138 mbsf. The analysis of these field well logs showed similar results. The use of Haar wavelet at low scales for continuous wavelet transform and the use of norm entropy criteria on wavelet coefficients identify the gas hydrates zone in hole U132A in a much better way than other available choices of wavelet, scale range and entropy criteria.