Analysis of reservoir heterogeneities and depositional environments: a new method

A new methodology is presented to quantify reservoir heterogeneities based on a Monte Carlo parameter estimation technique from sonic logs. The acoustic reservoir heterogeneities are then quantified as an indicator to differentiate depositional environments and reservoir facies for geological interpretation in this study. Fractal statistics provides us with a framework to model reservoir heterogeneities with different Hurst numbers, correlation lengths and fluctuation standard deviations from sonic logs. These fractal parameters are derived from the von Karman autocorrelation model and estimated by an improved methodology using a Monte Carlo parameter estimation technique. Unlike the regular estimation method, the Monte Carlo parameter estimation technique is more stable. The resulting Hurst numbers, correlation lengths and root mean square (RMS) heights from 20 sonic logs are mapped in a reservoir with sandstone–mudstone sequences over complex continental deposits in north-eastern China. The spatial distribution of estimated reservoir heterogeneity parameters are then correlated with depositional facies interpretation derived from seismic attributes and petrophysical properties based on prior geological knowledge. Numerical experiments show that the Monte Carlo parameter estimation technique is successful in recovering acoustic heterogeneity parameters from sonic logs. Results from qualitative correlational analysis of sonic logs from the complex continent deposits in north-eastern China show that these parameters can be key discriminants for depositional facies and environments and may be utilised as a constraint in reservoir characterisation. Maps of Hurst numbers and correlation lengths strongly correspond with reservoir facies distributions, whereas maps of RMS heights correlate significantly with fluvial depositional patterns. The results were generally uniform even when different depth ranges within the formation of interest were used as input for parameter estimation. Numerical values of reservoir heterogeneity parameters may not fully recover the geology of a given reservoir but their distribution in space is an important indication of the morphological features of depositional environments.