Clustering Analysis and Flow Zone Indicator for Electrofacies Characterization in the Upper Shale Member in Luhais Oil Field, Southern Iraq

Abstract

Electrofacies identification is a crucial procedure in reservoir characterization especially in the lack of lithofacies measurements from core analysis. Electrofacies classification is essential to improve permeability-porosity relationships in non-cored intervals. Flow Zone Indicator (FZI) is a conventional procedure for rock types classification whereas Clustering Analysis has been recently used as unsupervised machine learning technique to group a set of data objects into clusters with no predefined classes. In this paper, clustering analysis and flow zone indicator were adopted for the electrofacies characterization on a dataset obtained from incorporate of conventional core analysis and CPI logs (Effective Porosity, Water saturation and Shale volume) of three wells in the upper shale member/Zubair formation in Luhais oil field southern Iraq. The FZI attains reservoir quality evaluation through hydraulic flow unit and kozeny-carman equation, which correlates between permeability, porosity, pore throat, pore geometry and tortuosity factor. The reservoir intervals of similar or convergent range of FZI values belong to a single hydraulic flow unit and reflect the geological attributes of texture and mineralogy therefore they represent specific rock type. In clustering analysis, two approaches were adopted for electrofacies identification: k-mean and Ward's Hierarchical clustering method. These three techniques were implemented by the use of R language, which is a powerful statistical programming tool. The prepared R codes in this research can be utilized in any reservoir characterization process to define the different electrofacies The results of this research showed two implemented techniques have identified three key electrofacies with different levels of accuracy and approaches while four rock types have been predicted from the third technique. The k-mean clustering analysis was the most accurate method where each predicted electrofacies reflects the same vertical distribution of the lithofacies in the reservoir. On the other hand, the Ward's Hierarchical electrofacies prediction represents specific petrophysical properties with minor differences from real lithofacies distribution in the reservoir. Finally, electrofacies identification through the FZI technique has similar ranges of FZI values with specific porosity and permeability values. FZI technique showed difficulty in defining lithofacies boundaries within the reservoir and predicted four rock types.

Hydraulic Flow Unit Classification and Prediction Using Machine Learning Techniques: A Case Study from the Nam Con Son Basin, Offshore Vietnam

“…The data point similarity is measured by summing the square of the distances between them. The hierarchical clustering output is a dendrogram that shows the cluster hierarchy [35].…”

“…K-means clustering is a common unsupervised learning algorithm, which tries to group a data sets into K clusters so that items in the same cluster are close together and items in different clusters are more dispersed [35]. Hence, the K-means clustering process shrinks the distance between points in the same cluster and expands the distance between points when the clusters are different.…”

K-Mean Clustering Analysis and Logistic Boosting Regression for Rock Facies Characterization and Classification in Zubair Reservoir in Luhais Oil Field, Southern Iraq

“…Prior data partitioning, the number of clusters must be predicted and utilized as input for the K-Mean algorithm (Abbas & Al Lawe, 2019). Several statistical approaches can be used for determining clusters number.…”

“…Therefore, a variety of Advanced Data-Driven approaches have been proposed to estimate electrofacies or lithofacies distribution in non-cored wells. It can be summarized into two types: 1) The unsupervised machine learning techniques to partition well log response into different clusters and estimate electrofacies such as Model-Based Clustering (Woan et al, 2012), k-Mean Clustering (Abbas & Al Lawe, 2019) and Ward's hierarchical clustering (Al-Jafar & Al-Jaberi, 2019). 2) The supervised machine learning algorithms are used for lithofacies classification given well-logging and observed lithofacies from cored wells to predict their distribution in other non-cored wells.…”