The test study area is the Miocene reservoir of Nam Con Son Basin, offshore Vietnam. In the study we used unsupervised learning to automatically cluster hydraulic flow units (HU) based on flow zone indicators (FZI) in a core plug dataset. Then we applied supervised learning to predict HU by combining core and well log data. We tested several machine learning algorithms. In the first phase, we derived hydraulic flow unit clustering of porosity and permeability of core data using unsupervised machine learning methods such as Ward’s, K mean, Self-Organize Map (SOM) and Fuzzy C mean (FCM). Then we applied supervised machine learning methods including Artificial Neural Networks (ANN), Support Vector Machines (SVM), Boosted Tree (BT) and Random Forest (RF). We combined both core and log data to predict HU logs for the full well section of the wells without core data. We used four wells with six logs (GR, DT, NPHI, LLD, LSS and RHOB) and 578 cores from the Miocene reservoir to train, validate and test the data. Our goal was to show that the correct combination of cores and well logs data would provide reservoir engineers with a tool for HU classification and estimation of permeability in a continuous geological profile. Our research showed that machine learning effectively boosts the prediction of permeability, reduces uncertainty in reservoir modeling, and improves project economics.

The geoelectrical resistivity method is one of the most commonly used geophysical methods. This method uses different electrodes configuration, electrode array, depending on the purpose and conditions of the field, each type of array has its advantages and disadvantages. Due to the development of data acquisition technology, it is common for geoelectrical instruments enable to record data arising from different electrode arrays with negligible real-time construction. However, current software’s only allows to process for each individual electrode array. Inverted models of different electrode array can be integrated to build a common earth model. However, due to the nature of the geophysical inversion is non-unique solutions, it means that there will be an infinite of models that can be suitable for a measurement in a certain noise level. This leads to the same measurement data in an area with different electrode array may produce different geoelectrical models making the dificulty for integration process. To solve this problem, we utilise the simultaneous joint inversion algorithm of data sets arising from multiple electrode arrays. The test results on synthetic data show that this combination is better than the solution of each individual electrode array. The best result is a combination of pole - dipole (PD), dipole - pole (DP) and dipole - dipole (DD).

No abstract

No abstract