Alireza Shahkarami

Artificial Intelligence (AI) Assisted History Matching

History matching is the process of adjusting uncertain reservoir parameters until an acceptable match with the measured production data is obtained. Complexity and insufficient knowledge of reservoir characteristics makes this process timeconsuming with high computational cost. In the recent years, many efforts mainly referred as assisted history matching have attempted to make this process faster; nevertheless, the degree of success of these techniques continues to be a subject for debate. This study aims to examine the application of a unique pattern recognition technology to improve the time and efforts required for completing a successful history matching project. The pattern recognition capabilities of Artificial Intelligence and Data Mining (AI&DM) are used to develop Surrogate Reservoir Model (SRM) for utilization as the engine to drive the history matching process. SRM is an intelligent prototype of the full-field reservoir simulation model that runs in fractions of a second. SRM is built using a handful of geological realizations. In this study, a synthetic reservoir model of a heterogeneous oilfield with 24 production wells and 30 years of production history was used as the ground truth (the subject and the goal of the history match). An SRM was created to accurately represent this reservoir model. The history matching process for this field was performed using the SRM and by tuning static data (Permeability). The result of this study demonstrates the capabilities of SRM for fast track and accurate reproduction of the numerical model results. Speed and accuracy make SRM a fast and effective tool for assisted history matching.

Using Big Data and Smart Field Technology for Detecting Leakage in a CO2 Storage Project

Smart Fields are distinguished with two characteristics: Big Data and Real-Time access. A small smart field with only ten wells can generate more than a billion data points every year. This data is streamed in real-time while being stored in data historians. The challenge for operating a smart field is to be able to process this massive amount of information in ways that can be useful in reservoir management and relevant operations. In this paper we introduce a technology for processing and utilization of data generated in a smart field. The project is CO2 storage in Citronelle Dome, Alabama and the objective is to use smart field technology to build a real-time, long-term CO2 Intelligent Leakage Detection System (ILDS). The main concern for geologic CO2 sequestration is the capability of the underground carbon dioxide storage to confine and sustain the injected CO2 for very long time. If a leakage from a geological sink occurs, it is crucial to find the approximate location and amount of the leak in order to take on proper remediation activity. To help accommodate CO2 leak detection, two PDGs (Permanent Down-hole Gauges) have been installed in the observation well. A reservoir simulation model for CO2 sequestration in the Citronelle Dome was developed. Multiple scenarios of CO2 leakage is modeled and high frequency pressure data from the PDGs in the observation well are collected. The complexity of the pressure signal behaviors and the reservoir model makes the use of inverse solution of analytical models impractical. Therefore an alternate solution is developed for the ILDS, based on Machine Learning. High Frequency Data Streams are processed in real-time, summarized (by Descriptive Statistics) and transformed into a format appropriate for pattern recognition technology. Successful detection of location and amount of CO2 leaking from the reservoir using the real-time data streams demonstrates the power of pattern recognition and machine learning as a reservoir and operational management tool for smart fields.

Assisted history matching using pattern recognition technology

Assisted History Matching Using Pattern Recognition Technology

Application of Artificial Intelligence and Data Mining Techniques for Fast Track Modeling of Geologic Sequestration of Carbon Dioxide - Case Study of SACROC Unit

The complexities involved in the available reservoir simulation model for the geologic CO2 sequestration study at SACROC Unit, lead to a high computational cost nearly impractical for different types of reservoir studies. In this study, as an alternative to the full-field reservoir simulation model, we develop and examine the application of a new technology (Surrogate Reservoir Model – SRM) for fast track modeling of pressure and phase saturation distributions in the injection and post-injection time periods. The SRM is developed based on a few realizations of full-field reservoir simulation model, and it is able to generate the outputs in a very short time with reasonable accuracy. The SRM is developed using the pattern recognition capabilities of Artificial Intelligence and Data Mining (AI&DM) techniques. The SRM is trained based on the provided examples of the system and then verified using additional samples. The intricacy of simulating multiphase flow, having large number of time steps required to study injection and post-injection periods of CO2 sequestration, highly heterogeneous reservoir, and a large number of wells have led to a highly complicated reservoir simulation model for SACROC Unit. A single realization of this model takes hours to run. An in-depth understanding of CO2 sequestration process requires multiple realizations of the reservoir model. Consequently, using a conventional numerical simulator makes the computational cost of the analysis too high to be practical. On the other hand, the developed SRM for this case study runs in a matter of seconds. The comparison between the results of SRM and simulator, during training and verification steps of SRM development, demonstrates the ability of SRM in mimicking the behavior of numerical simulation model. The results of this study are intended to prove the potential of AI&DM based reservoir models, like SRM, to ease the obstacles involved in the conventional CO2 sequestration modeling.