document preview

Abstract

Developing an automated framework for real-time optimization (RTO) of the Steam Assisted Gravity Drainage (SAGD) process has significant potential because of the large number of parameters that must be monitored at a high frequency. However, the industry has not yet adopted a standard RTO framework for SAGD because of the intrinsic complexity of the process, the large number of parameters that must be monitored, harsh operating conditions, the lack of integration between various data acquisition systems, and the complex criteria required to optimize SAGD performance.

In this paper, a real-time monitoring workflow for SAGD is proposed that streams field data from multiple sources, including fiber optic distributed temperature sensing (DTS) directly into an engineering desktop application that has artificial intelligence (AI) and data mining capabilities. This system is used to derive advanced criteria to make decisions in a timely manner to improve the performance of the SAGD process.

It also demonstrates how subcool calculations can be effectively performed along the length of the horizontal well in real time and how the results are used to improve SAGD operation. Observations are compared “live” against simulated predictions from a multisegmented wellbore model that is fully coupled to a thermal/compositional reservoir simulator.

Introduction

This work explores the technologies and workflows that enable real-time optimization (RTO) of the SAGD process. In order to better understand the drivers and motivation for adapting RTO processes toSAGD, it is appropriate to first introduce the basic concepts and terminology of RTO for hydrocarbon systems in general.

Real-Time Production Optimization

The term “real-time” can be used to refer to as an interaction with any generic event happening in a short time scale.  In our industry, depending on the objective function to optimize it may be seconds, minutes, or even days.  A generic diagram of a closed-loop RTO system is presented on Firgure 1.  This closed loop begins with measuring the operational changes (measure), continuous well and reservoir model updating (analyze), detecting underperforming conditions and defining the optimal production strategy (detect/optimize), and changing operating parameters for surface equipment and wells (control).  This was developed based on the series of papers published by the Society of Petroleum Engineers (SPE) Real-Time Optimization Technical Interest Group (TIG) covering the technical and business-relevant aspects of RTO systems in the oil and gas industry (Ref. 1, 2, 3).

Many different closed-loops RTO classifications for a hydrocarbon producing system can be found in the literature. One typical example is provided on Figure 2 (Ref. 4).

The first level, surveillance, is about installing the right sensors, systems and instruments to collect sufficient real or near real-time data to support decisions. One important component of this level is the communications infrastructure needed to get the data from the sensors back to the key decision makers.

Once the information is received at the engineer’s desktop, a variety of engineering tools can be used to process the data into meaningful information by integrating it with existing models and providing real-time analysis. Use of intelligent tools to manage, visualize and analyze data is another key aspect of implementation, and is the second level: analysis. This data-to-information step can deliver significant benefits across a range of processes. At this level, we begin to use the information developed to improve work-processes and decision making.