Visual Analytics for Reservoir Analogues

Abstract

Geoscientists often deal with incomplete reservoir data when making business-critical decisions. One of the most common strategies is to use available known information from similar reservoirs (analogues) to better estimate the unknown information. However, there is a lack of analytics and visual tools for oil & gas industry that facilitate the investigation of similarities and analogies. Furthermore, for experts, facts that support their decisions are paramount for their daily work; thus, systems that work with artificial intelligence (AI) must enable to recover evidence that helps to explain the results. In this work, we present a workflow that brings the knowledge of the geoscientist along with AI. It applies machine learning (ML) algorithms to go through extensive datasets of reservoir characterization, allowing and empowering experts to visually explore such datasets, retrieve analogs, estimate unknown parameters and thus make better-informed decisions.

To proof the power of the proposed workflow, we implemented a system that geo-located the reservoirs on a map enabling experts to inspect and explore visually the data giving them insights about it. RAVA (Reservoir Analogues Visual Analytics) is an interactive system that allows users to explore a reservoir database and estimate unknown information of a target reservoir by automatically identifying analogues. Our workflow enables the experts to enter with their knowledge in all critical steps and test what-if scenarios. The similarity function, which defines the set of analogues, can be determined according to the final goal. For instance, the set of similar reservoirs appropriated to gather information about EOR techniques probably is different from those suitable to study porosity curves; thus, the system allows for the expert to set the similarity function for each scenario. Additionally, we create a method to examine the influence of each parameter on the analogues set. This technique enables the users to understand the reservoir's similarity characteristics more deeply. Finally, the user may explore the analogues list by a set of filters, such as establishing a similarity threshold or a range of parameter values that the analogues must fall in. Users may see the parameter distribution of the filtered analogues, analyzing the distribution to aid in the decision-making.

AAPG Datapages/Search and Discovery Article #90323 ©2018 AAPG Annual Convention and Exhibition, Salt Lake City, Utah, May 20-23, 2018