Event Sequence Interpretation of Structural Geomodels - A Knowledge-Based Approach for Extracting Tectonic Sequences.

Xianglin Zhan,Cai Lu,Guangmin Hu

DOI: https://doi.org/10.1007/978-3-030-39951-1_19

2020-01-01

Abstract:The tasks of obtaining past event occurrences and their temporal order information are important parts of the cognition of the external world. We call this kind of tasks Event Sequence Interpretations (ESI). In this work, we focus in the ESI in structural geomodels and propose a knowledge-based approach for extracting tectonic sequences, which is crucial for the cognition of structural geomodels. As a cognitive task, tectonic sequence interpretation has not been highly automated due to the need to use a large amount of expert knowledge for recognition and reasoning. Meanwhile, artificial ESI may introduce cognitive biases that ultimately lead to subjective uncertainty in the results, which affects the credibility of the interpretations and increases risks in oil and gas production. One potential solution is making personal knowledge better available for computers so that computers can also do ESI. Therefore, we proposed a meta-model for formally representing expert knowledge. The instance of the knowledge representation (KR) meta-model is called an Event Pattern (EP), which describes the associations between event occurrences and geometric features in the models. Moreover, we proposed a new pattern matching model called Joint Prototype Model (JPM) to find evidences of event occurrences from the raw geological data. The temporal relations of the events can be extracted according to the spatial topology of the geological objects. Our approach can also be extended from structural geomodels to other spatial geometric models. We show the effectiveness of the approach by an application to a real structural geomodel dataset.