Feature Extraction and Pattern Recognition in Time-Lapse Pressure Transient Responses

Abstract

Monitoring of modern wells equipped with flowmeters and permanent downhole gauges provides large datasets of pressure, temperature, and flow rate measurements. On a limited scale, selected events from these datasets are traditionally used in well performance analysis and monitoring, reservoir simulation, and many other tasks employing physics-based models. New methods that combine model- and data-driven approaches for full-scale analysis of these large datasets have recently attracted interest, suggesting new metrics for knowledge extraction. Most recent studies have used a combination of Pressure Transient Analysis (PTA), which is a key reservoir engineering tool, with hybrid methods, such as PTA metrics and AI-powered models, for the monitoring data interpretation. Although these approaches have great potential, there is a concern about the reliability of applying such methods, as most of them are dependent on proper data processing, interpretation, and feature engineering, calling for effective knowledge extraction and data engineering with regard to large datasets. This paper introduces a novel, feature extraction technique in combination with a pattern recognition method for analysis of time-lapse pressure transient responses and their Bourdet derivatives from large well monitoring datasets. This robust and scalable approach is designed for fast and autonomous extraction of features commonly used in PTA by field engineers. Furthermore, the developed methodology provides with identification of underlying patterns governed by sequences of these PTA-features in the transient responses. This information is crucial for well and reservoir performance analysis and monitoring and can be integrated with hybrid methods for knowledge extraction such as the PTA-metrics introduced previously. The article illustrates further such integration through field examples, demonstrating how it enhances the reliability of the PTA-metrics applications. The developed methodology, combining all the methods described above, serves as a new tool for knowledge extraction from big well monitoring datasets, available in the companies operating in the oil and gas industry, as well as the emerging industries such carbon capture and storage and geothermal energy production. The article concludes with a discussion of the main advantages and limitations of the suggested feature extraction and pattern recognition methods. Besides the combination of the methods considered in this article, the feature extraction and pattern recognition methods introduced may also be integrated with conventional physics-based approaches widely used in the industry for well data interpretation and reservoir simulations, improving their performance and efficiency for big data sets.

Keywords: Pressure Transient Analysis, Well Testing and Monitoring, Feature Extraction, Pattern Recognition, Time-Series Analysis