Proxy Model for EFDT Formtion Testing Sampling Basedon Numerical Simulation
Year: 2024
Type: article
Abstract: Using formation testing technology to obtain formation fluid samples is crucial for reservoir evaluation and study of fluid properties. With the development of formation testing technology, fluid sampling can be completed in complex environments. However, only using the downhole fluid analysis sensor data can not accurately quantify the pollution degree of oil reservoir under the condition of oil-based drilling fluid filtrate. Using the enhanced formation dynamic tester of China Oilfield Services Limited standard probes, numerous reservoir numerical simulation cases of downhole fluid sampling are generated, and big data are generated. Various machine learning methods are used to train the generated data with four formation testing sampling proxy models, and finally the model trained based on artificial neural networks is selected as the formation testing sampling proxy model. By changing formation and fluid properties, such as formation anisotropy, fluid viscosity, pumping flow rate and invasion depth, and analyzing their impact on contamination levels, the prediction of contamination levels during the EFDT standard probe sampling process is achieved. The practical application of the formation testing sampling proxy model shows that, the pump fluid density fitted based on the formation testing sampling proxy model has a high degree of agreement with the actual fluid density measured by EFDT, verifying the accuracy and practicality of the sampling proxy model. By inverting the data of the pump cleaning process, the variation of the contamination levels of the pump fluid over time is obtained, achieving real-time judgment of the sampling timing. It can be used for pre-job prediction and real-time assessment of contamination levels during sampling processes, which can improve operational efficiency and reducing risks.
Authors: YU Qiang, ZHOU Yanmin, CHU Xiaodong, NIE Yiying, ZUO Youxiang, KONG Sun, ZHOU Minggao
Institution: China University of Petroleum, Beijing
Language: English
Subfield: Ocean Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6924827349
FWCI:
Cites:
Cited by:
Related to: