Triad National Security, LLC

Chevron U.s.a. Inc.

Velimir V Vesselinov

Filip L Iliev

Ludmil B Alexandrov

Boian S Alexandrov

Valentin G Stanev

Margaretha Catharina Maria Rijken

Karim Shafik Zaki

Rajesh S Nair

Yan Li

A computer implemented method for prediction of geomechanical performance including productivity index decline and completion integrity for a well or a hydrocarbon reservoir using a geomechanics informed machine intelligence (GIMI) algorithm. The method includes running a geomechanical reservoir simulator to generate training datasets for the hydrocarbon reservoir and incorporating physical models and identified variables into the GIMI algorithm. The method further includes training a neural network of the GIMI algorithm by using correlated training datasets that correlate to the physical models to produce a resulting prediction model and performing sensitivity analysis on the resulting prediction model. Additionally, the method includes identifying dominant variables for damage mechanisms through design of experiment statistics and performing history matching and blind test on the resulting prediction model. Lastly, the method includes updating the identified variables and models incorporated into the GIMI algorithm.

Geomechanics informed machine intelligence

Machine-learning methods and apparatus are provided to solve blind source separation problems with an unknown number of sources and having a signal propagation model with features such as wave-like propagation, medium-dependent velocity, attenuation, diffusion, and/or advection, between sources and sensors. In exemplary embodiments, multiple trials of non-negative matrix factorization are performed for a fixed number of sources, with selection criteria applied to determine successful trials. A semi-supervised clustering procedure is applied to trial results, and the clustering results are evaluated for robustness using measures for reconstruction quality and cluster separation. The number of sources is determined by comparing these measures for different trial numbers of sources. Source locations and parameters of the signal propagation model can also be determined. Disclosed methods are applicable to a wide range of spatial problems including chemical dispersal, pressure transients, and electromagnetic signals, and also to non-spatial problems such as cancer mutation.

Source identification by non-negative matrix factorization combined with semi-supervised clustering

Growing community of inventors

Santa Fe, NM, United States of America

Velimir V Vesselinov