Abstract: This disclosure describes systems and methods for determining spatial distributions of petrophysical properties in a subsurface formation. A method includes obtaining input data including petrophysical data, geophysical data and geological data of the subsurface formation; selecting input features from the input data; forming a training dataset including the input features and corresponding labeled data representing a target petrophysical property of the subsurface formation; training, using the training dataset, an ensemble machine learning model; and determining a spatial distribution of the target petrophysical property based on the trained ensemble machine learning model.

Abstract: A petrophysically regularized time domain nuclear magnetic resonance (NMR) inversion includes using an NMR tool to acquire NMR data and inverting the acquired NMR data in a time domain using petrophysical constraints. The inverted NMR data is analyzed. The petrophysical constraints may be identified by: determining a number of porobodons to seek, defining a plurality of zones in which only a subset of porobodon sets is present, and stacking all NMR echoes in each zone satisfying discriminators. The number of porobodons to seek may be based on knowledge of core samples, logs, and NMR sensitivity. The discriminator logs may be logs sensitive to porosity partitioning. A computing system having a processor, a memory, and a program stored in memory may be configured to perform the method. The system may be conveyed downhole on a wireline, a while-drilling drill string, a coiled tubing, a slickline, or a wired drill pipe.

Abstract: A petrophysically regularized time domain nuclear magnetic resonance (NMR) inversion includes using an NMR tool to acquire NMR data and inverting the acquired NMR data in a time domain using petrophysical constraints. The inverted NMR data is analyzed. The petrophysical constraints may be identified by: determining a number of porobodons to seek, defining a plurality of zones in which only a subset of porobodon sets is present, and stacking all NMR echoes in each zone satisfying discriminators. The number of porobodons to seek may be based on knowledge of core samples, logs, and NMR sensitivity. The discriminator logs may be logs sensitive to porosity partitioning. A computing system having a processor, a memory, and a program stored in memory may be configured to perform the method. The system may be conveyed downhole on a wireline, a while-drilling drill string, a coiled tubing, a slickline, or a wired drill pipe.