A method of performing measurements of an earth formation includes disposing at least a first receiver and a second receiver in one or more monitoring boreholes in a formation, and injecting fluid into the formation from an injection borehole, wherein injecting includes operating a fluid control device to generate seismic and/or acoustic noise having an identifiable characteristic. The method also includes detecting seismic and/or acoustic signals at the first receiver and detecting seismic and/or acoustic signals at a second receiver, the seismic and/or acoustic signals corresponding to the seismic and/or acoustic noise, calculating an estimate of a Green's function between the first receiver and the second receiver by processing seismic and/or acoustic waves detected by the first receiver and the second receiver to at least partially reconstruct the Green's function, and estimating variations in a velocity of a region of the formation by determining variations in the reconstructed Green's function.

A method for correcting imbalance in IQ demodulation of optical fiber DAS data is disclosed. The method includes obtaining I/Q signals I0 and Q0 of each sampling on an optical fiber from a DAS acquisition instrument; performing DC bias correction on the I/Q signals I0 and Q0 to obtain I/Q signals I1 and Q1 after DC bias correction; performing amplitude imbalance correction on the I/Q signals I1 and Q1 after DC bias correction by using Hilbert transform to obtain I/Q signals I2 and Q2 after amplitude imbalance correction; and performing phase imbalance correction on the I/Q signals I2 and Q2 after amplitude imbalance correction by using Hilbert transform to obtain I/Q signals I3 and Q3 after phase imbalance correction. It is possible to perform accurate correction on I/Q signals, accurately and efficiently suppress optical fiber demodulation noise, and improve optical fiber DAS data acquisition quality.

A method for performing a seismic processing task using a machine-learned model developed using an in-domain adversarial attacker. The method includes obtaining a machine-learned model parameterized by a set of weights and generating a synthetic seismic dataset and associated target. The method further includes determining a noise profile for the synthetic seismic dataset in a frequency domain that when added, in a spatial-temporal domain, to the synthetic seismic dataset reduces a performance of the machine-learned model. The method further includes adding the noise profile to the synthetic seismic dataset forming a noisy seismic dataset and updating the set of weights of the machine-learned model based on the noisy seismic dataset and the target. The method further includes receiving a seismic dataset corresponding to a subsurface, processing the seismic dataset with the machine-learned model to form a predicted target, and developing a geological model for the subsurface using the predicted target.

The combination of one or more 3-component microseismic sensors deployed into a wellbore adjacent a microseismic event and a linear array of distributed fiber optic acoustic sensors deployed uphole thereof provides two sets of data for establishing noise-free signals for locating the microseismic event in the formation about the wellbore. The distributed fiber optic signals monitor noise transmitted along coiled tubing used to pump a completion operation or as a result of the fluid flowing through the casing or coiled tubing, or along wireline used to deploy the microseismic sensors. The noise is mapped and extrapolated for estimating noise at the 3-component sensors. The estimated noise is removed from the 3-component sensor data for producing clean signals representing the location of the microseismic events.

A technique includes receiving sensor data; sorting the data into a gather representation that corresponds to a plurality of shots of an energy source; and determining a signal cone based at least in part on at least one characteristic of the gather representation. The technique includes processing the sensor data in a processor-based machine to attenuate noise to generate data representing a signal based at least in part on the determined signal cone and the gather representation.

A system includes an acoustic attenuation interface disposed between a first acoustic transmission conduit and a second acoustic transmission conduit. An acoustic signal source acoustically coupled to the first acoustic transmission conduit generates an acoustic signal. An acoustic noise source acoustically coupled to the second acoustic transmission conduit generates an acoustic noise. A first sensor is configured to detect a first composite signal including the acoustic signal after transmission through at least a portion of the first acoustic transmission conduit and an attenuated acoustic noise. A second sensor is configured to detect a second composite signal including the acoustic signal after transmission through at least a portion of the first acoustic transmission conduit and attenuated by the acoustic attenuation interface and the acoustic noise. An acoustic signal processing system is configured to determine a noise-reduced signal from the first composite signal and the second composite signal.