Neural Network Seismic Signal Processing for High Vertical Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional seismic imaging tools face challenges in achieving high vertical resolution, particularly in carbonated subsoils, due to high seismic wavelet propagation rates, which can lead to significant uncertainties and loss of high-frequency information, hindering accurate hydrocarbon or gas volume estimation and reservoir modeling.
Innovation Solution
A method involving the use of a neural network to process seismic signals by identifying seismic wavelet reflections, determining wavelet length through autocorrelation, and training the network with sub-portions of seismic signals and geological data to enhance image resolution, ignoring wavelet variations and preserving high-frequency information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional seismic imaging tools are used with high propagation rates, then processing complexity is reduced, but vertical resolution of the seismic image deteriorates
Solution Approach 1:
The patent changes the parameter of wavelet length used in processing. By determining the actual wavelet length from the seismic signal itself and using this determined length for training the neural network and processing, rather than relying on conventional fixed-length assumptions, the method achieves both manageable processing complexity and improved vertical resolution.
Solution Approach 2:
The patent replaces conventional seismic imaging processing methods with a neural network-based system. The neural network is trained to recognize patterns and extract geological information directly from seismic signals, substituting traditional mechanical processing algorithms with an adaptive intelligent system that preserves high-frequency information while maintaining processing efficiency.
2Speed
If seismic wavelet is propagated at high rate in carbonated subsoil, then propagation speed increases, but measurement precision of geological information deteriorates
Solution Approach 1:
The patent introduces a neural network as an intermediary between the seismic signal and the geological information extraction. The neural network processes the high-rate propagated signals, learning to compensate for the effects of high propagation rates on measurement precision by training on labeled data that captures the true geological properties despite the distorted signals.
Solution Approach 2:
The patent employs feedback through the training process where the neural network learns from labeled well data corresponding to seismic signals. This feedback mechanism allows the system to adjust its processing to account for high propagation rate effects, improving measurement precision by learning the correction patterns from actual geological measurements.
3Device complexity
If stack seismic signals are used instead of pre-stack signals, then processing complexity is reduced, but loss of high-frequency information increases
Solution Approach 1:
The patent replaces the mechanical stacking process with a neural network-based processing system. Instead of summing pre-stack signals to create stack signals (which loses high-frequency information), the neural network processes individual signals or selectively combined signals, preserving high-frequency components while achieving the noise reduction and simplification benefits traditionally provided by stacking.
4Measurement precision
If wavelet length variations are considered in processing, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent performs preliminary determination of the wavelet length from the seismic signal before the main processing and interpretation steps. By calculating the actual wavelet length upfront and using this value to configure the neural network training and processing parameters, the system incorporates wavelet length variation considerations without adding complexity to the core processing pipeline.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly improves seismic image resolution, enabling more precise hydrocarbon volume estimation, reservoir modeling, and drilling operations by effectively utilizing high-frequency information previously lost in conventional processing methods.
Implementation Method 1
determining a length of the seismic wavelet; wherein the plurality of sub-portions of the at least one portion have a length dependent on the length of the seismic wavelet determined
Data Source
AI summary
A device, computer program and related method for processing a first seismic signal that includes identifying one portion of a second seismic signal and determining a length of a seismic wavelet. It is also possible to train a neural network by using a plurality of sub-portions of said portion a input variables and at least one second piece of information as a target variable. Said sub-portions of the portion have a length dependent on the length of the seismic wavelet determined. Finally, the method includes determining at least one first piece of geological information based on the first seismic signal using said trained neural network.