Neural Network Seismic Inversion for Non-Vertical Well Trajectory Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional seismic imaging tools face challenges in achieving high vertical resolution, especially in carbonated subsoils, due to uncertainties in seismic wavelet propagation rates, which are exacerbated in non-vertical wells, limiting the accuracy of hydrocarbon volume estimation and reservoir modeling.
Innovation Solution
A method utilizing a trained neural network to process seismic data from both vertical and non-vertical wells, correcting the trajectory of non-vertical wells based on comparison with observed well data, thereby enhancing seismic image definition and resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional seismic imaging tools are used to process seismic data, then the processing can be completed with standard methods, but the vertical resolution of the seismic image deteriorates due to high propagation rates and uncertainties in rate models
Solution Approach 1:
The method applies preliminary corrections to the trajectory of non-vertical wells before seismic image construction. By determining and correcting well trajectories using logged data and seismic data integration, the system prepares accurate spatial references in advance, which then enables high-resolution imaging without the degradation caused by propagation rate uncertainties
Solution Approach 2:
The system uses feedback loops where seismic data and well log data are continuously integrated and compared. The seismic images are used to update and refine well trajectory models, which in turn improve subsequent imaging. This iterative feedback process progressively enhances both trajectory accuracy and vertical resolution simultaneously
2Speed
If seismic wavelet is propagated at high rate in carbonated subsoil, then the propagation speed increases, but the vertical resolution and spatial accuracy deteriorate due to calculation uncertainties
Solution Approach 1:
The system introduces well trajectory correction as an intermediary element between the seismic data and the final image. By using logged trajectory data from non-vertical wells as a reference framework, the system mediates the relationship between high-speed propagation and spatial accuracy, allowing fast propagation without losing resolution
Solution Approach 2:
The method changes the parameter of well trajectory accuracy by applying corrections based on logged data. This parameter change compensates for the loss of spatial precision that would otherwise result from high propagation rates, effectively decoupling propagation speed from imaging resolution
3Measurement precision
If methods from WO 2015/067864 are applied to vertical wells, then seismic image resolution is enhanced, but the method cannot be applied to non-vertical wells within the reservoir zone
Solution Approach 1:
The system creates a universal method that works for both vertical and non-vertical wells. By integrating well log data and seismic data in a unified framework that handles arbitrary well orientations, the system makes the high-resolution imaging method applicable to all well types, not just vertical ones
Solution Approach 2:
The method transitions from handling only vertical well data to incorporating three-dimensional well trajectories. By adding the dimensional aspect of well orientation and using logged trajectory data in 3D space, the system extends the applicability of resolution enhancement to non-vertical wells while maintaining image quality
Data Source
Figure 1
Figure 2a~2b
Figure 3a~3b
AI summary
Disclosed is a method for processing seismic data relating to a reservoir zone comprising a plurality of substantially vertical wells and one or more non-vertical wells. The method comprises obtaining a trained neural network, having been trained to infer well data from seismic data and having been trained on training well data relating only to said substantially vertical wells using the trained neural network to invert seismic data relating to said reservoir zone, to obtain inversion well data. The inversion well data is compared to observed well data relating to the reservoir zone and a trajectory correction is determined to correct a trajectory of a non-vertical well of said one or more non-vertical wells based on the comparison.