AI Neural Networks for Seismic Subsurface Model Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Geologists' qualitative information about hydrocarbon reservoirs is difficult to integrate into seismic processing workflows, making it a manual and time-consuming process to construct subsurface images for drilling targets.
Innovation Solution
A method using artificial intelligence neural networks to generate subsurface model realizations and simulate synthetic seismic datasets, allowing for the prediction of inferred subsurface models from observed seismic data, thereby identifying drilling targets and determining wellbore paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional seismic processing techniques are used, then high resolution images can be obtained over large volume of space, but the process becomes computer memory and computation intensive and requires manual integration of geological information
Solution Approach 1:
The patent introduces an artificial intelligence model as an intermediary between seismic data and geological interpretation. The AI model automatically integrates geological information and velocity constraints, eliminating the need for manual processing steps while maintaining high-resolution imaging capabilities.
Solution Approach 2:
The patent replaces the manual mechanical processing workflow with an automated AI-based system. The AI model substitutes the need for geophysicists to manually integrate geological information and perform iterative velocity model building, thereby reducing computational complexity and processing time.
2Loss of information
If manual integration of geological information is performed, then qualitative geological knowledge can be incorporated, but the process becomes time-consuming and ad-hoc
Solution Approach 1:
The patent incorporates geological information and velocity constraints into the AI model training phase in advance. During actual seismic processing, the pre-trained model automatically applies this geological knowledge without requiring manual intervention, thereby eliminating time losses associated with ad-hoc information integration.
Solution Approach 2:
The AI model performs self-service by automatically integrating geological information and velocity constraints during the processing workflow. The system independently incorporates qualitative geological knowledge without requiring manual input from geologists, thereby eliminating time-consuming manual processes.
3Reliability
If traditional two-step seismic processing is used, then velocity model estimation can be performed, but the process lacks automated integration of geological constraints
Solution Approach 1:
The patent creates a universal AI model that performs multiple functions simultaneously: velocity model estimation, seismic image generation, and geological constraint integration. This single automated system replaces the traditional two-step process and adds automated geological constraint application, thereby improving reliability while increasing automation.
Solution Approach 2:
The patent implements feedback mechanisms where the AI model iteratively refines velocity models by comparing predicted seismic responses with actual data and adjusting according to geological constraints. This automated feedback loop improves velocity model accuracy while maintaining high automation levels throughout the processing workflow.
Data Source
AI summary
Systems and methods of identifying a drilling target are disclosed. The method includes obtaining a training set of base subsurface models and generating, using a first artificial intelligence neural network, a plurality of subsurface model realizations based on the base subsurface models. The method further includes simulating, for each subsurface model realization, a synthetic seismic dataset and training a second artificial intelligence neural network, using the plurality of subsurface model realizations and the corresponding synthetic seismic dataset for each subsurface model realization, to predict an inferred subsurface model from a seismic dataset. The method still further includes obtaining an observed seismic dataset for a subterranean region of interest, predicting, using the trained second artificial intelligence neural network, an inferred subsurface model from the observed seismic dataset, and identifying the drilling target based on the inferred subsurface model.