Lithology Prediction Using Geophysical Age Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting lithology in hydrocarbon exploration using seismic data are inaccurate due to noise and unrealistic stratigraphic interpretations, failing to account for variations in physical properties caused by diagenesis and fluid content, which affects the determination of wellsite locations and production risks.
Innovation Solution
A robust lithology prediction model that incorporates a geophysical age model into a supervised machine learning algorithm, using post-stack seismic reflection volumes and wellbores with calibrated TWT-depth relationships to differentiate seismic characters and reduce noise, thereby generating geologically realistic geological bodies and improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If seismic data alone is used for lithology prediction, then the workflow is simple, but the prediction accuracy is insufficient and cannot provide unequivocal determination of lithology
Solution Approach 1:
The patent combines multiple data sources including post-stack seismic reflection volumes, well data with lithology labels, and geophysical age models into a unified supervised machine learning framework. This integration allows the system to leverage complementary information from different sources to achieve more accurate lithology predictions while maintaining a coherent workflow structure.
Solution Approach 2:
The patent introduces a geophysical age model as an intermediary component that bridges seismic data and lithology predictions. This age model provides temporal context and geological realism to the predictions, acting as a mediator that enhances prediction accuracy while maintaining geological consistency throughout the volumetric prediction.
2Reliability
If traditional seismic attribute analysis is used, then the process is straightforward, but noise and unrealistic stratigraphic interpretations occur
Solution Approach 1:
The patent implements a supervised machine learning approach where the system learns from labeled well data and provides feedback to improve prediction accuracy. The model is trained on known lithology relationships and uses this feedback to refine predictions throughout the seismic volume, reducing noise and unrealistic interpretations through iterative learning.
Solution Approach 2:
The patent performs preliminary actions by training the machine learning model on well data and geophysical age models before applying it to the entire seismic volume. This pre-training phase establishes realistic stratigraphic relationships and noise reduction parameters that are then applied consistently across the volumetric prediction.
3Measurement precision
If seismic data without age modeling is used, then data processing is simpler, but geologically unrealistic predictions are generated
Solution Approach 1:
The patent generates the geophysical age model in advance before performing the main lithology prediction. This preliminary action establishes the temporal framework and geological constraints that guide subsequent predictions, ensuring geological realism without adding time pressure to the main prediction workflow.
Solution Approach 2:
The patent merges the geophysical age model with the supervised machine learning prediction process, integrating temporal and geological constraints directly into the prediction algorithm. This combination ensures that predictions are geologically realistic while efficiently utilizing the age model information throughout the volumetric analysis.
Data Source
AI summary
A lithology prediction that uses a geological age model as an input to a machine learning model. The geological age model is capable of separating and recoding different seismic packages derived from the horizon interpretation. Once the machine learning model has been trained, a validation may be performed to determine the quality of the machine learning model. The quality may be improved by refining the training of the machine learning model. The lithology prediction generated by the machine learning model that utilizes the geological age model provides an improved lithology prediction that more accurately reflects the subterranean formation of an area of interest.