Case Studies
Features
Webinars
Demo
AI-Agent Beta
AI-Agent Beta
High Resolution Seismic Conditioning
Mean Seismic Conditioning
Fault Interpretation
Horizon Interpretation
Structure-Oriented Seismic Conditioning
Paleo-Channels Detection
Salt Bodies Delineation
Resolving Sub-Salt Strata on Groningen Field
Improving Faults Imaging on Opunake 3D Data
Improving Seismic Data Quality on Waka Field
Delineation of Salt Bodies
Automatic Horizon Interpretation
Automatic Fault Tracking
Paleo Channels Discovery
Fault Determination
Problematics
Approach
Results
Tools
Business Effect

Tectonic disturbances are not only potential traps, but also barriers to fluid filtration, which critically affect the efficiency of deposit development. Building of a tectonic model using traditional methods is characterized by a high degree of subjectivity, which often leads to the creation of contradictory and structurally unreliable models.


The key business risks from under-reported faults are the following:

Problematics
Up to 20% Lost Profits
The irretrievable loss of reserves in undetected traps associated with low-amplitude faults and cracks
Operational Risks
An increase in the number of complications and accidents during drilling and hydraulic fracturing due to uncontrolled tectonic disturbances
Financial Losses
Decrease in oil recovery factor (up to 50%) due to errors in the well wiring scheme, leading to inefficient drainage.
Geoplat AI uses machine learning algorithms that have been applied to sets of synthetic data containing all potential kinds of faults. This way neural networks learn to recognize amplitude shifts, even with noise component, and determine what is the possibility to find there fault stick or surface.

Therefore, this solution provides a unified, coherent fault model for the entire seismic volume. It gives a specialist maximum information, while maintaining the control over the model.
Approach
Calculation speed
  • For base model - up to 1 day;
  • For advanced model - up to 2-3 days
Required parameter setting
  • 1 or 2 parameters for ML procedure;
  • 3-5 parameters for fault trajectory procedure
Required input data
  • Stacked seismic data;
  • Manual fault picks on 2–5 sections of the volume
Results
An accomplished result
That can be further used in Geoplat AI and other interpretation software
The ability to influence the result
Of tectonic model without performing long-term manual interpretation on hundreds of sections.
Full picture of the fault distribution
Including in areas with complex tectonics and low-quality data
Extracted 3D
fault surfaces
3D fault
probability volume

Tools

Fault Probability Calculation
Transfer Learning
Fault Surfaces Extraction
Fault Probability Calculation
provides a probability volume with values from 0 to 1 indicating fault zones
Distinguishing faults by their probability value
Full picture of the fault distribution in several hours
Automatic interpretation based on probability volume
Transfer Learning
Adds expert's knowledge to correct and update base model for current seismic survey
“Smart” extrapolation of manual picks on few sections to the whole data volume
Adaptation of models to geological and geophysical specifications of the area
Accounting for different types of fault systems within a single consistent model
Fault Surfaces Extraction
Extracts full 3D surfaces from fault probability cube that can be further used in other interpretaion softwares
Automatic generation of sticks, surfaces, and fault polygons
Parametric control of the extracted faults geometry
Export and integration of the result into other softwares
Go to Case Studies
Other AI Solutions
Seismic Facies Identification
Fault interpretation is one of the most difficult tasks within a general structural interpretation workflow.
More
Geobodies Detection
Liverpool city will open its doors to annual surrealistic show. Tenner Hall / August, 15.
More
Seismic Data Enhancement
Liverpool city will open its doors to annual surrealistic show. Tenner Hall / August, 15.
More
Contact Us