Stochastic Subsurface Modeling for Geometrical Coherency

The patent transitions from static deterministic modeling to dynamic stochastic modeling by introducing random perturbations that vary across different model realizations. The system dynamically generates multiple equiprobable models by applying stochastic perturbations to the preferred model, allowing the geometry to adapt and account for uncertainties while maintaining geological constraints through iterative refinement.

The patent changes the fundamental parameters of the modeling approach by introducing stochastic elements (random perturbations) to the geometrical parameters of horizons and faults. Instead of using fixed deterministic values, the system employs probability distributions and statistical variations to represent uncertainties in seismic data, transforming the modeling from a single-parameter solution to a multi-parameter probabilistic framework.

The patent performs preliminary action by first determining a preferred model from the seismic data, then using this preferred model as a baseline for generating multiple perturbed realizations. This preliminary step establishes the geological framework and constraints before introducing stochastic variations, ensuring that all subsequent models maintain geological realism while accounting for uncertainties.

The patent creates multiple copies of the preferred model through stochastic perturbation. Instead of building each model from scratch, the system generates equiprobable realizations by copying the preferred model's structure and applying controlled random variations, significantly reducing computational time compared to independent model building while maintaining accuracy.

The patent segments the modeling process into distinct phases: (1) determining the preferred model from seismic data, (2) generating multiple perturbed realizations through stochastic sampling, and (3) analyzing the ensemble of models. This segmentation allows efficient parallel processing of multiple realizations while maintaining a systematic workflow, improving productivity by organizing the computationally intensive stochastic modeling into manageable segments.

The patent creates a universal modeling framework that can generate multiple equiprobable models using a single preferred model as input. The stochastic perturbation methodology serves multiple functions simultaneously: it generates diverse realizations, quantifies uncertainties, maintains geological constraints, and provides statistical analysis capabilities, making the system highly productive despite the complexity of generating multiple models.