A seminar by Gabriel Godefroy at ENSG, Nancy.
Stochastic fault simulation described as a constraint statisfaction problem: brainstorming
On: Friday, 06th of June.
Structural modelling uses sparse and ambiguous field observations to build a model of
subsurface geometries. Stochastic approaches sample uncertainties by producing a large
set of equiprobable models. Geological concepts are often used a posteriori to validate and
update the structural models. Instead of validating each realization, we want to define
and integrate constraints during the stochastic simulation. Each constraint may either
represent a geological concept or a field observation. We propose to formalize stochastic
structural modelling as Constraint Satisfaction Problem (CSP). In this paper, we propose
a fault network simulator relying on Constraint Solving methods. Fault geometries honor
given fault observations and size and orientation distributions. Fault network topology
and kinematic are constrained to avoid inconsistent fault branchings and unlikely fault
displacement maps.
subsurface geometries. Stochastic approaches sample uncertainties by producing a large
set of equiprobable models. Geological concepts are often used a posteriori to validate and
update the structural models. Instead of validating each realization, we want to define
and integrate constraints during the stochastic simulation. Each constraint may either
represent a geological concept or a field observation. We propose to formalize stochastic
structural modelling as Constraint Satisfaction Problem (CSP). In this paper, we propose
a fault network simulator relying on Constraint Solving methods. Fault geometries honor
given fault observations and size and orientation distributions. Fault network topology
and kinematic are constrained to avoid inconsistent fault branchings and unlikely fault
displacement maps.