Reverse Time Migration Using Randomized Absorbing Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional reverse time migration techniques are limited by the use of saved source wavefields and boundary values at each time step, which restrict the full utilization of computational power offered by co-processors like GPUs and FPGAs, especially in handling absorbing and randomly perturbed boundary regions.
Innovation Solution
The method involves propagating source and receiver wavefields through an earth model with absorbing and randomly perturbed boundary regions, redefining these boundaries for forward and backward propagation to dampen and amplify wavefields respectively, thereby reducing artifacts and improving signal-to-noise ratios and image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional reverse time migration is used with saved source wavefields and boundary values at each time step, then imaging accuracy is maintained, but computational efficiency is limited and co-processor power cannot be fully utilized
Solution Approach 1:
The patent changes the boundary conditions from conventional absorbing boundaries to randomized absorbing boundaries with specific statistical properties. This parameter change allows the wavefield to be properly absorbed while enabling more efficient computational handling that can fully utilize co-processor power, thereby improving computational efficiency without sacrificing imaging accuracy
Solution Approach 2:
The patent applies preliminary randomized absorbing boundary conditions during the forward propagation phase. This preliminary action prepares the wavefield in a state that eliminates the need for saving boundary values at each time step, allowing direct utilization of computational resources during reverse time migration and improving overall processing efficiency
2Reliability
If absorbing boundary regions are used in conventional reverse time migration, then wavefield artifacts are reduced, but computational resources cannot be fully optimized
Solution Approach 1:
The patent modifies the absorbing boundary parameters by introducing randomization with specific statistical characteristics. This changes the boundary behavior to maintain wavefield absorption effectiveness for artifact reduction while simultaneously enabling more efficient computational processing that optimizes resource utilization
3Manufacturing precision
If source wavefield is propagated forward through earth model with absorbing and randomly perturbed boundaries, then wavefield damping and scattering occur, but storage access requirements increase
Solution Approach 1:
The patent changes the boundary interaction parameters by applying randomized absorbing conditions. This parameter change allows the wavefield to experience proper damping and scattering for accuracy while the randomized nature enables more efficient memory management during reverse propagation, reducing overall storage requirements
Solution Approach 2:
The randomized absorbing boundaries are applied during forward propagation as a preliminary action that shapes the wavefield in a way that reduces the need for extensive storage of intermediate wavefield states. This preliminary shaping allows efficient reverse propagation with reduced storage access requirements
Data Source
AI summary
Images relating to a subsurface region may be generated based at least in part on a backward propagated source wavefield and a receiver wavefield. A source wavefield may be propagated from an initial wavefield-state forward in time, from an initial time-state to a final time-state, through an earth model associated with the subsurface region. The backward propagated source wavefield may be determined by propagating the source wavefield backward in time, from the final time-state to the initial time-state, through the earth model to reconstruct the initial wavefield-state. The receiver wavefield may be propagated, from the final time-state, through the earth model. The earth model may include at least one boundary region that can be defined as having one or more of absorbing characteristics, boosting characteristics, randomly perturbed characteristics, and/or other characteristics. As such, wavefields may be dampened, amplified, randomly scattered, and/or otherwise altered at the at least one boundary region. These wavefields may be used for constructing images of subsurface regions with improved signal-to-noise ratios.