- Dona Paula, Goa, India.
- +91-0832- 2450327
- iiosc2020[at]nio[dot]org
Abstract Submission No. | ABS-2022-13-0335 |
Title of Abstract | 1D poroelastic full waveform inversion (FWI) for the estimation of pore fluid pressure distribution in KG basin, India. |
Authors | Aditya Sen*, Dibakar Ghosal, Pritam Chakraborty |
Organisation | Indian Institute of Technology, Kanpur |
Address | Flat 3821, Maikala, Vijaya Heritage 6th Phase, Marine Drive, Kadma Jamshedpur, Jharkhand, India Pincode: 831005 Mobile: 7980741911 E-mail: adityasen9519@gmail.com |
Country | India |
Presentation | Poster |
Abstract | The Krishna-Godavari (KG) basin is one of the most petroliferous basins located in the eastern continental margin of India. Presence of gas hydrate deposits have been proven in shallow marine sediments of the basin beyond water depth of 750 m. High resolution seismic datasets and bathymetry have shown slumping and sliding of upper and midslope sediments in the KG basin. Slumps and slides cause geohazards affecting offshore production platforms. Slumping and sliding occurs due to variation of subsurface pore fluid pressure from the movement of gas and fluids through regional fault systems. Hence estimation of pore fluid pressure from seismic datasets of offshore reservoirs is essential for prediction of occurrence of slumping and sliding and prevent calamities in development and production of offshore reservoirs. Acoustic and elastic reservoir modelling techniques employ indirect approximations for reservoir characterization as they do not incorporate the pore fluid content of reservoir rocks. Full waveform inversion (FWI) using Biots theory of seismic wave propagation in poroelastic media has been found to be most accurate for modelling and estimation of pore fluid pressure distribution in hydrocarbon reservoirs. In poroelastic FWI, the seismic wavefield contains four components i.e., two components each for the solid and fluid phases of the reservoir rock. This makes the inversion scheme more non-linear compared to elastic modelling techniques and hence, more likely to converge to local minima due to cycle skipping. To ensure convergence to global minimum and accurate inversion of reservoir parameters, global optimization techniques need to be used for poroelastic FWI. Unlike local gradient-based optimization techniques used in conventional FWI, global optimization techniques are computationally cheap and are independent of initial model parameters as they operate on the entire search space. In our abstract, we have used Biots poroelastic theory for one-dimensional poroelastic FWI of pore fluid pressure in marine reservoirs. For the forward modelling scheme, we discretized Biots poroelastic wave formulations using staggered-grid finite difference method of the second order in time and fourth order in space to simulate propagation of seismic waves in fluid-saturated porous media. Convolutional perfectly matched layer (C-PML) was placed at the edges of the computation model to absorb seismic energy incident at the boundaries. In the inversion scheme, we implemented damped-inertia-coefficient particle swarm optimization (PSO), a stochastic global optimization technique, to minimize the error between the observed and modelled seismic trace. The inversion scheme was applied on a 200m x 200 m synthetic model with poroelastic parameters of the Utsira formation and the pore pressure obtained after inversion was considerably accurate compared to their true values. This inversion scheme will be used for estimation of pore fluid pressure in hydrocarbon reservoirs in the KG basin. |