Comparison of Pure and Mixed Integration Schemes in Material Point Method Simulations of Hammer-Driven Conductor Installation in Undrained Cohesive Soils

Autores

  • Mávyla Sandreya Correia Tenório
  • João Paulo Lima Santos
  • Rafael Dias

Palavras-chave:

Material Point Method, Mixed Integration Scheme, Numerical Simulation, Impact Driving, Conductor Casing

Resumo

The installation of conductor casing is a critical step in offshore oil well construction, providing structural support, isolating seawater, and withstanding environmental loads. Among installation methods, hammer driving stands out for its minimal environmental impact and operational flexibility, involving three phases: self-weight penetration, suction, and hammering. Numerical simulations are essential for optimizing this process, but challenges like large soil deformations require advanced methods such as the Material Point Method (MPM). However, while MPM has proven effective for large-deformation geotechnical problems, its conventional formulation faces limitations in stress accuracy and numerical stability—particularly in undrained cohesive soils where pore pressure effects dominate. To address these challenges, this study evaluates the driving of conductor casing in undrained cohesive soils, typical of offshore environments, using two numerical methods: Material Point Method (pure MPM) and Mixed Integration (MPM with Gauss integration). The soil is modeled via Mohr-Coulomb criterion, with geotechnical parameters derived from real data from the Campos Basin in Brazil. The conductor is modeled as rigid body, and the analysis focuses on soil response during penetration. Results show that Mixed Integration provides superior stability under large deformations, reducing numerical problems and improving stress distribution accuracy. Field data comparisons validate the method’s efficiency, while pure MPM offers computational advantages for simpler models. The conclusion highlights the trade-off between accuracy and cost, with Mixed Integration being recommended for critical scenarios.

Publicado

2025-12-01

Edição

Seção

Artigos