Uncertainty Mitigation of Transformation Models in Geotechnical Parameters for Offshore Well Projects.

Resumo

Present work implements methodologies to mitigate uncertainty in design parameters derived from transformation models even when data is limited. Geotechnical inherent uncertainty arises from averaging soil properties due to the small number of tests available in both field and laboratory, as well as measurement errors. In top-hole design, soil characterization is essential for design planning, ensuring well integrity during construction and operation. It also impacts project costs, as conservative safety measures often increase expenses. Offshore oil well construction and completion in Brazil represents about 30% of total investment. Therefore, improving soil characterization can reduce costs without compromising safety. Published studies show that incorporating correlations between soil parameters in transformation models can reduce the coefficient of variation of design parameters. The literature review confirmed the feasibility of this approach with satisfactory results. This work assesses a local clay database to simulate more limited regional data, developing a conventional transformation model compared to one including parameter correlation. Since measured values are known, the “leave-one-out” method validates the models by comparing results to actual data. This approach promotes accurate and cost-effective practices by showing that relatively simple methods can contribute significantly to more efficient and safer well planning decisions.