A statistical investigation of ultrasonic logging data for a petroleum well casing integrity analysis

Resumo

The structure responsible to maintain well stability against borehole collapse as well as to prevent fluid
flow between formation and the borehole in an uncontrolled way is the casing string. It consists of several high
resistance steel tubulars segments connected along well trajectory. Due to severe conditions in-situ, casing integrity
may be compromised, impacting the capability to withstand loads. The scenario may be more critical concerning
to resistance reduction when casing wear is caused by drilling operation of deeper sections of the well. Casing
integrity is usually assessed by means of logging devices, such as the ultrasonic logging tool, to acquire casing
geometry data, like internal radius and wall thickness along well depth. This work proposes to apply a statistical
investigation on well logging datasets in order to characterize geometric data distribution of in-situ tubulars. It is
discussed ultrasonic logging data treatment to obtain reliable results as well as the comparison between statistical
variables from manufactured tubes presented in API TR 5C3:2008 code and those inferred by logging data. Python
codes are developed using statistic features to assist the data analysis. The statistical study is justified by its
contribution to improve the identification of imperfections and damage in logging data.