REVISION AND VALIDATION OF FILTERING TECHNIQUES FOR CASING WEAR PROFILES IN OIL WELLS

Resumo

Aimed to achieve a better petroleum well planning, engineers have to correlate several data
about drilling parameters, geological formations and reservoir properties, as well as many other details
collected from drilled wells databases. This information helps to generate well’s trajectory and improve
casing design. In the actual scenario of reaching deeper reservoirs and running complex well paths, a
common problem is casing wear. This phenomenon mainly results in material removal from casing inner
wall during well drilling process. Since this problem impacts well’s integrity and safety, the
improvement of casing design to consider accurate material loss is crucial. A regular procedure for
casing wear inspection is to use a logging device, like the ultrasonic logging tool, to collect several
discrete measures of inner radius and thickness variations. This data is used to determinate wear rate
along well depth. Therefore, wear profiles obtained from drilled wells, could be used to express
important information about wear intensity to future projects on planning phase. A problem that usually
appears in attempt to analyze this profiles, is the spurious spikes constant presence that corresponds to
unrealistic wear values, which are known as noise. The source of this meaningless data is mostly
attached to the presence of numerous connections between pipes, due to the great casing string
extension, and it is called casing collars. This study aims to use ultrasonic logging data to determine
more representatives wear profiles ignoring positions of unrealistic wear measurements. A revision and
validation of filtering methods is used to accomplish the wear profile improvement study. A Fast-Fourier
Transform low-pass filtering is implemented. The filter’s input is a raw profile with wear rate estimated
for several measure depth and with intrinsic noise that could be misinterpreted as casing wear. It is
discussed the effectiveness of the filtered profile when correlated to casing collars positions input
obtained from logging. Therefore, a study is performed using a regular industry’s procedure to
demonstrate the potential of the methodology to improve a casing wear representation.