US 12,346,365 B2
Noise reduction and validation of in-line inspection tool data
Rida Majed Bin Hashim, Safwa (SA); and Hassan Mohammadali Al Matouq, Al Qatif (SA)
Assigned to SAUDI ARABIAN OIL COMPANY, Dhahran (SA)
Filed by SAUDI ARABIAN OIL COMPANY, Dhahran (SA)
Filed on May 15, 2023, as Appl. No. 18/317,761.
Prior Publication US 2024/0386045 A1, Nov. 21, 2024
Int. Cl. G06F 16/35 (2019.01); G06F 16/355 (2025.01)
CPC G06F 16/355 (2019.01) 18 Claims
OG exemplary drawing
 
1. A method, comprising:
deploying an inline inspection tool in a pipe, the inline inspection tool comprising at least one sensor configured to measure a first target attribute in the pipe;
obtaining inline inspection tool data from the inline inspection tool comprising a first plurality of data points wherein each data point in the first plurality of data points comprises the first target attribute and a first clustering attribute;
receiving archival data comprising a second plurality of data points wherein each data point in the second plurality of data points comprises the first clustering attribute;
receiving a first weight and first local shift function parameterized by a first set of parameters for the first clustering attribute;
iteratively, until meeting an acceptance criterion:
assigning, by a clustering algorithm, each data point in the first and second pluralities of data points to a cluster of one or more clusters based on the first clustering attribute scaled using the first weight and the first local shift function;
subtracting a center of each of the one or more clusters from the first target attribute of each data point in the first plurality of data points, according to the respective cluster of each data point in the first plurality of data points, forming a plurality of differences;
constructing one or more subsets from the plurality of differences;
determining, at least, a subset score and a number of outliers for each of the one or more subsets, wherein the subset score of a given subset is composed of a normality score that indicates the normality of the given subset, an average value score that is based on the average of the given subset, and a similarity score that is based on the ratio of a standard deviation of the given subset and a standard deviation of the first target attribute of the first plurality of data points; and
adjusting the first weight and the first set of parameters based on the subset score and the number of outliers of each of the one or more subsets;
determining a first target attribute average for each of the one or more clusters;
reconstructing the first target attribute of each data point in the first plurality of data points by assigning the first target attribute for each data point to the first target attribute average according to the cluster to which each data point is assigned;
determining an area of corrosion based on the first plurality of data points with reconstructed first target attributes; and
repairing the area of corrosion.