Deep Learning Pipe Inspection Log Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The interpretation of inspection logs for underground casing and tubing, as well as surface metal pipes, is currently manual, time-consuming, subjective, and error-prone, which can lead to delayed detection of integrity issues like corrosion.
Innovation Solution
A computer-implemented method using deep learning models, such as UNet classifiers, to classify the condition of pipes based on inspection logs from tools like multi-finger caliper tools, EM phase shift tools, and ultrasonic imaging tools, thereby automating the interpretation process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual interpretation of inspection logs is used, then human expertise can be applied to assess pipe integrity, but the process becomes time-consuming and error-prone
Solution Approach 1:
The patent replaces the manual mechanical interpretation process with an automated image processing and machine learning system. The system uses EM phase shift tool logs to generate thickness images, which are then processed through a trained deep learning model (such as UNet or CNN) to automatically classify integrity states, substituting human manual analysis with automated computational methods that are both faster and more consistent.
Solution Approach 2:
The system enables self-service interpretation by training the deep learning model on labeled inspection logs created by human experts. Once trained, the model independently processes new inspection logs without requiring continuous human intervention, allowing the system to serve itself by automatically generating integrity assessments from raw inspection data.
2Reliability
If manual interpretation methods are used, then flexibility in handling various defect types is maintained, but consistency and objectivity decrease
Solution Approach 1:
The patent transforms the interpretation process by changing the parameters from subjective human judgment to objective quantitative image analysis. The EM phase shift tool generates thickness images with precise measurements, and the deep learning model processes these images using learned parameters from training data, converting qualitative expert assessment into quantitative automated classification with consistent results.
Solution Approach 2:
The system segments the interpretation task into distinct processing stages: (1) generating thickness images from EM phase shift logs, (2) preprocessing and normalizing the images, (3) applying the trained deep learning model for classification, and (4) outputting integrity state labels. This segmentation allows each stage to be optimized independently while maintaining overall system consistency.
3Productivity
If automated methods are implemented, then processing speed and consistency improve, but the complexity of the system increases
Solution Approach 1:
The deep learning model serves multiple functions: it can classify different types of defects (corrosion, deformation, wall loss), handle various inspection log formats, and provide both binary (intact/defective) and multi-class (integrity state) classifications. This universality allows a single automated system to replace multiple specialized manual interpretation processes, managing complexity through consolidation rather than proliferation of separate systems.
Data Source
AI summary
A computer-implemented method includes: accessing a first database holding information encoding a set of labels that specify a condition of at least one of: a surface pipe, or an underground enclosure that runs at a plurality of depth locations; accessing a second database holding a plurality of inspection logs that record measurement data of the surface pipe or underground enclosure; based on, at least in part, the labeling information and the plurality of inspection logs, training a deep learning model configured to classify, into the set of labels, the condition of the surface pipe or underground enclosure when presented with the inspection logs; applying the deep learning model to one or more newly received inspection logs containing measurement data of a new surface pipe or a new underground enclosure; and subsequently classifying, into the set of labels, the condition of the new surface pipe or the new underground enclosure.