Customer Case Study · Pipeline Defect Analytics
An inline inspection run produces thousands of features. HMC turned that data into a 3D defect twin of the pipeline, automated fitness-for-service checks for every metal-loss anomaly, forecast corrosion growth from repeat runs, and produced a risk-ranked dig plan, with GenAI writing the field briefs.
Figures indicative; validated metrics available on request.
Inline inspection vendors deliver huge feature lists. Turning them into safe, economical repair decisions was slow, manual and disconnected from what was around the pipe.
Each run arrived as spreadsheets of depth, length and clock position, reviewed line by line.
Fitness-for-service checks were run by hand on the deepest features, so long shallow defects slipped through.
Repeat runs were not matched, so corrosion rates and remaining life were guesses.
River crossings, villages, cathodic protection readings and design factors were not in the same view as the defects.
Standard fitness-for-service methods, applied to every feature, in context, with people making the final call.
Select a step to see what happened.
Worked alongside the pipeline integrity team on past runs.
Deliverable: Data model and assessment rulebook.
No black boxes for safety decisions.
Deliverable: Assessment engine specification.
One place from feature to field.
Deliverable: Working tool and Power BI views.
Engineers sign off every repair.
Deliverable: Validation report.
Integrity management as a routine.
Deliverable: Roll-out and governance plan.
A synthetic 24 km, 12-inch product pipeline segment with 73 reported anomalies. Run the pig, inspect any defect, assess it, project its growth and plan the digs.
Optional. Every ✦ button already works on built-in templates. Add an OpenAI key to ask the run anything in your own words and to have AI write dig briefs and run summaries.
The key goes only from this browser to OpenAI and stays in memory unless you tick "remember". Only the synthetic data on this page is sent.
The inspection tool travels the line and each anomaly appears where it was reported. Drag to rotate, scroll to zoom, click a defect for its details. Below, the same pipe is unrolled into a wall map.
Synthetic pipeline and inspection data. Defect positions and sizes are illustrative; the 3D pipe is exaggerated in diameter for visibility.
Modified B31G (0.85 dL) estimates the failure pressure of a corroded pipe from its depth and length. Pick a defect or move the sliders: the cross-section, the acceptance curve and the verdict update live.
Simplified Modified B31G for illustration. Real assessments also apply tool tolerances, interaction rules and RSTRENG effective-area methods where needed.
Matching the 2021 and 2026 runs gives a corrosion rate for every metal-loss defect. Play the years to watch defects deepen and cross the repair line. The risk matrix ranks the digs by likelihood and consequence.
Likelihood from remaining life; consequence from location (village, river and road crossings score highest). Click a cell to filter.
Pick a defect and get a dig brief for the field crew: what to expect, the assessment, likely cause, safety steps, inspections at the dig and the repair. Or summarise the whole run for management.
Pick a defect and press "Write dig brief".
Template briefs follow standard integrity practice; with a key, the AI writes from the same defect data and rules. Engineers review every brief before use.
Every anomaly now has a decision within days, and the analytics pointed cathodic protection and pigging fixes to the stretches that needed them.
Anomalies by decision status and type (circle size = count)
Horizon chart, mm per year by km; darker bands mean faster corrosion
Figures indicative; validated metrics available on request.
Automating a standard method removes sampling bias; long, shallow defects get the same scrutiny as deep ones.
Matched runs turn depth into remaining life, and remaining life into a dig calendar.
The same defect under a village or a river matters more than in an open field. Put location in the risk score.