Runs the ILI clustering algorithm

Clustering criteria for pipeline defects refer to the guidelines or criteria that are used to determine whether a group of defects in a pipeline should be considered a "cluster." These criteria are typically based on a combination of factors, including the size, location, and severity of the defects, as well as the spacing between them. There are several different clustering criteria that can be used for pipeline defects, depending on the specific context and application. Some common clustering criteria include:

• Percentage of wall loss: This criterion considers the percentage of the pipeline wall that has been lost due to corrosion or other defects. A group of defects that collectively represent a certain percentage of wall loss (e.g., 10%) may be considered a cluster.
• Spacing between defects: This criterion considers the distance between individual defects in a pipeline. If the spacing between defects falls below a certain threshold (e.g., 1 meter), they may be considered a cluster.
• Geographic location: This criterion considers the geographic location of defects within a pipeline. If defects are clustered within a certain area of the pipeline (e.g., a bend or a weld), they may be considered a cluster.
• Severity of defects: This criterion considers the severity or consequence of individual defects. If a group of defects collectively represent a significant risk of failure or leakage, they may be considered a cluster.

Pipeline operators typically use clustering criteria as part of their inspection and maintenance programs to identify and prioritize areas of the pipeline that require attention. By identifying and addressing clusters of defects, operators can help prevent failures and ensure the safe and reliable operation of their pipelines.

Examining the clustering of pipeline defects is a crucial aspect of pipeline integrity management. This involves evaluating the spacing between defects in a pipeline to determine the likelihood of larger defects clustering together. Specific conditions must be met for this distance to be considered acceptable. If these conditions are met, the possibility of defect clustering cannot be overlooked.

Clustering rules

Rules are used to configure the clustering algorithm. There are 11 standard rules and then a 12th custom rule.

Each rule is based on a length and circumferential criteria. If any two anomalies have a horizontal separation less than the length criteria and a circumferential separation less than the circumferential criteria the should be clustered.

The rules are details in the table below.

Rule no.	Length Criteria	Circumferential Criteria
Rule 1	Minimum (L1,L2)	Minimum (W1,W2)
Rule 2	Minimum (6*wall thickness, Minimum (L1,L2))	Minimum (6*wall thickness, Minimum (W1,W2))
Rule 3	Minimum (wall thickness, Minimum (L1,L2))	Minimum (wall thickness, Minimum (W1,W2))
Rule 4	Minimum (L1,L2)	Minimum (wall thickness, Minimum (W1,W2))
Rule 5	Average (L1,L2)	Average (W1,W2)
Rule 6	Wall thickness	Wall thickness
Rule 7	3*Wall thickness	3*Wall thickness
Rule 8	2*Minimum (L1,L2)	2*Minimum (W1,W2)
Rule 9	Minimum (6*wall thickness, Average (L1,L2))	Minimum (6*wall thickness, Average (W1,W2))
Rule 10	1 inch = 25.4 mm	6*wall thickness
Rule 11	1 inch = 25.4 mm	(3 × MAOP × OD) / (SF × SMYS × WJF)
Rule 12	User Configurations	User Configurations

Input

Arguments

• org_id: as string which should be a valid uuid for the organization
• project_id: the id of the data project where the pipeline data is found
• pipeline_id: an array of strings which should be valid UUIDs for pipelines
• size_settings: a string representing the size setting for clustering, possible values
  • original
  • tolerance
• interaction: a string representing the interaction setting for clustering, possible values
  • enabled
  • disabled
• surface_location: a string representing the surface location setting for clustering, possible values
  • matching
  • any
• rule: a string representing the clustering rule to use, possible values
  • rule1
  • rule2
  • rule3
  • rule4
  • rule5
  • rule6
  • rule7
  • rule8
  • rule9
  • rule10
  • rule11

Creating the function on the platform

To create this function on the platform using the cli set up the port forwarding as shown in README.

Then run the following command to create the function.

cargo run functions create \
  -f functions/ili_clustering \
  -d "Runs the ILI clustering algorithm" \
  -i org_id=string \
  -i project_id=string \
  -i pipeline_id=array \
  -i ili_id=array \
  -i size_settings=string \
  -i interaction=string \
  -i surface_location=string \
  -i rule=string

Testing the function locally

You can run and test the function locally by running

cargo run

Then you can check it work with curl as follows

curl localhost:8080 -d $(jq '. | tojson' functions/ili_clustering/example_input.json)