ILI Benchmarking – Ensuring an Adequate Understanding of ILI Capability Under Real- World Operating Conditions
Modern inline inspection tools offer an unequalled method to rapidly inspect pipeline systems
for a range of damage mechanisms and have undoubtably assisted reduction of incident rates
within the pipeline industry. Cost of inline inspection can be high; in costs of preparation such
as cleaning pig runs, increase of headcount on or in controlled installations, cost of reduced
production in addition to the actual cost of the inspection by the ILI tool.
Inspection tool efficacy is therefore a key consideration, particularly in cases where secondary
costs in addition to the actual ILI tool vendor cost are comparatively high. In these cases, the
overall cost benefit analysis of ILI tool can give an increased incentive to contract the best
technology and tool available. Commonly the data driving the cost benefit analysis is provided
by the specific ILI tool vendor; so how can operators verify the claims of the ILI vendors and
prove comparative ILI tool efficacy between vendors under real-world operating conditions?
Recent advances in computational capability of commonly available computer hardware has
enabled increased capability in analysis conducted on large datasets. This includes the
capability to conduct statistical analysis of every feature detected by multiple inline inspections across many individual pipelines.
The authors combine techniques commonly used to predict corrosion growth rate such as
statistical correlation of inline inspection results, with many inline inspection datasets and
advanced data science techniques. The various techniques utilised by the authors to analyse
the claimed probability of detection and sizing tolerances of various ILI tools are broadly based
on regression modelling with additional techniques where required for algorithm optimisation.
The authors will detail a typical anonymised output of statistical analysis of multiple ILI vendors and ILI tools and will explain the insights that this can provide, irrespective of ILI vendor claimed specifications. The authors will demonstrate that the methodology allows an unbiased analysis of ILI tool capability under real-world operating conditions, and how this can be used to better inform the cost benefit analysis process used when selecting a specific ILI vendor and ILI tool.