The MTC has demonstrated the benefits of guided wave structural health monitoring compared to traditional NDT by significantly improving inspection performance with the possibility of reducing inspection costs and minimizing health and safety risks to engineers.
Within the Power & Energy sector there a many high value and safety-critical pipelines which require periodic inspection to ensure they remain safe for operation. Many of these inspections are slow, with significant expense required to gain physical access to the assets which, in some cases, expose engineers to hazardous environments, such as radiation within nuclear power plants or working at height for pipeline bridge crossings. By moving to a Structural Health Monitoring (SHM) approach, permanently attached sensors can carry out the inspection, removing inspectors from hazardous environments and reducing the cost to access the asset. In addition, inspection performance can be improved using SHM signal processing techniques.
The MTC explored the potential benefits of implementing a SHM inspection framework using guided waves for a pipework asset. A 10 m steel pipe rig was constructed, where a number of artificial and real corrosion defects were initiated and grown. Throughout the accelerated defect lifecycles, guided wave inspection data was collected to build a library of asset health information. By applying SHM signal processing techniques, significant improvements in inspection sensitivity were achieved, the potential of machine learning for asset monitoring was demonstrated and a mixed reality visualisation tool created.
The Optimal Baseline Subtraction method, Continuous Optimal Baseline Subtraction method and Independent Component Analysis technique were successfully applied to the guided wave inspection datasets collected.
Using these methods, the sensitivity to corrosion defects was increased by up to 500% allowing for defects below 5% cross sectional wall loss to be detected and their growth tracked.
A mixed reality tool to visualise the inspection data was created that could allow remote analysis of data, display a range of inspection and asset metrics as well as guide remedial work.
Multiple machine and deep learning models were explored as alternative approaches to analyse SHM data to further improve performance
SHM has the potential to significantly reduce inspection costs, improve health and safety and increase inspection sensitivity.
This guided wave SHM approach can be deployed to inspect difficult to access, safety-critical and hazardously located pipeline assets within the nuclear power and energy utilities sectors.
The improved inspection results can be used to more accurately track structural integrity, implement intelligent maintenance scheduling and failure forecast modelling.
