In the pipeline - Ultrasonic phased array non-destructive testing of polyethylene pipes
The increase in the use of plastic as a pipe material to transport gas and water has been attributed to factors such as corrosion resistance, strength-to-weight ratio, lightness, flexibility and cost.
One joining process for pipes is electrofusion welding. As for all welding processes, defects can occur due to poor practices, which in turn affect the quality of the joint. To prevent this, an effective non-destructive technique is required to reduce environmental risk caused by leakage of dangerous chemicals from polyethylene pipe systems.
Polytec, a project from the European Commission and The Welding Institute (TWI), based in Cambridge, UK, has conducted exploratory work for in-manufacture, online, quality control of polyethylene electrofusion joints.
To inspect the joints, an ultrasonic phased array method has been developed. The technique can detect conventional defects and determines whether the weld has been correctly heated by analysing the position of the heat affected zone (HAZ).
Flaws and locations
Welding faults from bad preparation of the pipe, poor cleaning, the pipe surface not being scraped, badly clamped fittings, or from not respecting fusion time can generate defects such as:
• Voids – these flaws are totally subsurface and volumetric (not planar). They contain no material and appear above the wires.
• Lack of penetration (LOP) – caused by the pipe not being fully inserted.
• Lack of fusion (LOF) – the unfused matting surface of pipe and coupling, usually located below the wires.
• Not scraped – the pipe surface in the fusion area has not been removed and does not expose clean virgin material.
• Incorrect heating cycle – this flaw results in a weak joint due to an insufficient heating time.
The Welding Institute inspected the electrofusion joint in field conditions. The technique was developed on welded samples containing manufactured defects of known size and quantity, on pipes 125-250mm in diameter.
The image also shows the external varying shape of the connector sleeve and some of the inspection restrictions, such as labelling and fusion indicators. The only access is the external surface of the sleeve.
Detected ultrasonically, LOF defects are behind the heating wires. In this case, the copper wires have diameters ranging from 0.6-1.7mm with a spacing of two to four milimetres, depending on the coupling design. To resolve any defects from the wires, a tightly focused ultrasonic beam is necessary. Consequently, a contact inspection would not be possible and a water gap scan would be more appropriate. A single pass of the probe around each end of the coupling is desirable.
Three different types of flaws were intentionally developed – LOF, LOP and incorrect heating. These were generated by grease, cold welds, poor penetration and the pipe not being scraped. The majority of the welds were subjected to a crush test after non-destructive testing to verify the results gained with ultrasonic inspection.
The ultrasonic characteristics of the weld change with temperature. To provide stable inspection conditions, the elapsed time after welding was measured. The longitudinal ultrasound velocity in the coupling at 20°C was measured at 2,220m/s. It was concluded that the 250mm diameter needs to wait two hours after heating before being inspected, whereas a 125mm diameter coupling could be inspected one hour after heating.
To achieve the required flaw detection resolution, a seven megahertz phased array probe was designed by TWI and manufactured by Vermon, based in Tours, France, for the 125mm couplers. This was a compromise between material attenuation, focal spot size and defect resolution. A phased array probe with a 0.4mm focal spot was produced at the coupling fusion face to enable resolution of the wires and flaws.
Both ends of the electrofusion joint need to be inspected. To record the probe position and achieve an immersion inspection around the pipe, a prototype phased array non-destructive test manipulator has been manufactured, which provides a scan around the coupling on each end of the pipe. The prototype has been built to fit on a 125mm diameter pipe. The samples tested contained either no flaws or a selection of flaws such as LOP, LOF and heating time reduced.
From the trials, it was established that the defect types listed above were detected and sized. Incorrect scraping of the pipe ends generates a weak weld, but not a LOF. This defect has not been detected reliably by phased array inspection.
All the tests showed that LOF defects, which were four milimetres in diameter, could be reliably detected in the joint configurations.
Characterisation of LOP and LOF is provided by the position of the wires. Both flaws are under the heating wires, but for LOF the wires are in a straight line, whereas for LOP, misalignment can be seen.
Lack of heating can cause a reduction in joint strength, but it does not necessarily cause LOF. This defect is found by measuring the position of the HAZ. If the heating time is reduced, the zone appears closer to the wires on the ultrasonic image.
A correlation has been established between the physical measurement of the HAZ and the results obtained ultrasonically.
These prove that the phased array inspection method can provide a reliable value for the HAZ and diagnose an incorrect heating cycle.
The prototype manipulator has been trialled at Italgas, Italy, and proved easy to handle, with an inspection time of under two minutes for a complete coupling.
To assess the cohesion of a polyethylene pipe/electrofusion socket or saddle assembly, a crush test was performed by crushing the joint pipe adjacent to the coupling.
A comparison of the ultrasonic phased array results versus the intended flaws inserted into the pipe joints has been made at Italgas under blind inspection conditions. The crush tests demonstrated that the quality of the welds could be determined by the inspection system.
To view detailed images, graphs and diagrams referred to in this feature, please download the PDF of the original article using the link below.