Stochastic ultrasonic scattering from the tips of rough cracks

The inspection of safety-critical components in the nuclear power industry depends on procedures that can detect defects to a given threshold of severity; the acceptance process for this is known as inspection qualification. Inspection qualification in the UK is a highly developed formal activity, and is representative of the best practice in the world. However it can be very conservative if there is uncertainty in the expected measured response. A vital example is the scattering of ultrasound from the tips of rough cracks, such as thermal fatigue cracks or stress corrosion cracks. Ultrasound scattering from crack tips is widely exploited to measure crack sizes, but while the nature of the scattering is well understood for smooth cracks, scattering from the tips of rough cracks can differ significantly, and is not readily predictable. Consequently the qualification of ultrasound inspections for rough cracks has to be subject to severely conservative assumptions, and even so there remains a risk of misinterpreting findings.

This project aims to bring understanding to the nature of the scattering, and to develop predictive modelling tools, such that these conservative assumptions can be safely eroded and the reliability of inspections improved. This will enable industry to reduce the costs of manufacturing and repairing, and down-time from outages, as well as improving confidence in the safe operation of safety-critical plant. The project will build on a strong UK heritage of the knowledge of ultrasound scattering, including recent work by the proposers on the stochastic nature of wave reflections from rough surfaces. The key aim is to deliver a new analytical approach that will predict the statistically expected scattering from the tips of cracks of given characteristics of roughness. The work will also include experimental investigation of real cracks and numerical modelling studies. The new ideas will be applied to the primary ultrasound inspection techniques of Time-of-Flight-Diffraction, Pulse-Echo, and array imaging. The work will be undertaken as a collaboration between researchers in Mechanical Engineering and in Mathematics at Imperial College.

The proposal is being submitted within the UK Research Centre in NDE (RCNDE) to its targeted research programme. The proposal has been reviewed internally by the RCNDE, approved by the RCNDE board, and supported financially by five RCNDE industrial members.

Scattered signals from the tips of cracks are vital for the detection and sizing of cracks in safety critical components, and in particular in thick section pressure-containing components in nuclear power plant. The established techniques are Time of Flight Diffraction and Pulse-Echo, both of which rely on the signal from the tip of the crack to indicate its location. Emerging new techniques using ultrasound array imaging also aim to identify the location of the crack tip. The location of the tip indicates the crack length which is the most important information about a crack for the structural integrity evaluation.

Inspection qualification in the nuclear industry is a rigorous process that must be followed in order for a proposed inspection to be permitted. This uses evidence from a range of sources, largely experimental. In the absence of sufficient evidence, conservative assumptions must be made, and these can be very restrictive.

Until recently it has not been possible to manage the uncertainty of randomly varying signals such as are encountered from rough defects except by using very conservative safety factors. However recent progress in handling the stochastic nature of ultrasound scattering and in being able to perform simulations of large numbers of realisations has made it possible to address this. Success with this project in understanding and predicting the stochastic scattering from crack tips will enable a step change in the inspection of components that might contain rough cracks. This will reduce conservatism in the assessment, thus enabling huge reductions of costs. It will also improve confidence in the inspections.

The project is strongly supported by five industrial members of the RCNDE who are leading companies in the inspection and management of nuclear plant, and who share the same interest in pursuing this goal. They have identified this challenge to be one of the most important topics for research. It is also identified as a research priority by the UK's Office for Nuclear Regulation (ONR). This is vital work both for the existing fleet of power stations and as we move forward to the design and inspection planning for next generation power stations. The successful delivery of new capability will be of immediate use for exploitation by these companies.