Probability of Detection analysis to support Non-Destructive Testing

Applications for corrosion resistant alloys are being exposed to increasingly challenging operational environments, placing additional expectations on materials along the complete supply chain. Mannesmann Stainless Tubes have developed the use of Probably of Detection (POD) analysis to support Non Destructive Testing for additional quality validation of its seamless stainless and nickel alloy tubes and pipes.

By Dr Thomas Kremser, Mannesmann Stainless Tubes GmbH, Mülheim an der Ruhr, Germany.

Having extensive experience over nearly six decades, Mannesmann Stainless Tubes (MST) is committed to continuously improving its products through the latest NDT (Non-Destructive Testing) techniques. The validation of statistical evaluation of NDT data has become increasingly crucial, and now Probability of Detection (POD) analysis plays an essential role.

Customer preferences: a strategic choice

Nearly all sectors of industry are facing increasingly challenging product performance requirements. This has ledto the development of individual specifi-cations per product type, material grade, application, fabrication and installation. In the case of corrosion resistant alloy (CRA) seamless tubes and pipes – and especially in austenitic-ferritic materials such as duplex and super duplex – product specifications may include additional mechanical and metallurgical laboratory tests, as well as non-destructive testing (NDT). In some standards and specifications, NDT is either optional or not required at all.
In response to many end-users recommendations, Mannesmann Stainless Tubes (MST) have long since integrated NDT as a systematic control of products for key market segments, whether it is specified or not. Through its collaboration with end-users and industry partners, its acquired knowledge and experience is applied to the most suitable NDT method relative to the product application.

CRA Corrosion Resistant Alloy
DPT Dye Penetrant Test
ET Electric Test
FMEA Failure Mode and Effect Analysis
MT Magnetic Particle Test
NDT Non-Destructive Test
POD Probability of Detection
RT Radiography Test
UT Ultrasonic Test
VT Visual Test

Manufacturing process reliability

The reliability of the manufacturing process using both fundamental knowledge (R&D activities such as numerical simulation, advanced laboratory testing, in situ trials and training) and operational knowledge (manufacturing procedure qualification, risk analysis and action plans) is mandatory to monitor and define the correct parameters for the quality assurance of both processes and products.
Continuous improvement in MST’s manufacturing process through evolved Failure Mode and Effect Analysis (FEMA) techniques provides additional levels of assurance in the quality of the finished tubes and pipes. However, despite the reliability of such manufacturing processes to reach a point of complete product validation, with the detection of non-acceptable indications, further NDT procedures are used to provide customers with confidence in the final product.
In this context, understanding and characterizing the operational efficiency and sensitivity of the production mills NDT equipment is a pre-cursor for further continuous improvement. This also supports increasing requirements from both customers and specifications to provide statistical analysis data on the inspection process capabilities. POD in combination with PFA (false alarm rate) offers an important resource to give answers to these needs.

Selection of NDT technique critical

The range of NDT methods available is abundant, with the techniques often being complementary.
They can be divided into two families: volumetric control and surface control. The most commonly used volumetric methods are Ultrasonic testing (UT)
and Radiography testing (RT). The most commonly used surface methods besides Visual testing (VT) are the Electric test (ET), in our case, are Eddy Current
test, Magnetic Particle (MT) and Dye Penetrant (DPT).
For CRA seamless tubes and pipes, in addition to VT, ET and DPT, which would typically be considered near-surface product examinations, UT is specified as one of the most efficient industrial methods to validate soundness through the complete wall thickness of the product.
Ultrasounds are mechanical vibrations which propagate in solid or liquid media. The principle is to transmit an ultrasonic wave that propagates through the tested component and is reflected in the manner of an echo on the obstacles it encounters (anomalies, tested component boundaries). These waves are emitted by one or more sensors, each managed by an operator or via an automatic system. Today, advances in the miniaturization of connecting plugs, the segmentation of sensors and the improved speed of signal processing all combine to enhance the science of the production of the ultrasonic beams – leading to the technology that is ‘phased array’.
A phased array ultrasonic transducer contains several separate piezo-electric elements in a single housing, and phasing refers to how those elements are sequentially pulsed. A phased array system is generally based around a specialized ultrasonic transducer containing many individual elements which are pulsed separately in a programmed pattern. These transducers may be used with various types of wedges in either contact mode or immersion method. The shape of the sensors may be square, rectangular, or round, and the test frequencies are most commonly in the range from 1 to 10 MHz. Through optimization, the phase array technique allows the detection of very small indications, even when of an oblique configuration. The reason is that, by applying different delay laws generating the electronic deflection, it is possible to obtain a variation of incidence angles without mechanical adjustment of the sensor orientation. Each delay law provides a specific beam deviation which easily enables the detection of indications having different angle orientations with regard to the tube axis. Another valuable advantage is linked with balancing the translators through electronic compensation, ensuring a better onsistency/reproducibility in the control.
MST has chosen different UT benches in their various manufacturing locations to implement the POD analysis, for example see Figure 1.

Figure 1. Ultrasonic testing: semi immersion method while applying POD method.
Figure 1. Ultrasonic testing: semi immersion method while applying POD method.

Statistical analysis of results

After several test runs on precisely defined test tubes, a statistical analysis describes the potential parameters for optimizing the system.
Figure 2 shows an example of the results of a POD performed in one of MST’s mills. Five tubes with notch depths ranging from 0.05 mm to 0.15 mm were investigated. At the beginning, for reference, the standard tube with a notch depth of 0.10 mm was used for calibration. Following this setup, each tube was tested many times with defined UT bench parameters.
The figure shows the amplitude distribution for the investigated notch depths. The solid line represents a regression of the data points. The dashed line represents the 95% confidence bound on the regression line. The dotted line represents the 95% prediction bound, i.e. the assumed amplitude distribution in 95 out of 100 cases. Following this acquisition step, the data is processed and a Probability of Detection is calculated as a specific performance characteristic of the applied method for the given equipment and procedure.
In adopting this approach, the mills have identified the technical reasons for single variations and implemented corrective actions to continuously optimize them. Furthermore, a methodology for the quick evaluation of the POD results developed by SZMF (Salzgitter Mannesmann Forschung) has made the plants completely autonomous in the yearly evaluations that are performed. Following this successful implementation, the next focus for POD technique implementation are the Eddy Current and other continuously working NDT- systems.

Highly efficient technique

By continuous management using this method, MST obtains a full evaluation of its bench efficiency with regard to the applied specifications and limits. In addition, this approach is highly efficient compared to other techniques, which consistently lower the reference notch sizes regardless of their probability of detection in view of increasing the detectability of detrimental indications. Adopting this quality strategy has enabled MST to characterize the different types of indications associated with each product and process whilst also considering the needs and requirements of the ultimate end-use applications. From this, MST is able to define the best methods and techniques to be applied to each generic NDT requirement. In combination with its long-held philosophy of ensuring its NDT operators are trained to the highest industry certification standards, MST can meet its commitment to continuous improvement. This is reflected in the confidence of its customers.

Figure 2. Amplitude distribution for notch depths 0.05-0.15 mm (log. x-axis).
Figure 2. Amplitude distribution for notch depths 0.05-0.15 mm (log. x-axis).

About this Featured Story

This Featured Story first appeared in Stainless Steel World June 2022 magazine. To read many more articles like these on an (almost) monthly basis, subscribe to our magazine (available in print and digitally).

Want to contribute as author? Please contact Joanne.

Every week we share a new Featured Story with our Stainless Steel community. Join us and let’s share your Featured Story on Stainless Steel World online and in print.

Featured Story by James Chater

Previous articleBrown Metals meets customer’s requirements
Next articleNamibia reports the progress made with Hyphen
Stainless Steel World is part of The KCI Media Group, a group of companies focused on building and sustaining global communities in the flow control industries. We publish news on a daily basis and connect business-to-business professionals through our online communities, publications, conferences and exhibitions.