2nd International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, New Orleans May 2000.

EFFECT OF CLADDING ON INSPECTION RESULTS M. Bièth, L. Fabbri, J.L. Monjaret European Commission, JRC-IAM P.O. Box 2 1755 ZG Petten, The Netherlands Corresponding Author Contact: Email: bieth@jrc.nl, Web: http://www.jrc.nl

Abstract

The presence of a welded stainless steel cladding layer on carbon steel reactor vessel can seriously affect the ultrasonic inspection due to its coarse grained anisotropic structure and the surface irregularities, leading to beam skewing, scattering and mode conversion of ultrasonic waves.
This paper makes the assessment of the reliability of different inspection procedures based on several international round robin tests.
Blind trials conducted up to now show that the capability exists to detect underclad defects 5 mm or more through wall, with a 2 mm standard deviation for depth sizing.
Qualification of inspection procedures with pragmatic targets for defect detection and sizing is necessary to provide reliable information from field inspections for structural integrity assessment.

Introduction

Carbon steel components are cladded such as the reactor pressure vessel and other primary circuit components of nuclear power plants to protect them against the primary water corrosion.

Inspection procedures based on NDE are playing an important role in structural integrity assessment and this involves the knowledge of the defects present in the structure in the areas particularly subject to stresses induced by the pressurized thermal shock (PTS): sub clad defects of the planar type generally perpendicular to the surface (under clad cracks, weld cracks, lacks of fusion…), defects in the cladding and cladding defects.

Inspection systems are made of a mix of NDE techniques, setting or calibrating procedures, decision steps, scanning systems, recording and illustration tools and software’s. Moreover, inspection procedures often involve information interpretation of indications relying on the skills of the operator. Inspection procedures are thus not measurement techniques and their performances in defect detection; location, classification and sizing cannot be represented by simple figures and uncertainty margins.

To be able to use NDE inspection data in a structural integrity assessment reasoning or model, it is essential to know whether all the defects above a certain size were detected and reported, and the accuracy on the reported defect sizes in depth and in length and on the ligament, and finally the type of defect.

The present analysis aims at the answer to these questions in the specific case of the cladding inspection. It gives engineering trends, average values, which can be employed by the structural integrity engineer.

Collection of experimental data

Many reports are considered in this paper for the compilation of the NDE effectiveness data(1), particularly the following major NDE programmes:


• the three phases of PISC, including the parametric studies (2) (3)
• the DDT exercise in UK, (4)
• the NESC 1 project (5).

Defect families

The main types of defects susceptible to appear in the cladding area are summarized on tables 1 (manufacturing defects) and 2 (service-induced defects).

Defect type	Location of defect
Lack of bond/fusion	Clad-base metal interface
Inclusions	Between strip in strip cladding
Volumetric flaws	Clad-base metal interface
Cracks in cladding	Usually in the first layer of cladding
Crack in ferritic	Heat affected zone in ferritic steel. Parallel to the cladding interface
Table 1: Manufacturing defects





Type of flaw	Location of flaw
Loss of cladding thickness	Cladding
Surface-breaking transgranular Cracking	Clad to base metal interface and cladding
Surface-breaking intergranular Stress corrosion cracks	Clad to base metal interface and cladding
Table 2: Service induced defects

Despite of the fact that defects are of all kinds, the effectiveness of NDT refers mainly to planar type defects, generally perpendicular to the surface, with a tilt angle from -30 to +30 degrees. Such defects are correctly detected and sized when using specifically designed techniques or usual techniques set at high level of sensitivity. This condition of sensitivity leads to many indications and also to false calls.

Such defects are lacks of fusion and cracks (mechanical fatigue, thermal fatigue, corrosion, reheat, solidification).

Volumetric defects such as inclusions and porosities are also considered but more often as examples as the detection and sizing performance of NDE procedures is often better with such defects compared to planar ones.

Inspection procedures efect families

An inspection procedure based on NDE techniques is a process involving many (up to 22 in the PISC II exercise) individual techniques, decision steps, calibration rules, indication treatment, reporting guidelines…
The inspection procedures are also categorized in two groups corresponding to the families of inspection procedures.

• Usual industrial ultrasonic procedures giving often acceptable results (e.g. ASME XI,or with 20 or 10 % DAC). Such procedures are based on simple techniques: 0, 45, 60 degrees contact probes, 2 to 5 MHz, complemented with 70 degrees angle probes for near surface defects or with tandem probes for embedded defects.
• Advanced ultrasonic procedures adapted to sub clad defects detection, which are generally made of techniques adapted to the components and the defects to detect and which work at high sensitivity (noise level or 10 % DAC). They combine standard techniques (0, 45, 60, 70 degrees probes) in contact or focusing in immersion, with more advanced techniques or transducers such as, creeping waves, time of flight, phased arrays, eddy current probes, reconstruction algorithms… Such effective procedures are involving both detection and sizing of defects.
• Eddy current based procedures, which aims to detect defects (typically with a minimum size of 1 mm depth and 20 mm length, with a cladding thickness of 9 mm) on both cladding surfaces, but not in the parent metal. One of the most important parameters to take into account is the test frequency, this being closely related to the type of material to be inspected, defect type and location, and the depth of test penetration.

For usual reactor pressure vessels with wall thickness ranging from 100 to 250 mm it has to be admitted that no industrially applied NDE technique, even the advanced ones, can claim for better precision than + / - 1mm in the defect depth and length sizing.

Capability, effectiveness and reliability of an inspection system

Reliability of inspection procedures based on NDE techniques is made of three constitutive elements which are important to be defined correctly to understand the meaning of NDE evaluations and to use NDE reliability data correctly in view of structural integrity assessment.

R = reliability or total performance
IC =intrinsic capability of techniques / procedures
AP = application parameters
HF = human factors


The relation between these elements is summarized by the following formula:

R  =  f(IC ) - g( AP ) - h( HF )



The fact that the inspection is conducted on the right component or part of the component where defects can develop, is not considered here.

As human factor effects are not really predictable, discussion or use of NDE reliability data is misleading. What counts first is the knowledge of the effectiveness E of the inspection that could be defined as being

E = f( IC ) - g( AP )

A quality assurance programme, knowing that the operator effect could reduce the effectiveness to as little as zero in case of adverse situations, distraction or de-motivation must control the human factor.

This paper considers thus mainly inspection effectiveness and not inspection reliability. The inspection effectiveness is, in turn, considered stressing the value and limits of the intrinsic capability of the techniques. The limiting function g( AP ) of the parameters of the application such as access, surface roughness depends of the given situation.

Under-clad defects inspection effectiveness

The most relevant reference to this category of components is the PISC programme. Several other programmes were conducted but either produced results of lesser statistical significance or generally confirmed or anticipated the PISC results.
From PISC, conclusions were drawn on the detection, location and sizing capability:

• it was concluded that inspection procedures based on ultrasonic can meet the inspection objectives set in the frame of safety assessments. For planar defects of 10% of the wall thickness the expected detection rate is around 95%.
• sizing by ultrasonic procedures used after qualification, can lead to oversizing in depth (e.g. of 3 mm) but with a standard deviation of 5 mm, and errors on the length measurement of 3mm with a standard deviation of 10 mm.
Certain conditions have to be filled like the correct selection of techniques, correct access, and the necessary qualification programme.
For detection, a typical diagramme of PISC II, obtained in a very pragmatic manner but confirmed by several models, shows, for different types of defects, the difference of detection performance of ultrasonic procedures of good industrial practice type (Figure 1).

Fig 1: Detection probability of ASME-type procedures with recording level at 20% DAC as a function of the defect through wall size for the three categories of defects. A): smooth, planar for ultrasonic wavelength, sharp crack edges; B): hybrid defects or rough defects like hot tears; C): volumetric defects.




Inspection of the sub clad area with specialized procedures
Few blind tests results relevant with the under clad defects theme and suitable inspection procedures were produced up to now by RRTs conducted by independent institutions: PISC has included several near surface defects in the assemblies. However, either, the inspection procedures did not specifically address the sub clad defect in the context of PTS, or the statistical significance of results is not established: their evaluation was not conducted in the suitable context or only on very few sub clad defects.
The NESC exercise (5) addresses the case of the PTS type defects detection and sizing but the RRT involving specific inspection procedures was conducted on 20 defects only.
Very safe detection is reached for defects of a size in depth of about 10 mm for perpendicular planar defects using specific and advanced inspection techniques.
Information resulting from qualification programmes and trials in France, Sweden, USA, and Germany for European and Eastern plants, can be used as well.
Based on the above source of information, it is anticipated that inspection procedures are or will be qualified and applied along the new qualification requirements in several countries for the safe detection of sub clad defects as small as 5mm in depth and for sizing with uncertainties limited to few mm.

Inspection of the cladding
Again, few neutral exercises addressed the theme of cladding inspection in itself. PISC plates contained some defects in the cladding: porosities, few cracks, disbonding of the cladding.

Due to the frequent rough surface of the cladding, inspection of the cladding itself is generally ineffective. Very specific techniques show good detection performance if the clad surface is machined or ground. Such inspection techniques are the ones that can be used for thin walled cast austenitic components or welds.

In ideal conditions, defects of the order of 50% of the clad thickness are well detected. Sizing capability should not be considered.

Influential parameters

The reliability of inspection procedures depends heavily on the influential parameters:

Correct classification of defects
The results of RRTs, when considering the classification of defects, as a performance variable were generally not convincing. Results of good procedures indicate that the correct sequence to follow should be: detection - classification - sizing. For heavy section components, the use of specific techniques, on request, improved the differentiation between planar and volumetric flaws.

Defect characteristics
Several laboratory exercises were conducted to evaluate systematically the influence of defects variables like: class, surface roughness, aspect ratio, tilt and skew angles for planar cracks, position in depth,… Figure 1 is an example of the importance of the parameter ‘family of defect’, figure 2 demonstrates important changes of response in amplitude of the same planar defect as a function of the tilt and skew angles compared to the central plan of the weld.


Fig 2: Combined effect on tilt and skew angles on the defect response in amplitude, for smooth planar cracks with sharp crack tips in the vessel wall.

Defect location
Obviously, the knowledge of the ligament is essential to perform an integrity analysis. Performant NDE techniques can evaluate such values in the same way that they size a defect: by the detection and location of a crack tip. Ligament measurement capabilities are thus corresponding to the ones of capable sizing techniques in heavy section steel component.

Importance of the electronic equipment characteristics
Within the PISC II exercise, a parametric study has been performed aiming to evaluate the effect of a rather large number of variable parameters of the ultrasonic system: transducer electrical characteristics, cable parameters, ultrasonic equipment characteristics. Each of these parameters has been made to vary individually and independently from the others. The effect of these variations was measured on the defect detection performance, the sizing performance.

Conclusions were that most of the parameter influences are dominated or reduced by the calibration of the technique if done properly. Variations measured, after calibration, using the amplitude of response as variable are generally less than 6 dB.

Importance of the cladding
The effects of the cladding characteristics on the inspection performance of underclad zones were measured in PISC II (6). The importance of these effects depends on the type of cladding and on the NDE techniques used.

Ultrasonic techniques can find important disturbances created by the complex structureexture represented by the cladding strips or wires and heat affected zones. The back-wall echo, used as a measure of the attenuation, has shown in some cases of wrong selection of transducer frequencies, up to 30 dB drops in locations corresponding to the overlay of passes (channels). The figure 3 illustrates this influence. Beam distortions are also introducing important sizing errors or even false calls. Such local variations have to be considered in qualification plans.

Fig 3: Combined effect on tilt and skew angles on the defect response in amplitude, for smooth planar cracks with sharp crack tips in the vessel wall.

The PISC II parametric study concluded on this subject that the existence of a stainless steel cladding may have important effects on the beams of ultrasonic transducers, local variations of amplitude of 30 dB being observed at 2 MHz on a plane infinite reflector with a shear wave contact probe working in emission-reception.

Very tight defects
When defects are very tight with contacts between the faces, UT waves can propagate through the defect. The effect of compressive stresses on planar defect response to ultrasonic testing can lead to loss of signal amplitude, which is in some cases unacceptable for industrial inspection procedures.

Importance of human factors
Action 7 of PISC II analyzed some human factors that influence the inspection results. Conclusions were detailed in specific reports. Human errors were encountered at different levels of the inspection process during calibration, during scanning, by miss-observation of the screen, and during evaluation. These observations on the importance of human factors were noted in several exercises including the one in NESC and suggest that similar events can happen in the reality of industrial inspections.

Conclusions

Blind trials conducted up to now by independent organizations on thick-wall components show that the capability exists to detect underclad defects 5 mm or more in depth. Sizing errors in depth will always produce a standard deviation of about 2 mm. These results correspond to the effectiveness of several procedures desired for the purpose of inspecting cladding. However, qualification of inspection procedures via pragmatic targets for defect detection and sizing is seen as the only way of qualifying inspection performance and providing the structural integrity engineer with reliable information from field inspections.
The qualification process means:
• Defining targets resulting from safety objectives that are compatible with NDE capabilities,
• Executing the qualification with technical justification and open trials and not just as a simple verification exercise,
• Complementing the inspection procedure with a QA programme in order to maintain the performance level as during the qualification.

References

1. S. Crutzen, F. Franck, L. Fabbri, P. Lemaitre, C. Schneider, P. Visser, “SINTAP: compilation of NDE effectiveness data”
2. PISC II: R.W. Nichols, S. Crutzen, “Ultrasonic inspection of heavy section steel components” The PISC II final report, Elsevier Applied Science, Barking UK, ISBN 1-85116-155-7, 1988
3. PISC III: “Evaluation of the sizing results of 12 flaws of the full scale vessel installation”, PISC III report No 26 – Action 2 – Phase 1, JRC report: EUR 15371 EN, 1993
4. B. Watkins, R.E Ervine, K.J. Cowburn, “The UKAEA defect detection trial”, UKAEA DDT Symposium, Silver Birch Conference Centre, Birchwood, Warrington, UK, 7-8 Oct. 1982
5. B. Eriksen and all, “Inspection results obtained for he International NESC I project”,
   2^nd International Conference on NDE in relation to Structural Integrity for Nuclear and Pressurized Components, New Orleans, 24-26 May 2000
6. P. Benoist, PISC II Parametric studies on the effect of the cladding characteristics. Proceedings of a seminar on Reactor Safety Research, Contribution of the EC. EUR 12343 EN

© NDT.net - info@ndt.net

|Top|