Webinar: Watch now

Development of Effective Crack Management Strategies

This webinar discusses some of the options available for assessing cracks and demonstrate the importance of having comprehensive and accurate input data to support the assessment.

 

Abstract

Asset Management of existing infrastructure has become increasingly important as ageing makes the infrastructure increasingly vulnerable to failure and associated shut downs. Managing the long term integrity of pipelines that are susceptible to cracking requires a holistic strategy supported by experts within the fields of fracture mechanics, materials, welding and inspection.

During the planning stages of the inspection, it is important to establish the type(s) of cracking that the pipeline is susceptible to and what constitutes a critical crack size. The inspection strategy then needs to be defined considering the information available for a particular pipeline system including its operational history. The combination of in-line inspection technologies selected, the incorporation of historic inspection data and the adapted data analysis strategy allow for an effective and transparent inspection process.

Ultrasonic in-line inspection has proven to be the appropriate technology to address cracking in pipelines. Additional technologies like circumferential MFL provide valuable information about non crack like features and therefore substantially increase the correct identification of cracking.

Results from field verification or pipeline repair activities contribute to increase the confidence in the reported ILI results and are the basis for an interpolation of the ILI data to the entire pipeline.

Accurate and reliable inspection data forms a critical part of an overall crack management strategy, but this must be supported by a range of other integrity management activities, both before and after the inspection, in order to ensure a successful inspection campaign. Once the inspection data is available an appropriate assessment method should be selected to assess the significance of any reported cracks and prioritize them for in-field verification and investigation.

Presenter

Bob Andrews: Principal Consultant, Structural Integrity

Bob Andrews is a Principal Consultant at MACAW Engineering, part of the ROSEN Group, having previously worked at BMT Fleet Technology, British Gas research and The Welding Institute. He holds BEng and PhD degrees in Mechanical Engineering from the University of Sheffield. He has over 35 years’ experience in fracture, fatigue and structural integrity and is a Chartered Engineer and a Fellow of the Institution of Mechanical Engineers.

Thomas Beuker: Head of Businessline Advanced Pipeline Diagnostics

Thomas Beuker started with the R&D department at ROSEN in Germany. His expertise lies in data analysis, finite element modelling and in the development of new sensor technologies. He has more than 20 years experience in developing pipeline inspection solutions. Mr. Beuker holds a MSc in Geophysics from the University of Muenster, Germany.

Q&A Summary

1. RoCD Framework

Is ROCD a new service? What is ROSEN's track record for EMAT?

RoCD is not a new service per se. RoCD is a framework which allows a structured discussion of the needs and the individual measures to be taken by all parties involved to obtain optimum results when using in-line inspection systems and other related services, such as crack assessment or in the ditch validation. As of today, ROSEN has inspected approx. 45,000km (28,000mi).

Is the framework you presented also applicable for UT crack detection?

Yes, the framework is not limited to a specific ILI technology. UT-C, EMAT-C and MFL-C are all technologies that can be integrated into the framework as deemed necessary.

Rare events, is that not a topic handled through risk management processes?

In the context of this presentation, rare events was referenced with regards to the need to adapt or change the existing management system to address rare events. This has already been done by many operators in the recent past. The point was also made to emphasize how important it is that the ILI provider understand these changes and adapt their management systems.

How do you set expected tool confidence levels if no previous history is available, as in the case of a subsea pipeline?

In this case a valid model must be set up. This is also not unusual for onshore pipelines, since accessibility is not only an issue for subsea pipelines. Such a model can be derived, for example, from pre-inspection testing. This is also valid for toughness investigations.

Does ROSEN look into analyzing and reporting cracks associated with third party damage? Knowing that the crack tool is able to detect even shallow dents – that is valuable information for the operator.

Denting of the pipeline, e.g. due to third party damage, can be characterized by high resolution geometry ILI systems like the RoGeo XT service. The crack detection sensors on the ILI tools are supported by suspension systems that guide the sensors accurately across the dents.

How would you inspect for and assess cracks on a subsea structural member? As ILI the member cannot be found from ILI. What solution can you propose for ROV crack inspection?

EMAT-C and UT-C inspection technologies can be used to detect and size cracks for a variety of engineering structures. Handheld devices and robotic manipulators have been developed that can be used as a carrier in place of in-line inspection tools.

Is there a "trigger" crack or family of cracks in the preliminary analysis process where ROSEN will call the operator with a warning that this particular anomaly may require immediate action?

A "trigger" dimension for crack-like anomalies is set during the system selection module. If this "trigger" is exceeded, an immediate response is sent to the operator. This information is submitted right away, irrespective of the timing for any formal reporting.

Do you agree with Kiefner that hydrostatic testing at pressures near 100% SMYS reduce the probability of pressure reversals?

The selection of the hydrostatic test pressure is a balancing act between using high pressure to remove many defects (or equivalently demonstrate a large margin for future sub-critical growth) and using low pressure to minimize possible ductile tearing leading to reversals. Pressures around 100% probably represent a good compromise.

Does ROSEN's current fleet of ultrasonic crack technology and existing analysis capabilities allow for more confident feature type characterization of metal loss clusters with sharp edges or gouging from crack colonies?

The differentiation between sharp-edged corrosion and crack colonies is difficult, but it can be assessed and managed thru the RoCD framework. The differentiation between gouging and cracking is generally working well. In all cases, the combination of multiple ILI technologies in the inspection process is very beneficial.

What are the primary causes of crack formation in Europe - SCC or fatigue? What is the general state of cracking in seam welds - are we seeing any as the pipeline stock ages?

It's difficult to get hard data on this; unfortunately neither EGIG nor CONCAWE give sufficient detail on the causes of cracking in their reports. SCC generally is less common in Europe than in North America. So far we have not seen cracking in seam welds in Europe, although as you say the pipeline stock is ageing. However, I suspect that the stock has not yet accumulated sufficient fatigue damage for cracking to be detectable.

2. System Selection

Is it necessary to qualify crack tools for every inspection project?

No. As explained during the presentation, it is possible to use UT-C or EMAT-C systems without an explicit qualification program. The decision on the need for additional qualification will be made during the system selection module.  At this stage a gap analysis is run on the information available for a particular pipeline or pipeline system. An assessment is conducted determining the acceptable thresholds for the inspection.

How do you differentiate between cracks and mill features or laminations?

Multiple ILI technologies can be utilized within the RoCD framework for the inspection of one particular pipeline. MFL and ultrasonic technologies are complementary to each other and provide the analyst with different specific signal parameters to differentiate between cracks, laminations and mill features.  In some cases where similarities between the feature types are known in advance, it is possible to conduct pull tests in advance of the inspection to develop a more detailed understanding of a particular situation.

Can you explain verification and calibration as part of the system selection process?

The need and the opportunity for in the ditch verification as part of the project is discussed during system selection. Pipeline specific information and the history of the pipeline may already suggest a more stringent verification program. In other cases, where RoCD has already been utilized, the confidence level is already high enough to focus on pipeline repair activities only if required at all. The term calibration refers to feedback received from the field or from other sources, which justifies a re-assessment of the reported features. This is typically done when initial assessment was excessively conservative.

Does your company use 360 º testing and how is are the results utilized at ROSEN?

Yes, 360º testing in the ditch refers to an examination of the entire pipe joint from girth weld to girth weld. The entire joint is examined and documented with regards to cathodic protection, soil condition, coating condition, any anomalies identified thru MPI, pipe body and longseam. Any anomaly found (corrosion, mill related features and cracking) is further examined and documented. The results are fed back to data analysis to conduct a refinement of the reporting and to close out a project.

Can the API 579 method be used for both above ground and underground pipelines?

Yes, the API 579 (and BS 7910) methods can be used for both cases. You should adjust the input data, for example to take into account local stress concentrations at supports in an above-ground section. The different axial restraint conditions in an above-ground pipeline should also be included.

If an operator does not have detailed material properties for a pipeline, such as fracture toughness, how can a final report be generated with reasonable conclusions?

In the absence of detailed materials data, the conclusions of the report will have to be based on the best available data. The conclusions will still have some relative validity. If material becomes available from cutouts this can be tested to improve the inputs and the analyses.

Should fracture toughness be determined prior to committing to running crack tools?

Ideally a lower bound fracture toughness should be determined in advance, as this will allow the best assessment and prioritization of the ILI results. If this is not done, a fracture analysis of indications from ILI with an assumed toughness would still give relative priorities for investigation or repair.

3. In-Line Inspection

What is the performance specification of the crack inspection?

A performance specification for the individual technologies EMAT-C, UT-C or MFL-C can be provided upon request. The main benefit of using the RoCD framework or customized parts of it is an increase in statistical certainty – an increased statistical confidence level

What are the strengths and weaknesses of each MFL-C or EMAT-C?

The strength of EMAT-C and MFL-C lay in the combination of both complementary technologies. While the MFL-C technology is sensitive to volumetric features, the EMAT-C technology is very sensitive to any planar reflector. A weakness of the MFL-C technology is the limited capability to detect cracks, without a volumetric component. This is compensated by EMAT-C which picks up these features very well. This differing responses from each of the technologies allows for a better characterization of detected features.

What is POI for cracks at girth weld compared to cracks at pipe body?

The specified POI of axial cracks at the girthweld is the same as in the pipe body. However, the measurement is affected when the sensor passes across the girth weld; therefore the crack must have an extension of at least 50mm into the HAZ or base material.

Any solution to addressing circumferential SCC cracks as well as longitudinal in the same pipeline?

The technology presented today is capable of addressing longitudinal cracking. For the detection of circumferential cracking, the same technologies – whether EMAT, UT or MFL – can be used; however, the scanning direction of the sensors needs to be changed. MFL and UT tools are already available for the detection of circumferential anomalies. Implementation of the EMAT technology for circumferential features is currently under development at ROSEN.

What pipeline diameters are suitable for RoCD?

In-line inspection is an integral part of the RoCD framework. The various ILI technologies range from 4" to 56".

Can you discuss tool tolerance? How does tool tolerance play into predicted bust pressures? Is tool tolerance important to both preliminary and final analysis?

Tool tolerance is documented in the performance specification and in some case derived directly from a field verification program. The modules crack prioritization and crack assessment modules take tool tolerances into account and calculate with the more conservative values. In cases where quantitative risk assessments are conducted, the values from the actual findings are used. In this case, the tolerance is reflected in the calculation of the corresponding probabilities.

4. Crack Assessment

You discussed estimating the fracture toughness from Charpy energy – what are the alternative methods?

The only alternative to estimation is fracture toughness testing on material removed from the pipeline during repairs or diversions. This could be done during a repair/replacement or any time a cut out is made.

You showed a slide where assessments of the same defect using different codes gave answers of safe or potentially unsafe. What are the reasons for this large discrepancy?

There are two contributors to this difference. The greatest contribution, probably about 2/3, is the change from using specified minimum strengths in the BS 7910 analysis to measured strengths in the API 579 analysis. In this case there was a large difference between the values. This particularly affects the calculation of the load ratio; increasing the yield strength moves the assessment point to the left. The remainder of the difference is due to differences between the codes' methods of calculating the stress intensity factor and the reference stress.

How do you deal with required safety factors in correlation with allowable crack length/ depth, and how do you apply the PoD as used?

In principle safety factors can be applied to the inputs, although this is not normally done with API 579 or BS 7910 assessments if the recommended inputs are used, as the basic recommendations are intended to be conservative. Annex K of BS 7910 contains partial safety factors which could be used with mean or best estimate inputs, although we have not done this.

On your slide you inferred that you can model or predict crack growth due to sour corrosion. Is my understanding correct?

One slide showed a summary of recent published estimates of SCC growth rates; there has also been extensive laboratory work done to model SCC, but I believe the results are not yet conclusive. The slide was summarizing external SCC due to the environment. The last slide showed some measured data for sour crude (i.e. the internal environment) under fatigue loading. That particular data set is now quite dated; more data on the effects of sour environments has become publicly available in recent years. At present it is only possible to draw very general conclusions, so project-specific testing is often required unless your environment is a good match to a published case.

Regarding Charpy toughness used for crack assessment: pipe body or pipe seam weld for welded pipe? Typically we see a much lower toughness value for seam welds.

Yes, the seam weld Charpy energy and fracture toughness is typically lower than for the parent pipe, particularly in gas transmission pipelines where a high Charpy energy is often specified to ensure crack arrest in the pipe body. The appropriate value should be used if the ILI technology can identify the crack location relative to the seam weld. If the seam weld position cannot be identified (as in ERW or HFW pipe with a well-trimmed flange) then a conservative assumption will have to be made.

What standard do you recommend for cracks?

All things considered, I would recommend API 579 for cracks in pipelines as the stress intensity factor and reference stress solutions are better than those in BS 7910. There are some aspects of BS 7910 which are better such as the allowance for mechanical relaxation of secondary stress.

Have you found evidence of crack retardation?

I assume you are referring to retardation of crack growth due to high stress cycles, in particular for pipelines subject to hydrostest. I am not aware of crack retardation having been found in the field in pipelines, although it has been demonstrated in the laboratory and found in other applications. For fatigue there are more complex crack growth laws (such as Forman) than the Paris law shown in the last slide of the webinar which include retardation effects, but there is no data available for using these with pipeline steels.

Is it good practice to assume that corrosion flaws are like cracks in controlling final failure of the corrosion?

Volumetric corrosion flaws can be treated as cracks but the result will be a conservative assessment. It is better to analyze volumetric corrosion flaws using specific methods such as ASME B31G, RSTRENG or DNV RP-F101. However, cracks (whether SCC or fatigue) must be analyzed using fracture mechanics methods .

Given the uncertainty of SCC growth rates, how can the fracture mechanics approach be used to define remaining life and inspection periods?

At present only on the basis of upper bound rates, unless these can be refined by repeated in-ditch inspections or examination of crack faces from hydrotest failures or cutouts.

Is the Charpy test a good test for assuming the material's toughness?

The best test is fracture toughness testing to generate CTOD or J-integral values. Charpy impact energy can only be used to estimate the fracture toughness through correlations.

5. Field Verification and Rehabilitation

Does ROSEN believe that field NDE is sufficiently accurate to validate crack ILI? Do cutout and destructive testing have a place?

There was an exhaustive panel discussion around this topic at the last PPIM 2016 in Houston. The performance of NDE methods must be validated as with any other inspection technology. In cases where NDE tolerances have been derived, the NDE results are always integrated into the RoCD framework. Cutouts and destructive testing are the most reliable assessments of anomalies and are always appreciated. If cutouts are available, it is also recommended to determine the fracture toughness of the samples.

Do you report the false positives and false negatives to the customer?

The results from field verifications are consulted for the assessment of false positives and false negatives . Experience has shown so far that false negatives can be excluded in particular for critical crack sizes. The same procedure is used to reduce the number of false positives to an acceptable minimum.

What kinds of crack can be repaired?

The decision whether a certain crack or crack type must be repaired is based on the crack assessment. There several methodologies available to repair cracks. Whether a crack can be repaired or not depends mainly on the size of the crack. One of the most common methods is buffing of the cracks from the pipeline surface.