Table of Contents ECNDT '98
Session: Chemical, Petrochemical
The Need for Reliable NDT Measurements in Plant Management SystemsJ. Verkooijen
AEA Sonomatic BV, Krombraak 15, 4906 CR Oosterhout, The Netherlands.
Corresponding Author Contact:
|TABLE OF CONTENTS|
During recent years, much effort has been put into the development of Plant Life Management systems, to improve the effective operation. Although the introduction of Quality and Management Systems to capital intensive industries has led to major improvements in the reliability of plant as well as their availability, these systems can only be as good as the data used as input for these systems.
Optimisation of maintenance and inspection is shifting from certification of plant to verification of true condition, by means of systems such as Reliability Centered Maintenance (RCM) and Risk Based Inspection (RBI). They are aimed upon putting inspection and maintenance effort in these parts of the plant where this is needed, and not at the locations where this will give a minimal return. The assumption however is, that the inspection techniques are capable of detecting these defects that matter for the optimum functioning of plant, but also with an adequate accuracy, such that the required safety levels can be maintained and the remaining life time and interval extension calculations are optimised in accuracy.
In this paper examples are given of the methodology used by the authors' organisation to use techniques developed over the past years to achieve the above goals. Examples are given of a validation programme for the nuclear industry, whereby the performance of the Ultrasonic Time-Of-Flight Diffraction (TOFD) technique was independently demonstrated to be capable of achieving the required Probability of Detection (POD), False Call Rate (FCR) and accuracy for In Service Inspections (ISI) of piping welds, and a validation exercise for ISI of offshore piping, whereby the technique was again independently validated against destructive testing results and subsequently used during production to determine the optimum maintenance intervals.
Key to the success of the examples described is the methodology followed, whereby validation of the technique performance in detection, false call rate and accuracy of the techniques is proceduralised within an established quality system, including specific personnel training.
The lifetime and inspection intervals of most industrial plant is calculated based upon design data such as theoretical temperatures, pressures and material properties. During the whole lifecycle plant is inspected at regular intervals, in general to fulfil the demands by the certifying authority, the well known process of certification. Ideally however, these inspections should be used such as is the case with temperature and pressures, ie to verify the effects of the initially introduced parameters, such that amendments to the inspection interval and lifetime can be possible should the assumptions prove to be conservative.
In most cases convential technique are, based on the historical availability, introduced in the inspection specifications.The advantages of techniques such as radiography and manual ultrasonic examination are well-know and generally accepted . These techniques have however limited reliability and accuracy. Today more reliable and more accurate techniques are available that could fulfil the roll of a verification system much better such as mechanised ultrasonic examination and Time of Flight Diffraction (TOFD). Less then a decade ago these systems could be called expensive, cumbersome and inflexible but in the present time of ever faster computers and data acquisition systems combined with the increasing acceptance of these systems in specifications, these prejudices can no longer be held up.
One of the techniques used by AEA Sonomatic BV Sonomatic is Time of Flight Diffraction (TOFD).
On both sides of the weld an ultrasonic probe is positioned. One acts as an emitter of ultrasound, the other as a receiver. The longitudinal sound beam can encounter obstacles on its path, which cause reflected and diffracted signals. When the probes are moved parallel along the weld, the resultant waveforms are digitised, stored on hard-disk and displayed on the video-screen as a grey scale image. The image build up is in effect a through sectional view of the weld examined and can be used for accurate sizing and monitoring of indications.
Seescan (Micromap) combines conventional ultrasonic practice with the technology of miniature CCD (Charged Coupled Device) video cameras to provide a simple yet highly effective method of recording and presenting ultrasonic data during corrosion and erosion inspection applications. A brief description of these principles follows:
|Inspection setup Seescan (Micromap)|
on the specimen, the computer in Seescan constantly reviews the video image looking for the bright spot generated by the lamp. This bright spot is given a X-Y co-ordinate and shown on the real time video display. Simultaneously Seescan reviews a synchronised ultrasonic signal and converts the ultrasonic reading (thickness measurement) to an assigned colour. All this information is displayed on the video monitor as a composite colour image of the area examined.
The main advantage of the approach is the much higher reproduceability and accuracy for the following reasons:
The reliability of inspection techniques can be defined by determining the product of the probability of detection and the reciprocal value of the False Call Rate (1-FCR). During the NIL/KINT thin-plate project, the following values were derived for different inspection techniques as a result of examinations of welded plates in a thickness range of 6 to 15mm.
All plates were examined with the techniques shown here and later the results were verified by destructive testing. From these results it follows that manual pulse-echo ultrasonic examination will detect slightly over 50% of the defects present. It is also shown that in about 23% of the cases defects are reported where no defects could be confirmed (FCR).
In addition it is confirmed that X-ray techniques are more sensitive than Gamma-Ray techniques. However for this improved probability of detection one has to pay with a higher False Call Rate (FCR) which results in a reliability figure for X-ray which is only 2% higher then that for Gamma-Ray.
Mechanised Pulse Echo examination based upon line-scanning, whereby several probes are moved in parallel to the weld at different distances to the weld center line delivers a reliability comparable with the Radiographic techniques.
Two techniques however deliver a much higher probability of detection and reliability : TOFD and Pulse Echo Meanderscanning. For the examination during the project the AEA Sonomatic Microplus system was used. It was proven that the Mechanised Pulse Echo Meander Technique results in the highest probability of detection but TOFD delivers the lowest False Call Rate. The reliability derived from these figures is therefore only 1% different. However one has to bear in mind, that the time required for the examination with Pulse Echo meanderscanning, including the subsequent analysis of the data and the reporting thereof, took much more time than the same for the TOFD technique. As a result the costs for this type of examination will in most cases be a factor 10 higher then for Time of Flight Diffraction. The mean reason for the good score of the Pulse Echo Meanderscan is, that the whole weldvolume is examined with several angle probes (not only directed at the weldpreparation) such that defects of many different orientations are detected, this in contrast with linescanning, whereby a compromise is sought between the number of probes that would have to be used to look at all depths with all different angles of incidence and the number of probes that can be used within the physical and economical constraints. For TOFD of course, the main reason for the excellent score, is the fact that the existence of diffraction signals is almost independent of the orientation of the defect, such that in one linescan the whole weldvolume could be examined.
Using the figures derived from the NIL/KINT project, it becomes possible to quantify the risk of a implant shutdown or the failing of a component much better. If we by example assume that in the construction a certain chance of having an unacceptable defect exist (P1), than P1 times the number of welds, will have the potential of leading to a unplanned shutdown . Whether such a defect is detected depends on the probability of detection of the technique used.
A defect which is not acceptable to the code, of course does not necessarily always lead to a unplanned shutdown. Therefore another chance P2 is introduced that is defined as the chance a defect present really leads to an unplanned shutdown or failure. This chance is dependant upon the actual material properties and the actual operating parameters, but also depends on the actual size of the defect. P2 can be derived from Fracture Mechanics calculations, however the accuracy of this calculation is mainly determined by the accuracy with which the dimensions of the defect can be determined. In the following example the figures from the table have been used to calculate the probability of failure of the construction. A comparison is made between the different techniques .
|POD||FCR||Reliability||Probability of Failure|
|Probability of defect present :||5%
||Probability of defect leading to failure:
||Calculations based upon maximum 2 repairs
From these calculations it can be seen that the probability of failure or the probability of a unplanned shutdown can be brought down with at least a factor 2 to 3 in comparison with the convential techniques. No inspection at all would lead to a 5 times bigger chance of failure in the example chosen . Offcourse the actual figures of the examples should be adapted to the actual situation that is under consideration. AEA Sonomatic has software available to assess these risks.
This spread is important if we think of the further application of these results, by example to determine the remaining lifetime of an installation or as a basis for inspection interval extension programs. An example of this is a program to measure corrosion in the root of flowline welds. At certain intervals the remaining ligament of the weld is measured. The measurements at different moments in time results in a corrosion-rate. From this it can be derived when a certain minimum value would be reached. When this value is reached, the weld needs to be replaced or repaired. It is evident that the smaller the spread in the accuracy of the measurement the more accurate the replacement time can be calculated.
In the calculation of the remaining lifetime another important factor needs to be know, which is the stress present in the material, primaraly near the welds. Since some time AEA Sonomatic has a system available called MAPS, which has the capability of measuring stresses in materials in a non-destructive way.
|Example of biaxial calibrations derived theoretically from experimental uniaxial data, shown as MAPS screen images for (a) first impedance parameter, (b) second impedance parameter and (c) anisotropy parameter. Contour levels indicate parameter value at all stress states between -300 and +300 MPa|
|Maps screen image of biaxial stress distribution measured over surface of 25mm 1Cr 0.5Mo steel plate butt welded constraint. Arrows indicate sign, magnitude and direction of principal stress level weld position|
The equipment has been developed by AEA Technology at Culham (formally Harwell). Together with the accurately determined defects sizes, this can be used to improve the accuracy of the remaining lifetime calculations, by using the actual stress values, rather then the values derived from design parameters.
One of the things different in the NDE-industry to the established practice in the welding world, is validation. When welds are made, validation is very normal, when a fabricator of vessels wants to introduce a new weldingprocess , than the procedure is validated by determination of the sample's strength, hardness and toughness of the material. If these resulting values are within the specification then the welding procedure and therefore the new process will be accepted. In Non Destructive Examiniation this is generally not possible. One exception is the nuclear code, Asme XI and particularly appendix 8. In 1993 AEA Sonomatic BV went through a validation at the Electric Power Research Institute (EPRI) in Charlotte (NC). It was proven here that all defects present in a very number of testpieces were detected and also the sizes were determined within very accurate tolerances. In addition no false calls were reported. The result is that AEA Sonomatic BV was qualified, in accordance with appendix 8, as the first company in the world for the inspection of nuclear piping welds, which has led to an increased use of Time of Flight Diffraction for inservice-inspection of piping welds.
AEA Sonomatic has introduced a in-house qualification system in 1988. All operators are trained and qualified in accordance with this system, which is based upon ASNT-TC-1a, modified for TOFD. During 1996 and 1997, on behalf of AEA Technology's NDT group, the inspection validation centre (IVC) has developed an improved training- and certification system for TOFD, which is based on indepent examination and certification.
Using the probabilistic method, described earlier, it is possible to determine the total expected inspection costs for a weld, as illustrated in the following table.
|POD||FCR||Inspection price||Total expected inspection costs|
||Probability of defect present :
||Probability of defect leading to failure:
||Price of repair:
||Calculations based upon maximum 2 repairs
AEA Sonomatic has software available to calculate the total expected inspection costs for many cases, based upon the actual conditions. Dependent upon the actual conditions the inspection costs can differ, but also the consequences of failure can be different. From the results a adequate advice can be derived to determine the inspection strategy to be followed.
During the last years , TOFD has been introduced in various specifications . In 1993 BS7706 has been introduced, in which the method is described, followed by EN583pt6 which has been accepted as a pre-standard now. TOFD is also specified as a possible NDT technique in BS4515 for piping and in the European Pressure Equipment Directive. Also it is possible to apply TOFD in accordance with Asme XI appendix 8 as described, and TOFD is acceptable for wallthicknesses from 4" in accordance with Asme I and II, based upon the code case 2235 accepted early 1997. A new code case will be filed shortly to decrease the wall thickness to 1". A recent NIL-Kint project in the Netherlands has resulted in acceptance criteria for TOFD based upon good workmanship. This document will be introduced in the Dutch Stoomwezen rules and will also be proposed as a draft EN norm. Also initiatives have been developed to introduce TOFD in BS5500. In practise many thousands of welds have been examined with TOFD, in accordance with Stoomwezen and Tüv.
In conclusion, when TOFD is used within a controlled system as described, the use of TOFD leads to results at least equivalent to conventional techniques, and in many cases much better and more economically. TOFD is gaining acceptance rapidly, and is introduced in many standards already. Within the fully auditable system described, TOFD can verify plant condition, which through the high accuracy of the measurements can lead to a 5 times higher safety compared to no inspection or a 2 to 3 times higher safety when compared to convential techniques. Also it has been shown that the highest inspection price (per hour or per unit) will certainly not always lead to the highest inspection costs.