NOPIG A NEW TECHNIQUE IN PIPELINE INTEGRITY NOPIG UNE NOUVELLE TECHNIQUE DE DÉTERMINATION DE L INTÉGRITÉ DES PIPELINES

Transcription

1 NOPIG A NEW TECHNIQUE IN PIPELINE INTEGRITY G.G.J. Achterbosch, Gasunie Research, The Netherlands M. Broderick, N.P. Inspections Services GmbH, Germany SUMMARY Worldwide pipeline operators are searching for alternatives for intelligent pigging or hydro testing to prove integrity of their pipelines. N.P. Inspections Services GmbH offers an above ground inspection technique called NoPig to detect metal loss. In the last three years Gastransport Services was involved in investigations by means of projects but also by actual surveys on pipelines to determine the performance of the instrument in laboratory and field conditions. The investigations show a large progress in performance. It has been shown that corrosion defects that are critical for pipelines up to 12 can be detected when taking the specifications of the instrument into account. The paper describes the measurement principle of the NoPig technique and the main characteristics of the performance of the instrument as determined in the investigations and measurements. Furthermore, attention is given to the interaction of instrument performance, corrosion distribution based on pigrun data and the tolerances in defect assessment. Finally the possible use of the NoPig system in an inspection program for pipelines is discussed. The overall conclusion is that the NoPig inspection technique can offer a valuable contribution to the determination of pipeline integrity. NOPIG UNE NOUVELLE TECHNIQUE DE DÉTERMINATION DE L INTÉGRITÉ DES PIPELINES RÉSUMÉ Dans le monde entier, les exploitants de pipelines cherchent des solutions destinées à remplacer le racleur intelligent ou les tests hydrauliques pour déterminer l intégrité de leurs réseaux. N.P. Inspections Services GmbH propose une technique d inspection au-dessus du sol, appelée NoPig, pour la détection des défectuosités du métal. Au cours de ces trois dernières années, Gastransport Services été impliqué dans des recherches effectuées au moyen de projets mais aussi d examens effectifs de pipelines, destinés à déterminer les performances de l instrument dans des conditions de laboratoire ou sur le terrain. Ces recherches montrent le large progrès obtenu dans les performances. Il a ainsi été démontré que les défauts dus à la corrosion, qui sont critiques pour les pipelines d un diamètre atteignant jusqu à 12 pouces, peuvent être détectés lorsqu il est tenu compte des spécifications de l instrument. Cet article décrit le principe de mesure sur laquelle est basée la technique NoPig et les principales caractéristiques des performances de l instrument, telles que déterminées au cours des recherches et des mesures. De plus, l attention est portée sur l interaction entre les performances de l instrument, la répartition de la corrosion établie sur la base des données d un test de raclage et les tolérances dans la détermination des défectuosités. Enfin, les possibilités d utilisation du système NoPig dans un programme d inspection pour les pipelines sont discutées. La conclusion générale est que la technique d inspection NoPig peut offrir une contribution précieuse dans la détermination de l intégrité des pipelines.

2 NOPIG A NEW TECHNIQUE IN PIPELINE INTEGRITY G.G.J. Achterbosch, Gasunie Research, The Netherlands M. Broderick, N.P. Inspections Services GmbH, Germany 1. INTRODUCTION Pipelines are a crucial factor in the modern global economy. Billions of tons of different products are transported per year through pipelines both under and above ground. Pipelines are used because they are relatively cheap and safe when compared to other forms of transport. Although the safety record of pipelines is good, there is a growing pressure from (local) governments and regulators on pipeline owners and operators to prove the integrity of the pipeline. In the Netherlands, the safety awareness of both citizens and politicians was recently raised by 2 major incidents, not with pipelines but one with a fireworks factory and the other one was a fire in a pub. The most commonly used methods to prove the integrity of a pipeline is by inspection with intelligent pigs or hydro testing. For several reasons most companies prefer inspection by intelligent pigs. But a large portion of pipelines can not be inspected by intelligent pigs or only at very high cost. This leads to a world wide search for alternatives to prove pipeline integrity. Gastransport Services (part of N.V. Nederlandse Gasunie) is the main natural gas transmission company in the Netherlands with a pipeline grid of approximately km. Approximately half of this network is the main transmission grid (HTL) with pressures up to 66.7 bar and diameters ranging from 4 to 48. The other half is the medium pressure grid (RTL) with pressures between 8 40 bar and diameters ranging from Roughly speaking only the HTL grid can be inspected by intelligent pigs. Which means that approximately km of the network can not be inspected by intelligent pigs or only at very high cost. Since 1999 the policy of Gastransport Services is to regularly inspect the piggable pipelines by intelligent pigs. For the non-piggable lines there is yet no well defined inspection program. Realising that also the integrity of non-piggable lines should be well controlled, it was decided to investigate all possibilities to come to a solution for this integrity issue. In the search for alternative techniques Gastransport Services came across a technique called NoPig, being an above ground survey technique to detect metall loss, that was developed by FINO AG. In 2000 the principal characteristics of the technique were tested in a multi-sponsor project. Parameters that were investigated included defect size, depth of cover, defect position and their effect on detection by the system. External influences such as stray current and parallel lines were also investigated. The results confirmed the general specifications of the instrument at that time. Realising that the development of the instrument was still at an early stage and recognising the potential capabilities of the technique and the ongoing improvements by FINO, Gastransport Services decided to investigate the technique in more depth. In 2001 a project was defined, managed by Gasunie Research and sponsored by a consortium of 9 pipeline operators from Europe as well as from the USA. It was decided to investigate the performance of NoPig both by laboratory tests and field tests using both artificial, corrosion-like, defects and real corrosion defects. NoPig was used by Gastransport Services to partially survey 6 different pipelines with diameters in the range of 6 and 8. This paper will focus on the measurement principle, it s main characteristics, the progress in performance in time, a discussion on how this technique might be used and general conclusions.

3 2. THE NOPIG SYSTEM 2.1 Measurement principle The NoPig pipeline Inspection system analyses the magnetic field of a pipeline from above ground. To create this magnetic field, an inspection current is induced between two contact points on the pipeline. Preferably these contact points are existing points on the pipe: e.g. CP test posts. The inspection current is the superposition of a low (LF) and high frequency (HF): currently being 9 Hz (LF) and 620 Hz (HF). The low frequency fills up the whole cross section of the pipe wall while the high frequency only travels on the outer side of the pipe. Normally the magnetic field of a pipe without a defect consists of concentric circles around the geometric centre of the pipe. The magnetic field is equal to the magnetic field from the same current flowing through the centreline of the pipe (virtual current line). In case of metal loss the current distribution will be influenced by the defect. A defect on the outer side of the pipe will influence both the high and low frequency part of the current distribution while a defect on the inner side of the pipe will affect only the LF current. In both cases the virtual current axis (either LF, HF or both) will shift, resulting in a change of the magnetic field. By measuring this change of magnetic field above ground the metal loss can be detected (see figure 1). HF-current LF-current defect shift Magnetic flux lines Pipe without defect Pipe with 12 o`clock defect FIGURE 1: the magnetic field lines around a pipe with and without a defect The magnetic field is measured above ground by a sensor bar. This bar contains 4 lines with 6 magnetic field sensors per line. Each sensor measures the horizontal component of the magnetic field. This magnetic field is superposed on the existing earth magnetic field. In order to have a feeling for the magnitude of signals, lets view an example. Example: The magnetic field from a current of 1 Ampere at 1 meter distance is 0,2 µt while the earth magnetic field is approximately 50 µt. this means that the magnetic field of a pipeline without a defect is 250 times smaller then the existing earth magnetic field. Suppose that a defect will change the virtual current axis by 1 mm. This will then change the magnetic field at 1 meter distance by 0.1 %: 1 mm/1 m x 0,2 µt = 0,2 nt. This means that changes in the magnetic field as small as 1/ of the earth magnetic field will have to be detected. The signals from the sensors are measured, collected and stored in a data logger. From the measured signals of all sensors and sensor lines the virtual current axis for the LF and HF component, and changes in it, can be calculated. A shift in the magnetic field (and thus in the virtual current axis) will indicate a defect. Depending on the clock position of the defect the shift will either be vertical, horizontal or a combination of both. In fact, the calculation of the current axis gives the position of the pipeline and thus the NoPig system can act as a very precise pipelocator and determine the actual depth of cover! By displacing the sensor bar over the pipeline and measuring the signals the pipeline can be mapped.

4 2.2 The equipment The equipment consists of three modules: the sensor bar: an array of 4 sensor lines of 1 meter at 25 cm apart. the back pack unit: the back pack unit consists of the data logging and data storage equipment and a terminal to operate the sensor bar. It also provides the opportunity to install a GPS receiver for exact location information. The unit can either be mounted on a trolley or handled separately the current source module: a commercially available current generator that operates at the moment with a maximum current output of 10 Amperes 2.3 Inspection procedure In order to obtain most accurate and reliable results the system has to be calibrated daily before actually surveying a pipeline. Calibration is done by feeding a well defined current through a wire under a calibration table on which the sensor bar is placed. Knowing the exact distance between wire and sensor bar and the exact value of the current, the sensor bar is calibrated. For the actual survey the source module will be connected to the pipeline on two points thus creating a current loop. This will either be done on aboveground areas of the pipe, cathodic protection attachments or holes will have to be dug. The maximum distance between the two connection points is currently 500 metres. The sensor bar is connected to the back pack unit by a cable with the length of a few metres. After connection and calibration, the NoPig sensor array will be placed on the ground above the pipeline to be measured. The pipeline locator function of the system will give the precise location of the pipe under the ground as well as the depth. The crew will adjust the position of the sensor bar until the optimal position with respect to the pipeline is reached. The measurement is then started and will take approximately 11 seconds before the measurement stops. The measured signals, the pipe location, depth of cover and GPS coordinates (if applicable) are recorded. After the measurement the sensor bar will be moved 1 metre and the next measurement can be started. These steps are repeated until the desired or maximum survey length has been reached. 2.4 Data processing The in-depth analysis of signals is not done on site but after the survey has been completed. This evaluation process takes place in the office. Filtering techniques are used to select the specific low and high frequency signals and spatial filtering is used to eliminate interfering signals. In general the signal processing of an inspection technique can be described by at least three stages: 1. detection: is the signal significantly above the detection threshold which is related to noise levels and other factors? 2. sizing: what are the dimensions of a defect? 3. interpretation: what kind of defect is it? In this process of analysing the data, the software and adjustable parameters play an important role. But paramount is the expert knowledge of the characteristic signal of each kind of defect. Therefore fingerprints of known defects are stored in a database and used in the analysis of measured signals. After the raw data is collected from the pipeline the first step is the post processing of the inspection data. The post processing reduces external and internal interferences. External interferences include environmental magnetic noise, interference from power lines and influence of large metal objects directly adjacent to the pipeline. Internal interference include system random and systematic uncertainties. The second step consists of evaluating the post processed data in the Insight program. Based upon a given threshold the detection and sizing list is generated. In some regions the detection level will be adjusted if the depth of the pipeline is below the specified 1.5 metres.

5 2.5 Reporting The detection and sizing list is used to generate the Graphical analysis. It presents the schematic visualization as a bar of the inspected pipeline of approximately 100 meter length. The following elements are visualised in the bar using different colours: inspected and evaluated area with sensitive detection level 1 inspected and evaluated area with sensitive detection level 2 non-inspectable or not evaluated areas detection below detection level detection area metal loss or other information Below the bar the coverage of the pipe is shown in a separate graph (see figure 2) Photographs are made every 50 metres during inspection and enclosed in the report to illustrate peculiarities that might be relevant. Colour Legend Remarks Detection Area Detection below detection level Inspected Area, Level 1 Inspected Area, Level ,0 0,5 1,0 1,5 2,0 2,5 Coverage [m] FIGURE 2: the graphical presentation of results measured by the NoPig system 3. INSPECTION CONCEPT In order to determine whether the performance of the NoPig (and any) inspection technique is acceptable three items need to be adressed: 1. the expected corrosion distribution: what is the type and size of corrosion that can be expected on the pipelines to be investigated 2. the defect assessment model: which formulas and tolerances are being used when a defect is assessed to determine the remaining strength of the pipe 3. the performance of the instrument: what is the probability of detection or false calls and sizing accuracy for certain defect types and defect sizes. Only when these three elements are considered in combination, the question whether an instrument can be used, can be answered. To illustrate the above an example will be given. The example will only focus on external corrosion although the same holds for internal corrosion.

6 3.1 Expected corrosion distribution: Results from pigruns and excavations can be used to determine the type and sizes of corrosion on different categories of pipelines for a company. Each defect is characterised by its length (L), width (W) and depth (D). To simplify further presentation and simulation of defects it is possible to replace the length and width of a defect by a normalised radius (NR). This normalised radius is equal to the square root of the product of length and width: NR = B. L This normalised radius will resemble the real defect dimensions if B and L are about the same size. For large differences between B and L the normalised radius will be a less good substitute. Figure 3 shows upper and lower expected values for the corrosion distribution of a 36 pipeline based on the results of a pigrun. All defects with B/L> 6 or L/B > 6 have been left out. In the figure two curves have been drawn representing the 95 % confidence levels. Which means that it may be expected that 95 % of the corrosion on this pipeline have defect dimensions that are within the 2 curves. It can be seen from figure 3 that, with a chance of 95 %, it is likely that a defect with 3 mm reduction of wall loss will have a normalised radius of 18 < NR < 70 mm normalised radius corrosion r [log mm] corrosion expectation max. corrosion expectation min reduction wall thickness (mm) FIGURE 3: the expected external corrosion distribution, derived from an inspected 36 pipeline If one assumes that the dimensions of the corrosion are independent of the geometry of the pipe, then this corrosion distribution is valid not only for the inspected 36 pipeline but for other pipelines as well. 3.2 Defect assessment For defect assessment different models can be used: B31G, B31G modified, Rstreng and others.some of them have the same structure but use sligthly different factors. ASME B31G, and other evaluation methods, uses the following parameters: failure stress, flow stress, defect depth, defect axial length, wall thickness, pipe diameter and a Folias (bulging) factor. For a pipe with given geometrical and material properties for each defect, the resulting maximum pressure can be calculated and will only be depending on the length and depth of the defect and not on the width. Suppose that a pipeline grid is used for pressures up to 40 bar and that the acceptable stress is 300 N/mm 2 and that the BG-model is used to evaluate defects. Figure 4 shows the acceptable defect length (represented by the normalised radius) as a function of the reduction of wall thickness for 2 different pipes: a 6 pipe with wall thickness of 4.8 mm and a 12 pipe with wall thickness 7.1 mm.

7 1000 BG 6" BG 12" normalised radius NR (length L) [log mm] corrosion expectation max. corrosion expectation min reduction wall thickness (mm) FIGURE 4: corrosion distribution and corrosion tolerances based on the BG-defect assessment model for a 6 and 12 pipe. For the 6 pipe it can be seen that when the depth of the defect is below 3.75 mm (=78% of wall thickness) the defect length is not relevant for the remaining strength. While for a reduction of wall thickness above 3.75 mm the acceptable defect length is limited to very small values. The same holds for the 12 pipe with a threshold value of 5 mm reduction of wall thickness (= 70%). 3.3 Performance of the instrument The performance of the instrument and data analysis can be related to the items that were mentioned earlier: detection, sizing and interpretation. With respect to detection it is usual to use the term probability of detection (POD). This POD is the chance that the instrument will correctly detect a defect. And the POD is a function of the defectsize and is expressed as a percentage given a certain confidence level. Trying to improve the POD by changing detection levels often an increasing number of false calls (the chance that a detection is not correct) will arise. The sizes of the defect will be given with an uncertainty of Ud (at 95 % confidence levels). 3.4 Combining corrosion distribution, defect assessment, instrument performance From figure 4 it can be seen that for the 6 line a reduction of wall thickness of 3.75 mm is critical. One therefore wishes that the inspection technique will have a detection level of at least 3.75 mm. At least two additional factors will come into play in reality: the uncertainty Ud in defect depth and the corrosion rate CR of the defect. Combining the corrosion rate with the time interval to the next inspection one can calculate the extra margin that is needed in the minimum corrosion depth to be detected. Combining the measurement uncertainty of +/- 20% (of 3.75 mm), a corrosion rate of 0.1 mm/yr and an inspection interval of 5 yrs an extra margin of 1.25 mm is needed. Which means that the minimum detection level needs to be 2.5 mm for the 6 pipe (=52% of wall thickness). Similar calculations can be done for other pipe diameters and wall thickness.

8 4. PRINCIPAL CHARACTERISTICS OF THE NOPIG TECHNIQUE 4.1 External interferences From the measurement principle it can be understood that interferences on the magnetic field can be expected from stray currents or metal objects in the vicinity of the line. In general the experience is that static and slow dynamic interferences are hardly a problem. In most cases the post processing allows for the elimination of these disturbing signals like: parallel lines, AC and DC cables. Practical experience is that cars prevent measurement only during passing by. And metal objects connected to or very close to or above the pipeline will prevent an adequate measurement: reinforced concrete slabs, Tee s, valves, supports etc. 4.2 Environment From theory it can be understood that the medium between the sensor bar and pipeline is irrelevant as long as the medium is not magnetisable. So the type of soil and water(content) are not relevant as is proven in practice. 4.3 Ability to survey Just like other above ground inspection techniques the NoPig technique is partly limited in its possibility to be used as a survey technique because of the pipeline trace and of constraints of the technique itself. Obstacles over the line (e.g. fences, trees, reinforced concrete slabs, railroad- and watercrossings), near the line (e.g. large nearby metal objects like casings) or in the line (e.g. Tees, valves) will partly prevent measurements with the NoPig technique. Another factor is the dead zone near the connection points: the first few meters close to the connection point will be needed by the current to distribute evenly round the circumference of the pipe. And of course the pipeline depth will determine the probability of detection. 4.4 Detection and sizing performance Because of the confident nature of the above mentioned multi sponsor projects, details of the performance tests can not be given but the learning curve of NoPig can be seen in the figure underneath. The numbers are based on the technical specifications that were claimed for the NoPig system in the last 3 years % wall loss 50 % peak wall loss 0.5 m Pipe Diameter 3" to 8" at 1 m coverage 150 mm Pipe Diameter 3" to 12" at 1 m coverage % peak wall loss 50 mm Pipe Diameter 3" to 12" at 1 m coverage 2002 FIGURE 5: performance specifications for the NoPig system in 2000, 2001 and 2002

9 In 2000 a defect with dimensions of 50% wall loss and an area of typically an A-4 paper could be detected. And the range of diameters was limited to 8. At that time is was recognised that the system possessed a greater capability. Without modifications to the hardware but with learning experience from experiments and further field testing the performance of the post processing could be significantly improved resulting in a remarkable better specification of the instrument. The range of pipe diameters was expanded to 12 while the lower end of detection is now claimed to be in the range of the size of a business card with wall loss of 50%. For the sizing tolerance 20% of wall thickness is claimed with a minimum defect depth of 30% of wall thickness. Referring to figure 4 this means that the instrument is capable of detecting defects well below the critical limits for different pipe diameters. That the instrument still has additional potential for further improvement is deduced from the fact that there have been circumstances where NoPig correctly detected several welds in the pipeline plus the fact that sometimes defects with dimensions smaller than stated in the specifications have been found. 5. IMPLEMENTATION OF THE NOPIG TECHNIQUE IN AN INSPECTION PROGRAM For the implementation of the NoPig system there are several options: 1. use of the NoPig system as a stand alone survey technique 2. use of the NoPig system for spot measurements 3. use of the NoPig system in combination with other survey techniques It is very well possible to use the NoPig system as a stand alone technique for (parts of ) pipelines that fall within the specification of the technique. But it should be realised that the speed of surveillance is limited to approximately 500 metres a day (presently). Another approach could be the combination of a coating survey technique and NoPig either as a survey or spot measurement. One can decide to use a coating survey technique as a pre filter for the NoPig system. This means that the coating survey technique is used to select those spots where external corrosion is suspected. The NoPig technique can then be used either to survey around those spots or to measure at the spots to detect corrosion within the specification of the instrument. The second approach could be cost effective if there are parts of the line were no coating defects are detected and were no corrosion is expected. But it should be kept in mind that coating survey techniques also have limited probabilities of detection: meaning that not all coating defects are detected. And of course there can be places were external corrosion is present but no coating defect can be detected. 6. DISCUSSION Gastransport Services now has an experience of 3 years with the NoPig technique. And we feel there is an analogy with the development process of intelligent pigs. Comparing the performance of the intelligent pigs in the beginning of the eighties and of the pigs nowadays, one can see that there has been a tremendous progress in the performance. Which is a combination of improvement in hardware but probably even more in improved data processing and interpretation. Largely based on gaining experience from field operations. The same holds for the NoPig system: the progress in performance during the last 2 3 years is impressive. Not so much in hardware but also in data processing and interpretation. And also here the field experience is of paramount importance. Although there are similarities between intelligent pigs and NoPig it is good to realise the fundamental difference: intelligent pigs are in-line inspections devices and are in direct contact with the pipeline whereas NoPig is measuring aboveground and thereby always at a distance from the pipeline. This means that the ultimate performance of the NoPig system will never equal that of an intelligent pig. But the present performance of the NoPig system is good enough for the detection of

10 certain kinds/sizes of corrosion. Referring to figure 4 it is obvious that critical or sub critical defects fall within the specification of the NoPig system: either internal or external defects. Using the NoPig technique as an inspection technique requires, as most other aboveground survey techniques, a thorough preparation before actually starting the survey. Route map information will have to be used to determine (or at least estimate) which parts of the pipeline can and which parts can not be surveyed, what the depth of cover is, where connection points are available, what possible interference may be etc. etc. Once a survey is done one will obtain not only the information about metal loss but also the exact depth of cover and any peculiarities on the trace. 7. CONCLUSIONS The NoPig inspection system provides an additional tool for non piggable pipelines in the operators struggle to prove pipeline integrity. It has been shown that corrosion defects that are critical for pipelines up to 12 can be detected when taking the specifications of the instrument into account. How and how often the NoPig technique is used in an inspection regime will depend on the philosophy of the pipeline owner or pipeline operator. The decision will partly depend on the knowledge of the expected corrosion distribution on a pipeline, the estimate of corrosion rates and the defect assessment models used. But also on the possible use of other techniques. The NoPig pipeline inspection system can be used as a spot measurement or as a survey technique. Although much experience has been gained there are still questions open. It is for example not clear whether the technique detects the missing volume of metal loss or combinations of depth, length and width of defects. But these questions do not prevent the use of the system in the field as is proven in the surveys that have been done within Gastransport Services and other companies. 8. REFERENCE 1. Krivoi G.S., Kallmeyer J.P., Baranyak A., Broderick M. (November 2001). NoPig: metal-loss detection system for non-piggable pipelines, Paper presented at the IGRC conference, November 2001, Amsterdam