Biuletyn Instytutu Spawalnictwa No. 3/2013
Transkrypt
Biuletyn Instytutu Spawalnictwa No. 3/2013
BIULETYN ISSN 2300-1674 INSTYTUTU SPAWALNICTWA No. 3/2013 INSTITUTE OF WELDING BULLETIN BIULETYN INSTYTUTU SPAWALNICTWA No. 3 BIMONTHLY Volume 57 CONTENTS •T. Pfeifer, L. Aucott, A. Żak – Testing of solidification cracking susceptibility of MAG welded overlays on S355JR steel using a Transvarestraint test....................... 5 •W. Jamrozik, M. Fidali – Application of image fusion to identification of welding imperfections............................................................................................12 •A. Sawicki - Damping factor function in AC electric arc models. Part 2: Damping factor function in universal electric arc models with moderate cooling......... 22 •R. Krawczyk, P. Wojtas, K. Poch - Comparative assessment of selected welding imperfections in VT, PT and MT methods........................................................ 31 •M. Stachurski – Non-destructive testing of helically welded pipes made of thermomechanically rolled materials used for sending of combustibles........................ 37 This work is licenced under Creative Commons Attribution-NonCommercial 3.0 License INSTITUTE OF WELDING The International Institute of Welding and The European Federation for Welding, Joining and Cutting member Summaries of the articles T. Pfeifer, L. Aucott, A. Żak Testing of solidification cracking susceptibility of MAG welded overlays on S355JR steel using a Transvarestraint test A. Sawicki - Damping factor function in AC electric arc models. Part 2: Damping factor function in universal electric arc models with moderate cooling Application of the Transvarestraint test to assessment of solidification cracking susceptibility of overlay welds made by MAG method on S355JR steel has been presented. A specific test rig enabling to apply different strain of the test piece with high speed was designed. In the course of the research the weld pool temperature history was registered, cracking qualification was done and the examination of crack opening and HAZ using scanning electron microscopy and light microscopy was carried out. Solidification cracking temperature range of this material system was determined. Nonlinear function of the heat processes damping factor in thermal column plasma have been introduced to modified mathematical models of the AC electric arc. This way, universal arc models with moderate cooling and constant geometric sizes, using static characteristics, have been created. Similar method has been used to hybrid arc models with variable column length connecting Berger and Kulakov models. It has been proposed the introduction of nonlinear damping factor function to modified Voronin model using static characteristics. Thereby the extended range of current for application of these models has been achieved. W. Jamrozik, M. Fidali – Application of image fusion to identification of R. Krawczyk, P. Wojtas, K. Poch welding imperfections Comparative assessment of selected It has been presented the issues concerning welding imperfections in VT, PT the application of images fusion registered in and MT methods visible and infrared radiation for diagnostics of MIG/MAG processes. The sequences of visible and infrared images visualizing the arc burning during performing the weld were examined. Recorded during experiments image sequences were subjected to fusion and next to analysis in order to obtain diagnostic signals. On the basis of selected features of diagnostic signals it was carried out the classification of the welding process state and its results were compared with those obtained for diagnostic signal features determined independently for visible and infrared images. The results show that the application of images fusion enables to identify effectively various welding imperfections forming in welding processes. No. 3/2013 The issues connected with assessment of sensitivity in selected NDT methods are considered. In particular, the attention has been paid to the tests which make it possible to detect any imperfection of surface, i.e. visual testing, liquid-penetrant and magnetic-powder method. The aim of the studies was to reveal the analogy and differences between results obtained from series of tests (VT, PT and MT). M. Stachurski – Non-destructive testing of helically welded pipes made of thermomechanically rolled materials used for sending of combustibles In the first part of the paper it has been presented the short information on methods BIULETYN INSTYTUTU SPAWALNICTWA 3 of fabrication of helically welded pipes used for transporting of combustibles. In the second one it has been given the NDT methods used during inspection of the pipes in production plants. In the next parts it has been described the NDT methods for steel pipes focusing first of all on visual, ultrasonic and radiographic ones. The paper has been ended with the information about prospects of gas industry development in Poland and in the world. Biuletyn Instytutu Spawalnictwa ISSN 2300-1674 Publisher: Instytut Spawalnictwa (The Institute of Welding) Editor-in-chief: Prof. Jan Pilarczyk Managing editor: Alojzy Kajzerek Language editor: R. Scott Henderson Address: ul. Bł. Czesława 16-18, 44-100 Gliwice, Poland tel: +48 32 335 82 01(02); fax: +48 32 231 46 52 E-mail: [email protected]; [email protected]; [email protected] www.bis.is.gliwice.pl Prof. Jacek Senkara - Warsaw University of Technology, Biuletyn Scientific Council: Akademik Borys E. Paton - Institut Elektrosvarki im. E.O. Patona, Kiev, Ukraine; Nacionalnaia Akademiia Nauk Ukrainy (Chairman) Prof. Luisa Countinho - European Federation for Welding, Joining and Cutting, Lisbon, Portugal dr Mike J. Russel - The Welding Institute (TWI), Cambridge, England Prof. Andrzej Klimpel - Silesian University of Technology, Welding Department, Gliwice, Poland Prof. Jan Pilarczyk - Instytut Spawalnictwa, Gliwice, Poland Biuletyn Program Council: External members: Prof. Andrzej Ambroziak - Wrocław University of Technology, Prof. Andrzej Gruszczyk - Silesian University of Technology, Prof. Andrzej Kolasa - Warsaw University of Technology, Prof. Jerzy Łabanowski - Gdańsk University of Technology, Prof. Zbigniew Mirski - Wrocław University of Technology, Prof. Jerzy Nowacki - The West Pomeranian University of Technology, dr inż. Jan Plewniak - Częstochowa University of Technology, 4 Prof. Edmund Tasak - AGH University of Science and Technology, International members: Prof. Peter Bernasovsky - Výskumný ústav zváračský Priemyselný institút SR, Bratislava, Slovakia Prof. Alan Cocks - University of Oxford, England dr Luca Costa - Istituto Italiano della Saldatura, Genoa, Italy Prof. Petar Darjanow - Technical University of Sofia, Bulgaria Prof. Dorin Dehelean - Romanian Welding Society, Timisoara, Romania Prof. Hongbiao Dong - University of Leicester, England dr Lars Johansson - Swedish Welding Commission, Stockholm, Sweden Prof. Steffen Keitel - Gesellschaft für Schweißtechnik International mbH, Duisburg, Halle, Germany Eng. Peter Klamo - Výskumný ústav zváračský Priemyselný institút SR, Bratislava, Slovakia Prof. Slobodan Kralj - Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, Croatia Akademik Leonid M. Łobanow - Institut Elektrosvarki im. E.O. Patona, Kiev, Ukraine; dr Cécile Mayer - International Institute of Welding, Paris, France Prof. Dr.-Ing. Hardy Mohrbacher - NiobelCon bvba, Belgium Prof. Ian Richardson - Delft University of Technology, Netherlands Mr Michel Rousseau - Institut de Soudure, Paris, France Prof. Aleksander Zhelew - Schweisstechnische Lehr- und Versuchsanstalt SLV-München Bulgarien GmbH, Sofia Instytut Spawalnictwa members: dr inż. Bogusław Czwórnóg; dr hab. inż. Mirosław Łomozik prof. I.S.; dr inż. Adam Pietras; dr inż. Piotr Sędek prof. I.S.; dr hab. inż. Jacek Słania prof. I.S.; dr hab. inż. Eugeniusz Turyk prof. I.S. BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 Research Tomasz Pfeifer, Lee Aucott, Aleksander Żak Testing of solidification cracking susceptibility of MAG welded overlays on S355JR steel using a Transvarestraint test Abstract: Application of the Transvarestraint test to assessment of solidification cracking susceptibility of overlay welds made by MAG method on S355JR steel has been presented. A specific test rig enabling to apply different strain of the test piece with high speed was designed. In the course of the research the weld pool temperature history was registered, cracking qualification was done and the examination of crack opening and HAZ using scanning electron microscopy and light microscopy was carried out. Solidification cracking temperature range of this material system was determined Keywords: welding, cracking, MAG, Transvarestraint test Introduction Solidification hot cracking of welded joints is commonly regarded as one of the most hazardous of various types of cracking. For this reason the mechanisms of cracking and hot cracking resistance of different structural materials are a subject of thorough research. Such cracks, developing usually on grain boundaries, are generated in a weld at the end of a crystallisation process when a metal being the mixture of a liquid and a solid body is torn as a result of deformations related to weld material contraction accompanying crystallisation and self-cooling. One of reliable tests enabling the assessment of steel susceptibility to solidification cracking is the Transvarestraint test. The methodology of this test has been developed directly on the basis of the Varestraint test whose precursors were Savage and Lundin [1]. However, neither of these tests have become standardised and, as a result, today there are many different versions of them [2]. The purpose of research conducted using the Transvarestraint test was to determine a temperature range in which solidification cracks are generated and develop. The range is determined on the basis of the measurement of the maximum crack length (MCL) generated in a sample subjected to a high-rate strain after welding as well as on the basis of a recorded temperature course in a weld [3-8]. Most publications related to the Transvarestraint test describe tests conducted using TIG welding without a filler metal. However, the publication [9] refers to the possibility of welding with a filler metal, which enabled determining the hot cracking resistance of a fill- dr inż. Tomasz Pfeifer (Ph.D. Eng.) – Instytut Spawalnictwa, Zakład Technologii Spawalniczych /Welding Technologies Department/; M.Sc. Eng. Lee Aucott – University of Leicester; mgr inż. (M.Sc. Eng) Aleksander Żak – Frenzak ltd., Mikołów, Poland No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 5 er metal or of a specific material system (filler metal + base metal). In such cases the use of appropriate welding parameters, enabling the proper overlay weld fusion into the base metal is of great importance. This was the reason for undertaking tests aimed at the determination of hot cracking resistance and solidification cracking temperature range using the Transvarestraint for the following material system: electrode wire PN-EN ISO 14341-A-G3Si1 and steel S355JR (base metal). The assessment of susceptibility to hot cracking and the determination of solidification cracking temperature range of such a material system is important from the Mint-Weld project implementation point of view (contract no. NMP3-SL-2009-229108). The tests involve the preparation of a reliable and useful numerical model of the development of interfaces during the welding and operation of welded joints. The model should take into consideration molecular dynamics phenomena in the atomic scale (the properties of interfaces, the chemical composition and structure), phase models in the nano/micro scale (the structure of grain boundaries and the chemical composition of interface boundaries), mesoscopic models tracking the front bracket bending block rolls of interfaces (crystal growth and grain size distribution) as well as modelling using fluid dynamics in the macroscopic scale (heat and mass flow). The obtained results of the solidification cracking susceptibility assessment tests constituted necessary and very important input data for modelling the development of interphase boundaries in various scales. The tests and their results presented below are a part of the MintWeld project implemented at Instytut Spawalnictwa within the confines of the 7th Framework Programme. Course of tests and results Transvarestraint test The tests to determine the solidification cracking temperature range involved the use of a base metal in the form of 6 and 12mmthick S355JR steel plates (300 mm × 200 mm) and an electrode wire PN-EN ISO 14341 – A – G3Si1 with a diameter of 1 mm. A 180mmlong overlay weld was made on the aforesaid plates, using a robotised standard arc MAG method. In all the tests the following constant welding parameter values were used: current 200 A, arc voltage 25 V and a welding rate of 15 cm/min. The Transvarestraint tests were carried out with a device built especially for this purpose in the Frenzak Sp. z o.o. company in Mikołów, one of the MintWeld project hydraulic actuator Fig. 1. Device for Transvarestraint test and manner of mounting and deforming test piece 6 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 partners. The testing device is presented in Figure 1. The test plates were mounted in a repeatable manner under a fixing bracket on exchangeable blocks with a provided arch of appropriate radius (Fig. 1). Next, using the MAG method and adopted parameters a 180mm-long overlay weld Fig. 2. Results of temperature measurements in liquid metal pool during was made. During welding surfacing in Transvarestraint test the temperature of a weld pool was measured by a D-type thermocouple transformation of the solidification process. (tungsten-rhenium; dia. 0.25mm) directly in The temperature measurements revealed that the liquid pool. After welding (welding arc for the 12mm-thick plate the cooling rate in termination generated a signal for a bending the above range was approximately 40°C/s, device) the plate was deformed by means of whereas for the 6mm-thick plate it amounted a hydraulic actuator and bent on a steel block to approximately 19°C/s. Solidification cracks appear in the last perpendicularly in relation to the direction of welding; the rate of bending being 24 mm/s. crystallisation phase, at a temperature below Bend radiuses, made on a one edge of the which the whole material is fully solidified. steel block were selected in a manner mak- The 6mm-thick plates revealed a significanting it possible to generate (on the surface) ly greater number of solidification cracks a strain corresponding to a specific elonga- (Table 1) than in the 12-mm thick plates. The tion percentage, i.e. 0.5%, 1%, 2%, 5%, 10% reason behind such a difference is the fact and 20%. For this reason, for each S355JR for the same welding parameters the heat steel plate 6 tests were carried out each time emission from the 6mm-thick plate is much changing a steel block for another one, with slower. As a result, the weld remains significantly longer in the area above the solidifia different bend radius. cation temperature; on the basis of Figure 2 Temperature measurements it is possible to determine the cooling rate Temperature measurements carried out at the liquidus–solidus temperature range. In during welding revealed that the welds so- the aforesaid range the material is a mixture lidified at approximately 1460°C (Fig. 2) in of liquid and solid phases, and therefore is the case of both tested plate thicknesses. In significantly less susceptible to solidification the case of the 12mm-thick plate the max- cracking. After solidification, a cooling rate imum registered temperature of the liquid is approximately 74°C/s for the 12mm-thick metal was approximately 1622°C, which plate and 51°C/s for the 6mm-thick plate. significantly exceeds a liquidus temperature. Figure 2, presenting the results of a temper- Quantitative analysis of cracks ature measurement in the weld pool, shows After the bend test conducted transversea relatively low temperature gradient from ly in relation to the direction of surfacing, the maximum temperature to the solidifi- the analysis of cracks generated in the crater cation temperature. Such a gradient results was carried out. Table 1 presents the results from the heat release related to the phase obtained and Figure 3 shows the example of No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 7 Table 1. Results of analysis of solidification cracks in the craters of overlay welds made during the Transvarestraint tests Plate thickness, mm 6 Strain % Number MCL, of cracks mm 12 TCL, Number MCL, mm of cracks mm 0 0 0 0 0,5 0 0 0 1 1 10 10 2 2 11 14,5 5 5 21 36 10 16 20 36,5 20 34 21 40,5 Note: MCL – Maximum Crack Length TCL – Total Crack Length 0 0 0 1 1 3 4 0 0 0 8 3 8 9 0 0 0 8 3 19 26 Fig. 3. Solidification crack in the crater of weld made on 12mm-thick plate with 20% strain, used in Transvarestraint test the crater in the overlay weld made on the 12mm-thick plate with a 20% strain in which solidification cracks were observed. Visual tests and crack length measurements revealed that an increase in the strain degree is accompanied by an increase in the number of cracks as well as an increase in the maximum length of the solidification crack. Figure 3 presents a longitudinal crack in the weld axis formed at a significant strain. The crack was generated at the interface of grains, developing from the overlay weld edge towards the main axis of the weld and is a typical example of a solidification crack. Fig. 4. Structure of S355JR steel base metal, visible banded structure of ferrite and pearlite 8 TCL, mm The latter part of this publication presents the results of the metallographic tests of this crack and the areas adjacent to it. Metallographic tests Metallographic tests were carried out for the test pieces cut out of the 12mm-thick plates, which during surfacing were subjected to the greatest strain (20%). The tests included observations with a light microscope and the structural analysis of the base metal, heat affected zone and, in particular, the overlay weld line of fusion into the base metal. The tests also involved the observation of the crack surface using a scanning microscope. Figure 4 presents the structure of the base metal which was not exposed to any welding-related heat impact, and thus was free from any structural changes. Figure 5a presents the structure of the material approximately 1 mm away from the fusion line, in the HAZ area adjacent directly to the base metal. This area revealed only small changes in the microstructure (small amounts of bainite). Figure 5b presents the HAZ area near the fusion line. During surfacing, the area reached a temperature close to the melting point. The BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 HAZ has a coarse-grained structure com- significantly increase the likelihood of solidification cracking occurence. posed mainly of ferrite and bainite. At the next stage of the tests the surface Figure 6 presents the microstructure in the fusion line area both on the base met- of a crack in the overlay weld was examal side and on the overlay weld side. Dur- ined. The crack was sampled for a test piece ing the Transvarestraint test the area was which underwent ultrasonic cleaning and completely molten. Welding materials of a was observed by means of a scanning elecsimilar chemical composition is accompa- tron microscope FEI Sirion 200. Figure 7 nied by the epitaxial grain growth. In this presents the dendritic structure of the crack process a liquid crystallises forming new surface. The structure indicates that the grains directly from the solid phase from crack is a real solidification-type hot crack. the base metal located underneath. The a) b) grains are anisotropic as they grow in accordance with a heat flow direction. In this manner a characteristic columnar structure is obtained. Figure 6a presents Fig. 5. a) Structure of material in S355JR steel HAZ, a typical columnar in area adjacent to base metal, coarse-grained structure, b) Coarse-grained structure in HAZ adjacent to fusion line, structure, in which it visible bainite and small amounts of ferrite is possible to observe the Widmanstätten a) b) structure formed directly on the ferrite of grain boundaries. Figure 6b shows a structure composed mainly of acicular ferrite and polygonal ferrite. Impurities Fig. 6. Microstrucure in fusion line area, present in the steel or a) view from base metal side: visible acicular ferrite and Widmanstätten structure, weld metal such as b) view from overlay weld side: acicular ferrite and polygonal structure phosphorus and sulphur are usually sit- a) uated between dendrites formed during crystallisation. For this reason, impurities Fig. 7. Results of tests of usually concentrate in solidification crack surface; the weld axis at the a) sample preparation manner; end of the crystallib) image of tested surface obtained sation process, which from scanning electron microscope No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 9 Determination of solidification cracking temperature range In order to determine the temperature range of solidification cracking in the overlay welds made on the 12mm-thick plates, the longest crack recorded in all tests with a various strain degree was divided by the welding rate used in a given test. This was done in order to determine a time interval (crack length 9 mm, welding rate 2.5 mm/s, crack formation time interval 3.6 s). Afterwards, the time interval was plotted on a diagram presenting weld pool temperature values recorded during surfacing. The diagram enabled the determination of the temperature range (above a solidification temperature) in which a crack may develop when the weld material remains a mixture of liquid and solid phases. The wider this range, the higher the material susceptibility to solidification cracking. Figure 8 presents the methodology of determining a solidification cracking temperature range. G3Si1 on S355JR steel, subjected to a strain equal up to 20% elongation. The determined temperature range amounts to 130°C. The obtained value is significantly higher than values presented in reference publications (< 50°C). However, most of the described tests aimed to determine the hot cracking resistance of a base metal. For this reason the tests were conducted using the TIG method without a filler metal. Instead, the base metal underwent remelting. The higher SCTR values obtained in the tests refer to a specific material system (base metal + filler metal) and may result from greater stress concentration. Concluding remarks The applied testing methodology and the test rig used in the research made it possible to carry out a modified Transvarestraint test with a various strain degree and determine a Solidification Cracking Temperature Range during surfacing 6mm-thick and 12mm-thick S355JR steel plates. The modification of the Transvarestraint test consisted in MAG surfacing instead of remelting a base metal by means of the TIG method. The modified test proved useful for the assumed test objective, i.e. to develop input data for welding process modelling in respect of developing interfaces. The tests also enabled the development of a methodology for liquid metal temperature measurements during MAG surfacing as well as made it Fig. 8. Temperature of liquid metal pool and methodology of determining possible to record temperature Solidification Cracking Temperature Range (SCTR) for overlay welds changes during the liquid metmade of 12mm-thick S355JR steel plates al pool solidification and weld cooling. The tests revealed that heat flow impeding The tests revealed that the Transvarestraint test can be used to determine the So- conditions cause an overlay weld to remain a lidification Cracking Temperature Range for mixture of liquid and solid phases far longer MAG-welded overlay welds made with a wire than a material subjected to remelting only, 10 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 which favours hot cracking. The Transvarestraint test carried out for the 6mm-thick test pieces revealed a significantly greater number of cracks than that involving the use of the 12mm-thick test piece, applying the same surfacing parameters and bending conditions. The determined value of the Solidification Cracking Temperature Range (SCTR) for the MAG-surfaced 12mm-thick plate is significantly higher than the value determined in the Transvarestraint test conducted with the TIG method and without the addition of a filler metal. The use of the MAG method and making an overlay weld during a Transvarestraint tests increases the probability of hot cracking due to the greater concentration of stresses. References 1. Savage, W., and Lundin, C. (1965). The Varestraint Test. Welding Journal, (10). 2. Lippold, J. (2005). Recent Developments in Weldability Testing for Advanced Materials. Joining of Advanced and Specialty Materials, (7). 3. Dupont, J., and Lippold, J., et al. (2009). Welding metallurgy and weldability of nickel-base alloys. Hoboken, N.J.: Wiley. 4. Shankar, V. (2003). Solidification cracking in austentic stainless steel welds. No. 3/2013 Academy Proceedings in Engineering Sciences. Indian Academy of Sciences. 5. Finton T., and Lippold J. (2004). Standardization of the Transvarestraint test. Edison Welding Institute Summary Report Nr 04-05/2004 6. Kou, S. (2003). Welding metallurgy. Hoboken, N.J.: Wiley-Interscience. 7. Lippold, J., and Lin, W. (1996). “Weldability of commercial AlCu-Li alloys”, paper presented at Aluminum Alloys – Their Physical and Mechanical Properties ICAA5 Conference, Transtec Publications. 8. Nakata, K., and Matsuda, F., (1995). Evaluation of ductility characteristics and cracking susceptibility of Al alloys during welding. Trans of JWRI, (1). 9. Scotti, A. (2004). Performance assessment of the (Trans) Varestraint test for determining solidification cracking susceptibility when using welding processes with filler metal. Measurement Science and Technology, (11). 10.Keehan, E., et al, (2002). “Microstructural and mechanical effects of nickel and manganese on high strength steel weld metals”, paper presented at Trends in Welding Research International Conference, Pine Mountain, Georgia, USA. BIULETYN INSTYTUTU SPAWALNICTWA 11 Wojciech Jamrozik, Marek Fidali Application of image fusion to identification of welding imperfections Abstract: It has been presented the issues concerning the application of images fu- sion registered in visible and infrared radiation for diagnostics of MIG/MAG processes. The sequences of visible and infrared images visualizing the arc burning during performing the weld were examined. Recorded during experiments image sequences were subjected to fusion and next to analysis in order to obtain diagnostic signals. On the basis of selected features of diagnostic signals it was carried out the classification of the welding process state and its results were compared with those obtained for diagnostic signal features determined independently for visible and infrared images. The results show that the application of images fusion enables to identify effectively various welding imperfections forming in welding processes. Keywords: MAG welding, welding imperfections, image fusion, identification of welding process state classification Introduction Obtaining diagnostic signals is an important stage of diagnosing, the primary purpose of which is to recognise the state of an object or a process. Diagnostic signals are courses of certain quantities in time, on the basis of which it is possible to distinguish states. Diagnostic signals can simply be courses registered by means of various sensors or they can be obtained by processing and analysing such courses as the courses of values of features in time. Real diagnostic tasks, carried out for objects or processes characterised by high complexity require recording many process parameters. However, it should be noted using all such parameters to generate diagnostic signals, and further, to recognise the state of an object or a process is pointless as some signals carry no information about state changes and may only bring information noise to a recognition process. In welding process diagnostics, process parameters are mostly registered in the form of the time courses of the values of physical quantities affecting the quality of a joint, e.g. current, voltage, gas flow, sound flow etc. Recent years have seen a significant growth in the importance of visual techniques in diagnosing, as well as those of welding processes [7], which has created the possibility of recording signals in the form of image sequences. Such images are mainly recorded in the range of visible light (TV) as well as in the middle infrared band (IR, with a wavelength from approximately 0.9 to 14 μm). The use of a thermographic camera, operating in the infrared range enables a non-contact and multipoint temperature measurement. Thanks to this solution it is possible to obtain information about temperature distribution, mainly in the area of a burning arc, weld pool and a cooling weld. dr inż. Wojciech Jamrozik (Ph.D. Eng.), dr inż. Marek Fidali (Ph.D. Eng.) - Politechnika Śląska, Wydział Mechaniczny Technologiczny, Instytut Podstaw Konstrukcji Maszyn/Silesian University of Technology, Faculty of Mechanical Engineering, Institute of Fundamentals of Machinery Design 12 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 At the same time a given object can be observed by means of IR and TV cameras. In such cases the redundant representation of objects in an image is obtained. In addition to information relevant from a welding process recognition point of view, contained in IR and TV images, there is also some redundant information. The elimination of redundant information with simultaneous maintaining of all information important from a welding imperfection recognition point of view can be obtained by an image fusion, i.e. a multi-stage technique of merging many input images, presenting the same scene as one output image [4]. Images obtained in this way can constitute the basis for determining diagnostic signals enabling efficient diagnosing of welding processes [3]. Image fusion Image fusion is a technique which consists in forming a new synthetic image out of at least two input images recorded by devices of different parameters and different locations. It is important that input signals should represent a common scene (or its fragment). The process of image fusion is a multi-stage process and in most cases consists of three basic stages: image matching, image aggregation and post-processing, consisting mainly in spatial filtration sharpening a resultant image. Sometimes the third stage is omitted and the whole fusion includes matching and aggregation. There are three basic groups of methods enabling image fusion, i.e. methods operating on the level of pixels, methods operating on the level of features and methods operating on a symbolic level also referred to as a decision level. In practice, the most commonly used fusion methods are those operating on the level of pixels. These methods assume the existence of a relationship between individual pixels of all input images. Such an assumption requires performing a geometric matchNo. 3/2013 ing of input images. The results of an image fusion largely depend on the quality of the conducted matching. Geometric matching, being the first stage of image fusion, can be carried out in several ways [5]. The most common is image matching, i.e. matching of areas characterised by similar brightness or colour or matching based on the characteristic features of an image, e.g. the edges or points of objects visible in an image. Matching images registered in visible light and in infrared is a complex task as image contents differ, usually in pixel values and object shapes. An additional problem can be the lack of an unambiguous relationship between object colours and greyness shades in visible light and object colours and greyness shades in infrared. Reference publications contain descriptions of various matching algorithms, yet comparative tests [2], [5] of selected algorithms justify the application of hybrid algorithms, making use of both the brightness distribution of areas in an image and of edge-related information. One of such methods described in the publication [5] can be used with indexed images or images recorded in the scale of greyness shades. This method uses edge orientation maps. On the basis of these maps as well as on the basis of pixel brightness values it is possible to create a 3D bar chart. An objective function determines the uncertainty of matching objects located in input images corresponding to edges as well as the shade of individual areas. Such a solution enables the obtainment of a low level of matching errors. The next stage of the fusion of images is their aggregation aimed at joining previously matched input images into one output image. Similarly, as in the case if matching algorithms, there is also a significant number of well described algorithms for the aggregation of images. The algorithms on the level of pixels can be divided into four groups [7]: algorithms operating on a model of a colour BIULETYN INSTYTUTU SPAWALNICTWA 13 space, statistical/numerical algorithms, algo- • wavelet transformation, which is similar to pyramid decomposition methods, where rithms making use of a large-scale decompothe main difference is the amount of inforsition as well as spectral and radiometric almation contained in the hierarchical strucgorithms. For the purpose of merging vision ture of images. Pyramid transformations and thermographic images of a welding arc, lead to a redundant set of transformation the authors considered mainly algorithms factors, whereas a wavelet transform leads using a large-scale transform, the functionto a non-redundant image representation. ing of which is based on a general scheme of successive operations presented in Figure 1. Within the confines of previously conducted tests [3], the authors examined Decomposed and assessed various aggregaimage DA Image A tion algorithms for applicaAveraging tion in the fusion of welding Resultant or selection image DC arc vision and thermographic images. For the purpose of the tests Decomposed described in the article an agresultant gregation algorithm based on image DC a shift invariant Haar wavelet Image B Decomposed (SIH) [6] was used. Within image DB Fig. 1. Scheme of image aggregation this method, unlike in a disusing large-scale transformation (on the basis of [1]) crete wavelet transform, sucAs is shown in Figure 1, aggregation us- cessive image copies are not subjected to ing large-scale transformation consists of resolution reduction, which leads to obtaining a highly redundant image representation. three stages [1]: • decomposition, when each of the input This algorithm for joining information out of images is transformed into a hierarchical decomposed input images uses Burt’s adapstructure, whose successive stages are ap- tive rule [1]. As a result it is possible to join propriately transformed copies of an input images by averaging or selecting pixels depending on the changes of pixel brightness image, • joining of images, during which, by means variance in some vicinity of a given pixel. of a proper fusion principle using the se- The use of this method, to a significant exlection or averaging of pixels, a resultant tent, enables maintaining information about decomposed image representation is de- the shape, brightness distribution and temperature in the welding arc area and the weld termined, • synthesis, the result of which is a new syn- pool. thetic image, created by using a transform reverse to the one used at the decomposi- Welding process observation tion stage. The observation of a MAG welding proThe hierarchical structure is obtained by using: cess was carried out using an observation • pyramid transformation, used for deter- head composed of an uncooled thermographmining the set of images, in which each ic camera (Infratec VarioCam Head) with a of the images in a pyramid is two times bolometric detector of 640×480 px resolution smaller than an image on a lower level of and a lens with a focal length f=50 mm as the pyramid, well as a CCD camera operating in the visible 14 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 light range (ImagingSource DMK21AF04), with 640×480px resolution, equipped with a lens with a focal length f=25 mm. The cameras were provided with filters protecting optical systems against spatters and limiting the power of electromagnetic radiation emitted during welding. Images were recorded synchronically at a frequency of 50 images/s. The MAG welding tests were carried out at the Department of Welding of the Silesian Universi- Welding ty of Technology, using a ro- current [A] botised station for rectilinear MAG welding The station was 240 equipped with a welding tractor AS14a-1200 made by the OZAS company and provided with a set of supports for fixing a welding torch as well as with a semi-automatic welding machine TotalArc 5000 produced by Castolin. The testing station with the observation head is presented in Figure 2. Fig. 2. Mechanised station for MAG welding During experiments butt joints of steel S235JR (EN 10025-2) sheets with the dimensions of 300×150×5 mm were made. Apart from situations when special cases of imperfections were simulated, the edges of sheets were V-bevelled at an angle of α=60°; the distance between the sheets was b=1.0 mm. No. 3/2013 The butt joints of sheets were made on a copper plate equipped with mechanical clamping of sheets being joined. The welding process involved the use of a solid electrode wire (Castolin CastoMag 45255) with a diameter of 1.2 mm and a gas mixture M21 (82%Ar+18%CO2). Welding parameters are presented in Table 1. Table 1. MAG welding parameters Wire Shielding Arc Welding Exposed feeding gas voltage rate length of rate flow rate [V] [cm/min] wire [mm] [m/min] [l/min] 25 32 7,4 15 15 While making welded joints various welding parameters were deliberately simulated by destabilising arc using the following: • changes in the voltage and/or current of an electric arc, • changes in a shielding gas flow rate, • improper preparation of the surfaces of sheets to be joined. Welding imperfections were classified into the following 8 condition classes: • z0 – model process, no imperfections, • z1 – momentary disappearance of a gas shielding, • z2 – atmospheric corrosion of sheets to be joined, • z3 – weld groove imperfections, simulated by openings located at a specified distance from one another, in the axis of the weld groove, • z4 – improper welding current different by ±20% against model parameters, • z5 – improper distance between sheets being joined, amounting to 1.5 mm, 2.0 mm, 2.5 mm, 3.0 mm and 3.5 mm, • z6 – improper arc voltage, different by ±15% against model parameters, • z7 – changing geometry of the weld groove due to various values of a bevelling angle: 0°, 15°, 30°, 45°, 60°. BIULETYN INSTYTUTU SPAWALNICTWA 15 Fig. 3. Exemplary model of welded joint: a) view from the weld face b) view from the weld root Fig. 4. Exemplary welded joint, with an imperfection caused by the momentary disappearance of the gas shielding: a) view from the weld face b) view from the weld root Fig. 5: Examples of images presenting a welding arc recorded: a) in visible radiation range b) in infrared Fig. 6. Exemplary images presenting a welding arc obtained by a) averaging b) aggregation of vision and thermographic images matched by means of the SIH method 16 BIULETYN INSTYTUTU SPAWALNICTWA Figures 3 and 4 present an example of a model joint and a joint formed during the disappearance of a gas shielding. The conducted observations enabled obtaining sequences of thermographic images and images recorded in the visible light range. The examples of images showing a welding arc in visible light and in infrared are presented in Figure 5. The recorded sequences of vision and thermographic images underwent fusion. As a result, sequences of new, i.e. post-fusion images, were obtained. Figure 6 presents examples of post-fusion images, where the image in Figure 6a was obtained using the aforementioned aggregation algorithm based on a shift invariant discrete wavelet transform modification. In turn, for comparative purposes Figure 6b presents the effect of aggregation consisting in averaging the pixel values of input images. Both images contain information from the vision and thermographic images, yet as can be easily seen the averaged image is out of focus. In No. 3/2013 addition, the contrast of the area representing temperature distribution is reduced. The image obtained through SIH-based aggregation is characterised by a much higher contrast and sharpness of contours. The effect of the thermographic image on the resultant image is significant, which facilitates the description of the image both by means of measures taking into account welding arc luminous intensity and those based on temperature distribution. value image pixel. In the tests, the following parameters were taken into consideration (Fig.7): left external width and right external width (LEW, REW); left and right internal width (LIW, RIW); central width (CW); left and right half width (LHW, RHW) as well as total width (TW). The values of determined features were applied to determine diagnostic signals used for recognising the state of a welding process. Analysis of images after fusion The new images created through fusion were analysed in order to determine the set of features used for recognising welding process states. Three types of features were determined for the specified area of interest containing the welding arc area. The first type includes first- and second-grade statistical features based on the vector of the bar chart and Co-Occurrence Matrix (COM). The features in question include the average, variance, coarseness, kurtosis, contrast, correlation, entropy etc. The second type includes topological features describing the geometric properties of a previously binarised welding arc area. During the tests the following topological features were taken into consideration: shape factor, area, Feret coefficient, circumference, Malinowska coefficient, longer diagonal and shorter diagonal. The third type was composed of features resulting from the analysis of the horizontal profile of the welding arc area. The horizontal profile passing through the weld pool area has a pseudosymmetric character resulting from welding process properties. Disorders in the profile symmetry may indicate the instability of a welding process. This assumption made it possible to indicate the profile features which can be the widths between the abscissas of characteristic points, where characteristic points are the extrema of the horizontal profile passing through the maximum No. 3/2013 Fig. 7. Manner of determining features basing on the linear profile of post-fusion image pixel values The diagnostic signals determined as a result of post-fusion image analysis were estimated using the following point statistical features known in technical diagnostics: average value, root-mean-square value, peakto-peak value or kurtosis. The obtained numerical values of signal features were the basis for the operation of recognising welding process states by way of classification. Figure 8 presents the scheme of obtaining models describing various welding process states. Recognition of welding process state The recognition of a welding process state involved the development of models describing each of the welding process states simulated during the experiment. Each model was a one-dimensional vector consisting of the values of the features determined for a given welding process. The models were used to form both a training set and a testing set, in accordance with principles of building and teaching classifiers. BIULETYN INSTYTUTU SPAWALNICTWA 17 image sequence 1 2 bf2 bf1 3 bfa sf2 sf1 sfb tf2 tf1 tfc Diagnostic signals – courses of image feature values band cp1 1 cp2 bf1 ... statistical cp16 cp1 cp2 sf1 ... topological cp16 cp1 cp2 2 tf1 ... cp16 720 Classification Fig 8. Scheme of obtaining models describing welding process examples in the teaching and testing sets. The process of state recognition was considered for two cases: • classifier made an assignment to one of the states from z0 to z7, which means that both the examples, for which welding proceeded properly and those with simulated disturbances leading to welding imperfections were considered; • classifier made an assignment to one of the states from z1 to z7, which means that only the examples with simulated disturbances leading to welding imperfections were considered; in other words a distinction between undesired states was made. Prior to a state recognition, on the basis of the set of models determined on the basis of post-fusion images (FU), classification was also performed for the models determined separately for the thermographic images (IR) (1) and images recorded in the range of visible = radiation (TV). The models for the therwhere Ng is the number of properly recog- mographic images and vision images were nised examples and N is the number of all built on the same principle as in the case of For the purpose of state recognition it was necessary to use a minimum distance classifier using the k-nearest neighbour algorithm, where k amounted to 7. The value of parameter k was selected experimentally on the basis of preliminary tests. The estimation of a classifier error was carried out using n-fold cross-validation (leave-one-out). In this method the process of training and testing is repeated N-times, where N is a number of teaching examples. In each iteration the teaching set has N-1 elements, whereas the testing set has one element. In this method each element of the set of teaching examples is a testing example. The method leave-oneout was used because of a not too numerous set of teaching examples. The relative number of correct assignments was used as a measure of state recognition efficiency: 18 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 Figure 9a presents the comparison of the post-fusion images. The classification on the basis of separately determined features maximum, minimum and medium (marked of the thermographic images and vision im- with a square) values of the classification efages made it possible to assess the usability ficiency for the diagnostic signals determined of image fusion. Table 2 presents the values for individual sequences of images. The seof classification efficiency for considered quences of post-fusion images made it possiwelding process states obtained for select- ble to obtain the highest values of classificaed features and selected diagnostic signals tion efficiency eerr = 0.59, where the average determined on the basis of the sequence of value of efficiency is also higher than in the thermographic images (IR), vision images case of the diagnostic signals determined (TV) and post-fusion images (FU)TV. The from the sequences of TV and IR images. cases selected were those for which the aver- Taking into consideration only the states repage value of classification efficiency was the resenting the occurrence of welding imperhighest. The highest classification efficiency fections (Fig. 9b) one can notice that the best was achieved for the diagnostic signal ob- results were obtained for the signals generattained by determining the average value of ed from the sequences of the post-fusion imthe signal containing the values of the total ages, eerr = 0.61. Very poor results were obwidth of a characteristic band (TW), calcu- tained for the use of the signals obtained from lated for the sequence of images received as the sequences of the thermographic images a result of the fusion (FU). eerr = 0.44, with very similar average values. While analysTable 2. Classification efficiency in recognising individually identified welding process ing the achieved states for the classifier of the maximum total efficiency results one can State Mean Mean also notice that Signal for the images z0 z1 z2 z3 z4 z5 z6 z7 z0-z7 z1-z7 obtained as a re0,91 0,17 0,86 0,29 0,71 0,17 0,43 0,00 0,37 0,44 sult of fusion, the IR Variance classification reTV RIW 0,91 0,67 0,86 0,29 0,57 0,33 0,71 0,14 0,51 0,56 sult for the state RMS representing the properly con- FU TW 0,45 0,67 1,00 0,57 0,86 0,17 0,86 0,00 0,59 0,57 absolute mean ducted welding process is worse than the classification results obtained on the Using appropriate methods for the selecbasis of the sequence of IR and TV images tion of features it is possible to select feaonly. In the case of the states representing tures enabling (in many cases) error-free imperfections, the assessment of post-fusion recognition of some states. Such an approach images significantly improved classifica- makes it possible to increase the efficiency tion efficiency. The states of special interest of classification and thereby identify states are z5 and z7, for which very low results of which are most difficult to recognise. The classifier efficiency obtained for the thermo- tests resulted in increasing the efficiency of graphic images are directly reflected in the classification using groups of seven classiresults obtained for the classifiers based on fiers, each of which distinguished whether the signals generated from the sequence of a given case belonged to one of the identipost-fusion images. fied classes or not. Each classifier could opNo. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 19 tent of such signals. In the case of a multi-mode vision observation the reduction of the number of sequences to be analysed, and thus the limitation of the number of determined diagnostic signals can be achieved by merging images using image fusion methods. Fig. 9. Classification efficiency values: The conducted tests rea) during recognising all identified states,b) during recognising states vealed that the sequences with welding imperfections (set of signals; set of signals) of images obtained as a erate on a different previously determined result of the fusion of IR (thermographic) feature. Table 3 presents the maximum clas- and TV (recorded in visible radiation) imsification values obtained for states under ages can be the basis for obtaining better consideration. The error-free recognition of and more important diagnostic signals than states z0, z1, z2 and z4 was obtained for the those which can be obtained by analysing models created from the features of all con- the sequences of input images (IR and TV) sidered diagnostic signals. States z6 and z7 for the fusion process. The manner of imcreated as a result of welding arc instability age analysis significantly affects the qualand improper weld groove geometry were ity of the obtained diagnostic signals. All characterised by the lowest recognisability. imperfections such as, for instance, image It was not possible to achieve error-free rec- mismatching related to possible momentary and sudden displacements of cameras durognition for any of these states. ing image acquisition signifTable 3. Classification efficiency in recognising individually identified icantly deteriorate the final welding process states (maximum values) state recognition quality. It State should also be noted that the Signal z0 z1 z2 z3 z4 z5 z6 z7 selection of a feature used to IR 1,00 1,00 1,00 0,86 1,00 1,00 0,86 0,86 describe the whole sequence TV 1,00 1,00 1,00 0,86 1,00 0,67 0,71 0,71 significantly affects the final FU 1,00 1,00 1,00 1,00 1,00 0,83 0,86 0,86 diagnostic outcome. The obtained results indicate that Summary the use of image fusion enables the efficient Recognising state changes is the prima- identification of various welding imperfecry diagnostic objective often obtained by tion generated during welding. way of model recognition. For this reason obtaining proper diagnostic signals being Research work financed the basis for the generation of models is from funds for science an important, yet complicated task. Makin the years 2009-2012 ing use of too many diagnostic signals is not advantageous either, and therefore it is advisable to maximise the information cona) 20 b) BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 References 1. Burt, P., and Kolczynski, P (1993).“En- 5. Kim, Y. S., Lee, J. H., and Ra, J. B. (2008). Multi-sensor image registration based on hanced image capture through fusion”, th intensity and edge orientation information. paper presented at Computer Vision 4 Pattern Recognition, 41, pp. 3356-3365. International Conference, Proceedings, 6. Rockinger, O. (1997). “Image Sequence pp. 173–182. Fusion Using a Shift Invariant Wavelet 2. Jamrozik, W., and Fidali, M. (2011). Transform”, paper presented at InternaMetody dopasowania termogramów i tional Conference on Image Processing, obrazów wizyjnych dla dynamicznie Proceedings, vol.3, pp. 288-291. zmieniającej się struktury obserwowanej sceny. Measurement, Automation and 7. Nandhitha, N. M., Manoharan, N., Rani, B.S., Venkataraman, B., Sundaram, P. K., Monitoring, 57 (10), pp.1134-1137. and Ra, J.B. (2006). “Automatic Detec3. Jamrozik, W., and Fidali, M. (2012). tion and Quantification of Incomplete “Evaluation of the suitability of IR and Penetration in TIG Welding Through TV image aggregation algorithms for Segmentation and Morphological Image the purposes of welding process assessProcessing of Thermographs”, paper prement”, paper presented at 11th Quantitasented at National Seminar on Non-Detive InfraRed Thermography Conference, structive Evaluation, Hyderabad, 7-9 DeNaples, Italy, 11-14 June. cember, Proceedings, pp. 17-22. 4. Jamrozik, W., and Fidali M. (2011). Estimation of image fusion methods for purposes of vision monitoring of industrial process. Measurement, Automation and Monitoring, 57 (09), pp. 993-996. No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 21 Antoni Sawicki Damping factor function in AC electric arc models. Part 2: Damping factor function in universal electric arc models with moderate cooling Abstract: Nonlinear function of the heat processes damping factor in thermal col- umn plasma have been introduced to modified mathematical models of the AC electric arc. This way, universal arc models with moderate cooling and constant geometric sizes, using static characteristics, have been created. Similar method has been used to hybrid arc models with variable column length connecting Berger and Kulakov models. It has been proposed the introduction of nonlinear damping factor function to modified Voronin model using static characteristics. Thereby the extended range of current for application of these models has been achieved. Keywords: electric arc, damping factor, time constant, Mayr model, Cassie model, Berger model, Kulakov model, Voronin model, hybrid model Introduction Most of today’s arc and plasma-arc welding devices are designed as universal tools used for joining and cutting elements of various geometrical dimensions (thickness) and shapes made of various materials. Such processes require the use of diverse torches, electrodes, shielding gases, current pulses of various amplitudes, shape, packing degree, polarity and frequency. In variable operating conditions, the systems of automatic control ensure stable electric arc burning, high quality of technological processes, high production efficiency, low material consumption as well as minimum noxiousness to personnel, the environment etc. However, such versatility results in the occurrence of a wide range of state variables, the appearance of various non-linear static and dynamic characteristics of supply systems and electric discharge. Increasing advancements in computer-aided methods for designing electrotechnological and electrothermal welding devices require more and more accurate arc discharge models, precisely reproducing the non-linearity and dynamics of thermal and electric processes. However, the earlier imperfections of the experimental determination of the dynamic characteristics of arc discharges are accompanied by imperfections in their mathematical modelling [1]. Despite emerging compromises between the required accuracy of the reproduction of physical processes and the ease of measurements, the simplicity of interpretation, the low complexity and short computational time, the most popular arc models still fail to ensure appropriate precision. This is largely due to the wide range of exciting current changes and external effects causing changes to column geometrical dimensions. dr hab. inż. Antoni Sawicki (PH.D. Eng), professor at Częstochowa University of Technology Wydział Elektryczny /Faculty of Electrical Engineering/ 22 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 This study presents universal electric arc models with moderate cooling of the extended current range of approximations of static and dynamic characteristics. The objective of this work required the application of modified mathematical models using static characteristics and hybrid models, matching the Mayr and Cassie models. The use of the non-linear damping factor function corresponding to experimentation results was suggested [1] in the models. In modelling electric arc dynamic characteristics a damping function (or its specific value – time constant) along with other characteristics and dynamic parameters of a specific model constitute the whole complex of closely related quantities approximating the course of real physical processes with a pre-set accuracy. Due to random disorders, obtaining repeatable arc experimentation results with a pre-defined accuracy is very difficult. Even more problematic is matching several characteristics determined in various separately conducted experiments (e.g. static current-voltage characteristics, damping factor function) in a single model. This process cannot be carried out by means of simple analytical or even numerical methods but requires the use of complex procedures for the identification of electric arc mathematical model characteristics. Damping function in electric arc models with moderate cooling using static characteristics arc voltage, which in the case of arc burning coaxially with electrodes corresponds to electric field intensity Estat. The shapes of typical static characteristics of electric field voltage and intensity are presented in Figure 1. In the case of strong currents it is possible to assume that static voltage Ustat does not depend on current. Ustat(l) = as + bsl(2) According to formula (1), the generalised power static characteristic is described by the formula Pstat=Ustat(l, I)·I= = (as +bsl)·I+(cs+ dsl)·Ii-n (3) which in accordance with formula (2) in the case of string currents adopts a simpler form: Pstat=Ustat(l)·I=(as+bsl)·I(4) If and arc length does not change (l = const), the static voltage-current characteristics is the following: Ustat(I)=A+BI-n(5) where A=as+bsl = const; B = cs+dsl = const. Similarly, the power of an arc is a non-linear current function: Pstat(I)=Ustat·I=f(I)=AI+BI1-n(6) and in the case of strong currents the dependence becomes linear Pstat=Ustat·I=I·const(7) A modified arc model by Mayr [2, 3] can One of the most commonly known general be presented in a general conductance form static characteristics of an electric arc with 1 dg 1 Pkol (t ) (8) moderate cooling is the dependence provided g dt = θ (i(t )) P (t ) − 1 dys M by Nottingham: where g – column conductance; i – alternatcs + d s l U stat (l , I ) = as + bs l + (1) ing current intensity; Pkol – electric power In supplied to plasma column; Pdys – power of where l – column length; I – direct current energy dissipated from column; θM(i) – intensity; as – sum of near-cathode and near- damping function corresponding to the time -anode voltage drop; bs – gradient of deflected of relaxation of heat processes. Electric No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 23 where A’ = bsl = const. The power of losses in disequilibrium plasma PE = as |i| of very thin near-electrode areas is taken into consideration separately. More often applied are static voltage-current characteristics (1), (2) or (5), and then i2 Pdys (t ) = U stat (i ) ⋅ i = (12) G stat (i ) 0 After substituting formula (12) to formula (10) it is possible to obtain the modified Mayr equations in the conductance form 1 dg 1 Gstat (i ) = − 1 g dt θ M (i ) g (13) Fig. 1. Static characteristics of arc: Ustat(I) - voltage-current, Estat(I) – electric field intensity The shapes of the typical static characteristics of power and column conductance are power supplied to thermal plasma is ex- presented in Figure 2. pressed by the formula below: i2 Pkol (t ) = u kol i = g (9) where ukol – voltage drop on arc column. In the classical Mayr model it is assumed that Pdys(t) = const. In the range of stronger currents this condition cannot be met any longer and usually the Cassie model is used instead [3]. As heat dissipation processes react slowly to external disturbances, one can roughly assume that a power loss is principally determined by static characteristics [4], Pdys(t) ≈ P’stat(i(t)), i.e. 1 dg 1 ukol i = − 1 ' (i ) g dt θ M (i ) Pstat (10) Fig. 2. Arc static characteristics: Pstat(I) - power-current, Gstat(I) - conductance-current At this moment, when conductance does not change in time, the static characteristics of an arc in this model has the following form: where P’stat – takes into consideration power losses only in column plasma, without U (i ) = Pstat (i ) (14) stat i near-electrode areas. The function of loss power can be approximated by means of dei2 i2 pendences (6) or (7) allowing a generalised Gstat (i ) = = (15) ( ) ( ) P i U i ⋅ i stat stat dynamic arc model to be obtained: u i 1 dg 1 (11) = ' kol 1−n − 1 Thus, based on this it is possible to write g dt θ M (i ) A i + B i model (13) in a conductance form 24 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 i 1 dg 1 = − 1 ' (i ) g dt θ M (i ) g ⋅ U stat (16) • Cassie model g (t ) ≈ g C (t ) = u kol iC dg C , if i > I0 (21) − θ C U C2 dt which on the basis of formulas (5) and (10) where PM – constant power of the Mayr leads to the generalised dependence model; θM – time constant of the Mayr model (0<θC<<θM); UC – constant voltage of the i 1 dg 1 1 = − (17) Cassie model; θC – time constant of the Casg dt θ M (i ) g ⋅ A' + B i −n sie model (0<θC<<10-3 s). In comparison with the classical Mayr modFigure 3 presents the shapes of the typical el [2], models (11) and (17) can significant- static characteristics of voltage and power. ly more accurately reproduce the static and Adopted approximations in the form of the dynamic characteristics of an arc in a wid- constant values of Cassie model voltage and er range of current changes. This results not Mayr model power are marked against their only from the variation of a time constant background of the aforesaid static characterobtaining the form of the damping function istics. Both models also assume the constant θM(i), but also from applying a more accu- values of a damping factor, which is presentrate function approximating static voltage- ed in Figure 4 against the background of typical characteristics θ(i). -current or power-current characteristics. ( ) Damping function in arc TWV hybrid model matching Mayr and Cassie models In the TWV hybrid model of an arc [5] the fractions of currents flowing through two parallel non-linear conductances, corresponding to the Mayr and Cassie models, depend on their resultant value. Thus, it is possible to write i2 i (t ) = u kol g = u kol ⋅ g M exp − 2 + I0 (18) i2 + u kol ⋅ g C ⋅ 1 − exp − 2 I0 where I0 – limiting current between the Mayr and Cassie models. Hence we obtain i2 i2 g (t ) = g M (t ) ⋅ exp − 2 + g C (t ) ⋅ 1 − exp − 2 I0 I0 (19) Fig. 3. Static characteristics and dynamic parameters of arc: Ustat(I) - voltage-current, Pstat(I) – power-current, PM(I)=const – Mayr model power, UC(I) = const – Cassie model voltage, I0 – limiting current between the Mayr and Cassie models On the basis of formulas (18)-(21) a hybrid model is created [5] g = G min + [1 − ε (i )] u kol i i2 dg ( ) i + − θ MC (i ) ε (22) 2 PM dt UC The model selections conditions are the following: where θMC(i) – damping factor function of • Mayr model 2 i dg M , if i < I0 (20) TWV model. g (t ) ≈ g M (t ) = M − θ M In practical considerations it is usually asPM dt No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 25 corresponds to an increase in thermal power dQa dQa dl dl = = ql dt dl dt dt Fig. 4. Characteristics and dynamic parameters of arc: θ(i) – characteristics of damping factor, θM –Mayr model time constant, θC –Cassie model time constant (25) where ql – arc energy linear density. Therefore, in simplifying conditions [8] thermal power required for the generation of additional plasma volume is roughly proportional to the rate of a length increase. This is accompanied by some relaxation times resulting from gas thermal inertia and additional cooling of a column. As the arc voltage grows along with an increase in a column length, in publication [8] the following approach to determining Cassie model voltage component was suggested (21): sumed that Gmin = 0 [5, 6]. Then, formula u2 (l) = al(26) c (22) can be written in a simplified form where the parameter a is almost constant in u kol i u kol i 1 dg 1 ( ) ( ) [ ] = 1 − + − 1 i ε ε i (23) the wide range of current i changes. In turn, g dt θ MC (i ) PM gU C2 additional power pv(dl/dt) dissipated from where the designation of a tapering function the column is determined by u2c g; due to diswas introduced sipativity, the lack of even partial conservativity and the return of energy to the circuit i2 ε (i ) = exp − 2 (24) leads to the following dependence [8]: I0 dl dl The form of this function may vary and >0 if b (27) dl dt depends on the type of an electrotechnologidt pv = cal device and its operating conditions [7]. dl dt Damping function in variable length arc hybrid model matching Berger and Kulakov models 0 if dt ≤0 A modified Cassie equation with the dl variable value of voltage U C (t ) = U C l , While considering the models of an elec dt tric arc with the variable length of a plasma enables obtaining the conductance form of column it is possible to modify separate Cas- the Berger model [8] sie and Mayr models retaining the constant value of damping factors (time constants) 2 u kol 1 dg 1 and current ranges preferred by them. = − 1 (28) An increase in the geometrical dimensions g dt θ CB u 2 (l ) + 1 p dl v C of the plasma column of a high-current arc is g dt accompanied by an increase in energy necessary for the generation of additional plasma where pv(dl/dt) - power needed for the genvolume. The assumed axial-cylindrical shape eration of additional plasma volume; θCB – of a column and its tension by a length dl time constant of Cassie-Berger model. 26 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 Kulakov suggested the modification of a 1 dg = 1 [1 − ε (i )] u2kol + ε (i ) u kol − 1 (33) l ⋅ E stat (i ) uC (l ) low-current arc model (16) taking into con- g dt θ BK (i ) sideration column length changes. The After taking into consideration approximafirst-degree Mayr-Kulakov model in the con- tions (26) and (30) the equation is as follows: ductance form is the following [9]: 2 u kol 1 dg 1 = + [1 − ε (i )] 1 dl i 1 dg 1 ( ) g dt i al θ BK = − 1 − (29) (34) g dt θ MK g ⋅ l ⋅ Estat (i ) l dt u kol + ε (i ) − 1 where Estat(i) – non-linear static characteris−n l ⋅ bs + d s i tics of electric field intensity; θMK – time constant of the Mayr-Kulakov model. One of the The matching of the Berger and Kulakov imperfections of approximation by means of models as well as using the non-linear functhis model is overlooking the impact of plas- tion of a damping factor makes it possible to ma physical properties on conductance dy- use the hybrid model for simulating processes in electrotechnological devices operating namics during column length changes. Using the dependence approximating the with a wide range of current and an arc disstatic voltage-current characteristics (1) it is charge column length. possible to determine the static characterisDamping function in Voronin model tics of electric field intensity of arc with variable geometrical ∂U stat (l , I ) ds Estat (I ) = = bs + n (30) dimensions ∂l I The Voronin model makes it possible to Then, the Kulakov model has the following form: take into consideration an external influence exerted on the length and diameter of a cy i 1 dg 1 1 dl = − 1 − (31) lindrical column. In order to create such a −n g dt θ MK g ⋅ l ⋅ bs + d s i l dt model it is necessary to make a number of simplifying assumptions [9, 10]. The model The arc column hybrid model, taking into basis is a simplified equation of the thermal consideration the changes of an arc length, balance of a column. It is assumed that the matches models (28) and (29) in the manner dissipated power is proportional to the side (23) by means of an appropriate tapering area of an arc: function ε(i). Therefore the model has the P (l , S ) = p l 4πS (35) dysS S following form: As a result, an arc model with variable geomet 2 u kol 1 dg 1 rical dimensions S(t) and l(t) of the following = + [1 − ε (i )] 1 dl g dt θ BK (i ) general conductance form is obtained [10]: uC2 (l ) + pv g dt (32) 2 ( ( ) u kol 1 dl + ε (i ) − 1 − ε (i ) g ⋅ l ⋅ E stat (i ) l dt where θBK(i) – damping factor function of the hybrid Berger-Kulakov model. If a relatively low rate of arc length changes is assumed (dl/dt≈0), equation (32) can be written as No. 3/2013 ) 1 dg 1 ukol i = − 1 + g dt θ S (S ) PdysS (l , S ) − g l 1 dl 1 + ln + l dt K g S + g l 1 dS 1 − ln S dt K g S BIULETYN INSTYTUTU SPAWALNICTWA (36) 27 θ (S (i )) ∝ S (i ) ∝ d (i ). In turn, in theoretical deliberations [11] the relation θ (S (i )) ∝ S (i ) where a damping factor function is θ S (S ) = Q0 Sl Q = 0 PdysS (l , S ) pS S 4π (37) is assumed. However, experimental tests [5, 11] −1 reveal almost a reverse tendency θ (i ) ∝ i , and Q0 – reference factor, J/m3; Kg – coeffi- which should be recognised as a real one, cient of unitary conductance approximation, especially due to the fact that simplified cyS/m; l – length of arc column, m; pS – density lindrical arc models take into consideration of power dissipated by the side surface of a only selected dominant heat processes, and column, W/m2; S – area of arc cross section, entirely pass over gasodynamic and electrom2. All three parameters Q0, Kg, pS are deter- magnetic processes (e.g. effects of contracmined on the basis of experiments and are tion and gas pumping by a column). Figure 5 presents a typical shape of charassumed to be constant quantities. If the relative rate of arc length changes is acteristics d(i) and S(i). One can see that low (dl/dt≈0), equation (36) can be written as there is no correlation of these quantities with a damping factor function. To some ex 1 dg 1 ukol i tent an increase in θ in areas where current = − 1 + g dt θ S (S ) PdysS (l , S ) passes through zero can be explained by the (38) significant weakening of a contraction effect g l 1 dS 1 − ln + and momentary plasma expansion. S dt K S g Models (36) and (38), similarly as the Mayr model, reproduce arc characteristics in the low-current range relatively well. Therefore they can be used to calculate processes in devices with a relatively low temperature of the area surrounding a discharge. Such conditions occur in open arc welding or during melting of a charge at the initial stages of arc furnace operation. While considering a free or quasi-free arc, in which the area of column cross section is primarily determined by the module of current value, using equation (36) one obtains ukol i 1 dg 1 = − 1 + g dt θ S (i ) PdysS (l , S (i )) − g l 1 dl 1 + ln + l dt K g S (i ) + g l 1 dS di 1 − ln S (i ) di dt K g S (i ) While considering the case of a free or quasi-free welding arc, on the basis of experimentation and theoretical analysis [12] on can adopt a dependence related to an arc column diameter d=k|i|m(40) (39) Adopted assumptions simplifying model (36) lead to almost a linear dependence of a damping function (37) on an arc column diameter 28 Damping function in Voronin modified arc model using static characteristics Fig. 5. Arc static and dynamic characteristics of: θ(I) – damping factor, d(I) – column diameter, S(I) – area of column cross section BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 As heat dissipation processes react slowly to rameter, i.e. the surface density of thermally external disturbances, just as previously it is dissipated power: −n possible to roughly assume that bs + d s i Pdys(t) ≈ P’stat(i(t)), which on the basis of for- pS = (45) m −1 πk i mula (35) leads to the following dependence: PdysS (l , S (i )) = p S l 4πS (i ) = E stat (i ) ⋅ li (41) Concluding remarks Thus, the intensity of a field is expressed by 1. Introducing a damping function non-linthe following formula: early dependent on current extends the possiEstat(i) = πpsk|i|m-1 sign(i) = (bs+ds|i|-n)·sign(i) bilities of applying dynamic arc mathematical models using static characteristics. (42) 2. Introducing a damping function nonHence, on the basis of formula (36) the fol- -linearly dependent on current strengthens the versatility of arc hybrid models using lowing dependence is obtained: sub-models of a low and high-current arc. 1 dg 1 i 3. The developed mathematical models = − 1 + g dt θ S (i ) gEstat (i ) ⋅ l include a wide range of changes of current and those of plasma column geometrical di1 dl g l (43) mensions. As a result, they can be used to 1 + ln − + l dt K g S (i ) simulate processes in various arc electrotechnological and plasma devices (both weld1 dS (i ) di g l 1 − ln + ing and electrothermal ones) with moderate S (i ) di dt K g S (i ) cooling of a discharge area. If compared with the Kulakov model (29) this model significantly extends the possi- Research work financed from funds bility of approximating characteristics as it for science in the years 2010-2013 as takes into consideration not only the dynam- research project no. N N511 305038. ics of length changes but also the dynamics References of column cross section changes. Slow arc length changes (dl/dt ≈ 0) corre- 1. Sawicki, A. (2013). Damping Factor spond to the following equation: Function in AC Electrical Arc Models. Part 1: Heat Process Relaxation Phenom 1 dg 1 i = − 1 + ena, their Approximations and Measureg dt θ S (i ) gEstat (i ) ⋅ l ment. Biuletyn Instytutu Spawalnictwa, (44) 57 (2), pp. 37-41. 1 dS (i ) di g l 1 − ln + 2. Ciok, Z. (1987). Modele matematyczne S (i ) di dt K g S (i ) łuku łączeniowego. Warsaw: Warsaw University of Technology. If in formula (40) a typical value m= ⅔ [12] is assumed, using formula (42) the 3. Лесков, Г.И. (1970). Электрическая −1 / 3 сварочная дуга. Изд-во Машиностроdependence Estat (i ) ∝ i is obtained. This ение, Москва. dependence corresponds to applied approximations of electric field intensity in a 4. Sawicki, A. (2012). Modele dynamiczne łuku elektrycznego wykorzystujące chalow-current arc with moderate cooling of a rakterystyki statyczne. Śląskie Wiadomocolumn [4]. In addition, formula (42) can ści Elektryczne, (6), pp. 13-17. be used to determine the basic model paNo. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 29 5. Tseng, K.J., Wang, Y., and Vilathgamuwa, D.M. (1996). “Development of a Dynamic Model of Electric Arc for Power Electrics Simulations”, paper presented at IEEE Industrial Applications Conference, IAS’96, Proceedings, vol. 4, pp. 2173-2180. 6. Sawicki, A. (2012). Modified Habedank and TWV hybrid models of the arc with variable length for simulating processes in electrical devices. Biuletyn Instytutu Spawalnictwa, 56 (1), pp. 45-49. 7. Sawicki, A. (2012). Funkcje wagowe w modelach hybrydowych łuku elektrycznego. Śląskie Wiadomości Elektryczne, (5), pp. 15-19. 8. Berger, S. (2006). “Mathematical approach to model rapidly elongated 30 free-burning arcs in air in electric power circuits”, paper presented at ICEC 2006, 6-9 June, Sendai, Japan. 9. Sawicki, A. (2012). Modelowanie łuku elektrycznego o zaburzanej długości kolumny plazmowej. Śląskie Wiadomości Elektryczne, (1), pp. 9-17. 10.Воронин, А.А. (2009). Повышение эффективности контактно-дугогасительных систем сильноточных коммутационных аппаратов с удлиняющейся дугой. Автореферат дис. к.т.н. Самара. 11.Kalasek, V. (1971). Measurements of time constants on cascade D.C. arc in nitrogen. Technische Hogeschool Eindhoven, TH-Report 71-E-18. 12.Коперсак, В.М. (2011). Теорiя процесiв зварювания-1. КПI, Киiiв BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 Ryszard Krawczyk, Piotr Wojtas, Karolina Poch Comparative assessment of selected welding imperfections in VT, PT and MT methods Abstract: The issues connected with assessment of sensitivity in selected NDT methods are considered. In particular, the attention has been paid to the tests which make it possible to detect any imperfection of surface, i.e. visual testing, liquid-penetrant and magnetic-powder method. The aim of the studies was to reveal the analogy and differences between results obtained from series of tests (VT, PT and MT). Keywords: non-destructive testing, magnetic particle inspection (MT), penetration testing (PT) Introduction Importance of surface NDT The non-destructive testing of welding materials and welded joints is classified into two categories. The first category, comprising tests enabling the detection of surface imperfections is comprised of three methods, i.e. visual testing (VT), penetration testing (PT) and magnetic particle testing (MT). The second category, comprising non-destructive tests enabling the detection of internal imperfections, often referred to as volumetric imperfections, producing no superficial symptoms or indications, includes two methods, i.e. ultrasonic testing (UT) and radiographic testing (RT). The testing and assessment of surface imperfections of welding materials and welded joints are a very important element of technological and operating processes. Conducting tests makes it possible to verify that materials or structural elements being tested do not contain surface imperfections directly compromising the safety of operation and the possibility of further use of an element being tested. The visual test of welded joints is a basic test which must be carried out irrespective of what other tests will follow. The visual test consists of a thorough inspection of te element being tested in order to assess the state of its surface. The conditions which must be satisfied during VT are specified in detail in standard PN-EN ISO 17637:2011 ”Non-Destructive Testing of Welds – Visual Testing of Fusion-Welded Joints”. Penetration testing is based on the phenomenon of capillarity, i.e. water penetrating narrow spaces and rising within them despite inversely directed gravitational force. This process makes it possible to detect even the smallest discontinuities present inside a material and coming up to its surface. In turn, the essence of magnetic particle testing is focused on two aspects, the first of which involves introducing a magnetic field to the object being tested, whereas the other refers to the manner of interpreting revealed indications. Both elements are interdependent as in order to assess the effectiveness of the dr inż. Ryszard Krawczyk (Ph.D. Eng), mgr inż. Piotr Wojtas (M.Sc. Eng), mgr inż. Karolina Poch (M.Sc. Eng) - Częstochowa University of Technology, Zakład Spawalnictwa /Welding Technology Department / No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 31 magnetic field introduction it is necessary to apply one of the methods used to identify any detected imperfections. Each of these testing methods (VT, PT and MT) is based on different phenomena and is characterised by different testing sensitivity. It is assumed that VT is characterised by the lowest, PT by higher and MT by the highest testing sensitivity. Having taken this into account, an attempt aimed at comparing detection sensitivity of the aforesaid testing methods, particularly PT and MT, was undertaken. Fig. 1. Testing station prepared for tests Tests – preparation of sample and testing station wards, the sample underwent etching. The processing of the sample aimed to decrease the effect of weld reinforcements and significant roughness of the element, which could significantly deteriorate testing accuracy. The sizes of the sample used in tests were 220×130×8 mm. In order to carry out non-destructive testing (VT, PT and MT) it was necessary to prepare a testing station provided with measuring equipment following related standards (PN-EN 13927: 2009, PN-EN ISO 3452-4: 2006 and PN-EN ISO 9934-3: 2003). The testing station was equipped with the following devices and materials: • set of reagents for penetration testing: Diffu-Therm penetrant, remover and developer, • device for magnetising an object being tested – contour probe DA400S made by the Parker company, • set of reagents for magnetic particle testing: Diffu-Therm ferromagnetic coloured and fluorescent powder, priming paint and remover, • set of measuring instruments and additional devices: ◦◦ instrument for measuring tangent and residual areas, ◦◦ instrument for measuring the intensity of ultraviolet and visible light, ◦◦ Berthold magnetic field indicator, ◦◦ reference samples, ◦◦ ultraviolet lamp, ◦◦ straightedge, slide caliper, ◦◦ magnifying glass. The objective of the conducted experiments was to reveal differences and relations between the results of visual, penetration and magnetic particle testing methods carried out on the same object, i.e. the same imperfections. To this end, it was necessary to make test V-preparation butt joints of 10mm-thick S355 steel plates. During MAG welding the arc energy was significantly reduced, resulting in the VT TESTING lack of side fusion on both x y d No. fusion lines. The external [mm] [mm] [mm] surfaces of the sample were 1 124 4 1 subjected to grinding, by 2 173 4 0,3 means of which weld face 3 188 4 0,4 reinforcement and weld Fig. 2. Characteristics of imperfections revealed during VT root were removed. After32 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 The equipment and measuring devices used in the tests were provided with check certificates and the reagents had confirmed use-by dates. In order to verify the suitability of reagents, their sensitivity was additionally tested on standards appropriate for a given method. The prepared testing station is presented in Figure 1. The experimental tests (VT, PT and MT) were carried out according to a specified sequence. The first was a visual test consisting of a thorough inspection of the test surface both by the naked eye and using a magnifying glass 3x. The visual testing did not reveal any linear imperfections on the surface of the sample, yet it revealed three blowholes which due to weld face grinding are visible as pores with the following diameters d1=1 mm, d2=0.3 mm and d3=0.4 mm. The characteristics of point imperfections are presented in Figure 2. The second stage consisted of penetration testing by means of the dye penetrant method. The results of the test are presented in Figure 3 showing the sample following the test, the sample sketch with plotted indications and a table with full dimensional characteristics. The penetration testing revealed the presence of twenty one linear indications, the total length of which was 192 mm as well as identified three non-linear indications with the diameters of d10 =2.2 mm, d15 =0.8 mm and d17 =1.0 mm respectively. At the next stage fluorescent magnetic particle tests (MT-F) were carried out. The results of the tests are presented in Figure 4 showing the sample after the test, the sample sketch with plotted indications and a table with full dimensional characteristics. The magnetic particle testing revealed the presence of twenty linear indications, the total length of which was 270 mm. At the final stage, magnetic particle testing using the dye penetrant technique was carried out. The test revealed that the priming paint PT: dye penetrant method No. Fig. 3. View of the sample after PT, sketch of the sample with plotted indications and table with full dimensional characteristics No. 3/2013 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 x [mm] 24 44 48 78 87 98 106 120 124 123 130 141 144 152 172 180 187 44 66 92 112 130 170 213 BIULETYN INSTYTUTU SPAWALNICTWA y [mm] 5 5 5 5 5 5 5 5 5 3 5 5 5 5 4 5 4 -5 -5 -5 -5 -5 -5 -5 l [mm] d [mm] 12 2 10 2 2 6 5 2 2 --9 2 2 13 --2 --4 4 18 16 36 40 3 ------------------2,2 --------0,8 --1,0 --------------33 No. Fig. 4. View of the sample after MT-F, sketch of the sample with plotted indications and table with full dimensional characteristics can close the outlet of irregularities preventing subsequent, e.g. penetration, tests. Figure 5 presents the results of the test. The penetration testing revealed the presence of thirty eight linear indications, the total length of which was 247 mm, and also identified three non-linear indications having the diameters of d10 = 1.2 mm, d15 = 0.4 mm and d17 = 0.5 mm respectively. Analysis of results and summary The analysis involved the comparative assessment of welding imperfections detected during the following tests: • penetration testing with the dye penetrant technique (PT-D), • magnetic particle testing with the fluorescent technique MT-F, • magnetic particle testing with the dye penetrant technique (MT-D). The comparison of the presented test results reveals that detected imperfections and their detections differ in basic dimensions. Following the general assessment of the test sample area it is possible to draw a conclusion that MT-F was characterised by 34 x [mm] MT -F y l [mm] [mm] d [mm] -- 21 5 16 -- 48 5 4 -- 56 5 6 -- 68 5 4 -- 86 5 3 -- 92 5 2 -- 98 5 3 -- 110 5 10 -- 134 5 5 -- 142 5 32 -- 176 5 7 -- 185 5 25 -- 8 -5 4,5 -- 14 -5 4,5 -- 40 -5 16 -- 66 -5 11 -- 90 -5 5 -- 96 -5 2 -- 102 -5 34 -- 142 -5 76 the best detectability of linear imperfections. In this testing the total length of indications amounted to 270 mm, in MT-D 247 mm, and in the penetration method (PT-D) 192 mm. The best result obtained in MT-F is approximately 9% more advantageous than that obtained in MT-B, and 40% better than the result obtained in PT-D. It is largely due to the fact that penetration testing enables the detection of open imperfections which at the same time are present on the surface whilst in magnetic particle testing such requirements do not have to be satisfied. Only MT-F failed to uncover non-linear imperfections. The failure to detect these discontinuities is strictly related to the following factors: • shape of the imperfection – a small oval shape of the discontinuity does not cause sufficiently great interference in the magnetic field, • intensity – a small number of point discontinuities do not cause sufficiently great interference in the magnetic field either. The other two testing methods, in addition to detecting linear indications, also made it 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 MT-D No. Fig. 5. View of the sample after MT-D, sketch of the sample with plotted indications and table with full dimensional characteristics possible to uncover non-linear indications. In PT-D the diameter of obtained indications is, on the average, 130% greater in relation to actual imperfections detected in visual tests, whereas the indications in MT-D are on the average 25% greater than those observed in VT. This leads to the conclusion that indications detected by means of penetration testing do not present the real dimensions of imperfections but rather the indications obtained by the penetrant spread around imperfections on the developer surface. Comparing PT-D and MT-D, in which both types of indications were detected, it is possible to state that magnetic particle testing is characterised by better detectability and that its sensitivity is better by approximately 30%. In conclusion it is possible to state that testing techniques analysed in this study (VT, PT and MT) significantly differ as to the detectability of linear and non-linear imperfections. Therefore it is advisable to always use at least two surface testing methods. The best configuration includes visual testing followed by magnetic particle testing with the use of the dye penetrant method. No. 3/2013 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 x [mm] 28 44 51 64 82 88 96 102 120 124 126 130 140 146 154 164 172 177 173 186 188 193 198 10 16 38 48 58 62 67 73 79 81 92 102 112 125 132 164 172 212 BIULETYN INSTYTUTU SPAWALNICTWA y [mm] 5 5 5 5 5 5 5 5 5 4 5 5 5 5 5 5 5 5 3,5 5 4 5 5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 -5 l [mm] 10 5 8 2 3 6 4 4 2 --2 8 4 3 7 5 3 4 --4 --3 6 4 4 8 7 2 3 4 5 1,5 1,5 8 8 12 6 30 6 38 6 d [mm] ------------------1,2 ----------------0,4 --0,5 ----------------------------------------35 References przemysłowych. Wydawnictwo Instytutu 1. Czuchryj, J., and Sikora, S. (2007). Spawalnictwa w Gliwicach, Gliwice, Podstawy badań penetracyjnych wyPoland. robów przemysłowych. Wydawnictwo Instytutu Spawalnictwa w Gliwicach, 3. Krawczyk, R., Poch, K. (2011). Assessment of sensitivity of selected techniques Gliwice, Poland. for magnetic particle inspection. Biuletyn 2. Czuchryj, J., Drop, M., Sikora, S., and Instytutu Spawalnictwa w Gliwicach, 55 Szymański, A. (2011). Podstawy badań (4), pp. 33-39. magnetyczno–proszkowych wyrobów 36 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 Mariusz Stachurski Non-destructive testing of helically welded pipes made of thermomechanically rolled materials used for sending of combustibles Abstract: In the first part of the paper it has been presented the short information on methods of fabrication of helically welded pipes used for transporting of combustibles. In the second one it has been given the NDT methods used during inspection of the pipes in production plants. In the next parts it has been described the NDT methods for steel pipes focusing first of all on visual, ultrasonic and radiographic ones. The paper has been ended with the information about prospects of gas industry development in Poland and in the world. Keywords: welding, non-destructive testing, helially welded pipes Introduction The golden age of pipelines has come with a permanently increasing demand for raw materials; the excavation of which is an enormously complicated undertaking. As of today, the best solution related to this type of transportation is a subterranean grid of main pipelines made of steel pipes. Poland’s gas distribution system has approximately 17 thousand kilometres of high-pressure gas pipelines and approximately 79 thousand kilometres of distribution network, 23 gas pumping stations, and approximately 4 thousand 1st and 2nd degree pressure reducing and measuring stations. A growing consumption of energy combined with a growing demand for energy carriers, in particular for natural gas, shapes the development of the network of industrial pipelines. Great distances between the sources at which natural gas and oil are excavated and the places of their consumption require the construction of transit pipelines characterised by very high operating parameters, e.g. gas pipelines with a diameter of over 1000 mm and pressure of 10 MPa. Such pipelines must be made of pipes characterised by a strength of over 690 MPa (485MB, 555MB) and very good weldability. As a result, it is possible to observe the development of low-alloy steels (C 0.16%, Si 0.55%, Mn 1.9%), often with microadditions of Nb, Mo, or Ti. Such pipes are obtained by means of controller rolling referred to as thermoplastic or thermomechanical treatment. Pipes for the pipelines mentioned above are produced following the regulations contained in DIN standards or in their American (API5L) and European (EN 10208) counterparts. Large-diameter steel line pipes for transporting liquid and gaseous fuels are used as the following: • seamless pipes (S), obtained by means of hot or cold working methods, • welded pipes (W), produced by means of electric induction welding (EW) or butt welding (BW) as well as submerged arc welding (SAW) with a longitudinal weld (SAWL) or a helical weld (SAWH). Figure 1 presents the process of welding pipes with a helical weld (two-stage). Pipes can dr inż. Mariusz Stachurski (Ph.D. Eng) - Bureau Veritas Poland ltd. No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 37 also be welded using the combination of gas-shielded welding of root runs and submerged arc welding of filling layers (COW) [1]. Second stage submerged arc welding (with three automatic welding machines simultaneously) First stage tack weld – MAG welding Fig. 1. Two-stage helical welding of pipes used by the German company of Salzgitter Mannesmann Grossrohr GmbH [5] This article presents non-destructive testing of thermomechanically rolled helically submerged arc welded (SAWH) pipes L485MB and L555MB. These pipes were selected as they are typical and manufactured in many steelworks both in Poland and overseas. Following their manufacture the pipes undergo insulation application, usually with polyethylene or polypropylene and are painted inside with epoxy paints. Non-destructive testing used during pipe manufacture General information The basic purpose of non-destructive testing (NDT) is to assess the condition of a material or welded joint and, on this basis, to issue an opinion related to the quality, durability and safe operation of a structure. The effectiveness and quality of non-destructive testing depends on numerous factors: • personnel competence based on adequate training, • experience, • inspection procedures, • equipment, • environment in which testing is carried out, • psychological pressure, to which NDT personnel are exposed, • working time during the day, 38 • equipment features, • supervision and how it is exercised, • applied standards, requirements and guidelines, • systems applied for the certification of personnel, procedures and equipment [1]. The basic NDT methods used in the production of pipes are the following: • visual testing – VT, • radiographic testing of pipe ends – RT, • automatic fluoroscopic testing of a whole welded joint, • ultrasonic testing of a burden material in the form of a strip unwound from a coil (laminar imperfection test) – UT, • automatic ultrasonic testing of a whole welded joint – UT, • manual ultrasonic testing of a welded joint on a pipe end – UT, • manual ultrasonic testing of a welded joint being repaired – UT. In addition, in order to determine precisely the size of welding imperfections, also other NDT methods, i.e. penetration testing (PT) and magnetic particle testing (MT), can be used. The basic standard describing the selection of testing manners and assessment criteria is standard PN-EN 10208-2 “Steel pipes for pipelines for combustible fluids – Technical delivery conditions – Part 2: Pipes of requirements class B”. The standards describing testing manners and assessment criteria are the following: • PN-EN 10246-9 Non-destructive testing of Steel Tubes – Part 9: Automatic Ultrasonic Testing of the Weld Seam of Submerged Arc Welded Steel Tubes for the Detection of Longitudinal and/or Transverse Imperfections; • PN-EN 10246-10 Non-destructive testing of Steel Tubes – Part 10: Radiographic Testing of of the Weld Seam of Automatic Fusion Arc Welded Steel Tubes for the Detection of Imperfections; BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 • PN-EN 10246-15 Non-destructive testing of Steel Tubes – Part 15: Automatic Ultrasonic Testing of Strip/Plate Used in Manufacture of Welded Steel Tubes for the Detection of Laminar Imperfections; • PN-EN 10246-16 Non-destructive testing of Steel Tubes – Part 16: Automatic Ultrasonic Testing of the Area Adjacent to the Weld Seam of Welded Steel Tubes for the Detection of Laminar Imperfections; • PN-EN 10246-17 Non-destructive testing of Steel Tubes – Part 17: Ultrasonic Testing of Tube Ends of Seamless and Welded Steel Tubes for the Detection of Laminar Imperfections; • PN-EN ISO 10893-6 Non-destructive testing of Steel Tubes – Part 6: Radiographic Testing of the Weld Seam of Welded Steel Tubes for the Detection of Imperfections; • PN-EN ISO 10893-8 Non-destructive testing of Steel Tubes – Part 8: Automatic Ultrasonic Testing of Seamless and Welded Steel Tubes for the Detection of Laminar Imperfections; • PN-EN ISO 10893-9 Non-destructive testing of Steel Tubes – Part 9: Automated Ultrasonic Testing for the Detection of Laminar Imperfections in Strip/Plate Used in the Manufacture of Welded Steel Tubes; Table 1. Non-destructive tests used in manufacture of pipes [2] Types of imperfections detected during pipe manufacture Residual magnetism on pipe ends Laminar imperfection on pipe ends Longitudinal/transverse imperfections in welded joint Laminar imperfections on pipe body Laminar imperfections on strip edge where transverse weld was made Non-destructive testing of weld in pipe ends (areas not tested previously) / testing of areas being repaired No. 3/2013 Testing status Obligatory Optional Obligatory Optional Optional Obligatory Types of tests, requirements and acceptance levels Gaussmeter using Hall effect or equivalent; max. 30 gausses, random tests Ultrasonic testing according to PN-EN 10246-17 or PN-EN ISO 10893-8, inspection area: max. 6 mm from pipe ends on circumference Ultrasonic testing according to EN 10246-9 or PN-EN ISO 10893-11, acceptance level U2/U2H or calibration method ”two lamb” (also for welds joining strip ends in helically welded pipes) Radiographic testing according to PN-EN 10246-10 or PN-EN ISO 10893-6, image quality class R1, used for T-type joints of helically welded pipes Ultrasonic testing according to PN-EN 10246-15 or PN-EN ISO 10893-9, acceptance level U2 Ultrasonic testing according to PN-EN 10246-15 or PN-EN ISO 10893-9 or alternatively ultrasonic testing according to PN-EN 10246-16 or PN-EN ISO 10893-8, acceptance level U2 Ultrasonic testing according to PN-EN 10246-9 or PN-EN ISO 10893-11 for longitudinal imperfections, acceptance level U2/U2H or (if not specified otherwise) Radiographic testing according to PN-EN 10246-10 or PN-EN ISO 10893-6, image quality class R1for longitudinal imperfections and Ultrasonic testing according to PN-EN 10246-9, (PN-EN ISO 10893-11) or radiographic testing according to PN-EN 10246-10 (PN-EN ISO 10893-6) for transverse imperfections BIULETYN INSTYTUTU SPAWALNICTWA 39 • PN-EN ISO 10893-11 Non-destructive testing of Steel Tubes – Part 11: Automated Ultrasonic Testing of the Weld Seam of Welded Steel Tubes for the Detection of Longitudinal and/or Transverse Imperfections. The standards of the PN-EN 10246 series have been now replaced by the standards of the PN-EN ISO 10893 series. The basic product standard PN-EN 10208-2 (version in effect) still indicates older standards as those in effect. As a result, all the greater production plants continue to use these standards in the manufacturing of pipes. For this reason, in all cases this study refers to both standards as recommended for use. As far as ultrasonic manual testing is concerned, the standards applied are the basic standards concerned with the manner of the ultrasonic manual testing of welded joints and their assessment and are not discussed in this study. A detailed list of non-destructive tests applied in pipe manufacture is presented in Table 1. The residual magnetism at the ends of each pipe, parallel in relation to the pipe axis, should not exceed 30 G (3mT). The measurement of magnetism should be carried out on a random basis on the butting face of the pipe end, by means of a calibrated meter (gaussmeter) using the Hall effect or by means of equivalent devices. All non-destructive tests of welded joints should follow a hydraulic test of the pipe. Visual testing (VT) In accordance with the requirements of standard PN-EN 10208-2 each pipe should undergo a visual inspection of the entire outer surface. The inner surface of the pipe should be inspected on each side of the pipe if its outer diameter is 610 mm or less and in 100% of the entire inner surface if the outer diameter of the pipe is 610 mm or greater. Visual testing should be carried out by adequately trained personnel and with proper 40 lighting of the surface being viewed (the light intensity should amount to a minimum of 300 lx). Pipes should be free from imperfections as a finished product. The appearance of the outer and inner surfaces of the pipe should be typical for the production process and heat treatment applied. No surface imperfections requiring removal should be visible. Below are presented manners of dealing with imperfections detected on the outer surface: • imperfections with a depth equal to or below 12.5% of a required wall thickness, which do not reduce the nominal thickness of the pipe in the place of their occurrence should be treated as allowed imperfections (they can remain on the pipe surface or, in accordance with the manufacturer’s decision, can be removed by cosmetic grinding); • imperfections with a depth exceeding 12.5% of a required wall thickness, which do not reduce the nominal thickness of the pipe in the place of their occurrence are classified as imperfections and should be removed by grinding (after grinding it is necessary to apply a proper NDT method, e.g. MT, in order to check the pipe surface after grinding); • imperfections which reduce the nominal thickness of the pipe in the place of their occurrence should be treated as unallowed imperfections and repaired by welding or the part of the pipe containing such imperfections should be cut out (the required length of the time should be maintained). If the aforesaid solution is impossible, a pipe with an unallowed imperfection should be rejected. Geometrical deformations of the cylindrical contour of the pipe should not exceed the following: • 3 mm (flattening, metal excess, coldformed indentations with gentle edges); • 6 mm (other indentations). BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 Outer imperfections can be removed only by machining or grinding. The thickness of a pipe wall in the areas being repaired cannot be less than the required nominal wall thickness. All machined/ground areas should gently pass into the contour of the pipe [2]. Ultrasonic testing (UT) Ultrasonic testing of strip (pipe body) for laminar imperfections The first type of ultrasonic testing used during pipe checks is testing for laminar imperfections of the strip surface. In this case it is possible to accept ultrasonic testing carried out at the sheet manufacturer’s or test a (flat-formed) material used in the production of pipes, acting in accordance with the requirements of standards PN-EN 10246-15 or PN-EN ISO 10893-9 and acceptance level U2. Single laminar imperfections or laminar imperfections in groups exceeding acceptance level U2 are unallowed [2, 6, 11]. An example of a device for strip ultrasonic testing is presented in Figure 2. Fig. 2. Device applied in UT of strips for the presence of laminar imperfections, used in the Salzgitter Mannesmann Grossrohr GmbH compay, Germany [5] No. 3/2013 Ultrasonic testing for the detection of longitudinal and transverse imperfections in welded joints In order to ensure full NDT of welded joints it is required that helically welded pipes should undergo ultrasonic testing along their whole length. The test objective includes both longitudinal and transverse imperfections. Ultrasonic testing should be carried out in accordance with the requirements of standard PN-EN 10246-9 or PN-EN ISO 10893-11 [2]. As a rule, UT of pipes follows the completion of all the main operations of a production process. In order to ensure the reliability of ultrasonic testing, pipes should be sufficiently straight and free from foreign matter. In order to detect longitudinal and transverse imperfections, helical welds should undergo ultrasonic testing. In both cases the testing should be conducted in two opposite beam propagation directions. During the test the converter unit should remain in the proper position in relation to a weld so that the entire weld could be searched through. A searching rate should not oscillate by more than 10% than the adopted basic rate. Depending on the thickness and roughness of the pipe surface, the frequency used in UT should be contained within a 1 MHz÷15 MHz range. The maximum width of each single converter, measured parallel to the main axis of the pipe, should amount to 25 mm. By means of an ultrasonic monitor connected with a marking and/or sorting system, the equipment should enable the classification of pipes as accepted or rejected. If it is not possible to test welds on pipe ends by means of automatic ultrasonic testing equipment, the testing of pipe ends should be conducted by the manufacturer using either manual ultrasonic testing or radiographic testing. Ultrasonic equipment for detecting longitudinal imperfections should be calibrated by means of four longitudinal grooves, BIULETYN INSTYTUTU SPAWALNICTWA 41 A reference groove should be of an two on the outer surface and two in the inner surface of pipe masters and/or by means of ”N-type” (Fig. 4). The sides of the groove a datum hole (Fig. 3). Converters used for should be parallel to each other and the botdetecting transverse imperfections should tom should be perpendicular to the sides. be calibrated by means of a hole and/or two grooves transverse in relation to the weld, one on the outside and one inside a sample being tested. The decision whether to choose grooves or a hole is at the manufacturer’s discretion. Fig. 4. “N-groove” [3, 9]. w – width, d - depth Fig. 3. Arrangement of reference grooves and datum hole [3, 9] 1. longitudinal outer grooves 2. submerged arc made weld 3. pipe master or pipe section 4. right-through hole 5. longitudinal inner grooves A test sample should have the same nominal diameter, wall thickness, surface roughness, heat treatment state and similar acoustic properties (e.g. wave propagation velocity and wave damping coefficient) as the pipe being tested. The manufacturer should have the possibility of removing the inner and outer weld run in accordance with the pipe curvature. The outer and inner grooves as well as the datum hole should be sufficiently distant from the pipe ends and from each other so that clearly separate indications of signals coming from the grooves and the hole can be obtained. 42 Longitudinal reference grooves should be located in the parent metal near the edge of the weld and in parallel to the weld run (Fig. 3). A reference groove should be electrodischarge machined or made with another method. The dimensions of reference grooves should be the following: a)width, w, (Fig. 4) of a reference groove should not exceed 1.5 mm; b)depth, d, (Fig. 4) should be as specified in Table 2, subject to the following restrictions: ◦◦ minimum groove depth: 0.3 mm for pipes of categories U2 and U3 and 0.5 mm for pipes of category U4; ◦◦ maximum groove depth: 2.0 mm for pipes of categories U2 and U3 and 3 mm for pipes of category U4; c)depth tolerance for a reference groove should amount to ± 15% depth or ± 0.05 mm, whichever value is greater; d)length of reference grooves should be equal to at least double width of each single converter, but not more than 50 mm. A datum hole should be drilled through the wall of a test sample, in the weld centre, perpendicularly to the sample surface (Fig. 3). BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 The diameter of the drill used to make the datum hole should be selected in accordance with Table 3. The diameter of a datum hole should be checked and cannot exceed the nominal diameter of a drill by more than 0.2 mm. testing sensitivity by 3 dB (to compensate equipment indications), it is necessary to carry out equipment calibration again as well as to check all the pipes tested since the previous calibration verification. Table 2. Designation of acceptance level and related depth of reference groove [3, 9] Table 3. Designation of acceptance level and related drill diameter [3, 9] Acceptance level U2 U3 U4 U5 Groove depth in % of nominal wall thickness 5.0 10.0 12.5 15.0 The equipment should be calibrated so that it would ensure (e.g. in three subsequent passes of the test sample through the equipment) obtaining clearly distinguishable signals of the reference standard. The peak value of the amplitude of these signals should be used for adjusting the level of the gate/ alarm monitor in the equipment. The rate of the test sample travel in relation to the set of ultrasonic converters during the verification of calibration should be the same as during the tests of a product. Calibration can be verified by means of a semi-dynamic method. During the testing of pipes with the same nominal diameter, wall thickness, and made of the same steel grade, it is necessary to verify the calibration of the equipment at regular intervals. The equipment calibration is verified by putting the pipe master through the testing equipment. The verification of calibration should be carried out at least once every four hours as well as each time the equipment operator changes. Calibration should also be verified at the beginning and end of production. In addition, the equipment should be calibrated if any of the parameters set during the initial calibration have been altered. If while verifying calibration during production, calibration requirements are not satisfied even after increasing No. 3/2013 Acceptance level Drill diameter U2H 1.6 U3H 3.2 U4H 4.0 Following testing, a pipe revealing signals below the monitor/gate sensitivity threshold should be classified as meeting the requirements. And a pipe having signals equal to or greater than the monitor/gate sensitivity threshold should be classified as questioned or, at the manufacturer’s request, should undergo another test. If during the subsequent test of the same pipe no signal equal to or greater than the monitor/gate sensitivity threshold was obtained, the pipe should be regarded as meeting the requirements. Depending on the requirements of a product standard, pipes questioned during testing should be dealt with in accordance with one or several following manners: a)Subject to an agreement between the purchaser and the manufacturer, the area which has been questioned should be tested by means of another NDT technique or method (usually radiographic), on the basis of agreed acceptance criteria; b)The questioned area should be cut out. The manufacturer should guarantee that the whole questioned area has been removed; c)The pipe should be classified as not satisfying the requirements [3, 9]. Testing of helically welded pipes is carried out for acceptance level U2/U2H, taking into account the following inspection criteria: BIULETYN INSTYTUTU SPAWALNICTWA 43 a)the maximum groove depth should amount to 2.0 mm; b)Using inner and outer longitudinal grooves in the middle of the weld run for the purpose of equipment calibration is unallowed; c)In the case of acceptance level U2, for the purpose of detecting transverse imperfections, as an alternative to the datum hole one may use the inner and outer grooves positioned at a right angle and centred in the weld. In such a situation both inner and outer weld reinforcements should be ground evenly with the pipe contour, in the direct vicinity and on both sides of reference grooves. The grooves should be sufficiently distant from each other in a longitudinal direction. They should also be sufficiently distant from any other reinforcements. Such an approach is necessary to obtain a clear and distinguishable ultrasonic signal response. In order to determine the level of the gate/alarm of the measuring equipment it is necessary to use the complete amplitude of a signal from each of these grooves. d)For acceptance level U2 and by prior agreement, as an alternative to using reference grooves for calibration purposes, it is possible to use inner and outer grooves of a constant depth and increase testing sensitivity using electronic methods (by increasing dB amplification). In this case, also referred to as the „two lamb method”, the depth of grooves should be twice as big as the length of an ultrasonic wave. A required increase in testing sensitivity should depend on a pipe wall thickness. The manufacturer should adequately assure the purchaser that obtained testing sensitivity is basically equivalent to that obtained while using grooves for acceptance level U2. In the case of helically welded pipes the entire length of the weld joining strip edges should also undergo ultrasonic testing. Testing sen44 sitivity and parameters should be the same as during the initial testing of the pipe helical weld. In addition, cruciform joints, in which the weld ends of strip edges meet the helical weld should undergo radiographic testing [2]. Following the tests, the manufacturer should provide the ordering party with a report containing at least the following information: a)reference to standards PN-EN 10246-9 or PN-EN ISO 10893-11 and PN-EN 10208-2; b)date of test report; c)acceptance level; d)declaration of conformity; e)product specification, i.e. steel grade and dimensions; f) type and details of testing technique; g)description of a master/standard [3, 9]. An example of a device for UT of helically welded pipes is presented in Figure 5. Fig. 5. Automatic multichannel device for ultrasonic testing of helically welded pipe weld and strip-joining weld used in the Salzgitter Mannesmann Grossrohr GmbH company, Germany[5] Laminar imperfections in strip area adjacent to welded joint In such a case it is possible to accept ultrasonic testing carried out at steel sheet manufacturer’s or carry out post-weld testing at the pipe manufacturer’s, following the requirements of standard PN-EN 10246-16 or PN-EN ISO 10893-8 and acceptance level BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 U2. The tests are conducted within a 15mmwide zone along both longitudinal edges of a helical weld. In the case of transverse welds joining the edge of a strip testing is conducted along areas adjacent to this butt weld. Requirements concerned with ultrasonic testing of laminar imperfections in the area adjacent to the weld are dealt with in standard PNEN 10246-15 (PN-EN ISO 10893-9) or PNEN 10246-16 (PN-EN ISO 10893-8). Single laminar imperfections or imperfections in groups exceeding acceptance level U2 are not allowed [2]. Non-destructive testing of helically welded joints on pipe ends and in areas being repaired Sections of welded joints on pipe ends which cannot be tested by means of automatic ultrasonic equipment and repaired areas of welds should undergo the following tests: a)in order to detect longitudinal imperfections – manual or semiautomatic ultrasonic testing using the same testing parameters and sensitivity as in the case of the automatic testing of the whole helical weld. Radiographic testing is also possible. b)in order to detect transverse imperfections – manual or semiautomatic ultrasonic testing using the same testing parameters and sensitivity as in the case of the automatic testing of the whole helical weld. Radiographic testing is also possible. c)It is also necessary to carry out testing for the laminar imperfections of the pipe body areas on pipe ends which cannot be checked by means of automatic ultrasonic equipment. In such a case it is also possible to carry out manual ultrasonic testing using the same testing parameters and sensitivity as those used while testing areas adjacent to the weld. Laminar imperfections with a length ≥ 6 mm occurring in girth direction are unallowed if they are present 25 mm away from each pipe end; No. 3/2013 d)While carrying out manual ultrasonic testing a scanning rate should not exceed 150 mm/s [2]. Radiographic and radioscopic testing (RT) Another testing method used in the manufacture of helically welded pipes is radiographic testing, usually carried out on pipes after the completion of all principal production operations. In order to ensure an appropriate testing class, pipes intended for tests should be sufficiently straight and free from foreign matter impurities. The surfaces of welds and adjacent parent metal should be free from foreign matter and surface imperfections as they could impede the interpretation of radiograms. Grinding surfaces is allowed if it can ensure obtaining a surface condition acceptable for tests. If a weld reinforcement must be removed, it is advisable to place markers (usually lead arrows) in order to identify the weld location in a radiogram. In order to ensure the unambiguous identification of a given weld section, each section should be provided with identification symbols (usually lead letters) so that their images can be visible in a radiogram. It is necessary to apply permanent marking on the pipe surface from the side of the radiation source. This is done in order to provide reference points for the accurate assignment of each radiogram location. If it is impossible to strike markers due to the type of a product and/or its intended operating conditions, it is necessary to provide other appropriate means of radiogram assignment, e.g. by marking with paint or preparing accurate sketches. In order to ensure that each part of the weld undergoes testing, during X-raying a longer weld section with several films it is necessary that neighbouring films should overlap over a minimum length of 10 mm. Longitudinal or helical welds of pipes should undergo radiographic testing with BIULETYN INSTYTUTU SPAWALNICTWA 45 x-rays. The use of radioscopic methods is allowed (Fig. 6), yet only in cases when the manufacturer can demonstrate their appropriate sensitivity. Fig. 6. Example of device for radioscopic testing of welded joints There are two image quality classes, i.e. R1 and R2: • class R1 (class B): radiographic testing technique by means of X-radiation of heightened sensitivity; • class R2 (class A): radiographic testing technique by means of X-radiation of normal sensitivity. For image quality class R1 it is necessary to use at least fine-grained films (at least C4 class), whereas for image quality class R2 one should use at least medium-grained films (at least C5 class). For both image quality classes (R1 and R2), the thickness of front intensifying screens should be between 0.02 mm and 0.25 mm. Rear intensifying screens can have other thicknesses. Fluorescent intensifying screens should not be used. The amount of backscattered and internal X-radiation should be limited to a minimum. When in doubt as to the efficiency of protection against backscattered radiation, it is advisable to fix a characteristic sign (usually a 1.5mm-thick letter B) behind the film holder or film frame and make a radiogram in a nor46 mal manner. The density of the symbol image in the radiogram lower than the background density indicates that the protection against backscattered radiation is insufficient and that additional precautions should be used. A radiation beam should be directed onto the centre of a weld section being tested and should be perpendicular to the pipe surface at this point. A length subjected to an assessment should be so that the thickness X-rayed at the ends of the usable length of a radiogram does not exceed the thickness X-rayed in the radiogram centre by more than 10% for image quality class R1 and by more than 20% for image quality class R2. It is necessary to use a technique of X-raying through one wall. If, due to the size of the object, such a technique cannot be used, it is allowed to use a technique of X-raying through two walls (by prior agreement). The distance between the film and the weld surface should be as small as possible. The minimum distance between the source and the sample ,f, should be selected using appropriate formulas or diagrams. The conditions of exposition should be so that the density of a radiogram in the tested area of weld material free from imperfections is not lower than 2.0 for image quality class R1 (2.3 for image quality class B) and not lower than 1.7 for image quality class R2 (2.0 for image quality class B). The quality of an image should be determined using an image quality indicator (IQI) made of mild steel, grade specified in standards PN-EN 462-1 (PN-ISO 19232-1) and PNEN 462-2 (PN-ISO 19232-2). An IQI should be placed on the surface from the radiation source side or next to the weld or, in the case of a wire-type IQI – across the weld. Image quality indicators should be placed from the film side only if the surface from the source side is inaccessible. In such cases it is necessary to place a letter “F” next to an IQI and make a note of this change of the procedure in the test report [4, 10]. BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 diameter T/3 (T – wall thickness), whichever is smaller; c)The sum of the diameters of all acceptable single imperfections in each section of a weld with the following dimensions: a length of 150 mm or 12 T, whichever is smaller, should not exceed the smaller of the following values: 6.0 mm or 0.5 T, where the distance between single inclusions should be smaller than 4 T; d)Single longitudinal slag inclusions with a length of 12.0 mm or 1 T, whichever is smaller or with a width not exceeding 1.6 mm are acceptable. The total length of all acceptable single imperfections in each section of a weld with the following dimensions: a length of 150 mm or 12 T, whichever is smaller, should not exceed 12.0 mm, where the distance between single inclusions should be smaller than 4 T; e)Single undercuts of any length and with a maximum depth of 0.4 mm are acceptable; f) In addition, single undercuts with a maximum length of T/2 and a maximum depth of 0.8 mm which does not exceed 10% of a specified wall thickness are allowed provided that there are not more than two such undercuts on a weld section with a length of 300 mm and that all such undercuts will be levelled; g)All undercuts exceeding the above values should be repaired; in other cases the area under suspicion should be cut out or the pipe should be rejected; h) All undercuts, of any length and depth, on the inTable 4. Requirements for radiographic testing sensitivity and image quality class R1 according to PN-EN 10246-10 and PN-EN ISO 10893-6 [2] ner and outer side of the weld, overlapping in the lonWall thickness Required visibility gitudinal direction are unacOver [mm] Up to [mm] Hole diameter [mm] Wire diameter [mm] ceptable [2]. 10 0.40 0.16 4.5 Pipes not exceeding allowed 16 0.50 0.20 10.0 values/indications should be 25 0.63 0.25 16.0 classified as passing the test. 32 0.80 0.32 25.0 Pipes exceeding allowed val40 1.00 0.40 32.0 ues/indications should be clas- The radiographic testing of helically welded pipes is carried out following the requirements of standard PN-EN 10246-10 or PN-EN ISO 10893-6, image quality class R1, and in accordance with the following conditions: a)The conformity with sensitivity-related requirements presented in Table 4, determined on the basis of a base metal should be verified by means of a wire-type image quality indicator, in accordance with standard PN-EN 462-1 (PN-ISO 19232-1) or by means of an equivalent hole-type IQI according to PN-EN 462-2 (PN-ISO 19232-2); b)Radiographic testing should be carried out using only X-radiation, fine-grained and high contrast films as well as lead screens. It is possible to use fluoroscopic methods, yet only when the manufacturer can demonstrate their equivalence with a method based on X-ray films; c)The optical density of an X-ray photograph cannot be lower than 2.0 and should be selected so that the density in the thickest area of the weld is not lower than 1.5 and that the maximum contrast for a given film type is maintained. Requirements related to the assessment of welded joints are the following: a)Cracks, no joint penetration and incomplete fusion are unacceptable; b)Single slag inclusions and blowholes are acceptable if characterised by the following dimensions: size up to 3.0 mm or a No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 47 sified as questioned. Questioned pipes should be dealt with in one of the following manners: a)A questioned area should be levelled by means of an appropriate method. In order to ensure that the imperfection has been removed entirely, after checking that the remaining thickness is still within a tolerance range, the area should undergo another check using magnetic particle testing or penetration testing. Afterwards, the pipe should be classified as passing the test. If the recommended removal of the faulty area has decreased the thickness below the acceptable value, the questioned area should be repaired by welding carried out in accordance with an approved welding procedure specification. After that, the repaired area should undergo radiographic testing following the requirements of PNEN 10246-10 or PN-EN ISO 10893-6; b)A questioned area should be cut out. The manufacturer should ensure that the whole questioned section has been cut out; c)A pipe should be classified as failing to meet the requirements. In required cases the manufacturer should write a test report containing at least the following information: a)reference to PN-EN 10246-10 or PN-EN ISO 10893-6 and PN-EN 10208-2; b)test report date; c)declaration of conformity; d)designation of a product providing grade and dimensions; e)type of and detailed information about the testing technique used; f) each deviation from specified procedures, agreed or not agreed; g)image quality class; h)operator’s name, signature and certificate number [4, 10]. velopment owing to new gas pipelines under construction both in Poland and abroad. The production of these pipes requires the use of materials characterised by the highest mechanical properties in a given material group. NDT devices are usually provided with modern and recently developed technical solutions. Investment-related expectations for the years to come make the manufacture of welded pipes the area of interest for investors and manufacturers all over the world. References 1. Michałowski, W., and Trzop, S., et al. (2006). Rurociągi dalekiego zasięgu. Warsaw: Wydawnictwo Fundacja Odysseum. 2. PN-EN 10208-2:2011 Steel pipes for pipelines for combustible fluids - Technical delivery conditions - Part 2: Pipes of requirement class B 3. PN-EN 10246-9:2004 Non-destructive testing of steel tubes - Part 9: Automatic ultrasonic testing of the weld seam of submerged arc welded steel tubes for the detection of longitudinal and/or transverse imperfections 4. PN-EN 10246-10:2004 Non-destructive testing of steel tubes - Part 10: Radiographic testing of the weld seam of automatic fusion arc welded steel tubes for the detection of imperfections 5. Presentation of Salzgitter Mannesmann Grossrohr GmBH - Salzgitter 2012. 6. PN-EN 10246-15:2002 Non-destructive testing of steel tubes - Part 15: Automatic ultrasonic testing of strip/plate used in the manufacture of welded steel tubes for the detection of laminar imperfections 7. PN-EN 10246-16:2002 Non-destructive testing of steel tubes - Part 16: Automatic ultrasonic testing of the area adjacent to the weld seam of welded steel tubes for Conclusion the detection of laminar imperfections In recent years the manufacture of heli- 8. PN-EN 10246-17:2002 Non-destructive testing of steel tubes - Part 17: Ultrasoncally welded pipes has seen an intensive de48 BIULETYN INSTYTUTU SPAWALNICTWA No. 3/2013 ic testing of tube ends of seamless and 11.PN-EN ISO 10893-9:2011 Non-destructive testing of steel tubes - Part 9: Autowelded steel tubes for the detection of mated ultrasonic testing for the detection laminar imperfections of laminar imperfections in strip/plate 9. PN-EN ISO 10893-6:2011 Non-destrucused for the manufacture of welded steel tive testing of steel tubes - Part 6: Ratubes diographic testing of the weld seam of welded steel tubes for the detection of 12.PN-EN ISO 10893-11:2011 Non-destructive testing of steel tubes - Part 11: Autoimperfections mated ultrasonic testing of the weld seam 10.PN-EN ISO 10893-8:2011 Non-destrucof welded steel tubes for the detection of tive testing of steel tubes - Part 8: Autolongitudinal and/or transverse imperfecmated ultrasonic testing of seamless and tions welded steel tubes for the detection of laminar imperfections No. 3/2013 BIULETYN INSTYTUTU SPAWALNICTWA 49