Module 7 Residual Stress ance) Distortion www.cwbgroup.org © Copyright 2010 CWB Group - Industry ServicesMODULE 7 Residual Stress and Distortion 8260 Parkhill Drive, Milton, ON LOT 5V7 Tel: 1-800-844-6790 Fax: 905-542-1318 www.cwbgroup.org © Copyright 2010 The CWB Group Industry Services Revised September 2013 | All rights reserved ‘The works of authorship contained in this document, including but not limited to all design, text and images, are the property, except as otherwise expressly stated, of the CWB Group-Industry Services or one of its afiiates, the CWB (Canadian Welding Bureau) and QUASAR. This document, whether in current form or electronic, may not be copied, reproduced, transmitted, displayed, performed, distributed, rented, sublicensed, altered, stored for subsequent use, or otherwise used in whole or in part in any manner without the prior writen consent of the CWB Group industry Services, except tothe extent that such use ‘constitutes “fair use" under the Copyright Act of 1978 as amended, and except for a temporary copy ina single computer's ‘memory and one unaltered permanent copy to be used by the viewer or student for personal and non-commercial use only.Table of Contents Residual Stress and Distortion 1. Introduction 41. Terms and Definitions Expansion and Contraction of Metals Coefficient of Thermal Expansion and Thermal Stress 3.1 Compression (Restricted Expansion) 3.2 Summary Residual Stress 441 Residual Stresses Induced by Thermal Processes 4.2. Residual Stress Induced by Arc Welding 4.3, Residual Stress in Stee! Plates 4.4 Residual Stress of Mill Plates with As-Rolled Edges 4.5. Residual Stress Induced by Flame Cutting 4.6. Residual Stress in Welded Wide Flange Shapes 47 Residual Stress in Mill Rolled Shapes 4.8 Measurement of Residual Stresses 49 Summary 4.10 Estimation of Shrinkages Distortion 5.1. Distortion Caused by Oxy-fuel Gas Cutting 5.2. Distortion Caused by Welding 5.3. Transverse Contraction (Shrinkage)—Angular Distortion 5.4 Longitudinal Expansion and Contraction (Shrinkage) 5.5. Other Causes of Welding Distortion ‘Welding Procedure and Distortion 6.1. Welding Procedures Control and Correction of Distortions 7A Control of Distortion 7.2 Correction of Distortion Flame Straightening and Flame Forming 8.1 Introduction 8.2 Basic Principles 8.3 Maximum Temperature—Low Alloy Steels 8.4 Removing a Bulge 8.5 Line Heats 8.6 Strip Heats 8.7 Vee Heats 8.8 Beam Forming or Beam Cambering 8.9. Channel Forming 8.10 Angle Forming 28 28 38 38 a 42 42 43 45 45 48 47 488.11 Cambering Beams 49 8.12 Examples of Heat Straightening 52 8.13 Summary of Flame Forming 56 Guides and Exam Exercise 87 ‘© Copyright CWE Group - Industy ServicesMODULE 7 | Residual Sttess and Distortion @ 4. introduction ‘The objective of this module is to discuss: + the causes, behaviour and magnitude of residual stress and distortion + how to avoid or minimize residual stress and distortion + how to rectify welding distortion Residual stress is neither visible nor readily measurable, but distortion is both visible and measurable. The latter is always the manifestation of the stresses. The relationship between residual stress and distortion will bbe discussed in this module. In dealing with this subject, the adherence to established welding procedures and manufacturing plans is very important. Years of shop fabrication experience is still the best assurance to achieve the desired results. Knowledge of the fundamental theory and equations will help one grasp the nature of the problem, but precise control is not always achievable, Because there are many variables involved, experience and theory are the best tools to minimize distortion, Its often impossible to establish an exact, satisfactory welding procedure for an unsatisfactory design. The following lessons are of equal importance to both the designer and fabricator. The designer's work is not finished on the drawing board or at the computer, since the designer and fabricator must share responsibilty for the final product. ‘The development of welding procedures should begin with the understanding that the heat of welding will produce expansion, contraction and stress. Therefore the welding procedure's major objective should be to: 41) Produce sound weldments 2) Maintain dimensions by controlling distortion 3) Reduce and balance internal residual welding stresses 4) Allow access for welding and be economical ‘Awelding procedure will involve a welding process, base metals to be joined, joint design, preparation, filer ‘metals, power source, current, voltage, speed, welding technique and possibly heat treatment. However, of equal importance is the patter of heat input to the work, as determined by the sequence of assembly and progression of welding. To control distortion and residual stresses, the effect of each of these factors must be thoroughly understood and the welding procedure should be planned accordingly. The welding procedure, once planned, should be checked by a trial and modified when required. it should be clearly laid out and purposely followed by all workers. Itforms an important part of the shop's quality system. (© Copyright CW Group - Industry Services Page 1© 4A Terms and Definitions Angular Distortion Compression Contraction Distortion Ductitity Expansion Intermittent Welding Lateral Shrinkage Longitudinally Neutral Axis Peening Periphery Residual Stresses Permanent Deformation Restrain Restraint Shrinkage Force Strain MODULE 7 | Residual Stress and Distortion ‘To reduce in volume, or the state of being compressed. To force together or into less space. To become reduced in size. To pull together, shrink or wrinkle. The act of distorting. A physical change in shape by bending, twisting or warping Capability of being easily molded or shaped. Enlargement or increase. Increase in the dimensions of a body. ‘Welding that is not continuous over the length of the joint. ‘The process of being drawn together parallel to the line of the joint, Running lengthwise or pertaining to length. ‘Occupying a middle position of symmetrical shape. Through the centre of gravity. The act of inducing compression into the surface layers of an object. Often performed by striking the surfaces of a metal object with a ball shaped device, ie., ball-peen hammer, or blasting with specific procedures, i.e., shot peening. ‘The outermost part of an object or area. Stresses remaining after a process has been completed, Stress present in a material that is free of external forces or ‘thermal gradients, ‘Shape change that is not reversible by the removal of the forces (plastic deformation) that caused it. ‘movement. To prevent or The act of holding back. {An internal force exerted on a component due to uneven dimensional change during heating and cooling. ‘A shape change produced by stress. Page 2 (© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® Stress ‘An intemal force or system of forces that tends to strain or deform an object or body. Force per unit area. Symmetrical Exactly uniform in shape about a dividing line or about a centre or axis. ‘Transverse Situated or lying across. Tensile Stress stress that tends to cause elongation or extension. Upset ‘To make shorter or thicker. 2. Expansion and Contraction of Metals By nature, metals expand volumetrically when heat is applied. When the heat source is removed, the metal contracts in volume during cooling. The expansion and contraction of heated metal can be illustrated by considering the movement of a metal ball, Figure 2.1 shows a metal ball that expands freely during heating and contracts freely to its original volume and shape after cooling. Asteel ball is heated and allowed to expand freely Before Expanded by Heating A B c Figure 2.4: Free expansion and contraction. ‘The fundamental cause of welding distortion is local shape change next to the joint. The welding process heats a localized area. This area is restricted from freely expanding in all directions by the colder, stronger ‘material surrounding the welding point. Figure 2.2(A) shows a metal ball fit snugly between two rigid stops, which is then heated and cooled. Figure 2.2(B) shows that the ball is expanding in the open directions and restricted in the other direction by the stops. During this process the metal grains have undergone adjustment under force to expand to open space (restricted expansion). When the metal ball cools it contracts again, volumetrically. Consequently, it leaves ‘gaps between the stops (free contraction), as shown in Figure 2.2(C). A similar effect occurs during arc welding. (© Copyright CWB Group - Industry Services Page 3@® MODULE | Rei Sas nd Dorion A steel ball is heated between two barriers which cannot move e @ & Expanded by Heating After are B c ‘The grains (and atoms) in the material have rearranged themselves. Figure 2.2: Restricted expansion and free contraction. We have illustrated free expansion and free contraction, as well as restricted expansion and free contraction. The third condition—restricted expansion and restricted contraction—is shown in Figure 2.3. The ends of a round metal bar are rigidly gripped between two solid stops. Heat applied at any point on the bar causes it to expand, but tis not allowed to expand or contract lengthwise. Therefore, during heating, all the expansion takes place in the diameter of the heated portion because this part has to absorb all the ‘volume of metal: it cannot expand lengthwise. The upset (ie., the swelling of the heated part) is known as permanent deformation—it wil not disappear after the bar cools. ‘The bar is in compression during heating and in tension after cooling. During heating, the metal is softened and forced to locally increase in diameter. During cooling, the bar stretches. Tensile stress will be set up in the bar, and if not sufficiently elastic, the bar may break as shown in Figure 2.3(8). ‘Heated portion L , =" Eo tn nee a, Figure 2.3: Behaviour of metal bar when heated and coofed. (A) Restricted expansion, (B) Contraction prevented. Page 4 (© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® }.. Coefficient of Thermal Expansion and Thermal Stress From foregoing discussions we know that metals expand when the temperature is raised, Furthermore, Under the same temperature rise the amount of expansion differs for various metals. Table 3.1 shows the coefficient of some common metals. The unit ofthe coefficient is micrometer (10 meters) per meter per degree (°K or °C). The coefficient is not a constant, as can be seen In Figure 3.4, Table 3.1: The coofticiont of thermal expansion of some common metals. Metal Goafhcsentot expansion pian” Mild Stee! 18) ‘Stainless Stool: ‘Austenitic 145 Martensiic as Nicke! 133 Copper 185 Aluminum 234 Magnesium 27a Lead 293 Zine 30.7 ¥ 16 £ E15 a = a” 2 Average value from 20°C. B to indicated temperature. ‘Typical data for mild steel. B12 z 3 i u 10 100 200 300 400 500 600 700 800 ‘Temperature °C Figure 3.1: Typical values of the coefficient of thermal expansion for mild steels as @ function of temporature, ‘Thermal stress is the stress created by restricted thermal expansion or contraction. In Figure 2.3(A) for example, assume that the entire bar is heated uniformly. The expansion per unit length (ie., thermal strain) can be calculated: ‘Thermal strain e =a * AT Total expansion AL = Lx ¢=Lx ax AT {© Copyright CWB Group - Industry Series Page 5© MODULE 7 {Reso ese and Denon Where: = the coefficient of expansion, and AT= the change in temperature. When expansion is prevented, that means the metal bar is shortened by the same strain. In other words, the ‘metal must be under compression. From the stress and strain relation: Stress o (MPa) = (Ea * AT) Where: E= Young's modulus of elasti steel, E = 200,000 MPa a for steel = 11.8 x 10° 3.4 Compression (Restricted Expansion) ‘At room temperature (20°C) if the yield stress of the steel bar is 350 MPa (50 ksi), the temperature rise (AT) required to reach yielding in compression can be calculated: T= @ = ___350__ = 148°C (above room temperature). Ea 200000 x 118% 10° ‘When local temperature rise (47) is higher than 148°C, upset will occur. When the steel bar is cooled to room temperature, residual tensile stress will remain. Even a small temperature rise can cause distortion under the right condition of restraint. When arc welding, the local temperature rise is much greater than that calculated above. 3.2 Summary From the previous discussions, the following points can be drawn: 1) Expansion and contraction in a given material are proportional to the dimensions of the heated area and temperature change. 2) _ Restricted expansion and contraction will cause deformation and stresses in a metal. 3) _ Restricted contraction may set up internal tensile stress sufficient to cause failure if the metal is not ductile enough to accommodate the stretch. 4) Even @ relatively small local temperature rise can cause distortion. 4. Residual Stress The term residual stress means that some internal stress is created and stays inside the metal after the manufacturing processes are completed. These processes can include: + thermal cutting or heating + welding + mechanical forming + metallurgical changes such as heat treating This module deals mainly with the first two processes. Page 6 {© Copyright CWB Group - Indust ServicesMODULE 7 | Residual Stress and Distortion @ 4A Residual Stresses induced by Thermal Processes ‘The thermal effect on metals has been discussed in the foregoing paragraphs. We also explained how stress may be set up when expansion and/or contraction is restricted. To bring this analogy one step further, consider a large square steel plate spot-heated at the centre, as shown in Figure 4.1. At the heated area the metal becomes upset due to restricted expansion by the surrounding cold ‘metal mass. After cooling, the upset remains and the contraction induces tension around the heated area. This tensile stress remains if nothing else is done to the plate. This is why its called ‘residual stress’, to distinguish it from other stresses created by external loading, Prior to stress is Exaggerated Heat spot: The temperature is relatively the same throughout the thickness of the material but is localized. Figure 4.1: Residual stress causing shape change. it may help the student to understand the practical meaning and effect of residual stresses as internal compression and tension in the metal. For example, tensile forces are developed across a butt joint when the weld metal is unable to contract freely. The residual stresses are static and balanced—that 's, the overall tensile stressed areas are balanced by the compressive stressed areas, and no ‘movement results once the balance is attained. But, in the process of balancing, while the metal cools, | movement may happen. The result is “cistortion’. This is another important subject, which will be discussed later in this module. 4.2 Residual Stress Induced by Arc Welding Next, we shall investigate what happens when welding heat is applied to join two plates together, as shown in Figure 4.2. Two large, thick, rectangular plates of same size are welded together along their long sides, During welding the long edges are under intensive heat (actually melted) and go through thermal expansion, but the areas a short distance from the edges are relatively cool and do ‘not expand the same, or hardly expand because of the very steep thermal gradient. In other words, expansion is restricted by the plates themselves. Following the same reasoning, when the weld is cooling it goes through restricted contraction and sets up high tensile stresses along the length of the weld (see Figure 4.2(A)), a longitudinal residual stress patter. This high tensile stress stays with the plates if nothing else is done to them. This is (© Copyright CWB Group - Industry Services Page7ae ae MODULE 7 | Residual Stress and Distortion how residual stress is induced by welding. The residual stress in the longitudinal direction may be as high as yield stress of the plate (see Figure 4.2(A)). As explained previously, the thermal expansion ‘and concentration are in all directions (volumetric). Therefore, there is residual stress transverse to the weld line, as shown in Figure 4,2(B), a transverse residual stress pattern. Transverse stresses are not high except at the ends, where they are compressive. The most important residual stresses are the high longitudinal stresses along the length of the weld and in the heat-affected zone, x i, (A) Longitudinal residual stress pattern (B) Transverse residual stress pattern Figure 4.2: Typleal residual stress pattern in a weld in a flat plate. Further explanation of Figure 4.2 is illustrated with the aid of Figure 4.3 and Figure 4.4. Figure 4.3 shows that the plate edges along the weld joint undergo expansion during welding and the plates tend to bow outward. As the weld cools, as shown in Figure 4.4(A), the plates tend to bow in opposite directions because the plate edges contract with the weld. Since the weld holds the plates together, the middle of the joint will be under tension perpendicular to the weld line. The end regions are in ‘compression to balance the tension in the middle. This is the transverse residual stress pattern shown in Figure 4.2(8). Figure 4.4(8) ilustratps the formation of longitudinal residual stresses. Imagine the weld metals isolated from the plates, free to contract, and then stretched longer to fit the plate edges. This will result in tensile stress in the weld metal and part of the adjacent plate for a certain width on either side of the weld. This tension region must be balanced by compression regions outside this width on either side of the weld. This is the longitudinal residual stress pattern shown in Figure 4.2(A). Page 8 ‘© Copyright CWB Group - Industry ServicesMODULE | Read ress and Dsrion ® Expansion of plate edges along length of weld. Plates tend to bend applying ‘tension across ends of weld as it is deposited. Figure 4.3: When the weld is deposited, the edges of the plates get hot, expand, and tend to bow the Plates. Yielding occurs along the edges of the plates. ‘As weld cools, plates tend to ‘As weld and plate edges contract, bend back as hot edges contract it develops tension in welds and compression in plate | | n Final residual stress across weld Hot ‘Weld length (hot) Longitudinal (8) Figure 4.4: On cooling, the plates bow in the other direction but are held by the solidifying weld metal. Residual stresses, equivaient to a bending moment applied to the plate ends, result from the attempt to restrain the bow. ‘© Copyright CWB Group - Industry Services| Page 9© MODULE 7 | Reid Sess end Dsrion 4.3 Residual Stress in Steel Plates After the study of heat effect, expansion and contraction of metals, we are ready to investigate the steel plates we use to build pressure vessels, ships, buildings, machinery and automobiles. ‘We all know that steel plates are produced by thermal and mechanical processes and that it inevitably involves heating, rolling and cooling. We also know that heating and cooling, short of free expansion and contraction, creates residual stresses. In reality, steel slabs with large mass and size can hardly expand and contract freely unless through specially controlled cooling processes. In the following paragraphs we will discuss the behaviour of steel plates commonly used by fabricators. 4.4 Residual Stress of Mill Plates with As-Rolled Edges Figure 4.8 shows the pattern of residual stresses in a mill plate with as-rolled edges. It shows compressive residual stress at the edges and tensile residual stress in the middle of the plate. In comparing differences between rolled edges and that of flame-cut edges (see Figure 4.6) a distinctive contrast can be noticed. The flame-cut edges have tensile residual stress at the edges. The as-rolled edges have compressive residual stresses. From the discussion of the effect of heating and cooling, we know that the rolled edges will cool faster than the middle of the plate. Therefore, when the whole plate is cooled to room temperature, the edges are under compression. I Figure 4.6: Residual stress in plate with flame-cut edges. Page 10 (© Copyright CWB Group - Industry ServicesMODULE 7 | Resival Svess and Dstaton @® 4.5 Residua! Stress Induced by Flame Cutting In oxy-fuel gas cutting of steel, the temperature along the cutting surfaces approaches the melting point. The uneven heating and subsequent cooling will induce residual stress. Consider a plate that is cut with two torches simultaneously. The residual stress in the cutting edges is tensile. This is, of course, because of the restraining effect of the relatively cool areas adjacent to the cutting edges. As ‘a result, the adjacent areas are in compression. The distribution of the longitudinal stresses across the width of the plate is shown in Figure 4.8. Where only one side is cut at a time, the part may distort to balance the residual stresses. The edge that is cut will shorten. 4.6 Residual Stress in Welded Wide Flange Shapes From the previous ussions on residual stresses in as-rolled plates and flame-cut plates and on the influence of welding, we should be able to visualize the residual stress patterns in two welded wide-flange shapes. Figure 4.7 and Figure 4.8 show the built-up shapes with as-rolled and flame-cut stress patterns. /N a Figure 4.7: Longitudinal residual stresses in a welded built-up column with as-rolied flange plates (Flange edges in compression). Plate fame cut Plate fame cut before welding after welding a ‘Small compressive Large tensile or tensile stress stress Figure 4.8: Welded wide flanges with flame cut edges. © Copyright CWS Group - Industry Services Page 11® MODULE 7 | Residual Stress and Distortion ‘The residual stress patterns of a welded wide-flange shape with cover plates is shown in Figure 4.5. A large tensile residual stress is induced at the flange tips because of the localized welding heat and related local expansionicontraction. Cover plate Figure 4.9: Residual stress patterns in a welded assembly. Residual stress in a welded box section is shown in Figure 4.10. Applying the same principle, the areas next to the corners cool more slowly and end the cycle in tension. Figure 4.10: Longitudinal residual stresses in welded box column. Page 12 (© Copyright CWB Group - Industry Servicesissue 7 jade ones artnet @ 4.7 Residual Stress in Mill Rolled Shapes ‘After the discussion of residual stresses in plates and welded wide flanges, you would expect these forces to be present in hot-rolled shapes. From the discussion earlier, we recognize that residual stress is created by uneven heating and cooling. For the same reason, residual stresses are induced in hot-rolled |-shapes. As shown in Figure 4.11, the parts that coo! first or faster are the toes of flanges and the centre of the web. Compressive residual stress is formed. The parts that cool last or more slowly are the flange and web Junctions. Contraction is restrained by the parts that cooled first, and tensile residual stress is formed. Therefore, the pattern of residual stress is as shown in Figure 4.14 a Figure 4.11: Residual stress in hot-rolled I-shape. 4.8 Measurement dt Residual! Stresses ‘We have discussed the existence and the patterns of residual stresses in plates, welded wide flange and universal mill-rlled !-beams. This section addresses how residual stresses are measured. The methods used to measure residual stresses and plot their distribution are quite elaborate and time consuming. Only the principles involved in several measuring methods will be described. 4.8.1 Stress-Relaxation Method ‘The early method devised to measure residual stress is the sectioning method. Itis also called “stress-relaxation” method. When a piece of metal is removed from a base metal containing residual stress, the stress in the piece and its surrounding area are relaxed. By measuring the elastic strain change (dimensional change), the residual stress before removal can be calculated. To reveal the stress pattern the entire specimen is cut into strips and the strain changes are measured. © Copyright CW Group - Industry Services Page 13@ MODULE 7 Resa tess and Dison Several other measuring methods are based on the stress-relaxation or strain release principle: + Mather-Soete drilling + Gunnert driting * Rosenthal-Norton se ing + Photoelastic coating-driling A\ll these measuring techniques use strain gauges arranged in certain geometric shapes on the removed part of the base metal, and measure the strain changes, which are then converted to stresses. 4.8.2 X-ray Diffraction Method ‘The atomic structure of metals is crystalline. The crystals possess certain constant lattice parameters, which can be found by X-ray diffraction. When the metal is stressed, elastic strain is produced, changing the lattice parameter (atomic spacing), which can be converted to stress, Based on this principle, residual stresses are calculated when comparison is made to the unstressed lattice parameter. 4.9 Summary This module has shown the existence of residual stresses in plates and welded wide-flange shapes. ‘The distribution of these residual stresses does not follow an exact mathematical equation. Therefore, itis not possible to make a precise prediction of how the final product will turn out. However, the more we know about the nature of residual stress, the better we are able to design and fabricate welded products. The purpose of studying residual stresses has been in preparation for the understanding of distortion. 4.40 Estimation of Shrinkages Formulae are available for calculating the amount of contraction or shrinkage of welds. The exact amount of shrinkage is not always calculable because all the variables cannot be exactly controlled. However, these formulae do provide some telltale indicators of which variables exercise the most influence on shrinkage. In other words, these formulae can be used as a guide in practical shop fabrication in order to help keep distortion within the code allowance. Page 14 (© Copyright CW Group - Industiy ServicesMODULE 7 | Residual Stress and Distortion © 440.1 Transverse Shrinkage of Butt Joints ‘The following formula is applicable to carbon and low alloy steels: S=kAJt+ 0.080 Where: ‘S= Transverse shrinkage, mm or in A, = Cross-sectional area of weld metal, mm? or in? = Thickness of plate, mm or in d= Root opening between plates edges, mm or in k= 0.48 for 6 mm ts 25 mm (1/4 in s ts 4 in) k= 0.20 for t> 25 mm (t> 1 in) ‘The graph in Figure 4.12 shows the relationship between the plate thickness (joint volume) and. transverse shrinkage of the 60° included angle single and double V groove welds. It can be observed that a single V groove contracts more than double V groove of same thickness due to the larger joint volume. 40 Ee 6 E rey 9 25 g 60° £ \e ey & 15 0 25 50 75 400 I Thickness - mm Figure 4.12: Transverse shrinkage of single and double V groove welds. Figure 4.13 shows that it of the same thickness, the greater the weld metal area, the greater the transverse shrinkage. This graph shows that to reduce transverse shrinkage, a Joint should be designed with double grooves of the minimum included angle allowed by the applicable standard for the applicable welding process. ‘© Copyright CWB Group -Industy Sorices| Page 15® MODULE 7 | Residual Stress and Distortion Pees 3 7 et bcc eet eta ap ari ou 4.10.2 Longitudinal Shrinkage of Butt Joints ‘The following formula can be used to calculate the longitudinal shrinkage of butt joints (see Figure 4.14). AL= (AA) * 0.026L where: AL = Total longitudinal shrinkage, mm or in L= Length of weld joint, mm or in A, = Cross-sectional area of weld metal, mm? or in? A,= Cross-sectional area of restraining plates, mr? or in? Longitudin A Figure 4.14: Longitudinal shrinkage of a butt oint. Due to restraint, this formula loses accuracy if the cross-sectional area of the plate is greater than 20 times that of the weld. In such cases, the chart shown in Figure 4.15 may be used. It should be observed that in the curves in Figure 4.15, shrinkage attains high values for small Page 16 ‘© Copyright CW Group - Industry ServicesMODULE 7 | Reda Sess ard Dsterion ® resisting sections and falls extremely rapidly as the section increases. The shrinkage tends to become constant when the resisting section exceeds a certain value. The form these curves take should not be a surprise. In fact, the resistance to shrinkage offered by the resisting section increases very rapidly because the effect of shrinkage is at a maximum in a relatively narrow band symmetrical to the axis of the weld. Outside this band, only low temperatures are reached during welding, and the metal offers rapidly increasing resistance to the shrinkage arising from the heated area. The resisting section, once it is outside the section corresponding to the hot parts of the assembly, exerts essentially its maximum resistance. Further increase in the resisting section has scarcely any effect on shrinkage. ‘The following observation makes this phenorrienon more significant. When the cross-sectional area of the weld Is increased, the high temperature transverse portion that is acted upon by shrinkage is larger. The amount of material (section) necessary to completely prevent the effects of weld area shrinkage also becomes much larger. This is what the curves show. The dotted curve in Figure 4.15 shows the resisting section at which shrinkage becomes practically constant. 60 I 55 50 ast 40 30} 25 20 I i i ' i i os “Transverse cross-sectional area of plates joined - in? Figure 4.15: Each curve represents the variation of unit longitudinal shrinkage as a function of the transverse cross-sectional area of the welded assembly for a given cross section. (The shrinkage: tends to become stabilized when the sectional area of the assembly exceeds a certain value, which is Indicated by the dotted line). (© Copyright CWB Group - Industry Services Page 17© MODULE 7 | Residual Sess and Distorton 440.3. Transverse Shrinkage of Fillet Welds of Tee Joints Figure 4.16 shows transverse shrinkage due to the fillet welds of a tee joint. w Vena A ore Figure 4.1 Transverse shrinkage at the fillet welds of a tee joint. A flange may suffer angular distortion as a result of the web-to-flange welds. The extent varies with the flange thickness, since the thicker lange will bend less. A suggested formula for this distortion is: A=0193 WwI.3 112 where: A= The displacement as defined in Figure 4.16, mm W= Width of the flange, mm w= Fillet leg size, mm t= Plate thickness, mm. For the majority of practical cases, flange distortion predicted by this formula is within the tolerance specified in codes such as CSA W859 or AWS D1.1 5. Distortion I ‘The cause of distortion is unbalanced stress. Stresses are induced in materials in a number of ways. Stresses are induced during primary production processes such as cold rolling or casting. In part fabrication, stresses are induced by forming, cutting and driling/machining operations, A product distorts due to unbalanced residual stresses induced during fabrication. The stresses in any component must always be balanced—areas of tension in one area matched by areas of compression in another. When weld shrinkage is unbalanced, distortion occurs, as shown in the following figure of a machine base (see Figure 5.1),The machine base is welded only on one side of the thick plate. The resulting stresses cause shrinkage to occur on the welded side, Page 18 (© Copyright CWB Group -Industy ServicesMODULE 7 | Residual Stress and Distortion ® Figure 5.1: Distortion of machine base due to weld shrinkage. (Photo Courtesy of MacDonald Ste! Limited.) ‘The stresses induced by welding operations are due to uneven heating and cooling. A transferred arc welding process can be considered to be a moving-point heat source, which causes the joint surfaces to melt. Heating and melting is localized to a comparatively small area that expands in proportion to the temperature rise. ‘The following illustration describes the effect of heating and cooling one edge of a bar and the resulting shape change that occurs due to thermal stresses (Figure 5.2). Causes of Distortion Figure 5.2: Distortion results from thermal stresses. © Copyright CWB Group - Industry Services. Page 19© MODULE 7 | Residual Stess and Distortion ‘The edge of the bar expands in the directions itis free to expand. The bar is cooler and stronger farther away from the heat source, and this prevents the edge from expanding uniformly. Uneven expansion, as described in Section 1, causes flow of material. On cooling, the shape change is permanent. The phenomenon of distortion can best be understood through knowledge of the behaviour of residual stresses. 5.1 Distortion Caused by Oxy-fuel Gas Cutting (Oxy-fuel gas or plasma cutting is one of the major causes of distortion as the result of improper application of a thermal process. From the discussion of residual stresses we know that flame cutting induces residual stress and in turn causes distortion, Figure 5.3 shows the mode of distortion that can happen in flame cutting. To avoid this type of bowing (bending) distortion, two torches must be used to cut simultaneously, as shown in Figure 6.4. When cutting a long strip of steel plate of any thickness, for example 3 mm to 300 mm, two torches should be used to apply heat along both edges to keep the plate straight. Direction of cut B A a A (@) Before cutting * (b) During cutting 8 A 6) Cocmoe eee atc (0) After cutting Figure 5.3: Effect of cutting a flat plate with one torch (exaggerated). Page 20 ‘© Copyright CWB Group - Industy ServicesMODULE 7 | Residual Stress and Distortion @® Proposed lines of cuts Scrap (@) Before cutting B I A (b) After cutting Figure 5.4: Effect of cutting a flat plate with two torches. In Figure 8.3, when one torch is used and heat is applied to one edge only, bowing is inevitable after cooling. Bowing is caused by the tensile residual stress induced by the heat of cutting. When two torches are used, as shown in Figure 5.4, the plate stays straight after cutting because the residual stresses along the edges are balanced to each other and the resultant residual stress is equal along the centreline, or neutral axis of the plate. ‘When flame cutting shapes from plate, itis a rule that the workpiece must be kept with the remaining large plate unti the last severance is cut. This prevents the workpiece from moving away from the large plate due to thermal expansion. A good example is shown in Figure §.5—cutting a circular part out of a large plate. The cutting operation is controlled either numerically or by computer, and the cutting torch traverses a perfect circle regardless of the expansion movement. When the workpiece moves due to expansion it will end up slightly oval, and the torch will not return to the starting point. Therefore, where to start the cut and the sequence of cutting must be preplanned. In Figure 5.5, the cut should be started at point A, never point B, when proceeding in counterclockwise direction. Similarly, you can start at point B and proceed in a clockwise direction. ‘Another practical example is shown in Figure .6. In cutting a ring flange plate from a large plate, the first cut, the second cut (removal of the centre piece) and the final third cut are shown. This is different from Figure 5.5, as the cut is initiated by piercing inside the plate, not the edge. Even so, the cutting directions must be followed. Remember that the width of cutting kerf also provides room for expansion. It should be noted that the centre piece (scrap) should never be removed first; otherwise, the inside diameter will change (pull inward) and the width of the ring will vary. © Copyright CWB Group Industry Services Page 21® MODULE 7 | Residual Stress and Distortion Cy of cut B Figure 5.5: Method for cutting out a circle near the corner of a plate. Figure 6.6: Method for cutting out ring flanges near the edge of a plate. 5.2 Distortion Caused by Welding ‘As shown in Figure 1 and in the discussion of residual stress, we have leamed that welding heat causes residual stress and distortion. The most frequently seen types of welding distortions are shown in Figure 8.7. It should be recognized that when distortion occurs it is not always in the simple form of distortion as shown. Distortion quite often occurs in compounded forms, such as bending, twisting or angular, as well as any combination of the simple forms. Page 22 (© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion © ‘Shrinkage Longitudinal shrinkage and transverse shrinkage. Longitudinal shrinkage Trangverse ‘shrinkage “Angular Distortion ‘Caused by transverse shrinkage, ‘Bonding Distortion Caused by longitudinal shrinkage. Buckling ‘Caused by longitudinal shrinkage also to a minor degree by transverse shrinkage). Most often when welding lage, thin plates or sheets. Twisting a Caused by high longitudinal shrinkage; more lyin tin metal Figure 5.7: Types of distortion caused by welding. §.3 Transverse Contraction (Shrinkage)—Angular Distortion Consider a V-groove butt joint as shown in Figure 5.8(A), which is unrestrained (i.e, free to move in response to weld contraction). After welding, this joint will tend to assume the shape shown in Figure 5.8(B). The angular distortion results from the non-uniform contraction of weld metal due to the greater width of the face of the weld compared with the root. Ifthe weld metal could be deposited to form a more uniform section between the edges, as shown in Figure 5.8(C) and Figure §.8(D), there ‘would be (in theory) no angular distortion, only uniform contraction across the joi (© Copytight CWB Group - Industry Services Page 23,© MODULE 7| Rein Sros an Doron Likewise, in fillet welds the distortion resulting from contraction will be as shown in Figure §.8(F) and Figure 5.2(6) for a joint intially set up as shown in Figure 8.9(6). ["] KK {Hose Figure 5.8: Distortion of butt and tee joints due to weld metal contraction. 5.4 Longitudinal Expansion and Contraction (Shrinkage) When we consider movements along a joint the effect of expansion and contraction of the joint ‘edges becomes important because these movements are resisted by the comparatively cool metal surrounding the weld point. Under this kind of restraint, considerable stress is set up in the metal. This is ilustrated in Figure 5.9. In Figure 6.9(A), we assume that a portion of one edge has been rapidly heated. The result is the production of an effect similar to Figure 8.9(8). In this case, the ‘expansion of the heated zone is prevented by the comparatively cool metal. The result ofthis is that the increased volume of metal in the heated zone is absorbed by a slight thickening or upsetting of the plate edge. When cooling, contraction takes place and the edge shortens, producing the shape shown in Figure 6.9(B). This is exactly what is happening to any joint edge or surface during welding, and the magnitude of the cooling effect depends upon the size of the heated zone in relation to the overall size of the plate, Page 24 ‘© Copyright IMB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® Original tength Contracted length Figure 8.9: Figures A and B show how heating and cooling cause distortion of plate edge. Figure C ‘shows how contraction causes plates to take the shape shown by dotted lines. Ifthe edges are restrained, this effort to contract will, instead of causing distortion, set up stresses between the heated area (the weld) and the plate. This will happen if the parts being joined are ‘massive and rigid or if they are rigidly clamped or tacked in place, restricting movement. The effect of both the transverse and longitudinal contraction (shrinkage) of a butt joint where the plate 's not rigid is shown in Figure 6.9(C). The important point, which should be very clearly understood, that local heating always produces contraction during cooling of the base metal. This, with the nal contraction of the weld metal, causes concave bending (Le. sHortening), of the weld side of the joint both transversely and longitudinally. Another example illustrating the plate edge movement during welding is shown in Figure 8.10(A). This figure shows the far end of the joint moving closer during welding with the shielded metal arc welding process (SMAW). This is the result of low heat input and low travel speed, which allows the plate edges to contract. To prevent this from happening a wedge block is inserted at the far end to keep a constant root opening, Figure 8.10(B) shows the far end of the joint moving apart during welding with the submerged arc welding process (SAW). Contrary to SMAW, submerged arc welding is a high heat input and fast travel speed process which keeps the plate edges in an expanding state ahead of the arc during the welding process. In this case, a heavy tack weld or tack welded metal bar at the far end must be used ‘to maintain the constant root opening. Submerged arc welding can produce three times the heat input at five times the travel speed of SMAW. (© Copyright CWB Group - Industry Services Page 25,© MODULE7| Residual Stes nd Dorion ~\ = TT A (A) Unwelded portion of the joint (B) Unwelded portion of the joint ‘loses (in Shielded Metal Arc Welds) opens (Submerged Arc Welds) Figure 5.10: Contraction of two butt-welded plates—effect of travel speed. | 5.5 Other Causes of Welding Distortion ‘We have discussed distortion caused by residual stress, but residual stress alone does not cause bending or angular distortions. When the distribution of residual stresses is symmetrical about the neutral axes of the shape, bending or angular distortion will not occur. Note that longitudinal shortening will always exist and, in thin materials buckling occurs. The neutral axes of some common section profiles are shown in Figure 6.11. The neutral axis is. located through the centre of gravity of the cross-section of a shape. When residual stress is in symmetry about the neutral axis of a member it produces axial stress (tension or compression) only. When the residual stress is not balanced about the neutral axis, a moment is created (see Figure 8.12) and itis equal to force P * e (where e = eccentricity—the distance between the resultant of residual stress and the neutral axis). When the moment is large enough, a visible or unacceptable distortion will result. Figure 8.11: Neutral axes of various sections. Page 26 (© Copyright CWS Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® Figure 5.12: Bending distortion due to eccentricity. For complicated built-up shapes, such as the one shown in Figure 5.13, point A indicates the centre of gravity of the built-up shape and point B is the centre of gravity of the weld areas through which the ‘apparent shrinkage force acts. The distance between A and B is the eccentricity Welds actasa single B apparent shrinkage force ‘through the centre of gravity of the welds Figure 8.19: Bending distortion results when the net iongitudinal shrinkage force of|the welds acts in a line displaced from the neutral axis of the assembly. The line of action of the net apparent shrinkage force is approximately at the contre of gravity of the welds. From the previous discussion, we can conclude that there are five types of distortion: Longitudinal distortion—shortening in length Bending distortion—unbalanced residual stresses 1 2. 3. Angular distortion—transverse contraction 4 5. Buckling distortion—longitudinal plus transverse ‘Twisting distortion—longitudinal contraction likely in thin plates or sheet metal ‘© Copyright CW Group - Industry Services Page 27© MODULE 7 | Residual Stress and Distortion 6. Welding Procedure and Distortion ‘When a welding arc is passing along the surface of a steel plate it creates a very drastic change in temperature variation called a “thermal gradient’, as shown in Figure 6.1. It can be observed that within a few millimeters from the welding arc the temperature may be different by 1,000°C. The magnitude of the temperature drop in a given material is proportional to heat input and travel ‘speed. in a large assembly, distortion occurs because of uneven heating and rapid cooling. == Figure 8.1: Thermal gradient of a welding arc. In previous sections we have already shown several modes of distortion caused by welding. To contro! welding distortion-the relationship between distortion, welding procedures, joint design, preparation and fit- up must be fully understood. 6.1 Welding Procedures 6.11 Weld Pass—Single Pass, Multipass or Small Pass Generally speaking, multipass welding increases the angular distortion; that is, a large number ‘of small passes causes more distortion than a few large passes. The first pass forms a hinge point about which the contraction of subsequent passes takes place. Transverse shrinkage will also be greater because each pass will increase the number of upset areas along the plate edge. Therefore, the greater the number of passes, the greater the distortion tendency. In some cases, when the distortion in the longitudinal direction is more critical, the smaller the cross section of a bead, the less contraction force it can exert against the rigidity of the plates and the more it will stretch. In these cases, therefore, the number of passes should be increased rather than decreased. This apparently paradoxical relationship is a function of the thickness of the plate and its natural resistance to distortion. There is inherent rigidity against the longitudinal bending or shortening of a plate, providing the plate is thick enough. Light Page 28 (© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® ‘gauge sheets have litle rigidity in this direction and, therefore, will buckle easily. Uniess the ‘two plates to be welded are restrained, there is virtually no lateral rigidity since each of the ‘two plates is free to move with respect to the other. Qut-of-plane distortion is more common (buckling). 6.1.2 Travel Speed of Welding Arc ‘The distortion of a joint will be affected by the rate of welding (travel speed). As the arc travels along the joint the heat fans out in all directions from the weld point, as indicated in Figure: 6.2. Any heat that travels ahead of the weld point will distort the free joint edges and must, therefore, be kept to a minimum. Thhe/slower the rate of travel, the more time there is for the) heat to spread ahead of the weld point, as shown in Figure 8.2(A): The faster the travel, the Jess heat spread will occur ahead ofthe weld point, as shown in Figure 6.2(B): Direction of travel Figure 6.2: Arc travel speed and temperature distribution. 8.1.3 Uniformity of Heat Input Expansion and contraction of the metal in the heated zone is further complicated by the fact that the heat input to the joint is not uniform, but, as shown in Figure 6.2, is in the form of a concentrated zone (the weld point) which travels along the joint as the weld progresses. Atthe ‘weld point the heated joint edge is expanding and upsetting. The weld metal is deposited in the fully expanded condition. Behind the weld point, the joint edges and weld metal are cooling and ‘contracting. In front of the weld point, the joint edges are relatively cold and not yet subjected to expansion. Obviously it would be preferable if the heat could be applied to the joint uniformly and ‘simultaneously throughout the whole length. Although this is not practical in most fabrication ‘shops, preheat of work prior to welding does reduce the thermal gradient during cooling in ura” feducing distortion. 6.1.4 Joint Design, Preparation and Fit-Up Ithas already been noted in reference to Figure 5.8 that the more symmetrical the weld section and the more balanced the transverse contraction movements, the less the angular © Copyright CWB Group Industry Services Page 29® MODULE 7 Reside Stes nd Disrtion distortion will be. Joint design should, therefore, be as symmetrical as possible about the longitudinal centreline. From this viewpoint, joint D of Figure 5.8 is preferable to the joint shown in B. Similarly a U groove preparation is better than a V groove. Double-sided U- or J-type preparations are usually only cost effective in thicker material Since weld metal shrinkage is proportional to the amount of weld metal; it follows that the smaller the weld the better. It is therefore the responsibilty of the designer to detail weld sizes ©, matching the calculated strength requirements, and for the shop to make welds no greater than those shown on the drawings. A large fillet wll give more angular distortion than a smaller filet and a wide V groove more than a narrow groove since the contraction at the top will be greater (see Figure 5.8) ‘Therefore, V grooves should be designed for a minimum bevel, consistent with accessibility, and should be carefully prepared to see that this bevel is not exceeded. 6.1.8 Welding Sequence ‘Welding sequence is an essential part of any welding procedure. For example as shown in Figure 6.3 and Figure 6.4, for the same double V groove joint the sequence of weld metal deposited affects the outcome of distortion. Figure 6.3 shows a symmetrical double V groove weld preparation. In Figure 6.3(A), one side is welded completely. The joint is distorted as shown in Figure 6.3(A2). Then, when the other side is welded, the final joint geometry is distorted as shown in Figure 6.3(A4). This welding ‘sequence cannot eliminate the distortion that occurred in step A2 because the joint is locked rigidly, Figure 6.3(B) shows an alternate sequence. The numbering of the weld passes shows that at step 82 the distortion caused by pass 1 is partially eliminated. At step B3 the joint bends slightly upward. At step B4 the joint is recovered to a straight position. This is a satisfactory welding sequence; however, the plate assembly would have to be turned over and back a few times to achieve the final weld. {oe Symmetrical double V groove weld preparation showing the effect of welding procedure: balanced welding (right) prevents distortion. Page 30 (© Copyright CWE Group - Industry Services FigureMODULE 7 | Residual Stress and Distortion @ Figure 6.4 shows a double V groove weld with unequal depths. Figure 6.4(A) shows wel without root gouging and Figure 6.4(€) with root gouging. The one without gouging shows angular distortion and the one with gouging ends up straight. It should be noted that root, ‘gouging is generally performed on the shallow groove side for reduction of angular distortion ‘and cost. Again, the work must be turned over once for down-hand welding. _———— Figure 6.4: Asymmetrical double V groove weld preparation showing how gouging prevents distortion. It should be noted that these two figures (Figure 6.3 and Figure 6.4) do not show the welding Position, rather the sequence of weld passes are shown. All the weld passes are deposited in the flat position, ‘The sequence in which welds are carried out should be studied from the viewpoint of avoiding ‘complete restraint, which will inevitably introduce residual stresses in joints and could increase the probability of cracking if severe. \ ‘Another example is given in Figure 6.8, which illustrates the welding sequence to avoid excess restraint when welding structures consisting of plates and stiffeners. The welding sequence is given as follows: 1) Weld transverse filets. This allows plate A to shrink without restraint. 2) Weld butt joint in plate A—plate is free to move. 3) Butt ointiin stiffener may now be welded while it is free to move. 4) Stiffener may now be welded to plate A. 5) Brackets may be welded to vertical plate; bracket plate is free to move along stiffener. 6) Bracket may now be welded to stiffener. (© Copyright CWB Group - Industry Services Page 31® MODULE 7 Resa Stes and Dstorton Vertical plate PlateA Angle stiffeners Figure 6.5: An example of welding sequence in a structure combining plating and stiffeners. 6.1.6 Seam Welding ‘Seam welding is normally required in shipbuilding or in large fuel tanks when multiple plates are welded along the seams (horizontal and vertical seams) to form the hull or tank wall Correct welding procedures and/or sequence are necessary if smooth surfaces and joint geometry are to be maintained. ‘The simplest form of distortion control is exemplified by the well-known method for welding ‘a longitudinal seam. Start the weld some distance in from the end of the joint and make short weld first, as shown in Figure 6.8. In this way, the first weld pre-sets the joint edges and prevents the closing in of the joint as the main weld proceeds. First weld Main weld Start Figure 6.6: Simple welding sequence. Ithas already been mentioned that distortion control involves applying the proper pattern of heat distribution. We have seen how this principle may be applied by welding equal and ‘opposite welds. Also it has been noted that it would be desirable to apply heat uniformly and simultaneously throughout the entire length of a joint. As this is obviously not possible in arc welding, the next best thing is to weld at spaced intervals along the joint. (© Copyright CWS Group - Industry Services Page 32MODULE 7 Residual Stas an ison ® Figure 6.7 shows several sequences that apply this principle. A simple back-stepping method. Is shown at (A). This consists of starting a weld a short distanca from the end of a seam. The ‘next weld is then started a similar distance from the first weld and is fused in to the previous starting point, and so on, until the joint is completed. (B) is a minor variation of (A), leaving unwelded spaces in consecutive steps, called back-step and skip welding. On long joints the welder works outward from a central point, as shown in (C), (D) and (F). This is an important principle to follow. Still more elaborate variations of this procedure are the "staggered or “wandering” sequences shown at (E) and (F). These procedures consist of leaving spaces between each weld bead, progressing along the seam in this manner and then completing the unwelded spaces. D TCM 4 2 6 En ne $e a Gia Oe rm arate aeareecusa sal anal F mm cD SN sae are Figure 6.7: Seam welding technique. ‘With @ large plate fabrication, as shown in Figure 6.8, welding should start at a central point ‘and proceed outward; keeping the progress of welding as symmetrical about the centre as Possible, as shown by the numerical order. The principle of this technique is to progress in a way which allows each joint to have freedom of movement for the maximum amount of time. © Copyright GWE Group - Industry Services Page 33® MODULE 7 | Residual Stress and Distortion Figure 6.8: Suggested sequence for plate welding. ‘At the junction of a seam (horizontal) and a butt joint (vertical), the preferred welding sequence is shown in Figure 6.9. The seam adjacent to the butt joint should be left unwelded for a length ‘of 300 to 400 mm on each side and then completed after the vertical butt is welded. This ‘sequence allows contraction of the butt joint and avoids high rigidity. 300 mm to 400 mm Figure 6.9: Sequence for seams and butt joints. 6.1.7 Non-continuous Fillet Welds ‘The seam welding technique shown in Figure 6.7 may be used for both butt and tee joints, but in the latter case the welds may be staggered on both sides of the joint as shown in Figure 6.10. The main advantage of non-continuous fillet welds (or intermittent filet welds) is that heat input to the joint is considerably less, and therefore distortion and shrinkage stresses are reduced. Page 34 (© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® Figure 6.10: Intermittent fillet welds. Itis, of course, essential to make sure that a non-continuous weld will provide the required joint strength and Ifa complete filet weld is required, non-continuous welding cannot be adopted. Another advantage is that the heat is more uniformly distributed than it would be in the case of ‘a continuous weld. Moreover, the longitudinal weld shrinkage and, therefore, overall distortion, is only a small fraction of that produced by continuous welding. It has, in fact, been found that the reduction in these factors is far greater than would appear to be represented by the Proportion of intermittent to continuous welds. 6.1.8 Built-up Structures—Neutral Axis ‘The advantage of equal and opposite welding about a neutral axis has already been noted. ‘The neutral axis always passes through the centre of gravity and is usually defined as the line ‘on which there will be neither tension nor compression when the piece is flexed or bent. In the case of a plate, the neutral axis coincides with the centre plane of the plate (see Figure 5.11(A)); similarly, in the case of an I-beam, the neutral axis coincides with the centre of the web (see Figure 5.11(B)). In the case of a tee or angle section member, arranged as shown at Figure §.11(C) and Figure §.11(D), the neutral axis is not in the centre of the depth, but is near the flange. ‘A clear understanding of the position and function of the neutral axis is necessary if the effects of welding a plate, section or complete weldment are to be visualized. As previously mentioned, the simple act of depositing a bead of weld metal on the surface of a plate will cause that plate to bend with concavity on the welded side. This is due to the contraction of the weld metal exercising a shrinkage force, which is then offset from the neutral axis of the plate. If, on the other hand, beads were deposited simultaneously on opposite sides of the plate, the contractions of the two welds would be balanced about the neutral axis and there would be no bending or distortion. ‘© Copyright CWB Group - Industry Services Page 35@ MODULE 7 | Residual Stress and Distortion This balancing of welds about the neutral axis of a built-up section or structure is the most important fundamental point in reducing distortion. A further example is shown in Figure 6.14, ‘where various welds are arranged around the neutral axis of a built-up section. The sequence in which the welds should be made is indicated. Fe im Figure 6.11: Balancing the sequences of welds about the neutral axis of a section. Emphasis so far has been laid on the importance of welding equally about the neutral axis to maintain alignment. This assumes that the structure is true to begin with. In some cases this may not be so, and welding unequally about the axis may be used as a means of straightening. case in point is the construction of a beam from plate sections where the web plate has a curvature as received from the mill, This might be as much as 10 mm in a 1,500 mm plate. The following procedure may then be used to produce a straight beam (see Figure 6.12). oe ES * Fiange f, First joint to be completely ' — Sh 2] BZ Figure 6.12: Operations in welding a built-up I-beam with curved web. The flange plate f, is laid down on a siab and the web plate (with convex side down) is set up vertically on it, as shown in Figure 6.12(A). The flange is then pulled up to the web plate and strongly tacked, as shown in B. The welding ofthis flange to the web is then carried out until Page 36 ‘© Copysght CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion @ the web is not only straightened, but slightly bent in the opposite direction as shown in C. The second flange f, is now fitted to the web and tacked securely as shown in D. Welding is then ~ completed, preferably using two welders on opposite sides of the web and working in the same direction. With such a sequence the beam should be reasonably straight upon completion. Welding the first flange to the web before the second flange has been tacked in place results in considerable bending due to the shortening of the weld, as the beam is not strong or stable without the second flange. If, in performing such welding, the beam is slightly “overbent", the welding of the second flange ought to be just sufficient to pull the beam back to the straight position. Due to greater rigidity, the shrinkage effect will not be as great as under the conditions in. which the first flange was welded. 6.1.9 Complicated Weldments—Accurate Assembly In the case of complicated assemblies, the accuracy of preparing the various components requires careful consideration to enable dimensional tolerances to be kept to a minimum. An accumulation of tolerances over a number of components may create unacceptable results. ‘Obviously the more generous the tolerances, the greater the fit-up gaps, and an excessive ‘amount of weld metal will be needed, resulting in greater distortion than would otherwise be involved. ‘To avoid this, it may be desirable to machine components to size to obtain close tolerances and increase the accuracy of he final weldment. Itis also often possible to arrange the assembly ‘of components in such a way that cuulative tolerances can be controlled and prevented from adversely affecting the final accuracy of the structure (see Figure 6.13) x 208mm 3mm ‘Assembly A Assembly 8 Figure 6.13: Arranging components to ensure finished accuracy. ‘Where accurate location points are essential, the assembly arrangement of the structure should provide for some allowance in case the various subassembly allowances do not work ‘out to the degree of accuracy expected. For example, in the case of built-up I-beams, the ‘accumulated longitudinal contraction of the flange-to-web welds and the transverse contraction of the stiffener welds will result in appreciable shortening of the beam. It is usual to leave the flange and web plates overlength so that they may be finished to size after welding. ‘Similarly, for machine structures such as bedplates, engine frames, etc., those points that ‘must be located to close tolerances should be fixed only by the last weld, which affects their location. (© Copyright CWB Group - Industry Services Page 37© MODULE 7 | Residual Stress and Distortion With a tolerance of + 1.6 mm on plates X and Y of Figure 6.13, Assembly A would necessitate a tolerance of + 3 mm, whereas the accuracy of B could be # 0.8 mm. 7. Control and Correction of Distortions The causes and types of distortion have been discussed. What happens when the weldment is distorted beyond the allowance by the applicable codes or standards? What are the common measures used by the welding fabrication shops to prevent distortion? What corrective actions can be taken to eliminate distortion ‘once it has occurred? A brief discussion will be given in the following paragraphs. 7A Control of Distortion; In previous discussions of welding and distortion, several ways of preventing distortion have already been mentioned. The following is a summary of methods of distortion control by welding procedure: D} 2 3) 4) 5) 6) n 8) Ensuring accurate joint preparation and fit-up—this is one way to maintain minimum weld metal for the joint Using back-stepping or skip technique Welding progresses outward from a central point Balancing welds on either side of a centreiine, central point or about the neutral axis of a section Welding butts joints before fillets to allow large contraction to take place first Using intermittent filet welds instead of continuous fillet welds when allowed by design Arranging the weld sequence such that each the longest possible period Dividing a weldment into subassemblies to reduce cumulative distortions or shrinkage. In addition to the above procedures, which are aimed at reducing distortion, the following points should be given attention since they are particularly concerned with the production of an accurate weldment: 2) Applying welding so as to counteract plate edge curvature, if any b) _ Arranging components so as to avoid accumulation of errors due to tolerances on purchased shapes (beams, channel, plate, bar, etc.) ©) Where a high degree of overall accuracy is required, preparing components accurately to reduce the stacking of fit-up tolerances d) Arranging the sequence of welding so that location points requiring a high degree of accuracy are assembled and welded last e) Allowing for weld metal shrinkage f) Arranging for some latitude in assembly dimensions; a weldment can be machined to size if shrinkage and other allowances do not work out as expected ‘Other means of distortion control include, y 2) 3) Preheating—reduces shrinkage stress by providing more uniform heating and cooling Peening—reduces final shrinkage by stretching the weld metal. Restraint—any degree of restraint, external or self-weight, may be expected to reduce the ‘amount of shrinkage. Such restraint may be applied in any of the following ways: Page 38 (© Copyright CWB Group -Industy ServicesMODULE 7 | Residual Stress and Distortion ® @) Clamping b) Rigid tacking ©) Maintain minimum or zero root opening (reduces transverse shrinkage) 4) Cooling between weld passes (reduces the restraining required) 4) Mechanical control: a) resetting to allow recovery of angular distortion, as per Figure 7.4 b) —_Use of temporary stiffeners ©) Use of strong backs or special igs or fixtures 4) Artificial cooling Dotted line position Figure 7.1: Pre-setting of joint members to allow for contraction of weld metal. thas been already noted that distortion may be reduced by fixing components either by tacking, ‘clamping or assembling in jigs. However, complete rigidity by clamping increases residual stresses. ‘Therefore, unless the weld metal can be permitted to contract freely, for example as in a pre-set join, a balance must be found between the extremes of free movement and complete rigidity so that both distortion and stresses may be kept to a minimum. ‘Accurate edge preparation and joint fit-up has considerable influence on the praduction of stress-free Joints. A varying and unnecessarily wide joint causes considerable heat concentrations at the wide places, thus creating excessive residual stresses in the assembly. ‘Another preparation fault is excessive root face, particularly f accompanied by a tight fitting joint. Not only is complete fusion of the joint difficult to achieve, but shrinkage of the deposited metal will be Prevented. The result wil be high shrinkage stresses that could increase the probability of cracking during fabrication or, later during service. ‘The rigid alignment and complete restraint of joints by strongbacks, clamps and such devices should be avoided. Figure 7.2 shows several methods commonly used to align joints. in (A) the joint is made rigid and the method is entirely incorrect. In (B) and (C) the joint is free to contract and the methods are suitable, while (D) is correct ifthe jack is removed after tacking and before final welding, (© Copyright CWB Group - Industry Services| Page 39MODULE 7 | Residual Stress and Distortion Figure 7.2: Methods of joint alignment. ‘Weldments may be rigidly clamped to heavy sections of bases during welding. They may also be mounted in rigid fixtures or assembled and rigidly tacked for welding. Strongbacks and temporary stiffeners may be used to align and rigidly maintain edges and joints. Heavy slabs and fixtures will not only hold assemblies rigid, but also will withdraw the heat of welding from the weldments, thus further reducing distortion. A similar effect can be obtained by immersing assemblies in water or by spraying if the metallurgy allows it. However, none of the methods of restraint can be expected to fully retain alignment. Some springing and distortion will usually follow release from such superimposed control. Further, the greater the restraint against contraction, the greater the residual stresses induced and the ‘more likelihood that cracking will result. Heavy weldments of thick plate may, in themselves, offer great rigidity and restraint to welds. Figure 7.3 shows cover plates welded to H-sections. The filet welds will have a tendency to shorten due to their longitudinal contraction. This contraction will cause bending and a shortening of the sections. If they are tacked or clamped together as shown, this bending tendency in each will be counteracted by the other. SS |_— cramp or Teporay | Ss Figure 7.3: Eliminating distortion by balancing weld contractions of two similar weldments clamped or tacked back to back. Page 40 (© Copyright CW Group - Industry ServicesMODULE 7 | Residual Stress and Distortion @® ‘The procedure should be to start welding in short increments outward from the centre, alternating from one section to the other so that equal and opposite welds are made altematively and thus ‘counterbalance each other. Itis understandable that distortion will be increased in large assemblies where the welds are long. It therefore follows that ifthe job is broken down into a number of smaller weldments or subassemblies, the distortion in each will be less and can more easily be controlled and corrected. if necessary, each subassembly can be straightened or machined before final fitting and welding. Therefore, final fabrication from subassemblies is recommended, and the designer should bear this requirement in mind, Subassemblies make for easier and more efficient material handling and reduce the ‘accumulation of additive residual stresses. Experience has shown that control of distortion and reduction of welded-in stresses can be achieved by carefully planning the welding procedure. 7.2 Correction of Distortion Although the foregoing suggestions for control and minimizing of distortion are helpful, it should be appreciated that, despite the observance of all reasonable precautions, distortion may still occur. Any such distortion will, however, be much less severe than it would have been had no precautions been taken. When the distortion is greater than allowable, corrective measures are necessary. There are ‘two common methods used: 1) Mechanical straightening—uses mechanical devices, such as jacks, presses or specially designed straighteners, as shown in Figure 7.4 2) Application of heat—uses the principle of resisted expansion during heating and subsequent contraction on cooling Figure 7.4: Press for mechanical correction of distortion. (Courtesy of Canron Construction Corp.) 7.2.1 Mechanical Straightening It should be pointed out that each method is suited for certain applications. Most mechanical straighteners are suited for minor straightening, Heavy components require specially built straighteners that are not available in small fabrication shops. (© Copyright CWB Group - Industry Services Page 41® MODULE? |Restal Sess and Distortion In severe situations, mechanical straightening can cause cracking. Mechanical straightening applies external forces to further ‘stretch’ the fabrication. If the applied force working in ‘addition to the residual welding stresses exceeds the ultimate strength of the material, failure will occur. Care must always be taken when using mechanical methods to remove welding distortion 8. Flame Straightening and Flame Forming 841 Introduction ‘The term “flame straightening’ or “lame forming" denotes the action of changing the shape of ‘member by the application of localized flame heating followed by cooling. It may be for the purpose of straightening, correcting distortion or introducing a deliberate curve to the member such as in “beam ‘cambering’. Earler in this module we discussed how local heating of a plate gives rise to residual stresses on cooling. As the plate is locally heated, expansion is restrained by colder surrounding material, thus putting the heated area into compression. With a rise in temperature of a few hundred degrees the compression stresses exceed yield and the metal plastically upsets. On cooling, the heated zone contracts and goes into tension. The material is shortened across the plastically upset area and the residual tension stresses cause the plate to bend or otherwise change shape. From the above description it will be understood that any of the following procedures can be applied to both straightening and forming. 8.2 Basic Principles ‘To illustrate the general principle of flame forming, look at the example in Figure 8.1. Here a triangular wedge with base width D, and depth Wr2 is marked off as shown in Figure 8.1(A). When itis heated to about 650°C it expands, goes into compression and plastically deforms (see Figure 8.1(B)). At this point D, is greater than D,. On cooling, the wedge shrinks until it is smaller than its original size, and , is less than D, but is locally thicker because of the plastic upset. The resulting shape of the bar is bent, as illustrated in Figure 8.1(C). This is the basis of flame straightening or forming operations in which specific patterns on a member ‘are heated to induce desired shape changes. A general guide to flame straightening is as follows: + aheated member expands and grows in the unrestrained directions. + | restraining the expansion produces a compressive stress. 4. + ifthe restraining element is not stiff it could buckle under the compressive stress. + plastic flow must take place to induce residual stresses and change the shape. + when a heated member cools non-uniformly, the portions that cool first go into compression and the portions that coo! last go into tension Page 42 (© Copyright CW Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® 8.3 8.4 “Tiangular wedge Cold W > marked off on cold plete B . Wedge heated to about 650°C. Expands: D, > D, NZ Cold part restrains expansion so wedge isin Hot Compression, Thickness cof heating wedge increases due to plastic deformation, On cooling wedge shrinks: D, < D, ‘Shrinkage restrained by ‘old material so wedge is in tension. Thickness of wedge after cooling is ‘greater than original thickness because of Plastic deformation, Figure 8.1: The basic principle of flame forming in which bending of a member Is induced by heating local region until it upsets then allowing it to cool. Maximum Temperature—Low Alloy Steels ‘The temperature reached during heating is very important if metallurgical damage to the steel Is to be avoided. THe itiakimum temperature should be approximately 650°C; where the metal has just become a dull or cherry red color. Higher temperatures may cause metallurgical transformations, changes in properties, oxidation or surface damage. The colour. should never be allowed to become bright red or white-hot. Temperatures much higher than 650°C will have litte effect in inducing shape changes. Ifthe temperature is too iow, expansion will be insufficient to produce plastic deformation (upsetting). Before beginning a flame forming operation the procedure should be mapped out and the pattern of heat spots marked on the member. Heating should be done using oxy-fuel and conventional ‘equipment, and a rapid heating rate is desirable. When the steel shows a dull red, the flame is moved to the next pre-marked spot. Removing a Buige Bulges in plates can be flattened using spot heats about 50 mm in diameter. These are located on a spiral path chalked out on the crown of the bulge. The first spot is made at the centre and succeeding ‘spots follow the spiral outward, as shown in Figure 8.2. The heat is always applied to the convex side. ‘Spraying the underside with water or a jet of compressed air while heating the top will accelerate the flattening. Where access to the second side is not possible, the heat spot can be sprayed from the torch side. Note that water spray can only be used if no metallurgical damage will result. © Copyright OWE Group -Industy Sonvicns Page 43® MODULE 7 | Residual Stress and Distortion Heat applied to convex side 7 ———— oO —_—>___ ss W (* ‘50mm heat spot, Figure 8.2: Removing a bulge in a plate by spot heating in a spiral pattern. A variation of spot heats uses high external confinement forces produced by a press or sledge hammer and requires a larger spot, about 100 mm diameter. Heating a 100 mm diameter circle can ‘cause the plate to buckle and bulge locally towards the heat source, as can be seen in Figure 8.3. ‘When the heated bulge forms, the spot is forged by pounding with a sledgehammer or squeezing it fiat with a press. This displaces more material plastically, and radial shrinkage on cooling produces residual tension and flattening, Placing a steel or brass block on the bulge protects the member from. the hammer blows. Buco aotniran_ ‘ont Protecive Press a ~ 650°C ‘brass block a NZ Figure 8.3: Removing a bulge using external confinement forces. Page 44 {© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® 8.5 Line Heats A\line heat is rapid heating along a line chalked out on the convex surface to be shortened. A single orifice torch producing a narrow heat line is necessary for material less than 12 mm thick. A good technique results where about 3/4 of the plate thickness reaches the temperature needed to initiate Plastic flow. Since the higher temperature is on the torch side, the effect is similar to a vee heat projected through the thickness. This causes rotation, with the convex side shortening. 8.6 Strip Heats A strip heat is a broader version of a line heat, but the travel speed is slower since a slight weaving action is necessary. The plastic flow temperature should be reached through the total thickness using ‘a torch on each side if necessary. Strip heats are directional, pulling twice as much transverse to the heat as parallel to the axis of the strip, and they are more effective than spot heats on buckled plates if the buckles are directional, as Figure 8.4 illustrates. Strip heats 100 - 200mm, 3x thickness of panel Figure 8.4: Using strip heats to remove a buckle in a plate. Figure 8.5: Flame straightening using two torches. (Photo courtesy of Babcock & Wilcox Canada.) © Copyright CWB Group - Industry Services Page 45© MODULE 7 | Residual Stress and Distortion, 8.7 Vee Heats \Vee heats are used to produce a bend in a flat bar or plate. As shown in Figure 8.8, the heat source is applied first at point A, close to one edge. The point A should be located 2/3 to 3/4 of the member ‘width from the edge to be shortened but should not be directly on the edge. Point A is circled until the ‘material under the torch is a dull red colour (about 650°C), and is plastically upset. Since it is confined by surrounding cold material, plastic flow isin the direction of the thickness of the material. Now the torch is moved along an ever-widening path, keeping ahead of the dull red spot. The volume heated at any time is always confined by relatively cold metal. The path must be followed as a one-directional continuous trip. c op B Before heating A cop GF VW During heating ‘After cooling Figure 8.6: Procedure for inducing a bend in a member using vee heats. ‘When point B is reached, there is less cold material to resist expansion and by now the previously heated volumes are cooling and shrinking, producing compressive forces across C-D. This assists i upsetting the material, as there is less surrounding cold metal, but there must be no retracing of any step since this will remove the benefit of the confinement force resulting from shrinkage. nee the entire vee oh the fiat bar has been heated progressively there is nothing to resist cooling and shrinkage. The vee upsets proportionality from point A to the base C-D, with point A acting as a hinge. 8.8 Beam Forming or Beam Cambering ‘To camber a beam, a block heat is used on one flange and a vee heat is used on the web, as shown in Figure 8.7. The block heat is started at the point of maximum confinement, which is over the web on the top flange centreline. The heat line is blocked out symmetrically towards the flange edges. ‘To induce sweep in a beam a block heat is used in the web and vee heats are used on both flanges. ‘The beam should be supported at the ends, with the web horizontal and both flange vee heats applied simultaneously. Page 46 ‘© Copyright CWB Group - Industry ServicesoDULE7 Rein Stee a Dn ® To produce camber eral To produce sweep Figure 8.7: ilethod for producing a bend in a beam using a combination of vee and block heats. 8.9 Channel Forming Patterns used to form channels vary with the direction in which the movement is to take place. ‘Cambering a channel is identical to cambering a beam, with a block heat on the flange starting over the web and a vee heat in the web (see Figure 8.8). To produce sweep in a channel with the web on the outside of the curvature, only the flanges need be shortened by using vee heats (see Figure 8.9). ‘Sweeping in the other direction requires block heats to shorten the web and vee heats to shorten the flanges (see Figure 8.10). Note that the block heat on the web is applied in two sections. Each section starts at the edge of the web at the point of maximum restraint over the flange. Heating progresses toward the centre of the web and both sections should be done at the same time. For maximum effect, tm flange vee heats should be done first and allowed to cool completely before the web is block sale Ue a Figure 8.8: Method for producing camber in a channel by applying a block heat to the flange and a vee heat to the web. (© Copyright CWE Group - Industry Services Page 47® MODULE 7 | Resa! Stess an Distorton iS eo heats ony Figure 8.9: Method for producing sweep in a channel by applying vee heats to the flanges. No heat is applied to the web. The web is on the outside of the sweep. EEE: Swoop Figure 8.10: Method for producing sweep in a channel by applying vee heats to the flanges and a block heat to the web. The web is on the inside of the swoep. The web is heated in two separate blocks. 8.40 Angle Forming Forming an angle so the upstanding leg is on the outside radius of curvature, as shown in Figure 8.11, requires shrinkage of the horizontal leg only, with vee heats. They start at the inside of the leg and proceed to the free edge as shown. Block heats are not used. Forming an angle so the upstanding leg is on the inside radius of curvature requires shortening the upstanding leg with block heats and shortening the horizontal leg with vee heats (see Figure 8.12). Note that in this instance the block heat starts at the centre of the upstanding leg instead of over the other leg. This is necessary sitice the section is not symmetrical, and heating intially over the other leg will try to camber the angle downward. The vee heat should be made before making the block heat. Figure 8.11: Forming an angle with the upright leg on the outside of the curvature. No block heats aro needed. Page 48 (© Copyright CWB Group -Industy ServicesMODULE 7 | Residual Stress and Distortion © Figure 8.12: Forming an angie with the upright leg on the inside of the curvature. Both block heats and vee heats aro used. 8.11 Cambering Beams 8.11.1 Flame Cambering Beams There are two methods used to flame camber beams. The first uses strip heating along the full length of the flange directly over the web (see Figure 8.13). This is usually performed ‘automatically with a multi-orifice tip on a heating torch mounted on a carriage. On heavy sections with thick flanges and webs, three torches are used—all mounted on the carriage and travelling together—as shown in the figure. The flange-to-web intersection region upsets, and shrinking induces high residual tensile forces when cool. This produces high compressive stresses in the outside flange edges and along the top of the web. The beam cambers as the tensile and compressive forces reach equilibrium. FAN Tight Beams Heavy Beams Figure 8.13: Flame cambering a light beam (left) and a heavy beam (right) by heating over the web. On heavy sections three torches are used, mounted on the same carriage. {© Copyright CWB Group- Industry Services Page 49MODULE 7 | Residual Stress and Distortion ‘The advantage of this method is that cambering is done automatically with the torches on a carriage that travels at a constant speed. The procedure is easily duplicated with few variations in results. There is a clear advantage if several beams are cambered at the same time. Operator skills and experience are minimal, There are, however, a number of disadvantages. The method can result in buckling in the web land flange edges if they are too slender to resist the high residual compressive stresses. It can also be noted that the camber is held by intemal forces at or near the yield point. If the beams are shipped long distances by rail or over rough roads, much of the camber is lost because vibration or cyclic loading tends to relieve the residual stresses necessary to maintain the ‘camber. This also means the method is not recommended for fatigue loaded structures. 8.11.2 Alternate Method of Cambering Beams ‘The second method of cambering beams uses block heats on the flange and vee heats in the web as we have previously described. Heating is done manually with a single orifice tip, torch (or a torches on each side for thick flanges) and produces shrinkage in both the web and flange. This condition leaves minimal residual stresses in the cambered beam (see Figure 8.14). All vee heats made [*—*] Typically 125mm first. When cool block heats are made Figure 8.14: Alternate method of cambering a beam using vee and block heats. The technique is as follows: . Calculate the number of heats required to camber the beam. (See Section 8.13.)This is a function of the amount of camber required, the depth of the beam and the length of the beam. . ‘Support the beam at each end with the web vertical. The weight of the beam ‘compresses the top flange and assists the cool, confining material to resist the ‘expansion of the block heats, thus increasing the amount of upsetting in the flange. Page 50 (© Copyright CWB Group - Industry ServicesMODULE 7 | Residual Stress and Distortion ® + Mark the patterns on the beam top flange and on the web. * Heat the vee patterns in the web to cause plastic upsetting, * Cool the vee pattems completely with wet rags or a jet of compressed ait. This shrinks the vee, which is restrained by the surrounding web material. The resulting tension in the web induces balancing compressive forces in the flange, which add to those already induced by the weight of the beam. + Heat the flange using block patterns. The expansion of the blocks is resisted by the confining cold flange material and the forces described previously. Upsetting is accelerated since plastic flow starts at lower temperatures because of the high compressive confinement. * Cool the block patterns completely with wet rags or a jet of compressed air. On cooli the flanges shrink about the same amount as the web. The flanges are not restrained ‘on shrinking since the web shrinkage (vee heat) assists the flange and the restraint disappears in the process. This technique produces a permanently cambered beam that is virtually stress free. + Check the camber. Note that the beams have been cambered upside down. * Ifthe camber is insufficient, mark new patterns between the first heats and continue until the desired camber is achieved. ‘Advantages of this method are that the degree of camber can be easily calculated in advance and it results in a cambered beam good for fatigue as well as static applications. The disadvantage Is that itis a very time consuming method. 8.11.3 Heat Pattern Calculations Cambering a beam requires that one flange be shorter than the other. Therefore if we can calculate how much a flange must be shortened to give the desired camber, we can calculate the number of heats required in that length of beam to achieve the total shrinkage needed. The extent to which a heated member is shortened depends on a great many variables, Different Conditions of heat and restraint produce shrinkage strains that can vary between 0.004 mm/ ‘mm heated to 0.025 mm/mm heated. Experiments using the vee method described before suggest that when using the prescribed procedure on a beam the shrinkage is approximately 0.018 mmimm of flange block heated ‘The following calculation is based on the assumption that flange shrinkage is 0.02 mm/mm of heated block. This slight conservatism allows for the use of additional heats ifthe full specified camber is not achieved on the first run. It is assumed that the flange block heats are 125 mm long, so each block shrinks 2.6 mm, When a beam is to be cambered we must calculate the difference in length between the top ‘and bottom flanges to achieve the camber. This total shrinkage divided by 2.5 gives us the ‘fumber of 125 mm heat patterns needed. This is best illustrated with the example in Figure 48. Suppose we wish to put a 75 mm camber in a 1 m deep beam 12 min length, Through (© Copyright CWB Group Industry Services Page 51© MODULE 7 | Residual Stress and Distortion simple geometry it can be shown that the difference in length between the top and bottom flanges is 50 mm. Figure 8.15 shows the calculation of this. Thus the number of heat patterns required is 50+ 2.5=20. Calculating number of heat patterns If the beam has a camber A ina length L, then from simple geometry we have: (5)'= 20a ita< |. ea 2 Ce i z ace | creamer cian 5 ape poe «|e a ap fe tae fe 7 ier fen a 2 ee [a ¢ 2 [> | eet 0 a feeet eT «te Page 68 (© Copyright CWB Group - Industry Services© Copyright 2013 CWB Group - Industry Services ‘llrightsreserved. ‘CWB Group 1-800-844-6790 wnw.cwbgroup.org