Muhammad Saad Ali

1 Chapter 1: Introduction If we take a look around us we will see various objects such as computers, watches, cars, pencil, lights, fans and different other manmade products. We will notice that all these objects and their individual components are made from a variety of materials and have been produced and assembled into the finished item. This is done by a combination of processes called manufacturing. According to DIN 8580, manufacturing processes are classified into six main groups ➢ Primary shaping ➢ Material forming ➢ Dividing ➢ Joining ➢ Modifying material property ➢ Coating Fig 1.1: Classification of manufacturing processes [1] 2 1.1 Forming: According to DIN 8580 forming is defined as, “Manufacturing through the three-dimensional or plastic modification of a shape while retaining its mass and material cohesion.” Forming is the controlled modification of shape. Forming processes are characterized as chip less or non material removal processes. 1.2 Hydroforming: Hydroforming is a cold forming process that uses hydraulic pressure in place of hard tooling to plastically deform a given work piece into a desired shape. This technique is based on the inflation of, for instance a tube in presence of axial or radial compression by subsequent expansion and sizing against die wall. By this process we can manufacture geometrically highly complex parts at a reduced cost and increased strength as compared to stamping, forging or casting processes. 1.3 Types of Hydroforming: Hydroforming process can be categorized in two main types. • Sheet hydroforming • Tube hydroforming 1.3.1 Sheet Hydroforming: In sheet hydroforming a work piece is placed on a blank holder over a male punch. Then a hydraulic pressure surrounds the work piece at a relatively low initial pressure. Then pressure is increased to as high as 1000 bar, which forms the work piece against the punch. Then the pressure is released and punch is lifted and then work piece is removed. This technique allows much deeper draw which is necessary for manufacturing panels with complex curves. 3 Sheet hydroforming has further two types. • Bladder forming, where there is a bladder that contains fluid and no liquid contacts sheet. • Hydroforming in which fluid contacts the sheet and no bladder is present. Fig 1.2: (a) General description of simple sheet hydroforming operation, (b )Sheet hydroformed front hood (courtesy of Schuler Inc., 1998) [2] 1.3.2 Tube Hydroforming: Tube hydroforming is a material forming process in which tubes are formed into complex shapes using simultaneous application of internal pressure and axial compressive forces at both or either end or simply application of internal pressure (see fig). Internal pressure can be obtained by pumping a fluid into the tube or pressing a viscous medium such as elastomers, rubber or polyurethane in the tube. In tube hydroforming pressure is applied to the inside of a tube that is held by dies with the desired cross sections and forms. When the dies are closed, the tube ends are sealed by axial punches and the tube is filled with hydraulic fluid. The internal pressure can go up to a few thousands of bars and it causes the tube to calibrate against the dies. The fluid is injected into the tube through one of the two axial punches. Axial punches are movable and their action is 4 required to provide axial compression and to feed material towards the center of the bulging tube. This process is shown in fig. If we use a viscous medium such as an elastomer we will require some mechanism to apply force on elastomer inside tube. We can use a piston for this purpose. Figure 1.3 Hydroforming of an axisymmetrical part, Siegert et al. (2000) [26] Fig 1.4 Steps in a typical hydroforming process shown on a small tubular part (courtesy of Siempelkamp Pressen Systems) [3] 5 1.4 Advantages of Hydroforming: Hydroforming process has significant advantages over conventional forming techniques, stamping, forging and casting processes. 1.4.1 Saving of Tools: One advantage of hydroforming is saving of tools. For example for sheet hydroforming we only require draw ring and a punch or male die. There is no need for female die. Changes in material thickness can be usually made without change of tools. However dies need to be polished and for tube hydroforming a two piece die is required to allow opening and closing. 1.4.2 Complex Geometries: By hydroforming it is possible to produce highly complex geometries in single step. By sheet hydroforming it is possible to produce almost limitless geometries. However, the process is limited by very high closing force required in order to seal dies.Tube hydroforming (THF) can also produce many complex geometries, reducing the need for tube welding operations. 1.4.3 Tolerances and Surface Finish: Hydroforming is capable of producing parts within tight tolerances including aircraft tolerances where a common tolerance for sheet metal parts is within 0.76mm (thirty thousandths of an inch). Sheet metal hydroforming also allows for a smoother finish as draw marks produced by the traditional method of pressing a male and female die together are eliminated. 1.4.4 Effect on Work Material: When a blank is hydroformed the metal flows around the die rather than stretching, which produces less material thinning, and also reduces the rate of work hardening which helps eliminate the need for an annealing process on some parts that might otherwise require further forming operations. 6 1.5 Application of Hydroforming: The design and structural possibilities offered by hydroforming are used successfully in a number of applications for the manufacture of high strength components and assemblies with optimized service life and weight characteristics. In industry it is used for making various tubular parts. The production of a T piece by hydroforming is shown in following figure. Fig 1.5: Production of T fitting [1] On the basis of the various semi-finished product shapes (Fig. 5.2.8), hydroforming can be applied to a wide variety of fields. Three main user groups have emerged as the most significant over recent years 7 • the automotive engineering and supply industry • sanitary and domestic installation industry • pipe and pipe component manufacturers Fig 1.6 : Blank shapes [1] Table 1.1: Fields of application for the hydroforming technique [1] Sector Assembly Components automotive industry vehicles chassis, exhaust and intake cross members, side • road systems, add-on parts, drive members, manifolds, roof • water system, seats, frames/ rails, spoilers, gear shafts, • air bodywork, steering seat frame components, • rail A/B/C pillars, roof frame profiles, steering column with compensation 8 chemical, gas, oil industry, piping and tank components, T-fittings, reducers, power station construction pipe fittings housings, paneling domestic appliance industry fittings, machines tube bends, T-fittings, reducers, cross pieces, bends bicycle industry pipe fittings pedal bearings, joints, frames processing of sections frame construction, joints, girders, calibrated semifinished pipes, rail- tubes, sections, roof arches, bound vehicles, utility frames, structural members vehicles pumps and fittings housings intake pipes heating, ventilation, air fittings tube bends, reducers, T- conditioning furniture industry fittings, manifolds frames, molded elements legs, structural members, joints, shells, shelves lighting industry street lighting light masts, lamp shells optics telescopes, torches housings 9 Chapter 2 Literature Review The purpose of literature review is to identify the gaps in the previous work that provide a direction for our work. The previous work provides us baseline data for the future work. In the following section we will review literature on hydroforming process. 2.1 History of Tube Hydroforming: [4] Tube Hydroforming (THF) has been called with many other names depending on the time and country it was used and investigated. Bulge forming of tubes (BFTs) and liquid bulge forming (LBF) were two earlier terms, for instance. Hydraulic (or hydrostatic) pressure forming (HPF) was another form of name used for a while by some investigators. Internal high pressure forming (IHPF) has been mostly used within German manufacturers and researchers. [4] This process is comparatively new and is used in industry since 1990s but development of the techniques and establishment of the theoretical background goes back to 1940s. Manufacturing of seamless copper fittings with T branches was investigated using internal pressure and axial load by Grey et al. [5] Davis tested tubes of medium carbon steel under internal pressure and tensile axial load in order to determine their yield and fracture characteristics [6]. Experimental and numerical studies were conducted to find the bursting pressure of thick-walled cylinders by Faupel, Crossland and Dietmann during 1950s and 1960s [7-9]. Use of hydrostatic pressure for bulging tubular parts was first reported by Fuches [10]. He reported experimental studies on flanging and expansion of copper tubes using hydrostatic pressure. Al-Qureshi and his team [11] performed bulging and piercing experiments of different materials including copper, steel and aluminum using polyurethane to provide internal pressure. They did not use axial loading in their experiment. 10 During 1970s research on bulge forming continued both experimentally and theoretically by various researchers. New materials, different tooling configurations and new mechanics concepts were introduced but basics remained same. For example instead of using polyurethane Al-Qurashi used rubber for bulging tubes [12]. He presented that greater circumferential expansion of thin-walled tubes was obtained using rubber forming methods than hydraulic technique. Limb and his team [13] performed bulge forming of different materials by changing tube thicknesses. They reported that we can get best results by gradually increasing pressure and application of axial load. In addition, experimentation of different lubricants such as PTFE film, colloidal graphite and Rocol RTD spray were carried out. Limb et al. [14] used oil as pressurizing medium in their experiments to investigate the forming of copper, aluminum, low carbon steel and brass Tee-shaped tubular parts. Results of lubricant and material evaluations were reported in terms of protrusion height attainable. In 1980s research continued on material properties and their effect on bulge forming operations. Manabe and Nishimura [15] investigated influence of the strain-hardening exponent and anisotropy on forming of tubes in hydraulic bulging and nosing processes. They briefly presented the maximum internal pressure as a function of tube radius, thickness, strain hardening exponent, and strength coefficient assuming that there was no axial loading. Fuchizawa [16] analyzed bulge forming of finite-length, thin-walled cylinders under internal pressure using incremental plasticity theory. He presented the influence of strain-hardening exponent on limits of bulge height. 11 2.2 Work on Hydroforming since 1990s: Hydroforming has been used in industry extensively since 1990s. Much research has been done in this field since 90s. S.Thiruvarudchelvan worked on bulging of tubes using urethane rod. In 1990 he published his work on bulging of copper tubes by internal pressurization in a closed die under different conditions of lubrication[17]. The component he formed was bottle shaped. In his experiment he concluded that three modes of deformation are possible using three different conditions of friction. Fig 2.1: schematic diagram of tube bulging using urethane rod [17] 12 Fig 2.2: showing stages of deformation when bulging under least friction [17] Fig2.3: showing stages of deformation when bulging under greatest friction [17] 13 Fig 2.4: showing stages of deformation when bulging under intermediate friction [17] Asnafi worked on analytical modeling of hydroforming [18]. He worked on free forming and constructed its analytical model and used this model to predict forming limits and see the effects of different materials and different loading patterns. He discussed three modes of failure in hydroforming process i.e. buckling, wrinkling and fracture, shown in following figure 2.5. 14 Fig 2.5: failure modes in hydroforming [18] In his paper he discussed buckling and fracture. The risk of buckling is greatest specially in the presence of axial loading. It depends upon the free length, tube diameter and wall thickness. If buckling occurs it is not possible to continue hydroforming process since the process can no longer be controlled. To avoid bucking Asnafi suggested using the guidelines shown in figure 2.6. The limits of different modes of failure are shown in figure 2.7 schematically. 15 Fig 2.6: to avoid buckling these rules should be followed [18] Fig 2.7: the limits and the working range in tube hydroforming [18] 16 The hydroforming operation can be divided into two stages: free forming and calibration. The part of the hydroforming operation, in which the tube expands without tool contact, is called free forming. Calibration starts as soon as tool contact is established, Fig2.8. during calibration the tube is forced to confirm inner corners of tube using internal pressure only. Fig 2.8: the tube hydroforming consists of free forming and calibration [18] Yingyot Aue-U-Lan and his team worked on optimizing tube hydroforming process using simulation and experimental verification [19]. The success of this process is dependent on a number of variables such as loading paths, lubrication conditions and material formability. A suitable combination of all these is required to avoid failure. Depending on the complexity of the part, the THF process window can be very small thus making it 17 difficult to obtain the right loading paths. Hence, it is imperative to establish a systematic way for determining loading paths the finite element analysis (FEA) [19]. The pressure loading path is estimated by using three pressure components namely yield pressure 𝑃𝑦𝑖𝑒𝑙𝑑 , the bursting pressure 𝑃𝑏𝑢𝑟𝑠𝑡𝑖𝑛𝑔 , and the calibration pressure 𝑃𝑐𝑎𝑙𝑖𝑏𝑟𝑎𝑡𝑖𝑜𝑛 , as given in following equations [20]. 𝑃𝑦𝑖𝑒𝑙𝑑 = 𝜎𝑦 2𝑡0 𝐷0 − 𝑡0 𝑃𝑏𝑢𝑟𝑠𝑡𝑖𝑛𝑔 = 𝜎𝑢 𝑃𝑐𝑎𝑙𝑖𝑏𝑟𝑎𝑡𝑖𝑜𝑛 = 2 √3 4𝑡0 𝐷𝑝 − 𝑡0 𝜎𝑓 [𝑙𝑛 𝑟𝑏 ] 𝑟𝑏 − 𝑡 Where 𝑡0 is the initial tube wall thickness, 𝐷0 the tube diameter,𝐷𝑝 the protrusion diameter, 𝑟𝑏 the smallest die cornerradius, 𝜎𝑓 the flow stress of the material, 𝜎𝑦 the yieldstrength of the material and 𝜎𝑢 the ultimate tensile strength of the material. From these three pressure components a linear pressure curve can be constructed. However for complex parts this curve may not agree with the process window. Considerable research has been done to find robust and cost effective techniques to find optimized loading paths. Three techniques have been proposed in literature, self feeding (SF), optimization and adaptive simulation (AS). Yingyot [19] discussed experimental verification of SF and AS approaches. SF method designed to restrict the search for the loading path to a proper family of curves and select the optimum within that family. This method contains two steps. The first step is used to determine the relationship between internal pressure (P) and axial feed (dax), where the process is simulated by imposing only the internal pressure versus time. The friction at the interface is assumed to be zero. Then, the displacement versus time at the 18 node located at the ends of the tube and the maximum thinning on the deforming tube are determined. This information is used to estimate approximately how much the axial feed should be in order to avoid excessive thinning of the hydroformed tube. In the second simulation step, a friction coefficient is prescribed and the axial feed is increased by a certain amount using a scale factor, as shown in Fig. 2.9. This scale factor is varied until a successful part is formed [21] Fig 2.9: SF loading paths: α is a scale factor to increase the amount of axial feeding. [19] The principal idea of AS approach is to feed material into deformation zones as much as possible without any fracture or wrinkles. At the beginning of the simulation, the tube is “deformed” by pressuring to the yield pressure. Then, axial feeding is provided in the simulation, while maintaining the yield pressure, until wrinkles are detected. The wrinkles are then eliminated by pressuring the tube without any axial feeding. Once the wrinkles are eliminated the tube is fed by axial feeding at a constant pressure (see Fig. 2). These steps are repeated until a part without wrinkles and excessive thinning is obtained. 19 Fig 2.10 : schematic procedure of Adaptive simulation [19] J.P. Abrantes, A. Szabo-Ponce and G.F. Batalha simulated two tube hydroforming processes [22]. The first simulation case is the THF of a free aluminum tube without thrust feed force, that does not regard any die reaction forces. The second one is a study of THF calibration into a closed die. The initial pressure required to start free bulging can be found by 𝑃0 = 𝜎𝑦 𝑡0 𝑟 Where 𝑟 is the average radius of the tube. A finite element model for free bulging in LSDYNA was constructed. M. Imaninejad and his team worked on loading path optimization of tube hydroforming process [23]. He proved the importance of optimization of loading paths through his experiment. 20 Fig 2.11: Samples of closed-die hydroformed tubes employing optimized (a) lower pressure optimized loading paths and (b) higher pressure loading paths [23] 21 Chapter 3 Preparation of Experimental Setup This study is based on experimentation and simulation. Hence an experiment was designed for hydroforming process. An experimental setup was designed, analyzed and manufactured for this study. Design, control and maintenance of the tube hydroforming system is of special importance since high hydraulic pressure levels, very large tooling and equipment with high capital cost are involved to ensure continuous mass production of complex shaped parts. The system needed for a typical hydroforming consists of the following [23] • Clamping devices for closing and holding the dies: presses (hydraulic) • Tooling: dies, inserts, etc. • Pressure system; pumps, intensifier, valves, sensors/transducers, controls, • Hydraulic cylinders and punches: for sealing the tube and move the material • Process control systems; computers, data acquisition, transducers, etc. • Hydraulic conditioners: coolers, filters, additives, etc. We decided to use an elastomer such as rubber instead of using oil or water to apply internal pressure. Hence we did not require pumps, seals and valves as they are required when oil, water or any other liquid is used. We manufactured die, die fixture and piston (inserts) to force rubber in die. 22 3.1 Die Design, Analysis and Manufacturing: 3.1.1 Design of Die: The die was modeled in pro engineers. We can hydroform a tube of length 47 mm and of outer diameter 18 mm in it. Fig 3.1 : one half of split die The two parts of die were to be assembled using four M10 bolts and nuts as shown in figure 3.2. 23 Fig 3.2 : assembled die 3.1.2 Analysis of Die: The die was to be analyzed against maximum applied load for safety. In first step the die was analyzed on Ansys against a load of 2 tons. First analysis was done by assuming the die rigid (no bolted connection). The die material was assumed to be structural steel and a pressure of approximately 3000 bars was applied on the internal surface of die. This pressure will be generated by application of 2 ton force. Material properties used for structural steel are shown in following table. 24 Property Value units Density 7850 Kg/m^3 Yield Strength 400 MPa Ultimate Tensile Strength 460 MPa Table 3.1: Properties of Structural Steel In this analysis die was found safe as its safety factor was 2.4 as shown in following figure. Fig3.3: safety factor of die Equivalent von-Mises stresses were also found as shown in following figure 3.4. 25 Fig 3.4: Von-Mise Stresses 3.1.3 Analysis of Bolted Connection:[24] We have analyzed die in Ansys and it was found safe but we also have to analyze the bolted connection in die. The two parts of die will be connected by four M10 nuts and bolts. We have used grade 8.8 bolts and grade 8 nuts. Properties of M10 grade 8.8 bolt according to ISO 898-1 are shown in table and properties of M10 grade 8 nut according to ISO 898-2 are shown in table. Property Value Units Pitch (p) 1.5 mm Stress area (At) 58.8 mm^2 Proof stress (Sp) 580 N/mm^2 26 Proof load (Fp) 34.1 kN Tensile stress (Ft) 800 N/mm^2 Torque (T) 45.8 Nm Hardness 22-32 HRC Elongation 12 % Table 3.2: properties of grade 8.8 M10 bolt Property Value Units Proof stress 880 N/mm^2 Proof load 51.7 kN Hardness 30 max HRC Table 3.3: properties of grade 8 M10 nut. The calculations for bolted connection are following. 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙 𝑙𝑜𝑎𝑑 = 𝑃𝑡 = 2 𝑡𝑜𝑛 = 20𝑘𝑁 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑏𝑜𝑙𝑡𝑠 𝑡𝑜 𝑏𝑒 𝑢𝑠𝑒𝑑 = 4 𝑡𝑒𝑛𝑠𝑖𝑙𝑒 𝑙𝑜𝑎𝑑 𝑝𝑒𝑟 𝑏𝑜𝑙𝑡 = 𝑃 = 5𝑘𝑁 𝑚𝑎𝑗𝑜𝑟 𝑑𝑖𝑎 𝑜𝑓 𝑏𝑜𝑙𝑡 = 𝑑 = 10𝑚𝑚 Properties of M10 bolts and nuts are taken from tables. Material for bolts is carbon steel. 𝑝𝑟𝑒𝑙𝑜𝑎𝑑 = 𝐹𝑖 = 0.75F𝑝 = 25.6𝑘𝑁 𝑔𝑟𝑖𝑝 𝑙𝑒𝑛𝑔𝑡ℎ = 𝑙 = 48𝑚𝑚 𝑡ℎ𝑟𝑒𝑎𝑑𝑒𝑑 𝑙𝑒𝑛𝑔𝑡ℎ = 𝐿𝑡 = 26𝑚𝑚 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑢𝑛𝑡ℎ𝑟𝑒𝑎𝑑𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑖𝑛 𝑔𝑟𝑖𝑝 = 𝑙𝑑 = 37.9𝑚𝑚 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑡ℎ𝑟𝑒𝑎𝑑𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑖𝑛 𝑔𝑟𝑖𝑝 = 𝑙𝑡 = 10.1𝑚𝑚 27 𝑎𝑟𝑒𝑎 𝑜𝑓 𝑡ℎ𝑟𝑒𝑎𝑑𝑒𝑑 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 = 𝐴𝑑 = 78𝑚𝑚2 𝑚𝑜𝑑𝑢𝑙𝑢𝑠 𝑜𝑓 𝑒𝑙𝑎𝑠𝑡𝑖𝑐𝑖𝑡𝑦 𝑜𝑓 𝑏𝑜𝑙𝑡 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 = 𝐸 = 207𝐺𝑝𝑎 𝑠𝑡𝑖𝑓𝑓𝑛𝑒𝑠𝑠 𝑜𝑓 𝑏𝑜𝑙𝑡 = 𝑘𝑏 = 𝐴𝑑 𝐴𝑡 𝐸 𝐴𝑑 𝑙𝑡 + 𝐴𝑡 𝑙𝑑 𝑘𝑏 = 316356𝑘𝑁/𝑚 From table 8.8 of shingley we find for carbon steel. 𝐴 = 0.78715 𝐵 = 0.62873 Using these constants we can find member stiffness km using equation 8-23 of Shingley 𝑏𝑑 𝑘𝑚 = 𝐸𝑑𝐴𝑒𝑥𝑝 ( ) =-𝑘𝑁/𝑚 𝑙 Fraction of external load carried by bolt can be found by 𝑓𝑟𝑎𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙 𝑙𝑜𝑎𝑑 𝑃 𝑐𝑎𝑟𝑟𝑖𝑒𝑑 𝑏𝑦 𝑏𝑜𝑙𝑡 = 𝐶 = 𝑘𝑏 𝑘𝑏 + 𝑘𝑚 𝐶 = 0.145 Using equation 8-24 of Shingley we can find portion of load carried by bolt 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑙𝑜𝑎𝑑 𝑐𝑎𝑟𝑟𝑖𝑒𝑑 𝑏𝑦 𝑏𝑜𝑙𝑡 = 𝐹𝑏 = 𝐶𝑃 + 𝐹𝑖 𝐹𝑏 = 26.3𝑘𝑁 Our joint is a statically loaded tension joint with preload. The yielding factor of safety guarding against the static stress exceeding the proof strength is 𝑛𝑝 = 𝑆𝑝 𝐴𝑡 = 1.3 𝐶𝑃 + 𝐹𝑖 28 Since it is common to load a bolt close to the proof strength, the yielding factor of safety is often not much greater than unity. Another indicator of yielding that is sometimes used is a load factor, which is applied only to the load P as a guard against overloading. From equation 8-29 of Shingley we have 𝑛𝑙 = 𝑆𝑝 𝐴𝑡 − 𝐹𝑖 = 12 𝐶𝑃 It is also essential for a safe joint that the external load be smaller than that needed to cause the joint to separate. If separation does occur, then the entire external load is imposed on the bolt. Using equation 8-30 of Shingley we can find factor of safety against joint separation (n0) 𝑛0 = 𝐹𝑖 =6 𝑃(1 − 𝐶) As all factors of safety were found to be greater than 1, so our bolted joint of die is safe for our applied load. 29 3.1.4 Manufacturing of die: The die was manufactured in university using electric discharge machining and wire cut. A copper electrode was used for EDM. This electrode was made by turning on lathe. The rectangular block for die was cut on wire cut (fig 3.5) and then it was bisected from middle. Then the cavity in die on its each half was made by EDM (fig 3.6) using electrode shown in fig 3.7. Then holes for bolts were drilled on vertical drill machine and holes in one half of die were threaded using tap. Fig 3.5: Wire cut machine 30 Fig 3.6: EDM machine Fig 3.7: Electrode for EDM of die 31 During experimentation we experienced some difficulties due to die, so we had to do some minor modifications in the die. In the above fig we can see that the corners of the contour which was meant to be formed by die are sharp, so during experiments our tubes were cut at this edge. So we had to round off these edges. Hence a radius of 3 mm was made on this edge also by EDM. The die was also polished to reduce friction and aid material flow. (see fig 3.10) Fig 3.8: Finished die with electrode and bolts 32 Fig 3.9: A finished half of die Fig 3.10: Die with polished inner surface and rounded edges 33 3.2 Mechanism for application of pressure: Instead of using a fluid for applying internal pressure, we had decided to use an elastomer for this purpose. Hence we needed some setup to force elastomer in the die and apply required for on it. Elastomer such as polyurethane, rubber etc. are also used for applying internal pressure in hydroforming process. We have used a rubber mandrel for this purpose. To apply load on rubber we decided to use MTS 810 material testing system present in fracture mechanics lab. This machine can apply 10 ton compressive load, where as our maximum possible load is 2 tons. Hence this machine was able to fulfill our requirement easily. Fig 3.11: MTS 810 material testing system. 34 Another alternate mode of application of pressure was the hydraulic press present in machine tool lab of university. This press had a capacity to apply maximum load of approximately 1.2 tons. Fig 3.12: Hydraulic press in machine tools lab 35 3.3 Fixture for die: Our die could not be directly fitted on MTS 810. Hence a fixture was designed to fit die on machine and applying load. 3.3.1 Design of Fixture: The solid model of fixture was made on PTC Pro/ENGINEER Wildfire 5.0 and it was analyzed in Ansys against maximum possible load. Fig 3.13: model of fixture for die in Pro/ENGINEER Wildfire 5.0 The die was to be place in this fixture and four bolts of M16 were used to hold and place die in right position. This fixture was manufactured in three parts i.e. base plate on which die was to be placed, four posts for M16 bolts which were then welded to base plate and a square block 36 which has a female thread to M27 × 3 was to be press fitted on the bottom of base plate. The fixture was to be mounted on MTS 810 through this M27 thread of square block. Fig 3.14: base plate of fixture Fig 3.15: post for M16 bolt and bottom square block. 37 3.3.2 Analysis of fixture: This fixture was then analyzed in Ansys against maximum possible load. The base plate of fixture will be subjected to vertical load which will be applied on die and the bottom square block will be supported on MTS machine. Hence to simulate these conditions in Ansys we have fixed the bottom block and applied a vertical force of 2 ton on base plate. The minimum factor of safety was found to be 4. Hence our fixture is safe. Fig 3.16: Analysis of fixture in Ansys 38 3.3.3 Manufacturing of fixture: The fixture was manufactured at a workshop in Rawalpindi. As described before it was made in three parts. Base plate was made on shaper and then posts for bolts were welded to it. Bottom square block was press fitted to the base plate. At the end whole fixture was powder coated to protect it from rust and other weather effects. Fig 3.17: Die fixture with bolts 39 3.4 Rubber Mandrel: We decided to use an elastomer such as rubber for applying internal pressure in the tube. In general a greater degree of forming is possible while using a hydraulic fluid but many parts can also be formed using elastomer. Use of elastomer has following advantages. • There is no need of an elaborate control system to co-ordinate the axial compression with the hydraulic pressure, the axial compression is generated by the frictional forces between tube and rubber. • The sealing problems associated with a hydraulic fluid are not present in case of use of rubber. • Only simple tooling and press is required in case of rubber where as specialized machinery is required if we use a fluid. • As the possibility of leakage at high pressure is not present so the process is safer. • There is no need for the filling and then removal of oil, or for the cleaning of the bulged tube of oil after forming. Hence the use of rubber is convenient and quick. Fig 3.18: different lengths of rubber used during experimentation. 40 We used a rubber rod of diameter 16 mm for application of internal pressure. Different lengths used during experimentation are shown in following fig 3.16. These rubber mandrels were place inside the tube in die and load was applied on it through an insert or piston so that it expands and applies hydrostatic pressure on tube to form it. 41 3.5 Inserts: To force rubber into the tubes in die and insert was required. As we decided to use MTS 810 hence an insert was designed according to this machine. This insert had a male thread of M27 × 3 at one end so that it could be fitted in machine. The other end was a cylindrical portion of diameter 18 turned according to die, so that it can pressurize rubber in die. It was made from mild steel. Fig 3.19: Insert which was used for pressurizing rubber. Same insert was used on hydraulic press in machine tools lab as most of the experimentation was done on this press. During later stage of experimentation we required an insert of longer cylindrical portion to pressurize rubber, so a second insert was manufactured. This insert was made from aluminum. 42 Fig 3.20: Insert with longer cylindrical portion. 43 Chapter 4 Experimentation The next step after preparation of experimentation setup was to conduct experiments. For this procedure first step was to produce work pieces. 4.1 Preparation of work pieces: For experiment of hydroforming tubes of following dimensions were manufactured. 𝑙𝑒𝑛𝑔𝑡ℎ = 47𝑚𝑚 𝑜𝑢𝑡𝑒𝑟 𝑑𝑖𝑎 = 18𝑚𝑚 𝑖𝑛𝑛𝑒𝑟 𝑑𝑖𝑎 = 17𝑚𝑚 Tubes of following materials were produced. ➢ Aluminum AA1145 ➢ Aluminum AA2024 ➢ Copper ➢ Stainless Steel AISI304 Fig 4.1: Tubes used for hydroforming. Left one is aluminum, middle one is copper and right one is SS 304 44 The selected dimensions of out tubes can be verified according to fig 2.6 [18], (guidelines to avoid buckling). In our case 𝑑0 = 18𝑚𝑚 𝑡0 = 0.5𝑚𝑚 𝑑0 = 36 𝑡0 Therefore free length must be smaller than 2𝑑0which is equal to 36 mm. our free length is 10 mm. Hence this criterion is fulfilled. These dimensions of tubes were chosen according to size of die. A tube in die is shown in following picture. Fig 4.2: SS tube in die 45 4.2 Pressure required to start bulging: The theoretical pressure necessary to start the THF bulging process can be determined by the following expression: 𝑃0 = 𝜎𝑦 𝑡0 𝑟 ( 1) Here for all tubes we have 𝑟 = 8.75 And 𝑡0 = 0.5 𝑚𝑚 Yield stress values for these materials are given in following table Material Yield stress (Mpa) AA1145-O (annealed) 35 AA2024-O (annealed) 75 Copper (annealed) 50 AISI 304 240 Table 4.1: Yield Stresses of materials Now using equation 1 we can find the pressure required to start bulging process. These starting pressures are listed in table. Material Minimum pressure required to start bulging (bars) AA1145-O (annealed) 20 AA2024-O (annealed) 43 Copper 28.5 AISI 304 137 Table 4.2: Minimum pressure required to start bulging 46 The area on which we have to apply load to generate pressure is of 17 mm diameter (inner diameter of tube). So we can calculate our required load. We know that 𝐹 = 𝑃𝐴𝑖 (2) Where F is the force, P is pressure and 𝐴𝑖 is area on which force is applied (area of inner dia of tube). The area which we will use is the perpendicular area of inner dia (𝑑𝑖 ) of tube. So we have 𝑑𝑖 = 17𝑚𝑚 𝐴𝑖 = 𝜋 2 𝑑 =-𝑚2 4 𝑖 Now using equation 2 we can calculate minimum load required to start bulging. These loads are listed in following table Material Minimum load required to start bulging tonnes AA1145-O (annealed) 0.046 AA2024-O (annealed) 0.1 Copper 0.065 AISI 304 0.31 Table 4.3: Minimum load required to start bulging These are the loads which are required to start bulging. Calibration and complete hydroforming will require much greater load. We decided to use hydraulic press of 1.2 tonne capacity for experimentation. 47 4.3 Initial Experiments: Initially at start we first conducted experiments with aluminum AA2024 and copper tubes. First experiment was done using 2024 tube. The tube was place in die. Then die was closed and bolts were tightened. The rubber rod was inserted in the die and tube. Then steel insert was used to apply load on rubber in the die. This process was done on hydraulic press in machine tool lab. The rubber mandrel used and the piston used to force it into the tube are shown in fig. When the load was applied aluminum tube failed. It experienced fracture very early. Fig 4.3: rubber mandrel and insert used to force it in die. Then we decided to use annealed copper tubes as copper is more ductile. Several experiments were done but copper tubes also failed. The failed tubes are shown in following picture. These tubes not only fractured but also got cut at the base of forming couture. 48 Fig4.4: Failed copper and aluminum tubes. Same experiment was also done with SS304 tubes. These tubes also failed. They also not only fractured but were cut at the edge of die. Fig 4.5: Failed SS tube. Due to this we concluded that one problem is the edges in the die. These were causing the tubes to cut. 49 Sharp edges Fig 4.6: tubes were cut due to sharp edges of die. Tube is cut due to sharp edges 50 4.4 Modifications with Die: To solve this problem we removed sharp edges from die by making radiuses of 3 mm by EDM. To reduce friction between die and tube we also polished the die. Radiuses were drawn on edges Fig 4.7: Die with radius on edges and polished surface. 51 4.5 Experimentation after Die Modification: After modification of die we continued our experimentation with AISI 304 tubes as they were found strongest. Actually the press which we were using was not very precise. MTS 810 was precise but it had the capacity of 10 ton load and we required approximately 1 ton load. So if accidently more load was applied bolted joint of our die could fail. So this was not safe. Hence we decided to use hydraulic press of approximately 1.2 ton capacity as it was simply not able to provide too much load. Copper and aluminum tubes required much less load for hydroforming, and our press was not very precise to be restricted to these small loads such as 0.4 ton. Every time the load exceeds this limit, so we decided to use AISI 304 tubes for further experimentation as it can withstand greater loads. Now after these modifications with die these SS tubes were not cut due to sharp edges of die, however they were again fractured. Fig 4.8: failed AISI 304 tubes 52 4.6 Experimentation using Play-Doh: Then we used another material, play-doh instead of rubber to internally pressurize tubes. Fig 4.9: Play-Doh Play-Doh is a modeling compound used by young children for art and craft projects at home and in school. It is composed of flour, water, salt, boric acid, and mineral oil. We tested AISI 304 and AA 1145 tubes using play-don for pressurization. These tubes also fractured. In case of play doh we can see how much the tube is formed before facture,(figures 4.10 & 4.11) because when fracture occurs play doh will leak and no more internal pressure will be built. So the tube will not form anymore. Whereas in case of rubber, it cannot leak and it will keep on expanding when load is applied on it. So the tube will keep on forming in shape of die even after fracture. Play doh was also used for forming of AA 1145. We can see how little forming was done before fracture in this case. 53 Fig 4.10: fractured AISI 304 tube. Play-Doh was used for forming in this case. Fig 4.11: fractured AA 1145 tube. 54 4.7 Experimentation using rubber mandrels of multiple lengths: In our experiments we examined that the length of tubes was almost remaining constant, whereas in order for successful forming material must flow into the die cavity and so the length of tube must decrease. As the material was not flowing into the die cavity so necking was occurring and subsequently tubes were fracturing. After some examination of die and our test samples we concluded that this was happening because the expanding rubber holds the tube at both ends tightly against die not allowing any material to flow. To solve this problem we decided to use rubber mandrels of smaller lengths, so that rubber will only be in the portion which is to be hydroformed and one end of tube will not have rubber in it. So material on this end will flow into the die cavity. In this case our previous insert could not be used because it will not reach rubber. So a new insert was manufactured from an aluminum rod. The new insert and rubber lengths used are shown in fig 4.12 Fig 4.12: rubber lengths used and new insert. 55 First the smallest length of rubber was put in die and load was applied on it. Due to which tube was formed to some extent. Then die was opened and rubber of greater length was put in tube and again load was applied. So the tube formed more. Then again this rubber was removed and an even greater length was used. This time the tube completely formed. The length of the tube was also decreased as material flew into die cavity. Fig 4.13: Successfully hydroformed tube 56 The decrease in length is shown in fig 4.14. One hydroformed tube was sliced on wire cut and then the dimensions of hydroformed portion were measured. These measured dimensions are shown in fig 4.15. 44 mm Fig 4.14: The length of hydroformed tube was decreased by 3 mm 47 mm 57 Fig 4.15: Dimensions of hydroformed portion. 58 Chapter 5 Simulation For simulation of hydroforming process we used finite element software ABAQUS. This analysis was done in two dimensions. In solving our problem solution mapping capabilities of ABAQUS will be used. When the strains become large in a geometrically nonlinear analysis, the elements often become so severely distorted that they no longer provide a good discretization of the problem. When this occurs, it is necessary to map the solution onto a new mesh that is better designed to continue the analysis. 5.1 Modeling of problem: The first step in simulation is to model the problem. We have to model die, tube and hydraulic pressure. Tube is modeled as an axisymmetric deformable shell. As we have modeled this problem in two dimensions, so to model tube its rectangular cross section which is a rectangle of width 0.5 mm and length 47 mm and at a distance of 8.5 mm from central axis was drawn. This is shown in figure. The element type CAX4R is used. This is a 4-node quadrilateral element. This element type is chosen because this element is good for problem involving non linear constitutive behavior. Die is modeled as an axisymmetric analytical rigid wire. This means that it is not deformable and it will act as a rigid surface. A material was then defined and assigned to the tube. This is explained in next section. An assembly of die and tube is defined. This is shown in fig. 59 Fig 5.1: models of tube and die in ABAQUS Fig 5.2: tube and die assembly in Abaqus 60 5.2 Material Definition: We needed to define material of tube for our analysis. For our analysis we had to define both elastic and plastic properties of the material. In elastic properties Abaqus requires modulus of elasticity and poison’s ratio. Where as in plastic properties it requires values from the true stress strain curve of the material. The true stress strain curve of AISI 304 was obtained from ASM Atlas of Stress Strain Curves. Fig 5.3: 304 stainless steel bar, true stress strain curve at room and elevated temperatures [25] 61 The values which were used in Abaqus are listed in following table. Stress (Pa) True Strain 1 - 0 2 - 0.0015 3 - 0.003 4 - 0.004 5 - 0.012 6 - 0.1 7 - 0.8 Table 5.1: Values from true stress strain curve of SS304 The elastic modulus was taken 210 GPa and poison’s ratio is 0.3. 5.3 Interaction between Tube and Die: After the creation of assembly of tube and die, the next step is to define the interaction between surfaces of tube and die. For this first we need to define surfaces which will interact with each other. In our case the inner surface of die and outer surface of tube will interact with each other. After defining these surfaces the interaction between them is defined. These surfaces will interact with each other both normally and tangentially. Normally a hard contact between these two surfaces was defined. On the other hand tangentially two types of behaviors were used. In one analysis tangentially frictionless behavior was defined whereas in the other analysis a friction coefficient of 0.2 was defined between these two surfaces. 5.4 Boundary Conditions and Loading: The next step is to apply boundary conditions, constraints and loads. First we constrained our die as a rigid analytical surface with its reference point. This means that position of die is fixed with respect to this point in the analysis and it will not change. In our experiment we observed that one end of tube in which rubber is present is held tightly against die when load is applied 62 on rubber so we fixed that end with die. The other end of tube can move in y direction so we allowed that end to move in y direction and restricted its other movements. Fixed end of tube Fixed die Restricted end of tube. It can move in y direction Fig 5.4: Boundary conditions. 63 To simulate the pressure generated by expansion of rubber we applied a pressure load of 50 MPa on the inner surface of tube. 500 bar pressure is applied on the inner surface of tube Fig 5.5: Pressure application 64 5.5 Analysis and Results: After apply all conditions and meshing this job was submitted for analysis. When analysis was done results were examined in the result window of Abaqus. Fig 5.6: Result window of Abaqus Different dimensions of hydroformed tube were measured. These dimensions are shown in following figure. 65 Fig 5.7: Different dimensions of hydroformed tube By measuring displacement of tube in y direction we can measure how much the length of tube is reduced. By this method we found out that the length of tube is reduced by 2.2 mm. 66 47 mm Fig 5.8: lengths of tube before and after hydroforming 44.8 mm 67 5.6 Analysis using Aluminum: Same problem was also analyzed using aluminum 1100 as material. The modulus of elasticity for AA1100 was taken 70 GPa and Poison’s ratio 0.3. The plastic properties were taken from Atlas of Stress Strain Curves. The values from true stress strain curve of fig 5.8 were put in Abaqus. Fig 5.9: 1100-0 stress strain curves [27] 68 In this case we examined that the simulation was aborted after some time. The reason for this was that even after maximum deformation the tube did not made contact with the die surface in die cavity. The strain value was greater than the maximum true stain which we input in Abaqus from true stress-strain curve. Fig 5.10: tube is not fully deformed in case of AA1100 69 Chapter 6 Comparison of Results: The results of our simulation were compared with the experimental values. To measure values from simulation we measured nodes displacements form Abaqus and then modeled the deformed tube in Pro Engineers, so that we can measure these values easily. The results of our simulation were compared with the experimental values and the results were found very close. Fig 6.1: Comparison of various dimensions of hydroformed tube with Abaqus simulation The length of hydroformed tube predicted in simulation was very close to experimental value. 70 Fig 6.2: Simulated lengths 71 Fig 6.3: Experimentally measured lengths 72 Chapter 7 Conclusion Tube hydroforming process is one of the most effective forming processes and is now in use in industry for recent twenty years. It is regarded as chip less or non-material removal process. But many of the parameters involved in this process are still not known. Many theories have been established to discuss various failure modes such as buckling, fracture and wrinkling in this process and forming limits. Analytical modeling for this process is very difficult. The finite element simulation of this process is the most effective and accurate virtual simulation method in tube hydroforming process. It also takes the least computational time. The computational time depends upon various factor such as mesh size, factors considered etc. If we consider too many factors and choose a small mesh size the computational time will be greater but our results will be more accurate. Whereas if we choose large mesh size and consider less parameters the computational time will be less but the results will be less accurate. So there is compromise between accuracy and computational time. During our experimentation first we concluded that hydroforming cannot be done if die has sharp corners. For this process there must be radiuses on all edges. Secondly we must make sure that the material is flowing in the die. If there is some hindrance in material flow fracture will occur. Material formability is a very important factor in this process. It the material does not have enough formability hydroforming will not occur. By using elastomer for generating hydrostatic pressure in the tube we can avoid certain fluid relating problems such as sealing problems. However rubber has a limit and more complex shapes cannot be formed using rubber. In our research we constructed a simulation of tube hydroforming process in Abaqus. The conditions of experiment were modeled in Abaqus and boundary conditions and surface interactions were defined. Then compared its results with actual experimental results. A reasonable agreement between these two results was found. The simulation done with SS 304 73 showed complete hydroforming. The tube was fully deformed against die. Same was observed in experiment. Whereas in case of aluminum it showed little forming. In experimentation it also fractured after very little forming. So this model can be used to simulate tube hydroforming process before expensive and time consuming manufacturing of dies and preparation of other equipment for this process. It will help in optimizing die design according to shape of our required component. 74 References [1] Schuler Metal Forming Handbook, ISBN- [2] Muammer Koç, Hydroforming for advanced manufacturing, p-6 [3] Muammer Koç, Hydroforming for advanced manufacturing, p-2 [4] MuammerKoc, TaylanAltan (2001), An overall review of the tube hydroforming (THF) technology, Journal of Materials Processing Technology [5] J.E. Grey, A.P. Devereaux, W.N. Parker, Apparatus for makingwrought metal T's, US Patent 2,203,868 June (1939) [6] E.A. Davis, Yield and fracture of medium-carbon steel under combined stress, J. Appl. Mech. (March 1945) [7]J.H. Faupel, Yield and bursting characteristics of heavy wallcylinders, Trans. ASME ( July 1956 [8]H. Dietmann, The flow behavior of thick-walled cylinders underinternal pressure, BanderBlecheRohre 8 (3) (1967) [9]B. Crossland, S.M. Jorgensen, J.A. Bones, The strength of thick-walledcylinders, ASME J. Eng. Ind. (May 1959) [10]F.J. Fuchs, Hydrostatic pressure - its role in metal forming, Mech.Eng. (April 1966) [11] H.A. Al-Qureshi, P.B. Mellor, S. Garber, Application of polyurethaneto the bulging and piercing of thin-walled tubes, in: Proceedings ofthe Ninth International MTDR Conference, Birmingham, UK, September 1968 [12] H.A. Al-Qureshi, Comparison between the bulging of thin-walledtubes using rubber forming technique and hydraulic forming process,Sheet Metal Ind. (July 1970) 75 [13] M.E. Limb, J. Chakrabarty, S. Garber, P.B. Mellor, The forming ofaxisymmetric and asymmetric components from tube, in: Proceedingsof the 14th International MTDR Conference, 1973 [14] M.E. Limb, J. Chakrabarty, S. Garber, W.T. Roberts, Hydraulicforming of tubes, Sheet Metal Ind. (November 1976) [15] K. Manabe, H. Nishimura, Influence of material properties informing of tubes, BanderBlecheRohre 9 (1983) [16] S. Fuchizawa, Influence of strain hardening exponent on thedeformation of thin-walled tube of finite length subjected tohydrostatic external pressure, Adv. Technol. Plasticity 1 (1984) [17]S.Thiruvarudchelvan and F.W. Travis, Tube bulging with a urethane rod, journal Material Processing Technology, 23 (1990) [18]Nader Asnafi, Analytical modelling of tube hydroforming, Thin-Walled Structures 34 (1999) [19] YingyotAue-U-Lan, Gracious Ngaile, TaylanAltan, Optimizing tube hydroforming using processsimulation and experimental verification, Journal of Materials Processing Technology 146 (2004) [20]T. Altan, S. Jirathearanat, M. Strano, Adaptive FEM process simulationfor hydroforming tubes, in: Proceedings of the InternationalConference on Hydroforming, Fellbach-Stuttgart, Germany, November2001. [21] M. Strano, S. Jirathearanat, Adaptive simulation concept for tube hydroforming,in: Proceedings of the Innovation in Tube HydroformingTechnology, Troy, June 2000. [22] J.P. Abrantes, A. Szabo-Ponce, G.F. Batalha, Experimental and numerical simulation of tube hydroforming (THF), Journal of Materials Processing Technology 164–165 (2005) [23] MuammerKoç, Hydroforming for advanced manufacturing, p-10 76 [24]Shingley’s Mechanical Engineering Design, 9th edition, Chapter 8 [25] Atlas of Stress Strain Curves, ASM International, 2nd edition, p-185, SS046 [26] Joakim Lundqvist, Numerical Simulation of Tube Hydroforming Adaptive Loading Paths, Licentiate Thesis 2004:26 [27] Atlas of Stress Strain Curves, ASM International, 2nd edition, p-301, WA.005