Experimental work

Equipment. The experimental work was done at MEFOS in Luleil. The hydraulic 5 MN pull-down forging press de- signed by Pahnke was used. The stroke speed maximum was 110 mm/s. The tools used in the forging press were made of Nimonic. They were 250 mm wide with a radius of 25 mm. The press is equipped with a manipulator designed for a maximum ingot weight of 500 kg and a deflection torque of 10kNm. An integrated control system is used for the manipulator and the forging press. For heating, MEFOS propane fire chamber furnace was used. The outline and nomenclature used is shown in figure  1.

Procedure. In the first series of experiment rolled 150 mm square blooms of carbon steel were used. Holes 300 mm long with a diameter of eight mm were drilled in the axis of the specimens. The holes were sealed with steel bolts and welded. The specimens were heated to 1100 °C. The specimens were transported to the forging press with a fork lift equipped with tongs. The specimens were forged one pass with different bites of 30, 60, 90, 120 and 150 mm. The set point height was 120 mm. The specimens were forged one pass and then cut at the edge and centre of the tool marks. The widths and heights of the holes were meas- ured by vernier calipers. The lengths of the holes are meas- ured from the cut under the edge of the tool to closed part under the centre of the tool. In those cases where the hole was not closed under the centre of the tool, the length was considered to be half the bite. The cold dimension of the blooms after heating and forging were measured at the part kept by the manipulator tongs to 148.5 mm. The height af- ter forging measured on cold blooms was 116 mm corre- sponding to a height reduction of 22 9'o.

Finite element simulations

Finite element code. The finite element code used for the simulations was the implicit MSC.MARC2003. The ele— ment type used was a three-dimensional arbitrarily distort—

Figure 1. Outline of bloom at forging.

ed brick. This element (type 7 in MARC) uses tri-linear in- terpolation and therefore the strain tends to be constant throughout the element, which can result in a poor repre- sentation of the shear behaviour. This element is still pre- ferred over higher order element in contact analysis. The as- sociated heat-transfer element (type 43 in MARC), also uses tri-linear interpolation and therefore the thermal gradi- ents tend to be constant throughout the element. Eight-point Gaussian integration is used to form both the stiffness and the conductivity in the elements.

Geometry. The geometry was modelled fully three-di- mensional with hexahedral eight node elements. Because of the two symmetry planes, only one quarter of the cross sec- tion was modelled. The mesh was made very fine in the centre around the hole, much coarser inside the bloom and finer in the contact surface. The tools were considered to be rigid and were modelled with surfaces. The motion of the forging press used in the experiments was approximated with a sine function.

Thermal simulation. The  billet  was  first  heated to 1100 °C, and then it cooled down while it was mounted in the manipulator and also during the scale breaking by  wa-

Table 1. Material constants [30].

98.2 12.231 0.00250 0.0494 0. 174 0.0726 0. 139

Table 2. Elastic and thermal material properties.

ter. This gave a significant temperature gradient before the forging began. First the billet was cooled due to radiation and free convection to the air for 30 s, then descaled by wa- ter for 2 s, and finally cooled in air again for 120 s. The cor- responding  connective  boundary  condition is: