Extrusion International 2-2023

41 Extrusion International 2/2023 ed to the biaxial orientation state and the crystallisation state of the material. For the orientation of thermoplastics, they must be stretched above their softening temperature (re- tardation). In blown film extrusion, this occurs in the direction of extrusion if the take-off speed of the film above the frost line is great- er than the exit speed of the melt from the die (take up ratio > 1). Similarly, orienta- tion of the film in the circumferential direction occurs when the blow-up ratio is greater than one (blow-up ratio > 1), which is given by nozzle and bubble diam- eter. The faster the material is solidified, the lower the orientation retardation and therefore, the lower the decay of molecular orientation. It can thus be stated that the interplay of the cooling process and the course of longitudinal and transverse stretching determines the biaxial orientation state of the film. If the orienta- tion behaviour of different materials is investigated, a qualitative correlation can be found: Materials with a lower molecular weight relax faster. In general, a high- er degree of orientation of the macromolecules leads to an increase in the mechanical properties in the di - rection of orientation. In addition to the formation of orientations, crystal- lisation can exert an influence on the mechanical prop- erties. The more the film is stretched in the tube forma- tion zone, the more spherulite growth is inhibited and existing superstructures are destroyed. The result is a hidden structure with a high degree of crystallisation. The faster the melt is cooled, the finer the resulting mi- crostructure. To what extent thermoplastics crystallise at all and to what degree depends crucially on the struc- ture of the macromolecule. Semi-crystalline thermoplas- tics with a linear chain structure, e.g. PE-HD, can reach a degree of crystallisation of up to 70%, while branched chain structures (e.g. PE-LD) are not capable of a high packing density. However, a generally valid statement between the degree of crystallisation or the structure that forms and the resulting mechanical properties can- not be made [MHMS14]. Modelling for the prediction of mechanical properties in blown film extrusion Analogous to other technical processes, there are dif- ferent possibilities for choosing the input and output variables a model for predicting the mechanical film properties. A distinction must be made between ma - chine, process and quality parameters. If the machine parameters are used as input parameters of the model, the model might not be transferable to other machines. Similarly, a process model reflects the relationship be- tween process and quality parameters. Both physical and statistical modelling can be applied for both vari- ants. Furthermore model building methods, e.g. "grey box models”, allow a combination. Picture 1 shows the possibilities of model building for the calculation of me- chanical properties. The first possibility consists in the purely physical modelling, which represents the ideal solution for the prediction of the mechanical properties with sufficient model quality. Hauk [Hau99] uses two-stage modelling to predict tensile strength. He succeeded in the first partial step, the physical calculation of the process state from the machine parameters, however, multiscale simulations are known which can derive the film prop- erties physically conclusively from the process state, but currently they have unfeasibly high computational de- mands [Hau99]. As already explained, the mechanical properties depend significantly on the material- and process-dependent degree of orientation, crystallisa- tion and their interaction, for which physically reasoned correlations, would have to be available. Due to the above-mentioned considerations, the second possibil- ity, the physical modelling with selection of the process parameters as input variables, is not yet feasible today. A widely used method to predict film properties is the statistical modelling of the quality variables starting from the machine parameters. Since no process knowl - edge is developed about the behaviour of the melt in the tube formation zone, this model cannot be trans- ferred to other machines. Furthermore, not all influenc- es, such as material, machine or tool behaviour, are suf- ficiently captured and modelled. A statistical mapping from process to quality variables, which are valid for all processes and plants, allows the transfer of the model to other production plants if the modelling quality is suffi- cient. The disadvantage is also the experimental effort required to parameterise the model. However, since the modelling approach should be able to determine the mechanics based on the process state independently of the machine, the approach according to Ohlendorf is presented in the following and potential for improve- ment is shown. Picture 1: Possibilities of modelling for the prediction of film properties [Ohl04]