Advanced mechanical FEM analyses of polymer gears:

Polymer gears are increasingly becoming a viable alternative to conventional metal gears in many industrial applications ranging from automotive to household appliances due to substantial improvements in the quality and durability of polymers. These types of gears offer a much cheaper solution as compared to metal gears, especially in serial production, along with many other additional benefits. In general they are however still comparably less reliable than metal gear due to their non-elastic properties, lower stiffness and lower heat resistance. Especially the temperature state of thermoplastic polymers plays a major role in their response to cyclic mechanical loads and the overall service life of thermoplastic components. An increased temperature at the gear contact interface can result in accelerated wear and even melting in the case of excessive loads. The temperature increase in polymers due to mechanical loads can occur as a result of self-heating due to non-elastic hysteresis losses in the material structure.

Company Dr. Duhovnik d.o.o. has lot of experience with numerical modelling and simulations of the polymer gears. But since they don’t have access to the HPC their simulations are simplified so they could be performed on common workstations. It was decided that proposal describing technical aspects and actions for FEM simulation of polymer gears will be prepared and computational resources will be requested by Director Discretion scheme at IT4Innovations. After the initial contact at Workshop in Ljubljana several e-mails between mr. Černe and Tomas Karasek were exchanged. Tomas Karasek and Mr. Černe from Dr. Duhovnik d.o.o. prepared applications for Director Discretion scheme which was submitted by Tomas Karasek. After approval access for Mr. Černe to IT4Innovations infrastructure was granted.

For pilot project “Advanced mechanical FEM analyses of polymer gears” by Dr. Duhovnik d.o.o. computational resources from IT4Innovations through Director Discretion scheme were provided as a in-kind contribution to the InnoHPC project. In this case communication will proceed and established collaboration between SMEs and HPC center will continue even after the InnoHPC project end. In case of “Advanced mechanical FEM analyses of polymer gears” further simulations will be carried out.

Main activities include development of accurate and computationally efficient models for flow friction in gas pipes, and also development and optimization of pipelines systems for distribution of water, natural gas and hot water for heating purposes. The results can have applications also in ventilation networks. Our recent efforts to solve the certain problems with flow frictions have already been affiliated with the IT4Inovations Czech’s National Supercomputing Centre from Ostrava.

For this research we used different computation tools, such as Padé approximations, switching functions of various types, Symbolic Regression analysis using different software tools such as Eureqa, or HeuristicLab, as well iterative methods recently developed in mathematics. For example, novel three-point methods were successfully tested for the purpose of efficient solution of flow friction problems. For all mentioned tasks, further extensive computational tests should be performed, especially when uncertainty of input parameters is introduced. Fortunately, a parallel implementation of a Monte-Carlo simulation technique is straightforward. Consequently, the proposed algorithms can exploit the special infrastructure of the IT4Inovations Supercomputing Centre in Ostrava.

Further results are expanded in simulations of flow friction using special functions such as the Lambert W-function and the Wright Ω-function. Moreover, an optimization of the corresponding fast rational approximations using symbolic regression techniques requires a supercomputing power, too. All presented finding in the next step should be applied on a complex network of pipes for distribution of water and gas and for heating purposes. Such findings can be useful in order to reduce losses in the systems.