Material Characterization and Blade Impact Simulation

Blades used on brushcutters and lawn mowers are subjected to a wide variety of working conditions. Besides continuous loads from cutting grass, the blades are also subjected to accidental impacts of branches, stones and structures. Due to exceptionally high rotational velocities, these types of impacts involve blade deformation at high strain rates. This master’s thesis aims to improve understanding and predictability of blade properties for design of future blades. The project is aimed at characterization of the mechanical response of steel used for brushcutter blades and developing a simulation model of a blade impact load case. Thus, the problem was divided into two main parts: firstly, material characterization, and secondly, numerical modeling. The objective of the material characterization part was to determine the rate dependence of the flow stress for two hardened steels. Experimental compression tests were performed at quasi-static strain rates (10-4 - 10-2 s-1) and at high strain rates (102 - 104 s-1) in order to characterize the rate dependence of each material. The objective of the numerical modeling part was to develop simulation models of an impact load case for the purpose of recreating tests performed with an experimental test setup. The simulation models were aimed to include material models for the blade based on the experimental tests performed for the two hardened steels. In preparation for the compression tests, cylindrical specimens were acquired through electrical discharge machining involving material removal from blades intended for brushcutters. Compression tests at high strain rates were performed utilizing a split-Hopkinson pressure bar apparatus which resulted in strain rates in the order of 1000 s-1 and 3000 s-1. Compression tests at quasi-static strain rates were performed with an electro-mechanical loading machine and implementation of two-dimensional digital image correlation for strain measurements. With this method, strain rates in the order of 5 · 10-2 s-1 and 5 · 10-4 s-1 were achieved. The acquired results from the experimental tests included the response of the two materials at four different strain rates in the form of true stress-true strain curves. The results were indicative of small strain rate dependency for each of the two hardened steels with a slight increase in yield stress for increasing strain rates. Both materials exhibited closely similar characteristics. At quasi-static rates, the response of both materials exhibited work-hardening of closely similar characteristics. At high strain rates, the response of both materials exhibited a close to identical decrease in stress for values of strain exceeding 10 %. This behavior was suggested to be a consequence of adiabatic heating. At all four achieved strain rates, the results were indicative of a higher yield stress with higher subsequent stresses for one of the hardened steels in comparison to the other. The impact load case aimed to be simulated involved one swing of a brushcutter against a 25 mm diameter steel rod according to standard SS-EN ISO 11806-1:2011. The steel rod was specified to be impacted horizontally by the blade at an approaching translational velocity of 1 m/s and a blade rotational velocity of 8500 rpm. The multi-physics simulation software LS-DYNA was used to develop simulation models which consisted of two main parts, the blade and the rod and included two different blade geometries. As a result of a study regarding the suitability of different discretization techniques, the decision was made to implement the mesh-free particle method Smoothed Particle Galerkin (SPG) and to perform coupling with the finite element method (FEM). Two material models were developed based on the measured stress-strain response obtained through high strain rate compression testing. Several numerical models of the impact load case were produced, all of which entailed different sets of parameters. These included selection of blade material, failure strain, rod length and blade angle relative to the horizontal plane. Finally, two models were developed which were opposite in terms of assigned element formulation for the blade tip and the rod and otherwise identical. The results of the different models were then compared, namely in terms of resulting material failure of the blade after impact. It was concluded that SPG was the most suitable method of choice for the impact load case aimed to be simulated due to its ability to handle large deformation and the inclusion of the a bond-based failure mechanism. Furthermore, implementation of the SPG method resulted in deformation and failure considered to be of greater agreement to experimental test results compared to FEM