Trochoidal Milling and Neural Networks Simulation of Magnesium Alloys

This paper set out to investigate the effect of cutting speed v(c) and trochoidal step s(tr) modification on selected machinability parameters (the cutting force components and vibration). In addition, for a more detailed analysis, selected surface roughness parameters were investigated. The research was carried out for two grades of magnesium alloys—AZ91D and AZ31—and aimed to determine stable machining parameters and to investigate the dynamics of the milling process, i.e., the resulting change in the cutting force components and in vibration. The tests were performed for the specified range of cutting parameters: v(c) = 400–1200 m/min and s(tr) = 5–30%. The results demonstrate a significant effect of cutting data modification on the parameter under scrutiny—the increase in v(c) resulted in the reduction of the cutting force components and the displacement and level of vibration recorded in tests. Selected cutting parameters were modelled by means of Statistica Artificial Neural Networks (Radial Basis Function and Multilayered Perceptron), which, furthermore, confirmed the suitability of neural networks as a tool for prediction of the cutting force and vibration in milling of magnesium alloys.