Surface roughness is a critical factor for the operational efficacy and lifespan of the manufactured components, often serving as a key metric for evaluating manufacturing processes and determining acceptance of a component. Traditional surface roughness measurements are highly sensitive to the fixed standard cut-off length, influencing the scale at which surface texture features are analyzed. Thus, developing techniques with more reliable filtering of surface features is essential for accurate surface metrology. This research proposes a novel method that combines Empirical Mode Decomposition (EMD) and Fast Fourier Transform (FFT) for surface roughness characterization. We applied our method to Nitinol shape memory alloy (equal atomic proportions of nickel and titanium) surfaces processed using micro-Wire Electrical Discharge Machining (μ-WEDM). Our results demonstrate that the EMD-FFT method significantly outperforms traditional filtering, especially for surfaces without distinct patterns like μ-WEDM. This improvement is particularly valuable for materials like Nitinol, which is widely used in critical applications where precise surface roughness control is essential. By accurately assessing the effects of μ-WEDM parameters on surface roughness, the proposed method reveals that for a constant pulse on time and a constant discharge current, the variation of pulse off time is more important than servo voltage. Thus, this method can contribute to optimizing manufacturing processes and producing higher-quality components with improved performance and durability.
Read full abstract