Abstract
The quality of surface morphology can reflect the electrical performance of silver-based contacts. Existing research on the correlation of morphological-electrical performance is based solely on empirical models from traditional visual inspections and only considers the impact of visually observable macro-textural features on electrical performance. However, the influence of micro-textural features on electrical performance should not be overlooked. This paper establishes a contactless surface morphology acquisition device based on a laser profilometer to address the assembly characteristics of contact components. Various original profile signals such as surface roughness, waviness, and surface form error are calculated using wavelet transformation, and a robust weight function is introduced to separate micro-textural features from macro-textural features. After the morphological parameters affecting electrical performance are quantified, the variation laws of single and composite morphological parameters on electrical performance are clarified. Parameter optimization iterations and parameter space distribution optimization are performed using a decision tree, and the optimized predictive model forecasts specific electrical parameter values. The predicted results are quantitatively evaluated, establishing evaluation metrics that reflect the errors and degree of fit between the model predictions and actual values from different perspectives. From the experimental results, the accuracy of the predictive model established in this study exceeds 97%.