Prediction of Wear Rate in Al/SiC Metal Matrix Composites Using a Neurosymbolic Artificial Intelligence (NSAI)-Based Algorithm

Abstract

This research paper delves into an innovative utilization of neurosymbolic programming for forecasting wear rates in aluminum-silicon carbide (Al/SiC) metal matrix composites (MMCs). The study scrutinizes compositional transformations in MMCs with various weight percentages of SiC (0%, 3%, and 5%), employing comprehensive spectroscopic analysis. The effect of SiC integration on the compositional distribution and ratio of elements within the composite is meticulously examined. In a novel move for this field of research, the study introduces and applies neurosymbolic programming as a novel computational modeling approach. The performance of this cutting-edge methodology is compared to a traditional simple artificial neural network (ANN). The neurosymbolic algorithm exhibits superior performance, providing lower mean squared error (MSE) values and higher R-squared (R2) values across both training and validation datasets. This highlights its potential for delivering more precise and resilient predictions, marking a significant development in the field. Despite the promising results, the study recognizes that the performance of the model might vary based on specific characteristics of the composite material and operational conditions. Thus, it encourages future studies to authenticate and expand these innovative findings across a wider spectrum of materials and conditions. This research represents a substantial advancement towards a more profound understanding of wear rates in Al/SiC MMCs and emphasizes the potential of the novel neurosymbolic programming in predictive modeling of complex material systems.