Data-driven efficient and intelligent prediction of mechanical properties for low-carbon basalt fiber composites in automotive applications

This article aims to use data-driven methods to study the changes in mechanical properties of basalt fiber (BF) reinforced polylactic acid (PLA) composite materials, and establish a high-precision and scalable performance prediction and low-carbon material design method to promote the application of such green composite materials in lightweight and low-carbon components of automobiles, such as front bumpers. By preparing samples through injection molding process and systematically changing the diameter, length, and mass fraction of BF, it was found that the mechanical properties showed a quadratic trend of increasing first and then decreasing with each parameter. The study is based on the concept of total quantity theory, with fiber parameters and their interaction terms as independent variables. Firstly, Kruskal Wallis test is used to screen for significant factors, and then a multiple superposition regression model is established to predict the tensile, bending, and impact strength of materials. After optimizing using the least squares method, the average error value of tensile strength was found to be 4.29%, the average error value of bending strength was 4.41%, and the average error value of impact strength was 4.72%. The comparison with the multiple linear regression model validates the superiority of the improved model in handling nonlinear data. The established model was applied to the low-carbon design and lightweight development of the front bumper of a certain vehicle model. Through model cycle prediction and quasi conjugate gradient method acceleration screening, the BF/PLA material ratio that meets the target performance was obtained. Finite element analysis and validation of actual vehicle components show that compared with raw materials, the newly designed bumper structure has improved static performance by 28%, frontal low-speed collision performance by 14.3%, weight reduction by 9.7%, and is expected to reduce carbon emissions throughout its entire lifecycle. The results indicate that the proposed data-driven prediction and low-carbon design method based on an improved multiple superposition regression model can guide the formulation design and performance optimization of BF/PLA composite materials, improving their mechanical properties and environmental benefits while achieving product lightweight.