Design Optimization and Metamodel-based Sensitivity Analysis of Various Capacity Sterilization Shredder

Since COVID-19, a significant amount of highly infectious medical waste has been generated from medical facilities. The typical method for processing such waste involves a sterilization-based shredding system. However, these systems are often overly designed and lack optimization based on each facility's capacity. To address this challenge, a data-driven metamodel-based sensitivity analysis and design optimization approach is proposed. The proposed method used Latin Hypercube Sampling (LHS) to construct an efficient metamodel encompassing all relevant information about the design space. This metamodel, generated from finite element analysis (FEA) data, serves as an effective stress estimation tool. This stress estimation model was used to perform global sensitivity analysis (GSA) and optimization processes. The proposed approach significantly reduces the number of simulations required for sensitivity analysis, leading to a substantial decrease in computational time. The optimization is demonstrated for shredders with two different shredding capacities, which showed significant weight savings.