We present an approach for modelling a deterministic simulation process using design of experiments. Five design of experiments models - Central Composite Design (CCD), Face-Centered Central Composite Design (FC-CCD), Box-Behnken Design (BBD), 34 Factorial Design and Latin Hypercube Design (LHD) - are evaluated to identify the regions of predictability for modelling the performance of a 1.0 mm multifunctional forceps-scissors surgical instrument for minimally invasive surgery. The objective is to develop metamodels to accurately and efficiently evaluate surgical tool performance, and to determine the feasibility of employing design of experiments in similar problems. The instrument chosen for this framework was selected because of its potentially broad application in surgical procedures. Through this case study we found that the bending stress metamodels had an accuracy of ±5%. On the other hand, the models approximating blocked forces are less accurate, indicating that there is much more curvature in the relationship between these blocked force responses and design variables.
All Science Journal Classification (ASJC) codes
- Industrial and Manufacturing Engineering