ABSTRACT:

To approximate the dynamic responses of a system with respect to specified design variables, kriging metamodel technique approximating the responses is employed because it has found much success in approximation of highly nonlinear computer simulation. An efficient sampling technique and the kriging metamodel to predict the motion of a tracked vehicle traveling on soft soil are suggested. Based on the simulations at the limited pre-sampled points, the kriging metamodel is constructed and the accuracy of the proposed metamodel is discussed.

INTRODUCTION

Simulation based detailed design of a complex system, for instance, more specifically deep-sea ocean mining collector, is increasingly becoming a distributed design activity involving multiple decisions and complicate trade-offs among various design variables. Therefore multidisciplinary design optimization can embody a set of methodologies that provide a means of coordinating efforts and possibly conflicting recommendations of multidisciplinary design teams from early design phases to the detailed design phase. Dynamic analysis of tracked vehicles on soft soil is quite expensive and very time consuming because of its numerical complexity and nonlinearity. A simplified 3-dimensional dynamic model of a tracked vehicle that crawls on cohesive soil such as deep seabed is simulated and its responses are approximated by means of the so-called the kriging metamodel. The vehicle is assumed as rigid body with 6-degree of freedom to investigate the dynamic behavior of the vehicle traveling on cohesive soil. The cohesive soil is modeled through relations of pressure to sinkage, shear displacement to shear stress, and shear stress to dynamic sinkage. Equations of motion are derived with respect to the body-fixed coordinate system to investigate 3-dimensional dynamics of the tracked vehicle. The dynamic analysis that is very time consuming work is executed for a tracked vehicle model and the performances required in the concept design phase are computed.

This content is only available via PDF.
You can access this article if you purchase or spend a download.