We consider an Arctic drilling platform, which must withstand seasonal ice floes and be protected against extreme ice features. Adequate ice floe thickness estimates are not discernible from satellite images, so autonomous underwater vehicles (AUVs) equipped with sonar scan the thickness of the floes. We approximate this system by an M/G/k+G queue and develop a high fidelity simulation model. We develop and solve a stochastic facility location model, which addresses the design and operational decisions related to the system. Finally, we create a visualization tool to graphically represent the docking station configurations, ice floe arrivals, and the AUV scanning process. These simulations were part of a desktop study and have not been validated through field trials, which we hope to accomplish in the next phase of this study.