Refinery operation planning is a complex task since refinery processes and inventories are tightly interconnected. We study deterministic refinery planning when ships are loaded with a blend of components. Any delay in ship arrival may result in overfull component tanks which results in less efficient blending alternatives, reduced process operations or even shut downs. In this paper we formulate a planning model and in order to analyze the performance of the model ship arrivals are simulated over a period of 28 days. From numerical results, we can show that the model indeed can model the integrated problem. The practical problem also include uncertainties, and in particular for ship arrival times. Safety stock and blending flexibility are utilized in order to deal with uncertainty. The impact of uncertainty is also studied in the numerical experiments.