Coke Formation Model in Crude Oil Furnace for Maintenance Scheduling

Abstract

Abstract A crude oil distillation unit (CDU) is one of the most important unit in petroleum industry. Its main function is to separate the crude oil into many kinds of petroleum products. Generally, the CDU’s design includes the crude oil preheater, which are either cabin or vertical cylindrical furnace, for adjusting the crude feed properties and increasing feed temperature. Carbon coking inside the furnace during the preheating process leads to accumulating of the coke, deteriorating of the product quality, increasing of the pressure drop across the furnace, and increasing of the energy consumption. The de-coking process is normally executed on demand based on the measured heat loss or performed every fixed period of time. However, due to the different rate of coke formation of various crude oil grades, as well as the difference in cost of product, process operation, and maintenance operation; both on-demand and fixed maintenance practices are not optimal method for de-coking. Model of the coking rate and accumulation inside of the furnace could be a useful tool for scheduling the decoking. The model includes the balance equations for the heat generated, the heat transfer in forms of convection and radiation, the changes of the temperature and the amount of crude oil and coke, the constitutive equations for the coke formation and accumulation, and the conditional equations for optimization of the de-coking schedule. The model parameters are fitted to the data provided by the refinery in Thailand with absolute average deviation below 3%; the operation and maintenance costs are also estimated from the financial activity report of the similar sources. It is found that with different sources of crude oil, the optimal furnace maintenance schedules are different. The improvements, in term of cost per maintenance, are found to be 15 - 34% depending on the operation scenario. Although, the initial results look promising and the initial goal is accomplished, the application of the model toward multiple crude oil feed for better operation is under development.