납사 크래킹 센터의 납사 입고 및 이송 계획을 위한 수학적 계획모델 및 휴리스틱 방법의 개발

Abstract

This thesis addresses the feeding scheduling problem in a Naphtha Cracking Center. To tackle this problem, two mathematical programming models and one heuristic algorithm are suggested. Chapter I addresses a naphtha feeding problem for Naphtha Cracking Center (NCC). The rapid increase of petroleum prices compelled to petro-chemical industries to figure out ways to remove any potential redundancies in and out of their network. The increasing attention on integrating activities that have been addressed separately is in line with this trend. The naphtha feeding problem involves two key operations: delivering naphtha from refineries to NCC and blending naphtha in storage tanks before feeding it to NCC. While the first is concerned with selection sources and scheduling the loading and unloading of naphtha, the latter involves the transfer of the naphtha from storage tanks to a charging tank. The both issues are simultaneously considered by transforming them into a single mixed integer linear programming problem of minimizing the cost function of naphtha prices, shipping expenses, and unloading costs, etc. A numerical example of a real industrial case is presented to illustrate the applicability of the proposed mathematical model. In Chapter II, we propose a decision-supporting framework for a feeding problem in the petrochemical industry. The problem is concerned with delivering materials from suppliers to plants, unloading and storing in storage tanks, and mixing the materials before directly feeding into main processes. Most of the previous works in the literature have addressed these concepts, based on the assumption that the delivery of raw materials is given and fixed. From a joint investigation with industry partners, we have determined that the purchase of feedstock and its delivery also are critical issues in the feed scheduling problem of real-world plants. Thereafter, we takes into account previously addressed issues separately and simultaneously, to increase the overall efficiency. The corresponding decision-making problem is mathematically transformed to a mixed-integer nonlinear programming (MINLP) problem. The solution of the problem is computed using the iterative framework between that of a relaxed mixed-integer linear programming (MILP) problem and that of a nonlinear programming (NLP) problem, to prevent compositional discrepancy. An industry-coworked example of the naphtha case is presented, to illustrate the applicability of the proposed framework.In Chapter III, a heuristic method is developed to address the unloading and blending scheduling problem in NCC. This method is based on the rules which are applied to mange tanks in NCC. By applying this method, the massive scheduling problem is successfully solved in a short time compared to the mathematical programming method. By simulating the schedule solution generated with the proposed method, site-operators can estimate the long-term plans (vessel arrival events) predetermined by the head quarter. If the simulation shows the bad results, site-operators can require that the predetermined long-term plans should be modified in order to maintain NCC at stable and efficient condition. This is the additional advantage to apply the proposed method.