(607f) Reactive Scheduling of Short-Term Crude Oil Operations Under Tank Availability and Demand Uncertainties

Nowadays, petroleum industry is confronted with versatile market, unstable crude prices and strict environmental restrictions, which results in lower margins and tighter surviving space for refineries. And due to the globalization and dynamic environment of petroleum industry, the supply chain is more vulnerable to unexpected events which lead to interruption of normal operations and incur adverse outcomes such as failure to fulfill the customer orders, process shutdown and waste of materials. Uncertainties can be defined as any abrupt incidents that cause abnormal activities on supply chain to deviate from what are planned, which will result in infeasible solution and thus, necessitate rescheduling to be employed. There are mainly two categories of rescheduling strategies existing, one is called proactive and the other is reactive. The major difference between the two is that the latter puts more emphasis on solution quality instead of flexibility, and it makes amendments whenever the deviations actually become realized. In addition, proactive rescheduling usually itemizes the potential uncertainties and largely depends on probabilistic strategies.

Literature review suggests there lacks in-depth discussions on reactive scheduling without aids of heuristics for short-term crude oil operations. In this work, a new two-stage solution procedure is developed to swiftly generate updated optimal schedules in face of various uncertainties. At the first stage, the deterministic model is solved to derive the optimal schedule based on the current problem data. The scope of the deterministic model includes single-parcel vessels that carry different crudes, together with storage and charging tanks which are used to unload and mix the crudes, respectively, and also distillation units which receive the blended crudes and carry out further processing. Since short-term crude scheduling is a multi-objective problem inherently, for which the top priority is to ensure continuous operation of distillation columns throughout the scheduling horizon, the related hard constraints are applied. Then, objective function is formulated based on operational costs consisting of vessel demurrage and unloading, column changeover and tank inventory costs. As for the second stage, a set of compatibility constraints is developed that comprehensively incorporates the potential amendment options to achieve smooth transition from original schedule to reschedule. In order to improve the operability of reschedule, the objective function is modified by including a penalty term to keep the reschedules as close as possible to the original ones. In this paper, two types of uncertainties are taken into account such as tank unavailability and demand change. For tank unavailability, the decision variables are fixed to the original schedule if the associated operations take place during the period before tank breakdown or at some units other than the tank that fails. However, the operations in progress at the tank which malfunctions are discarded and the starting times are shifted after the tank finishes maintenance. For the scenario of demand change, it’s assumed that an important client places a rush order for a certain product downstream of a CDU. In this light, corresponding column demands need to be altered which may lead to infeasibility given the limited refinery capacity. Therefore, the analysis about inherent flexibility of a refinery under demand uncertainty is conducted.