(531a) Comprehensive Planning of Annual Delivery Program for LNG Suppliers

The demand for natural gas has increased over the decade as it is the cleanest fossil fuel. However, it is economically infeasible to transport natural gas over long distances. Therefore, it is transported in liquefied form at cryogenic conditions using specialized ships known as LNG carriers. LNG supply chain planning is done on three levels namely strategic, tactical and operational planning. Strategic planning involves making long term decision that are effective over 20-30 years of duration. For e.g., selection of a contract portfolio, investment decisions associated with both liquefaction and regasification terminals, fleet composition and size etc. Tactical planning involves decision making which are effective over a horizon of 12 - 18 months. Creation of an Annual Delivery Program (ADP) by the supplier is an important tactical decision. Operational planning involves decision making which are effective over a horizon of 1 – 3 months. Rescheduling due to unforeseen events such as equipment breakdown or ship delays due to bad weather etc. are some examples of operational decisions. In this paper our focus is on tactical planning. LNG is procured by customers through Long Term Contracts (LTC) and spot market. LTC is an agreement signed between a supplier and a customer usually spanning up to 20-25 years of duration. So, with increasing demand, fleet sizes, customers around the world and other complicated decisions to make it becomes important for the supplier to plan deliveries to the customers at the accepted delivery dates throughout the supply period. To plan for deliveries for a coming year an Annual Delivery Program (ADP) is created by the supplier. ADP is the list of scheduled voyages which contains information about the long-term contract served, ships nominated to serve the contracts, regasification terminals nominated to receive delivery, date of ship loading etc. In this paper we present an LNG supply chain planning problem with an objective of creating an ADP to maximize the revenue generated by selling LNG in spot market and minimize the cost incurred by the supplier in delivering LNG to the long-term customers.

Currently, most of the literature present on creating ADPs is to satisfy the LTC demands only considering the inventory and berth management at the supplier’s port. Rakke et al. (2011) developed a rolling horizon heuristic (RHH) for creating LNG annual delivery programs. A Mixed Integer Programming (MIP) formulation was proposed¹. StÃ¥lhane et al. (2012) developed a construction and improvement heuristic (CIH) to create ADPs². Their models allowed under supply of LNG and did not consider inventory and berth management at the customer’s terminal. Mutlu et al. (2016) proposed a MIP model which incorporated split delivery for delivering LNG cargos. Commercial optimizers were not providing even a feasible solution to the MIP model in reasonable time. Small instances were tested but even those instances could not be solved to optimality in a 24-hour run time. So, they proposed a Vehicle Routing Heuristic (VRH) which gave cost effective solutions and outperformed commercial optimizers³. Sheikhtajian et. al. (2020) considered the model proposed by Mutlu et. al. (2015) and developed a model considering ship’s speed as a fuzzy parameter. The aim of the study was to compare shipping expenses of split and non-split delivery strategies considering both deterministic and uncertain situations. They also proposed a meta heuristic as a solution method to solve the problem⁴. To summarize, some of the models in literature allowed under-supply neglecting inventory and berth management at the customer port, some models did not consider production and delivery of multiple grades of LNG, while others were not able to give optimal solution within reasonable time. We seek to address these gaps through our model.

In this paper, we present a Mixed Integer Linear Programming (MILP) model for delivering LNG shipments from the supplier to LTC customers and a regional spot market. The problem has one supplier and multiple customers spread around the world. The supplier produces and delivers multiple grades of LNG. Our objective is to create an ADP which maximizes profit for the supplier while considering ongoing long-term contracts and regional spot market. The supplier generates revenue by selling excess LNG in spot market. The cost incurred by the supplier in delivering LNG is the sum of transportation and penalty costs if it fails to abide by LTCs. The minimum and maximum production rates, storage capacities at the supplier and customer terminals, transportation time from supplier to customer ports and the slot wise demand at the customers terminal are known a priori. Our model allows over-supply to customers in a reasonable limit and penalizes the supplier for delivering more than the demand. In this paper, we also focus on the inventory and the berth management at the customer’s port. The supplier is responsible for scheduling the heterogeneous fleet of ships. Our model explicitly accounts for loading of ships at the supplier’s port, inventory of LNG at the supplier’s and customer’s terminal. The supplier sends excess LNG to the spot market over the planning horizon. Maintenance of ships is also incorporated in our model. The MILP model is demonstrated on various scenarios having different demands. We get an optimal ship schedule from our model which satisfies the inventory and berth constraints at the supplier and customer ports while fulfilling customers demand with reasonable over-supply. The MILP model for LNG delivery has been implemented in and solved using CPLEX. We believe that the results from this paper would help develop critical insights regarding the tactical planning aspects of the LNG supply chain from the perspective of both supplier and customer.

References

1. J. G. Rakke et al., “A rolling horizon heuristic for creating a liquefied natural gas annual delivery program,” Transp. Res. Part C Emerg. Technol., vol. 19, no. 5, pp. 896–911, Aug. 2011, doi: 10.1016/j.trc.2010.09.006.

2. M. Stålhane, J. G. Rakke, C. R. Moe, H. Andersson, M. Christiansen, and K. Fagerholt, “A construction and improvement heuristic for a liquefied natural gas inventory routing problem,” Comput. Ind. Eng., vol. 62, no. 1, pp. 245–255, Feb. 2012, doi: 10.1016/j.cie.2011.09.011.

3. F. Mutlu, M. K. Msakni, H. Yildiz, E. Sönmez, and S. Pokharel, “A comprehensive annual delivery program for upstream liquefied natural gas supply chain,” Eur. J. Oper. Res., vol. 250, no. 1, pp. 120–130, Apr. 2016, doi: 10.1016/j.ejor.2015.10.031.

4. Sheikhtajian, Sanaz, Ali Nazemi, and Majid Feshari. "Marine Inventory-Routing Problem for Liquefied Natural Gas under Travel Time Uncertainty." International Journal of Supply and Operations Management7, no. 1 (2020): 93-111.