High Performance Computing based system for optimization and analyses in the LNG business
Publication Date: 2014-Jan-14
The IP.com Prior Art Database
The current LNG business is driven by long-term contracts and planning. Currently, annual delivery schedules for each LNG project are planned and agreed upon by various parties before the beginning of each year. In addition, an updated 90-day delivery schedule is developed by the LNG producer and provided to customers every month to account for deviations from the annual schedule. Agreement on these delivery plans can involve significant negotiation and coordination of operations by several parties. Consequently, operating an LNG production terminal involves significant long-term planning which can greatly benefit from robust planning and optimization tools.
Looking to the future, increasing liquidity in the LNG market may cause the global LNG business to evolve from a long-term contracts based business to one with significantly more flexibility and short-term sales. This will complicate the management of projects since operations will have to be optimized not only to satisfy contractual obligations but also to maximize profitability by exploiting contractual flexibility and market opportunities.
Background: Computational Technology for LNG business
Known attempts to manage LNG projects via computational technology have fallen short because of substantially reduced scope, reduced capabilities of the proposed solutions, and/or a lack of the technology utilized. The following paragraphs discuss known attempts as they relate to various aspects of the disclosed methodologies and techniques.
Ship Scheduling. Many LNG projects currently tend to use simple spreadsheets for scheduling ships. The schedule has to be populated manually and does not provide any optimization functionality. Even in the more detailed systems, there are no known integrated models for lifting schedule generation combined with ship schedule optimization. This can lead to sub-optimal plans manifested in over-utilization of spot vessels for satisfying contractual demands. Further, generating a feasible shipping schedule could require a great number of iterations between the capacity planning and the ship scheduling components. Additionally, the ship scheduling components of the more sophisticated models do not seek to optimize schedules for selling spot cargoes, and do not account for transportation losses in cargo (e.g. boil-off, fuel) and consequently the generated ship schedules have discrepancies when attempting to satisfy contractual obligations related to annual volume delivered.
Existing technology has addressed problems of developing Annual Development Plans (ADPs) for larger LNG projects. While Rakke reports results for problems with multiple ships and a one year planning horizon, the optimization model and solution methods are fairly simplified. For example, the model is built for a case with only one producing terminal, boil-off and heel calculations are not integr...