Speaking at The MathWorks’ Energy Virtual Conference last month, Tristram Scott described how consultants Quantitative Energy Models use Matlab to study LNG terminal operations. LNG tankers take several weeks to reach their terminal destination for offloading, storage and or re-gasification for local markets. Price differentials between local and remote markets mean that operators may opt to divert shipments to more lucrative markets.
QEM’s Matlab-based models consider month ahead prices and shipment diversion costs. Diversions are considered only if they are clearly profitable—i.e. when a certain threshold probability is exceeded. To determine this, a ‘price duration curve’ is generated from gas prices—the PDC provides an estimate of the time that prices are expected to be at or above a certain level. Scott’s paper provides an in-depth look at the Matlab code behind the PDC and other indicators including maximum storage capacity, release and injection rates.
Another option for LNG operators is as feedstock to hydroelectric storage (pumping water to an uphill reservoir for later turbining). Here QEM uses stochastic dynamic programming to simulate arrivals and storage options. Scott concluded that LNG terminals fit well into framework of the standard hydro model and that the flexibility of having storage at an LNG terminal has considerable value.
~
Another Matlab event this month highlighted RWE AG’s use of Matlab to develop an energy pricing engine integrated with its SAP enterprise resource planning tool. The risk management system for natural gas trading a.k.a. ‘EwITA’ (Entwicklung IT-Zielarchitektur Gas) takes account of linear and non linear factors affecting price volatility and replaces previous ‘inconsistent implementations’ of the formulae throughout the enterprise in the legacy system and in Excel. The Matlab Compiler and Builder JA were used to create the production version of the system and interface with SAP. Since February 2009, when EwITA went in production, it has had zero downtime. A recording of the webinar is available at links/1004_8 (login required). More from mathworks.com.
© Oil IT Journal - all rights reserved.