Improving Shipping Distribution at Exxon

Exxon use WITNESS simulation software from the Lanner Group to ship increased production of plastic in Belgium

Background

Exxon Chemicals is one of the world's largest and most successful petrochemical producers. Key to its European operations is an integrated system of three chemical plants, two in Britain and one in France, which feed large plastics plants in Belgium via the Exxon terminal in Antwerp. At the U.K. refinery and chemicals complexes in Fife, Fawley and LeHavre, crude oil is refined into liquified petroleum gas, lubricating oils and bitumen. In all, some 30 product groups are produced. One of the chemical products, Ethylene, is shipped to Antwerp for conversion into plastic. A few years ago, planning was in progress to increase the production of plastic in Belgium and to increase ethylene output at Fife. A key factor in the success of this multi-million dollar plan would be the ability to ship the increased production around a shipping circuit involving France, Britain and the Benelux countries.

The following report shows how Exxon Chemicals used WITNESS to evaluate its alternatives. The simulation had to take into account some two thousand ship movements per year between the four ports involved and allow for the effects of weather and tides, storage tank capacities, pipe sizes and pumping rates. As with all business plans, it is invaluable to be able to review a number of alternative strategies, each of which will have its own strengths and weaknesses, and simulation has long been regarded as a good method of doing this. WITNESS helped Exxon quickly build large and sophisticated models so the company had more time to explore alternatives and test for the best possible option.

Two separate areas were modelled: the shipping circuit between the three dispatching ports and Antwerp and, as part of an ongoing jetty automation project, the detailed handling of products at the Fawley site.

The shipping circuit was causing concern because it was approaching berth capacity at some locations, and this had the potential to result in unacceptable queuing times whenever there was bad weather. On top of this, the key site at Fife (targeted for increased production) was particularly susceptible to disruption through high winds. Three options were available: increasing shipping capacity, investing in jetty operational improvements, and increasing tankage. Millions of dollars were at stake, both through the high cost of operating disruption and the probable capital costs of improvements.

Exxon used WITNESS simulation to build a model that enabled extensive sensitivity analysis of the competing options. By careful use of the visual and interactive modelling capabilities it was possible not just to test already known alternatives, but to develop others previously unsuspected. Exxon engineers were able to watch the interaction of multiple events in an animated graphical form, and then to make instant changes.

In the case of the shipping circuit, the events that were modelled included:

• size and number of ships
• schedule of production and consumption
• schedule of ships
• pattern of ships for other products
• tank sizes at the production and user sites
• ship journey times 
• product pumping rates
• use of additional jetties at the producer sites
• use of third party tanks at user areas
• weather effects on the jetty operations

The four major results being monitored were:

• the risk of operating disruption to either a producing or consuming plant 
• the implied cost of capital to meet the simulated outcomes
• the operational costs
• the impact on the supply of non-modelled products from the plants

The information for the simulation was read into WITNESS from a spreadsheet. This included the proposed schedule of shipping that WITNESS would then attempt to meet after simulating what could happen during the time span of the schedule. Any event that might reasonably be expected to effect the circuit and therefore the results was statistically modelled. This included: wind speed and direction, tides, fog, and other shipping movements.

The most complex part of the model was allocating ships to berths. This involved testing some 15 decision conditions before choosing which ship would berth next and in what berth. Conditions considered included: weather, tides, tugs and pilots, product availability, line preparation, and ship size and type.

Results

The model confirmed that there was an operating problem that required corrective action. The use of the model has allowed Exxon to identify and evaluate several potential solutions involving, in combination:

1. Changes in shipping patterns
2. Changes in nonethylene inventories to reduce interference problems
3. A review of terminal flexibility with the Benelux countries

Exxon's experience demonstrates how flexible WITNESS simulation can be, and in particular the faith with which major companies entrust simulation as a means of planning both small incremental improvements to daily running and major strategic initiatives that can make or break a company's competitiveness.