4.0 Evaluating Storage Facilities


4.1 Overview


Storage facilities and stock levels play a crucial part in inventory management. The more inventory one has on hand, the more likelihood of meeting demand. On the other hand, there are associated holding and capital costs involved with keeping a high stock level. The cost of a stock-out depends on how long it lasts.


Currently, New Plymouth and each of the three depots, Auckland, Christchurch and Dunedin, have a permanent tank, each of a different size. Liquigas Ltd has been considering extending the size of its permanent storage tank in Christchurch by one of two possible sizes. The following storage options are being considered:


              I.      Keeping the status quo

           II.      Increasing the storage capacity in Christchurch from 2000 tonnes to XXXX tonnes

         III.      Increasing the storage capacity in Christchurch from 2000 tonnes to YYYY tonnes


For each different storage option/scenario configuration, including the status quo, the ability to avoid stock-outs was analysed by using the Simulation Model, the monetary costs associated with implementing the option were evaluated and corresponding risks and limitations were considered.


4.2 Future Scenarios


A ‘scenario’ in our evaluation of storage options, refers to the future demand and supply situation of LPG. Previously, we had thought of using our daily demand/supply forecasting models to forecast one such possible scenario. Since our analysis in the previous section concluded that this was not appropriate, we decided to create simpler scenarios, with which to provide different situations to access the usefulness of different storage options.


Scenario One


Scenario One assumes a decreasing supply of LPG and an increasing demand of LPG. We investigated using the average growth of demand so far, to project the daily demand of LPG into the future.  Using all the historical data we had up to the end of 2005, we first tried totalling the monthly demands from September 2001 to December 2005. For each month, the average percentage increase/decrease from year to year was calculated. Then using the 2005 data, every day in that month in 2005 was increased/decreased as appropriate. A similar approach was applied using seasonally and yearly percentage increases/decreases.  The daily forecasts for 2006 were evaluated against the real 2006 daily data, by calculating the prediction errors. In addition, the forecasts and the real data were summed into monthly figures, and the monthly errors were calculated and analysed. We followed the same procedure for supply.

After the daily errors and monthly errors were assessed for each ‘method’, it was decided that using the average monthly growth percentage would be the most appropriate method to project a future scenario. Using the 2005 data as the base year, since that was the last year that we had a full year’s worth of daily data, projections for supply and demand were made into the future up to the end of 2010.


Scenario Two


Scenario One assumes that supply is steadily decreasing to a point that supply cannot meet demand. Scenario Two assumes that the supply situation does not get to such a drastic stage, that is, Liquigas Ltd is able to source LPG from elsewhere to help relieve the situation.

The Simulation Model only allows one source of supply at the current stage, therefore when we looked at additional supply; the only option was to add to the Maul supply input file. That would assume that any source of supply would go through the Maui pipelines into New Plymouth. Assuming that all future supply of LPG is from within New Zealand (when new fields open, and start producing LPG), this is a reasonable representation of reality, since domestic suppliers are close to New Plymouth (and therefore Maui) anyway.

Scenario Two takes the 2005 figures for LPG supply and keeps them constant from year to year until the end of 2010. Daily demand of LPG is still assumed to be growing, and is the same figures as used in Scenario One. In this case, over the course of a year and most months, it is assumed that there is enough supply to meet total demand. Of course, on a particular day, supply may not meet total demand and at any case, even if it did, there are lead times due to shipping that would prevent the stock being able to meet demand at its depot, where it is needed.


Main Page

1.0 Background

2.0 Forecasting Supply

3.0 Forecasting Demand