Simulating the weather to help LNG operations

When we conduct simulation studies of LNG terminal operations, or indeed any other type of shipping terminal, the weather patterns are obviously a key factor. Bad weather will delay ships that are attempting to enter or leave the port, which could ultimately lead to higher inventory and potential tank-tops. This means that accurately reflecting the effects of the weather is crucial when assessing the correct investment in storage and berthing capacity.

Where does the data actually come from?

There are a number of sources of weather data, some locations have an abundance while others have little information available. There are normally two main types: 

1. Detailed hourly recordings from a nearby weather station, probably showing wind speed and direction, visibility and current. There could well be other factors that are recorded as well, for example the number of lightning strikes.

2. Historic data of channel closures, perhaps recorded by the local pilots and showing when transits were not possible.

These data sources can generally be used effectively to support a model of the weather, but we need to be careful in some areas too: 

• If we have weather station data, are these for the correct location? For example the local airport may have extensive historical data which is quite tempting to use because of the quantity and detail, but the visibility at sea could be significantly different to that inland
• How well are the weather constraints understood, can we always say that transits are stopped if the wind speed exceeds 20 knots, or does it also depend on wind direction?
• A good log of historical closures from the pilots can be an excellent source, but it is likely that they will only log a closure when there are ships actually waiting to transit. Some bad weather periods might be missed, especially when looking at an area that has a relatively low number of ship transits
• Similarly, the pilots’ log of closures will relate to the types of ship that have historically used the channel; if LNG ships are new to the area, they may be subject to different transit rules.

A sophisticated weather model is possible

So we know that there are a variety of different weather statistics but how do we choose the ones that are relevant to the model in question? Firstly, we need to find out which specific aspects of the weather are going to affect operation of the terminal. In most cases we try to characterise the weather by determining the average duration of a weather closure, and the average time between weather closures; normally different by month of the year.

Sometimes we need to use a more sophisticated weather model. An example would be where ships are not able to approach the berth if the wind speed exceeds 25 knots. In addition, loading must be suspended if the wind speed exceeds 30 knots. In this case these factors cannot be input as two independent constraints within the model, otherwise we would end up with a 30 knot disconnection at a time when transits were still allowed. Here we need to build up a model of how the wind speed varies over time and use that in the model’s decision logic instead.

Often we link weather characteristics together and model the correlation between the two. It could be that ships are not allowed to berth if wind speed is above 20 knots, nor are they allowed to berth if the wave height is above 2 metres, but the periods of high waves are quite likely to coincide with periods of high wind.

LNG Logistics Simulator

Accurately modelling the weather patterns that affect a shipping terminal is key for efficient planning and decision making in LNG operations, and will impact hugely on the inventory levels required for smooth operation. Experience has shown us that we can build up different weather models that appear to match the actual data available, but which still allow for different terminal behaviours. In my experience it is definitely worth investing time into detailed weather analysis to get the model right so decision makers have the confidence that their preferred strategies will perform as planned and time and money isn’t wasted.

If you interested in finding out more on how LNG operators can model the weather and a variety of other influencing factors, take a look at Lanner’s LNG Logistics Simulator which generates data driven results across the whole LNG value chain.