Bread, Aeroplanes and LNG
The increasing global demand for natural gas calls for shipping hundreds of millions of tons of it every year—liquefied and cooled down to -160°C. LNG transportation is a very complex and demanding area of shipping, not least because of operating with the low temperatures and high pressures required to keep the cargo liquid. However, managing the performance of the transportation effort of this cold substance sometimes seems to be based more on a warm gut feeling than cold facts.
For myself—coming from outside the shipping industry—this observation has been somewhat puzzling. Being involved with business analytics for years, I have learned that even the most common bread I eat has seen more decisions based on analytics along its way from the rye field to my mouth than shipping has. The volume of the bread and the price of the product at my nearest supermarket are most likely forecasted and optimized. The route and the schedule of the delivery truck are typically based on some sort of optimization as well. The manufacturer certainly has a statistical quality control system in place. Even the decision to farm the rye in a specific spot may have been a result of a statistical analysis of spaces and yields.
If rye bread doesn’t feel familiar, an example of very advanced utilization of analytics touching a large and complex means of transportation can be adopted from aviation. Airlines and turbine manufacturers have for years been developing methods for predictive asset maintenance. This doesn’t mean that the maintenance visits are just planned ahead and done early enough. It means that the data received from the turbine and its surroundings are put under a sophisticated analysis with an aim to identify patterns that indicate forthcoming breakdowns. This kind of method enables the saving of significant amounts of money by fixing equipment just when it’s needed. As a consumer of flights, I’m of course more interested in the safety side: this approach will increase the chances that turbine issues are not occurring when I’m thousands of meters above the ground.
Going back to LNG shipping, one may easily argue that analytics is already commonplace. However, analytics is not an on or off thing; it comes in many flavors, that is, levels of advancement. Yes, reports are being made and (static) figures are being followed, but as long as they are based on postward-looking noon-to-noon values or sea trials done ages ago, we can hardly say they are very advanced nor valuable. The progress in LNG shipping is not an exception. Other industries have gone down the same path: first reporting highly aggregated figures from the past, then drilling down to details, followed by analyzing relationships, moving on to predicting what is likely to happen and ending up defining optimal solutions to certain decisions.
In the two examples above, there were two important viewpoints we can address and adopt in LNG shipping as well: operational performance (e.g., manufacturing and delivering efficiently) and asset performance (keeping physical assets in a good condition). In terms of operations, it is possible for an LNG ship, for example, to be kept in optimal trim at any point in time based on the laws of physics and very detailed actual measurements. One can also optimize the speed, taking into account the weather conditions and engines and fuels used, in order to arrive at the destination just in time, consuming as little energy as possible. Even the specific concern of LNG ships, the boil-off, can be estimated for its amount at any time and the actions for using or getting rid of it most efficiently managed. From the asset point of view, being able to mathematically model the energy decomposition of an LNG vessel enables tracking the development of its performance or that of its components such as engines. One of the most typical examples is fouling, which can be estimated very accurately if the ship operator knows how much the ship consumes and what it should consume in the prevailing conditions.
The reasons for some industries to pioneer in analytics have been at least two: cost pressure and safety. Margins for bread are small, and airplanes…well, they just better not drop. When any reason becomes compelling enough, gut feeling may not be sufficient to cope with it, and analytics may need to be called to the rescue. Is the pressure enough for LNG shipping to take the next step as well?