\u201cPeople talk about \u2018LNG\u2019 but it\u2019s really a mixture of gases,\u201d says Christos Chryssakis of DNV GL, \u201cand there\u2019s large differences depending on where it\u2019s sourced\u201d. While what\u2019s brought onboard in one area could be almost pure methane, in another it could contain up to 13% ethane, 3% propane, 1% butane and maybe a few percent nitrogen he explains.<\/p>\n
However, regional variation isn\u2019t the only thing playing around with the energy content of the gas; the \u2018boil-off\u2019 of a number of elements can significantly alter its composition during a voyage. \u201cWe knew the quality of the gas onboard was changing,\u201d says Mr Chryssakis, \u201cbut we had no idea by how much\u201d he admits, adding the variation was discovered \u201calmost by chance\u201d when the team installed a gas chromatograph \u2013 a separation technique that yields an on-the-spot analysis of volatile compounds \u2013 during an energy efficiency research programme.<\/p>\n
However, his colleague Sander Gersen adds that this change is \u201cnot straightforward\u201d by any means as other gases \u2013 as well as methane – will evaporate at different points along the journey. Somewhat surprisingly, this means the quality of the fuel may not be diminished but could be raised over the duration of the voyage. \u201cNitrogen, for example, needs -196\u00b0C to keep liquid, rather lower than LNG\u2019s \u2212163 \u00b0C, and thus boils off first,\u201d he says, pointing out that nitrogen is not particularly useful for combustion. Moreover, different types of tank produce widely divergent effects while the conditions en-route will impact sloshing: this too plays a part in the boil-off sequence.<\/p>\n
The end result is a product that\u2019s almost constantly changing, explains Mr Gersen, Mr Chryssakis adding these variations are big enough to prompt at least one long-distance carrier to fit chromatographs or other gas sensors to the newest of its fleet: determining quality, not just quantity, of delivery means more appropriate billing.<\/p>\n
Even though the calorific value of cargo on delivery is important enough, the impact onboard is proving to be even more of an issue as many carriers have fitted dual-fuel engines to make use of the boil-off gas for propulsion.<\/p>\n
Mr Chryssakis recalls: \u201cWe saw the engines were only running up to 80% load, but these engines are at their most efficient when the loading is even higher… which is when we started to look at other, power industry applications to see where we could bring in some improvement.\u201d<\/p>\n
Mr Gersen explains that one of the most critical issues for these type of engines is spontaneous, early combustion \u2013 knocking. This is really down to an uneven ignition, linked to the other, non-methane mixtures in the gas which have different burn characteristics.<\/p>\n
\u201cIt\u2019s a problem common to spark-ignited gasoline and pilot-fuelled, spark ignited and pre-chamber natural gas engines,\u201d he says, adding that it\u2019s a well known issue in the power generation industry which also has to deal with gas imports of varying quality. \u201cAs it\u2019s to be avoided at all costs the result is that manufacturers adjust their engines to the \u2018worst-case\u2019 fuel… which means there are large efficiency penalties to be paid.\u201d<\/p>\n
Therefore, the hunt has been on for a system that would allow a better, more responsive tuning of the engine to the fuel while still keeping well away from the potential for knock.<\/p>\n
However, it has taken some time. He explains a breakthrough needed two elements: one being in sensor technology. Although the classification society was in a very good position to keep an eye on new developments, \u201cuntil fairly recently there just wasn\u2019t a real time gas sensor available that could do the job\u201d says Mr Gersen. \u201cWe kept testing them out in our lab, but only a year or so ago did we find one that could measure propane and butane well.\u201d<\/p>\n
The other element was a more reliable way to predict knock resistance of fuels \u2013 and this DNV GL has been working on for a while. Current techniques suffer from a number of shortcomings he explains: the common methane-number methods are derived from data using a stoichiometric engine from the 60s, \u201cbut today things are different, for example modern engines tend to operate fuel-lean for a substantial amount of the time\u201d. Therefore a lot of hard work has gone into a combustion algorithm that more accurately maps knocking against different fuel qualities. \u201cIt\u2019s a very sharp tool,\u201d says Mr Gersen.<\/p>\n
Put all this together with a fast-response gas composition sensor which can sample the fuel line, and the result is \u201cnear real-time responsiveness to what\u2019s going into the engine\u201d says Mr Chryssakis. And this, according to both researchers, could lead to up to 6% fuel saving as the engines could be loaded much more efficiently \u2013 and it also gives the operator the flexibility to shut off one engine entirely allowing the others to do the job, reducing running hours and maintenance.<\/p>\n