Cheap; fast; reliable. An engineer will tell you: pick two. The promise of additive manufacturing (AM) has been in tackling the first couple; but recently, there has emerged a distinct possibility of resolving the third, too.
AM is so-called because it replaces ‘subtractive’ manufacturing techniques — in which the raw material of manufacturing is arranged to a vague shape, before being machined down into its final form, wasting upwards of 90% of it. Theoretically then, AM could use less raw material, and be cheaper than conventional manufacturing methods.
But up to this point, 3D-printed parts have been limited to mock-ups and prototypes, thanks to their lack of strength. Working in plastic, building an object out of layers from the bottom up does not imbue them with much adherence between those layers. For metal, this technique is almost useless, which is why a different process, Direct Metal Laser-Sintering (DMLS), is used instead. This involves firing a laser at a powder bed, welding the particles together.
It is fraught with its own set of complications. Millenia of metallurgy have taught us that the strongest metal structures are formed when the entire component is heated and cooled at once. But the technology is gradually improving.
Nature knows best
The manner of this improvement has to do with the intersection of materials science and computational power, and the name engineers have chosen for this is “topology optimisation”. It refers to the practice of achieving the optimal possible geometry for load-bearing equipment by situating all the material along axes of stress – called ‘Michell structures,’ after the father of topology optimisation, Anthony George Maldon Michell.
Topology optimisation is nearly impossible using traditional manufacturing. Perhaps that is why the structures it produces look so deeply strange. In a recent example, carmaker Porsche, working with German compatriots Mahle, Trumpf, and Zeiss, developed a new type of additive-manufactured piston for its engines, weighing 20% less than a cast or forged equivalent, and able to withstand higher RPMs and temperatures. On closer inspection of the piston crown, it is possible to see structures bearing an eerie resemblance to the human ear. Porsche calls this ‘bionic design,’ because in nature, structures like the skeleton — or the ear – tend to use material sparingly, in locations where forces are transferred.
Clearly, this approach works. In testing, the pistons proved capable of running on a full engine load of 7300RPM, for over 135 hours – quite the achievement, for an AM part. In fact, at the very cutting-edge of AM, it has proven possible to use DMLS to make structures three times stronger than cast stainless steel – though it requires operating at the micron-scale, well outside the scope of almost every practitioner.
Of course, the value offering of AM for racing cars is wildly different from what it could bring to shipping: tankers are rarely seen trying to lap one another between ports, and engines are designed not to be rebuilt every 24 hours, but to operate for 30 years. It is certain, though, that AM will have a business case to make in the maritime sector, and it has to do with getting parts to where they are needed.
Layer by layer
In 2014, Maersk announced that it would be embarking on a project to 3D-print spare parts aboard its ships as and when they are needed, instead of carrying every conceivable spare part on board. A small number of non-weight bearing parts — valves, pump components, tubes and pipes – are suitable for in-situ AM. Maersk is not the only one to think of this: the US Navy has embarked on a similar project.
Not much has been heard from Maersk since on the project’s success or failure, but recently there has been a flurry of activity elsewhere. Late last year, ABS certified a 3D-printed gear set, shaft and boiler supply fuel pump, as well as a flexible pump coupling and ejector nozzle for a freshwater generator, which were installed on board a Sembcorp Marine tanker and operated for six months. ClassNK released guidelines for AM, including specifications for the feedstocks, and the manufacturing, testing and inspection processes.
In February, DNV awarded Type Approval, DNV-CP-291 “Additive Manufacturing Feedstock,” to a manufacturer of solid wire feedstock for metal printing systems.
Components in shipping are more often one-offs than in most industries, and getting them onto ships can cost thousands of dollars. Many believe that ports will soon have local AM facillities which can produce spare parts on demand, an issue on which Maritime and Port Authority of Singapore (MPA) and Singapore’s National Additive Manufacturing Innovation Cluster (NAMIC) have been leading the charge.
The Port of Rotterdam has the same idea. Both envision an ecosystem where intellectual property, and not tangible objects, become the purchases of the future. Residing in the cloud, the effort of applying AI and topology optimisation to such designs to be able to build them cheaper, faster, and more reliable would be trivial compared to what it takes today.
But a PricewaterhouseCoopers (PwC) research paper, 2015 Commercial Transportation Trends, is worth remembering. It anticipated that AM would lead to a 37% decrease in container shipping volumes, due to a process of ‘deglobalisation’. Most of a decade later, we are not yet awash with cheap, locally-sourced consumer goods; but with the threads connecting the continents visibly fraying, it may yet emerge that PwC had their cause-and-effect the wrong way around. Shipowners would be well-advised keep a close eye on these developments.