EYE ON DESIGNBY EVERT-JAN FOETH New Propeller Optimization Process Can Analyze 10,000 Designs a DayA newly developed propeller optimizer makes it much faster to identify propeller-hull reactions and reach design decisionsn the EU project STREAMLINE drance from cavitation is as low as pos- Avoiding the Generation of Unsuit- thrust requirement and a larger propeller the fuel ef? ciency of a chemical sible. This is a time-consuming process able Propellers diameter, and lastly, an increase in pro-tanker was improved by optimizing that is particularly ill-suited to ? nding Naturally, the proof of the propeller is peller ef? ciency by an optimized loading Ithe hull form and propeller. To this the best propeller when considering 500 in the computing. For STREAMLINE, distribution in the wake of the ship. end, MARIN used its hull optimization hull form variations can be generated, as all propeller geometries were analyzed The propeller optimization procedure program PARNASSOS EXPLORER was the case here. It was clear that a tool with MARIN’s Boundary-Element can quickly show the relations between and a newly developed propeller op- that can quickly cover a wide range of Method PROCAL. Although each pro- all the parameters and propulsion ef? -timizer. The hull form itself was opti- design choices and return ballpark ? g- peller was unique, differing in blade ciency for a given wake ? eld of a ship. mized for low resistance and for having ures for ef? ciency, including the propel- shape as well as blade area ratio, diam- But not only did MARIN ? nd an opti-a wake ? eld as rotationally symmetric ler-hull interaction effects, was required. eter or blade number, they all generated mum loading distribution and diameter, as possible, while simultaneously main- When a large number of propeller ge- the same thrust at the same ship speed it also knew how much skew to apply to taining the displacement and suf? cient ometries are to be generated, the range and shaft rate of revolutions by automat- reduce thrust loading ? uctuations. When space for the engine room. Meanwhile, should be as large as possible but cer- ically adapting the effective pitch. This results of the optimization with a given the propeller was required to deliver the tainly not by including every imaginable last requirement avoids the generation of propeller design are compared, it is pos-right thrust at the same ship speed and shape. ‘trivial’ propellers that operate at a lower sible to see how much its ef? ciency can engine rate of revolutions. Therefore, the propeller blade geome- rate of revolutions and nearly always improve and which parameters to tune. Normally, the propeller designer starts try was built up from radial distributions have a better ef? ciency, masking the ef- Now that this ? rst test case has been when the hull lines have been set and the for the pitch, chord length, skew, rake, fect of improved blade design and intro- performed, MARIN is continuing to ex-main propeller parameters have been de- camber and thickness, and these distri- ducing a bias in the comparison. plore the automated propeller design. cided upon. butions were described by newly devel- Using MARIN’s computer cluster, it The process will be sped up by the use These parameters follow from series oped parametric functions. was possible to generate and analyze of genetic algorithms and integrated into analysis, such as the Wageningen B- In order to remain close to the shape of more than 10,000 new designs per day. the daily design process. The EU project Series, plus hands-on experience and real propellers, MARIN analyzed more The best propeller-hull combination SONIC already presents the next chal-rules of thumb to estimate the effect of than 1,200 unique propeller designs in was tested for its performance by model lenge: optimizing for both high ef? cien-propeller-hull interaction. For instance, its database and used the results to train scale tests: a 7.5% power reduction at cy, as well as low radiated noise. the sector knows that the best propeller the parameters. Not only could the ma- the design speed was identi? ed and the References: Foeth, E.J. & Lafeber, F.H., diameter in a wake ? eld is smaller than jority of propellers be well captured, the shaft rate of revolutions did not change. Systematic propeller optimization using an given by the propeller series. Once the generation of a new blade by randomly This improvement is attributed to a re-unsteady Boundary-Element Method, 12th main dimensions are selected, the de- changing the parameter values nearly duction in resistance, an improvement Int. Sym. Practical Design of Ships and Oth-signer continues to mold the geometry always resulted in a plausible blade out- in propeller-hull interaction by a lower er Floating Structures, Korea, 2013.until the ef? ciency is high and the hin- line. propeller loading due to both the lower 18 Maritime Reporter & Engineering News • MARCH 2014MR #3 (18-25).indd 18 MR #3 (18-25).indd 18 3/5/2014 9:42:29 AM3/5/2014 9:42:29 AM