CFD Simulations for Ships with Rotating Propeller - Self propulsion, Cavitation & Ship radiated noise - http://www.aukevisser.nl/inter-2/id427.htm NMRI, Tokyo JAPAN N.Sakamoto and H.Kamiirisa 1
Table of Contents 1.Background and objective 2.Overview of CFD solver 3.Test cases 4.Computational setup 5.Results and discussion 6.Concluding remarks 2
1. Background Propulsive performance of commercial vessels EEDI regulations (IMO) over over Design exploration in Hull form Energy saving device Propeller Powering estimation by CFD Complex geometry/physics Grid generation Propeller-hull interaction Class NK Annual Fall Seminar (2012) Kawakita et al. (2012) Japan Shipbuilding Digest (2013) 3
1. Objective: Propulsive performance 1:CFD analyses in Resistance and self-propulsion Propeller open water 2:Validation in Self-propulsion factors (SPFs) Local flow field in model scale 3:Estimate Effect of energy saving device Accuracy of the present methods 4
1. Background Propeller cavitation and noise Marine environmental protection Merchant ship in St.Lawrence at 1 mile (http://ocr.org/portfolio/shipping-noise/) Regulations ICES CRR No.209 (1995) IMO MEPC.1/Circ.833 on 2014. For commercial ships, yet non-mandatory Regulate SPL (ICES) Regulate Kp i (IMO): Kp 1 3kPa, Kp 2 2kPa for C B <0.65 Kp 1 5kPa, Kp 2 3kPa for C B >0.65 Kamiirisa and Goto (2005) ICES 5
1. Objective: Propeller cavitation and noise 1:CFD analyses in Propeller performance Propeller cavitation 2:Validation in Cavitation pattern Near field cavitation noise in model and full scale Feasibility of empirical formula with CFD 3:Understand Present capability of CFD to ship radiated noise 6
2. Overview of CFD solver STAR-CCM+ 10.06 (double-precision ver.) k-ω SST (all turbulent), DES for cavitation in fine grid - Low Rn near wall treatment (y+~1) Cavitation model by Schnerr and Sauer (2001) - NO hydrostatic component at this time VoF for interface capturing Overset for propeller rotation SIMPLE for v-p coupling MPI parallelization 2 nd order in space Propeller rotation in 3deg(self-propulsion), 1deg(cavitation) per time-step. 7
3. Test cases: Propulsive performance Japan Bulk Carrier (JBC) Energy saving duct at stern Resistance and self-propulsion data Local flow measurement by PIV without/with propeller rotation! Geometries and exp. data open to public at http://www.t2015.nmri.go.jp/jbc.html Hull (model scale) L pp (m) 7.0 B (m) 1.125 d (m) 0.4125 C B 0.858 Propeller (MPNO.687) D p (m) 0.203 ae 0.500 P/D 0.750 Z 5 Blade section AU 8
3. Test cases: Propeller cavitation and noise Propellers for TS Seiun-1 st CP HSP2 D p (m) 0.221 0.220 ae 0.650 0.700 P/D 0.950 0.944 Z 5 Blade section Mod. MAU Mod. SRI-B - Decreased pitch - Tip-unloaded HSP2 retrofitted after CP to reduce hull vibration Both model and full scale measurement data available. 9
4. Computational setup: Propulsive performance Flow condition (based on CFDWS Tokyo 2015) (Fn, Rn)=(0.0, 7.46E+06), n p (rps)=7.8(w.o. duct), 7.5(with duct) Grid trimmed cell, overset linear scheme for calculating interpolation coefficients from donors to receptors. flux correction activated. Stator(hull) 2.02M Rotator(prop.) 3.95M Total 5.97M 10
4. Computational setup: Propeller cavitation and noise Flow condition (based on Kudo et al. 1989) n p (rps) Rn(kempf) K T σ n CP 6.5E+05 0.207 3.06 20.0 HSP2 7.0E+05 0.201 2.99 Grid (sliding/overset) Local refinement to resolve sheet and tip-vortex cavitation Approx. 19M cells in total. 0.7D p (Hasuike et al. 2010) 11
5. Results: Propulsive performance Effect of the duct to resistance and SPFs C tm x10-3 Exp. CFD w.o. duct 4.289 4.160 with duct 4.263 4.131 Relative difference very well predicted! Contribute for further design explorations. K T -identity method for self-propulsion analysis 12
5. Results: Propulsive performance Thrust distribution (θ=0 o, 24 o, 48 o ), back side Pitch distribution can be increased up to r/r~0.6 for the propeller WITH duct. n p may decrease yet yields the same thrust. 13
5. Results: Propulsive performance Vortical structure Tip and hub vortices are the same. Flow separation at the bottom of the duct wake gain valid in full scale? Difference in vortices behind the propeller effect of n p? 14
5. Results: Propeller cavitation Cavitation pattern (Void fraction=0.1) Nuclei radius=1.0e-6(m) Nuclei density=1.0e+14(1/m 3 ) Click to animate Sheet cavitation well resolved, TVC needs more resolution. Phase & amplitude difference in K p and V c due to skew. 15
5. Results: Propeller cavitation Validation in cavitation extent port starboard port starboard θ=0 θ=20 θ=40 θ=60 θ=0 θ=20 θ=40 θ=60 Exp.data: Kurobe et al. (1983) 16
5. Results: Propeller cavitation noise Overview of the validation data (SR183 Final report) B&K 8103 Levkovskii s scaling law (model to full) Wire mesh wake in model scale n=163rpm in full scale σ ns = σ nm Geo-sym hydrophone location B&K 8103 Brown s formula for empirical method K: empirical const. (=163) B: # of blade, Dp: prop.dia., A D : blade area Compute Ac/A D by CFD! 17
5. Results: Propeller cavitation noise Direct estimation from CFD solution Model scale Full scale Tonal noise well resolved. Relative difference between CP and HSP2 captured well in tonal noise. ( Wake and sheet cavitation patterns are well reproduced.) Broadband noise are fair. ( Bubble growth and collapse cannot be well resolved.) 18
5. Results: Propeller cavitation noise Indirect estimation from CFD(model scale)+brown s formula Brown s formula gives appropriate upper-bound. ( A c /A D well predicted.) 19
6. Concluding remarks Propulsive performance Still needs diagnostics in POT (both exp. and CFD) Relative difference in SPFs very well predicted with and without duct configurations. Local physics helps for design exploration. Propeller cavitation noise Sheet cavitation pattern well predicted. Quantification of cavitation extent feasible by CFD. Tonal noise well predicted. Broadband noise fair, yet empirical formula is still useful together with CFD. 20