Print Friendly, PDF & Email

The vane wheel consists of two portions, a turbine – designed using BEMT – and a propeller portion – using CFD. This approach is relatively fast and accurate. CFD in full scale indicated even better performance.
IMO has made the EEDI (Energy Efficiency Design Index) mandatory for new ships to reduce GHG (Green House Gas) emissions[ds_preview] from shipping by 30% by 2025. Along with high fuel prices, the regulatory pressure has greatly increased the interest in energy efficiency as such, and in ESD (Energy Saving Device) in particular.

Our interest here is on the vane wheel. Professor Otto Grim developed the vane wheel (Grimsches Leitrad), a freely rotating device located behind the propeller. It consists of two portions, turbine portion and propeller portion. The main function of the vane wheel is to extract energy from the propeller slipstream in the turbine portion and to convert this energy into additional thrust in the propeller portion.

The diameter of the vane wheel is about 20% larger than that of the propeller and the distance between propeller and vane wheel is 25% of the propeller diameter. Typically, the vane wheel has more than six blades and rotates at 30% to 50% of the propeller rpm. It is similar to CRP (contra-rotating propeller) that also recovers some of the rotational energy for the propeller.

The vane wheel has been shown to save 5% and more, making it more efficient than most ESDs. Based on many model tests, Grim asserted that a propeller with vane wheel gains 4.4% to 9.2% depending on the vane wheel diameter. Jörg Blaurock gives up to 10% improvement, depending on the specific thrust loading of the propeller.

However, after some 60 vane wheels were installed, it has not been applied anymore, due to fears that the relatively slender blades might break off. 15% of the installations reported such problems. Particularly, wave loads on the vane wheel resulted in damages to hub and blades.

We applied now CFD (Computational Fluid Dynamics) and BEMT (Blade Element Momentum Theory) to design and analyze a vane wheel. Calculations used the commercial CFD software STAR-CCM+. The turbine portion was designed using BEMT, the propeller portion using CFD in open-water condition. This condition requires less computational time and still has a high degree of accuracy. Model tests on a tanker model measured both thrust and torque of the vane wheel and have uncertainty on measuring in free rotating condition. Therefore, the vane wheel was forced to rotate to satisfy the operating condition (zero torque on the vane wheel).

Cases with and without vane wheel were compared. There is no extrapolation method to full scale available for the vane wheel. Therefore, numerical simulation for full scale predicted the scale effect between model and full scale.

The vane wheel design

The vane wheel should satisfy the following relationships:

dT and dQ are, respectively, thrust and torque produced by the vane wheel. R, rt and r0 are vane wheel radius, transition (turbine-to-propeller portion) radius and the boss radius, respectively. Turbine and propeller portion torques must balance, ignoring the small frictional component due to the low rotational speed, and the net thrust of the vane wheel must be positive.

The vane wheel was designed for the tanker model in design draft at 15.5kn and with 103.9 propeller rpm. The propeller had 7.2m diameter and four blades. There are key parameters for the efficiency of the vane wheel: Rotational speed and diameter. The diameter of the vane wheel is limited by the baseline of the ship and by risk of piercing the water surface in ballast.

The diameter of the preliminary vane wheel was set to 20% larger than the propeller diameter considering limitation of the vane wheel diameter. The radius of the turbine portion should be selected based on the slipstream contraction at the vane wheel, 85% of the vane wheel radius in this case. The vane wheel should have as many blades as possible for good efficiency. Nine blades were selected considering manufacture and efficiency. The rotational speed of the vane wheel was selected at 30% propeller rpm.

The pitch distribution followed the patent by Grim in 1986. The camber distribution of the vane wheel was designed similar to the general design of turbines and propellers. The chord distribution was designed such that it increased in the turbine portion to absorb energy from the propeller slipstream and decreased in the transient position between propeller and turbine portions because this part should neither generate thrust nor absorb energy. Small EAR (Expanded Area Ratio) was chosen to increase efficiency and decrease loads. The thickness distribution was based on strength, cavitation, and blade weight. Blade sections of NACA66 were chosen for low drag and good cavitation behavior. The vane wheel had zero rake and small skew.

The turbine portion was designed using BEMT. This approach is simple yet quite accurate. Turbine portion is then divided into 2D wing sections. Lift and drag coefficients are calculated on these sections. Summing lift and drag forces over the 2D strips gives then blade lift and drag. These were converted into thrust and torque to analyze the performance of the turbine portion.

The propeller portion was designed using CFD in open-water condition, as for single propellers. This design condition requires less computational time than the self-propulsion condition. It is justifiable to design the vane wheel in open water since the vane wheel is mainly influenced by propeller slipstream. Before using CFD extensively in design, we validated the approach against model tests.

Calculated and experimental data differed, but had similar tendencies, e.g. showing negative thrust of the vane wheel. Although CFD results did not match the experimental data, it allowed relative comparisons among new vane wheel designs. In such a way, we designed our final vane wheel using open-water CFD results. In total, five variations of the vane wheel design were investigated. The main change was in the pitch distribution. The open-water efficiency of the vane wheel in CFD was high if the turbine portion was designed having higher power and lower thrust in BEMT. In order to obtain higher efficiency and thrust, the propeller portion was increased. Some further fine-tuning of the propeller portion gave another 1.87% increase in open-water efficiency.

Performance of the vane wheel was finally evaluated in open water and self-propulsion tests in the HMRI (Hyundai Maritime Research Institute) towing tank. Again, the vane wheel was forced to rotate with changing rpm until we had zero torque on the vane wheel. The experimental accuracy is estimated to be 2% for thrust and torque. Full-scale extrapolation was based on CFD simulations for model scale and full scale.

Computations and Experiments

The finite-volume solver Star-CCM+ provided by CD-adapco was employed. The fluid is assumed viscous and incompressible, and the flow turbulent, satisfying the Reynolds-averaged Navier-Stokes (RANS) equations with the Reynolds-stress model for turbulence modelling. The solution domain is subdivided into a finite number of control volumes. In the first step of the study, the required grid resolution was investigated. The final grid consisted of 9.2 million cells for the simulations in model scale and 10 million cells for full-scale simulations. We assumed a single-fluid flow, i.e. double-body model with symmetry at the free surface. Cavitation effects were not considered in the simulations.

To implement the rotating motion of the propeller and vane wheel, a moving mesh approach was used (»sliding interfaces«). The propeller and vane wheel grids were moved independently from each other. The revolution rates of the propeller and vane wheel were set manually and adjusted during simulations; the rpm of the vane wheel was varied until the total moment around its rotation axis was (almost) zero.

Simulations were carried out for design draught and speed, with and without vane wheel. The RANS simulations showed model-scale results, indicating savings due to vane wheel of 1.5% (as opposed to 4.5% measured in the model tests). Open-water tests were performed in HMRI towing tank using CRP dynamometer to measure both thrust and torque.

Thrust coefficient, torque coefficient and open water efficiency with or without the vane wheel based on the model test for different advanced coefficients JA were calculated. At the design point, for JA = 0.398, the open-water was 3.7% higher than for a single propeller. Overall, the predicted values agree well with the measurements. The self-propulsion test was performed on the tanker model using a CRP dynamometer in HMRI towing tank. The propeller rpm was lower and the vane wheel produced additional thrust. Overall, the vane wheel improved efficiency by 4.8% compared to case without the vane wheel, in model scale.

Calculated savings in full scale of 2.9% are greater predicted for model scale. The reason is probably the relatively lower influence of the viscous wake in full scale, together with a better performance of the vane wheel in a uniform flow compared to a non-uniform flow, following from the design and optimization procedure.

Conclusions

The results indicate that correct prediction of the wake field is necessary in simulations to obtain good agreement with model tests. The wake field in model scale seems to be much more sensitive to the employed turbulence model than in full scale. The differences in the wake field between full and model scale might explain the differences in performance of the vane wheel between model and full scale.

The resistance of the rudder is much higher for the vessel with the vane wheel than for the variant without vane wheel. This decreases the overall efficiency gain due to the vane wheel. This needs to be taken into account in design and optimisation of the vane wheel, e.g. by taking rudder in the optimisation loop.

The thrust created by the vane wheel over time is very non-uniform. The vane wheel creates thrust on starboard side of the vessel and drag on port side. The reason for this is the coupled effect of the non-uniform wake and propeller-induced flow rotation. This non-uniform loading of vane wheel blades should be taken into account in design to obtain larger improvements.

Authors: Woo-Chan Seok und

Hyoungsuk Lee, Hyundai Heavy

Industries, Tobias Zorn und

Vladimir Shigunov, DNV GL

vladimir.shigunov@dnvgl.com


Woo-Chan Seok, Hyoungsuk Lee,Tobias Zorn, Vladimir Shigunov