Elsevier

Applied Ocean Research

Volume 104, November 2020, 102328
Applied Ocean Research

Improved body force propulsion model for ship propeller simulation

https://doi.org/10.1016/j.apor.2020.102328Get rights and content

Highlights

  • Wake flow is taken into account in new body force.

  • Open water results of the different methods are very close and new body force shows higher accuracy than other body forces regardless of flow field or open water results.

  • New body force results for open water and wake flow effect are in good agreement with discretized propeller and experiment.

  • New body-force is reliable and supports rapid and accurate prediction of self-propulsion.

Abstract

Body force propulsion model has been used for ship self-propulsion CFD simulation for a long time due to its higher computational efficiency and less computer source required. This paper presents a new body force method coupled the blade element momentum theory (BEMT) considering the three-dimensional viscous effects with the RANS solvers. In-house code HUST-Ship was used to solve RANS equations with finite difference method and PISO arithmetic for propulsion model and ship hydrodynamic simulation. KP505 propeller and KCS ship model were used as the numerical model for the studies. Open water characteristics of discretized propeller model were obtained after uncertainty analysis of CFD results. The lift coefficients CL and the drag coefficients CD of different radius sections on the propeller blade were obtained from CFD results of open water characteristics. The multivariate regression method was used to get the correlations between CL, CD and propeller parameters for a propeller model. Then the BEMT method would adopt the correlations with the local velocity at the virtual disk from CFD simulation results simultaneously to calculate the thrust and torque distributions on the blade and thrust and torque of the propulsion model were obtained with integration along the blade. The comparisons of open water characteristics of CFD results between three different body force models and discretized propeller model were performed. The CFD simulation for open water characteristics and velocity distribution around propulsion models are almost same for new body force model and discretized propeller model. The self-propulsion simulations of KCS with new body force model were performed. The comparisons of the KCS self-propulsion simulations results with new body force model, discretized propeller model and EFD results showed good matches with each other.

Introduction

It is very important for the ship design stage to predict and assess the self-propulsion performance of various vessels. In addition, CFD (Computational Fluid Dynamics) is a higher fidelity method based on physical principles and the results are typically more accurate with almost no need for empirical inputs (L. Guo et al., 2018). Compared with experiment study, ship self-propulsion free running CFD simulation not only reduces time and cost, but can provide more information of ship flow, ship wave and other details related to the ship hydrodynamic performance. Usually there are two ways to model the propeller during simulating the free running ship: discretized propeller model and body force propulsion model.

Carrica et al. (Carrica et al., 2010; Castro et al., 2011; Chase and Carrica, 2013) and Song et al. (Song et al., 2020) proposed a method using PI speed controller and discretized propellers that could be used to predict self-propulsion performance for KVLCC, ONRT, KCS and DARPA SUBOFF generic submarine hull in model scale. Zhang et al. (Zhang et al., 2020) had analyzed the hydrodynamic performance and structural response of composite propellers using CFD and FEM and then proposed an objective function for selecting the most suitable material attribute scheme for propeller based on existing propeller theories. The flow around a rotating propeller is complex and transient, so the discretized propeller model requires higher mesh resolution and smaller time step to capture the flow features around propeller blades, which leads to complex meshes, more simulation time and cost. The body-force method makes self-propulsion simulation more easily and quickly when the detailed propeller flow is not essential for some studies like interaction of propeller-hull-rudder system. DURANTE et al. (DURANTE et al., 2013) had compared different body-force models to analyze the propeller unsteady loads in an axial, non-uniform inflow. The models were validated through numerical comparisons with propeller loads predicted by a fully 3-D panel-method BEM solver and represented a good trade-off between accuracy and cost. Villa et al. (Villa et al., 2019) proposed a new procedure to extract the self-propulsion coefficients and the method was proven to be reliable in KCS self-propulsion and maneuver conditions. A simplified hybrid method was proposed by Villa et al. (Villa et al., 2018) to address propeller and rudder interactions with reducing computational efforts and these discrepancies of the body force might be accepted due to the lower influence on the overall ship maneuverability performance. Bradford et al. (Knight and Maki, 2019) trained a semi-empirical algorithm to accurately prescribe the unsteady body force to model the propeller and the predictions of the algorithm had a good match with the CFD results. Saettone et al. (Saettone et al., 2020) selected full-scale KVLCC2 propeller to compare the propeller performance in waves using the quasi-steady and the fully unsteady approach.

The body-force methods can usually be classified into descriptive and iterative body force based on the methods of obtaining the body-force source term. The descriptive body force is based on the propeller open water curve and discretizes the thrust and torque of the propeller using a specified distribution which is first put forward by Hough and Ordway (Hough and Ordway, 1964) . Choi et al. (Choi et al., 2010), Kunihide et al. (Ohashi et al., 2018) and Jin et al. (Yuting et al., 2019) had performed propeller-hull interaction simulations using the body-force method. The results obtained through the method showed that prediction of propeller-hull interaction using CFD method was feasible. However, the vessel velocity or modified vessel velocity is usually used as the advance velocity in this method, so the interaction of the propeller and hull isn't under consideration. But in iterative body-force method, the local instantaneous velocity is used as advance velocity in every iteration step so that the virtual propeller and RANS codes can interact with each other towards a solution and the bidirectional coupling between the local flow field and the virtual propeller can be taken into account. This body-force methods generally originate from some potential flow methods, such as lift-line theory and surface singularity potential flow models. KUM (Kyushu University Method) proposed by Yamazaki (Yamazaki, 1977) is a typical potential flow method based on a simplified QBET (quasi-steady blade element theory) with the finite-bladed propeller model. Tokgoz (Tokgoz, 2015) developed viscous flow method OUM (Osaka University Method), which uses infinite-bladed propeller model coupled with RANS based on the idea of KUM. Ernesto Benini (Benini, 2004) proposed CMBET (combined momentum-blade element theory) method which is intrinsically two-dimensional following the methods of Goldstein (Goldstein and Sydney 1929) and Tachmindji and Milan (Tachmindji and Milam, 1957), and explained that the difference between RANS and CMBET is due to the difference between two-dimensional and three-dimensional airfoil. In addition, CL and CD are vital for the methods based on BET. In the KUM, OUM and CMBET, CL and CD are obtained by empirical formulas, two-dimensional data or potential flow methods and can't incorporate viscous effects.

In this paper, the new body force model follows OUM and CMBET, with however, an important difference regarding the acquisition of CL and CD. The basic idea of this method is BEMT (Blade element momentum theory) based on infinite-bladed propeller model coupled with URANS. CL and CD closer to the actual ones will be obtained directly from CFD. Then functions of CL and CD will be used in BEMT to predict thrust and torque. The key objective of this study is to develop a new body-force method which coupled BEMT with the RANS solver which will improve the accuracy and reliability for ship propulsion simulation. KP505 propeller model was used for this study. The lift coefficients CL and the drag coefficients CD of different radius sections on the propeller blade are obtained from the CFD simulation for open water characteristics of the propeller. The multivariate regression method is used to get the fitting curve of CL, CD and propeller parameters. Then the recommended range of the time step and the thickness of the body domain for the new body-force method are given. The reliability of the new body-force method are studied for the open water performance. All simulations are carried out by HUST-Ship RANS solver in this paper. RANS equations are discretized by finite difference method and solved by PISO algorithm. Computations have been made using structured grid with overset technology.

Section snippets

Governing equations

All computations are carried out by HUST-Ship RANS solver in this paper. It solves unsteady incompressible RANS equations:Uixi=0Uit+UjUixj=p^xi+1Re2Uixj2xjui,uj,¯+fbi*In which, Ui=(U,V,W) is Reynolds time-averaged velocity component, p^=(ppρU02+zFr2) is the time-averaged pressure and ui,uj,¯ is Reynolds stress tensor. To simulate the thrust and torque of the discretized propeller, the body-force source term fbi* is added to momentum equation. fbi*=(fbiLPP/ρU02) is dimensionless

The geometry of propellers

The open water characteristics of KP505 are available from Gothenburg 2010 (Gothenburg, 2010). The main parameters of the KP505 model are presented in Table 1.

Computational domain and boundary conditions

Boundary conditions are set according to previous studies (Chen et al., 2018; Wei et al., 2018): The front and rear ends of computational domain along the X axis are set to the inlet and exit boundary condition respectively; The two sides and top are both selected as Farfield #2 (zero pressure gradient) and the bottom is Farfield #1(zero

New body force model

Based on discretized propeller model simulation results, the blade is divided into several parts to get the CL and the CD on each section. In addition, the hydrodynamic performance sensitivity analysis on blade divisions is also discussed for the verification of the numerical method. Then CL and CD will be used for predicting the KT and KQ of the propulsion model based on the BEMT using the relationships between the performance of the propeller sections (CL and CD) and the influencing factors

Simulation of open water characteristic using new body force model

Before comparing the new body force model with discretize propeller model, the uncertainty analysis for the new body force model were conducted.

New body force model with wake flow

In this section, simulation of propeller behind hull using the new body force model would be conducted at first. According to the self-propulsion test data, the incoming flow velocity and rotating speed are set to 2.2 m/s and 9.5 r/s respectively. The results of the new body force model would be compared with the results of the discretized propeller model to verify the effectiveness of the new body force model.

Then self-propulsion simulation using new body-force method for KCS would be

Conclusions

In the present study, a new body force model is developed for propulsion model technology. The feasibility of the new body force model were studied for open water characteristics and wake flow conditions. The reliability of the new body force model was proven in simulations of open water and propeller behind hull. By analyzing the CFD results and comparing with the EFD data. The conclusions are following as:

  • 1

    This method uses the direct coupling of BEMT and RANS so that the iteration loop to

Declaration of Competing Interest

We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and company. We confirm that this work is original and has not been published elsewhere, nor is it currently under consideration for publication elsewhere.

Acknowledgements

This research was sponsored by the Advanced Research Common Technology Project of CHINA CMC (41407010401, 41407020502). The essential support is greatly acknowledged.

References (34)

  • Gothenburg2010CFD Workshop organizers. Gothenburg 2010 Cfd Workshop website....
  • Guo, L., Wei, P., Zhang, Z., Sun, Y., and Yu, J.Numerical simulation of surface ship motion in regular head waves....
  • Guo, L., Yu, J., Chen, J., Jiang, K., and Feng, D.Unsteady viscous cfd simulations of kcs behaviour and performance in...
  • G.R. Hough et al.

    The generalized actuator disk

    Developments in Theoretical and Appl. Mech.

    (1964)
  • International Towing Tank Conference (ITTC) (2008). Uncertainty Analysis in CFD Verification and Validation Methodology...
  • Practical Guidelines for Ship Cfd Applications

    (2011)
  • Li, Z., Yu, J., Feng, D., Jiang, K., and Zhou, Y.Research on the improved body-force method based on viscous flow....
  • Cited by (33)

    View all citing articles on Scopus
    View full text