Crest of the Royal National Institution of Naval Architects - Click to return to the homepage

The Royal Institute of Naval Architects

Jotun 09/11/2016

Bentley Systems 10/11/16

Anmax and LR team up in CFD scaling project

Naval Architect: May 2016

The value of ship scale CFD to ship designers and ship owners goes beyond predicting accurate performance at the conceptual stage. It can be used to help identify causes of poor performance in existing vessels, predict the effectiveness of energy saving measures that improve the hydrodynamics, and simultaneously account for aerodynamic flows around the vessel.

Accurate ship scale flow modelling can also assist in failure investigations such as propeller cavitation and other sources of ship noise and vibration. Undoubtedly, some of these investigations can be performed through model scale testing; however a number of assumptions are often required due to the inability to match all relevant scales, such as Reynolds and Froude, simultaneously.

Software developments and the decrease in price of computational power have made ship scale simulations widely available. Despite this, as mentioned for example by CD-Adapco’s marine director, Dr. Dejan Radosavljević (The Naval Architect July/August 2015), the uptake by the industry is not high, especially among ship owners. One of the reasons for this reluctance is the lack of comprehensive validation that would give both the users and the ship owners confidence in the methods.

Due to the well-established practices of conducting model tests in a towing tank, there are a number of public domain cases for model scale CFD validation (such as the well-known containership and VLCC tested by KRISO). However, there is a severe shortage of fully documented cases for ship scale CFD validation available in the public domain. This is due to the high cost of the trials required and the challenges of obtaining permission from the stakeholders to publish the hull and propeller geometries.

Lloyd’s Register (LR) has a number of loyal clients who not only class their ships with LR, but are willing to contribute to the development of the marine industry by participating in challenging research projects. One of these clients, the Anmax Trading Corporation, has recently provided the LR Technical Investigations Department (TID) with access to one of its ships to perform in-service performance measurements for the purpose of ship scale benchmarking that can be shared with the marine research community.

The subject vessel, Regal, is an ideal candidate for the ship scale CFD benchmarking as it is a single screw general cargo vessel (150m x 23m x 8m) with a simple configuration: no thrusters, no bulbous bulb and no energy saving devices. The vessel is equipped with a four bladed fixed pitch propeller and a semi spade rudder.

The objective of the trials was to collect information related to ship hydrodynamics that could be further used for ship scale CFD simulation validation of self-propulsion and cavitation in deep and shallow water. Apart from ship performance characteristics such as speed, heading, slip angle, propeller speed, propeller thrust and torque, the environmental conditions defined by wave height and direction, wind and current were also recorded. The trials were carried out on the Black Sea in the summer of 2015 to minimise the risk of harsh weather conditions.

Prior to the trials, the vessel was dry-docked for hull cleaning and the propeller surface was polished. The hull, rudder and propeller were then 3D laser scanned to get an accurate geometric representation of the in-service geometry. While the vessel was in drydock, LR TID specialists installed performance measuring equipment onboard. All the equipment was checked and calibrated in the TID laboratory prior to installation. The following were installed onboard:

  • Shaft optical sensor to record Propeller RPM
  • Independent GPS unit to record vessel speed, position and slip angle
  • Full bridge in one strain gauge on the shaft line to record Propeller Torque 1
  • Full bridge in one strain gauge on the same shaft line to record Propeller Torque 2 (independent to Torque 1 in order to cross check with Torque 1)
  • Full bridge in two strain gauges on the shaft line to record Propeller Thrust 1
  • Full bridge in two strain gauges on the same shaft to record Propeller Thrust 2 (independent to Thrust 1 in order to cross check with Thrust 1)
  • Acoustic emission on the shaft ­ two independent sets to record acoustic signal from cavitation events on the propeller
  • Acoustic emission on the bearing house ­ two independent sets
  • Two borescopes (one on port side, one on starboard side) with high speed cameras to observe and video cavitation events on the propeller.
  • Draw wire sensor to independently record Rudder angle
  • Two independent sensors to measure the vessel’s dynamic trim
  • Two independent sensors to measure the vessel’s Roll
  • GoPro cameras to record the environmental conditions

For the borescope installation, which is used to observe cavitation, the shell plate in the engine room was drilled and tapped at specific locations facilitating port and starboard views of the propeller. Video recordings were made on starboard and port side simultaneously with two high-speed video cameras, which were set to 200 frames per second and mounted on a borescope.

High speed video data was synchronised with acoustic emissions (AE) signals measured on the shaft bearing houses and the shaft line using four 150kHz resonant AE sensors. The signal was sampled at 10kHz with National Instruments “cDAQ” technology. The AE data provides a relative, quantitative assessment of the aggressiveness of energy imparted to the propeller surface during the growth and collapse phases of cavitation development. This technique was developed by LR TID and has been proved to be an effective technique on over 20 projects in the past 10 years.

Due to the size of the data file generated by the high video frame rate, video and AE signal recordings were limited to 20 seconds for each section. During data analysis and post-processing, the AE signals and high-speed videos were synchronised using MATLAB. As thrust and torque measurements do not require high frequency output, the signal was sampled at 2kHz and was recorded during all runs.

In order to record the shaft speed, a reflecting tape was glued on the shaft line and an optical sensor was mounted close to the shaft. Each time the tape passed the sensor window a voltage signal was transmitted to a recording device. For ease of reference, the tape was fitted to the shaft in line with the sensor window when the key blade was located in the top dead centre, resulting in a simultaneous recording of the key blade position.

The propeller thrust and torque were measured by strain gauges similar to those used in Formula 1 cars for various measurements, including the wheel axis moment. The installation of the gauges on the shaft was the most difficult task and required careful treatment due to gauge sensitivity. Before each gauge was fitted, the corresponding area of the shaft was manually polished with sand paper and longitudinal and transversal axes were drawn. Before and after the installation, all strain gauges were tested by the strain indicator and the gauges out of balance were identified. The amplifier and battery pack were also mounted to the shaft (Figure 5). Six batteries supplied power to the shaft amplifier for the duration of the sea trials. The aerial was glued around the shaft and the signal pickup was mounted close to the aerial so that the amplified signal could be transmitted to the recording software.

The measurement of the propeller torque requires only one full-bridged strain gauge and the values are usually attained to a high degree of accuracy. The torsional deformation of the shaft (usually on the order of a hundred micro strains) is generally much higher than equipment tolerances (usually 10 micro strains). The propeller thrust on the other hand is the most challenging value to measure, as the longitudinal deformation of the shaft is in the same order of magnitude as the equipment tolerances. Hence, even careful installation of the thrust strain gauges does not guarantee a high accuracy of measurements. In order to build a full bridge for the thrust measurements, two strain gauges were glued on the opposite sides of the shaft and a heating compensator was applied.

Before commencement of thrust and torque measurements, it is necessary to determine the zero level when the shaft is stationary and there are no forces or moments acting on the shaft. Theoretically, this can be done after the installation whilst the ship is alongside. However, the zero level on the “cold” shaft will not necessarily be the same as on the “warmed” shaft during the trials, hence the heating impact would be ignored. Stopping the shaft at sea for a short time to determine the zero level of a heated shaft is also impractical, as the ship movement due to the inertia would even create forces and moments on the stopped shaft. The practical solution, which was applied in the current case, was to record zero thrust and torque after about 30 minutes when the vessel was stopped at sea. The recorded strain values were then post-processed in order to obtain the corrected values for the thrust and torque.

Since redundant strain gauges were installed on the shaft, it was possible to cross check the measured values between them. As stated earlier, the torque results are usually the most reliable and the cross check confirmed this, revealing an acceptable maximum difference of 5% from two independent torque sensors. For the thrust cross check the difference from two independent installations was larger so the thrust values should be used indicatively.

Apart from the strain gauges, a draw wire sensor was mounted on the rudder stock to measure the rudder angles. After the installation, and whilst the ship was alongside, the rudder was applied in the range of 20⁰ port to 20⁰ starboard, in 5⁰ steps in order to calibrate the sensor and to determine the dependence of the wire displacement on the rudder angle. After calibration, the resulting coefficient was applied in the recording software in order to acquire the signal directly in degrees.

During the periods of the trial when the vessel was stopped, the draughts (forward, aft and middle on port and starboard sides) were recorded by a camera, water temperature was measured and water samples taken in order to keep the record of water density.

The speed tests were conducted under various power conditions at ballast draught at deep and shallow water. The aforementioned quantities were measured for all tests. The weather conditions recorded during the trial show reasonably calm conditions, which helps to minimise the uncertainties related to environmental impact. The ship speed was recorded for all double runs by an independent GPS system installed on the navigation bridge.

The trial has resulted in a collection of ship scale data that defines the vessel speed/power relationship, dynamic trim, and cavitation behaviour including acoustic response for a range of environmental conditions and manoeuvres. Together with the high resolution hull and propeller geometries, this data forms a valuable resource for validation and developing the science of naval architecture.

LR in cooperation with Anmax Trading Corporation is now proposing to raise the level of trust for ship scale CFD modelling in the marine industry by releasing some of this data to the community through an international workshop.

Participants in the workshop will be provided with the hull and propeller geometries as well as the sea trial conditions. They will be asked to submit CFD simulations of the ship scale vessel which will then be shared and discussed at the workshop together with the ship scale trials data.  An outcome of the workshop will be an extensive database of CFD results for one case that highlights the impact of modelling approaches, leading to the development of best practice for ship scale marine CFD. Moreover, it will provide a forum for researchers and commercial practitioners of marine CFD to network and develop ideas for future collaborative work.

The workshop is planned for November 2016 at Lloyd’s Register’s Global Technology Centre in Southampton, UK. Further details will be released on the LR website in the near future.

LR would like to thank Anmax Trading Corporation PTE. LTD shipping for their generous support throughout this project.

Alfa Laval 260x120

Stone Marine 260x120

Rolls Royce Marine 2016

JETS VAC July 2015


Drew Safety