Skip to content. | Skip to navigation

Sections
A Green Technology Company.

A GL company

Log in


Subscribe to our newsletter



 

Case Study Identification of Main Dimensions

Resistance reduction in the preliminary design of a container vessel

hawse.jpg

 
The preliminary design phase substantially determines the course for a ship design and thereby whether it can become an optimal ship. At this point in time, the main dimensions can be determined freely—as freely and unrestricted as at no later point in the whole design process. This is when the total energy efficiency of the ship can be positively influenced without the strong constraints that usually come about during the course of a project. Prediction methods on the basis of regression analyses facilitate estimates about the total resistance and propulsive efficiency. Fed by only few input variables, the calculations usually last only several seconds per design. Quite often, those methods are available in simple spreadsheets like MS Excel. Sometimes they comprise solely the collected designs of a single yard, while some regression methods contain large data sets from many years of towing tank testing, see e.g. Hollenbach [1]. Naturally, they allow a hindsight on existing vessels only. Combined with formal optimization, they can nonetheless be utilized to identify favorable main dimensions within a mere hour of work.

The FRIENDSHIP-Framework with its advanced technologies for variation and optimization is the perfect tool for this process: the Windows COM-interface supported by the FRIENDSHIP-Framework allows direct access to any MS Excel file. Readily available as an Excel file, the regression analysis by Hollenbach [1], HSVA, was coupled to the FRIENDSHIP-Framework. Starting from an existing baseline, investigation found a reduction of total resistance of 7 percent by adapting the main dimensions. This illustrates the potential and may serve as a promising starting point for a subsequent (and more demanding) simulation-driven design process.

Regression methods for performance prediction usually do not require detailed hull form definitions. However, if available important characteristics such as wetted surface area need not be estimated, but can be directly taken from the existing geometry. It was therefore decided to utilize a fully parametric model in the process, which yields two benefits: the input data are of higher quality; for each variant a full shape definition is readily accessible.

The procedure was divided into the following steps:

1.      Selection of a suitable parametric model from FRIENDSHIP-Framework’s model harbor, in this case a fully-parametric container carrier

2.      Coupling of this fully-parametric model to the regression analysis by Hollenbach via the COM-interface

3.      Determination of the total resistance according to Hollenbach’s resistance prediction

4.      Investigation of a series of variants and identification of most favorable main dimensions


The design conditions for the study were:

-        Variation of LPP in the range of -5% to +10% starting from 153.6m

-        Variation of CP in the range of 0.65 to 0.69 starting from 0.67

-        Variation of the draft in a range of 10.35m to 11.1m starting from 10.6m

-        Volume no less than 29000m³

-        Design speed of 19 knots

 
checkmaindimhollenbach01.png

Fig. 1: COM-interface coupling of the FRIENDSHIP-Framework and MS Excel is realized by means of FRIENDSHIP SYSTEMS’ Feature technology

 

For the analysis, the following input data are needed [1]:

-        Ship type (single screw design draft, single screw ballast draft, twin screw design draft)

-        Length between perpendiculars                          LPP

-        Length of the waterline                                      LWL

-        Molded breadth                                                B

-        Design draught                                                 D

-        Trim                                                                 T

-        Block coefficient                                               CB

-        Design speed                                                    V

-        Wetted surface                                                SHULL

-        The length of the bulbous bow, the number of propellers and the propeller diameter

The regression data is provided within an MS Excel file, which serves as an external computation for formal optimization. The link between the FRIENDSHIP-Framework and MS Excel is realized by a Feature definition that accesses MS Excel via the COM-interface. COM [Component Object Model] is an interface developed by Microsoft and enables the access to all software supporting this standard (including MS Excel). Details are covered in [2]. (For an integration between the FRIENDSHIP-Framework and Hydromax by Formation Design Systems see [2]; this too was realized through the COM-interface, allowing for inclusion of e.g. damage stability calculations.)

 

hlbfeature.png

Fig. 2: Input to the Excel based regression analysis via a Feature, giving ready access to all required data and allowing quick investigations

 

As shown in Fig. 2, the design conditions specify input values of e.g. the length between perpendiculars, breadth, draught and speed. Some input data—like block coefficient or wetted surface—naturally depend on both the main dimensions and the hull shape itself. Therefore, a fully parametric hull model for this type of vessel was used to extract these data items with higher accuracy than through pure regression.

comparevariants.png

Fig. 3: Comparison of the baseline (blue) with the best design from an exploration via Design-of-Experiment, DoE, (green) and the best design from an exploitation by formal optimization (magenta)

 

In order to identify favorable main dimensions, both an exploration and an exploitation were undertaken. The generated variants indicate a prolongation of the length between perpendiculars by about 9 percent to 167.8m. Concurrently, the value for draft decreases by about 2 percent to 10.35 m. This leads, in comparison to the initial ship, to a more slender shape with positive effects on resistance. The slenderness is also mirrored in the reduction of the prismatic coefficient from 0.67 to 0.65. All in all, a reduction of total resistance of about 7 percent was achieved (see Fig 3).

In addition to the estimate for total resistance, the Hollenbach method indicates the range from worst to best expectable performance (see Fig 4). This helps to judge the likely potential for advanced optimizations on the basis of first-principle methods—ultimately, it helps to offer the best possible main dimensions along with the most beneficial lines for a particular trade.

minhlb.png

Fig. 4: Predicted resistance (mean) along with expectations for maximum and minimum performance

 

For more information about the optimization, do not hesitate to get in contact with FRIENDSHIP SYSTEMS.

 

[1]           Hollenbach K.U.: “Weiterentwicklung eines Verfahrens zur Abschätzung von Widerstand und Propulsion von Ein- und Zweischraubenschiffen im Vorentwurf”, Yearbook Schiffbautechnische Gesellschaft Vol.92, 1998

[2]           Abt, C.; Harries, S.; Wunderlich, S.; Zeitz, B.: “Flexible Tool Integration for Simulation-driven Design”, 8th International Conference on Computer and IT Applications in the Maritime Industries (COMPIT’09), Budapest, 10-12 May 2009 

back to main page

Document Actions