Feature-based estimation of steel weight in shipbuilding

Lin, Cheng Kuan; Shaw, Heiu-Jou

doi:10.1016/j.oceaneng.2015.06.013

Cited by 5 publications

(2 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This newly developed method includes several existing methods, such as the cubic number method, the rate per meter method, and the rational equation method, and is recommended by various classification societies. Lin and Shaw [11] presented a featurebased estimation (FSE) method to estimate a ship's steel weight and centre of gravity in the preliminary designing phase. Their approach utilises principal component analysis (PCA) to identify principal parameters from the ship's parameters and its main structural components as well as to develop equations for estimating steel weight.…”

Section: Introductionmentioning

confidence: 99%

Estimation of Tanker Ships’ Lightship Displacement Using Multiple Linear Regression and XGBoost Machine Learning

Frančić

Hasanspahić

Mandušić

et al. 2023

JMSE

View full text Add to dashboard Cite

It is of the utmost importance to accurately estimate different ships’ weights during their design stages. Additionally, lightship displacement (LD) data are not always easily accessible to shipping stakeholders, while other ships’ dimensions are within hand’s reach (for example, through data from the online Automatic Identification System (AIS)). Therefore, determining lightship displacement might be a difficult task, and it is traditionally performed with the help of mathematical equations developed by shipbuilders. Distinct from the traditional approach, this study offers the possibility of employing machine learning methods to estimate lightship displacement weight as accurately as possible. This paper estimates oil tankers’ lightship displacement using two ships’ dimensions, length overall, and breadth. The dimensions of oil tanker ships were collected from the INTERTANKO Chartering Questionnaire Q88, available online, and, because of similar block coefficients, all tanker sizes were used for estimation. Furthermore, multiple linear regression and extreme gradient boosting (XGBoost) machine learning methods were utilised to estimate lightship displacement. Results show that XGBoost and multiple linear regression machine learning methods provide similar results, and both could be powerful tools for estimating the lightship displacement of all types of ships.

show abstract

Section: Introductionmentioning

confidence: 99%

Estimation of Tanker Ships’ Lightship Displacement Using Multiple Linear Regression and XGBoost Machine Learning

Frančić

Hasanspahić

Mandušić

et al. 2023

JMSE

View full text Add to dashboard Cite

show abstract

“…Because it requires more information due to the inclusion of the bulkheads, this approach is better suited to cases in which the general arrangement and subdivisions are already approximately known. In another study, a feature-based method based on principal component analysis (PCA) was presented to analyze and estimate the segment weight of ships [16]. We adopt this concept aiming to design a new parametric method that divides the main hull into several estimated sections.…”

Section: Introductionmentioning

confidence: 99%

Preliminary parametric estimation of steel weight for new ships

Lin

Shaw

2015

J Mar Sci Technol

Self Cite

View full text Add to dashboard Cite

This paper presents an agile method for accurately estimating the steel weight of a new ship in the preliminary design stage. A container ship database within a three-dimensional computer-aided design system was used to facilitate the division of a ship into several estimated structural sections, including the forward, cargo, engine room, aft, etc. This procedure is an improvement over the traditional parametric estimation process, which only estimates a rough weight for an entire ship based on specific empirical formulas and then assigns a weight to each part of the ship manually. To estimate the steel weight of a new ship, the parameters of the individual sections are referenced from a similar vessel automatically selected using the principal dimension. Next, principal component analysis is applied to define the principal parameters, i.e., the principal dimension, the feature of each section and the fitting coefficient of the principal parameters. The general equations with the highest correlative predictor variables for the individual sections are then obtained by regression. Comparing the estimation variability, the section estimation provides a more accurate steel weight and constitutes a new means of efficiently obtaining the weight distribution to support the various aspects of preliminary design, i.e., the determination of the center of gravity, the design of the ship lines and ship performance evaluation.

show abstract