Machine learning approach to ship fuel consumption: A case of container vessel

Uyanık, Tayfun; Karatuğ, Çağlar; Arslanoğlu, Yasin

doi:10.1016/j.trd.2020.102389

Cited by 112 publications

(69 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• the Pearson's and Spearman's correlation analysis showed that from a total of 18 variables in the dataset 14 Based on the conducted investigation, it can be concluded that the GP algorithm can be used for the estimation of CODLAG propulsion system-specific variables. The use of decay state coefficients in symbolic expressions can produce more realistic symbolic expressions which potentially could be used to predict possible performance degradation of the CODLAG propulsion system.…”

Section: Discussionmentioning

confidence: 89%

“…The proposed research achieves an R 2 score of 0.96 in both observed cases. Uyanik et al (2020) [14] proposed an ML approach to the prediction of a container vessel fuel consumption. Through the application of multiple algorithms, such as Multiple Linear Regression, Ridge and LASSO Regression, Support Vector Regression, Tree-Based Algorithms, and Boosting Algorithms are applied and evaluated using R 2 .…”

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Use of Genetic Programming for the Estimation of CODLAG Propulsion System Parameters

Anđelić

Šegota

Lorencin

et al. 2021

JMSE

View full text Add to dashboard Cite

In this paper, the publicly available dataset for the Combined Diesel-Electric and Gas (CODLAG) propulsion system was used to obtain symbolic expressions for estimation of fuel flow, ship speed, starboard propeller torque, port propeller torque, and total propeller torque using genetic programming (GP) algorithm. The dataset consists of 11,934 samples that were divided into training and testing portions in an 80:20 ratio. The training portion of the dataset which consisted of 9548 samples was used to train the GP algorithm to obtain symbolic expressions for estimation of fuel flow, ship speed, starboard propeller, port propeller, and total propeller torque, respectively. After the symbolic expressions were obtained the testing portion of the dataset which consisted of 2386 samples was used to measure estimation performance in terms of coefficient of correlation (R2) and Mean Absolute Error (MAE) metric, respectively. Based on the estimation performance in each case three best symbolic expressions were selected with and without decay state coefficients. From the conducted investigation, the highest R2 and lowest MAE values were achieved with symbolic expressions for the estimation of fuel flow, ship speed, starboard propeller torque, port propeller torque, and total propeller torque without decay state coefficients while symbolic expressions with decay state coefficients have slightly lower estimation performance.

show abstract

Section: Discussionmentioning

confidence: 89%

Section: Literature Reviewmentioning

confidence: 99%

Use of Genetic Programming for the Estimation of CODLAG Propulsion System Parameters

Anđelić

Šegota

Lorencin

et al. 2021

JMSE

View full text Add to dashboard Cite

show abstract

“…They stated that their results may be useful for more accurate prediction of PM 2.5 components in the air. Uyanık et al [9] performed the fuel consumption optimisation of a container ship with machine learning using multiple linear regression, ridge and lasso regression, support vector regression, tree-based algorithms and boosting algorithms. They compared the prediction models in their studies and they found that parameters such as the main engine rpm, cylinder values, scavenge air and shaft indicators are highly correlated with fuel consumption, and stated that they found the most accurate estimate with multiple regression and ridge regression.…”

Section: Introductionmentioning

confidence: 99%

Modelling Ships Main and Auxiliary Engine Powers with Regression-Based Machine Learning Algorithms

Okumuş

Ekmekçioğlu

Kara

2021

Polish Maritime Research

View full text Add to dashboard Cite

Based on data from seven different ship types, this paper provides mathematical relationships that allow us to estimate the main and auxiliary engine power of new ships. With these mathematical relationships we can estimate the power of the engine based on the ship’s length (L), gross tonnage (GT) and age. We developed these approaches using simple linear regression, polynomial regression, K-nearest neighbours (KNN) regression and gradient boosting machine (GBM) regression algorithms. The relationships presented here have a practical application: during the pre-parametric design of new ships, our mathematical relationships can be used to estimate the power of the engines so that more environmentally friendly ships may be built. In addition, with the machine learning methodology, the prediction of the main engine (ME) and auxiliary engine (AE) powers used in the numerical calculation of ship-based emissions provides data for researchers working on emission calculations. We conclude that the GBM regression algorithm provides more accurate solutions to estimate the main and auxiliary engine power of a ship than other algorithms used in the study.

show abstract

“…They concluded that the requested model could reduce ship fuel consumption by 2-7% and the reduction in fuel consumption will also lead to lower CO2 emissions. Uyanık et al (Uyanık, Karatuğ, and Arslanoğlu 2020) studied that the fuel consumption optimization of a container ship with the help of multiple regression, ridge and lasso regression, support vector regression, treebased algorithms and reinforcement algorithms. They compared the prediction models and stated that the predictions made by multiple regression and ridge regression yielded more accurate results.…”

Section: Introductionmentioning

confidence: 99%

Modeling of General Cargo Ship’s Main Engine Powers With Regression Based Machine Learning Algorithms: Comparative Research

Okumuş

Ekmekçioğlu

2021

Mersin University Journal of Maritime Faculty

View full text Add to dashboard Cite

This study, which allows estimating main engine power of new ships based on data from general cargo ships, consists of a series of mathematical relationships. Thanks to these mathematical relationships, it can be predicted main engine power according to length (L), gross tonnage (GT) and age of a general cargo ship. In this study, polynomial regression, K-Nearest Neighbors (KNN) regression and Gradient Boosting Machine (GBM) regression algorithms are used. By this means the relationships presented here, it is aimed to build ships that are environmentally friendly and can be sustained at a lower cost by using the main engine power of the new ships with high accuracy. In addition, the relationships presented here provide validation for computational fluid dynamics (CFDs) and other studies with empirical statements. As a result of the study, polynomial regression gives similar results with other studies in the literature. We also concluded that while KNN regression yields fast results, GBM regression algorithm provides more accurate solutions to estimate the ship's main engine power.

show abstract

Machine learning approach to ship fuel consumption: A case of container vessel

Cited by 112 publications

References 37 publications

Use of Genetic Programming for the Estimation of CODLAG Propulsion System Parameters

Use of Genetic Programming for the Estimation of CODLAG Propulsion System Parameters

Modelling Ships Main and Auxiliary Engine Powers with Regression-Based Machine Learning Algorithms

Modeling of General Cargo Ship’s Main Engine Powers With Regression Based Machine Learning Algorithms: Comparative Research

Contact Info

Product

Resources

About