Alternative Empirical Formula for Predicting the Frictional Drag Penalty due to Fouling on the Ship Hull using the Design of Experiments (DOE) Method

Biofouling is known as one of the main problems in the maritime sector because it can increase the surface roughness of the ship’s hull, which will increase the hull’s frictional resistance  and consequently, the ship’s fuel consumption and emissions. It is thus important to re...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Muhammad Luqman Hakim, Bagus Nugroho, I Ketut Suastika, I Ketut Aria Pria Utama
Formato: article
Lenguaje:EN
Publicado: Universitas Indonesia 2021
Materias:
T
Acceso en línea:https://doaj.org/article/014fb52bf30247debe24a97af0b25414
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Descripción
Sumario:Biofouling is known as one of the main problems in the maritime sector because it can increase the surface roughness of the ship’s hull, which will increase the hull’s frictional resistance  and consequently, the ship’s fuel consumption and emissions. It is thus important to reduce the impact of biofouling by predicting the value of . Such prediction using existing empirical methods is still a challenge today, however. Granville’s similarity law scaling method can predict accurately because it can be adjusted for all types of roughness using the roughness function  variable as the input, but it requires iterative calculations using a computer, which is difficult for untrained people. Other empirical methods are more practical to use but are less flexible because they use only one  input. The variance of  is very important to represent the biofouling roughness that grew randomly. This paper proposes an alternative formula for predicting the value of  that is more practical and flexible using the modern statistical method, the Design of Experiments (DOE), particularly two-level full factorial design. For each factor, the code translation method using nonlinear regression combined with optimization of constants was utilized. The alternative formula was successfully created and subjected to a validation test. Its error, calculated against the result of the Granville method, had a coefficient of determination R2= 0.9988 and an error rate of ±7%, which can even become ±5% based on 93.9% of 1,000 random calculations.