Web Application with Data Centric Approach to Ship Powering Prediction using Deep Learning

Prerequisite

Interpreter ver.: python 3.6.15
Container ver.: Docker 20.10.10
Application ver.: v1.0
Dependencies: As per listed in requirements.txt

Introduction

This web application is developed based on the joint work between the Marine Technology Centre , Universiti Teknologi Malaysia (UTM) and the Hiekata Laboratory, University of Tokyo (UTokyo) under the UTM research project "Predictive Model for Ship Design and Configuration Using Machine Learning : Model Development and Validation", project number: Q.J130000.3851.19J88.

The details of the ship powering prediction modelling is first described in the research paper presented in the International Conference on Design and Concurrent Engineering 2021 & Manufacturing System Conference 2021

Adi Maimun, Hiekata Kazuo, Jauhari Khairuddin, Siow Chee Loon and Arifah Ali, : "Estimation of Ship Powering in Preliminary Ship Design Using Graph Theory and Machine Learning Method". In the International Conference on Design and Concurrent Engineering 2021 & Manufacturing System Conference 2021, Sep 2021, Japan

Getting started

Clone the repo by running this command in the terminal:

git clone https://github.com/jtkhair/aiship_webapp

Make sure to run the command git pull (if you already cloned this repo). In the cloned repo directory, run this command in the terminal:

git pull https://github.com/jtkhair/aiship_webapp

Build docker image by running below commands in the terminal (make sure docker is running):

cd aiship_webapp
docker build -t aishipwebapp:1.0 .

Run the docker container by running below command in the terminal:

docker run -d --name aishipwebapp -p 80:80 aishipwebapp:1.0

Go to the link http://0.0.0.0:80 or http://localhost:80 to use the web app
Input dataset *.csv file and click submit to perform the prediction. Note that the *.csv file must follow the set schema: (['ID', 'LWL', 'B', 'T', 'L/B', 'B/T', 'Disp', 'CB', 'Vs'. 'Fn', 'P']) and example data file as in the src/data directory

Input data format, range and description

Parameter	LWL	B	T	L/B	B/T	Disp	CB	Vs	Fn	P
Range	80 - 240	15 - 32	3 - 8	3.5 - 9.0	3.0 - 5.5	2500 - 32000	0.5 - 0.7	14.5 - 30.5	0.20 - 0.40	3000 - 70000

Input data example

ID	LWL	B	T	L/B	B/T	Disp	CB	Vs	Fn	P
INPUT_80_1	80	12.4	4.4	6.45	2.82	1712	0.39	10	0.18	Nan

Acronym

Waterline length in m, LWL
Breadth at waterline in m, B
Draught in m, T
Length-to-breadth ratio, L/B
Breadth-to-draught ratio, B/T
Displacement in t, Disp
Block coefficient, CB
Service speed in kn, Vs
Froude number, Fn
Brake kilowatt power in kW, P

Acknowledgement

This work is funded by the Universiti Teknologi Malaysia (UTM) under the Research University Grant (RUG). Project number: Q.J130000.3851.19J88 and titled: “Predictive Model for Ship Design and Configuration Using Machine Learning - Model Development and Validation”