2 Quick Guidance For Ship Trim Optimization - Based On Machine Learning Method
2 Quick Guidance For Ship Trim Optimization - Based On Machine Learning Method
Based on Machine
2017/11/24 15:15 115.0 18.0 11.6 1.1
2017/11/24 15:30 117.6 18.1 11.6 1.1
2017/11/24 15:45 120.7 18.7 11.6 1.0 ●● Click the button in the
2017/11/24 16:00 120.2 18.7 11.6 1.0 Data Management Window, the
2017/11/24 16:15 116.0 18.2 11.6 1.1
Learning Method
2017/11/24 16:30 115.5 17.9 11.5 1.2 three-dimensional relation among
2017/11/24 16:45 120.9 18.6 11.5 1.1
2017/11/24 17:00 121.0 18.7 11.5 1.1 fuel consumption of main engine,
2017/11/24 17:15 121.2 18.7 11.6 1.1
2017/11/24 17:30 122.8 18.8 11.5 1.1 trim and speed at given mean draft
2017/11/24 17:45 124.1 18.8 11.5 1.1
2017/11/24 18:00 121.0 18.8 11.5 1.1
can be clearly shown.
2017/11/24 18:15 121.1 18.9 11.5 1.0
2017/11/24 18:30 121.3 18.9 11.5 1.1
2017/11/24 18:45 122.8 18.9 11.5 1.1
This publication was produced by MTCC-Asia with the financial assistance of the European Union. The
contents of this publication are the sole responsibility of MTCC-Asia and can in no way be taken to reflect the
views of the European Union.
Optimum trim Displacement D(t)
Method of acquiring past navigational basic data of the ship
22000 4 visual area
Speed:15 kn
According to the principle of hydrodynamics, the Start
20000
ship resistance is related to the displacement (draft),
yes 2 parameters input area
speed and trim provided that the external environ- 18000
Electronic data
ment is not taken into account. In theory, there is a recording system
16000
trim which minimizes the ship resistance, the main no
14000
engine power and the fuel consumption at the same Optimum
yes
speed. This theoretical trim is called the optimum 12000 Reasonable Noon report
Establish electronic
trim for this ship when it is sailing in this navigational data recording no 3 value display area
10000
status. It can be divided into static optimal operation system
yes
and dynamic optimal operation. 8000
Common
Sister ship
Interface of a ship trim optimization system
no
change ship structure and install additional devices; Visual area: show the corresponding relationship between fuel consumption and trim. The
It is strongly recommended that the methods should be comprehensively considered as
●● Easy operation. It can be easily achieved by adjusting the fore draft and aft draft lowest point of the curve is the fuel consumption corresponding to the optimum trim.
practicable as possible for the sake of acquiring the past navigational basic data of the ship
manually or automatically; in a prompt and accurate manner.
Precautions of using ship trim optimization system based on machine
●● Little influence. It neither reduces the ship cargo loading capacity and speed nor learning method
Also, it is recommended that the electronic data recording system using sensors to auto-
compromises the ship’s safety of machinery and cargo; and
matically record the navigational basic data of the ship should be appropriately established ●● By no means should the trim optimization operation hinder the ship’s safety and
●● Remarkable effect. It can save 2%-6% energy for ships. for the ships which have yet installed this system on board. pollute the marine environment.
Steps of determining the ship optimum trim through using a machine ●● The ship optimum trim should meet the requirements of the permissible stress and
Ship trim optimization system based on machine learning method
learning method Requirements of training data set the allowable hull strength.
●● Acquire the past ship navigational basic data and set up initial model of the ship The system can provide reliable conclusions that fit the objective rule only if the training ●● While the ship optimum trim is against with the ship safety requirements, the trim
trim optimization; and data reaches an effective amounts and is continually updated. The magnitude of training which is within the safety limits but closest to the optimum trim could be adopted.
data amount is directly related to the adopted algorithm, which can be tested and validated
●● Collect continuously the ship navigational basic data, synchronize data iteration, ●● For static trim optimization operations, the application of the optimum trim should
during the system development process.
and improve the ship trim optimization model. also consider the parameters such as varied wind conditions and water densities in
the various sea areas that the ship is intended to pass during the entire voyage.
Types of ship navigational basic data that need be collected System functions
●● For container ship, the optimum trim should be considered as the containers and
To obtain the ship optimum trim through machine learning method based on big data anal- The user interface of a ship trim optimization system based on machine learning method is
ballast water are handling.
ysis, the ship navigational basic data at least need to be acquired as follows: date, time, normally comprised by three areas: parameters input area, value display area and visual
navigational status, mean draft (or fore draft, aft draft), speed, fuel consumption of main area. ●● For oil tanker, the free surface effect of tanks should be considered during the ad-
engine, and trim. justment of optimum trim.