지도학습을 이용한 전기추진 스마트 선박의 분당회전수 및 축 동력 추정 연구
Prediction of RPM and Shaft Power of an Electric-Propulsion Smart Ship from Data-Driven Analysis Using Supervised Learning
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
IMO has been progressively strengthening regulations aimed at reducing greenhouse gas emissions and improving energy efficiency. As a result, there is a growing demand for methods that can quantitatively evaluate the full-scale performance of ships. Although various approaches have been proposed for estimating the performance, there are limitations in determining the relationships between speed–RPM and speed–power. Moreover, it is necessary to assess the effects of various factors on full-scale performance. In this paper, it is aimed to predict RPM and shaft power of an electric-propulsion smart ship from data-driven analysis using supervised learning. First, post-processing procedure for full-scale measurements is revised to improve the feasibility of deriving speed-RPM and speed-power relationships from the proposed model. Second, design variables for analyzing full-scale measurements based on supervised learning are established. Finally, it is possible to predict RPM and power with improved accuracy.
Abstract
IMO has been progressively strengthening regulations aimed at reducing greenhouse gas emissions and improving energy efficiency. As a result, there is a growing demand for methods that can quantitatively evaluate the full-scale performance of ships. Although various approaches have been proposed for estimating the performance, there are limitations in determining the relationships between speed–RPM and speed–power. Moreover, it is necessary to assess the effects of various factors on full-scale performance. In this paper, it is aimed to predict RPM and shaft power of an electric-propulsion smart ship from data-driven analysis using supervised learning. First, post-processing procedure for full-scale measurements is revised to improve the feasibility of deriving speed-RPM and speed-power relationships from the proposed model. Second, design variables for analyzing full-scale measurements based on supervised learning are established. Finally, it is possible to predict RPM and power with improved accuracy.