Research on prediction algorithm of ship equipment heath condition. (1st April 2022)

• Record Type: Journal Article
• Authors: Chen, Dejun; Tang, Tao; Yao, Yongan
• Is Part Of: Ocean engineering. Volume 249(2022)
• Publication Date: 2022-04-01
• Subjects: MTGNN -- DTW -- Ship equipment correlation -- Prediction operation condition; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.110750
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: The prediction of ship equipment operation status is of great significance to ship equipment maintenance and ship navigation safety. How to use ship equipment operation data to predict ship equipment operation status is the most important part. Ship equipment operation status prediction is divided into equipment operation data feature extraction and prediction. And ship is a highly coupled system, so this article focuses on extracting ship equipment correlation to predict the operating status of ship equipment. In the research of this article, the UCI_CBMNPP dataset and UCI_CMHS dataset are used as the basic data set of this article, UCI_CBMNPP is generated from a complex gas turbine simulator, and UCI_CMHS is obtained through experiments on a hydraulic test bench. Analyze the characteristics of ship equipment operation data through two data sets. The improved dynamic time warping (DTW algorithm is a dynamic programming algorithm to calculate the similarity of two time series.) is used to extract the relationship between devices, this paper changes multivariate time series forecasting with graph neural networks (MTGNN: Multivariate Time Series Forecasting with Graph Neural Networks.) structure to establish predictive model, named ImDTW-MTGNN. Experiments demonstrated that ImDTW-MTGNN has achieved better results on the UCI_CBMNPP and UCI_CMHS. Highlights: Improve the dynamic time warping algorithm to calculate the correlation of ship equipment. Use ship equipment correlation to establish ship equipment correlation diagrams. Study the multi-objective graph neural network prediction model, use it to predicte the ship equipment operating conditions.