The load forecasting for the next few hours can serve as an important reference for ship navigation planning, but the uncertainty of ship navigation load poses significant challenges. Based on ships' navigation data and the hydrological and meteorological information of waterway waters, this article proposes a classification prediction strategy for ship propulsion load and ship service load. Firstly, the Spearman correlation coefficient is used to screen variables with strong correlations with two types of loads as input feature sequences for the model. Subsequently, a combined convolutional gated recurrent unit (Com-ConvGRU) deep network architecture was proposed to classify and predict two types of predicted loads. The model adopts a design architecture of parallel input and feature merging, strictly following the temporal relationship between navigation data and hydrological and meteorological data. The proposed method was validated on a hydrogen-powered ship, and a multi-step load prediction was achieved for the ship in the inland waterway. The experimental results indicate that the proposed method has more accurate load forecasting performance than multiple typical prediction models.
A Com-ConvGRU Model Combined with Classification Feature Selection for Hydrogen Ship Load Prediction
2024-09-06
1691885 byte
Conference paper
Electronic Resource
English
Needle Tip Force Estimation Using an OCT Fiber and a Fused convGRU-CNN Architecture
BASE | 2018
|Ship detention prediction via feature selection scheme and support vector machine (SVM)
Taylor & Francis Verlag | 2022
|SHIP LOAD OBSERVATION APPARATUS, SHIP LOAD OBSERVATION METHOD, AND SHIP LOAD OBSERVATION PROGRAM
European Patent Office | 2023
|