ITCNet: long sequence time-series forecasting of underwater excavation states for cutter suction dredgers

Huang, Wenjing, Wei, Changyun, Shen, Kaijun, Liu, Zenghui, Ji, Ze ORCID: https://orcid.org/0000-0002-8968-9902 and Guan, Dawei 2025. ITCNet: long sequence time-series forecasting of underwater excavation states for cutter suction dredgers. Maritime Engineering 10.1680/jmaen.25.00007

Full text not available from this repository.

Abstract

In cutter suction dredger (CSD) operations, direct visual assessment of underwater soil composition and terrain is unavailable to operators. This limitation necessitates reliance on indirect indicators, such as concentration meter readings, interpreted through the operator’s accumulated experience. However, concentration meters are typically installed at the stern of CSDs rather than near the cutter head, introducing significant time delays in the recorded signals. Consequently, these signals fail to provide real-time feedback on the cutter head’s excavation states. To address this challenge, this paper proposes a novel framework based on long sequence time-series forecasting for multi-step prediction of dredged slurry concentration. The framework aims to enable early assessment of underwater excavation states, thereby supporting timely decision making by operators. The study begins by identifying characteristic indicators related to slurry concentration through a detailed analysis of the excavation process. These indicators are then screened to construct a subset of relevant features. Furthermore, a velocity-integrated time compensation method is used to temporally align concentration data with other feature data. The proposed ITCNet model is evaluated against four baseline models using real-world construction data from CSD operations. Results demonstrate that the ITCNet delivers consistent multi-step predictions with minimal errors over a 15 s horizon, providing operators with ample time to respond to dynamic changes. Compared to baseline models, the ITCNet achieves superior accuracy across key metrics, empowering operators to make proactive and informed decisions.

Item Type:	Article
Date Type:	Published Online
Status:	In Press
Schools:	Schools > Engineering
Publisher:	ICE Publishing
ISSN:	1741-7597
Date of Acceptance:	26 April 2025
Last Modified:	01 Jul 2025 09:15
URI:	https://orca.cardiff.ac.uk/id/eprint/179435

Actions (repository staff only)

Edit Item