Nonlinear fault-accommodation thrust allocation for over-activated vessels using artificial neural network and multivariate analysis

This study seeks to develop a better understanding of the thrust allocation process of over-actuated marine vessels. A two-stage analysis framework is proposed for thrust allocation when thrusters experience faults. In Stage-I, sequential quadratic programming with a multi-start mechanism (SQP-MultiStart) is used to find the optimal solution to the nonlinear fault-accommodation thrust allocation problem. According to the solution feasibility, another two different optimization models are considered in Stage-II. A newly developed meta-heuristic optimization approach, neural network algorithm (NNA), and SQP-MultiStart method are employed to find optimal solutions. Histories of decision variables and thrust allocation status are obtained. Self-organization mapping (SOM) and principal component analysis (PCA) are utilized to mine the hidden information in the thrust allocation. A deepwater pipe-laying crane vessel with 7 azimuth thrusters is selected to illustrate the analysis procedure. The features of thrust allocation are investigated and the effectiveness of the proposed framework is verified. Simulation results extend the understanding of the thrust allocation process and assist in choosing the thrust allocation model in different cases. The framework is suitable for fault-free and faulty cases. Neural network technique and multivariate analysis help to mine the knowledge about the thrust allocation performances.

• Record URL:
  https://doi.org/10.1016/j.oceaneng.2022.112936
• Record URL:
  http://www.sciencedirect.com/science/article/pii/S0029801822022193
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/issn/00298018
• Supplemental Notes:
  • © 2022 Elsevier Ltd. All rights reserved. Abstract reprinted with permission of Elsevier.
• Authors:
  • Xuebin, Li
  • Luchun, Yang
• Publication Date: 2022-12-15

Language

• English

Media Info

• Media Type: Web
• Features: Figures; References; Tables;
• Pagination: 112936
• Serial:
  • Ocean Engineering
  • Volume: 266
  • Issue Number: 0
  • Publisher: Pergamon
  • ISSN: 0029-8018
  • EISSN: 1873-5258
  • Serial URL: http://www.sciencedirect.com/science/journal/00298018

Subject/Index Terms

• TRT Terms: Algorithms; Fault monitoring; Floating cranes; Multivariate analysis; Neural networks; Thrust
• Subject Areas: Marine Transportation; Vehicles and Equipment;

Filing Info

• Accession Number: 01864557
• Record Type: Publication
• Files: TRIS
• Created Date: Nov 21 2022 4:21PM