ISC combined optimisation-based predictive control with Machine Learning (ML) improve the control of the power of an fuel cell (FC) powered vessel. The vessel power system model included a set of power sources (FC and battery), systems to be supplied (auxiliary loads and propulsion system), and a command system representing the controller that sets the propulsion power demand. The project demonstrated that combining Linear Parameter-Varying (LPV) - Model Predictive Controller with ML can offer improved control performance with reduced FC degradation and battery stress.

Download Project Note >

ISC was initially contracted to assess the performance attainable for a hydraulically actuated gangway, when compensating agai415518nst boat motions. This involved modelling the various sources of errors - from sensing, actuation and control performance - to achieve the target performance whilst minimising cost and complexity. ISC were then engaged to implement the entire control system, which included MMI using touch panel computers, real-time control and I/O, solving kinematics, and extensive monitoring and safety logic. Full factory testing and Sea Trials were completed. The system was installed and on active service.

Download Presentation
See also Development of Control System for 70-Ton Gripper Arms

Gain scheduled roll stabilisation controllers were optimised using models of vessel and wave dynamics. Our solution was tested on a nonlinear model of the vessel, and out-performed more advanced methods developed by academic groups.