Optimization of fuel efficiency and control of diesel electric marine vessels subject to data communication constraints
Contact: Professor Tor Arne Johansen
Power Management Systems are crucial control and information systems for both fuel-efficient and reliable power generation and distribution on diesel electric vessels as well as stand-alone offshore structures. For dynamically positioned vessels one important task of the Power Management System is to maintain operational safety by avoiding uncontrolled vessel drift-off due to full or partial power blackout. Automatic start and stop of engines and generators, fast automatic load reduction and heavy consumer handling aims to maintain sufficient power generation capacity subject to variations in loads, the possibility of a sudden generator or switchboard failure, while at the same time minimizes fuel consumption and emissions by operating the diesels engines and gas turbines at high efficiencies. Thruster and propulsion systems can be operated in different modes in order to minimize power consumption and tear and wear of the mechanical components of the thruster and propulsion systems.
Integrated vessel control systems implementing the above mentioned functionality are enabled by advanced communication and information structures, supporting automatic and manual decision making at different time scales. Fuel consumption can typically be optimized on scales from minutes to hours. The vessel motion is optimized on scales from second to minutes. In order to avoid power blackout in case of sudden loss of power consumers or generators, or sudden power demands, response times on the scale of milliseconds to seconds are usually required.
Today, such systems are realized by a range of different controllers with limited communication and coordination, each performing separate functions. Communication solutions range from industrial Ethernet on the high level solutions, to Fieldbus networks at intermediate levels, and dedicated drive bus and hardwired communication solutions at the lowest and fastest levels. As the technical complexity and high degree of redundancy and operational flexibility of the power plants keep growing together with even stronger requirements for safety, reliability, availability and fuel-efficiency, there is an interest to look into more systematic methods for implementing highly integrated Power Management Systems. The proposed project will develop a study on the use of model predictive control and real-time optimization of the power plant operation as a method to implement integrated Power Management Systems. Due to the combination of analog and digital decision variables, this will be implemented by numerical mixed-integer programming. Since data must be updated, and instruments sampled, at different frequencies the approach is based on information structures and a computational control topology that supports the data communication constraints of a communication network and fieldbus topology.
The objective is to develop an optimization based Power Management System strategy for improved fuel efficiency and safety that supports typical automation network and fieldbus communication topologies and real-time information processing requirement. The study will be based on realistic computer simulations. The work will build on the following references