Optimization of Marine Boilers using Model-based Multivariable Control

During the years marine boilers have traditionally been controlled using classical SISO controllers. To optimize marine boiler performance, reduce new installation time and to minimize the physical dimensions of these large steel constructions a more comprehensive and coherent control strategy is needed. This project deals with applying advanced control to a specific class of marine boilers combining well known design methods for MIMO systems.

To achieve this a model of the boiler system is constructed using first principles. The model is verified using measurements data. Performance requirements are put up and a control strategy is chosen for the specific boiler family.

For the controller design a number of different methods are available e.g. robust and optimal control, MPC and the theory of self-tuning and adaptive control. This project focuses on the theory of MPC possible with robustness demands incorporated. The control system should be invariant to the environment the boiler is placed in and may furthermore not require manual tuning meaning that methods for parameter estimation suitable for the problem are designed. This is to be included in some self-tuning or adaptive control configuration.

The designed controllers will be evaluated by performing tests on full scale boilers.

It is expected that the results from the project provide a scalable model of the specific boiler type. A control strategy is laid out for the family of boilers concerned and a controller is designed.