Application of stochastic, multivariable integral control to marine diesel/controllable pitch propeller propulsion systems

This dissertation studies the limit cycle characteristics of a class of existing bridge control systems and the application of stochastic, multivariable integral control theory to develop improved integrated control for marine diesel/controllable pitch propeller propulsion systems. Existing bridge control systems employed in Great Lakes bulk carriers have proven to be difficult to adjust and in many cases they have been expensive to maintain in adjustment. In addition, troublesome continuous oscillations (limit cycles) in engine speed and propeller pitch have been experienced during steady steaming with a constant control handle position. In the main part of the dissertation, the ship propulsion control problem is formulated as multivariable, linear state variable control problem subject to measurement noises and system disturbances. System disturbances, such as the fluctuating part of propeller torque and thrust and the added resistance of the ship in a seaway, are modeled through linear time-invariant systems, or shaping filters, driven by stationary Gaussian white noise. Desired output selection is made to ensure that each engine operating point is safe and near maximum engine efficiency. The multivariable integral control method advanced by Holley and Bryson in their 1973 study is selected and extended. The augmented state is estimated using a Kalman filter. The controller provides zero steady state error to a constant commanded set point with constant disturbances. The controller and filter gains are designed at discrete handle positions and then interpolated as a function of the commanded handle position. Overall system performance in the vicinity of the design points and in response to large commanded changes show promise for an effective and practical ship propulsion control concept.