Mean Value Model and Control of a Marine Turbocharged Diesel Engine 2005-01-3889

Use of microprocessor-based controllers in place of traditional governor-based controllers for diesel engines is motivated by the requirement of meeting the increasingly stringent legislations on exhaust emissions and fuel economy. Such controllers can also give improved transient performance. In this paper a fourth order nonlinear mean value model of a 600 HP turbocharged diesel engine is developed for the controller design. Differential equations for various subsystems have been derived using first principles, experimental data and characteristic maps. The model is implemented in MATLAB™ and Simulink™ environment for simulation and controller design. The model is generic and can be modified with a little effort for other heavy-duty turbocharged diesel engines.

The nonlinear model is linearized at sixteen operating points covering wide operating range. These models are reduced to second and first order models using a balanced realization. PI controller is designed using the reduced first order model by the pole placement method. An anti-windup strategy that improves the response during large transients is also presented. The second order linear model is used to design an ITAE optimal PID controller. The optimal controller has been found to substantially reduce overshoot and settling time.