Model Order Reduction and Control of an Organic Rankine Cycle Waste Heat Recovery System

As fuel efficiency and emissions requirements continue to rise, auto manufacturers are continuously striving to adopt new technologies to help reach these goals.With methods such as turbocharging, direct fuel injection, variable valve actuation, and engine stop-start now common in mass production vehicles, the next step forward could be in waste heat recovery.In most vehicles today, more than 60 percent of fuel energy is lost to waste heat in the cooling system and exhaust.The higher temperature heat energy in the exhaust can be recovered using an Organic Rankine Cycle (ORC).Past research on ORC's focused on creating highly detailed models for performance prediction or controlling extremely simple models.Neither of these options are ideal for use in operating a real system.The detailed model is too slow and the controls based on the simple model are not accurate enough to predict what the real system will do.This thesis takes a highly detailed model and uses model order reduction to create a reduced order model which retains most of the prediction accuracy of the full model but is now smaller and faster.This new reduced model has been used with feedforward and feedback controls, but it also has the potential to be used in advanced model based controls such as model predictive control (MPC).