Online

Wisotzki, Sam

Electrical and Biomedical Engineering

2019

M.S.

In automakers' never-ending quest to reduce emissions and improve performance, the turbocharger represents a major step in advancing these goals. By repurposing waste exhaust and compressing the air intake, they are able to increase overall power. One critical control loop in the turbocharger is control of boost pressure via the wastegate. This is a highly nonlinear process and experimental data has shown that a gain-scheduled PID (proportional integral derivative) controller developed with IMC (internal model control) tuning methodology is an effective means to control boost pressure. Motivated by this successful implementation of IMC-PID tuning in the automotive world, this work hopes to extend and analyze that framework. Traditionally, the success of an IMC controller depends on the accuracy of the plant model. This research challenges this view and investigates using IMC with a gain-integrator-delay (GID) model identified at a critical frequency, regardless of the actual plant. The GID model is useful because of its simplicity to characterize and its ability to be translated to the ubiquitous PID controller easily. Three design techniques are developed: (1) design for post-hoc tuning, (2) design for closed loop bandwidth, and (3) design for phase margin. In addition, these techniques are investigated via a Monte Carlo simulation to determine efficacy for when there exists plant/model mismatch. Finally, the three techniques are applied to control the speed of an inertia disk on the Quanser Servo 2 device.