Dynamical system modeling of a micro gas turbine engine

Abstract

Since 1995, MIT has been developing the technology for a micro gas turbine engine capable of producing tens of watts of power in a package less than one cubic centimeter in volume. The demo engine developed for this research has low and diabtic component performance and severe heat transfer from the turbine side to the compressor side. The goals of this thesis are developing a dynamical model and providing a simulation platform for predicting the microengine performance and control design, as well as giving an estimate of the microengine behavior under current design. The thesis first analyzes and models the dynamical components of the microengine. Then a nonlinear model, a linearized model, and corresponding simulators are derived, which are valid for estimating both the steady state and transient behavior. Simulations are also performed to estimate the microengine performance, which include steady states, linear properties, transient behavior, and sensor options. A parameter study and investigation of the startup process are also performed. Analysis and simulations show that there is the possibility of increasing turbine inlet temperature with decreasing fuel flow rate in some regions. Because of the severe heat transfer and this turbine inlet temperature trend, the microengine system behaves like a second-order system with low damping and poor linear properties. This increases the possibility of surge, over-temperature and over-speed. This also implies a potentially complex control system. The surge margin at the design point is large, but accelerating directly from minimum speed to 100% speed still causes surge. Investigation of the sensor options shows that temperature sensors have relatively fast response time but give multiple estimates of the engine state. Pressure sensors have relatively slow response time but they change monotonically with the engine state. So the future choice of sensors may be some combinations of the two. For the purpose of feedback control, the system is observable from speed, temperature, or pressure measurements. Parameter studies show that the engine performance doesn't change significantly with changes in either nozzle area or the coefficient relating heat flux to compressor efficiency. It does depend strongly on the coefficient relating heat flux to compressor pressure ratio. The value of the compressor peak efficiency affects the engine operation only when it is inside the range of the engine operation. Finally, parameter studies indicate that, to obtain improved transient behavior with less possibility of surge, over-temperature and over-speed, and to simplify the system analysis and design as well as the design and implementation of control laws, it is desirable to reduce the ratio of rotor mechanical inertia to thermal inertia, e.g. by slowing the thermal dynamics. This can in some cases decouple the dynamics of rotor acceleration and heat transfer. Several methods were shown to improve the startup process: higher start speed, higher start spool temperature, and higher start fuel flow input. Simulations also show that the efficiency gradient affects the transient behavior of the engine significantly, thereby effecting the startup process. Finally, the analysis and modeling methodologies presented in this thesis can be applied to other engines with severe heat transfer. The estimates of the engine performance can serve as a reference of similar engines as well.