Modeling, estimation, and control of a piezoelectric actuated fuel injector

Chris A Satkoski, Purdue University

Abstract

Improving tradeoffs between noise, fuel consumption, and emissions in future internal combustion engines will require the development of increasingly flexible fuel injection systems which can deliver more complex injection profiles. Piezoelectric injectors have the ability to deliver multiple, tightly spaced injections in each cycle and continuously variable injection called “rate shaping,” but are highly dynamic systems requiring careful voltage input modulation to achieve sophisticated flow profiles. Closed-loop control could prove to be a key enabler for this technology, but will require on-line estimation of the injected fuel flow rate to be realized because of the inability to measure injector flow rate on an actual engine. This work outlines the development of an 11 state simulation model for a piezoelectric fuel injector and associated driver that can be used for injector design and control system verification. Non-measureable states of the model are plotted and analyzed, while measurable quantities including injection rate, piezo stack voltage, and piezo stack current are validated against experimental injector rig data for two different rail pressures. This thesis also describes the development of a physics based fuel flow estimator which can be used as feedback to a closed-loop controller. Available measurements of piezo stack voltage and rail-to-injector line pressure are used for dynamic state estimation. Estimator results are compared against both open-loop simulation and experimental data for a variety of profiles at different rail pressures, and show improvement, particularly for more complex multi-pulse profiles. Also the error seen in the simulated injected fuel quantity is generally reduced by over 50% with the estimator for almost every pulse. Internal states of the estimator are used to evaluate pulse-to-pulse interaction phenomena that make control of multi-pulse profiles difficult to achieve. A hypothesis for pulse-to-pulse interaction is illustrated by dividing the needle lift vs. fuel flow resistance relationship into regimes, and correlating the pulse-to-pulse behavior to the regime switching that occurs for tightly spaced pulses. Finally, this work summarizes the use of estimation algorithms for cycle-to-cycle determination of an injection flow profile capable of being used as feedback for a closed-loop control algorithm. While the estimation equations are complex and require a small time step, a strategy is proposed which captures important estimation feedback during the injection period and delays the integration of state variables across the engine cycle to more efficiently utilize a real-time processor. Lastly, a simplified model is developed to represent the dynamics of simultaneously controlling the quantity of pulses as well as the realized dwell time in between pulses. The model accounts for the coupling between the two, and a control law is developed and refined to provide an overdamped response of both the pulse quantities and realized dwell time, in order to prevent pulse bleeding. Transient response of the controller is shown in simulation and validated against experimental data with good correlation. Overall, this thesis describes a generalizable method to produce cycle-to-cycle virtual measurements of fuel injection for the purpose of subsequent controller design, which could provide a platform for making substantial improvements in combustion related efficiency and emissions by increasing the flexibility of the fuel injection system.

Degree

M.S.M.E.

Advisors

Shaver, Purdue University.

Subject Area

Mechanical engineering

COinS