PAGE
IN CONSTRUCTIONDr. Hanz Richter, Mechanical Engineering Department, Cleveland State University
PAGE
IN CONSTRUCTION
Projects:
Multiplexed Implementation of Model-Predictive Control for Aircraft Engines
In this NASA-sponsored project, we investigate ways to reduce the computational complexity of model predictive control, so that implementation is feasible in real-time, using aircraft onboard processing. We demonstrated the feasibility of introducing a cyclic sequence for actuator updates within the framework of MPC. By doing this, the quadratic program that must be solved at each sampling interval is reduced to a one-degree-of-freedom search (corresponding to the actuator being updated). Since quadratic programs have a complexity that grows with the cube of the number of controls being optimized, substantial computation savings are obtained. We demonstrated that the speed of computation could be easily doubled without significant performance losses. We provided analytical results concerning the effects of multiplexing in a system, including the use of observers.
Set Invariance Methods for Constrained Variable-Structure Control
This control-theoretical effort incorporates set invariance concepts in the area of variable-structure control, with the aim of designing controllers which guarantee safe operation in the presence of state and control constraints and uncertainties. The theory allows a designer to specify a sliding-mode controller which is guaranteed to keep the system within the boundaries of an invariant set which results from the intersection of a cylinder and the state constraints. A designer computes the control gains and invariant sets by solving a Linear Matrix Inequality (LMI). Current results are being applied to two problems: optimal transfer of liquid containers with slosh constraints and to hybrid controls of rotorcraft drivetrains (NASA Glenn).
Experimental
setup to study liquid transfer with slosh constraints.
The tank is mounted on a controlled linear slide and is fitted with a magnetostrictive sensor to measure liquid level. The problem is to transfer the tank from one position to another as quickly as possible without exceeding a prescribed water level (for instance the edge of the tank).
After modeling the system by a combination of physical principles and system identification techniques, we design a variable-structure controller and apply our set-invariance theory to ensure that the water level and other variables will stay within acceptable limits.
Control of Smart Structures with Limited Hardware
Piezoelectric actuators have been successfully used in vibration control, high-precision positioning and shape control. In systems requiring large numbers of actuators (multi degree-of-freedom space structures), the cost and bulk of the power amplifiers becomes a limitation. In this research we explore the idea of sharing a reduced number of power units among the actuators according to some schedule, for instance multiplexing. Preliminary work suggests that it might be possible, for example, to simultaneously stabilize two cantilever beams with just one amplifier, by following a multiplexing arrangement. Theory has been developed by lab members to design linear controllers that operate under multiplexing and guarantee stable operation. This theory is being applied to an experimental setup consisting of an array of beams to be simultaneously stabilized. This research also extends the multiplexing approach to arbitrary switching among the amplifiers, piezo patches and passive shunts. Hybrid dynamical theory is being used as a framework.
Cantilever beam with piezoelectric actuator and non-contact vibration sensing.
Strain-Sensing with Piezoelectric Biopolymers
In this project, we investigate the potential uses of certain piezoelectric biopolymers as biomedical sensors. Many biopolymers such as collagen, chitin and cellulose are known to exhibit piezoelectric activity. Some of these materials are already being used in biomedical applications for surgical threads and wound dressings, due to their excellent mechanical properties and biocompatibility. In this project we consider exploiting their piezoelectric activity to create biocompatible sensors. We collaborate with the CSU Physics Dept. in depositing thin metallic layers on biopolymer strips to create electrodes.
Attitude Control with Cold-Gas Thrusters
The focus of this project is to develop robust ON-OFF strategies for rotational attitude control of small satellites using cold-gas thrusters. We built a testbed consisting of a rotating platform propelled by Nitrogen from a high-pressure tank. A basic time-fuel optimal law was implemented in the system with good results. Many issues remain regarding the robustness of the control law and the number of firings. According to time-optimal theory, at most two firings should occur in a time-optimal law for the double integrator. But in practice, chattering of the solenoid valves is observed, especially near the target attitude. Existing theory has limitations which make it inapplicable when severe uncertainty is present (Jing and McInnes, Automatica, 2002). We are seeking to develop an improved strategy aimed at the elimination of chattering when severe uncertainties influence system behavior.