A Novel Robust Feedback Controller Design Methodology for Exploiting Directional Preferences

Texas A&M Engineering Experiment Station, College Station TX

Investigators

Abstract

The main objective of the proposed work is to develop a new feedback design methodology that takes advantage of directional preferences that are possible in multi-input multi-output systems. Over the last six decades a considerable body of knowledge has accumulated in the design of feedback control systems. In spite of the many developments that have occurred, classical control still occupies a central role in feedback control design. Proposed work will exploit the true advantages offered by classical design. One fundamental difference between SISO and MIMO designs is that there are directional influences in the MIMO problem whereas there is no such influence in the SISO problem. On closer examination of existing design methodologies such as and QFT it is clear that available flexibility with directional properties are not taken advantage of. As a matter of fact all these methodologies essentially enforce performance in all possible directions. The reason for this is that design or synthesis is focussed on a closed loop system transfer function. For example, even when an output disturbance is known to be in a specific direction the sensitivity transfer function matrix is made small in all possible directions although it is not necessary. We propose a new paradigm of directional feedback control of MIMO systems. This research would provide new tools to enable the solution of numerous advanced control problems that may not be solvable with existing methods. For example, with the advent of new control effectors and sensors it is anticipated that very high performance systems will possess highly redundant actuators and sensors that will enable the incorporation of many hitherto not considered performance requirements. High performance aircraft will likely incorporate highly redundant control actuators, may not have vertical tails, may be capable of very high g-maneuvers with large attitude rates and will operate over extended flight envelopes with highly nonlinear aerodynamics. Reconfigurable control will become ever so important. It can be envisioned that reconfigurable controls will allow design and implementation of systems that will perform automatic, on-line optimization of system performance in the event of damage or system failures. The successful resolution of proposed objectives will have an immediate effect on fault diagnosis and fault isolation in machinery and structures, redundancy management and reconfigurable control system design by making available a new set of design tools. Condition based monitoring of complex expensive equipment such as automated manufacturing systems, turbomachinery, and drive trains can improve safety and reliability as well as reduce the staggering O&M costs. The annual cost of poor maintenance to US industries is currently estimated at $50 billion.