Reconfigurable Control System Design for Future Life Support Systems 2008-01-1976
A reconfigurable control system is an intelligent control system that detects faults within the system and adjusts its performance automatically to avoid mission failure, save lives, and reduce system maintenance costs. The concept was first successfully demonstrated by NASA between December 1989 and March 1990 on the F-15 flight control system (SRFCS), where software was integrated into the aircraft's digital flight control system to compensate for component loss by reconfiguring the remaining control loop. This was later adopted in the Boeing X-33. Other applications include modular robotics, reconfigurable computing structure, and reconfigurable helicopters.
The motivation of this work is to test such control system designs for future long term space missions, more explicitly, the automation of life support systems. Due to the complexity of the system, a large amount of automation will be required and the corresponding control system will need to perform normally even in the presence of drastic changes in the system dynamics due to abrupt system component failures (sensors or actuators) or rapid change in operating conditions (temperature or energy).
Utilization of such a technique is essential for systems expected to experience severe component failures during exceedingly long missions, tightly limited on resupply. Several pre-programmed control laws, or contingency plans, will be stored onboard for immediate adjustment when failure occurs for anticipated challenges; in addition, artificial intelligent learning techniques can also be utilized for unpredicted scenarios where an automated on-line failure accommodation technique will help to ensure mission success and crew safety.
A conceptual framework for formulating ECLSS design problem utilizing artificial intelligent reconfiguration control system is presented. The paper will focus on the literature review, problem formulation, and design procedures description. Future work regarding the optimality test for determining the benefits and costs of reconfigurable control utilizing the previously developed BioSim modeling tool will also be proposed.