A statistical model from reliability theory is used to estimate overall engine damage accumulation and failure risk.

We present the design of a thrust vector controller for a space shuttle vehicle with multiple engines. This controller will maintain vehicle trajectory and thrust vector while minimizing risk and damage to each engine and to the propulsion system as a whole by independently controlling the thrust magnitude and exhaust cone gimbal angles of each engine. A statistical model from reliability theory is used to estimate overall engine damage accumulation and failure risk. An intelligent control system framework, which functionally decomposes the control task into a formal hierarchical structure composed of nominally independent task coordinators, is used to design and analyze the control structure for the entire propulsion system. One subtask that appears repeatedly within the hierarchy is that of distributing the required control actions among multiple engines or actuators dynamically in response to changing damage, risk, and status information. We identify this as the control relegation problem and solve it using linear quadratic optimal control techniques with variable weighting matrices. The individual controllers, when integrated into the formal structure, achieve the control objectives while minimizing mission risk and component stress and maximizing operating efficiency. © 1991-2012 IEEE.