State estimation of in-flight aircraft centre of gravity

For all types of aircraft (civil, military, manned, unmanned) the aircraft designer specifies a safe range for the centre of gravity (cg) of the aircraft and designs the aircraft to operate safely within these limits. Changes to the cg may affect aircraft stability, performance and fuel economy so it is an important system parameter. Changes to the in-flight cg position have traditionally been estimated by calculating fuel burn and from that calculating the change in weight and hence change in cg. Other techniques to estimate in-flight aircraft cg are included in the literature review. The motivation for additional cg estimation techniques arise from the potential benefits they offer to a Flight Control System (FCS). These benefits include the potential for improved fault detection and an improved FCS design with better aircraft performance and fuel economy. State estimation using Kalman filters has been used since the 1960 s in many fields of application including the aerospace industry. This thesis will introduce the concept of using state estimation to detect the unexpected angular acceleration associated with a cg change. This state estimation concept is applied to a linear Phantom aircraft model and then to a complex non-linear aircraft model of a delta-canard military aircraft, called ADMIRE. The most common state estimation approach used with non-linear systems is the Extended Kalman Filter (EKF), but an alternative approach is proposed in which the pitching and roll moment coefficient derivatives are selectively modified based upon the aircraft angle of attack, speed and altitude. Both longitudinal and lateral cg estimators are described and examples of their performance are provided and compared with an EKF version of the estimator. A discrete version of the estimator is also described and used with a hardware fuel rig. Faults are applied to the fuel rig and it is proposed that the estimator could aid the fault diagnosis. In a real implementation the aircraft will not be precisely modelled, therefore a sequence of robustness tests are included to identify the critical aircraft parameters affecting the estimator. The results show that a cg estimator based upon a Kalman filter, and using a selective coefficient correction approach, can satisfy the performance requirements specified by the industrial sponsor, BAE Systems.