Physical insights, characteristics and diagnosis of structural freeplay nonlinearity in transonic aeroelastic systems: a system identification based approach

The Next Generation of aircraft sustainment is based on an emerging paradigm known as Prognostics and Health Management. PHM encompasses numerous innovative concepts which shape the future of air asset sustainment according to pre-emptive condition-based maintenance, intelligence-based individual aircraft tracking, and damage/fault prognosis. Smart Diagnostics is an integral component of the SPHM paradigm, and refers to the detection, localisation and tracking of nonlinear structural anomalies that occur in various forms across the airframe structure or within mechanical interfaces. Control surface damage/ failure scenarios, such as, nonlinear hinge stiffness, backlash, and structural freeplay, are a class of structural anomaly which plague modern aircraft and introduce a range of dangerous nonlinear dynamic behaviours, such as, chaotic response and limit cycle oscillation. As a result, the freeplay structural anomaly can reduce fatigue life and is problematic for the stakeholder on many levels, including the management of structural health, maintenance practices, asset availability, mission capability, and sustainment provisions. The traditional approach to handling freeplay-type nonlinear events is based on avoidance and pre-emptive repetitive maintenance practices which, despite being over-conservative, inefficient and expensive, have remained unchanged for more than half a century. As the aerospace sector begins to adopt modern aircraft design and sustainment practices, including the realisation of SPHM-based technologies, there is an urgent requirement for contemporary solutions towards the diagnosis and tracking of structural freeplay anomalies. The research presented in this thesis is pursued with the global objective of contributing towards contemporary structural health monitoring technology through a nonlinear system identification framework for rapid control surface freeplay diagnostics. The proposed framework is driven by the fundamental assumption that all information pertaining to the freeplay event is contained within the time-histories extracted from an aircraft¿s sensory network. It is shown that through careful adaptation of well-established nonlinear system identification methods, namely the Higher-Order Spectra (HOS) and Hilbert-Huang Transform (HHT), rapid detection, localisation and magnitude tracking of the freeplay event is realisable, through a truly data-driven framework, with no inherent dependency of knowledge of the airframe structure, the flight parameters, the aerodynamic condition, or uncertainties. A novel and systematic approach is used to characterise the freeplay event, where nonlinear aeroelastic predictions (numerical aeroelastic models of increasing complexity) are considered to study the isolated physical freeplay mechanism in a nonlinear system identification setting, to understand how its physical action on an aeroelastic system can be exploited for diagnostics purposes. The findings are adapted to formulate temporal and spectral characteristic signatures, then implemented as a basis for the data-driven diagnostics strategy. A flight test case study is used to show that the signature-based diagnostics framework which is formulated using numerical cases with well-defined parameters, remains valid when diagnosing freeplay in a real-world aircraft system. The freeplay is detected and isolated, then a single tuned algorithm is shown to efficiently track the freeplay magnitude over the course of three years with several maintenance/ repair cycles, using a sensor with significant spatial discrepancy to the freeplay source. It is shown that rapid actionable diagnostics information can be extracted with a high level of robustness, demonstrated and verified by making consistent predictions despite: i) a large deviation in Mach number and angle-of-attack (with high angle manoeuvres), ii) highly nonlinear aerodynamic conditions, iii) no knowledge of uncertainty bounds, iv) mixture between stationary nonstationary response, and iv) little information available pertaining to the aircraft structural properties or geometry (a single geometric vector is used). In developing the diagnostics framework, numerous freeplay induced nonlinear phenomena are revisited, providing a new understanding of the structural freeplay physical mechanism. Several freeplay-induced nonlinear phenomena are defined, quantified and related according to a consolidated underlying nonlinear mechanism, founded upon empirically derived correlations. In showing that data-driven signature-based diagnostics is feasible for freeplay, this research makes a significant contribution towards the fields of nonlinear system identification, applied nonlinear dynamics and aircraft structural health monitoring. This provides a clear pathway to extend this signature-based system identification diagnostics strategy to capture other discrete nonlinear mechanisms in aircraft systems, or any relevant mechanical systems across the engineering disciplines. Requirements and limiting aspects of the data-driven approach are thoroughly discussed, predominantly related to sensory network requirements, and recommendations on how to address the limitations and progress with this research are clearly outlined