X-01D TVC BWB System Modelling Reference

This technical report provides the complete system modelling reference for the X-01D, a conceptual 237 kg, 6.4 m span blended-wing-body (BWB) unmanned aerial vehicle developed as the common research platform for a series of fault-tolerant control (FTC) papers submitted from the School of Engineering, Monash University Malaysia. The X-01D is derived from the NASA X-48B aerodynamic configuration, with the conventional aerodynamic surfaces and empennage replaced by two overhead-mounted JetCat P200-RX turbojet engines equipped with four-actuator ball-joint thrust-vectoring nozzles. The vehicle has no rudder, no ailerons, no elevons, and no vertical tail. All lateral-directional moment generation is provided exclusively by the four effective thrust vector control (TVC) channels produced by the two gimballed nozzles. The report documents the full derivation of the nonlinear lateral-directional plant model in State-Dependent Coefficient (SDC) form, including the Gamma-formulation for exact Ixz inertia cross-coupling arising from the overhead engine installation. Aerodynamic coefficients are modelled as piecewise polynomial functions of angle of attack, fitted to the NASA X-48B wind tunnel and flight test dataset. The dominant aerodynamic nonlinearity is the sign reversal of the yaw stability derivative Cnbeta at approximately 17 degrees angle of attack, which transforms the open-loop plant from directionally stable to directionally unstable in the high-demand flight regime. This nonlinearity is retained in full in the SDC plant matrix and is the primary motivation for the angle-of-attack-scheduled control designs documented in the companion papers. The TVC geometry is derived from first principles using the measured moment arms of the JetCat P200-RX installation: longitudinal moment arm XTV = -1.81 m and lateral offset YTV = +/-0.64 m. The yaw moment authority at maximum deflection is 297.1 Nm and the roll moment authority is 105.1 Nm. A TVC mechanism selection analysis comparing two-axis gimballed nozzles, jet vanes, fluidic secondary injection, and dual-throat nozzles is included, with the gimballed nozzle selected on the basis of bandwidth, maximum deflection, thrust loss, and retrofit feasibility criteria. Vehicle parameters including the full inertia tensor (Ixx = 548.58 kgm², Izz = 613.54 kgm², Ixz = 38.40 kgm²) are derived from the NASA X-48B parameter estimation database. Model validation is provided through eigenvalue locus analysis across the full angle-of-attack and altitude operating envelope, a nonlinear free-response simulation from an initial sideslip perturbation of 5 degrees with no TVC actuation, and a 100-run Monte Carlo robustness study applying independent +/-15% uniform random perturbations to all seven aerodynamic stability derivatives simultaneously. All 100 Monte Carlo runs produced stable configurations, with the worst-case Dutch Roll real part at -0.132 per second against a nominal value of -0.164 per second. The environmental disturbance model covers three sources: an International Standard Atmosphere (ISO 2533) density model across the 0 to 3000 m service envelope, a steady crosswind model derived from direction cosine matrix transformation, and a Dryden stochastic turbulence model implemented as a second-order state-space filter per MIL-F-8785C and MIL-HDBK-1797B. The lateral and vertical turbulence channels each contribute two filter states to the simulation. The structured fault model defines a six-type taxonomy covering degradation, saturation, stuck-at, complete failure, transport delay, and random erratic faults, formalised through a fault effectiveness matrix F = diag(f1, f2, f3, f4) acting on the four TVC channels. Seven simulation scenarios spanning single, dual, mixed, and worst-case fault conditions are catalogued. A fault severity corridor is derived analytically, defining the controllable operating envelope as a joint function of fault effectiveness, turbulence intensity, and service altitude. The corridor confirms that under severe turbulence at 3000 m with a single-engine fault, TVC authority is sufficient to counteract only 30% of the peak disturbance moment, rendering that condition analytically unrecoverable. The actuator dynamics chain models each TVC channel as a first-order lag (time constant 0.05 s) preceded by a rate limiter at 100 degrees per second and followed by a first-order Pade-approximated transport delay. A sensor corruption model covers additive Gaussian noise, exponential bias drift, and scaling error, combined into a unified measurement equation suitable for use as the observation model of an Unscented Kalman Filter. The full augmented state vector comprises four lateral-directional plant states and four Dryden turbulence filter states, giving an 8-state simulation model. All numerical parameters, polynomial coefficients, aerodynamic derivatives, eigenvalue tables, TVC authority ratios, and simulation settings required to reproduce all results are provided within this report. This report is deposited as an embargoed institutional technical report on Monash University Bridges pending journal acceptance of the companion modelling paper. The embargo protects the novel TVC-embedded SDC plant model, which constitutes the primary intellectual contribution of the companion journal submission, prior to peer review and formal publication. The aerodynamic source data are derived entirely from publicly available NASA X-48B wind tunnel and flight test publications and carry no additional access restrictions. No human subjects, animal subjects, or restricted data were involved in this research. Ethical approval was not required. The simulation model was developed entirely in MATLAB using publicly documented numerical methods and standard aerospace modelling conventions per MIL-HDBK-1797B and MIL-F-8785C.