Desarrollo de un modelo CFD - 6DOF para el análisis de la estabilidad lateral – direccional de un UAV de tipo ala volante.
No Thumbnail Available
Date
2026
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
Las aeronaves no tripuladas (UAV) de tipo ala volante presentan ventajas aerodinámicas significativas frente a configuraciones convencionales, pero su ausencia de empenaje compromete su estabilidad lateral-direccional, particularmente en el modo dinámico Dutch Roll, caracterizado por un amortiguamiento casi nulo. Frente a esta problemática, el objetivo de este estudio fue desarrollar un modelo CFD-6DOF validado experimentalmente para analizar la estabilidad dinámica lateral-direccional de un UAV de tipo ala volante.
Para esto, se utilizó como plataforma de referencia un Skywalker X8, instrumentado con un autopiloto Pixhawk 6C, un GPS y un sensor de velocidad de viento, entre otros. Se ejecutaron ensayos de vuelo aplicando una perturbación impulsiva en los alerones, registrando la respuesta transitoria de la aeronave para su posterior caracterización. En paralelo, se implementó un modelo CFD estacionario en ANSYS Fluent, empleando el modelo de turbulencia 𝑘 − 𝜔 𝑆𝑆𝑇, el que fue validado respecto a coeficientes aerodinámicos reportados en la literatura especializada. A partir de este modelo, se caracterizaron las cargas equivalentes generadas por la deflexión de los elevones, las que fueron incorporadas mediante una Función Definida por el Usuario (UDF) en un modelo dinámico de seis grados de libertad (6DOF) con mallado dinámico híbrido, permitiendo simular la respuesta de la aeronave ante la misma perturbación aplicada en vuelo.
La validación estática mostró concordancia satisfactoria para la mayoría de los coeficientes aerodinámicos, con errores RMSE normalizados iguales o inferiores a un 11.9%, aunque se identificaron discrepancias mayores en los coeficientes de momento de roll y yaw, de hasta un 20.4%, atribuibles a incertidumbres en la conversión de momentos respecto al punto de referencia aerodinámico utilizado en la literatura. El modelo dinámico logró reproducir satisfactoriamente los modos Roll y Spiral, así como los parámetros característicos del modo oscilatorio Dutch Roll, con errores RMSE normalizados de las variables de actitud lateral-direccional entre 6.4% y 17.5%, y diferencias relativas en frecuencia y amplitud de oscilación entre 8.0% y 21.1% respecto a los datos de vuelo.
Se concluye que el modelo CFD-6DOF desarrollado queda validado, representando correctamente la física dominante del problema y las tendencias generales de la dinámica lateral-direccional del UAV, con discrepancias reducidas y dentro de márgenes de tolerancia aceptables. Este resultado verifica la hipótesis planteada, estableciendo una base metodológica viable para el análisis de estabilidad de UAV de ala volante en etapa de diseño, reduciendo la dependencia de ensayos físicos costosos.
Flying-wing unmanned aerial vehicles (UAVs) offer significant aerodynamic advantages over conventional configurations, but the absence of an empennage compromises their lateral-directional stability, particularly regarding the Dutch Roll dynamic mode, which is characterized by near-zero damping. In response to this problem, the objective of this study was to develop an experimentally validated CFD-6DOF model to analyze the lateral-directional dynamic stability of a flying-wing UAV. For this purpose, a Skywalker X8 was used as the reference platform, instrumented with a Pixhawk 6C autopilot, a GPS, and an airspeed sensor, among others. Flight tests were conducted applying an impulsive perturbation to the ailerons, recording the aircraft's transient response for subsequent characterization. In parallel, a steady-state CFD model was implemented in ANSYS Fluent using the 𝑘 − 𝜔 𝑆𝑆𝑇 turbulence model, which was validated against aerodynamic coefficients reported in the specialized literature. Based on this model, the equivalent loads generated by elevon deflection were characterized and incorporated through a User-Defined Function (UDF) into a dynamic six-degrees-of-freedom (6DOF) model with a hybrid dynamic mesh, enabling simulation of the aircraft's response to the same perturbation applied in flight. Static validation showed satisfactory agreement for most aerodynamic coefficients, with normalized RMSE errors equal to or below 11.9%, although larger discrepancies were identified in the roll and yaw moment coefficients, of up to 20.4%, attributable to uncertainties in the conversion of moments with respect to the aerodynamic reference point used in the literature. The dynamic model successfully reproduced the Roll and Spiral modes, as well as the characteristic parameters of the oscillatory Dutch Roll mode, with normalized RMSE errors of the lateral-directional attitude variables between 6.4% and 17.5%, and relative differences in oscillation frequency and amplitude between 8.0% and 21.1% compared to flight data. It is concluded that the developed CFD-6DOF model is validated, correctly representing the dominant physics of the problem and the general trends of the UAV's lateral-directional dynamics, with reduced discrepancies within acceptable tolerance margins. This result verifies the proposed hypothesis, establishing a viable methodological basis for the stability analysis of flying-wing UAVs during the design stage, reducing reliance on costly physical testing.
Flying-wing unmanned aerial vehicles (UAVs) offer significant aerodynamic advantages over conventional configurations, but the absence of an empennage compromises their lateral-directional stability, particularly regarding the Dutch Roll dynamic mode, which is characterized by near-zero damping. In response to this problem, the objective of this study was to develop an experimentally validated CFD-6DOF model to analyze the lateral-directional dynamic stability of a flying-wing UAV. For this purpose, a Skywalker X8 was used as the reference platform, instrumented with a Pixhawk 6C autopilot, a GPS, and an airspeed sensor, among others. Flight tests were conducted applying an impulsive perturbation to the ailerons, recording the aircraft's transient response for subsequent characterization. In parallel, a steady-state CFD model was implemented in ANSYS Fluent using the 𝑘 − 𝜔 𝑆𝑆𝑇 turbulence model, which was validated against aerodynamic coefficients reported in the specialized literature. Based on this model, the equivalent loads generated by elevon deflection were characterized and incorporated through a User-Defined Function (UDF) into a dynamic six-degrees-of-freedom (6DOF) model with a hybrid dynamic mesh, enabling simulation of the aircraft's response to the same perturbation applied in flight. Static validation showed satisfactory agreement for most aerodynamic coefficients, with normalized RMSE errors equal to or below 11.9%, although larger discrepancies were identified in the roll and yaw moment coefficients, of up to 20.4%, attributable to uncertainties in the conversion of moments with respect to the aerodynamic reference point used in the literature. The dynamic model successfully reproduced the Roll and Spiral modes, as well as the characteristic parameters of the oscillatory Dutch Roll mode, with normalized RMSE errors of the lateral-directional attitude variables between 6.4% and 17.5%, and relative differences in oscillation frequency and amplitude between 8.0% and 21.1% compared to flight data. It is concluded that the developed CFD-6DOF model is validated, correctly representing the dominant physics of the problem and the general trends of the UAV's lateral-directional dynamics, with reduced discrepancies within acceptable tolerance margins. This result verifies the proposed hypothesis, establishing a viable methodological basis for the stability analysis of flying-wing UAVs during the design stage, reducing reliance on costly physical testing.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Aeroespacial.
Keywords
Dinámica de fluidos computacional, Aeronaves Pruebas, Drones, Aerodinámica