Stability and Control Analysis for an Unmanned Aircraft Configuration Using System-Identification Techniques

Computing Stability Derivatives and Their Gradients for Aerodynamic Shape Optimization

“…Note also that the steady and time-spectral results for our flow solver coincide, validating the use of the leading coefficient from the linear regression to predict the coefficients. Figure 8 shows a comparison of the C L derivatives with experimental results from Rohlf et al [55,41]. C Lα matches well, with the time-spectral stability derivative method predicting a value within approximately 2% of the experimental result at 5 degrees.…”

“…Murman observes that the ability to simulate the forced oscillation with a single frequency arises because the forcing frequency of the oscillation largely determines the frequency of the output and because there is a limited potential for energy transfer between the modes when using the Euler equations. We also note that, as demonstrated by Rohlf et al [41], smaller amplitude oscillations tend to produce more sinusoidal, if also noisier, output in the experiments. Since noise is not a major issue for time-periodic CFD solutions, we have chosen to use small-amplitude oscillations in this work, on the order of 0.5 to 2 degrees.…”

“…Using the 1 482 752 cell mesh and a RANS flow solver, we ran a series of comparisons with derivatives computed from experimental data by Rohlf et al [55,41] and CFD results from Le Roy and Morgand [23]. The goal of this comparison is to place the results from the time-spectral computation in context with physical results, thereby validating the method we propose for computing the stability derivatives.…”

“…These two components are motivated by the complex numberα derivative methodology outlined by Etkin [40] and are based on the idea that the transient derivatives are related to the time lag in the development of the aerodynamic forces when an aircraft changes orientation. A similar linear regression method for determining the stability derivatives is used with experimental flow results by Rohlf et al [41].…”

Multi-Objective Optimization for Wind Estimation and Aircraft Model Identification

“…In system theory, this process is known as system identification and the objective is to obtain dynamic models from observed input and output signals [3]. In particular, system identification methods have been used for flight-test evaluations [4][5][6][7][8][9][10], control analysis and design [11,12] and advanced simulation [13][14][15].…”